linux/Documentation/power/runtime_pm.txt
<<
>>
Prefs
   1Runtime Power Management Framework for I/O Devices
   2
   3(C) 2009-2011 Rafael J. Wysocki <rjw@sisk.pl>, Novell Inc.
   4(C) 2010 Alan Stern <stern@rowland.harvard.edu>
   5
   61. Introduction
   7
   8Support for runtime power management (runtime PM) of I/O devices is provided
   9at the power management core (PM core) level by means of:
  10
  11* The power management workqueue pm_wq in which bus types and device drivers can
  12  put their PM-related work items.  It is strongly recommended that pm_wq be
  13  used for queuing all work items related to runtime PM, because this allows
  14  them to be synchronized with system-wide power transitions (suspend to RAM,
  15  hibernation and resume from system sleep states).  pm_wq is declared in
  16  include/linux/pm_runtime.h and defined in kernel/power/main.c.
  17
  18* A number of runtime PM fields in the 'power' member of 'struct device' (which
  19  is of the type 'struct dev_pm_info', defined in include/linux/pm.h) that can
  20  be used for synchronizing runtime PM operations with one another.
  21
  22* Three device runtime PM callbacks in 'struct dev_pm_ops' (defined in
  23  include/linux/pm.h).
  24
  25* A set of helper functions defined in drivers/base/power/runtime.c that can be
  26  used for carrying out runtime PM operations in such a way that the
  27  synchronization between them is taken care of by the PM core.  Bus types and
  28  device drivers are encouraged to use these functions.
  29
  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
  32runtime PM are described below.
  33
  342. Device Runtime PM Callbacks
  35
  36There are three device runtime PM callbacks defined in 'struct dev_pm_ops':
  37
  38struct dev_pm_ops {
  39        ...
  40        int (*runtime_suspend)(struct device *dev);
  41        int (*runtime_resume)(struct device *dev);
  42        int (*runtime_idle)(struct device *dev);
  43        ...
  44};
  45
  46The ->runtime_suspend(), ->runtime_resume() and ->runtime_idle() callbacks
  47are executed by the PM core for the device's subsystem that may be either of
  48the following:
  49
  50  1. PM domain of the device, if the device's PM domain object, dev->pm_domain,
  51     is present.
  52
  53  2. Device type of the device, if both dev->type and dev->type->pm are present.
  54
  55  3. Device class of the device, if both dev->class and dev->class->pm are
  56     present.
  57
  58  4. Bus type of the device, if both dev->bus and dev->bus->pm are present.
  59
  60If the subsystem chosen by applying the above rules doesn't provide the relevant
  61callback, the PM core will invoke the corresponding driver callback stored in
  62dev->driver->pm directly (if present).
  63
  64The PM core always checks which callback to use in the order given above, so the
  65priority order of callbacks from high to low is: PM domain, device type, class
  66and bus type.  Moreover, the high-priority one will always take precedence over
  67a low-priority one.  The PM domain, bus type, device type and class callbacks
  68are referred to as subsystem-level callbacks in what follows.
  69
  70By default, the callbacks are always invoked in process context with interrupts
  71enabled.  However, the pm_runtime_irq_safe() helper function can be used to tell
  72the PM core that it is safe to run the ->runtime_suspend(), ->runtime_resume()
  73and ->runtime_idle() callbacks for the given device in atomic context with
  74interrupts disabled.  This implies that the callback routines in question must
  75not block or sleep, but it also means that the synchronous helper functions
  76listed at the end of Section 4 may be used for that device within an interrupt
  77handler or generally in an atomic context.
  78
  79The subsystem-level suspend callback, if present, is _entirely_ _responsible_
  80for handling the suspend of the device as appropriate, which may, but need not
  81include executing the device driver's own ->runtime_suspend() callback (from the
  82PM core's point of view it is not necessary to implement a ->runtime_suspend()
  83callback in a device driver as long as the subsystem-level suspend callback
  84knows what to do to handle the device).
  85
  86  * Once the subsystem-level suspend callback (or the driver suspend callback,
  87    if invoked directly) has completed successfully for the given device, the PM
  88    core regards the device as suspended, which need not mean that it has been
  89    put into a low power state.  It is supposed to mean, however, that the
  90    device will not process data and will not communicate with the CPU(s) and
  91    RAM until the appropriate resume callback is executed for it.  The runtime
  92    PM status of a device after successful execution of the suspend callback is
  93    'suspended'.
  94
  95  * If the suspend callback returns -EBUSY or -EAGAIN, the device's runtime PM
  96    status remains 'active', which means that the device _must_ be fully
  97    operational afterwards.
  98
  99  * If the suspend callback returns an error code different from -EBUSY and
 100    -EAGAIN, the PM core regards this as a fatal error and will refuse to run
 101    the helper functions described in Section 4 for the device until its status
 102    is directly set to  either'active', or 'suspended' (the PM core provides
 103    special helper functions for this purpose).
 104
 105In particular, if the driver requires remote wakeup capability (i.e. hardware
 106mechanism allowing the device to request a change of its power state, such as
 107PCI PME) for proper functioning and device_run_wake() returns 'false' for the
 108device, then ->runtime_suspend() should return -EBUSY.  On the other hand, if
 109device_run_wake() returns 'true' for the device and the device is put into a
 110low-power state during the execution of the suspend callback, it is expected
 111that remote wakeup will be enabled for the device.  Generally, remote wakeup
 112should be enabled for all input devices put into low-power states at run time.
 113
 114The subsystem-level resume callback, if present, is _entirely_ _responsible_ for
 115handling the resume of the device as appropriate, which may, but need not
 116include executing the device driver's own ->runtime_resume() callback (from the
 117PM core's point of view it is not necessary to implement a ->runtime_resume()
 118callback in a device driver as long as the subsystem-level resume callback knows
 119what to do to handle the device).
 120
 121  * Once the subsystem-level resume callback (or the driver resume callback, if
 122    invoked directly) has completed successfully, the PM core regards the device
 123    as fully operational, which means that the device _must_ be able to complete
 124    I/O operations as needed.  The runtime PM status of the device is then
 125    'active'.
 126
 127  * If the resume callback returns an error code, the PM core regards this as a
 128    fatal error and will refuse to run the helper functions described in Section
 129    4 for the device, until its status is directly set to either 'active', or
 130    'suspended' (by means of special helper functions provided by the PM core
 131    for this purpose).
 132
 133The idle callback (a subsystem-level one, if present, or the driver one) is
 134executed by the PM core whenever the device appears to be idle, which is
 135indicated to the PM core by two counters, the device's usage counter and the
 136counter of 'active' children of the device.
 137
 138  * If any of these counters is decreased using a helper function provided by
 139    the PM core and it turns out to be equal to zero, the other counter is
 140    checked.  If that counter also is equal to zero, the PM core executes the
 141    idle callback with the device as its argument.
 142
 143The action performed by the idle callback is totally dependent on the subsystem
 144(or driver) in question, but the expected and recommended action is to check
 145if the device can be suspended (i.e. if all of the conditions necessary for
 146suspending the device are satisfied) and to queue up a suspend request for the
 147device in that case.  If there is no idle callback, or if the callback returns
 1480, then the PM core will attempt to carry out a runtime suspend of the device;
 149in essence, it will call pm_runtime_suspend() directly.  To prevent this (for
 150example, if the callback routine has started a delayed suspend), the routine
 151should return a non-zero value.  Negative error return codes are ignored by the
 152PM core.
 153
 154The helper functions provided by the PM core, described in Section 4, guarantee
 155that the following constraints are met with respect to runtime PM callbacks for
 156one device:
 157
 158(1) The callbacks are mutually exclusive (e.g. it is forbidden to execute
 159    ->runtime_suspend() in parallel with ->runtime_resume() or with another
 160    instance of ->runtime_suspend() for the same device) with the exception that
 161    ->runtime_suspend() or ->runtime_resume() can be executed in parallel with
 162    ->runtime_idle() (although ->runtime_idle() will not be started while any
 163    of the other callbacks is being executed for the same device).
 164
 165(2) ->runtime_idle() and ->runtime_suspend() can only be executed for 'active'
 166    devices (i.e. the PM core will only execute ->runtime_idle() or
 167    ->runtime_suspend() for the devices the runtime PM status of which is
 168    'active').
 169
 170(3) ->runtime_idle() and ->runtime_suspend() can only be executed for a device
 171    the usage counter of which is equal to zero _and_ either the counter of
 172    'active' children of which is equal to zero, or the 'power.ignore_children'
 173    flag of which is set.
 174
 175(4) ->runtime_resume() can only be executed for 'suspended' devices  (i.e. the
 176    PM core will only execute ->runtime_resume() for the devices the runtime
 177    PM status of which is 'suspended').
 178
 179Additionally, the helper functions provided by the PM core obey the following
 180rules:
 181
 182  * If ->runtime_suspend() is about to be executed or there's a pending request
 183    to execute it, ->runtime_idle() will not be executed for the same device.
 184
 185  * A request to execute or to schedule the execution of ->runtime_suspend()
 186    will cancel any pending requests to execute ->runtime_idle() for the same
 187    device.
 188
 189  * If ->runtime_resume() is about to be executed or there's a pending request
 190    to execute it, the other callbacks will not be executed for the same device.
 191
 192  * A request to execute ->runtime_resume() will cancel any pending or
 193    scheduled requests to execute the other callbacks for the same device,
 194    except for scheduled autosuspends.
 195
 1963. Runtime PM Device Fields
 197
 198The following device runtime PM fields are present in 'struct dev_pm_info', as
 199defined in include/linux/pm.h:
 200
 201  struct timer_list suspend_timer;
 202    - timer used for scheduling (delayed) suspend and autosuspend requests
 203
 204  unsigned long timer_expires;
 205    - timer expiration time, in jiffies (if this is different from zero, the
 206      timer is running and will expire at that time, otherwise the timer is not
 207      running)
 208
 209  struct work_struct work;
 210    - work structure used for queuing up requests (i.e. work items in pm_wq)
 211
 212  wait_queue_head_t wait_queue;
 213    - wait queue used if any of the helper functions needs to wait for another
 214      one to complete
 215
 216  spinlock_t lock;
 217    - lock used for synchronisation
 218
 219  atomic_t usage_count;
 220    - the usage counter of the device
 221
 222  atomic_t child_count;
 223    - the count of 'active' children of the device
 224
 225  unsigned int ignore_children;
 226    - if set, the value of child_count is ignored (but still updated)
 227
 228  unsigned int disable_depth;
 229    - used for disabling the helper funcions (they work normally if this is
 230      equal to zero); the initial value of it is 1 (i.e. runtime PM is
 231      initially disabled for all devices)
 232
 233  unsigned int runtime_error;
 234    - if set, there was a fatal error (one of the callbacks returned error code
 235      as described in Section 2), so the helper funtions will not work until
 236      this flag is cleared; this is the error code returned by the failing
 237      callback
 238
 239  unsigned int idle_notification;
 240    - if set, ->runtime_idle() is being executed
 241
 242  unsigned int request_pending;
 243    - if set, there's a pending request (i.e. a work item queued up into pm_wq)
 244
 245  enum rpm_request request;
 246    - type of request that's pending (valid if request_pending is set)
 247
 248  unsigned int deferred_resume;
 249    - set if ->runtime_resume() is about to be run while ->runtime_suspend() is
 250      being executed for that device and it is not practical to wait for the
 251      suspend to complete; means "start a resume as soon as you've suspended"
 252
 253  unsigned int run_wake;
 254    - set if the device is capable of generating runtime wake-up events
 255
 256  enum rpm_status runtime_status;
 257    - the runtime PM status of the device; this field's initial value is
 258      RPM_SUSPENDED, which means that each device is initially regarded by the
 259      PM core as 'suspended', regardless of its real hardware status
 260
 261  unsigned int runtime_auto;
 262    - if set, indicates that the user space has allowed the device driver to
 263      power manage the device at run time via the /sys/devices/.../power/control
 264      interface; it may only be modified with the help of the pm_runtime_allow()
 265      and pm_runtime_forbid() helper functions
 266
 267  unsigned int no_callbacks;
 268    - indicates that the device does not use the runtime PM callbacks (see
 269      Section 8); it may be modified only by the pm_runtime_no_callbacks()
 270      helper function
 271
 272  unsigned int irq_safe;
 273    - indicates that the ->runtime_suspend() and ->runtime_resume() callbacks
 274      will be invoked with the spinlock held and interrupts disabled
 275
 276  unsigned int use_autosuspend;
 277    - indicates that the device's driver supports delayed autosuspend (see
 278      Section 9); it may be modified only by the
 279      pm_runtime{_dont}_use_autosuspend() helper functions
 280
 281  unsigned int timer_autosuspends;
 282    - indicates that the PM core should attempt to carry out an autosuspend
 283      when the timer expires rather than a normal suspend
 284
 285  int autosuspend_delay;
 286    - the delay time (in milliseconds) to be used for autosuspend
 287
 288  unsigned long last_busy;
 289    - the time (in jiffies) when the pm_runtime_mark_last_busy() helper
 290      function was last called for this device; used in calculating inactivity
 291      periods for autosuspend
 292
 293All of the above fields are members of the 'power' member of 'struct device'.
 294
 2954. Runtime PM Device Helper Functions
 296
 297The following runtime PM helper functions are defined in
 298drivers/base/power/runtime.c and include/linux/pm_runtime.h:
 299
 300  void pm_runtime_init(struct device *dev);
 301    - initialize the device runtime PM fields in 'struct dev_pm_info'
 302
 303  void pm_runtime_remove(struct device *dev);
 304    - make sure that the runtime PM of the device will be disabled after
 305      removing the device from device hierarchy
 306
 307  int pm_runtime_idle(struct device *dev);
 308    - execute the subsystem-level idle callback for the device; returns an
 309      error code on failure, where -EINPROGRESS means that ->runtime_idle() is
 310      already being executed; if there is no callback or the callback returns 0
 311      then run pm_runtime_suspend(dev) and return its result
 312
 313  int pm_runtime_suspend(struct device *dev);
 314    - execute the subsystem-level suspend callback for the device; returns 0 on
 315      success, 1 if the device's runtime PM status was already 'suspended', or
 316      error code on failure, where -EAGAIN or -EBUSY means it is safe to attempt
 317      to suspend the device again in future and -EACCES means that
 318      'power.disable_depth' is different from 0
 319
 320  int pm_runtime_autosuspend(struct device *dev);
 321    - same as pm_runtime_suspend() except that the autosuspend delay is taken
 322      into account; if pm_runtime_autosuspend_expiration() says the delay has
 323      not yet expired then an autosuspend is scheduled for the appropriate time
 324      and 0 is returned
 325
 326  int pm_runtime_resume(struct device *dev);
 327    - execute the subsystem-level resume callback for the device; returns 0 on
 328      success, 1 if the device's runtime PM status was already 'active' or
 329      error code on failure, where -EAGAIN means it may be safe to attempt to
 330      resume the device again in future, but 'power.runtime_error' should be
 331      checked additionally, and -EACCES means that 'power.disable_depth' is
 332      different from 0
 333
 334  int pm_request_idle(struct device *dev);
 335    - submit a request to execute the subsystem-level idle callback for the
 336      device (the request is represented by a work item in pm_wq); returns 0 on
 337      success or error code if the request has not been queued up
 338
 339  int pm_request_autosuspend(struct device *dev);
 340    - schedule the execution of the subsystem-level suspend callback for the
 341      device when the autosuspend delay has expired; if the delay has already
 342      expired then the work item is queued up immediately
 343
 344  int pm_schedule_suspend(struct device *dev, unsigned int delay);
 345    - schedule the execution of the subsystem-level suspend callback for the
 346      device in future, where 'delay' is the time to wait before queuing up a
 347      suspend work item in pm_wq, in milliseconds (if 'delay' is zero, the work
 348      item is queued up immediately); returns 0 on success, 1 if the device's PM
 349      runtime status was already 'suspended', or error code if the request
 350      hasn't been scheduled (or queued up if 'delay' is 0); if the execution of
 351      ->runtime_suspend() is already scheduled and not yet expired, the new
 352      value of 'delay' will be used as the time to wait
 353
 354  int pm_request_resume(struct device *dev);
 355    - submit a request to execute the subsystem-level resume callback for the
 356      device (the request is represented by a work item in pm_wq); returns 0 on
 357      success, 1 if the device's runtime PM status was already 'active', or
 358      error code if the request hasn't been queued up
 359
 360  void pm_runtime_get_noresume(struct device *dev);
 361    - increment the device's usage counter
 362
 363  int pm_runtime_get(struct device *dev);
 364    - increment the device's usage counter, run pm_request_resume(dev) and
 365      return its result
 366
 367  int pm_runtime_get_sync(struct device *dev);
 368    - increment the device's usage counter, run pm_runtime_resume(dev) and
 369      return its result
 370
 371  void pm_runtime_put_noidle(struct device *dev);
 372    - decrement the device's usage counter
 373
 374  int pm_runtime_put(struct device *dev);
 375    - decrement the device's usage counter; if the result is 0 then run
 376      pm_request_idle(dev) and return its result
 377
 378  int pm_runtime_put_autosuspend(struct device *dev);
 379    - decrement the device's usage counter; if the result is 0 then run
 380      pm_request_autosuspend(dev) and return its result
 381
 382  int pm_runtime_put_sync(struct device *dev);
 383    - decrement the device's usage counter; if the result is 0 then run
 384      pm_runtime_idle(dev) and return its result
 385
 386  int pm_runtime_put_sync_suspend(struct device *dev);
 387    - decrement the device's usage counter; if the result is 0 then run
 388      pm_runtime_suspend(dev) and return its result
 389
 390  int pm_runtime_put_sync_autosuspend(struct device *dev);
 391    - decrement the device's usage counter; if the result is 0 then run
 392      pm_runtime_autosuspend(dev) and return its result
 393
 394  void pm_runtime_enable(struct device *dev);
 395    - decrement the device's 'power.disable_depth' field; if that field is equal
 396      to zero, the runtime PM helper functions can execute subsystem-level
 397      callbacks described in Section 2 for the device
 398
 399  int pm_runtime_disable(struct device *dev);
 400    - increment the device's 'power.disable_depth' field (if the value of that
 401      field was previously zero, this prevents subsystem-level runtime PM
 402      callbacks from being run for the device), make sure that all of the pending
 403      runtime PM operations on the device are either completed or canceled;
 404      returns 1 if there was a resume request pending and it was necessary to
 405      execute the subsystem-level resume callback for the device to satisfy that
 406      request, otherwise 0 is returned
 407
 408  int pm_runtime_barrier(struct device *dev);
 409    - check if there's a resume request pending for the device and resume it
 410      (synchronously) in that case, cancel any other pending runtime PM requests
 411      regarding it and wait for all runtime PM operations on it in progress to
 412      complete; returns 1 if there was a resume request pending and it was
 413      necessary to execute the subsystem-level resume callback for the device to
 414      satisfy that request, otherwise 0 is returned
 415
 416  void pm_suspend_ignore_children(struct device *dev, bool enable);
 417    - set/unset the power.ignore_children flag of the device
 418
 419  int pm_runtime_set_active(struct device *dev);
 420    - clear the device's 'power.runtime_error' flag, set the device's runtime
 421      PM status to 'active' and update its parent's counter of 'active'
 422      children as appropriate (it is only valid to use this function if
 423      'power.runtime_error' is set or 'power.disable_depth' is greater than
 424      zero); it will fail and return error code if the device has a parent
 425      which is not active and the 'power.ignore_children' flag of which is unset
 426
 427  void pm_runtime_set_suspended(struct device *dev);
 428    - clear the device's 'power.runtime_error' flag, set the device's runtime
 429      PM status to 'suspended' and update its parent's counter of 'active'
 430      children as appropriate (it is only valid to use this function if
 431      'power.runtime_error' is set or 'power.disable_depth' is greater than
 432      zero)
 433
 434  bool pm_runtime_active(struct device *dev);
 435    - return true if the device's runtime PM status is 'active' or its
 436      'power.disable_depth' field is not equal to zero, or false otherwise
 437
 438  bool pm_runtime_suspended(struct device *dev);
 439    - return true if the device's runtime PM status is 'suspended' and its
 440      'power.disable_depth' field is equal to zero, or false otherwise
 441
 442  bool pm_runtime_status_suspended(struct device *dev);
 443    - return true if the device's runtime PM status is 'suspended'
 444
 445  void pm_runtime_allow(struct device *dev);
 446    - set the power.runtime_auto flag for the device and decrease its usage
 447      counter (used by the /sys/devices/.../power/control interface to
 448      effectively allow the device to be power managed at run time)
 449
 450  void pm_runtime_forbid(struct device *dev);
 451    - unset the power.runtime_auto flag for the device and increase its usage
 452      counter (used by the /sys/devices/.../power/control interface to
 453      effectively prevent the device from being power managed at run time)
 454
 455  void pm_runtime_no_callbacks(struct device *dev);
 456    - set the power.no_callbacks flag for the device and remove the runtime
 457      PM attributes from /sys/devices/.../power (or prevent them from being
 458      added when the device is registered)
 459
 460  void pm_runtime_irq_safe(struct device *dev);
 461    - set the power.irq_safe flag for the device, causing the runtime-PM
 462      callbacks to be invoked with interrupts off
 463
 464  void pm_runtime_mark_last_busy(struct device *dev);
 465    - set the power.last_busy field to the current time
 466
 467  void pm_runtime_use_autosuspend(struct device *dev);
 468    - set the power.use_autosuspend flag, enabling autosuspend delays
 469
 470  void pm_runtime_dont_use_autosuspend(struct device *dev);
 471    - clear the power.use_autosuspend flag, disabling autosuspend delays
 472
 473  void pm_runtime_set_autosuspend_delay(struct device *dev, int delay);
 474    - set the power.autosuspend_delay value to 'delay' (expressed in
 475      milliseconds); if 'delay' is negative then runtime suspends are
 476      prevented
 477
 478  unsigned long pm_runtime_autosuspend_expiration(struct device *dev);
 479    - calculate the time when the current autosuspend delay period will expire,
 480      based on power.last_busy and power.autosuspend_delay; if the delay time
 481      is 1000 ms or larger then the expiration time is rounded up to the
 482      nearest second; returns 0 if the delay period has already expired or
 483      power.use_autosuspend isn't set, otherwise returns the expiration time
 484      in jiffies
 485
 486It is safe to execute the following helper functions from interrupt context:
 487
 488pm_request_idle()
 489pm_request_autosuspend()
 490pm_schedule_suspend()
 491pm_request_resume()
 492pm_runtime_get_noresume()
 493pm_runtime_get()
 494pm_runtime_put_noidle()
 495pm_runtime_put()
 496pm_runtime_put_autosuspend()
 497pm_runtime_enable()
 498pm_suspend_ignore_children()
 499pm_runtime_set_active()
 500pm_runtime_set_suspended()
 501pm_runtime_suspended()
 502pm_runtime_mark_last_busy()
 503pm_runtime_autosuspend_expiration()
 504
 505If pm_runtime_irq_safe() has been called for a device then the following helper
 506functions may also be used in interrupt context:
 507
 508pm_runtime_idle()
 509pm_runtime_suspend()
 510pm_runtime_autosuspend()
 511pm_runtime_resume()
 512pm_runtime_get_sync()
 513pm_runtime_put_sync()
 514pm_runtime_put_sync_suspend()
 515pm_runtime_put_sync_autosuspend()
 516
 5175. Runtime PM Initialization, Device Probing and Removal
 518
 519Initially, the runtime PM is disabled for all devices, which means that the
 520majority of the runtime PM helper funtions described in Section 4 will return
 521-EAGAIN until pm_runtime_enable() is called for the device.
 522
 523In addition to that, the initial runtime PM status of all devices is
 524'suspended', but it need not reflect the actual physical state of the device.
 525Thus, if the device is initially active (i.e. it is able to process I/O), its
 526runtime PM status must be changed to 'active', with the help of
 527pm_runtime_set_active(), before pm_runtime_enable() is called for the device.
 528
 529However, if the device has a parent and the parent's runtime PM is enabled,
 530calling pm_runtime_set_active() for the device will affect the parent, unless
 531the parent's 'power.ignore_children' flag is set.  Namely, in that case the
 532parent won't be able to suspend at run time, using the PM core's helper
 533functions, as long as the child's status is 'active', even if the child's
 534runtime PM is still disabled (i.e. pm_runtime_enable() hasn't been called for
 535the child yet or pm_runtime_disable() has been called for it).  For this reason,
 536once pm_runtime_set_active() has been called for the device, pm_runtime_enable()
 537should be called for it too as soon as reasonably possible or its runtime PM
 538status should be changed back to 'suspended' with the help of
 539pm_runtime_set_suspended().
 540
 541If the default initial runtime PM status of the device (i.e. 'suspended')
 542reflects the actual state of the device, its bus type's or its driver's
 543->probe() callback will likely need to wake it up using one of the PM core's
 544helper functions described in Section 4.  In that case, pm_runtime_resume()
 545should be used.  Of course, for this purpose the device's runtime PM has to be
 546enabled earlier by calling pm_runtime_enable().
 547
 548If the device bus type's or driver's ->probe() callback runs
 549pm_runtime_suspend() or pm_runtime_idle() or their asynchronous counterparts,
 550they will fail returning -EAGAIN, because the device's usage counter is
 551incremented by the driver core before executing ->probe().  Still, it may be
 552desirable to suspend the device as soon as ->probe() has finished, so the driver
 553core uses pm_runtime_put_sync() to invoke the subsystem-level idle callback for
 554the device at that time.
 555
 556Moreover, the driver core prevents runtime PM callbacks from racing with the bus
 557notifier callback in __device_release_driver(), which is necessary, because the
 558notifier is used by some subsystems to carry out operations affecting the
 559runtime PM functionality.  It does so by calling pm_runtime_get_sync() before
 560driver_sysfs_remove() and the BUS_NOTIFY_UNBIND_DRIVER notifications.  This
 561resumes the device if it's in the suspended state and prevents it from
 562being suspended again while those routines are being executed.
 563
 564To allow bus types and drivers to put devices into the suspended state by
 565calling pm_runtime_suspend() from their ->remove() routines, the driver core
 566executes pm_runtime_put_sync() after running the BUS_NOTIFY_UNBIND_DRIVER
 567notifications in __device_release_driver().  This requires bus types and
 568drivers to make their ->remove() callbacks avoid races with runtime PM directly,
 569but also it allows of more flexibility in the handling of devices during the
 570removal of their drivers.
 571
 572The user space can effectively disallow the driver of the device to power manage
 573it at run time by changing the value of its /sys/devices/.../power/control
 574attribute to "on", which causes pm_runtime_forbid() to be called.  In principle,
 575this mechanism may also be used by the driver to effectively turn off the
 576runtime power management of the device until the user space turns it on.
 577Namely, during the initialization the driver can make sure that the runtime PM
 578status of the device is 'active' and call pm_runtime_forbid().  It should be
 579noted, however, that if the user space has already intentionally changed the
 580value of /sys/devices/.../power/control to "auto" to allow the driver to power
 581manage the device at run time, the driver may confuse it by using
 582pm_runtime_forbid() this way.
 583
 5846. Runtime PM and System Sleep
 585
 586Runtime PM and system sleep (i.e., system suspend and hibernation, also known
 587as suspend-to-RAM and suspend-to-disk) interact with each other in a couple of
 588ways.  If a device is active when a system sleep starts, everything is
 589straightforward.  But what should happen if the device is already suspended?
 590
 591The device may have different wake-up settings for runtime PM and system sleep.
 592For example, remote wake-up may be enabled for runtime suspend but disallowed
 593for system sleep (device_may_wakeup(dev) returns 'false').  When this happens,
 594the subsystem-level system suspend callback is responsible for changing the
 595device's wake-up setting (it may leave that to the device driver's system
 596suspend routine).  It may be necessary to resume the device and suspend it again
 597in order to do so.  The same is true if the driver uses different power levels
 598or other settings for runtime suspend and system sleep.
 599
 600During system resume, the simplest approach is to bring all devices back to full
 601power, even if they had been suspended before the system suspend began.  There
 602are several reasons for this, including:
 603
 604  * The device might need to switch power levels, wake-up settings, etc.
 605
 606  * Remote wake-up events might have been lost by the firmware.
 607
 608  * The device's children may need the device to be at full power in order
 609    to resume themselves.
 610
 611  * The driver's idea of the device state may not agree with the device's
 612    physical state.  This can happen during resume from hibernation.
 613
 614  * The device might need to be reset.
 615
 616  * Even though the device was suspended, if its usage counter was > 0 then most
 617    likely it would need a runtime resume in the near future anyway.
 618
 619If the device had been suspended before the system suspend began and it's
 620brought back to full power during resume, then its runtime PM status will have
 621to be updated to reflect the actual post-system sleep status.  The way to do
 622this is:
 623
 624        pm_runtime_disable(dev);
 625        pm_runtime_set_active(dev);
 626        pm_runtime_enable(dev);
 627
 628The PM core always increments the runtime usage counter before calling the
 629->suspend() callback and decrements it after calling the ->resume() callback.
 630Hence disabling runtime PM temporarily like this will not cause any runtime
 631suspend attempts to be permanently lost.  If the usage count goes to zero
 632following the return of the ->resume() callback, the ->runtime_idle() callback
 633will be invoked as usual.
 634
 635On some systems, however, system sleep is not entered through a global firmware
 636or hardware operation.  Instead, all hardware components are put into low-power
 637states directly by the kernel in a coordinated way.  Then, the system sleep
 638state effectively follows from the states the hardware components end up in
 639and the system is woken up from that state by a hardware interrupt or a similar
 640mechanism entirely under the kernel's control.  As a result, the kernel never
 641gives control away and the states of all devices during resume are precisely
 642known to it.  If that is the case and none of the situations listed above takes
 643place (in particular, if the system is not waking up from hibernation), it may
 644be more efficient to leave the devices that had been suspended before the system
 645suspend began in the suspended state.
 646
 647The PM core does its best to reduce the probability of race conditions between
 648the runtime PM and system suspend/resume (and hibernation) callbacks by carrying
 649out the following operations:
 650
 651  * During system suspend it calls pm_runtime_get_noresume() and
 652    pm_runtime_barrier() for every device right before executing the
 653    subsystem-level .suspend() callback for it.  In addition to that it calls
 654    __pm_runtime_disable() with 'false' as the second argument for every device
 655    right before executing the subsystem-level .suspend_late() callback for it.
 656
 657  * During system resume it calls pm_runtime_enable() and pm_runtime_put_sync()
 658    for every device right after executing the subsystem-level .resume_early()
 659    callback and right after executing the subsystem-level .resume() callback
 660    for it, respectively.
 661
 6627. Generic subsystem callbacks
 663
 664Subsystems may wish to conserve code space by using the set of generic power
 665management callbacks provided by the PM core, defined in
 666driver/base/power/generic_ops.c:
 667
 668  int pm_generic_runtime_suspend(struct device *dev);
 669    - invoke the ->runtime_suspend() callback provided by the driver of this
 670      device and return its result, or return -EINVAL if not defined
 671
 672  int pm_generic_runtime_resume(struct device *dev);
 673    - invoke the ->runtime_resume() callback provided by the driver of this
 674      device and return its result, or return -EINVAL if not defined
 675
 676  int pm_generic_suspend(struct device *dev);
 677    - if the device has not been suspended at run time, invoke the ->suspend()
 678      callback provided by its driver and return its result, or return 0 if not
 679      defined
 680
 681  int pm_generic_suspend_noirq(struct device *dev);
 682    - if pm_runtime_suspended(dev) returns "false", invoke the ->suspend_noirq()
 683      callback provided by the device's driver and return its result, or return
 684      0 if not defined
 685
 686  int pm_generic_resume(struct device *dev);
 687    - invoke the ->resume() callback provided by the driver of this device and,
 688      if successful, change the device's runtime PM status to 'active'
 689
 690  int pm_generic_resume_noirq(struct device *dev);
 691    - invoke the ->resume_noirq() callback provided by the driver of this device
 692
 693  int pm_generic_freeze(struct device *dev);
 694    - if the device has not been suspended at run time, invoke the ->freeze()
 695      callback provided by its driver and return its result, or return 0 if not
 696      defined
 697
 698  int pm_generic_freeze_noirq(struct device *dev);
 699    - if pm_runtime_suspended(dev) returns "false", invoke the ->freeze_noirq()
 700      callback provided by the device's driver and return its result, or return
 701      0 if not defined
 702
 703  int pm_generic_thaw(struct device *dev);
 704    - if the device has not been suspended at run time, invoke the ->thaw()
 705      callback provided by its driver and return its result, or return 0 if not
 706      defined
 707
 708  int pm_generic_thaw_noirq(struct device *dev);
 709    - if pm_runtime_suspended(dev) returns "false", invoke the ->thaw_noirq()
 710      callback provided by the device's driver and return its result, or return
 711      0 if not defined
 712
 713  int pm_generic_poweroff(struct device *dev);
 714    - if the device has not been suspended at run time, invoke the ->poweroff()
 715      callback provided by its driver and return its result, or return 0 if not
 716      defined
 717
 718  int pm_generic_poweroff_noirq(struct device *dev);
 719    - if pm_runtime_suspended(dev) returns "false", run the ->poweroff_noirq()
 720      callback provided by the device's driver and return its result, or return
 721      0 if not defined
 722
 723  int pm_generic_restore(struct device *dev);
 724    - invoke the ->restore() callback provided by the driver of this device and,
 725      if successful, change the device's runtime PM status to 'active'
 726
 727  int pm_generic_restore_noirq(struct device *dev);
 728    - invoke the ->restore_noirq() callback provided by the device's driver
 729
 730These functions can be assigned to the ->runtime_idle(), ->runtime_suspend(),
 731->runtime_resume(), ->suspend(), ->suspend_noirq(), ->resume(),
 732->resume_noirq(), ->freeze(), ->freeze_noirq(), ->thaw(), ->thaw_noirq(),
 733->poweroff(), ->poweroff_noirq(), ->restore(), ->restore_noirq() callback
 734pointers in the subsystem-level dev_pm_ops structures.
 735
 736If a subsystem wishes to use all of them at the same time, it can simply assign
 737the GENERIC_SUBSYS_PM_OPS macro, defined in include/linux/pm.h, to its
 738dev_pm_ops structure pointer.
 739
 740Device drivers that wish to use the same function as a system suspend, freeze,
 741poweroff and runtime suspend callback, and similarly for system resume, thaw,
 742restore, and runtime resume, can achieve this with the help of the
 743UNIVERSAL_DEV_PM_OPS macro defined in include/linux/pm.h (possibly setting its
 744last argument to NULL).
 745
 7468. "No-Callback" Devices
 747
 748Some "devices" are only logical sub-devices of their parent and cannot be
 749power-managed on their own.  (The prototype example is a USB interface.  Entire
 750USB devices can go into low-power mode or send wake-up requests, but neither is
 751possible for individual interfaces.)  The drivers for these devices have no
 752need of runtime PM callbacks; if the callbacks did exist, ->runtime_suspend()
 753and ->runtime_resume() would always return 0 without doing anything else and
 754->runtime_idle() would always call pm_runtime_suspend().
 755
 756Subsystems can tell the PM core about these devices by calling
 757pm_runtime_no_callbacks().  This should be done after the device structure is
 758initialized and before it is registered (although after device registration is
 759also okay).  The routine will set the device's power.no_callbacks flag and
 760prevent the non-debugging runtime PM sysfs attributes from being created.
 761
 762When power.no_callbacks is set, the PM core will not invoke the
 763->runtime_idle(), ->runtime_suspend(), or ->runtime_resume() callbacks.
 764Instead it will assume that suspends and resumes always succeed and that idle
 765devices should be suspended.
 766
 767As a consequence, the PM core will never directly inform the device's subsystem
 768or driver about runtime power changes.  Instead, the driver for the device's
 769parent must take responsibility for telling the device's driver when the
 770parent's power state changes.
 771
 7729. Autosuspend, or automatically-delayed suspends
 773
 774Changing a device's power state isn't free; it requires both time and energy.
 775A device should be put in a low-power state only when there's some reason to
 776think it will remain in that state for a substantial time.  A common heuristic
 777says that a device which hasn't been used for a while is liable to remain
 778unused; following this advice, drivers should not allow devices to be suspended
 779at runtime until they have been inactive for some minimum period.  Even when
 780the heuristic ends up being non-optimal, it will still prevent devices from
 781"bouncing" too rapidly between low-power and full-power states.
 782
 783The term "autosuspend" is an historical remnant.  It doesn't mean that the
 784device is automatically suspended (the subsystem or driver still has to call
 785the appropriate PM routines); rather it means that runtime suspends will
 786automatically be delayed until the desired period of inactivity has elapsed.
 787
 788Inactivity is determined based on the power.last_busy field.  Drivers should
 789call pm_runtime_mark_last_busy() to update this field after carrying out I/O,
 790typically just before calling pm_runtime_put_autosuspend().  The desired length
 791of the inactivity period is a matter of policy.  Subsystems can set this length
 792initially by calling pm_runtime_set_autosuspend_delay(), but after device
 793registration the length should be controlled by user space, using the
 794/sys/devices/.../power/autosuspend_delay_ms attribute.
 795
 796In order to use autosuspend, subsystems or drivers must call
 797pm_runtime_use_autosuspend() (preferably before registering the device), and
 798thereafter they should use the various *_autosuspend() helper functions instead
 799of the non-autosuspend counterparts:
 800
 801        Instead of: pm_runtime_suspend    use: pm_runtime_autosuspend;
 802        Instead of: pm_schedule_suspend   use: pm_request_autosuspend;
 803        Instead of: pm_runtime_put        use: pm_runtime_put_autosuspend;
 804        Instead of: pm_runtime_put_sync   use: pm_runtime_put_sync_autosuspend.
 805
 806Drivers may also continue to use the non-autosuspend helper functions; they
 807will behave normally, not taking the autosuspend delay into account.
 808Similarly, if the power.use_autosuspend field isn't set then the autosuspend
 809helper functions will behave just like the non-autosuspend counterparts.
 810
 811Under some circumstances a driver or subsystem may want to prevent a device
 812from autosuspending immediately, even though the usage counter is zero and the
 813autosuspend delay time has expired.  If the ->runtime_suspend() callback
 814returns -EAGAIN or -EBUSY, and if the next autosuspend delay expiration time is
 815in the future (as it normally would be if the callback invoked
 816pm_runtime_mark_last_busy()), the PM core will automatically reschedule the
 817autosuspend.  The ->runtime_suspend() callback can't do this rescheduling
 818itself because no suspend requests of any kind are accepted while the device is
 819suspending (i.e., while the callback is running).
 820
 821The implementation is well suited for asynchronous use in interrupt contexts.
 822However such use inevitably involves races, because the PM core can't
 823synchronize ->runtime_suspend() callbacks with the arrival of I/O requests.
 824This synchronization must be handled by the driver, using its private lock.
 825Here is a schematic pseudo-code example:
 826
 827        foo_read_or_write(struct foo_priv *foo, void *data)
 828        {
 829                lock(&foo->private_lock);
 830                add_request_to_io_queue(foo, data);
 831                if (foo->num_pending_requests++ == 0)
 832                        pm_runtime_get(&foo->dev);
 833                if (!foo->is_suspended)
 834                        foo_process_next_request(foo);
 835                unlock(&foo->private_lock);
 836        }
 837
 838        foo_io_completion(struct foo_priv *foo, void *req)
 839        {
 840                lock(&foo->private_lock);
 841                if (--foo->num_pending_requests == 0) {
 842                        pm_runtime_mark_last_busy(&foo->dev);
 843                        pm_runtime_put_autosuspend(&foo->dev);
 844                } else {
 845                        foo_process_next_request(foo);
 846                }
 847                unlock(&foo->private_lock);
 848                /* Send req result back to the user ... */
 849        }
 850
 851        int foo_runtime_suspend(struct device *dev)
 852        {
 853                struct foo_priv foo = container_of(dev, ...);
 854                int ret = 0;
 855
 856                lock(&foo->private_lock);
 857                if (foo->num_pending_requests > 0) {
 858                        ret = -EBUSY;
 859                } else {
 860                        /* ... suspend the device ... */
 861                        foo->is_suspended = 1;
 862                }
 863                unlock(&foo->private_lock);
 864                return ret;
 865        }
 866
 867        int foo_runtime_resume(struct device *dev)
 868        {
 869                struct foo_priv foo = container_of(dev, ...);
 870
 871                lock(&foo->private_lock);
 872                /* ... resume the device ... */
 873                foo->is_suspended = 0;
 874                pm_runtime_mark_last_busy(&foo->dev);
 875                if (foo->num_pending_requests > 0)
 876                        foo_process_requests(foo);
 877                unlock(&foo->private_lock);
 878                return 0;
 879        }
 880
 881The important point is that after foo_io_completion() asks for an autosuspend,
 882the foo_runtime_suspend() callback may race with foo_read_or_write().
 883Therefore foo_runtime_suspend() has to check whether there are any pending I/O
 884requests (while holding the private lock) before allowing the suspend to
 885proceed.
 886
 887In addition, the power.autosuspend_delay field can be changed by user space at
 888any time.  If a driver cares about this, it can call
 889pm_runtime_autosuspend_expiration() from within the ->runtime_suspend()
 890callback while holding its private lock.  If the function returns a nonzero
 891value then the delay has not yet expired and the callback should return
 892-EAGAIN.
 893
lxr.linux.no kindly hosted by Redpill Linpro AS, provider of Linux consulting and operations services since 1995.