blob: 44fe1d28a16327d0cb04771f05a4df4f7f175e0d [file] [log] [blame]
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +02001Runtime Power Management Framework for I/O Devices
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +02002
Rafael J. Wysocki9659cc02011-02-18 23:20:21 +01003(C) 2009-2011 Rafael J. Wysocki <rjw@sisk.pl>, Novell Inc.
Alan Stern7490e442010-09-25 23:35:15 +02004(C) 2010 Alan Stern <stern@rowland.harvard.edu>
Rafael J. Wysockif71495f2014-05-16 02:47:37 +02005(C) 2014 Intel Corp., Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +02006
71. Introduction
8
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +02009Support for runtime power management (runtime PM) of I/O devices is provided
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +020010at the power management core (PM core) level by means of:
11
12* The power management workqueue pm_wq in which bus types and device drivers can
13 put their PM-related work items. It is strongly recommended that pm_wq be
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +020014 used for queuing all work items related to runtime PM, because this allows
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +020015 them to be synchronized with system-wide power transitions (suspend to RAM,
16 hibernation and resume from system sleep states). pm_wq is declared in
17 include/linux/pm_runtime.h and defined in kernel/power/main.c.
18
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +020019* A number of runtime PM fields in the 'power' member of 'struct device' (which
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +020020 is of the type 'struct dev_pm_info', defined in include/linux/pm.h) that can
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +020021 be used for synchronizing runtime PM operations with one another.
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +020022
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +020023* Three device runtime PM callbacks in 'struct dev_pm_ops' (defined in
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +020024 include/linux/pm.h).
25
26* A set of helper functions defined in drivers/base/power/runtime.c that can be
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +020027 used for carrying out runtime PM operations in such a way that the
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +020028 synchronization between them is taken care of by the PM core. Bus types and
29 device drivers are encouraged to use these functions.
30
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +020031The runtime PM callbacks present in 'struct dev_pm_ops', the device runtime PM
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +020032fields of 'struct dev_pm_info' and the core helper functions provided for
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +020033runtime PM are described below.
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +020034
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200352. Device Runtime PM Callbacks
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +020036
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +020037There are three device runtime PM callbacks defined in 'struct dev_pm_ops':
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +020038
39struct dev_pm_ops {
40 ...
41 int (*runtime_suspend)(struct device *dev);
42 int (*runtime_resume)(struct device *dev);
Rafael J. Wysockie1b19032009-12-03 21:04:08 +010043 int (*runtime_idle)(struct device *dev);
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +020044 ...
45};
46
Ming Lei2fb242a2011-10-09 11:40:25 +080047The ->runtime_suspend(), ->runtime_resume() and ->runtime_idle() callbacks
Rafael J. Wysocki5841eb62011-11-23 21:18:39 +010048are executed by the PM core for the device's subsystem that may be either of
49the following:
50
51 1. PM domain of the device, if the device's PM domain object, dev->pm_domain,
52 is present.
53
54 2. Device type of the device, if both dev->type and dev->type->pm are present.
55
56 3. Device class of the device, if both dev->class and dev->class->pm are
57 present.
58
59 4. Bus type of the device, if both dev->bus and dev->bus->pm are present.
60
Rafael J. Wysocki35cd1332011-12-18 00:34:13 +010061If the subsystem chosen by applying the above rules doesn't provide the relevant
62callback, the PM core will invoke the corresponding driver callback stored in
63dev->driver->pm directly (if present).
64
Rafael J. Wysocki5841eb62011-11-23 21:18:39 +010065The PM core always checks which callback to use in the order given above, so the
66priority order of callbacks from high to low is: PM domain, device type, class
67and bus type. Moreover, the high-priority one will always take precedence over
68a low-priority one. The PM domain, bus type, device type and class callbacks
69are referred to as subsystem-level callbacks in what follows.
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +020070
Alan Sternc7b61de2010-12-01 00:14:42 +010071By default, the callbacks are always invoked in process context with interrupts
Rafael J. Wysocki35cd1332011-12-18 00:34:13 +010072enabled. However, the pm_runtime_irq_safe() helper function can be used to tell
73the PM core that it is safe to run the ->runtime_suspend(), ->runtime_resume()
74and ->runtime_idle() callbacks for the given device in atomic context with
75interrupts disabled. This implies that the callback routines in question must
76not block or sleep, but it also means that the synchronous helper functions
77listed at the end of Section 4 may be used for that device within an interrupt
78handler or generally in an atomic context.
Alan Sternc7b61de2010-12-01 00:14:42 +010079
Rafael J. Wysocki35cd1332011-12-18 00:34:13 +010080The subsystem-level suspend callback, if present, is _entirely_ _responsible_
81for handling the suspend of the device as appropriate, which may, but need not
82include executing the device driver's own ->runtime_suspend() callback (from the
Rafael J. Wysockia6ab7aa2009-12-22 20:43:17 +010083PM core's point of view it is not necessary to implement a ->runtime_suspend()
84callback in a device driver as long as the subsystem-level suspend callback
85knows what to do to handle the device).
86
Rafael J. Wysocki35cd1332011-12-18 00:34:13 +010087 * Once the subsystem-level suspend callback (or the driver suspend callback,
88 if invoked directly) has completed successfully for the given device, the PM
89 core regards the device as suspended, which need not mean that it has been
90 put into a low power state. It is supposed to mean, however, that the
91 device will not process data and will not communicate with the CPU(s) and
92 RAM until the appropriate resume callback is executed for it. The runtime
93 PM status of a device after successful execution of the suspend callback is
94 'suspended'.
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +020095
Rafael J. Wysocki35cd1332011-12-18 00:34:13 +010096 * If the suspend callback returns -EBUSY or -EAGAIN, the device's runtime PM
97 status remains 'active', which means that the device _must_ be fully
98 operational afterwards.
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +020099
Rafael J. Wysocki35cd1332011-12-18 00:34:13 +0100100 * If the suspend callback returns an error code different from -EBUSY and
101 -EAGAIN, the PM core regards this as a fatal error and will refuse to run
102 the helper functions described in Section 4 for the device until its status
103 is directly set to either'active', or 'suspended' (the PM core provides
104 special helper functions for this purpose).
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200105
Rafael J. Wysocki35cd1332011-12-18 00:34:13 +0100106In particular, if the driver requires remote wakeup capability (i.e. hardware
Rafael J. Wysockia6ab7aa2009-12-22 20:43:17 +0100107mechanism allowing the device to request a change of its power state, such as
108PCI PME) for proper functioning and device_run_wake() returns 'false' for the
109device, then ->runtime_suspend() should return -EBUSY. On the other hand, if
Rafael J. Wysocki35cd1332011-12-18 00:34:13 +0100110device_run_wake() returns 'true' for the device and the device is put into a
111low-power state during the execution of the suspend callback, it is expected
112that remote wakeup will be enabled for the device. Generally, remote wakeup
113should be enabled for all input devices put into low-power states at run time.
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200114
Rafael J. Wysocki35cd1332011-12-18 00:34:13 +0100115The subsystem-level resume callback, if present, is _entirely_ _responsible_ for
116handling the resume of the device as appropriate, which may, but need not
117include executing the device driver's own ->runtime_resume() callback (from the
118PM core's point of view it is not necessary to implement a ->runtime_resume()
119callback in a device driver as long as the subsystem-level resume callback knows
120what to do to handle the device).
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200121
Rafael J. Wysocki35cd1332011-12-18 00:34:13 +0100122 * Once the subsystem-level resume callback (or the driver resume callback, if
123 invoked directly) has completed successfully, the PM core regards the device
124 as fully operational, which means that the device _must_ be able to complete
125 I/O operations as needed. The runtime PM status of the device is then
126 'active'.
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200127
Rafael J. Wysocki35cd1332011-12-18 00:34:13 +0100128 * If the resume callback returns an error code, the PM core regards this as a
129 fatal error and will refuse to run the helper functions described in Section
130 4 for the device, until its status is directly set to either 'active', or
131 'suspended' (by means of special helper functions provided by the PM core
132 for this purpose).
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200133
Rafael J. Wysocki35cd1332011-12-18 00:34:13 +0100134The idle callback (a subsystem-level one, if present, or the driver one) is
135executed by the PM core whenever the device appears to be idle, which is
136indicated to the PM core by two counters, the device's usage counter and the
137counter of 'active' children of the device.
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200138
139 * If any of these counters is decreased using a helper function provided by
140 the PM core and it turns out to be equal to zero, the other counter is
141 checked. If that counter also is equal to zero, the PM core executes the
Rafael J. Wysocki35cd1332011-12-18 00:34:13 +0100142 idle callback with the device as its argument.
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200143
Rafael J. Wysocki35cd1332011-12-18 00:34:13 +0100144The action performed by the idle callback is totally dependent on the subsystem
145(or driver) in question, but the expected and recommended action is to check
Rafael J. Wysockia6ab7aa2009-12-22 20:43:17 +0100146if the device can be suspended (i.e. if all of the conditions necessary for
147suspending the device are satisfied) and to queue up a suspend request for the
Alan Stern43d51af2013-06-03 21:49:59 +0200148device in that case. If there is no idle callback, or if the callback returns
Ulf Hanssond66e6db2013-10-15 22:25:08 +02001490, then the PM core will attempt to carry out a runtime suspend of the device,
150also respecting devices configured for autosuspend. In essence this means a
151call to pm_runtime_autosuspend() (do note that drivers needs to update the
152device last busy mark, pm_runtime_mark_last_busy(), to control the delay under
153this circumstance). To prevent this (for example, if the callback routine has
154started a delayed suspend), the routine must return a non-zero value. Negative
155error return codes are ignored by the PM core.
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200156
157The helper functions provided by the PM core, described in Section 4, guarantee
Rafael J. Wysocki35cd1332011-12-18 00:34:13 +0100158that the following constraints are met with respect to runtime PM callbacks for
159one device:
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200160
161(1) The callbacks are mutually exclusive (e.g. it is forbidden to execute
162 ->runtime_suspend() in parallel with ->runtime_resume() or with another
163 instance of ->runtime_suspend() for the same device) with the exception that
164 ->runtime_suspend() or ->runtime_resume() can be executed in parallel with
165 ->runtime_idle() (although ->runtime_idle() will not be started while any
166 of the other callbacks is being executed for the same device).
167
168(2) ->runtime_idle() and ->runtime_suspend() can only be executed for 'active'
169 devices (i.e. the PM core will only execute ->runtime_idle() or
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200170 ->runtime_suspend() for the devices the runtime PM status of which is
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200171 'active').
172
173(3) ->runtime_idle() and ->runtime_suspend() can only be executed for a device
174 the usage counter of which is equal to zero _and_ either the counter of
175 'active' children of which is equal to zero, or the 'power.ignore_children'
176 flag of which is set.
177
178(4) ->runtime_resume() can only be executed for 'suspended' devices (i.e. the
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200179 PM core will only execute ->runtime_resume() for the devices the runtime
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200180 PM status of which is 'suspended').
181
182Additionally, the helper functions provided by the PM core obey the following
183rules:
184
185 * If ->runtime_suspend() is about to be executed or there's a pending request
186 to execute it, ->runtime_idle() will not be executed for the same device.
187
188 * A request to execute or to schedule the execution of ->runtime_suspend()
189 will cancel any pending requests to execute ->runtime_idle() for the same
190 device.
191
192 * If ->runtime_resume() is about to be executed or there's a pending request
193 to execute it, the other callbacks will not be executed for the same device.
194
195 * A request to execute ->runtime_resume() will cancel any pending or
Alan Stern15bcb91d2010-09-25 23:35:21 +0200196 scheduled requests to execute the other callbacks for the same device,
197 except for scheduled autosuspends.
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200198
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +02001993. Runtime PM Device Fields
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200200
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200201The following device runtime PM fields are present in 'struct dev_pm_info', as
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200202defined in include/linux/pm.h:
203
204 struct timer_list suspend_timer;
Alan Stern15bcb91d2010-09-25 23:35:21 +0200205 - timer used for scheduling (delayed) suspend and autosuspend requests
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200206
207 unsigned long timer_expires;
208 - timer expiration time, in jiffies (if this is different from zero, the
209 timer is running and will expire at that time, otherwise the timer is not
210 running)
211
212 struct work_struct work;
213 - work structure used for queuing up requests (i.e. work items in pm_wq)
214
215 wait_queue_head_t wait_queue;
216 - wait queue used if any of the helper functions needs to wait for another
217 one to complete
218
219 spinlock_t lock;
220 - lock used for synchronisation
221
222 atomic_t usage_count;
223 - the usage counter of the device
224
225 atomic_t child_count;
226 - the count of 'active' children of the device
227
228 unsigned int ignore_children;
229 - if set, the value of child_count is ignored (but still updated)
230
231 unsigned int disable_depth;
Masanari Iida1f999d12014-11-08 17:54:51 +0900232 - used for disabling the helper functions (they work normally if this is
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200233 equal to zero); the initial value of it is 1 (i.e. runtime PM is
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200234 initially disabled for all devices)
235
Geert Uytterhoevene1a29762014-03-24 21:31:27 +0100236 int runtime_error;
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200237 - if set, there was a fatal error (one of the callbacks returned error code
Masanari Iida1f999d12014-11-08 17:54:51 +0900238 as described in Section 2), so the helper functions will not work until
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200239 this flag is cleared; this is the error code returned by the failing
240 callback
241
242 unsigned int idle_notification;
243 - if set, ->runtime_idle() is being executed
244
245 unsigned int request_pending;
246 - if set, there's a pending request (i.e. a work item queued up into pm_wq)
247
248 enum rpm_request request;
249 - type of request that's pending (valid if request_pending is set)
250
251 unsigned int deferred_resume;
252 - set if ->runtime_resume() is about to be run while ->runtime_suspend() is
253 being executed for that device and it is not practical to wait for the
254 suspend to complete; means "start a resume as soon as you've suspended"
255
Rafael J. Wysocki7a1a8eb2009-12-03 21:19:18 +0100256 unsigned int run_wake;
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200257 - set if the device is capable of generating runtime wake-up events
Rafael J. Wysocki7a1a8eb2009-12-03 21:19:18 +0100258
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200259 enum rpm_status runtime_status;
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200260 - the runtime PM status of the device; this field's initial value is
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200261 RPM_SUSPENDED, which means that each device is initially regarded by the
262 PM core as 'suspended', regardless of its real hardware status
263
Rafael J. Wysocki87d1b3e2010-03-06 21:28:17 +0100264 unsigned int runtime_auto;
265 - if set, indicates that the user space has allowed the device driver to
266 power manage the device at run time via the /sys/devices/.../power/control
267 interface; it may only be modified with the help of the pm_runtime_allow()
268 and pm_runtime_forbid() helper functions
269
Alan Stern7490e442010-09-25 23:35:15 +0200270 unsigned int no_callbacks;
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200271 - indicates that the device does not use the runtime PM callbacks (see
Alan Stern7490e442010-09-25 23:35:15 +0200272 Section 8); it may be modified only by the pm_runtime_no_callbacks()
273 helper function
274
Alan Sternc7b61de2010-12-01 00:14:42 +0100275 unsigned int irq_safe;
276 - indicates that the ->runtime_suspend() and ->runtime_resume() callbacks
277 will be invoked with the spinlock held and interrupts disabled
278
Alan Stern15bcb91d2010-09-25 23:35:21 +0200279 unsigned int use_autosuspend;
280 - indicates that the device's driver supports delayed autosuspend (see
281 Section 9); it may be modified only by the
282 pm_runtime{_dont}_use_autosuspend() helper functions
283
284 unsigned int timer_autosuspends;
285 - indicates that the PM core should attempt to carry out an autosuspend
286 when the timer expires rather than a normal suspend
287
288 int autosuspend_delay;
289 - the delay time (in milliseconds) to be used for autosuspend
290
291 unsigned long last_busy;
292 - the time (in jiffies) when the pm_runtime_mark_last_busy() helper
293 function was last called for this device; used in calculating inactivity
294 periods for autosuspend
295
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200296All of the above fields are members of the 'power' member of 'struct device'.
297
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +02002984. Runtime PM Device Helper Functions
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200299
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200300The following runtime PM helper functions are defined in
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200301drivers/base/power/runtime.c and include/linux/pm_runtime.h:
302
303 void pm_runtime_init(struct device *dev);
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200304 - initialize the device runtime PM fields in 'struct dev_pm_info'
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200305
306 void pm_runtime_remove(struct device *dev);
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200307 - make sure that the runtime PM of the device will be disabled after
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200308 removing the device from device hierarchy
309
310 int pm_runtime_idle(struct device *dev);
Alan Stern43d51af2013-06-03 21:49:59 +0200311 - execute the subsystem-level idle callback for the device; returns an
312 error code on failure, where -EINPROGRESS means that ->runtime_idle() is
313 already being executed; if there is no callback or the callback returns 0
Ulf Hanssond66e6db2013-10-15 22:25:08 +0200314 then run pm_runtime_autosuspend(dev) and return its result
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200315
316 int pm_runtime_suspend(struct device *dev);
Rafael J. Wysockia6ab7aa2009-12-22 20:43:17 +0100317 - execute the subsystem-level suspend callback for the device; returns 0 on
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200318 success, 1 if the device's runtime PM status was already 'suspended', or
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200319 error code on failure, where -EAGAIN or -EBUSY means it is safe to attempt
Rafael J. Wysocki632e2702011-07-01 22:29:15 +0200320 to suspend the device again in future and -EACCES means that
321 'power.disable_depth' is different from 0
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200322
Alan Stern15bcb91d2010-09-25 23:35:21 +0200323 int pm_runtime_autosuspend(struct device *dev);
324 - same as pm_runtime_suspend() except that the autosuspend delay is taken
325 into account; if pm_runtime_autosuspend_expiration() says the delay has
326 not yet expired then an autosuspend is scheduled for the appropriate time
327 and 0 is returned
328
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200329 int pm_runtime_resume(struct device *dev);
Thadeu Lima de Souza Cascardode8164f2010-01-17 19:22:28 -0200330 - execute the subsystem-level resume callback for the device; returns 0 on
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200331 success, 1 if the device's runtime PM status was already 'active' or
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200332 error code on failure, where -EAGAIN means it may be safe to attempt to
333 resume the device again in future, but 'power.runtime_error' should be
Rafael J. Wysocki632e2702011-07-01 22:29:15 +0200334 checked additionally, and -EACCES means that 'power.disable_depth' is
335 different from 0
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200336
337 int pm_request_idle(struct device *dev);
Rafael J. Wysockia6ab7aa2009-12-22 20:43:17 +0100338 - submit a request to execute the subsystem-level idle callback for the
339 device (the request is represented by a work item in pm_wq); returns 0 on
340 success or error code if the request has not been queued up
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200341
Alan Stern15bcb91d2010-09-25 23:35:21 +0200342 int pm_request_autosuspend(struct device *dev);
343 - schedule the execution of the subsystem-level suspend callback for the
344 device when the autosuspend delay has expired; if the delay has already
345 expired then the work item is queued up immediately
346
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200347 int pm_schedule_suspend(struct device *dev, unsigned int delay);
Rafael J. Wysockia6ab7aa2009-12-22 20:43:17 +0100348 - schedule the execution of the subsystem-level suspend callback for the
349 device in future, where 'delay' is the time to wait before queuing up a
350 suspend work item in pm_wq, in milliseconds (if 'delay' is zero, the work
351 item is queued up immediately); returns 0 on success, 1 if the device's PM
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200352 runtime status was already 'suspended', or error code if the request
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200353 hasn't been scheduled (or queued up if 'delay' is 0); if the execution of
354 ->runtime_suspend() is already scheduled and not yet expired, the new
355 value of 'delay' will be used as the time to wait
356
357 int pm_request_resume(struct device *dev);
Rafael J. Wysockia6ab7aa2009-12-22 20:43:17 +0100358 - submit a request to execute the subsystem-level resume callback for the
359 device (the request is represented by a work item in pm_wq); returns 0 on
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200360 success, 1 if the device's runtime PM status was already 'active', or
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200361 error code if the request hasn't been queued up
362
363 void pm_runtime_get_noresume(struct device *dev);
364 - increment the device's usage counter
365
366 int pm_runtime_get(struct device *dev);
367 - increment the device's usage counter, run pm_request_resume(dev) and
368 return its result
369
370 int pm_runtime_get_sync(struct device *dev);
371 - increment the device's usage counter, run pm_runtime_resume(dev) and
372 return its result
373
374 void pm_runtime_put_noidle(struct device *dev);
375 - decrement the device's usage counter
376
377 int pm_runtime_put(struct device *dev);
Alan Stern15bcb91d2010-09-25 23:35:21 +0200378 - decrement the device's usage counter; if the result is 0 then run
379 pm_request_idle(dev) and return its result
380
381 int pm_runtime_put_autosuspend(struct device *dev);
382 - decrement the device's usage counter; if the result is 0 then run
383 pm_request_autosuspend(dev) and return its result
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200384
385 int pm_runtime_put_sync(struct device *dev);
Alan Stern15bcb91d2010-09-25 23:35:21 +0200386 - decrement the device's usage counter; if the result is 0 then run
387 pm_runtime_idle(dev) and return its result
388
Alan Sternc7b61de2010-12-01 00:14:42 +0100389 int pm_runtime_put_sync_suspend(struct device *dev);
390 - decrement the device's usage counter; if the result is 0 then run
391 pm_runtime_suspend(dev) and return its result
392
Alan Stern15bcb91d2010-09-25 23:35:21 +0200393 int pm_runtime_put_sync_autosuspend(struct device *dev);
394 - decrement the device's usage counter; if the result is 0 then run
395 pm_runtime_autosuspend(dev) and return its result
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200396
397 void pm_runtime_enable(struct device *dev);
Rafael J. Wysockie358bad2011-07-06 10:52:06 +0200398 - decrement the device's 'power.disable_depth' field; if that field is equal
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200399 to zero, the runtime PM helper functions can execute subsystem-level
Rafael J. Wysockie358bad2011-07-06 10:52:06 +0200400 callbacks described in Section 2 for the device
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200401
402 int pm_runtime_disable(struct device *dev);
Rafael J. Wysockie358bad2011-07-06 10:52:06 +0200403 - increment the device's 'power.disable_depth' field (if the value of that
404 field was previously zero, this prevents subsystem-level runtime PM
Geert Uytterhoeven91e63cc2014-03-24 21:31:28 +0100405 callbacks from being run for the device), make sure that all of the
406 pending runtime PM operations on the device are either completed or
407 canceled; returns 1 if there was a resume request pending and it was
408 necessary to execute the subsystem-level resume callback for the device
409 to satisfy that request, otherwise 0 is returned
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200410
Rafael J. Wysockie358bad2011-07-06 10:52:06 +0200411 int pm_runtime_barrier(struct device *dev);
412 - check if there's a resume request pending for the device and resume it
413 (synchronously) in that case, cancel any other pending runtime PM requests
414 regarding it and wait for all runtime PM operations on it in progress to
415 complete; returns 1 if there was a resume request pending and it was
416 necessary to execute the subsystem-level resume callback for the device to
417 satisfy that request, otherwise 0 is returned
418
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200419 void pm_suspend_ignore_children(struct device *dev, bool enable);
420 - set/unset the power.ignore_children flag of the device
421
422 int pm_runtime_set_active(struct device *dev);
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200423 - clear the device's 'power.runtime_error' flag, set the device's runtime
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200424 PM status to 'active' and update its parent's counter of 'active'
425 children as appropriate (it is only valid to use this function if
426 'power.runtime_error' is set or 'power.disable_depth' is greater than
427 zero); it will fail and return error code if the device has a parent
428 which is not active and the 'power.ignore_children' flag of which is unset
429
430 void pm_runtime_set_suspended(struct device *dev);
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200431 - clear the device's 'power.runtime_error' flag, set the device's runtime
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200432 PM status to 'suspended' and update its parent's counter of 'active'
433 children as appropriate (it is only valid to use this function if
434 'power.runtime_error' is set or 'power.disable_depth' is greater than
435 zero)
436
ShuoX Liufbadc582013-01-23 21:49:37 +0100437 bool pm_runtime_active(struct device *dev);
438 - return true if the device's runtime PM status is 'active' or its
439 'power.disable_depth' field is not equal to zero, or false otherwise
440
Rafael J. Wysockid690b2c2010-03-06 21:28:37 +0100441 bool pm_runtime_suspended(struct device *dev);
Rafael J. Wysockif08f5a02010-12-16 17:11:58 +0100442 - return true if the device's runtime PM status is 'suspended' and its
443 'power.disable_depth' field is equal to zero, or false otherwise
Rafael J. Wysockid690b2c2010-03-06 21:28:37 +0100444
Kevin Hilmanf3393b62011-07-12 11:17:09 +0200445 bool pm_runtime_status_suspended(struct device *dev);
446 - return true if the device's runtime PM status is 'suspended'
447
Rafael J. Wysockif71495f2014-05-16 02:47:37 +0200448 bool pm_runtime_suspended_if_enabled(struct device *dev);
449 - return true if the device's runtime PM status is 'suspended' and its
450 'power.disable_depth' field is equal to 1
451
Rafael J. Wysocki87d1b3e2010-03-06 21:28:17 +0100452 void pm_runtime_allow(struct device *dev);
453 - set the power.runtime_auto flag for the device and decrease its usage
454 counter (used by the /sys/devices/.../power/control interface to
455 effectively allow the device to be power managed at run time)
456
457 void pm_runtime_forbid(struct device *dev);
458 - unset the power.runtime_auto flag for the device and increase its usage
459 counter (used by the /sys/devices/.../power/control interface to
460 effectively prevent the device from being power managed at run time)
461
Alan Stern7490e442010-09-25 23:35:15 +0200462 void pm_runtime_no_callbacks(struct device *dev);
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200463 - set the power.no_callbacks flag for the device and remove the runtime
Alan Stern7490e442010-09-25 23:35:15 +0200464 PM attributes from /sys/devices/.../power (or prevent them from being
465 added when the device is registered)
466
Alan Sternc7b61de2010-12-01 00:14:42 +0100467 void pm_runtime_irq_safe(struct device *dev);
468 - set the power.irq_safe flag for the device, causing the runtime-PM
Rafael J. Wysocki64584eb2011-08-25 15:31:05 +0200469 callbacks to be invoked with interrupts off
Alan Sternc7b61de2010-12-01 00:14:42 +0100470
Krzysztof Kozlowski3fb15812014-11-14 09:47:25 +0100471 bool pm_runtime_is_irq_safe(struct device *dev);
472 - return true if power.irq_safe flag was set for the device, causing
473 the runtime-PM callbacks to be invoked with interrupts off
474
Alan Stern15bcb91d2010-09-25 23:35:21 +0200475 void pm_runtime_mark_last_busy(struct device *dev);
476 - set the power.last_busy field to the current time
477
478 void pm_runtime_use_autosuspend(struct device *dev);
479 - set the power.use_autosuspend flag, enabling autosuspend delays
480
481 void pm_runtime_dont_use_autosuspend(struct device *dev);
482 - clear the power.use_autosuspend flag, disabling autosuspend delays
483
484 void pm_runtime_set_autosuspend_delay(struct device *dev, int delay);
485 - set the power.autosuspend_delay value to 'delay' (expressed in
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200486 milliseconds); if 'delay' is negative then runtime suspends are
Alan Stern15bcb91d2010-09-25 23:35:21 +0200487 prevented
488
489 unsigned long pm_runtime_autosuspend_expiration(struct device *dev);
490 - calculate the time when the current autosuspend delay period will expire,
491 based on power.last_busy and power.autosuspend_delay; if the delay time
492 is 1000 ms or larger then the expiration time is rounded up to the
493 nearest second; returns 0 if the delay period has already expired or
494 power.use_autosuspend isn't set, otherwise returns the expiration time
495 in jiffies
496
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200497It is safe to execute the following helper functions from interrupt context:
498
499pm_request_idle()
Alan Stern15bcb91d2010-09-25 23:35:21 +0200500pm_request_autosuspend()
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200501pm_schedule_suspend()
502pm_request_resume()
503pm_runtime_get_noresume()
504pm_runtime_get()
505pm_runtime_put_noidle()
506pm_runtime_put()
Alan Stern15bcb91d2010-09-25 23:35:21 +0200507pm_runtime_put_autosuspend()
508pm_runtime_enable()
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200509pm_suspend_ignore_children()
510pm_runtime_set_active()
511pm_runtime_set_suspended()
Alan Stern15bcb91d2010-09-25 23:35:21 +0200512pm_runtime_suspended()
513pm_runtime_mark_last_busy()
514pm_runtime_autosuspend_expiration()
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200515
Alan Sternc7b61de2010-12-01 00:14:42 +0100516If pm_runtime_irq_safe() has been called for a device then the following helper
517functions may also be used in interrupt context:
518
Ming Lei2e6ba512011-09-21 22:31:33 +0200519pm_runtime_idle()
Alan Sternc7b61de2010-12-01 00:14:42 +0100520pm_runtime_suspend()
521pm_runtime_autosuspend()
522pm_runtime_resume()
523pm_runtime_get_sync()
Kevin Hilman02b26772011-08-05 21:45:20 +0200524pm_runtime_put_sync()
Alan Sternc7b61de2010-12-01 00:14:42 +0100525pm_runtime_put_sync_suspend()
Colin Cross311aab72011-08-08 23:39:36 +0200526pm_runtime_put_sync_autosuspend()
Alan Sternc7b61de2010-12-01 00:14:42 +0100527
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +02005285. Runtime PM Initialization, Device Probing and Removal
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200529
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200530Initially, the runtime PM is disabled for all devices, which means that the
Masanari Iida1f999d12014-11-08 17:54:51 +0900531majority of the runtime PM helper functions described in Section 4 will return
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200532-EAGAIN until pm_runtime_enable() is called for the device.
533
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200534In addition to that, the initial runtime PM status of all devices is
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200535'suspended', but it need not reflect the actual physical state of the device.
536Thus, if the device is initially active (i.e. it is able to process I/O), its
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200537runtime PM status must be changed to 'active', with the help of
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200538pm_runtime_set_active(), before pm_runtime_enable() is called for the device.
539
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200540However, if the device has a parent and the parent's runtime PM is enabled,
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200541calling pm_runtime_set_active() for the device will affect the parent, unless
542the parent's 'power.ignore_children' flag is set. Namely, in that case the
543parent won't be able to suspend at run time, using the PM core's helper
544functions, as long as the child's status is 'active', even if the child's
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200545runtime PM is still disabled (i.e. pm_runtime_enable() hasn't been called for
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200546the child yet or pm_runtime_disable() has been called for it). For this reason,
547once pm_runtime_set_active() has been called for the device, pm_runtime_enable()
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200548should be called for it too as soon as reasonably possible or its runtime PM
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200549status should be changed back to 'suspended' with the help of
550pm_runtime_set_suspended().
551
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200552If the default initial runtime PM status of the device (i.e. 'suspended')
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200553reflects the actual state of the device, its bus type's or its driver's
554->probe() callback will likely need to wake it up using one of the PM core's
555helper functions described in Section 4. In that case, pm_runtime_resume()
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200556should be used. Of course, for this purpose the device's runtime PM has to be
Rafael J. Wysocki5e928f72009-08-18 23:38:32 +0200557enabled earlier by calling pm_runtime_enable().
558
Ulf Hanssonea309942013-11-08 06:13:51 +0100559It may be desirable to suspend the device once ->probe() has finished.
560Therefore the driver core uses the asyncronous pm_request_idle() to submit a
561request to execute the subsystem-level idle callback for the device at that
562time. A driver that makes use of the runtime autosuspend feature, may want to
563update the last busy mark before returning from ->probe().
Rafael J. Wysockif5da24d2011-07-02 14:27:11 +0200564
565Moreover, the driver core prevents runtime PM callbacks from racing with the bus
566notifier callback in __device_release_driver(), which is necessary, because the
567notifier is used by some subsystems to carry out operations affecting the
568runtime PM functionality. It does so by calling pm_runtime_get_sync() before
569driver_sysfs_remove() and the BUS_NOTIFY_UNBIND_DRIVER notifications. This
570resumes the device if it's in the suspended state and prevents it from
571being suspended again while those routines are being executed.
572
573To allow bus types and drivers to put devices into the suspended state by
574calling pm_runtime_suspend() from their ->remove() routines, the driver core
575executes pm_runtime_put_sync() after running the BUS_NOTIFY_UNBIND_DRIVER
576notifications in __device_release_driver(). This requires bus types and
577drivers to make their ->remove() callbacks avoid races with runtime PM directly,
578but also it allows of more flexibility in the handling of devices during the
579removal of their drivers.
Alan Sternf1212ae2009-12-22 20:43:40 +0100580
Rafael J. Wysocki87d1b3e2010-03-06 21:28:17 +0100581The user space can effectively disallow the driver of the device to power manage
582it at run time by changing the value of its /sys/devices/.../power/control
583attribute to "on", which causes pm_runtime_forbid() to be called. In principle,
584this mechanism may also be used by the driver to effectively turn off the
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200585runtime power management of the device until the user space turns it on.
586Namely, during the initialization the driver can make sure that the runtime PM
Rafael J. Wysocki87d1b3e2010-03-06 21:28:17 +0100587status of the device is 'active' and call pm_runtime_forbid(). It should be
588noted, however, that if the user space has already intentionally changed the
589value of /sys/devices/.../power/control to "auto" to allow the driver to power
590manage the device at run time, the driver may confuse it by using
591pm_runtime_forbid() this way.
592
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +02005936. Runtime PM and System Sleep
Alan Sternf1212ae2009-12-22 20:43:40 +0100594
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200595Runtime PM and system sleep (i.e., system suspend and hibernation, also known
Alan Sternf1212ae2009-12-22 20:43:40 +0100596as suspend-to-RAM and suspend-to-disk) interact with each other in a couple of
597ways. If a device is active when a system sleep starts, everything is
598straightforward. But what should happen if the device is already suspended?
599
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200600The device may have different wake-up settings for runtime PM and system sleep.
601For example, remote wake-up may be enabled for runtime suspend but disallowed
Alan Sternf1212ae2009-12-22 20:43:40 +0100602for system sleep (device_may_wakeup(dev) returns 'false'). When this happens,
603the subsystem-level system suspend callback is responsible for changing the
604device's wake-up setting (it may leave that to the device driver's system
605suspend routine). It may be necessary to resume the device and suspend it again
606in order to do so. The same is true if the driver uses different power levels
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200607or other settings for runtime suspend and system sleep.
Alan Sternf1212ae2009-12-22 20:43:40 +0100608
Rafael J. Wysocki455716e2011-07-01 22:29:05 +0200609During system resume, the simplest approach is to bring all devices back to full
610power, even if they had been suspended before the system suspend began. There
611are several reasons for this, including:
Alan Sternf1212ae2009-12-22 20:43:40 +0100612
613 * The device might need to switch power levels, wake-up settings, etc.
614
615 * Remote wake-up events might have been lost by the firmware.
616
617 * The device's children may need the device to be at full power in order
618 to resume themselves.
619
620 * The driver's idea of the device state may not agree with the device's
621 physical state. This can happen during resume from hibernation.
622
623 * The device might need to be reset.
624
625 * Even though the device was suspended, if its usage counter was > 0 then most
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200626 likely it would need a runtime resume in the near future anyway.
Alan Sternf1212ae2009-12-22 20:43:40 +0100627
Rafael J. Wysocki455716e2011-07-01 22:29:05 +0200628If the device had been suspended before the system suspend began and it's
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200629brought back to full power during resume, then its runtime PM status will have
Rafael J. Wysocki455716e2011-07-01 22:29:05 +0200630to be updated to reflect the actual post-system sleep status. The way to do
631this is:
Alan Sternf1212ae2009-12-22 20:43:40 +0100632
633 pm_runtime_disable(dev);
634 pm_runtime_set_active(dev);
635 pm_runtime_enable(dev);
636
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200637The PM core always increments the runtime usage counter before calling the
Rafael J. Wysocki1e2ef052011-07-06 10:51:58 +0200638->suspend() callback and decrements it after calling the ->resume() callback.
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200639Hence disabling runtime PM temporarily like this will not cause any runtime
Rafael J. Wysocki1e2ef052011-07-06 10:51:58 +0200640suspend attempts to be permanently lost. If the usage count goes to zero
641following the return of the ->resume() callback, the ->runtime_idle() callback
642will be invoked as usual.
643
Rafael J. Wysocki455716e2011-07-01 22:29:05 +0200644On some systems, however, system sleep is not entered through a global firmware
645or hardware operation. Instead, all hardware components are put into low-power
646states directly by the kernel in a coordinated way. Then, the system sleep
647state effectively follows from the states the hardware components end up in
648and the system is woken up from that state by a hardware interrupt or a similar
649mechanism entirely under the kernel's control. As a result, the kernel never
650gives control away and the states of all devices during resume are precisely
651known to it. If that is the case and none of the situations listed above takes
652place (in particular, if the system is not waking up from hibernation), it may
653be more efficient to leave the devices that had been suspended before the system
654suspend began in the suspended state.
655
Rafael J. Wysockif71495f2014-05-16 02:47:37 +0200656To this end, the PM core provides a mechanism allowing some coordination between
657different levels of device hierarchy. Namely, if a system suspend .prepare()
658callback returns a positive number for a device, that indicates to the PM core
659that the device appears to be runtime-suspended and its state is fine, so it
660may be left in runtime suspend provided that all of its descendants are also
661left in runtime suspend. If that happens, the PM core will not execute any
662system suspend and resume callbacks for all of those devices, except for the
663complete callback, which is then entirely responsible for handling the device
664as appropriate. This only applies to system suspend transitions that are not
665related to hibernation (see Documentation/power/devices.txt for more
666information).
667
Rafael J. Wysocki1e2ef052011-07-06 10:51:58 +0200668The PM core does its best to reduce the probability of race conditions between
669the runtime PM and system suspend/resume (and hibernation) callbacks by carrying
670out the following operations:
671
Rafael J. Wysocki4ec6a9c2014-05-16 13:05:59 +0200672 * During system suspend pm_runtime_get_noresume() is called for every device
673 right before executing the subsystem-level .prepare() callback for it and
674 pm_runtime_barrier() is called for every device right before executing the
675 subsystem-level .suspend() callback for it. In addition to that the PM core
676 calls __pm_runtime_disable() with 'false' as the second argument for every
677 device right before executing the subsystem-level .suspend_late() callback
678 for it.
Rafael J. Wysocki1e2ef052011-07-06 10:51:58 +0200679
Rafael J. Wysocki4ec6a9c2014-05-16 13:05:59 +0200680 * During system resume pm_runtime_enable() and pm_runtime_put() are called for
681 every device right after executing the subsystem-level .resume_early()
682 callback and right after executing the subsystem-level .complete() callback
Rafael J. Wysocki9f6d8f62012-12-22 23:59:01 +0100683 for it, respectively.
Rafael J. Wysocki1e2ef052011-07-06 10:51:58 +0200684
Rafael J. Wysockid690b2c2010-03-06 21:28:37 +01006857. Generic subsystem callbacks
686
687Subsystems may wish to conserve code space by using the set of generic power
688management callbacks provided by the PM core, defined in
689driver/base/power/generic_ops.c:
690
Rafael J. Wysockid690b2c2010-03-06 21:28:37 +0100691 int pm_generic_runtime_suspend(struct device *dev);
692 - invoke the ->runtime_suspend() callback provided by the driver of this
Geert Uytterhoeven39c29f32014-03-24 21:31:29 +0100693 device and return its result, or return 0 if not defined
Rafael J. Wysockid690b2c2010-03-06 21:28:37 +0100694
695 int pm_generic_runtime_resume(struct device *dev);
696 - invoke the ->runtime_resume() callback provided by the driver of this
Geert Uytterhoeven39c29f32014-03-24 21:31:29 +0100697 device and return its result, or return 0 if not defined
Rafael J. Wysockid690b2c2010-03-06 21:28:37 +0100698
699 int pm_generic_suspend(struct device *dev);
700 - if the device has not been suspended at run time, invoke the ->suspend()
701 callback provided by its driver and return its result, or return 0 if not
702 defined
703
Rafael J. Wysockie5291922011-07-01 22:12:59 +0200704 int pm_generic_suspend_noirq(struct device *dev);
705 - if pm_runtime_suspended(dev) returns "false", invoke the ->suspend_noirq()
706 callback provided by the device's driver and return its result, or return
707 0 if not defined
708
Rafael J. Wysockid690b2c2010-03-06 21:28:37 +0100709 int pm_generic_resume(struct device *dev);
710 - invoke the ->resume() callback provided by the driver of this device and,
711 if successful, change the device's runtime PM status to 'active'
712
Rafael J. Wysockie5291922011-07-01 22:12:59 +0200713 int pm_generic_resume_noirq(struct device *dev);
714 - invoke the ->resume_noirq() callback provided by the driver of this device
715
Rafael J. Wysockid690b2c2010-03-06 21:28:37 +0100716 int pm_generic_freeze(struct device *dev);
717 - if the device has not been suspended at run time, invoke the ->freeze()
718 callback provided by its driver and return its result, or return 0 if not
719 defined
720
Rafael J. Wysockie5291922011-07-01 22:12:59 +0200721 int pm_generic_freeze_noirq(struct device *dev);
722 - if pm_runtime_suspended(dev) returns "false", invoke the ->freeze_noirq()
723 callback provided by the device's driver and return its result, or return
724 0 if not defined
725
Rafael J. Wysockid690b2c2010-03-06 21:28:37 +0100726 int pm_generic_thaw(struct device *dev);
727 - if the device has not been suspended at run time, invoke the ->thaw()
728 callback provided by its driver and return its result, or return 0 if not
729 defined
730
Rafael J. Wysockie5291922011-07-01 22:12:59 +0200731 int pm_generic_thaw_noirq(struct device *dev);
732 - if pm_runtime_suspended(dev) returns "false", invoke the ->thaw_noirq()
733 callback provided by the device's driver and return its result, or return
734 0 if not defined
735
Rafael J. Wysockid690b2c2010-03-06 21:28:37 +0100736 int pm_generic_poweroff(struct device *dev);
737 - if the device has not been suspended at run time, invoke the ->poweroff()
738 callback provided by its driver and return its result, or return 0 if not
739 defined
740
Rafael J. Wysockie5291922011-07-01 22:12:59 +0200741 int pm_generic_poweroff_noirq(struct device *dev);
742 - if pm_runtime_suspended(dev) returns "false", run the ->poweroff_noirq()
743 callback provided by the device's driver and return its result, or return
744 0 if not defined
745
Rafael J. Wysockid690b2c2010-03-06 21:28:37 +0100746 int pm_generic_restore(struct device *dev);
747 - invoke the ->restore() callback provided by the driver of this device and,
748 if successful, change the device's runtime PM status to 'active'
749
Rafael J. Wysockie5291922011-07-01 22:12:59 +0200750 int pm_generic_restore_noirq(struct device *dev);
751 - invoke the ->restore_noirq() callback provided by the device's driver
752
Geert Uytterhoevenfd6fe822014-03-24 21:31:30 +0100753These functions are the defaults used by the PM core, if a subsystem doesn't
754provide its own callbacks for ->runtime_idle(), ->runtime_suspend(),
Rafael J. Wysockie5291922011-07-01 22:12:59 +0200755->runtime_resume(), ->suspend(), ->suspend_noirq(), ->resume(),
756->resume_noirq(), ->freeze(), ->freeze_noirq(), ->thaw(), ->thaw_noirq(),
Geert Uytterhoevenfd6fe822014-03-24 21:31:30 +0100757->poweroff(), ->poweroff_noirq(), ->restore(), ->restore_noirq() in the
758subsystem-level dev_pm_ops structure.
Rafael J. Wysockid690b2c2010-03-06 21:28:37 +0100759
760Device drivers that wish to use the same function as a system suspend, freeze,
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200761poweroff and runtime suspend callback, and similarly for system resume, thaw,
762restore, and runtime resume, can achieve this with the help of the
Rafael J. Wysockid690b2c2010-03-06 21:28:37 +0100763UNIVERSAL_DEV_PM_OPS macro defined in include/linux/pm.h (possibly setting its
764last argument to NULL).
Alan Stern7490e442010-09-25 23:35:15 +0200765
7668. "No-Callback" Devices
767
768Some "devices" are only logical sub-devices of their parent and cannot be
769power-managed on their own. (The prototype example is a USB interface. Entire
770USB devices can go into low-power mode or send wake-up requests, but neither is
771possible for individual interfaces.) The drivers for these devices have no
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200772need of runtime PM callbacks; if the callbacks did exist, ->runtime_suspend()
Alan Stern7490e442010-09-25 23:35:15 +0200773and ->runtime_resume() would always return 0 without doing anything else and
774->runtime_idle() would always call pm_runtime_suspend().
775
776Subsystems can tell the PM core about these devices by calling
777pm_runtime_no_callbacks(). This should be done after the device structure is
778initialized and before it is registered (although after device registration is
779also okay). The routine will set the device's power.no_callbacks flag and
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200780prevent the non-debugging runtime PM sysfs attributes from being created.
Alan Stern7490e442010-09-25 23:35:15 +0200781
782When power.no_callbacks is set, the PM core will not invoke the
783->runtime_idle(), ->runtime_suspend(), or ->runtime_resume() callbacks.
784Instead it will assume that suspends and resumes always succeed and that idle
785devices should be suspended.
786
787As a consequence, the PM core will never directly inform the device's subsystem
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200788or driver about runtime power changes. Instead, the driver for the device's
Alan Stern7490e442010-09-25 23:35:15 +0200789parent must take responsibility for telling the device's driver when the
790parent's power state changes.
Alan Stern15bcb91d2010-09-25 23:35:21 +0200791
7929. Autosuspend, or automatically-delayed suspends
793
794Changing a device's power state isn't free; it requires both time and energy.
795A device should be put in a low-power state only when there's some reason to
796think it will remain in that state for a substantial time. A common heuristic
797says that a device which hasn't been used for a while is liable to remain
798unused; following this advice, drivers should not allow devices to be suspended
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200799at runtime until they have been inactive for some minimum period. Even when
Alan Stern15bcb91d2010-09-25 23:35:21 +0200800the heuristic ends up being non-optimal, it will still prevent devices from
801"bouncing" too rapidly between low-power and full-power states.
802
803The term "autosuspend" is an historical remnant. It doesn't mean that the
804device is automatically suspended (the subsystem or driver still has to call
Rafael J. Wysocki62052ab2011-07-06 10:52:13 +0200805the appropriate PM routines); rather it means that runtime suspends will
Alan Stern15bcb91d2010-09-25 23:35:21 +0200806automatically be delayed until the desired period of inactivity has elapsed.
807
808Inactivity is determined based on the power.last_busy field. Drivers should
809call pm_runtime_mark_last_busy() to update this field after carrying out I/O,
810typically just before calling pm_runtime_put_autosuspend(). The desired length
811of the inactivity period is a matter of policy. Subsystems can set this length
812initially by calling pm_runtime_set_autosuspend_delay(), but after device
813registration the length should be controlled by user space, using the
814/sys/devices/.../power/autosuspend_delay_ms attribute.
815
816In order to use autosuspend, subsystems or drivers must call
817pm_runtime_use_autosuspend() (preferably before registering the device), and
818thereafter they should use the various *_autosuspend() helper functions instead
819of the non-autosuspend counterparts:
820
821 Instead of: pm_runtime_suspend use: pm_runtime_autosuspend;
822 Instead of: pm_schedule_suspend use: pm_request_autosuspend;
823 Instead of: pm_runtime_put use: pm_runtime_put_autosuspend;
824 Instead of: pm_runtime_put_sync use: pm_runtime_put_sync_autosuspend.
825
826Drivers may also continue to use the non-autosuspend helper functions; they
827will behave normally, not taking the autosuspend delay into account.
828Similarly, if the power.use_autosuspend field isn't set then the autosuspend
829helper functions will behave just like the non-autosuspend counterparts.
830
Alan Stern886486b2011-11-03 23:39:18 +0100831Under some circumstances a driver or subsystem may want to prevent a device
832from autosuspending immediately, even though the usage counter is zero and the
833autosuspend delay time has expired. If the ->runtime_suspend() callback
834returns -EAGAIN or -EBUSY, and if the next autosuspend delay expiration time is
835in the future (as it normally would be if the callback invoked
836pm_runtime_mark_last_busy()), the PM core will automatically reschedule the
837autosuspend. The ->runtime_suspend() callback can't do this rescheduling
838itself because no suspend requests of any kind are accepted while the device is
839suspending (i.e., while the callback is running).
840
Alan Stern15bcb91d2010-09-25 23:35:21 +0200841The implementation is well suited for asynchronous use in interrupt contexts.
842However such use inevitably involves races, because the PM core can't
843synchronize ->runtime_suspend() callbacks with the arrival of I/O requests.
844This synchronization must be handled by the driver, using its private lock.
845Here is a schematic pseudo-code example:
846
847 foo_read_or_write(struct foo_priv *foo, void *data)
848 {
849 lock(&foo->private_lock);
850 add_request_to_io_queue(foo, data);
851 if (foo->num_pending_requests++ == 0)
852 pm_runtime_get(&foo->dev);
853 if (!foo->is_suspended)
854 foo_process_next_request(foo);
855 unlock(&foo->private_lock);
856 }
857
858 foo_io_completion(struct foo_priv *foo, void *req)
859 {
860 lock(&foo->private_lock);
861 if (--foo->num_pending_requests == 0) {
862 pm_runtime_mark_last_busy(&foo->dev);
863 pm_runtime_put_autosuspend(&foo->dev);
864 } else {
865 foo_process_next_request(foo);
866 }
867 unlock(&foo->private_lock);
868 /* Send req result back to the user ... */
869 }
870
871 int foo_runtime_suspend(struct device *dev)
872 {
873 struct foo_priv foo = container_of(dev, ...);
874 int ret = 0;
875
876 lock(&foo->private_lock);
877 if (foo->num_pending_requests > 0) {
878 ret = -EBUSY;
879 } else {
880 /* ... suspend the device ... */
881 foo->is_suspended = 1;
882 }
883 unlock(&foo->private_lock);
884 return ret;
885 }
886
887 int foo_runtime_resume(struct device *dev)
888 {
889 struct foo_priv foo = container_of(dev, ...);
890
891 lock(&foo->private_lock);
892 /* ... resume the device ... */
893 foo->is_suspended = 0;
894 pm_runtime_mark_last_busy(&foo->dev);
895 if (foo->num_pending_requests > 0)
Geert Uytterhoevenfe982452014-03-24 21:31:31 +0100896 foo_process_next_request(foo);
Alan Stern15bcb91d2010-09-25 23:35:21 +0200897 unlock(&foo->private_lock);
898 return 0;
899 }
900
901The important point is that after foo_io_completion() asks for an autosuspend,
902the foo_runtime_suspend() callback may race with foo_read_or_write().
903Therefore foo_runtime_suspend() has to check whether there are any pending I/O
904requests (while holding the private lock) before allowing the suspend to
905proceed.
906
907In addition, the power.autosuspend_delay field can be changed by user space at
908any time. If a driver cares about this, it can call
909pm_runtime_autosuspend_expiration() from within the ->runtime_suspend()
910callback while holding its private lock. If the function returns a nonzero
911value then the delay has not yet expired and the callback should return
912-EAGAIN.