Documentation/power/runtime_pm.rst

   1 ==================================================
   2 Runtime Power Management Framework for I/O Devices
   3 ==================================================
   4
   5 (C) 2009-2011 Rafael J. Wysocki <rjw@sisk.pl>, Novell Inc.
   6
   7 (C) 2010 Alan Stern <stern@rowland.harvard.edu>
   8
   9 (C) 2014 Intel Corp., Rafael J. Wysocki <rafael.j.wysocki@intel.com>
  10
  11 1. Introduction
  12 ===============
  13
  14 Support for runtime power management (runtime PM) of I/O devices is provided
  15 at the power management core (PM core) level by means of:
  16
  17 * The power management workqueue pm_wq in which bus types and device drivers can
  18   put their PM-related work items.  It is strongly recommended that pm_wq be
  19   used for queuing all work items related to runtime PM, because this allows
  20   them to be synchronized with system-wide power transitions (suspend to RAM,
  21   hibernation and resume from system sleep states).  pm_wq is declared in
  22   include/linux/pm_runtime.h and defined in kernel/power/main.c.
  23
  24 * A number of runtime PM fields in the 'power' member of 'struct device' (which
  25   is of the type 'struct dev_pm_info', defined in include/linux/pm.h) that can
  26   be used for synchronizing runtime PM operations with one another.
  27
  28 * Three device runtime PM callbacks in 'struct dev_pm_ops' (defined in
  29   include/linux/pm.h).
  30
  31 * A set of helper functions defined in drivers/base/power/runtime.c that can be
  32   used for carrying out runtime PM operations in such a way that the
  33   synchronization between them is taken care of by the PM core.  Bus types and
  34   device drivers are encouraged to use these functions.
  35
  36 The runtime PM callbacks present in 'struct dev_pm_ops', the device runtime PM
  37 fields of 'struct dev_pm_info' and the core helper functions provided for
  38 runtime PM are described below.
  39
  40 2. Device Runtime PM Callbacks
  41 ==============================
  42
  43 There are three device runtime PM callbacks defined in 'struct dev_pm_ops'::
  44
  45   struct dev_pm_ops {
  46         ...
  47         int (*runtime_suspend)(struct device *dev);
  48         int (*runtime_resume)(struct device *dev);
  49         int (*runtime_idle)(struct device *dev);
  50         ...
  51   };
  52
  53 The ->runtime_suspend(), ->runtime_resume() and ->runtime_idle() callbacks
  54 are executed by the PM core for the device's subsystem that may be either of
  55 the following:
  56
  57   1. PM domain of the device, if the device's PM domain object, dev->pm_domain,
  58      is present.
  59
  60   2. Device type of the device, if both dev->type and dev->type->pm are present.
  61
  62   3. Device class of the device, if both dev->class and dev->class->pm are
  63      present.
  64
  65   4. Bus type of the device, if both dev->bus and dev->bus->pm are present.
  66
  67 If the subsystem chosen by applying the above rules doesn't provide the relevant
  68 callback, the PM core will invoke the corresponding driver callback stored in
  69 dev->driver->pm directly (if present).
  70
  71 The PM core always checks which callback to use in the order given above, so the
  72 priority order of callbacks from high to low is: PM domain, device type, class
  73 and bus type.  Moreover, the high-priority one will always take precedence over
  74 a low-priority one.  The PM domain, bus type, device type and class callbacks
  75 are referred to as subsystem-level callbacks in what follows.
  76
  77 By default, the callbacks are always invoked in process context with interrupts
  78 enabled.  However, the pm_runtime_irq_safe() helper function can be used to tell
  79 the PM core that it is safe to run the ->runtime_suspend(), ->runtime_resume()
  80 and ->runtime_idle() callbacks for the given device in atomic context with
  81 interrupts disabled.  This implies that the callback routines in question must
  82 not block or sleep, but it also means that the synchronous helper functions
  83 listed at the end of Section 4 may be used for that device within an interrupt
  84 handler or generally in an atomic context.
  85
  86 The subsystem-level suspend callback, if present, is _entirely_ _responsible_
  87 for handling the suspend of the device as appropriate, which may, but need not
  88 include executing the device driver's own ->runtime_suspend() callback (from the
  89 PM core's point of view it is not necessary to implement a ->runtime_suspend()
  90 callback in a device driver as long as the subsystem-level suspend callback
  91 knows what to do to handle the device).
  92
  93   * Once the subsystem-level suspend callback (or the driver suspend callback,
  94     if invoked directly) has completed successfully for the given device, the PM
  95     core regards the device as suspended, which need not mean that it has been
  96     put into a low power state.  It is supposed to mean, however, that the
  97     device will not process data and will not communicate with the CPU(s) and
  98     RAM until the appropriate resume callback is executed for it.  The runtime
  99     PM status of a device after successful execution of the suspend callback is
 100     'suspended'.
 101
 102   * If the suspend callback returns -EBUSY or -EAGAIN, the device's runtime PM
 103     status remains 'active', which means that the device _must_ be fully
 104     operational afterwards.
 105
 106   * If the suspend callback returns an error code different from -EBUSY and
 107     -EAGAIN, the PM core regards this as a fatal error and will refuse to run
 108     the helper functions described in Section 4 for the device until its status
 109     is directly set to  either 'active', or 'suspended' (the PM core provides
 110     special helper functions for this purpose).
 111
 112 In particular, if the driver requires remote wakeup capability (i.e. hardware
 113 mechanism allowing the device to request a change of its power state, such as
 114 PCI PME) for proper functioning and device_can_wakeup() returns 'false' for the
 115 device, then ->runtime_suspend() should return -EBUSY.  On the other hand, if
 116 device_can_wakeup() returns 'true' for the device and the device is put into a
 117 low-power state during the execution of the suspend callback, it is expected
 118 that remote wakeup will be enabled for the device.  Generally, remote wakeup
 119 should be enabled for all input devices put into low-power states at run time.
 120
 121 The subsystem-level resume callback, if present, is **entirely responsible** for
 122 handling the resume of the device as appropriate, which may, but need not
 123 include executing the device driver's own ->runtime_resume() callback (from the
 124 PM core's point of view it is not necessary to implement a ->runtime_resume()
 125 callback in a device driver as long as the subsystem-level resume callback knows
 126 what to do to handle the device).
 127
 128   * Once the subsystem-level resume callback (or the driver resume callback, if
 129     invoked directly) has completed successfully, the PM core regards the device
 130     as fully operational, which means that the device _must_ be able to complete
 131     I/O operations as needed.  The runtime PM status of the device is then
 132     'active'.
 133
 134   * If the resume callback returns an error code, the PM core regards this as a
 135     fatal error and will refuse to run the helper functions described in Section
 136     4 for the device, until its status is directly set to either 'active', or
 137     'suspended' (by means of special helper functions provided by the PM core
 138     for this purpose).
 139
 140 The idle callback (a subsystem-level one, if present, or the driver one) is
 141 executed by the PM core whenever the device appears to be idle, which is
 142 indicated to the PM core by two counters, the device's usage counter and the
 143 counter of 'active' children of the device.
 144
 145   * If any of these counters is decreased using a helper function provided by
 146     the PM core and it turns out to be equal to zero, the other counter is
 147     checked.  If that counter also is equal to zero, the PM core executes the
 148     idle callback with the device as its argument.
 149
 150 The action performed by the idle callback is totally dependent on the subsystem
 151 (or driver) in question, but the expected and recommended action is to check
 152 if the device can be suspended (i.e. if all of the conditions necessary for
 153 suspending the device are satisfied) and to queue up a suspend request for the
 154 device in that case.  If there is no idle callback, or if the callback returns
 155 0, then the PM core will attempt to carry out a runtime suspend of the device,
 156 also respecting devices configured for autosuspend.  In essence this means a
 157 call to __pm_runtime_autosuspend() (do note that drivers needs to update the
 158 device last busy mark, pm_runtime_mark_last_busy(), to control the delay under
 159 this circumstance).  To prevent this (for example, if the callback routine has
 160 started a delayed suspend), the routine must return a non-zero value.  Negative
 161 error return codes are ignored by the PM core.
 162
 163 The helper functions provided by the PM core, described in Section 4, guarantee
 164 that the following constraints are met with respect to runtime PM callbacks for
 165 one device:
 166
 167 (1) The callbacks are mutually exclusive (e.g. it is forbidden to execute
 168     ->runtime_suspend() in parallel with ->runtime_resume() or with another
 169     instance of ->runtime_suspend() for the same device) with the exception that
 170     ->runtime_suspend() or ->runtime_resume() can be executed in parallel with
 171     ->runtime_idle() (although ->runtime_idle() will not be started while any
 172     of the other callbacks is being executed for the same device).
 173
 174 (2) ->runtime_idle() and ->runtime_suspend() can only be executed for 'active'
 175     devices (i.e. the PM core will only execute ->runtime_idle() or
 176     ->runtime_suspend() for the devices the runtime PM status of which is
 177     'active').
 178
 179 (3) ->runtime_idle() and ->runtime_suspend() can only be executed for a device
 180     the usage counter of which is equal to zero _and_ either the counter of
 181     'active' children of which is equal to zero, or the 'power.ignore_children'
 182     flag of which is set.
 183
 184 (4) ->runtime_resume() can only be executed for 'suspended' devices  (i.e. the
 185     PM core will only execute ->runtime_resume() for the devices the runtime
 186     PM status of which is 'suspended').
 187
 188 Additionally, the helper functions provided by the PM core obey the following
 189 rules:
 190
 191   * If ->runtime_suspend() is about to be executed or there's a pending request
 192     to execute it, ->runtime_idle() will not be executed for the same device.
 193
 194   * A request to execute or to schedule the execution of ->runtime_suspend()
 195     will cancel any pending requests to execute ->runtime_idle() for the same
 196     device.
 197
 198   * If ->runtime_resume() is about to be executed or there's a pending request
 199     to execute it, the other callbacks will not be executed for the same device.
 200
 201   * A request to execute ->runtime_resume() will cancel any pending or
 202     scheduled requests to execute the other callbacks for the same device,
 203     except for scheduled autosuspends.
 204
 205 3. Runtime PM Device Fields
 206 ===========================
 207
 208 The following device runtime PM fields are present in 'struct dev_pm_info', as
 209 defined in include/linux/pm.h:
 210
 211   `struct timer_list suspend_timer;`
 212     - timer used for scheduling (delayed) suspend and autosuspend requests
 213
 214   `unsigned long timer_expires;`
 215     - timer expiration time, in jiffies (if this is different from zero, the
 216       timer is running and will expire at that time, otherwise the timer is not
 217       running)
 218
 219   `struct work_struct work;`
 220     - work structure used for queuing up requests (i.e. work items in pm_wq)
 221
 222   `wait_queue_head_t wait_queue;`
 223     - wait queue used if any of the helper functions needs to wait for another
 224       one to complete
 225
 226   `spinlock_t lock;`
 227     - lock used for synchronization
 228
 229   `atomic_t usage_count;`
 230     - the usage counter of the device
 231
 232   `atomic_t child_count;`
 233     - the count of 'active' children of the device
 234
 235   `unsigned int ignore_children;`
 236     - if set, the value of child_count is ignored (but still updated)
 237
 238   `unsigned int disable_depth;`
 239     - used for disabling the helper functions (they work normally if this is
 240       equal to zero); the initial value of it is 1 (i.e. runtime PM is
 241       initially disabled for all devices)
 242
 243   `int runtime_error;`
 244     - if set, there was a fatal error (one of the callbacks returned error code
 245       as described in Section 2), so the helper functions will not work until
 246       this flag is cleared; this is the error code returned by the failing
 247       callback
 248
 249   `unsigned int idle_notification;`
 250     - if set, ->runtime_idle() is being executed
 251
 252   `unsigned int request_pending;`
 253     - if set, there's a pending request (i.e. a work item queued up into pm_wq)
 254
 255   `enum rpm_request request;`
 256     - type of request that's pending (valid if request_pending is set)
 257
 258   `unsigned int deferred_resume;`
 259     - set if ->runtime_resume() is about to be run while ->runtime_suspend() is
 260       being executed for that device and it is not practical to wait for the
 261       suspend to complete; means "start a resume as soon as you've suspended"
 262
 263   `enum rpm_status runtime_status;`
 264     - the runtime PM status of the device; this field's initial value is
 265       RPM_SUSPENDED, which means that each device is initially regarded by the
 266       PM core as 'suspended', regardless of its real hardware status
 267
 268   `enum rpm_status last_status;`
 269     - the last runtime PM status of the device captured before disabling runtime
 270       PM for it (invalid initially and when disable_depth is 0)
 271
 272   `unsigned int runtime_auto;`
 273     - if set, indicates that the user space has allowed the device driver to
 274       power manage the device at run time via the /sys/devices/.../power/control
 275       `interface;` it may only be modified with the help of the
 276       pm_runtime_allow() and pm_runtime_forbid() helper functions
 277
 278   `unsigned int no_callbacks;`
 279     - indicates that the device does not use the runtime PM callbacks (see
 280       Section 8); it may be modified only by the pm_runtime_no_callbacks()
 281       helper function
 282
 283   `unsigned int irq_safe;`
 284     - indicates that the ->runtime_suspend() and ->runtime_resume() callbacks
 285       will be invoked with the spinlock held and interrupts disabled
 286
 287   `unsigned int use_autosuspend;`
 288     - indicates that the device's driver supports delayed autosuspend (see
 289       Section 9); it may be modified only by the
 290       pm_runtime{_dont}_use_autosuspend() helper functions
 291
 292   `unsigned int timer_autosuspends;`
 293     - indicates that the PM core should attempt to carry out an autosuspend
 294       when the timer expires rather than a normal suspend
 295
 296   `int autosuspend_delay;`
 297     - the delay time (in milliseconds) to be used for autosuspend
 298
 299   `unsigned long last_busy;`
 300     - the time (in jiffies) when the pm_runtime_mark_last_busy() helper
 301       function was last called for this device; used in calculating inactivity
 302       periods for autosuspend
 303
 304 All of the above fields are members of the 'power' member of 'struct device'.
 305
 306 4. Runtime PM Device Helper Functions
 307 =====================================
 308
 309 The following runtime PM helper functions are defined in
 310 drivers/base/power/runtime.c and include/linux/pm_runtime.h:
 311
 312   `void pm_runtime_init(struct device *dev);`
 313     - initialize the device runtime PM fields in 'struct dev_pm_info'
 314
 315   `void pm_runtime_remove(struct device *dev);`
 316     - make sure that the runtime PM of the device will be disabled after
 317       removing the device from device hierarchy
 318
 319   `int pm_runtime_idle(struct device *dev);`
 320     - execute the subsystem-level idle callback for the device; returns an
 321       error code on failure, where -EINPROGRESS means that ->runtime_idle() is
 322       already being executed; if there is no callback or the callback returns 0
 323       then run pm_runtime_autosuspend(dev) and return its result
 324
 325   `int pm_runtime_suspend(struct device *dev);`
 326     - execute the subsystem-level suspend callback for the device; returns 0 on
 327       success, 1 if the device's runtime PM status was already 'suspended', or
 328       error code on failure, where -EAGAIN or -EBUSY means it is safe to attempt
 329       to suspend the device again in future and -EACCES means that
 330       'power.disable_depth' is different from 0
 331
 332   `int pm_runtime_autosuspend(struct device *dev);`
 333     - same as pm_runtime_suspend() except that the autosuspend delay is taken
 334       `into account;` if pm_runtime_autosuspend_expiration() says the delay has
 335       not yet expired then an autosuspend is scheduled for the appropriate time
 336       and 0 is returned
 337
 338   `int pm_runtime_resume(struct device *dev);`
 339     - execute the subsystem-level resume callback for the device; returns 0 on
 340       success, 1 if the device's runtime PM status is already 'active' (also if
 341       'power.disable_depth' is nonzero, but the status was 'active' when it was
 342       changing from 0 to 1) or error code on failure, where -EAGAIN means it may
 343       be safe to attempt to resume the device again in future, but
 344       'power.runtime_error' should be checked additionally, and -EACCES means
 345       that the callback could not be run, because 'power.disable_depth' was
 346       different from 0
 347
 348   `int pm_runtime_resume_and_get(struct device *dev);`
 349     - run pm_runtime_resume(dev) and if successful, increment the device's
 350       usage counter; return the result of pm_runtime_resume
 351
 352   `int pm_request_idle(struct device *dev);`
 353     - submit a request to execute the subsystem-level idle callback for the
 354       device (the request is represented by a work item in pm_wq); returns 0 on
 355       success or error code if the request has not been queued up
 356
 357   `int pm_request_autosuspend(struct device *dev);`
 358     - schedule the execution of the subsystem-level suspend callback for the
 359       device when the autosuspend delay has expired; if the delay has already
 360       expired then the work item is queued up immediately
 361
 362   `int pm_schedule_suspend(struct device *dev, unsigned int delay);`
 363     - schedule the execution of the subsystem-level suspend callback for the
 364       device in future, where 'delay' is the time to wait before queuing up a
 365       suspend work item in pm_wq, in milliseconds (if 'delay' is zero, the work
 366       item is queued up immediately); returns 0 on success, 1 if the device's PM
 367       runtime status was already 'suspended', or error code if the request
 368       hasn't been scheduled (or queued up if 'delay' is 0); if the execution of
 369       ->runtime_suspend() is already scheduled and not yet expired, the new
 370       value of 'delay' will be used as the time to wait
 371
 372   `int pm_request_resume(struct device *dev);`
 373     - submit a request to execute the subsystem-level resume callback for the
 374       device (the request is represented by a work item in pm_wq); returns 0 on
 375       success, 1 if the device's runtime PM status was already 'active', or
 376       error code if the request hasn't been queued up
 377
 378   `void pm_runtime_get_noresume(struct device *dev);`
 379     - increment the device's usage counter
 380
 381   `int pm_runtime_get(struct device *dev);`
 382     - increment the device's usage counter, run pm_request_resume(dev) and
 383       return its result
 384
 385   `int pm_runtime_get_sync(struct device *dev);`
 386     - increment the device's usage counter, run pm_runtime_resume(dev) and
 387       return its result;
 388       note that it does not drop the device's usage counter on errors, so
 389       consider using pm_runtime_resume_and_get() instead of it, especially
 390       if its return value is checked by the caller, as this is likely to
 391       result in cleaner code.
 392
 393   `int pm_runtime_get_if_in_use(struct device *dev);`
 394     - return -EINVAL if 'power.disable_depth' is nonzero; otherwise, if the
 395       runtime PM status is RPM_ACTIVE and the runtime PM usage counter is
 396       nonzero, increment the counter and return 1; otherwise return 0 without
 397       changing the counter
 398
 399   `int pm_runtime_get_if_active(struct device *dev);`
 400     - return -EINVAL if 'power.disable_depth' is nonzero; otherwise, if the
 401       runtime PM status is RPM_ACTIVE, increment the counter and
 402       return 1; otherwise return 0 without changing the counter
 403
 404   `void pm_runtime_put_noidle(struct device *dev);`
 405     - decrement the device's usage counter
 406
 407   `int pm_runtime_put(struct device *dev);`
 408     - decrement the device's usage counter; if the result is 0 then run
 409       pm_request_idle(dev) and return its result
 410
 411   `int pm_runtime_put_autosuspend(struct device *dev);`
 412     - does the same as __pm_runtime_put_autosuspend() for now, but in the
 413       future, will also call pm_runtime_mark_last_busy() as well, DO NOT USE!
 414
 415   `int __pm_runtime_put_autosuspend(struct device *dev);`
 416     - decrement the device's usage counter; if the result is 0 then run
 417       pm_request_autosuspend(dev) and return its result
 418
 419   `int pm_runtime_put_sync(struct device *dev);`
 420     - decrement the device's usage counter; if the result is 0 then run
 421       pm_runtime_idle(dev) and return its result
 422
 423   `int pm_runtime_put_sync_suspend(struct device *dev);`
 424     - decrement the device's usage counter; if the result is 0 then run
 425       pm_runtime_suspend(dev) and return its result
 426
 427   `int pm_runtime_put_sync_autosuspend(struct device *dev);`
 428     - decrement the device's usage counter; if the result is 0 then run
 429       pm_runtime_autosuspend(dev) and return its result
 430
 431   `void pm_runtime_enable(struct device *dev);`
 432     - decrement the device's 'power.disable_depth' field; if that field is equal
 433       to zero, the runtime PM helper functions can execute subsystem-level
 434       callbacks described in Section 2 for the device
 435
 436   `int pm_runtime_disable(struct device *dev);`
 437     - increment the device's 'power.disable_depth' field (if the value of that
 438       field was previously zero, this prevents subsystem-level runtime PM
 439       callbacks from being run for the device), make sure that all of the
 440       pending runtime PM operations on the device are either completed or
 441       canceled; returns 1 if there was a resume request pending and it was
 442       necessary to execute the subsystem-level resume callback for the device
 443       to satisfy that request, otherwise 0 is returned
 444
 445   `int pm_runtime_barrier(struct device *dev);`
 446     - check if there's a resume request pending for the device and resume it
 447       (synchronously) in that case, cancel any other pending runtime PM requests
 448       regarding it and wait for all runtime PM operations on it in progress to
 449       complete; returns 1 if there was a resume request pending and it was
 450       necessary to execute the subsystem-level resume callback for the device to
 451       satisfy that request, otherwise 0 is returned
 452
 453   `void pm_suspend_ignore_children(struct device *dev, bool enable);`
 454     - set/unset the power.ignore_children flag of the device
 455
 456   `int pm_runtime_set_active(struct device *dev);`
 457     - clear the device's 'power.runtime_error' flag, set the device's runtime
 458       PM status to 'active' and update its parent's counter of 'active'
 459       children as appropriate (it is only valid to use this function if
 460       'power.runtime_error' is set or 'power.disable_depth' is greater than
 461       zero); it will fail and return error code if the device has a parent
 462       which is not active and the 'power.ignore_children' flag of which is unset
 463
 464   `void pm_runtime_set_suspended(struct device *dev);`
 465     - clear the device's 'power.runtime_error' flag, set the device's runtime
 466       PM status to 'suspended' and update its parent's counter of 'active'
 467       children as appropriate (it is only valid to use this function if
 468       'power.runtime_error' is set or 'power.disable_depth' is greater than
 469       zero)
 470
 471   `bool pm_runtime_active(struct device *dev);`
 472     - return true if the device's runtime PM status is 'active' or its
 473       'power.disable_depth' field is not equal to zero, or false otherwise
 474
 475   `bool pm_runtime_suspended(struct device *dev);`
 476     - return true if the device's runtime PM status is 'suspended' and its
 477       'power.disable_depth' field is equal to zero, or false otherwise
 478
 479   `bool pm_runtime_status_suspended(struct device *dev);`
 480     - return true if the device's runtime PM status is 'suspended'
 481
 482   `void pm_runtime_allow(struct device *dev);`
 483     - set the power.runtime_auto flag for the device and decrease its usage
 484       counter (used by the /sys/devices/.../power/control interface to
 485       effectively allow the device to be power managed at run time)
 486
 487   `void pm_runtime_forbid(struct device *dev);`
 488     - unset the power.runtime_auto flag for the device and increase its usage
 489       counter (used by the /sys/devices/.../power/control interface to
 490       effectively prevent the device from being power managed at run time)
 491
 492   `void pm_runtime_no_callbacks(struct device *dev);`
 493     - set the power.no_callbacks flag for the device and remove the runtime
 494       PM attributes from /sys/devices/.../power (or prevent them from being
 495       added when the device is registered)
 496
 497   `void pm_runtime_irq_safe(struct device *dev);`
 498     - set the power.irq_safe flag for the device, causing the runtime-PM
 499       callbacks to be invoked with interrupts off
 500
 501   `bool pm_runtime_is_irq_safe(struct device *dev);`
 502     - return true if power.irq_safe flag was set for the device, causing
 503       the runtime-PM callbacks to be invoked with interrupts off
 504
 505   `void pm_runtime_mark_last_busy(struct device *dev);`
 506     - set the power.last_busy field to the current time
 507
 508   `void pm_runtime_use_autosuspend(struct device *dev);`
 509     - set the power.use_autosuspend flag, enabling autosuspend delays; call
 510       pm_runtime_get_sync if the flag was previously cleared and
 511       power.autosuspend_delay is negative
 512
 513   `void pm_runtime_dont_use_autosuspend(struct device *dev);`
 514     - clear the power.use_autosuspend flag, disabling autosuspend delays;
 515       decrement the device's usage counter if the flag was previously set and
 516       power.autosuspend_delay is negative; call pm_runtime_idle
 517
 518   `void pm_runtime_set_autosuspend_delay(struct device *dev, int delay);`
 519     - set the power.autosuspend_delay value to 'delay' (expressed in
 520       milliseconds); if 'delay' is negative then runtime suspends are
 521       prevented; if power.use_autosuspend is set, pm_runtime_get_sync may be
 522       called or the device's usage counter may be decremented and
 523       pm_runtime_idle called depending on if power.autosuspend_delay is
 524       changed to or from a negative value; if power.use_autosuspend is clear,
 525       pm_runtime_idle is called
 526
 527   `unsigned long pm_runtime_autosuspend_expiration(struct device *dev);`
 528     - calculate the time when the current autosuspend delay period will expire,
 529       based on power.last_busy and power.autosuspend_delay; if the delay time
 530       is 1000 ms or larger then the expiration time is rounded up to the
 531       nearest second; returns 0 if the delay period has already expired or
 532       power.use_autosuspend isn't set, otherwise returns the expiration time
 533       in jiffies
 534
 535 It is safe to execute the following helper functions from interrupt context:
 536
 537 - pm_request_idle()
 538 - pm_request_autosuspend()
 539 - pm_schedule_suspend()
 540 - pm_request_resume()
 541 - pm_runtime_get_noresume()
 542 - pm_runtime_get()
 543 - pm_runtime_put_noidle()
 544 - pm_runtime_put()
 545 - pm_runtime_put_autosuspend()
 546 - __pm_runtime_put_autosuspend()
 547 - pm_runtime_enable()
 548 - pm_suspend_ignore_children()
 549 - pm_runtime_set_active()
 550 - pm_runtime_set_suspended()
 551 - pm_runtime_suspended()
 552 - pm_runtime_mark_last_busy()
 553 - pm_runtime_autosuspend_expiration()
 554
 555 If pm_runtime_irq_safe() has been called for a device then the following helper
 556 functions may also be used in interrupt context:
 557
 558 - pm_runtime_idle()
 559 - pm_runtime_suspend()
 560 - pm_runtime_autosuspend()
 561 - pm_runtime_resume()
 562 - pm_runtime_get_sync()
 563 - pm_runtime_put_sync()
 564 - pm_runtime_put_sync_suspend()
 565 - pm_runtime_put_sync_autosuspend()
 566
 567 5. Runtime PM Initialization, Device Probing and Removal
 568 ========================================================
 569
 570 Initially, the runtime PM is disabled for all devices, which means that the
 571 majority of the runtime PM helper functions described in Section 4 will return
 572 -EAGAIN until pm_runtime_enable() is called for the device.
 573
 574 In addition to that, the initial runtime PM status of all devices is
 575 'suspended', but it need not reflect the actual physical state of the device.
 576 Thus, if the device is initially active (i.e. it is able to process I/O), its
 577 runtime PM status must be changed to 'active', with the help of
 578 pm_runtime_set_active(), before pm_runtime_enable() is called for the device.
 579
 580 However, if the device has a parent and the parent's runtime PM is enabled,
 581 calling pm_runtime_set_active() for the device will affect the parent, unless
 582 the parent's 'power.ignore_children' flag is set.  Namely, in that case the
 583 parent won't be able to suspend at run time, using the PM core's helper
 584 functions, as long as the child's status is 'active', even if the child's
 585 runtime PM is still disabled (i.e. pm_runtime_enable() hasn't been called for
 586 the child yet or pm_runtime_disable() has been called for it).  For this reason,
 587 once pm_runtime_set_active() has been called for the device, pm_runtime_enable()
 588 should be called for it too as soon as reasonably possible or its runtime PM
 589 status should be changed back to 'suspended' with the help of
 590 pm_runtime_set_suspended().
 591
 592 If the default initial runtime PM status of the device (i.e. 'suspended')
 593 reflects the actual state of the device, its bus type's or its driver's
 594 ->probe() callback will likely need to wake it up using one of the PM core's
 595 helper functions described in Section 4.  In that case, pm_runtime_resume()
 596 should be used.  Of course, for this purpose the device's runtime PM has to be
 597 enabled earlier by calling pm_runtime_enable().
 598
 599 Note, if the device may execute pm_runtime calls during the probe (such as
 600 if it is registered with a subsystem that may call back in) then the
 601 pm_runtime_get_sync() call paired with a pm_runtime_put() call will be
 602 appropriate to ensure that the device is not put back to sleep during the
 603 probe. This can happen with systems such as the network device layer.
 604
 605 It may be desirable to suspend the device once ->probe() has finished.
 606 Therefore the driver core uses the asynchronous pm_request_idle() to submit a
 607 request to execute the subsystem-level idle callback for the device at that
 608 time.  A driver that makes use of the runtime autosuspend feature may want to
 609 update the last busy mark before returning from ->probe().
 610
 611 Moreover, the driver core prevents runtime PM callbacks from racing with the bus
 612 notifier callback in __device_release_driver(), which is necessary because the
 613 notifier is used by some subsystems to carry out operations affecting the
 614 runtime PM functionality.  It does so by calling pm_runtime_get_sync() before
 615 driver_sysfs_remove() and the BUS_NOTIFY_UNBIND_DRIVER notifications.  This
 616 resumes the device if it's in the suspended state and prevents it from
 617 being suspended again while those routines are being executed.
 618
 619 To allow bus types and drivers to put devices into the suspended state by
 620 calling pm_runtime_suspend() from their ->remove() routines, the driver core
 621 executes pm_runtime_put_sync() after running the BUS_NOTIFY_UNBIND_DRIVER
 622 notifications in __device_release_driver().  This requires bus types and
 623 drivers to make their ->remove() callbacks avoid races with runtime PM directly,
 624 but it also allows more flexibility in the handling of devices during the
 625 removal of their drivers.
 626
 627 Drivers in ->remove() callback should undo the runtime PM changes done
 628 in ->probe(). Usually this means calling pm_runtime_disable(),
 629 pm_runtime_dont_use_autosuspend() etc.
 630
 631 The user space can effectively disallow the driver of the device to power manage
 632 it at run time by changing the value of its /sys/devices/.../power/control
 633 attribute to "on", which causes pm_runtime_forbid() to be called.  In principle,
 634 this mechanism may also be used by the driver to effectively turn off the
 635 runtime power management of the device until the user space turns it on.
 636 Namely, during the initialization the driver can make sure that the runtime PM
 637 status of the device is 'active' and call pm_runtime_forbid().  It should be
 638 noted, however, that if the user space has already intentionally changed the
 639 value of /sys/devices/.../power/control to "auto" to allow the driver to power
 640 manage the device at run time, the driver may confuse it by using
 641 pm_runtime_forbid() this way.
 642
 643 6. Runtime PM and System Sleep
 644 ==============================
 645
 646 Runtime PM and system sleep (i.e., system suspend and hibernation, also known
 647 as suspend-to-RAM and suspend-to-disk) interact with each other in a couple of
 648 ways.  If a device is active when a system sleep starts, everything is
 649 straightforward.  But what should happen if the device is already suspended?
 650
 651 The device may have different wake-up settings for runtime PM and system sleep.
 652 For example, remote wake-up may be enabled for runtime suspend but disallowed
 653 for system sleep (device_may_wakeup(dev) returns 'false').  When this happens,
 654 the subsystem-level system suspend callback is responsible for changing the
 655 device's wake-up setting (it may leave that to the device driver's system
 656 suspend routine).  It may be necessary to resume the device and suspend it again
 657 in order to do so.  The same is true if the driver uses different power levels
 658 or other settings for runtime suspend and system sleep.
 659
 660 During system resume, the simplest approach is to bring all devices back to full
 661 power, even if they had been suspended before the system suspend began.  There
 662 are several reasons for this, including:
 663
 664   * The device might need to switch power levels, wake-up settings, etc.
 665
 666   * Remote wake-up events might have been lost by the firmware.
 667
 668   * The device's children may need the device to be at full power in order
 669     to resume themselves.
 670
 671   * The driver's idea of the device state may not agree with the device's
 672     physical state.  This can happen during resume from hibernation.
 673
 674   * The device might need to be reset.
 675
 676   * Even though the device was suspended, if its usage counter was > 0 then most
 677     likely it would need a runtime resume in the near future anyway.
 678
 679 If the device had been suspended before the system suspend began and it's
 680 brought back to full power during resume, then its runtime PM status will have
 681 to be updated to reflect the actual post-system sleep status.  The way to do
 682 this is:
 683
 684          - pm_runtime_disable(dev);
 685          - pm_runtime_set_active(dev);
 686          - pm_runtime_enable(dev);
 687
 688 The PM core always increments the runtime usage counter before calling the
 689 ->suspend() callback and decrements it after calling the ->resume() callback.
 690 Hence disabling runtime PM temporarily like this will not cause any runtime
 691 suspend attempts to be permanently lost.  If the usage count goes to zero
 692 following the return of the ->resume() callback, the ->runtime_idle() callback
 693 will be invoked as usual.
 694
 695 On some systems, however, system sleep is not entered through a global firmware
 696 or hardware operation.  Instead, all hardware components are put into low-power
 697 states directly by the kernel in a coordinated way.  Then, the system sleep
 698 state effectively follows from the states the hardware components end up in
 699 and the system is woken up from that state by a hardware interrupt or a similar
 700 mechanism entirely under the kernel's control.  As a result, the kernel never
 701 gives control away and the states of all devices during resume are precisely
 702 known to it.  If that is the case and none of the situations listed above takes
 703 place (in particular, if the system is not waking up from hibernation), it may
 704 be more efficient to leave the devices that had been suspended before the system
 705 suspend began in the suspended state.
 706
 707 To this end, the PM core provides a mechanism allowing some coordination between
 708 different levels of device hierarchy.  Namely, if a system suspend .prepare()
 709 callback returns a positive number for a device, that indicates to the PM core
 710 that the device appears to be runtime-suspended and its state is fine, so it
 711 may be left in runtime suspend provided that all of its descendants are also
 712 left in runtime suspend.  If that happens, the PM core will not execute any
 713 system suspend and resume callbacks for all of those devices, except for the
 714 .complete() callback, which is then entirely responsible for handling the device
 715 as appropriate.  This only applies to system suspend transitions that are not
 716 related to hibernation (see Documentation/driver-api/pm/devices.rst for more
 717 information).
 718
 719 The PM core does its best to reduce the probability of race conditions between
 720 the runtime PM and system suspend/resume (and hibernation) callbacks by carrying
 721 out the following operations:
 722
 723   * During system suspend pm_runtime_get_noresume() is called for every device
 724     right before executing the subsystem-level .prepare() callback for it and
 725     pm_runtime_barrier() is called for every device right before executing the
 726     subsystem-level .suspend() callback for it.  In addition to that the PM core
 727     calls __pm_runtime_disable() with 'false' as the second argument for every
 728     device right before executing the subsystem-level .suspend_late() callback
 729     for it.
 730
 731   * During system resume pm_runtime_enable() and pm_runtime_put() are called for
 732     every device right after executing the subsystem-level .resume_early()
 733     callback and right after executing the subsystem-level .complete() callback
 734     for it, respectively.
 735
 736 7. Generic subsystem callbacks
 737 ==============================
 738
 739 Subsystems may wish to conserve code space by using the set of generic power
 740 management callbacks provided by the PM core, defined in
 741 driver/base/power/generic_ops.c:
 742
 743   `int pm_generic_runtime_suspend(struct device *dev);`
 744     - invoke the ->runtime_suspend() callback provided by the driver of this
 745       device and return its result, or return 0 if not defined
 746
 747   `int pm_generic_runtime_resume(struct device *dev);`
 748     - invoke the ->runtime_resume() callback provided by the driver of this
 749       device and return its result, or return 0 if not defined
 750
 751   `int pm_generic_suspend(struct device *dev);`
 752     - if the device has not been suspended at run time, invoke the ->suspend()
 753       callback provided by its driver and return its result, or return 0 if not
 754       defined
 755
 756   `int pm_generic_suspend_noirq(struct device *dev);`
 757     - if pm_runtime_suspended(dev) returns "false", invoke the ->suspend_noirq()
 758       callback provided by the device's driver and return its result, or return
 759       0 if not defined
 760
 761   `int pm_generic_resume(struct device *dev);`
 762     - invoke the ->resume() callback provided by the driver of this device and,
 763       if successful, change the device's runtime PM status to 'active'
 764
 765   `int pm_generic_resume_noirq(struct device *dev);`
 766     - invoke the ->resume_noirq() callback provided by the driver of this device
 767
 768   `int pm_generic_freeze(struct device *dev);`
 769     - if the device has not been suspended at run time, invoke the ->freeze()
 770       callback provided by its driver and return its result, or return 0 if not
 771       defined
 772
 773   `int pm_generic_freeze_noirq(struct device *dev);`
 774     - if pm_runtime_suspended(dev) returns "false", invoke the ->freeze_noirq()
 775       callback provided by the device's driver and return its result, or return
 776       0 if not defined
 777
 778   `int pm_generic_thaw(struct device *dev);`
 779     - if the device has not been suspended at run time, invoke the ->thaw()
 780       callback provided by its driver and return its result, or return 0 if not
 781       defined
 782
 783   `int pm_generic_thaw_noirq(struct device *dev);`
 784     - if pm_runtime_suspended(dev) returns "false", invoke the ->thaw_noirq()
 785       callback provided by the device's driver and return its result, or return
 786       0 if not defined
 787
 788   `int pm_generic_poweroff(struct device *dev);`
 789     - if the device has not been suspended at run time, invoke the ->poweroff()
 790       callback provided by its driver and return its result, or return 0 if not
 791       defined
 792
 793   `int pm_generic_poweroff_noirq(struct device *dev);`
 794     - if pm_runtime_suspended(dev) returns "false", run the ->poweroff_noirq()
 795       callback provided by the device's driver and return its result, or return
 796       0 if not defined
 797
 798   `int pm_generic_restore(struct device *dev);`
 799     - invoke the ->restore() callback provided by the driver of this device and,
 800       if successful, change the device's runtime PM status to 'active'
 801
 802   `int pm_generic_restore_noirq(struct device *dev);`
 803     - invoke the ->restore_noirq() callback provided by the device's driver
 804
 805 These functions are the defaults used by the PM core if a subsystem doesn't
 806 provide its own callbacks for ->runtime_idle(), ->runtime_suspend(),
 807 ->runtime_resume(), ->suspend(), ->suspend_noirq(), ->resume(),
 808 ->resume_noirq(), ->freeze(), ->freeze_noirq(), ->thaw(), ->thaw_noirq(),
 809 ->poweroff(), ->poweroff_noirq(), ->restore(), ->restore_noirq() in the
 810 subsystem-level dev_pm_ops structure.
 811
 812 Device drivers that wish to use the same function as a system suspend, freeze,
 813 poweroff and runtime suspend callback, and similarly for system resume, thaw,
 814 restore, and runtime resume, can achieve this with the help of the
 815 UNIVERSAL_DEV_PM_OPS macro defined in include/linux/pm.h (possibly setting its
 816 last argument to NULL).
 817
 818 8. "No-Callback" Devices
 819 ========================
 820
 821 Some "devices" are only logical sub-devices of their parent and cannot be
 822 power-managed on their own.  (The prototype example is a USB interface.  Entire
 823 USB devices can go into low-power mode or send wake-up requests, but neither is
 824 possible for individual interfaces.)  The drivers for these devices have no
 825 need of runtime PM callbacks; if the callbacks did exist, ->runtime_suspend()
 826 and ->runtime_resume() would always return 0 without doing anything else and
 827 ->runtime_idle() would always call pm_runtime_suspend().
 828
 829 Subsystems can tell the PM core about these devices by calling
 830 pm_runtime_no_callbacks().  This should be done after the device structure is
 831 initialized and before it is registered (although after device registration is
 832 also okay).  The routine will set the device's power.no_callbacks flag and
 833 prevent the non-debugging runtime PM sysfs attributes from being created.
 834
 835 When power.no_callbacks is set, the PM core will not invoke the
 836 ->runtime_idle(), ->runtime_suspend(), or ->runtime_resume() callbacks.
 837 Instead it will assume that suspends and resumes always succeed and that idle
 838 devices should be suspended.
 839
 840 As a consequence, the PM core will never directly inform the device's subsystem
 841 or driver about runtime power changes.  Instead, the driver for the device's
 842 parent must take responsibility for telling the device's driver when the
 843 parent's power state changes.
 844
 845 Note that, in some cases it may not be desirable for subsystems/drivers to call
 846 pm_runtime_no_callbacks() for their devices. This could be because a subset of
 847 the runtime PM callbacks needs to be implemented, a platform dependent PM
 848 domain could get attached to the device or that the device is power managed
 849 through a supplier device link. For these reasons and to avoid boilerplate code
 850 in subsystems/drivers, the PM core allows runtime PM callbacks to be
 851 unassigned. More precisely, if a callback pointer is NULL, the PM core will act
 852 as though there was a callback and it returned 0.
 853
 854 9. Autosuspend, or automatically-delayed suspends
 855 =================================================
 856
 857 Changing a device's power state isn't free; it requires both time and energy.
 858 A device should be put in a low-power state only when there's some reason to
 859 think it will remain in that state for a substantial time.  A common heuristic
 860 says that a device which hasn't been used for a while is liable to remain
 861 unused; following this advice, drivers should not allow devices to be suspended
 862 at runtime until they have been inactive for some minimum period.  Even when
 863 the heuristic ends up being non-optimal, it will still prevent devices from
 864 "bouncing" too rapidly between low-power and full-power states.
 865
 866 The term "autosuspend" is an historical remnant.  It doesn't mean that the
 867 device is automatically suspended (the subsystem or driver still has to call
 868 the appropriate PM routines); rather it means that runtime suspends will
 869 automatically be delayed until the desired period of inactivity has elapsed.
 870
 871 Inactivity is determined based on the power.last_busy field.  Drivers should
 872 call pm_runtime_mark_last_busy() to update this field after carrying out I/O,
 873 typically just before calling __pm_runtime_put_autosuspend().  The desired
 874 length of the inactivity period is a matter of policy.  Subsystems can set this
 875 length initially by calling pm_runtime_set_autosuspend_delay(), but after device
 876 registration the length should be controlled by user space, using the
 877 /sys/devices/.../power/autosuspend_delay_ms attribute.
 878
 879 In order to use autosuspend, subsystems or drivers must call
 880 pm_runtime_use_autosuspend() (preferably before registering the device), and
 881 thereafter they should use the various `*_autosuspend()` helper functions
 882 instead of the non-autosuspend counterparts::
 883
 884         Instead of: pm_runtime_suspend    use: pm_runtime_autosuspend;
 885         Instead of: pm_schedule_suspend   use: pm_request_autosuspend;
 886         Instead of: pm_runtime_put        use: __pm_runtime_put_autosuspend;
 887         Instead of: pm_runtime_put_sync   use: pm_runtime_put_sync_autosuspend.
 888
 889 Drivers may also continue to use the non-autosuspend helper functions; they
 890 will behave normally, which means sometimes taking the autosuspend delay into
 891 account (see pm_runtime_idle).
 892
 893 Under some circumstances a driver or subsystem may want to prevent a device
 894 from autosuspending immediately, even though the usage counter is zero and the
 895 autosuspend delay time has expired.  If the ->runtime_suspend() callback
 896 returns -EAGAIN or -EBUSY, and if the next autosuspend delay expiration time is
 897 in the future (as it normally would be if the callback invoked
 898 pm_runtime_mark_last_busy()), the PM core will automatically reschedule the
 899 autosuspend.  The ->runtime_suspend() callback can't do this rescheduling
 900 itself because no suspend requests of any kind are accepted while the device is
 901 suspending (i.e., while the callback is running).
 902
 903 The implementation is well suited for asynchronous use in interrupt contexts.
 904 However such use inevitably involves races, because the PM core can't
 905 synchronize ->runtime_suspend() callbacks with the arrival of I/O requests.
 906 This synchronization must be handled by the driver, using its private lock.
 907 Here is a schematic pseudo-code example::
 908
 909         foo_read_or_write(struct foo_priv *foo, void *data)
 910         {
 911                 lock(&foo->private_lock);
 912                 add_request_to_io_queue(foo, data);
 913                 if (foo->num_pending_requests++ == 0)
 914                         pm_runtime_get(&foo->dev);
 915                 if (!foo->is_suspended)
 916                         foo_process_next_request(foo);
 917                 unlock(&foo->private_lock);
 918         }
 919
 920         foo_io_completion(struct foo_priv *foo, void *req)
 921         {
 922                 lock(&foo->private_lock);
 923                 if (--foo->num_pending_requests == 0) {
 924                         pm_runtime_mark_last_busy(&foo->dev);
 925                         __pm_runtime_put_autosuspend(&foo->dev);
 926                 } else {
 927                         foo_process_next_request(foo);
 928                 }
 929                 unlock(&foo->private_lock);
 930                 /* Send req result back to the user ... */
 931         }
 932
 933         int foo_runtime_suspend(struct device *dev)
 934         {
 935                 struct foo_priv foo = container_of(dev, ...);
 936                 int ret = 0;
 937
 938                 lock(&foo->private_lock);
 939                 if (foo->num_pending_requests > 0) {
 940                         ret = -EBUSY;
 941                 } else {
 942                         /* ... suspend the device ... */
 943                         foo->is_suspended = 1;
 944                 }
 945                 unlock(&foo->private_lock);
 946                 return ret;
 947         }
 948
 949         int foo_runtime_resume(struct device *dev)
 950         {
 951                 struct foo_priv foo = container_of(dev, ...);
 952
 953                 lock(&foo->private_lock);
 954                 /* ... resume the device ... */
 955                 foo->is_suspended = 0;
 956                 pm_runtime_mark_last_busy(&foo->dev);
 957                 if (foo->num_pending_requests > 0)
 958                         foo_process_next_request(foo);
 959                 unlock(&foo->private_lock);
 960                 return 0;
 961         }
 962
 963 The important point is that after foo_io_completion() asks for an autosuspend,
 964 the foo_runtime_suspend() callback may race with foo_read_or_write().
 965 Therefore foo_runtime_suspend() has to check whether there are any pending I/O
 966 requests (while holding the private lock) before allowing the suspend to
 967 proceed.
 968
 969 In addition, the power.autosuspend_delay field can be changed by user space at
 970 any time.  If a driver cares about this, it can call
 971 pm_runtime_autosuspend_expiration() from within the ->runtime_suspend()
 972 callback while holding its private lock.  If the function returns a nonzero
 973 value then the delay has not yet expired and the callback should return
 974 -EAGAIN.