aboutsummaryrefslogtreecommitdiffstats
path: root/Documentation/power/runtime_pm.txt
diff options
context:
space:
mode:
Diffstat (limited to 'Documentation/power/runtime_pm.txt')
-rw-r--r--Documentation/power/runtime_pm.txt268
1 files changed, 179 insertions, 89 deletions
diff --git a/Documentation/power/runtime_pm.txt b/Documentation/power/runtime_pm.txt
index 22accb3eb40e..6066e3a6b9a9 100644
--- a/Documentation/power/runtime_pm.txt
+++ b/Documentation/power/runtime_pm.txt
@@ -1,39 +1,39 @@
1Run-time Power Management Framework for I/O Devices 1Runtime Power Management Framework for I/O Devices
2 2
3(C) 2009-2011 Rafael J. Wysocki <rjw@sisk.pl>, Novell Inc. 3(C) 2009-2011 Rafael J. Wysocki <rjw@sisk.pl>, Novell Inc.
4(C) 2010 Alan Stern <stern@rowland.harvard.edu> 4(C) 2010 Alan Stern <stern@rowland.harvard.edu>
5 5
61. Introduction 61. Introduction
7 7
8Support for run-time power management (run-time PM) of I/O devices is provided 8Support for runtime power management (runtime PM) of I/O devices is provided
9at the power management core (PM core) level by means of: 9at the power management core (PM core) level by means of:
10 10
11* The power management workqueue pm_wq in which bus types and device drivers can 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 12 put their PM-related work items. It is strongly recommended that pm_wq be
13 used for queuing all work items related to run-time PM, because this allows 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, 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 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. 16 include/linux/pm_runtime.h and defined in kernel/power/main.c.
17 17
18* A number of run-time PM fields in the 'power' member of 'struct device' (which 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 19 is of the type 'struct dev_pm_info', defined in include/linux/pm.h) that can
20 be used for synchronizing run-time PM operations with one another. 20 be used for synchronizing runtime PM operations with one another.
21 21
22* Three device run-time PM callbacks in 'struct dev_pm_ops' (defined in 22* Three device runtime PM callbacks in 'struct dev_pm_ops' (defined in
23 include/linux/pm.h). 23 include/linux/pm.h).
24 24
25* A set of helper functions defined in drivers/base/power/runtime.c that can be 25* A set of helper functions defined in drivers/base/power/runtime.c that can be
26 used for carrying out run-time PM operations in such a way that the 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 27 synchronization between them is taken care of by the PM core. Bus types and
28 device drivers are encouraged to use these functions. 28 device drivers are encouraged to use these functions.
29 29
30The run-time PM callbacks present in 'struct dev_pm_ops', the device run-time PM 30The runtime PM callbacks present in 'struct dev_pm_ops', the device runtime PM
31fields of 'struct dev_pm_info' and the core helper functions provided for 31fields of 'struct dev_pm_info' and the core helper functions provided for
32run-time PM are described below. 32runtime PM are described below.
33 33
342. Device Run-time PM Callbacks 342. Device Runtime PM Callbacks
35 35
36There are three device run-time PM callbacks defined in 'struct dev_pm_ops': 36There are three device runtime PM callbacks defined in 'struct dev_pm_ops':
37 37
38struct dev_pm_ops { 38struct dev_pm_ops {
39 ... 39 ...
@@ -54,11 +54,10 @@ referred to as subsystem-level callbacks in what follows.
54By default, the callbacks are always invoked in process context with interrupts 54By default, the callbacks are always invoked in process context with interrupts
55enabled. However, subsystems can use the pm_runtime_irq_safe() helper function 55enabled. However, subsystems can use the pm_runtime_irq_safe() helper function
56to tell the PM core that a device's ->runtime_suspend() and ->runtime_resume() 56to tell the PM core that a device's ->runtime_suspend() and ->runtime_resume()
57callbacks should be invoked in atomic context with interrupts disabled 57callbacks should be invoked in atomic context with interrupts disabled.
58(->runtime_idle() is still invoked the default way). This implies that these 58This implies that these callback routines must not block or sleep, but it also
59callback routines must not block or sleep, but it also means that the 59means that the synchronous helper functions listed at the end of Section 4 can
60synchronous helper functions listed at the end of Section 4 can be used within 60be used within an interrupt handler or in an atomic context.
61an interrupt handler or in an atomic context.
62 61
63The subsystem-level suspend callback is _entirely_ _responsible_ for handling 62The subsystem-level suspend callback is _entirely_ _responsible_ for handling
64the suspend of the device as appropriate, which may, but need not include 63the suspend of the device as appropriate, which may, but need not include
@@ -72,11 +71,11 @@ knows what to do to handle the device).
72 not mean that the device has been put into a low power state. It is 71 not mean that the device has been put into a low power state. It is
73 supposed to mean, however, that the device will not process data and will 72 supposed to mean, however, that the device will not process data and will
74 not communicate with the CPU(s) and RAM until the subsystem-level resume 73 not communicate with the CPU(s) and RAM until the subsystem-level resume
75 callback is executed for it. The run-time PM status of a device after 74 callback is executed for it. The runtime PM status of a device after
76 successful execution of the subsystem-level suspend callback is 'suspended'. 75 successful execution of the subsystem-level suspend callback is 'suspended'.
77 76
78 * If the subsystem-level suspend callback returns -EBUSY or -EAGAIN, 77 * If the subsystem-level suspend callback returns -EBUSY or -EAGAIN,
79 the device's run-time PM status is 'active', which means that the device 78 the device's runtime PM status is 'active', which means that the device
80 _must_ be fully operational afterwards. 79 _must_ be fully operational afterwards.
81 80
82 * If the subsystem-level suspend callback returns an error code different 81 * If the subsystem-level suspend callback returns an error code different
@@ -104,7 +103,7 @@ the device).
104 103
105 * Once the subsystem-level resume callback has completed successfully, the PM 104 * Once the subsystem-level resume callback has completed successfully, the PM
106 core regards the device as fully operational, which means that the device 105 core regards the device as fully operational, which means that the device
107 _must_ be able to complete I/O operations as needed. The run-time PM status 106 _must_ be able to complete I/O operations as needed. The runtime PM status
108 of the device is then 'active'. 107 of the device is then 'active'.
109 108
110 * If the subsystem-level resume callback returns an error code, the PM core 109 * If the subsystem-level resume callback returns an error code, the PM core
@@ -130,7 +129,7 @@ device in that case. The value returned by this callback is ignored by the PM
130core. 129core.
131 130
132The helper functions provided by the PM core, described in Section 4, guarantee 131The helper functions provided by the PM core, described in Section 4, guarantee
133that the following constraints are met with respect to the bus type's run-time 132that the following constraints are met with respect to the bus type's runtime
134PM callbacks: 133PM callbacks:
135 134
136(1) The callbacks are mutually exclusive (e.g. it is forbidden to execute 135(1) The callbacks are mutually exclusive (e.g. it is forbidden to execute
@@ -142,7 +141,7 @@ PM callbacks:
142 141
143(2) ->runtime_idle() and ->runtime_suspend() can only be executed for 'active' 142(2) ->runtime_idle() and ->runtime_suspend() can only be executed for 'active'
144 devices (i.e. the PM core will only execute ->runtime_idle() or 143 devices (i.e. the PM core will only execute ->runtime_idle() or
145 ->runtime_suspend() for the devices the run-time PM status of which is 144 ->runtime_suspend() for the devices the runtime PM status of which is
146 'active'). 145 'active').
147 146
148(3) ->runtime_idle() and ->runtime_suspend() can only be executed for a device 147(3) ->runtime_idle() and ->runtime_suspend() can only be executed for a device
@@ -151,7 +150,7 @@ PM callbacks:
151 flag of which is set. 150 flag of which is set.
152 151
153(4) ->runtime_resume() can only be executed for 'suspended' devices (i.e. the 152(4) ->runtime_resume() can only be executed for 'suspended' devices (i.e. the
154 PM core will only execute ->runtime_resume() for the devices the run-time 153 PM core will only execute ->runtime_resume() for the devices the runtime
155 PM status of which is 'suspended'). 154 PM status of which is 'suspended').
156 155
157Additionally, the helper functions provided by the PM core obey the following 156Additionally, the helper functions provided by the PM core obey the following
@@ -171,9 +170,9 @@ rules:
171 scheduled requests to execute the other callbacks for the same device, 170 scheduled requests to execute the other callbacks for the same device,
172 except for scheduled autosuspends. 171 except for scheduled autosuspends.
173 172
1743. Run-time PM Device Fields 1733. Runtime PM Device Fields
175 174
176The following device run-time PM fields are present in 'struct dev_pm_info', as 175The following device runtime PM fields are present in 'struct dev_pm_info', as
177defined in include/linux/pm.h: 176defined in include/linux/pm.h:
178 177
179 struct timer_list suspend_timer; 178 struct timer_list suspend_timer;
@@ -205,7 +204,7 @@ defined in include/linux/pm.h:
205 204
206 unsigned int disable_depth; 205 unsigned int disable_depth;
207 - used for disabling the helper funcions (they work normally if this is 206 - used for disabling the helper funcions (they work normally if this is
208 equal to zero); the initial value of it is 1 (i.e. run-time PM is 207 equal to zero); the initial value of it is 1 (i.e. runtime PM is
209 initially disabled for all devices) 208 initially disabled for all devices)
210 209
211 unsigned int runtime_error; 210 unsigned int runtime_error;
@@ -229,10 +228,10 @@ defined in include/linux/pm.h:
229 suspend to complete; means "start a resume as soon as you've suspended" 228 suspend to complete; means "start a resume as soon as you've suspended"
230 229
231 unsigned int run_wake; 230 unsigned int run_wake;
232 - set if the device is capable of generating run-time wake-up events 231 - set if the device is capable of generating runtime wake-up events
233 232
234 enum rpm_status runtime_status; 233 enum rpm_status runtime_status;
235 - the run-time PM status of the device; this field's initial value is 234 - the runtime PM status of the device; this field's initial value is
236 RPM_SUSPENDED, which means that each device is initially regarded by the 235 RPM_SUSPENDED, which means that each device is initially regarded by the
237 PM core as 'suspended', regardless of its real hardware status 236 PM core as 'suspended', regardless of its real hardware status
238 237
@@ -243,7 +242,7 @@ defined in include/linux/pm.h:
243 and pm_runtime_forbid() helper functions 242 and pm_runtime_forbid() helper functions
244 243
245 unsigned int no_callbacks; 244 unsigned int no_callbacks;
246 - indicates that the device does not use the run-time PM callbacks (see 245 - indicates that the device does not use the runtime PM callbacks (see
247 Section 8); it may be modified only by the pm_runtime_no_callbacks() 246 Section 8); it may be modified only by the pm_runtime_no_callbacks()
248 helper function 247 helper function
249 248
@@ -270,16 +269,16 @@ defined in include/linux/pm.h:
270 269
271All of the above fields are members of the 'power' member of 'struct device'. 270All of the above fields are members of the 'power' member of 'struct device'.
272 271
2734. Run-time PM Device Helper Functions 2724. Runtime PM Device Helper Functions
274 273
275The following run-time PM helper functions are defined in 274The following runtime PM helper functions are defined in
276drivers/base/power/runtime.c and include/linux/pm_runtime.h: 275drivers/base/power/runtime.c and include/linux/pm_runtime.h:
277 276
278 void pm_runtime_init(struct device *dev); 277 void pm_runtime_init(struct device *dev);
279 - initialize the device run-time PM fields in 'struct dev_pm_info' 278 - initialize the device runtime PM fields in 'struct dev_pm_info'
280 279
281 void pm_runtime_remove(struct device *dev); 280 void pm_runtime_remove(struct device *dev);
282 - make sure that the run-time PM of the device will be disabled after 281 - make sure that the runtime PM of the device will be disabled after
283 removing the device from device hierarchy 282 removing the device from device hierarchy
284 283
285 int pm_runtime_idle(struct device *dev); 284 int pm_runtime_idle(struct device *dev);
@@ -289,9 +288,10 @@ drivers/base/power/runtime.c and include/linux/pm_runtime.h:
289 288
290 int pm_runtime_suspend(struct device *dev); 289 int pm_runtime_suspend(struct device *dev);
291 - execute the subsystem-level suspend callback for the device; returns 0 on 290 - execute the subsystem-level suspend callback for the device; returns 0 on
292 success, 1 if the device's run-time PM status was already 'suspended', or 291 success, 1 if the device's runtime PM status was already 'suspended', or
293 error code on failure, where -EAGAIN or -EBUSY means it is safe to attempt 292 error code on failure, where -EAGAIN or -EBUSY means it is safe to attempt
294 to suspend the device again in future 293 to suspend the device again in future and -EACCES means that
294 'power.disable_depth' is different from 0
295 295
296 int pm_runtime_autosuspend(struct device *dev); 296 int pm_runtime_autosuspend(struct device *dev);
297 - same as pm_runtime_suspend() except that the autosuspend delay is taken 297 - same as pm_runtime_suspend() except that the autosuspend delay is taken
@@ -301,10 +301,11 @@ drivers/base/power/runtime.c and include/linux/pm_runtime.h:
301 301
302 int pm_runtime_resume(struct device *dev); 302 int pm_runtime_resume(struct device *dev);
303 - execute the subsystem-level resume callback for the device; returns 0 on 303 - execute the subsystem-level resume callback for the device; returns 0 on
304 success, 1 if the device's run-time PM status was already 'active' or 304 success, 1 if the device's runtime PM status was already 'active' or
305 error code on failure, where -EAGAIN means it may be safe to attempt to 305 error code on failure, where -EAGAIN means it may be safe to attempt to
306 resume the device again in future, but 'power.runtime_error' should be 306 resume the device again in future, but 'power.runtime_error' should be
307 checked additionally 307 checked additionally, and -EACCES means that 'power.disable_depth' is
308 different from 0
308 309
309 int pm_request_idle(struct device *dev); 310 int pm_request_idle(struct device *dev);
310 - submit a request to execute the subsystem-level idle callback for the 311 - submit a request to execute the subsystem-level idle callback for the
@@ -321,7 +322,7 @@ drivers/base/power/runtime.c and include/linux/pm_runtime.h:
321 device in future, where 'delay' is the time to wait before queuing up a 322 device in future, where 'delay' is the time to wait before queuing up a
322 suspend work item in pm_wq, in milliseconds (if 'delay' is zero, the work 323 suspend work item in pm_wq, in milliseconds (if 'delay' is zero, the work
323 item is queued up immediately); returns 0 on success, 1 if the device's PM 324 item is queued up immediately); returns 0 on success, 1 if the device's PM
324 run-time status was already 'suspended', or error code if the request 325 runtime status was already 'suspended', or error code if the request
325 hasn't been scheduled (or queued up if 'delay' is 0); if the execution of 326 hasn't been scheduled (or queued up if 'delay' is 0); if the execution of
326 ->runtime_suspend() is already scheduled and not yet expired, the new 327 ->runtime_suspend() is already scheduled and not yet expired, the new
327 value of 'delay' will be used as the time to wait 328 value of 'delay' will be used as the time to wait
@@ -329,7 +330,7 @@ drivers/base/power/runtime.c and include/linux/pm_runtime.h:
329 int pm_request_resume(struct device *dev); 330 int pm_request_resume(struct device *dev);
330 - submit a request to execute the subsystem-level resume callback for the 331 - submit a request to execute the subsystem-level resume callback for the
331 device (the request is represented by a work item in pm_wq); returns 0 on 332 device (the request is represented by a work item in pm_wq); returns 0 on
332 success, 1 if the device's run-time PM status was already 'active', or 333 success, 1 if the device's runtime PM status was already 'active', or
333 error code if the request hasn't been queued up 334 error code if the request hasn't been queued up
334 335
335 void pm_runtime_get_noresume(struct device *dev); 336 void pm_runtime_get_noresume(struct device *dev);
@@ -367,22 +368,32 @@ drivers/base/power/runtime.c and include/linux/pm_runtime.h:
367 pm_runtime_autosuspend(dev) and return its result 368 pm_runtime_autosuspend(dev) and return its result
368 369
369 void pm_runtime_enable(struct device *dev); 370 void pm_runtime_enable(struct device *dev);
370 - enable the run-time PM helper functions to run the device bus type's 371 - decrement the device's 'power.disable_depth' field; if that field is equal
371 run-time PM callbacks described in Section 2 372 to zero, the runtime PM helper functions can execute subsystem-level
373 callbacks described in Section 2 for the device
372 374
373 int pm_runtime_disable(struct device *dev); 375 int pm_runtime_disable(struct device *dev);
374 - prevent the run-time PM helper functions from running subsystem-level 376 - increment the device's 'power.disable_depth' field (if the value of that
375 run-time PM callbacks for the device, make sure that all of the pending 377 field was previously zero, this prevents subsystem-level runtime PM
376 run-time PM operations on the device are either completed or canceled; 378 callbacks from being run for the device), make sure that all of the pending
379 runtime PM operations on the device are either completed or canceled;
377 returns 1 if there was a resume request pending and it was necessary to 380 returns 1 if there was a resume request pending and it was necessary to
378 execute the subsystem-level resume callback for the device to satisfy that 381 execute the subsystem-level resume callback for the device to satisfy that
379 request, otherwise 0 is returned 382 request, otherwise 0 is returned
380 383
384 int pm_runtime_barrier(struct device *dev);
385 - check if there's a resume request pending for the device and resume it
386 (synchronously) in that case, cancel any other pending runtime PM requests
387 regarding it and wait for all runtime PM operations on it in progress to
388 complete; returns 1 if there was a resume request pending and it was
389 necessary to execute the subsystem-level resume callback for the device to
390 satisfy that request, otherwise 0 is returned
391
381 void pm_suspend_ignore_children(struct device *dev, bool enable); 392 void pm_suspend_ignore_children(struct device *dev, bool enable);
382 - set/unset the power.ignore_children flag of the device 393 - set/unset the power.ignore_children flag of the device
383 394
384 int pm_runtime_set_active(struct device *dev); 395 int pm_runtime_set_active(struct device *dev);
385 - clear the device's 'power.runtime_error' flag, set the device's run-time 396 - clear the device's 'power.runtime_error' flag, set the device's runtime
386 PM status to 'active' and update its parent's counter of 'active' 397 PM status to 'active' and update its parent's counter of 'active'
387 children as appropriate (it is only valid to use this function if 398 children as appropriate (it is only valid to use this function if
388 'power.runtime_error' is set or 'power.disable_depth' is greater than 399 'power.runtime_error' is set or 'power.disable_depth' is greater than
@@ -390,7 +401,7 @@ drivers/base/power/runtime.c and include/linux/pm_runtime.h:
390 which is not active and the 'power.ignore_children' flag of which is unset 401 which is not active and the 'power.ignore_children' flag of which is unset
391 402
392 void pm_runtime_set_suspended(struct device *dev); 403 void pm_runtime_set_suspended(struct device *dev);
393 - clear the device's 'power.runtime_error' flag, set the device's run-time 404 - clear the device's 'power.runtime_error' flag, set the device's runtime
394 PM status to 'suspended' and update its parent's counter of 'active' 405 PM status to 'suspended' and update its parent's counter of 'active'
395 children as appropriate (it is only valid to use this function if 406 children as appropriate (it is only valid to use this function if
396 'power.runtime_error' is set or 'power.disable_depth' is greater than 407 'power.runtime_error' is set or 'power.disable_depth' is greater than
@@ -400,6 +411,9 @@ drivers/base/power/runtime.c and include/linux/pm_runtime.h:
400 - return true if the device's runtime PM status is 'suspended' and its 411 - return true if the device's runtime PM status is 'suspended' and its
401 'power.disable_depth' field is equal to zero, or false otherwise 412 'power.disable_depth' field is equal to zero, or false otherwise
402 413
414 bool pm_runtime_status_suspended(struct device *dev);
415 - return true if the device's runtime PM status is 'suspended'
416
403 void pm_runtime_allow(struct device *dev); 417 void pm_runtime_allow(struct device *dev);
404 - set the power.runtime_auto flag for the device and decrease its usage 418 - set the power.runtime_auto flag for the device and decrease its usage
405 counter (used by the /sys/devices/.../power/control interface to 419 counter (used by the /sys/devices/.../power/control interface to
@@ -411,14 +425,13 @@ drivers/base/power/runtime.c and include/linux/pm_runtime.h:
411 effectively prevent the device from being power managed at run time) 425 effectively prevent the device from being power managed at run time)
412 426
413 void pm_runtime_no_callbacks(struct device *dev); 427 void pm_runtime_no_callbacks(struct device *dev);
414 - set the power.no_callbacks flag for the device and remove the run-time 428 - set the power.no_callbacks flag for the device and remove the runtime
415 PM attributes from /sys/devices/.../power (or prevent them from being 429 PM attributes from /sys/devices/.../power (or prevent them from being
416 added when the device is registered) 430 added when the device is registered)
417 431
418 void pm_runtime_irq_safe(struct device *dev); 432 void pm_runtime_irq_safe(struct device *dev);
419 - set the power.irq_safe flag for the device, causing the runtime-PM 433 - set the power.irq_safe flag for the device, causing the runtime-PM
420 suspend and resume callbacks (but not the idle callback) to be invoked 434 callbacks to be invoked with interrupts off
421 with interrupts disabled
422 435
423 void pm_runtime_mark_last_busy(struct device *dev); 436 void pm_runtime_mark_last_busy(struct device *dev);
424 - set the power.last_busy field to the current time 437 - set the power.last_busy field to the current time
@@ -431,7 +444,7 @@ drivers/base/power/runtime.c and include/linux/pm_runtime.h:
431 444
432 void pm_runtime_set_autosuspend_delay(struct device *dev, int delay); 445 void pm_runtime_set_autosuspend_delay(struct device *dev, int delay);
433 - set the power.autosuspend_delay value to 'delay' (expressed in 446 - set the power.autosuspend_delay value to 'delay' (expressed in
434 milliseconds); if 'delay' is negative then run-time suspends are 447 milliseconds); if 'delay' is negative then runtime suspends are
435 prevented 448 prevented
436 449
437 unsigned long pm_runtime_autosuspend_expiration(struct device *dev); 450 unsigned long pm_runtime_autosuspend_expiration(struct device *dev);
@@ -468,78 +481,95 @@ pm_runtime_suspend()
468pm_runtime_autosuspend() 481pm_runtime_autosuspend()
469pm_runtime_resume() 482pm_runtime_resume()
470pm_runtime_get_sync() 483pm_runtime_get_sync()
484pm_runtime_put_sync()
471pm_runtime_put_sync_suspend() 485pm_runtime_put_sync_suspend()
472 486
4735. Run-time PM Initialization, Device Probing and Removal 4875. Runtime PM Initialization, Device Probing and Removal
474 488
475Initially, the run-time PM is disabled for all devices, which means that the 489Initially, the runtime PM is disabled for all devices, which means that the
476majority of the run-time PM helper funtions described in Section 4 will return 490majority of the runtime PM helper funtions described in Section 4 will return
477-EAGAIN until pm_runtime_enable() is called for the device. 491-EAGAIN until pm_runtime_enable() is called for the device.
478 492
479In addition to that, the initial run-time PM status of all devices is 493In addition to that, the initial runtime PM status of all devices is
480'suspended', but it need not reflect the actual physical state of the device. 494'suspended', but it need not reflect the actual physical state of the device.
481Thus, if the device is initially active (i.e. it is able to process I/O), its 495Thus, if the device is initially active (i.e. it is able to process I/O), its
482run-time PM status must be changed to 'active', with the help of 496runtime PM status must be changed to 'active', with the help of
483pm_runtime_set_active(), before pm_runtime_enable() is called for the device. 497pm_runtime_set_active(), before pm_runtime_enable() is called for the device.
484 498
485However, if the device has a parent and the parent's run-time PM is enabled, 499However, if the device has a parent and the parent's runtime PM is enabled,
486calling pm_runtime_set_active() for the device will affect the parent, unless 500calling pm_runtime_set_active() for the device will affect the parent, unless
487the parent's 'power.ignore_children' flag is set. Namely, in that case the 501the parent's 'power.ignore_children' flag is set. Namely, in that case the
488parent won't be able to suspend at run time, using the PM core's helper 502parent won't be able to suspend at run time, using the PM core's helper
489functions, as long as the child's status is 'active', even if the child's 503functions, as long as the child's status is 'active', even if the child's
490run-time PM is still disabled (i.e. pm_runtime_enable() hasn't been called for 504runtime PM is still disabled (i.e. pm_runtime_enable() hasn't been called for
491the child yet or pm_runtime_disable() has been called for it). For this reason, 505the child yet or pm_runtime_disable() has been called for it). For this reason,
492once pm_runtime_set_active() has been called for the device, pm_runtime_enable() 506once pm_runtime_set_active() has been called for the device, pm_runtime_enable()
493should be called for it too as soon as reasonably possible or its run-time PM 507should be called for it too as soon as reasonably possible or its runtime PM
494status should be changed back to 'suspended' with the help of 508status should be changed back to 'suspended' with the help of
495pm_runtime_set_suspended(). 509pm_runtime_set_suspended().
496 510
497If the default initial run-time PM status of the device (i.e. 'suspended') 511If the default initial runtime PM status of the device (i.e. 'suspended')
498reflects the actual state of the device, its bus type's or its driver's 512reflects the actual state of the device, its bus type's or its driver's
499->probe() callback will likely need to wake it up using one of the PM core's 513->probe() callback will likely need to wake it up using one of the PM core's
500helper functions described in Section 4. In that case, pm_runtime_resume() 514helper functions described in Section 4. In that case, pm_runtime_resume()
501should be used. Of course, for this purpose the device's run-time PM has to be 515should be used. Of course, for this purpose the device's runtime PM has to be
502enabled earlier by calling pm_runtime_enable(). 516enabled earlier by calling pm_runtime_enable().
503 517
504If the device bus type's or driver's ->probe() or ->remove() callback runs 518If the device bus type's or driver's ->probe() callback runs
505pm_runtime_suspend() or pm_runtime_idle() or their asynchronous counterparts, 519pm_runtime_suspend() or pm_runtime_idle() or their asynchronous counterparts,
506they will fail returning -EAGAIN, because the device's usage counter is 520they will fail returning -EAGAIN, because the device's usage counter is
507incremented by the core before executing ->probe() and ->remove(). Still, it 521incremented by the driver core before executing ->probe(). Still, it may be
508may be desirable to suspend the device as soon as ->probe() or ->remove() has 522desirable to suspend the device as soon as ->probe() has finished, so the driver
509finished, so the PM core uses pm_runtime_idle_sync() to invoke the 523core uses pm_runtime_put_sync() to invoke the subsystem-level idle callback for
510subsystem-level idle callback for the device at that time. 524the device at that time.
525
526Moreover, the driver core prevents runtime PM callbacks from racing with the bus
527notifier callback in __device_release_driver(), which is necessary, because the
528notifier is used by some subsystems to carry out operations affecting the
529runtime PM functionality. It does so by calling pm_runtime_get_sync() before
530driver_sysfs_remove() and the BUS_NOTIFY_UNBIND_DRIVER notifications. This
531resumes the device if it's in the suspended state and prevents it from
532being suspended again while those routines are being executed.
533
534To allow bus types and drivers to put devices into the suspended state by
535calling pm_runtime_suspend() from their ->remove() routines, the driver core
536executes pm_runtime_put_sync() after running the BUS_NOTIFY_UNBIND_DRIVER
537notifications in __device_release_driver(). This requires bus types and
538drivers to make their ->remove() callbacks avoid races with runtime PM directly,
539but also it allows of more flexibility in the handling of devices during the
540removal of their drivers.
511 541
512The user space can effectively disallow the driver of the device to power manage 542The user space can effectively disallow the driver of the device to power manage
513it at run time by changing the value of its /sys/devices/.../power/control 543it at run time by changing the value of its /sys/devices/.../power/control
514attribute to "on", which causes pm_runtime_forbid() to be called. In principle, 544attribute to "on", which causes pm_runtime_forbid() to be called. In principle,
515this mechanism may also be used by the driver to effectively turn off the 545this mechanism may also be used by the driver to effectively turn off the
516run-time power management of the device until the user space turns it on. 546runtime power management of the device until the user space turns it on.
517Namely, during the initialization the driver can make sure that the run-time PM 547Namely, during the initialization the driver can make sure that the runtime PM
518status of the device is 'active' and call pm_runtime_forbid(). It should be 548status of the device is 'active' and call pm_runtime_forbid(). It should be
519noted, however, that if the user space has already intentionally changed the 549noted, however, that if the user space has already intentionally changed the
520value of /sys/devices/.../power/control to "auto" to allow the driver to power 550value of /sys/devices/.../power/control to "auto" to allow the driver to power
521manage the device at run time, the driver may confuse it by using 551manage the device at run time, the driver may confuse it by using
522pm_runtime_forbid() this way. 552pm_runtime_forbid() this way.
523 553
5246. Run-time PM and System Sleep 5546. Runtime PM and System Sleep
525 555
526Run-time PM and system sleep (i.e., system suspend and hibernation, also known 556Runtime PM and system sleep (i.e., system suspend and hibernation, also known
527as suspend-to-RAM and suspend-to-disk) interact with each other in a couple of 557as suspend-to-RAM and suspend-to-disk) interact with each other in a couple of
528ways. If a device is active when a system sleep starts, everything is 558ways. If a device is active when a system sleep starts, everything is
529straightforward. But what should happen if the device is already suspended? 559straightforward. But what should happen if the device is already suspended?
530 560
531The device may have different wake-up settings for run-time PM and system sleep. 561The device may have different wake-up settings for runtime PM and system sleep.
532For example, remote wake-up may be enabled for run-time suspend but disallowed 562For example, remote wake-up may be enabled for runtime suspend but disallowed
533for system sleep (device_may_wakeup(dev) returns 'false'). When this happens, 563for system sleep (device_may_wakeup(dev) returns 'false'). When this happens,
534the subsystem-level system suspend callback is responsible for changing the 564the subsystem-level system suspend callback is responsible for changing the
535device's wake-up setting (it may leave that to the device driver's system 565device's wake-up setting (it may leave that to the device driver's system
536suspend routine). It may be necessary to resume the device and suspend it again 566suspend routine). It may be necessary to resume the device and suspend it again
537in order to do so. The same is true if the driver uses different power levels 567in order to do so. The same is true if the driver uses different power levels
538or other settings for run-time suspend and system sleep. 568or other settings for runtime suspend and system sleep.
539 569
540During system resume, devices generally should be brought back to full power, 570During system resume, the simplest approach is to bring all devices back to full
541even if they were suspended before the system sleep began. There are several 571power, even if they had been suspended before the system suspend began. There
542reasons for this, including: 572are several reasons for this, including:
543 573
544 * The device might need to switch power levels, wake-up settings, etc. 574 * The device might need to switch power levels, wake-up settings, etc.
545 575
@@ -554,18 +584,50 @@ reasons for this, including:
554 * The device might need to be reset. 584 * The device might need to be reset.
555 585
556 * Even though the device was suspended, if its usage counter was > 0 then most 586 * Even though the device was suspended, if its usage counter was > 0 then most
557 likely it would need a run-time resume in the near future anyway. 587 likely it would need a runtime resume in the near future anyway.
558
559 * Always going back to full power is simplest.
560 588
561If the device was suspended before the sleep began, then its run-time PM status 589If the device had been suspended before the system suspend began and it's
562will have to be updated to reflect the actual post-system sleep status. The way 590brought back to full power during resume, then its runtime PM status will have
563to do this is: 591to be updated to reflect the actual post-system sleep status. The way to do
592this is:
564 593
565 pm_runtime_disable(dev); 594 pm_runtime_disable(dev);
566 pm_runtime_set_active(dev); 595 pm_runtime_set_active(dev);
567 pm_runtime_enable(dev); 596 pm_runtime_enable(dev);
568 597
598The PM core always increments the runtime usage counter before calling the
599->suspend() callback and decrements it after calling the ->resume() callback.
600Hence disabling runtime PM temporarily like this will not cause any runtime
601suspend attempts to be permanently lost. If the usage count goes to zero
602following the return of the ->resume() callback, the ->runtime_idle() callback
603will be invoked as usual.
604
605On some systems, however, system sleep is not entered through a global firmware
606or hardware operation. Instead, all hardware components are put into low-power
607states directly by the kernel in a coordinated way. Then, the system sleep
608state effectively follows from the states the hardware components end up in
609and the system is woken up from that state by a hardware interrupt or a similar
610mechanism entirely under the kernel's control. As a result, the kernel never
611gives control away and the states of all devices during resume are precisely
612known to it. If that is the case and none of the situations listed above takes
613place (in particular, if the system is not waking up from hibernation), it may
614be more efficient to leave the devices that had been suspended before the system
615suspend began in the suspended state.
616
617The PM core does its best to reduce the probability of race conditions between
618the runtime PM and system suspend/resume (and hibernation) callbacks by carrying
619out the following operations:
620
621 * During system suspend it calls pm_runtime_get_noresume() and
622 pm_runtime_barrier() for every device right before executing the
623 subsystem-level .suspend() callback for it. In addition to that it calls
624 pm_runtime_disable() for every device right after executing the
625 subsystem-level .suspend() callback for it.
626
627 * During system resume it calls pm_runtime_enable() and pm_runtime_put_sync()
628 for every device right before and right after executing the subsystem-level
629 .resume() callback for it, respectively.
630
5697. Generic subsystem callbacks 6317. Generic subsystem callbacks
570 632
571Subsystems may wish to conserve code space by using the set of generic power 633Subsystems may wish to conserve code space by using the set of generic power
@@ -590,40 +652,68 @@ driver/base/power/generic_ops.c:
590 callback provided by its driver and return its result, or return 0 if not 652 callback provided by its driver and return its result, or return 0 if not
591 defined 653 defined
592 654
655 int pm_generic_suspend_noirq(struct device *dev);
656 - if pm_runtime_suspended(dev) returns "false", invoke the ->suspend_noirq()
657 callback provided by the device's driver and return its result, or return
658 0 if not defined
659
593 int pm_generic_resume(struct device *dev); 660 int pm_generic_resume(struct device *dev);
594 - invoke the ->resume() callback provided by the driver of this device and, 661 - invoke the ->resume() callback provided by the driver of this device and,
595 if successful, change the device's runtime PM status to 'active' 662 if successful, change the device's runtime PM status to 'active'
596 663
664 int pm_generic_resume_noirq(struct device *dev);
665 - invoke the ->resume_noirq() callback provided by the driver of this device
666
597 int pm_generic_freeze(struct device *dev); 667 int pm_generic_freeze(struct device *dev);
598 - if the device has not been suspended at run time, invoke the ->freeze() 668 - if the device has not been suspended at run time, invoke the ->freeze()
599 callback provided by its driver and return its result, or return 0 if not 669 callback provided by its driver and return its result, or return 0 if not
600 defined 670 defined
601 671
672 int pm_generic_freeze_noirq(struct device *dev);
673 - if pm_runtime_suspended(dev) returns "false", invoke the ->freeze_noirq()
674 callback provided by the device's driver and return its result, or return
675 0 if not defined
676
602 int pm_generic_thaw(struct device *dev); 677 int pm_generic_thaw(struct device *dev);
603 - if the device has not been suspended at run time, invoke the ->thaw() 678 - if the device has not been suspended at run time, invoke the ->thaw()
604 callback provided by its driver and return its result, or return 0 if not 679 callback provided by its driver and return its result, or return 0 if not
605 defined 680 defined
606 681
682 int pm_generic_thaw_noirq(struct device *dev);
683 - if pm_runtime_suspended(dev) returns "false", invoke the ->thaw_noirq()
684 callback provided by the device's driver and return its result, or return
685 0 if not defined
686
607 int pm_generic_poweroff(struct device *dev); 687 int pm_generic_poweroff(struct device *dev);
608 - if the device has not been suspended at run time, invoke the ->poweroff() 688 - if the device has not been suspended at run time, invoke the ->poweroff()
609 callback provided by its driver and return its result, or return 0 if not 689 callback provided by its driver and return its result, or return 0 if not
610 defined 690 defined
611 691
692 int pm_generic_poweroff_noirq(struct device *dev);
693 - if pm_runtime_suspended(dev) returns "false", run the ->poweroff_noirq()
694 callback provided by the device's driver and return its result, or return
695 0 if not defined
696
612 int pm_generic_restore(struct device *dev); 697 int pm_generic_restore(struct device *dev);
613 - invoke the ->restore() callback provided by the driver of this device and, 698 - invoke the ->restore() callback provided by the driver of this device and,
614 if successful, change the device's runtime PM status to 'active' 699 if successful, change the device's runtime PM status to 'active'
615 700
701 int pm_generic_restore_noirq(struct device *dev);
702 - invoke the ->restore_noirq() callback provided by the device's driver
703
616These functions can be assigned to the ->runtime_idle(), ->runtime_suspend(), 704These functions can be assigned to the ->runtime_idle(), ->runtime_suspend(),
617->runtime_resume(), ->suspend(), ->resume(), ->freeze(), ->thaw(), ->poweroff(), 705->runtime_resume(), ->suspend(), ->suspend_noirq(), ->resume(),
618or ->restore() callback pointers in the subsystem-level dev_pm_ops structures. 706->resume_noirq(), ->freeze(), ->freeze_noirq(), ->thaw(), ->thaw_noirq(),
707->poweroff(), ->poweroff_noirq(), ->restore(), ->restore_noirq() callback
708pointers in the subsystem-level dev_pm_ops structures.
619 709
620If a subsystem wishes to use all of them at the same time, it can simply assign 710If a subsystem wishes to use all of them at the same time, it can simply assign
621the GENERIC_SUBSYS_PM_OPS macro, defined in include/linux/pm.h, to its 711the GENERIC_SUBSYS_PM_OPS macro, defined in include/linux/pm.h, to its
622dev_pm_ops structure pointer. 712dev_pm_ops structure pointer.
623 713
624Device drivers that wish to use the same function as a system suspend, freeze, 714Device drivers that wish to use the same function as a system suspend, freeze,
625poweroff and run-time suspend callback, and similarly for system resume, thaw, 715poweroff and runtime suspend callback, and similarly for system resume, thaw,
626restore, and run-time resume, can achieve this with the help of the 716restore, and runtime resume, can achieve this with the help of the
627UNIVERSAL_DEV_PM_OPS macro defined in include/linux/pm.h (possibly setting its 717UNIVERSAL_DEV_PM_OPS macro defined in include/linux/pm.h (possibly setting its
628last argument to NULL). 718last argument to NULL).
629 719
@@ -633,7 +723,7 @@ Some "devices" are only logical sub-devices of their parent and cannot be
633power-managed on their own. (The prototype example is a USB interface. Entire 723power-managed on their own. (The prototype example is a USB interface. Entire
634USB devices can go into low-power mode or send wake-up requests, but neither is 724USB devices can go into low-power mode or send wake-up requests, but neither is
635possible for individual interfaces.) The drivers for these devices have no 725possible for individual interfaces.) The drivers for these devices have no
636need of run-time PM callbacks; if the callbacks did exist, ->runtime_suspend() 726need of runtime PM callbacks; if the callbacks did exist, ->runtime_suspend()
637and ->runtime_resume() would always return 0 without doing anything else and 727and ->runtime_resume() would always return 0 without doing anything else and
638->runtime_idle() would always call pm_runtime_suspend(). 728->runtime_idle() would always call pm_runtime_suspend().
639 729
@@ -641,7 +731,7 @@ Subsystems can tell the PM core about these devices by calling
641pm_runtime_no_callbacks(). This should be done after the device structure is 731pm_runtime_no_callbacks(). This should be done after the device structure is
642initialized and before it is registered (although after device registration is 732initialized and before it is registered (although after device registration is
643also okay). The routine will set the device's power.no_callbacks flag and 733also okay). The routine will set the device's power.no_callbacks flag and
644prevent the non-debugging run-time PM sysfs attributes from being created. 734prevent the non-debugging runtime PM sysfs attributes from being created.
645 735
646When power.no_callbacks is set, the PM core will not invoke the 736When power.no_callbacks is set, the PM core will not invoke the
647->runtime_idle(), ->runtime_suspend(), or ->runtime_resume() callbacks. 737->runtime_idle(), ->runtime_suspend(), or ->runtime_resume() callbacks.
@@ -649,7 +739,7 @@ Instead it will assume that suspends and resumes always succeed and that idle
649devices should be suspended. 739devices should be suspended.
650 740
651As a consequence, the PM core will never directly inform the device's subsystem 741As a consequence, the PM core will never directly inform the device's subsystem
652or driver about run-time power changes. Instead, the driver for the device's 742or driver about runtime power changes. Instead, the driver for the device's
653parent must take responsibility for telling the device's driver when the 743parent must take responsibility for telling the device's driver when the
654parent's power state changes. 744parent's power state changes.
655 745
@@ -660,13 +750,13 @@ A device should be put in a low-power state only when there's some reason to
660think it will remain in that state for a substantial time. A common heuristic 750think it will remain in that state for a substantial time. A common heuristic
661says that a device which hasn't been used for a while is liable to remain 751says that a device which hasn't been used for a while is liable to remain
662unused; following this advice, drivers should not allow devices to be suspended 752unused; following this advice, drivers should not allow devices to be suspended
663at run-time until they have been inactive for some minimum period. Even when 753at runtime until they have been inactive for some minimum period. Even when
664the heuristic ends up being non-optimal, it will still prevent devices from 754the heuristic ends up being non-optimal, it will still prevent devices from
665"bouncing" too rapidly between low-power and full-power states. 755"bouncing" too rapidly between low-power and full-power states.
666 756
667The term "autosuspend" is an historical remnant. It doesn't mean that the 757The term "autosuspend" is an historical remnant. It doesn't mean that the
668device is automatically suspended (the subsystem or driver still has to call 758device is automatically suspended (the subsystem or driver still has to call
669the appropriate PM routines); rather it means that run-time suspends will 759the appropriate PM routines); rather it means that runtime suspends will
670automatically be delayed until the desired period of inactivity has elapsed. 760automatically be delayed until the desired period of inactivity has elapsed.
671 761
672Inactivity is determined based on the power.last_busy field. Drivers should 762Inactivity is determined based on the power.last_busy field. Drivers should