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(C) 2014 Intel Corp., Rafael J. Wysocki <rafael.j.wysocki@intel.com> 6 71. Introduction 8 9Support for runtime power management (runtime PM) of I/O devices is provided 10at the power management core (PM core) level by means of: 11 12* The power management workqueue pm_wq in which bus types and device drivers can 13 put their PM-related work items. It is strongly recommended that pm_wq be 14 used for queuing all work items related to runtime PM, because this allows 15 them to be synchronized with system-wide power transitions (suspend to RAM, 16 hibernation and resume from system sleep states). pm_wq is declared in 17 include/linux/pm_runtime.h and defined in kernel/power/main.c. 18 19* A number of runtime PM fields in the 'power' member of 'struct device' (which 20 is of the type 'struct dev_pm_info', defined in include/linux/pm.h) that can 21 be used for synchronizing runtime PM operations with one another. 22 23* Three device runtime PM callbacks in 'struct dev_pm_ops' (defined in 24 include/linux/pm.h). 25 26* A set of helper functions defined in drivers/base/power/runtime.c that can be 27 used for carrying out runtime PM operations in such a way that the 28 synchronization between them is taken care of by the PM core. Bus types and 29 device drivers are encouraged to use these functions. 30 31The runtime PM callbacks present in 'struct dev_pm_ops', the device runtime PM 32fields of 'struct dev_pm_info' and the core helper functions provided for 33runtime PM are described below. 34 352. Device Runtime PM Callbacks 36 37There are three device runtime PM callbacks defined in 'struct dev_pm_ops': 38 39struct dev_pm_ops { 40 ... 41 int (*runtime_suspend)(struct device *dev); 42 int (*runtime_resume)(struct device *dev); 43 int (*runtime_idle)(struct device *dev); 44 ... 45}; 46 47The ->runtime_suspend(), ->runtime_resume() and ->runtime_idle() callbacks 48are executed by the PM core for the device's subsystem that may be either of 49the following: 50 51 1. PM domain of the device, if the device's PM domain object, dev->pm_domain, 52 is present. 53 54 2. Device type of the device, if both dev->type and dev->type->pm are present. 55 56 3. Device class of the device, if both dev->class and dev->class->pm are 57 present. 58 59 4. Bus type of the device, if both dev->bus and dev->bus->pm are present. 60 61If the subsystem chosen by applying the above rules doesn't provide the relevant 62callback, the PM core will invoke the corresponding driver callback stored in 63dev->driver->pm directly (if present). 64 65The PM core always checks which callback to use in the order given above, so the 66priority order of callbacks from high to low is: PM domain, device type, class 67and bus type. Moreover, the high-priority one will always take precedence over 68a low-priority one. The PM domain, bus type, device type and class callbacks 69are referred to as subsystem-level callbacks in what follows. 70 71By default, the callbacks are always invoked in process context with interrupts 72enabled. However, the pm_runtime_irq_safe() helper function can be used to tell 73the PM core that it is safe to run the ->runtime_suspend(), ->runtime_resume() 74and ->runtime_idle() callbacks for the given device in atomic context with 75interrupts disabled. This implies that the callback routines in question must 76not block or sleep, but it also means that the synchronous helper functions 77listed at the end of Section 4 may be used for that device within an interrupt 78handler or generally in an atomic context. 79 80The subsystem-level suspend callback, if present, is _entirely_ _responsible_ 81for handling the suspend of the device as appropriate, which may, but need not 82include executing the device driver's own ->runtime_suspend() callback (from the 83PM core's point of view it is not necessary to implement a ->runtime_suspend() 84callback in a device driver as long as the subsystem-level suspend callback 85knows what to do to handle the device). 86 87 * Once the subsystem-level suspend callback (or the driver suspend callback, 88 if invoked directly) has completed successfully for the given device, the PM 89 core regards the device as suspended, which need not mean that it has been 90 put into a low power state. It is supposed to mean, however, that the 91 device will not process data and will not communicate with the CPU(s) and 92 RAM until the appropriate resume callback is executed for it. The runtime 93 PM status of a device after successful execution of the suspend callback is 94 'suspended'. 95 96 * If the suspend callback returns -EBUSY or -EAGAIN, the device's runtime PM 97 status remains 'active', which means that the device _must_ be fully 98 operational afterwards. 99 100 * If the suspend callback returns an error code different from -EBUSY and 101 -EAGAIN, the PM core regards this as a fatal error and will refuse to run 102 the helper functions described in Section 4 for the device until its status 103 is directly set to either 'active', or 'suspended' (the PM core provides 104 special helper functions for this purpose). 105 106In particular, if the driver requires remote wakeup capability (i.e. hardware 107mechanism allowing the device to request a change of its power state, such as 108PCI PME) for proper functioning and device_can_wakeup() returns 'false' for the 109device, then ->runtime_suspend() should return -EBUSY. On the other hand, if 110device_can_wakeup() returns 'true' for the device and the device is put into a 111low-power state during the execution of the suspend callback, it is expected 112that remote wakeup will be enabled for the device. Generally, remote wakeup 113should be enabled for all input devices put into low-power states at run time. 114 115The subsystem-level resume callback, if present, is _entirely_ _responsible_ for 116handling the resume of the device as appropriate, which may, but need not 117include executing the device driver's own ->runtime_resume() callback (from the 118PM core's point of view it is not necessary to implement a ->runtime_resume() 119callback in a device driver as long as the subsystem-level resume callback knows 120what to do to handle the device). 121 122 * Once the subsystem-level resume callback (or the driver resume callback, if 123 invoked directly) has completed successfully, the PM core regards the device 124 as fully operational, which means that the device _must_ be able to complete 125 I/O operations as needed. The runtime PM status of the device is then 126 'active'. 127 128 * If the resume callback returns an error code, the PM core regards this as a 129 fatal error and will refuse to run the helper functions described in Section 130 4 for the device, until its status is directly set to either 'active', or 131 'suspended' (by means of special helper functions provided by the PM core 132 for this purpose). 133 134The idle callback (a subsystem-level one, if present, or the driver one) is 135executed by the PM core whenever the device appears to be idle, which is 136indicated to the PM core by two counters, the device's usage counter and the 137counter of 'active' children of the device. 138 139 * If any of these counters is decreased using a helper function provided by 140 the PM core and it turns out to be equal to zero, the other counter is 141 checked. If that counter also is equal to zero, the PM core executes the 142 idle callback with the device as its argument. 143 144The action performed by the idle callback is totally dependent on the subsystem 145(or driver) in question, but the expected and recommended action is to check 146if the device can be suspended (i.e. if all of the conditions necessary for 147suspending the device are satisfied) and to queue up a suspend request for the 148device in that case. If there is no idle callback, or if the callback returns 1490, then the PM core will attempt to carry out a runtime suspend of the device, 150also respecting devices configured for autosuspend. In essence this means a 151call to pm_runtime_autosuspend() (do note that drivers needs to update the 152device last busy mark, pm_runtime_mark_last_busy(), to control the delay under 153this circumstance). To prevent this (for example, if the callback routine has 154started a delayed suspend), the routine must return a non-zero value. Negative 155error return codes are ignored by the PM core. 156 157The helper functions provided by the PM core, described in Section 4, guarantee 158that the following constraints are met with respect to runtime PM callbacks for 159one device: 160 161(1) The callbacks are mutually exclusive (e.g. it is forbidden to execute 162 ->runtime_suspend() in parallel with ->runtime_resume() or with another 163 instance of ->runtime_suspend() for the same device) with the exception that 164 ->runtime_suspend() or ->runtime_resume() can be executed in parallel with 165 ->runtime_idle() (although ->runtime_idle() will not be started while any 166 of the other callbacks is being executed for the same device). 167 168(2) ->runtime_idle() and ->runtime_suspend() can only be executed for 'active' 169 devices (i.e. the PM core will only execute ->runtime_idle() or 170 ->runtime_suspend() for the devices the runtime PM status of which is 171 'active'). 172 173(3) ->runtime_idle() and ->runtime_suspend() can only be executed for a device 174 the usage counter of which is equal to zero _and_ either the counter of 175 'active' children of which is equal to zero, or the 'power.ignore_children' 176 flag of which is set. 177 178(4) ->runtime_resume() can only be executed for 'suspended' devices (i.e. the 179 PM core will only execute ->runtime_resume() for the devices the runtime 180 PM status of which is 'suspended'). 181 182Additionally, the helper functions provided by the PM core obey the following 183rules: 184 185 * If ->runtime_suspend() is about to be executed or there's a pending request 186 to execute it, ->runtime_idle() will not be executed for the same device. 187 188 * A request to execute or to schedule the execution of ->runtime_suspend() 189 will cancel any pending requests to execute ->runtime_idle() for the same 190 device. 191 192 * If ->runtime_resume() is about to be executed or there's a pending request 193 to execute it, the other callbacks will not be executed for the same device. 194 195 * A request to execute ->runtime_resume() will cancel any pending or 196 scheduled requests to execute the other callbacks for the same device, 197 except for scheduled autosuspends. 198 1993. Runtime PM Device Fields 200 201The following device runtime PM fields are present in 'struct dev_pm_info', as 202defined in include/linux/pm.h: 203 204 struct timer_list suspend_timer; 205 - timer used for scheduling (delayed) suspend and autosuspend requests 206 207 unsigned long timer_expires; 208 - timer expiration time, in jiffies (if this is different from zero, the 209 timer is running and will expire at that time, otherwise the timer is not 210 running) 211 212 struct work_struct work; 213 - work structure used for queuing up requests (i.e. work items in pm_wq) 214 215 wait_queue_head_t wait_queue; 216 - wait queue used if any of the helper functions needs to wait for another 217 one to complete 218 219 spinlock_t lock; 220 - lock used for synchronization 221 222 atomic_t usage_count; 223 - the usage counter of the device 224 225 atomic_t child_count; 226 - the count of 'active' children of the device 227 228 unsigned int ignore_children; 229 - if set, the value of child_count is ignored (but still updated) 230 231 unsigned int disable_depth; 232 - used for disabling the helper functions (they work normally if this is 233 equal to zero); the initial value of it is 1 (i.e. runtime PM is 234 initially disabled for all devices) 235 236 int runtime_error; 237 - if set, there was a fatal error (one of the callbacks returned error code 238 as described in Section 2), so the helper functions will not work until 239 this flag is cleared; this is the error code returned by the failing 240 callback 241 242 unsigned int idle_notification; 243 - if set, ->runtime_idle() is being executed 244 245 unsigned int request_pending; 246 - if set, there's a pending request (i.e. a work item queued up into pm_wq) 247 248 enum rpm_request request; 249 - type of request that's pending (valid if request_pending is set) 250 251 unsigned int deferred_resume; 252 - set if ->runtime_resume() is about to be run while ->runtime_suspend() is 253 being executed for that device and it is not practical to wait for the 254 suspend to complete; means "start a resume as soon as you've suspended" 255 256 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_autosuspend(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 int pm_runtime_get_if_in_use(struct device *dev); 372 - return -EINVAL if 'power.disable_depth' is nonzero; otherwise, if the 373 runtime PM status is RPM_ACTIVE and the runtime PM usage counter is 374 nonzero, increment the counter and return 1; otherwise return 0 without 375 changing the counter 376 377 void pm_runtime_put_noidle(struct device *dev); 378 - decrement the device's usage counter 379 380 int pm_runtime_put(struct device *dev); 381 - decrement the device's usage counter; if the result is 0 then run 382 pm_request_idle(dev) and return its result 383 384 int pm_runtime_put_autosuspend(struct device *dev); 385 - decrement the device's usage counter; if the result is 0 then run 386 pm_request_autosuspend(dev) and return its result 387 388 int pm_runtime_put_sync(struct device *dev); 389 - decrement the device's usage counter; if the result is 0 then run 390 pm_runtime_idle(dev) and return its result 391 392 int pm_runtime_put_sync_suspend(struct device *dev); 393 - decrement the device's usage counter; if the result is 0 then run 394 pm_runtime_suspend(dev) and return its result 395 396 int pm_runtime_put_sync_autosuspend(struct device *dev); 397 - decrement the device's usage counter; if the result is 0 then run 398 pm_runtime_autosuspend(dev) and return its result 399 400 void pm_runtime_enable(struct device *dev); 401 - decrement the device's 'power.disable_depth' field; if that field is equal 402 to zero, the runtime PM helper functions can execute subsystem-level 403 callbacks described in Section 2 for the device 404 405 int pm_runtime_disable(struct device *dev); 406 - increment the device's 'power.disable_depth' field (if the value of that 407 field was previously zero, this prevents subsystem-level runtime PM 408 callbacks from being run for the device), make sure that all of the 409 pending runtime PM operations on the device are either completed or 410 canceled; returns 1 if there was a resume request pending and it was 411 necessary to execute the subsystem-level resume callback for the device 412 to satisfy that request, otherwise 0 is returned 413 414 int pm_runtime_barrier(struct device *dev); 415 - check if there's a resume request pending for the device and resume it 416 (synchronously) in that case, cancel any other pending runtime PM requests 417 regarding it and wait for all runtime PM operations on it in progress to 418 complete; returns 1 if there was a resume request pending and it was 419 necessary to execute the subsystem-level resume callback for the device to 420 satisfy that request, otherwise 0 is returned 421 422 void pm_suspend_ignore_children(struct device *dev, bool enable); 423 - set/unset the power.ignore_children flag of the device 424 425 int pm_runtime_set_active(struct device *dev); 426 - clear the device's 'power.runtime_error' flag, set the device's runtime 427 PM status to 'active' and update its parent's counter of 'active' 428 children as appropriate (it is only valid to use this function if 429 'power.runtime_error' is set or 'power.disable_depth' is greater than 430 zero); it will fail and return error code if the device has a parent 431 which is not active and the 'power.ignore_children' flag of which is unset 432 433 void pm_runtime_set_suspended(struct device *dev); 434 - clear the device's 'power.runtime_error' flag, set the device's runtime 435 PM status to 'suspended' and update its parent's counter of 'active' 436 children as appropriate (it is only valid to use this function if 437 'power.runtime_error' is set or 'power.disable_depth' is greater than 438 zero); it will fail and return an error code if the device has a child 439 which is active and the 'power.ignore_children' flag is unset 440 441 bool pm_runtime_active(struct device *dev); 442 - return true if the device's runtime PM status is 'active' or its 443 'power.disable_depth' field is not equal to zero, or false otherwise 444 445 bool pm_runtime_suspended(struct device *dev); 446 - return true if the device's runtime PM status is 'suspended' and its 447 'power.disable_depth' field is equal to zero, or false otherwise 448 449 bool pm_runtime_status_suspended(struct device *dev); 450 - return true if the device's runtime PM status is 'suspended' 451 452 void pm_runtime_allow(struct device *dev); 453 - set the power.runtime_auto flag for the device and decrease its usage 454 counter (used by the /sys/devices/.../power/control interface to 455 effectively allow the device to be power managed at run time) 456 457 void pm_runtime_forbid(struct device *dev); 458 - unset the power.runtime_auto flag for the device and increase its usage 459 counter (used by the /sys/devices/.../power/control interface to 460 effectively prevent the device from being power managed at run time) 461 462 void pm_runtime_no_callbacks(struct device *dev); 463 - set the power.no_callbacks flag for the device and remove the runtime 464 PM attributes from /sys/devices/.../power (or prevent them from being 465 added when the device is registered) 466 467 void pm_runtime_irq_safe(struct device *dev); 468 - set the power.irq_safe flag for the device, causing the runtime-PM 469 callbacks to be invoked with interrupts off 470 471 bool pm_runtime_is_irq_safe(struct device *dev); 472 - return true if power.irq_safe flag was set for the device, causing 473 the runtime-PM callbacks to be invoked with interrupts off 474 475 void pm_runtime_mark_last_busy(struct device *dev); 476 - set the power.last_busy field to the current time 477 478 void pm_runtime_use_autosuspend(struct device *dev); 479 - set the power.use_autosuspend flag, enabling autosuspend delays; call 480 pm_runtime_get_sync if the flag was previously cleared and 481 power.autosuspend_delay is negative 482 483 void pm_runtime_dont_use_autosuspend(struct device *dev); 484 - clear the power.use_autosuspend flag, disabling autosuspend delays; 485 decrement the device's usage counter if the flag was previously set and 486 power.autosuspend_delay is negative; call pm_runtime_idle 487 488 void pm_runtime_set_autosuspend_delay(struct device *dev, int delay); 489 - set the power.autosuspend_delay value to 'delay' (expressed in 490 milliseconds); if 'delay' is negative then runtime suspends are 491 prevented; if power.use_autosuspend is set, pm_runtime_get_sync may be 492 called or the device's usage counter may be decremented and 493 pm_runtime_idle called depending on if power.autosuspend_delay is 494 changed to or from a negative value; if power.use_autosuspend is clear, 495 pm_runtime_idle is called 496 497 unsigned long pm_runtime_autosuspend_expiration(struct device *dev); 498 - calculate the time when the current autosuspend delay period will expire, 499 based on power.last_busy and power.autosuspend_delay; if the delay time 500 is 1000 ms or larger then the expiration time is rounded up to the 501 nearest second; returns 0 if the delay period has already expired or 502 power.use_autosuspend isn't set, otherwise returns the expiration time 503 in jiffies 504 505It is safe to execute the following helper functions from interrupt context: 506 507pm_request_idle() 508pm_request_autosuspend() 509pm_schedule_suspend() 510pm_request_resume() 511pm_runtime_get_noresume() 512pm_runtime_get() 513pm_runtime_put_noidle() 514pm_runtime_put() 515pm_runtime_put_autosuspend() 516pm_runtime_enable() 517pm_suspend_ignore_children() 518pm_runtime_set_active() 519pm_runtime_set_suspended() 520pm_runtime_suspended() 521pm_runtime_mark_last_busy() 522pm_runtime_autosuspend_expiration() 523 524If pm_runtime_irq_safe() has been called for a device then the following helper 525functions may also be used in interrupt context: 526 527pm_runtime_idle() 528pm_runtime_suspend() 529pm_runtime_autosuspend() 530pm_runtime_resume() 531pm_runtime_get_sync() 532pm_runtime_put_sync() 533pm_runtime_put_sync_suspend() 534pm_runtime_put_sync_autosuspend() 535 5365. Runtime PM Initialization, Device Probing and Removal 537 538Initially, the runtime PM is disabled for all devices, which means that the 539majority of the runtime PM helper functions described in Section 4 will return 540-EAGAIN until pm_runtime_enable() is called for the device. 541 542In addition to that, the initial runtime PM status of all devices is 543'suspended', but it need not reflect the actual physical state of the device. 544Thus, if the device is initially active (i.e. it is able to process I/O), its 545runtime PM status must be changed to 'active', with the help of 546pm_runtime_set_active(), before pm_runtime_enable() is called for the device. 547 548However, if the device has a parent and the parent's runtime PM is enabled, 549calling pm_runtime_set_active() for the device will affect the parent, unless 550the parent's 'power.ignore_children' flag is set. Namely, in that case the 551parent won't be able to suspend at run time, using the PM core's helper 552functions, as long as the child's status is 'active', even if the child's 553runtime PM is still disabled (i.e. pm_runtime_enable() hasn't been called for 554the child yet or pm_runtime_disable() has been called for it). For this reason, 555once pm_runtime_set_active() has been called for the device, pm_runtime_enable() 556should be called for it too as soon as reasonably possible or its runtime PM 557status should be changed back to 'suspended' with the help of 558pm_runtime_set_suspended(). 559 560If the default initial runtime PM status of the device (i.e. 'suspended') 561reflects the actual state of the device, its bus type's or its driver's 562->probe() callback will likely need to wake it up using one of the PM core's 563helper functions described in Section 4. In that case, pm_runtime_resume() 564should be used. Of course, for this purpose the device's runtime PM has to be 565enabled earlier by calling pm_runtime_enable(). 566 567Note, if the device may execute pm_runtime calls during the probe (such as 568if it is registers with a subsystem that may call back in) then the 569pm_runtime_get_sync() call paired with a pm_runtime_put() call will be 570appropriate to ensure that the device is not put back to sleep during the 571probe. This can happen with systems such as the network device layer. 572 573It may be desirable to suspend the device once ->probe() has finished. 574Therefore the driver core uses the asynchronous pm_request_idle() to submit a 575request to execute the subsystem-level idle callback for the device at that 576time. A driver that makes use of the runtime autosuspend feature, may want to 577update the last busy mark before returning from ->probe(). 578 579Moreover, the driver core prevents runtime PM callbacks from racing with the bus 580notifier callback in __device_release_driver(), which is necessary, because the 581notifier is used by some subsystems to carry out operations affecting the 582runtime PM functionality. It does so by calling pm_runtime_get_sync() before 583driver_sysfs_remove() and the BUS_NOTIFY_UNBIND_DRIVER notifications. This 584resumes the device if it's in the suspended state and prevents it from 585being suspended again while those routines are being executed. 586 587To allow bus types and drivers to put devices into the suspended state by 588calling pm_runtime_suspend() from their ->remove() routines, the driver core 589executes pm_runtime_put_sync() after running the BUS_NOTIFY_UNBIND_DRIVER 590notifications in __device_release_driver(). This requires bus types and 591drivers to make their ->remove() callbacks avoid races with runtime PM directly, 592but also it allows of more flexibility in the handling of devices during the 593removal of their drivers. 594 595Drivers in ->remove() callback should undo the runtime PM changes done 596in ->probe(). Usually this means calling pm_runtime_disable(), 597pm_runtime_dont_use_autosuspend() etc. 598 599The user space can effectively disallow the driver of the device to power manage 600it at run time by changing the value of its /sys/devices/.../power/control 601attribute to "on", which causes pm_runtime_forbid() to be called. In principle, 602this mechanism may also be used by the driver to effectively turn off the 603runtime power management of the device until the user space turns it on. 604Namely, during the initialization the driver can make sure that the runtime PM 605status of the device is 'active' and call pm_runtime_forbid(). It should be 606noted, however, that if the user space has already intentionally changed the 607value of /sys/devices/.../power/control to "auto" to allow the driver to power 608manage the device at run time, the driver may confuse it by using 609pm_runtime_forbid() this way. 610 6116. Runtime PM and System Sleep 612 613Runtime PM and system sleep (i.e., system suspend and hibernation, also known 614as suspend-to-RAM and suspend-to-disk) interact with each other in a couple of 615ways. If a device is active when a system sleep starts, everything is 616straightforward. But what should happen if the device is already suspended? 617 618The device may have different wake-up settings for runtime PM and system sleep. 619For example, remote wake-up may be enabled for runtime suspend but disallowed 620for system sleep (device_may_wakeup(dev) returns 'false'). When this happens, 621the subsystem-level system suspend callback is responsible for changing the 622device's wake-up setting (it may leave that to the device driver's system 623suspend routine). It may be necessary to resume the device and suspend it again 624in order to do so. The same is true if the driver uses different power levels 625or other settings for runtime suspend and system sleep. 626 627During system resume, the simplest approach is to bring all devices back to full 628power, even if they had been suspended before the system suspend began. There 629are several reasons for this, including: 630 631 * The device might need to switch power levels, wake-up settings, etc. 632 633 * Remote wake-up events might have been lost by the firmware. 634 635 * The device's children may need the device to be at full power in order 636 to resume themselves. 637 638 * The driver's idea of the device state may not agree with the device's 639 physical state. This can happen during resume from hibernation. 640 641 * The device might need to be reset. 642 643 * Even though the device was suspended, if its usage counter was > 0 then most 644 likely it would need a runtime resume in the near future anyway. 645 646If the device had been suspended before the system suspend began and it's 647brought back to full power during resume, then its runtime PM status will have 648to be updated to reflect the actual post-system sleep status. The way to do 649this is: 650 651 pm_runtime_disable(dev); 652 pm_runtime_set_active(dev); 653 pm_runtime_enable(dev); 654 655The PM core always increments the runtime usage counter before calling the 656->suspend() callback and decrements it after calling the ->resume() callback. 657Hence disabling runtime PM temporarily like this will not cause any runtime 658suspend attempts to be permanently lost. If the usage count goes to zero 659following the return of the ->resume() callback, the ->runtime_idle() callback 660will be invoked as usual. 661 662On some systems, however, system sleep is not entered through a global firmware 663or hardware operation. Instead, all hardware components are put into low-power 664states directly by the kernel in a coordinated way. Then, the system sleep 665state effectively follows from the states the hardware components end up in 666and the system is woken up from that state by a hardware interrupt or a similar 667mechanism entirely under the kernel's control. As a result, the kernel never 668gives control away and the states of all devices during resume are precisely 669known to it. If that is the case and none of the situations listed above takes 670place (in particular, if the system is not waking up from hibernation), it may 671be more efficient to leave the devices that had been suspended before the system 672suspend began in the suspended state. 673 674To this end, the PM core provides a mechanism allowing some coordination between 675different levels of device hierarchy. Namely, if a system suspend .prepare() 676callback returns a positive number for a device, that indicates to the PM core 677that the device appears to be runtime-suspended and its state is fine, so it 678may be left in runtime suspend provided that all of its descendants are also 679left in runtime suspend. If that happens, the PM core will not execute any 680system suspend and resume callbacks for all of those devices, except for the 681complete callback, which is then entirely responsible for handling the device 682as appropriate. This only applies to system suspend transitions that are not 683related to hibernation (see Documentation/power/admin-guide/devices.rst for more 684information). 685 686The PM core does its best to reduce the probability of race conditions between 687the runtime PM and system suspend/resume (and hibernation) callbacks by carrying 688out the following operations: 689 690 * During system suspend pm_runtime_get_noresume() is called for every device 691 right before executing the subsystem-level .prepare() callback for it and 692 pm_runtime_barrier() is called for every device right before executing the 693 subsystem-level .suspend() callback for it. In addition to that the PM core 694 calls __pm_runtime_disable() with 'false' as the second argument for every 695 device right before executing the subsystem-level .suspend_late() callback 696 for it. 697 698 * During system resume pm_runtime_enable() and pm_runtime_put() are called for 699 every device right after executing the subsystem-level .resume_early() 700 callback and right after executing the subsystem-level .complete() callback 701 for it, respectively. 702 7037. Generic subsystem callbacks 704 705Subsystems may wish to conserve code space by using the set of generic power 706management callbacks provided by the PM core, defined in 707driver/base/power/generic_ops.c: 708 709 int pm_generic_runtime_suspend(struct device *dev); 710 - invoke the ->runtime_suspend() callback provided by the driver of this 711 device and return its result, or return 0 if not defined 712 713 int pm_generic_runtime_resume(struct device *dev); 714 - invoke the ->runtime_resume() callback provided by the driver of this 715 device and return its result, or return 0 if not defined 716 717 int pm_generic_suspend(struct device *dev); 718 - if the device has not been suspended at run time, invoke the ->suspend() 719 callback provided by its driver and return its result, or return 0 if not 720 defined 721 722 int pm_generic_suspend_noirq(struct device *dev); 723 - if pm_runtime_suspended(dev) returns "false", invoke the ->suspend_noirq() 724 callback provided by the device's driver and return its result, or return 725 0 if not defined 726 727 int pm_generic_resume(struct device *dev); 728 - invoke the ->resume() callback provided by the driver of this device and, 729 if successful, change the device's runtime PM status to 'active' 730 731 int pm_generic_resume_noirq(struct device *dev); 732 - invoke the ->resume_noirq() callback provided by the driver of this device 733 734 int pm_generic_freeze(struct device *dev); 735 - if the device has not been suspended at run time, invoke the ->freeze() 736 callback provided by its driver and return its result, or return 0 if not 737 defined 738 739 int pm_generic_freeze_noirq(struct device *dev); 740 - if pm_runtime_suspended(dev) returns "false", invoke the ->freeze_noirq() 741 callback provided by the device's driver and return its result, or return 742 0 if not defined 743 744 int pm_generic_thaw(struct device *dev); 745 - if the device has not been suspended at run time, invoke the ->thaw() 746 callback provided by its driver and return its result, or return 0 if not 747 defined 748 749 int pm_generic_thaw_noirq(struct device *dev); 750 - if pm_runtime_suspended(dev) returns "false", invoke the ->thaw_noirq() 751 callback provided by the device's driver and return its result, or return 752 0 if not defined 753 754 int pm_generic_poweroff(struct device *dev); 755 - if the device has not been suspended at run time, invoke the ->poweroff() 756 callback provided by its driver and return its result, or return 0 if not 757 defined 758 759 int pm_generic_poweroff_noirq(struct device *dev); 760 - if pm_runtime_suspended(dev) returns "false", run the ->poweroff_noirq() 761 callback provided by the device's driver and return its result, or return 762 0 if not defined 763 764 int pm_generic_restore(struct device *dev); 765 - invoke the ->restore() callback provided by the driver of this device and, 766 if successful, change the device's runtime PM status to 'active' 767 768 int pm_generic_restore_noirq(struct device *dev); 769 - invoke the ->restore_noirq() callback provided by the device's driver 770 771These functions are the defaults used by the PM core, if a subsystem doesn't 772provide its own callbacks for ->runtime_idle(), ->runtime_suspend(), 773->runtime_resume(), ->suspend(), ->suspend_noirq(), ->resume(), 774->resume_noirq(), ->freeze(), ->freeze_noirq(), ->thaw(), ->thaw_noirq(), 775->poweroff(), ->poweroff_noirq(), ->restore(), ->restore_noirq() in the 776subsystem-level dev_pm_ops structure. 777 778Device drivers that wish to use the same function as a system suspend, freeze, 779poweroff and runtime suspend callback, and similarly for system resume, thaw, 780restore, and runtime resume, can achieve this with the help of the 781UNIVERSAL_DEV_PM_OPS macro defined in include/linux/pm.h (possibly setting its 782last argument to NULL). 783 7848. "No-Callback" Devices 785 786Some "devices" are only logical sub-devices of their parent and cannot be 787power-managed on their own. (The prototype example is a USB interface. Entire 788USB devices can go into low-power mode or send wake-up requests, but neither is 789possible for individual interfaces.) The drivers for these devices have no 790need of runtime PM callbacks; if the callbacks did exist, ->runtime_suspend() 791and ->runtime_resume() would always return 0 without doing anything else and 792->runtime_idle() would always call pm_runtime_suspend(). 793 794Subsystems can tell the PM core about these devices by calling 795pm_runtime_no_callbacks(). This should be done after the device structure is 796initialized and before it is registered (although after device registration is 797also okay). The routine will set the device's power.no_callbacks flag and 798prevent the non-debugging runtime PM sysfs attributes from being created. 799 800When power.no_callbacks is set, the PM core will not invoke the 801->runtime_idle(), ->runtime_suspend(), or ->runtime_resume() callbacks. 802Instead it will assume that suspends and resumes always succeed and that idle 803devices should be suspended. 804 805As a consequence, the PM core will never directly inform the device's subsystem 806or driver about runtime power changes. Instead, the driver for the device's 807parent must take responsibility for telling the device's driver when the 808parent's power state changes. 809 8109. Autosuspend, or automatically-delayed suspends 811 812Changing a device's power state isn't free; it requires both time and energy. 813A device should be put in a low-power state only when there's some reason to 814think it will remain in that state for a substantial time. A common heuristic 815says that a device which hasn't been used for a while is liable to remain 816unused; following this advice, drivers should not allow devices to be suspended 817at runtime until they have been inactive for some minimum period. Even when 818the heuristic ends up being non-optimal, it will still prevent devices from 819"bouncing" too rapidly between low-power and full-power states. 820 821The term "autosuspend" is an historical remnant. It doesn't mean that the 822device is automatically suspended (the subsystem or driver still has to call 823the appropriate PM routines); rather it means that runtime suspends will 824automatically be delayed until the desired period of inactivity has elapsed. 825 826Inactivity is determined based on the power.last_busy field. Drivers should 827call pm_runtime_mark_last_busy() to update this field after carrying out I/O, 828typically just before calling pm_runtime_put_autosuspend(). The desired length 829of the inactivity period is a matter of policy. Subsystems can set this length 830initially by calling pm_runtime_set_autosuspend_delay(), but after device 831registration the length should be controlled by user space, using the 832/sys/devices/.../power/autosuspend_delay_ms attribute. 833 834In order to use autosuspend, subsystems or drivers must call 835pm_runtime_use_autosuspend() (preferably before registering the device), and 836thereafter they should use the various *_autosuspend() helper functions instead 837of the non-autosuspend counterparts: 838 839 Instead of: pm_runtime_suspend use: pm_runtime_autosuspend; 840 Instead of: pm_schedule_suspend use: pm_request_autosuspend; 841 Instead of: pm_runtime_put use: pm_runtime_put_autosuspend; 842 Instead of: pm_runtime_put_sync use: pm_runtime_put_sync_autosuspend. 843 844Drivers may also continue to use the non-autosuspend helper functions; they 845will behave normally, which means sometimes taking the autosuspend delay into 846account (see pm_runtime_idle). 847 848Under some circumstances a driver or subsystem may want to prevent a device 849from autosuspending immediately, even though the usage counter is zero and the 850autosuspend delay time has expired. If the ->runtime_suspend() callback 851returns -EAGAIN or -EBUSY, and if the next autosuspend delay expiration time is 852in the future (as it normally would be if the callback invoked 853pm_runtime_mark_last_busy()), the PM core will automatically reschedule the 854autosuspend. The ->runtime_suspend() callback can't do this rescheduling 855itself because no suspend requests of any kind are accepted while the device is 856suspending (i.e., while the callback is running). 857 858The implementation is well suited for asynchronous use in interrupt contexts. 859However such use inevitably involves races, because the PM core can't 860synchronize ->runtime_suspend() callbacks with the arrival of I/O requests. 861This synchronization must be handled by the driver, using its private lock. 862Here is a schematic pseudo-code example: 863 864 foo_read_or_write(struct foo_priv *foo, void *data) 865 { 866 lock(&foo->private_lock); 867 add_request_to_io_queue(foo, data); 868 if (foo->num_pending_requests++ == 0) 869 pm_runtime_get(&foo->dev); 870 if (!foo->is_suspended) 871 foo_process_next_request(foo); 872 unlock(&foo->private_lock); 873 } 874 875 foo_io_completion(struct foo_priv *foo, void *req) 876 { 877 lock(&foo->private_lock); 878 if (--foo->num_pending_requests == 0) { 879 pm_runtime_mark_last_busy(&foo->dev); 880 pm_runtime_put_autosuspend(&foo->dev); 881 } else { 882 foo_process_next_request(foo); 883 } 884 unlock(&foo->private_lock); 885 /* Send req result back to the user ... */ 886 } 887 888 int foo_runtime_suspend(struct device *dev) 889 { 890 struct foo_priv foo = container_of(dev, ...); 891 int ret = 0; 892 893 lock(&foo->private_lock); 894 if (foo->num_pending_requests > 0) { 895 ret = -EBUSY; 896 } else { 897 /* ... suspend the device ... */ 898 foo->is_suspended = 1; 899 } 900 unlock(&foo->private_lock); 901 return ret; 902 } 903 904 int foo_runtime_resume(struct device *dev) 905 { 906 struct foo_priv foo = container_of(dev, ...); 907 908 lock(&foo->private_lock); 909 /* ... resume the device ... */ 910 foo->is_suspended = 0; 911 pm_runtime_mark_last_busy(&foo->dev); 912 if (foo->num_pending_requests > 0) 913 foo_process_next_request(foo); 914 unlock(&foo->private_lock); 915 return 0; 916 } 917 918The important point is that after foo_io_completion() asks for an autosuspend, 919the foo_runtime_suspend() callback may race with foo_read_or_write(). 920Therefore foo_runtime_suspend() has to check whether there are any pending I/O 921requests (while holding the private lock) before allowing the suspend to 922proceed. 923 924In addition, the power.autosuspend_delay field can be changed by user space at 925any time. If a driver cares about this, it can call 926pm_runtime_autosuspend_expiration() from within the ->runtime_suspend() 927callback while holding its private lock. If the function returns a nonzero 928value then the delay has not yet expired and the callback should return 929-EAGAIN. 930