blob: d0e79d5820a5e7a7237f36a1cb5776ec589552d8 [file] [log] [blame]
David Brownell4fc08402006-08-10 16:38:28 -07001Most of the code in Linux is device drivers, so most of the Linux power
2management code is also driver-specific. Most drivers will do very little;
3others, especially for platforms with small batteries (like cell phones),
4will do a lot.
Linus Torvalds1da177e2005-04-16 15:20:36 -07005
David Brownell4fc08402006-08-10 16:38:28 -07006This writeup gives an overview of how drivers interact with system-wide
7power management goals, emphasizing the models and interfaces that are
8shared by everything that hooks up to the driver model core. Read it as
9background for the domain-specific work you'd do with any specific driver.
Linus Torvalds1da177e2005-04-16 15:20:36 -070010
11
David Brownell4fc08402006-08-10 16:38:28 -070012Two Models for Device Power Management
13======================================
14Drivers will use one or both of these models to put devices into low-power
15states:
16
17 System Sleep model:
18 Drivers can enter low power states as part of entering system-wide
19 low-power states like "suspend-to-ram", or (mostly for systems with
20 disks) "hibernate" (suspend-to-disk).
21
22 This is something that device, bus, and class drivers collaborate on
23 by implementing various role-specific suspend and resume methods to
24 cleanly power down hardware and software subsystems, then reactivate
25 them without loss of data.
26
27 Some drivers can manage hardware wakeup events, which make the system
28 leave that low-power state. This feature may be disabled using the
29 relevant /sys/devices/.../power/wakeup file; enabling it may cost some
30 power usage, but let the whole system enter low power states more often.
31
32 Runtime Power Management model:
33 Drivers may also enter low power states while the system is running,
34 independently of other power management activity. Upstream drivers
35 will normally not know (or care) if the device is in some low power
36 state when issuing requests; the driver will auto-resume anything
37 that's needed when it gets a request.
38
39 This doesn't have, or need much infrastructure; it's just something you
40 should do when writing your drivers. For example, clk_disable() unused
41 clocks as part of minimizing power drain for currently-unused hardware.
42 Of course, sometimes clusters of drivers will collaborate with each
43 other, which could involve task-specific power management.
44
45There's not a lot to be said about those low power states except that they
46are very system-specific, and often device-specific. Also, that if enough
47drivers put themselves into low power states (at "runtime"), the effect may be
48the same as entering some system-wide low-power state (system sleep) ... and
49that synergies exist, so that several drivers using runtime pm might put the
50system into a state where even deeper power saving options are available.
51
52Most suspended devices will have quiesced all I/O: no more DMA or irqs, no
53more data read or written, and requests from upstream drivers are no longer
54accepted. A given bus or platform may have different requirements though.
55
56Examples of hardware wakeup events include an alarm from a real time clock,
57network wake-on-LAN packets, keyboard or mouse activity, and media insertion
58or removal (for PCMCIA, MMC/SD, USB, and so on).
Linus Torvalds1da177e2005-04-16 15:20:36 -070059
60
David Brownell4fc08402006-08-10 16:38:28 -070061Interfaces for Entering System Sleep States
62===========================================
63Most of the programming interfaces a device driver needs to know about
64relate to that first model: entering a system-wide low power state,
65rather than just minimizing power consumption by one device.
Linus Torvalds1da177e2005-04-16 15:20:36 -070066
David Brownell4fc08402006-08-10 16:38:28 -070067
68Bus Driver Methods
69------------------
70The core methods to suspend and resume devices reside in struct bus_type.
71These are mostly of interest to people writing infrastructure for busses
72like PCI or USB, or because they define the primitives that device drivers
73may need to apply in domain-specific ways to their devices:
Linus Torvalds1da177e2005-04-16 15:20:36 -070074
75struct bus_type {
David Brownell4fc08402006-08-10 16:38:28 -070076 ...
77 int (*suspend)(struct device *dev, pm_message_t state);
78 int (*suspend_late)(struct device *dev, pm_message_t state);
79
80 int (*resume_early)(struct device *dev);
81 int (*resume)(struct device *dev);
Linus Torvalds1da177e2005-04-16 15:20:36 -070082};
83
David Brownell4fc08402006-08-10 16:38:28 -070084Bus drivers implement those methods as appropriate for the hardware and
85the drivers using it; PCI works differently from USB, and so on. Not many
86people write bus drivers; most driver code is a "device driver" that
87builds on top of bus-specific framework code.
Linus Torvalds1da177e2005-04-16 15:20:36 -070088
David Brownell4fc08402006-08-10 16:38:28 -070089For more information on these driver calls, see the description later;
90they are called in phases for every device, respecting the parent-child
91sequencing in the driver model tree. Note that as this is being written,
92only the suspend() and resume() are widely available; not many bus drivers
93leverage all of those phases, or pass them down to lower driver levels.
Linus Torvalds1da177e2005-04-16 15:20:36 -070094
95
David Brownell4fc08402006-08-10 16:38:28 -070096/sys/devices/.../power/wakeup files
97-----------------------------------
98All devices in the driver model have two flags to control handling of
99wakeup events, which are hardware signals that can force the device and/or
100system out of a low power state. These are initialized by bus or device
101driver code using device_init_wakeup(dev,can_wakeup).
Linus Torvalds1da177e2005-04-16 15:20:36 -0700102
David Brownell4fc08402006-08-10 16:38:28 -0700103The "can_wakeup" flag just records whether the device (and its driver) can
104physically support wakeup events. When that flag is clear, the sysfs
105"wakeup" file is empty, and device_may_wakeup() returns false.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700106
David Brownell4fc08402006-08-10 16:38:28 -0700107For devices that can issue wakeup events, a separate flag controls whether
108that device should try to use its wakeup mechanism. The initial value of
109device_may_wakeup() will be true, so that the device's "wakeup" file holds
110the value "enabled". Userspace can change that to "disabled" so that
111device_may_wakeup() returns false; or change it back to "enabled" (so that
112it returns true again).
Linus Torvalds1da177e2005-04-16 15:20:36 -0700113
Linus Torvalds1da177e2005-04-16 15:20:36 -0700114
David Brownell4fc08402006-08-10 16:38:28 -0700115EXAMPLE: PCI Device Driver Methods
116-----------------------------------
117PCI framework software calls these methods when the PCI device driver bound
118to a device device has provided them:
Linus Torvalds1da177e2005-04-16 15:20:36 -0700119
David Brownell4fc08402006-08-10 16:38:28 -0700120struct pci_driver {
121 ...
122 int (*suspend)(struct pci_device *pdev, pm_message_t state);
123 int (*suspend_late)(struct pci_device *pdev, pm_message_t state);
124
125 int (*resume_early)(struct pci_device *pdev);
126 int (*resume)(struct pci_device *pdev);
127};
128
129Drivers will implement those methods, and call PCI-specific procedures
130like pci_set_power_state(), pci_enable_wake(), pci_save_state(), and
131pci_restore_state() to manage PCI-specific mechanisms. (PCI config space
132could be saved during driver probe, if it weren't for the fact that some
133systems rely on userspace tweaking using setpci.) Devices are suspended
134before their bridges enter low power states, and likewise bridges resume
135before their devices.
136
137
138Upper Layers of Driver Stacks
139-----------------------------
140Device drivers generally have at least two interfaces, and the methods
141sketched above are the ones which apply to the lower level (nearer PCI, USB,
142or other bus hardware). The network and block layers are examples of upper
143level interfaces, as is a character device talking to userspace.
144
145Power management requests normally need to flow through those upper levels,
146which often use domain-oriented requests like "blank that screen". In
147some cases those upper levels will have power management intelligence that
148relates to end-user activity, or other devices that work in cooperation.
149
150When those interfaces are structured using class interfaces, there is a
151standard way to have the upper layer stop issuing requests to a given
152class device (and restart later):
153
154struct class {
155 ...
156 int (*suspend)(struct device *dev, pm_message_t state);
157 int (*resume)(struct device *dev);
158};
159
160Those calls are issued in specific phases of the process by which the
161system enters a low power "suspend" state, or resumes from it.
162
163
164Calling Drivers to Enter System Sleep States
165============================================
166When the system enters a low power state, each device's driver is asked
167to suspend the device by putting it into state compatible with the target
168system state. That's usually some version of "off", but the details are
169system-specific. Also, wakeup-enabled devices will usually stay partly
170functional in order to wake the system.
171
172When the system leaves that low power state, the device's driver is asked
173to resume it. The suspend and resume operations always go together, and
174both are multi-phase operations.
175
176For simple drivers, suspend might quiesce the device using the class code
177and then turn its hardware as "off" as possible with late_suspend. The
178matching resume calls would then completely reinitialize the hardware
179before reactivating its class I/O queues.
180
181More power-aware drivers drivers will use more than one device low power
182state, either at runtime or during system sleep states, and might trigger
183system wakeup events.
184
185
186Call Sequence Guarantees
187------------------------
188To ensure that bridges and similar links needed to talk to a device are
189available when the device is suspended or resumed, the device tree is
190walked in a bottom-up order to suspend devices. A top-down order is
191used to resume those devices.
192
193The ordering of the device tree is defined by the order in which devices
194get registered: a child can never be registered, probed or resumed before
195its parent; and can't be removed or suspended after that parent.
196
197The policy is that the device tree should match hardware bus topology.
198(Or at least the control bus, for devices which use multiple busses.)
199
200
201Suspending Devices
202------------------
203Suspending a given device is done in several phases. Suspending the
204system always includes every phase, executing calls for every device
205before the next phase begins. Not all busses or classes support all
206these callbacks; and not all drivers use all the callbacks.
207
208The phases are seen by driver notifications issued in this order:
209
210 1 class.suspend(dev, message) is called after tasks are frozen, for
211 devices associated with a class that has such a method. This
212 method may sleep.
213
214 Since I/O activity usually comes from such higher layers, this is
215 a good place to quiesce all drivers of a given type (and keep such
216 code out of those drivers).
217
218 2 bus.suspend(dev, message) is called next. This method may sleep,
219 and is often morphed into a device driver call with bus-specific
220 parameters and/or rules.
221
222 This call should handle parts of device suspend logic that require
223 sleeping. It probably does work to quiesce the device which hasn't
224 been abstracted into class.suspend() or bus.suspend_late().
225
226 3 bus.suspend_late(dev, message) is called with IRQs disabled, and
227 with only one CPU active. Until the bus.resume_early() phase
228 completes (see later), IRQs are not enabled again. This method
229 won't be exposed by all busses; for message based busses like USB,
230 I2C, or SPI, device interactions normally require IRQs. This bus
231 call may be morphed into a driver call with bus-specific parameters.
232
233 This call might save low level hardware state that might otherwise
234 be lost in the upcoming low power state, and actually put the
235 device into a low power state ... so that in some cases the device
236 may stay partly usable until this late. This "late" call may also
237 help when coping with hardware that behaves badly.
238
239The pm_message_t parameter is currently used to refine those semantics
240(described later).
241
242At the end of those phases, drivers should normally have stopped all I/O
243transactions (DMA, IRQs), saved enough state that they can re-initialize
244or restore previous state (as needed by the hardware), and placed the
245device into a low-power state. On many platforms they will also use
246clk_disable() to gate off one or more clock sources; sometimes they will
247also switch off power supplies, or reduce voltages. Drivers which have
248runtime PM support may already have performed some or all of the steps
249needed to prepare for the upcoming system sleep state.
250
251When any driver sees that its device_can_wakeup(dev), it should make sure
252to use the relevant hardware signals to trigger a system wakeup event.
253For example, enable_irq_wake() might identify GPIO signals hooked up to
254a switch or other external hardware, and pci_enable_wake() does something
255similar for PCI's PME# signal.
256
257If a driver (or bus, or class) fails it suspend method, the system won't
258enter the desired low power state; it will resume all the devices it's
259suspended so far.
260
261Note that drivers may need to perform different actions based on the target
262system lowpower/sleep state. At this writing, there are only platform
263specific APIs through which drivers could determine those target states.
264
265
266Device Low Power (suspend) States
267---------------------------------
268Device low-power states aren't very standard. One device might only handle
269"on" and "off, while another might support a dozen different versions of
270"on" (how many engines are active?), plus a state that gets back to "on"
271faster than from a full "off".
272
273Some busses define rules about what different suspend states mean. PCI
274gives one example: after the suspend sequence completes, a non-legacy
275PCI device may not perform DMA or issue IRQs, and any wakeup events it
276issues would be issued through the PME# bus signal. Plus, there are
277several PCI-standard device states, some of which are optional.
278
279In contrast, integrated system-on-chip processors often use irqs as the
280wakeup event sources (so drivers would call enable_irq_wake) and might
281be able to treat DMA completion as a wakeup event (sometimes DMA can stay
282active too, it'd only be the CPU and some peripherals that sleep).
283
284Some details here may be platform-specific. Systems may have devices that
285can be fully active in certain sleep states, such as an LCD display that's
286refreshed using DMA while most of the system is sleeping lightly ... and
287its frame buffer might even be updated by a DSP or other non-Linux CPU while
288the Linux control processor stays idle.
289
290Moreover, the specific actions taken may depend on the target system state.
291One target system state might allow a given device to be very operational;
292another might require a hard shut down with re-initialization on resume.
293And two different target systems might use the same device in different
294ways; the aforementioned LCD might be active in one product's "standby",
295but a different product using the same SOC might work differently.
296
297
298Meaning of pm_message_t.event
299-----------------------------
300Parameters to suspend calls include the device affected and a message of
301type pm_message_t, which has one field: the event. If driver does not
302recognize the event code, suspend calls may abort the request and return
303a negative errno. However, most drivers will be fine if they implement
304PM_EVENT_SUSPEND semantics for all messages.
305
306The event codes are used to refine the goal of suspending the device, and
307mostly matter when creating or resuming system memory image snapshots, as
308used with suspend-to-disk:
309
310 PM_EVENT_SUSPEND -- quiesce the driver and put hardware into a low-power
311 state. When used with system sleep states like "suspend-to-RAM" or
312 "standby", the upcoming resume() call will often be able to rely on
313 state kept in hardware, or issue system wakeup events. When used
314 instead with suspend-to-disk, few devices support this capability;
315 most are completely powered off.
316
317 PM_EVENT_FREEZE -- quiesce the driver, but don't necessarily change into
318 any low power mode. A system snapshot is about to be taken, often
319 followed by a call to the driver's resume() method. Neither wakeup
320 events nor DMA are allowed.
321
322 PM_EVENT_PRETHAW -- quiesce the driver, knowing that the upcoming resume()
323 will restore a suspend-to-disk snapshot from a different kernel image.
324 Drivers that are smart enough to look at their hardware state during
325 resume() processing need that state to be correct ... a PRETHAW could
326 be used to invalidate that state (by resetting the device), like a
327 shutdown() invocation would before a kexec() or system halt. Other
328 drivers might handle this the same way as PM_EVENT_FREEZE. Neither
329 wakeup events nor DMA are allowed.
330
331To enter "standby" (ACPI S1) or "Suspend to RAM" (STR, ACPI S3) states, or
332the similarly named APM states, only PM_EVENT_SUSPEND is used; for "Suspend
333to Disk" (STD, hibernate, ACPI S4), all of those event codes are used.
334
335There's also PM_EVENT_ON, a value which never appears as a suspend event
336but is sometimes used to record the "not suspended" device state.
337
338
339Resuming Devices
340----------------
341Resuming is done in multiple phases, much like suspending, with all
342devices processing each phase's calls before the next phase begins.
343
344The phases are seen by driver notifications issued in this order:
345
346 1 bus.resume_early(dev) is called with IRQs disabled, and with
347 only one CPU active. As with bus.suspend_late(), this method
348 won't be supported on busses that require IRQs in order to
349 interact with devices.
350
351 This reverses the effects of bus.suspend_late().
352
353 2 bus.resume(dev) is called next. This may be morphed into a device
354 driver call with bus-specific parameters; implementations may sleep.
355
356 This reverses the effects of bus.suspend().
357
358 3 class.resume(dev) is called for devices associated with a class
359 that has such a method. Implementations may sleep.
360
361 This reverses the effects of class.suspend(), and would usually
362 reactivate the device's I/O queue.
363
364At the end of those phases, drivers should normally be as functional as
365they were before suspending: I/O can be performed using DMA and IRQs, and
366the relevant clocks are gated on. The device need not be "fully on"; it
367might be in a runtime lowpower/suspend state that acts as if it were.
368
369However, the details here may again be platform-specific. For example,
370some systems support multiple "run" states, and the mode in effect at
371the end of resume() might not be the one which preceded suspension.
372That means availability of certain clocks or power supplies changed,
373which could easily affect how a driver works.
374
375
376Drivers need to be able to handle hardware which has been reset since the
377suspend methods were called, for example by complete reinitialization.
378This may be the hardest part, and the one most protected by NDA'd documents
379and chip errata. It's simplest if the hardware state hasn't changed since
380the suspend() was called, but that can't always be guaranteed.
381
382Drivers must also be prepared to notice that the device has been removed
383while the system was powered off, whenever that's physically possible.
384PCMCIA, MMC, USB, Firewire, SCSI, and even IDE are common examples of busses
385where common Linux platforms will see such removal. Details of how drivers
386will notice and handle such removals are currently bus-specific, and often
387involve a separate thread.
388
389
390Note that the bus-specific runtime PM wakeup mechanism can exist, and might
391be defined to share some of the same driver code as for system wakeup. For
392example, a bus-specific device driver's resume() method might be used there,
393so it wouldn't only be called from bus.resume() during system-wide wakeup.
394See bus-specific information about how runtime wakeup events are handled.
395
396
397System Devices
398--------------
Linus Torvalds1da177e2005-04-16 15:20:36 -0700399System devices follow a slightly different API, which can be found in
400
401 include/linux/sysdev.h
402 drivers/base/sys.c
403
David Brownell4fc08402006-08-10 16:38:28 -0700404System devices will only be suspended with interrupts disabled, and after
405all other devices have been suspended. On resume, they will be resumed
406before any other devices, and also with interrupts disabled.
407
408That is, IRQs are disabled, the suspend_late() phase begins, then the
409sysdev_driver.suspend() phase, and the system enters a sleep state. Then
410the sysdev_driver.resume() phase begins, followed by the resume_early()
411phase, after which IRQs are enabled.
412
413Code to actually enter and exit the system-wide low power state sometimes
414involves hardware details that are only known to the boot firmware, and
415may leave a CPU running software (from SRAM or flash memory) that monitors
416the system and manages its wakeup sequence.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700417
418
419Runtime Power Management
David Brownell4fc08402006-08-10 16:38:28 -0700420========================
421Many devices are able to dynamically power down while the system is still
422running. This feature is useful for devices that are not being used, and
423can offer significant power savings on a running system. These devices
424often support a range of runtime power states, which might use names such
425as "off", "sleep", "idle", "active", and so on. Those states will in some
426cases (like PCI) be partially constrained by a bus the device uses, and will
427usually include hardware states that are also used in system sleep states.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700428
David Brownell4fc08402006-08-10 16:38:28 -0700429However, note that if a driver puts a device into a runtime low power state
430and the system then goes into a system-wide sleep state, it normally ought
431to resume into that runtime low power state rather than "full on". Such
432distinctions would be part of the driver-internal state machine for that
433hardware; the whole point of runtime power management is to be sure that
434drivers are decoupled in that way from the state machine governing phases
435of the system-wide power/sleep state transitions.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700436
Linus Torvalds1da177e2005-04-16 15:20:36 -0700437
David Brownell4fc08402006-08-10 16:38:28 -0700438Power Saving Techniques
439-----------------------
440Normally runtime power management is handled by the drivers without specific
441userspace or kernel intervention, by device-aware use of techniques like:
Linus Torvalds1da177e2005-04-16 15:20:36 -0700442
David Brownell4fc08402006-08-10 16:38:28 -0700443 Using information provided by other system layers
444 - stay deeply "off" except between open() and close()
445 - if transceiver/PHY indicates "nobody connected", stay "off"
446 - application protocols may include power commands or hints
Linus Torvalds1da177e2005-04-16 15:20:36 -0700447
David Brownell4fc08402006-08-10 16:38:28 -0700448 Using fewer CPU cycles
449 - using DMA instead of PIO
450 - removing timers, or making them lower frequency
451 - shortening "hot" code paths
452 - eliminating cache misses
453 - (sometimes) offloading work to device firmware
Linus Torvalds1da177e2005-04-16 15:20:36 -0700454
David Brownell4fc08402006-08-10 16:38:28 -0700455 Reducing other resource costs
456 - gating off unused clocks in software (or hardware)
457 - switching off unused power supplies
458 - eliminating (or delaying/merging) IRQs
459 - tuning DMA to use word and/or burst modes
Linus Torvalds1da177e2005-04-16 15:20:36 -0700460
David Brownell4fc08402006-08-10 16:38:28 -0700461 Using device-specific low power states
462 - using lower voltages
463 - avoiding needless DMA transfers
Linus Torvalds1da177e2005-04-16 15:20:36 -0700464
David Brownell4fc08402006-08-10 16:38:28 -0700465Read your hardware documentation carefully to see the opportunities that
466may be available. If you can, measure the actual power usage and check
467it against the budget established for your project.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700468
Linus Torvalds1da177e2005-04-16 15:20:36 -0700469
David Brownell4fc08402006-08-10 16:38:28 -0700470Examples: USB hosts, system timer, system CPU
471----------------------------------------------
472USB host controllers make interesting, if complex, examples. In many cases
473these have no work to do: no USB devices are connected, or all of them are
474in the USB "suspend" state. Linux host controller drivers can then disable
475periodic DMA transfers that would otherwise be a constant power drain on the
476memory subsystem, and enter a suspend state. In power-aware controllers,
477entering that suspend state may disable the clock used with USB signaling,
478saving a certain amount of power.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700479
David Brownell4fc08402006-08-10 16:38:28 -0700480The controller will be woken from that state (with an IRQ) by changes to the
481signal state on the data lines of a given port, for example by an existing
482peripheral requesting "remote wakeup" or by plugging a new peripheral. The
483same wakeup mechanism usually works from "standby" sleep states, and on some
484systems also from "suspend to RAM" (or even "suspend to disk") states.
485(Except that ACPI may be involved instead of normal IRQs, on some hardware.)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700486
David Brownell4fc08402006-08-10 16:38:28 -0700487System devices like timers and CPUs may have special roles in the platform
488power management scheme. For example, system timers using a "dynamic tick"
489approach don't just save CPU cycles (by eliminating needless timer IRQs),
490but they may also open the door to using lower power CPU "idle" states that
491cost more than a jiffie to enter and exit. On x86 systems these are states
492like "C3"; note that periodic DMA transfers from a USB host controller will
493also prevent entry to a C3 state, much like a periodic timer IRQ.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700494
David Brownell4fc08402006-08-10 16:38:28 -0700495That kind of runtime mechanism interaction is common. "System On Chip" (SOC)
496processors often have low power idle modes that can't be entered unless
497certain medium-speed clocks (often 12 or 48 MHz) are gated off. When the
498drivers gate those clocks effectively, then the system idle task may be able
499to use the lower power idle modes and thereby increase battery life.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700500
David Brownell4fc08402006-08-10 16:38:28 -0700501If the CPU can have a "cpufreq" driver, there also may be opportunities
502to shift to lower voltage settings and reduce the power cost of executing
503a given number of instructions. (Without voltage adjustment, it's rare
504for cpufreq to save much power; the cost-per-instruction must go down.)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700505
Linus Torvalds1da177e2005-04-16 15:20:36 -0700506
David Brownell4fc08402006-08-10 16:38:28 -0700507/sys/devices/.../power/state files
508==================================
509For now you can also test some of this functionality using sysfs.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700510
David Brownell4fc08402006-08-10 16:38:28 -0700511 DEPRECATED: USE "power/state" ONLY FOR DRIVER TESTING, AND
512 AVOID USING dev->power.power_state IN DRIVERS.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700513
David Brownell4fc08402006-08-10 16:38:28 -0700514 THESE WILL BE REMOVED. IF THE "power/state" FILE GETS REPLACED,
515 IT WILL BECOME SOMETHING COUPLED TO THE BUS OR DRIVER.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700516
David Brownell4fc08402006-08-10 16:38:28 -0700517In each device's directory, there is a 'power' directory, which contains
518at least a 'state' file. The value of this field is effectively boolean,
519PM_EVENT_ON or PM_EVENT_SUSPEND.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700520
David Brownell4fc08402006-08-10 16:38:28 -0700521 * Reading from this file displays a value corresponding to
522 the power.power_state.event field. All nonzero values are
523 displayed as "2", corresponding to a low power state; zero
524 is displayed as "0", corresponding to normal operation.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700525
David Brownell4fc08402006-08-10 16:38:28 -0700526 * Writing to this file initiates a transition using the
527 specified event code number; only '0', '2', and '3' are
528 accepted (without a newline); '2' and '3' are both
529 mapped to PM_EVENT_SUSPEND.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700530
David Brownell4fc08402006-08-10 16:38:28 -0700531On writes, the PM core relies on that recorded event code and the device/bus
532capabilities to determine whether it uses a partial suspend() or resume()
533sequence to change things so that the recorded event corresponds to the
534numeric parameter.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700535
David Brownell4fc08402006-08-10 16:38:28 -0700536 - If the bus requires the irqs-disabled suspend_late()/resume_early()
537 phases, writes fail because those operations are not supported here.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700538
David Brownell4fc08402006-08-10 16:38:28 -0700539 - If the recorded value is the expected value, nothing is done.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700540
David Brownell4fc08402006-08-10 16:38:28 -0700541 - If the recorded value is nonzero, the device is partially resumed,
542 using the bus.resume() and/or class.resume() methods.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700543
David Brownell4fc08402006-08-10 16:38:28 -0700544 - If the target value is nonzero, the device is partially suspended,
545 using the class.suspend() and/or bus.suspend() methods and the
546 PM_EVENT_SUSPEND message.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700547
David Brownell4fc08402006-08-10 16:38:28 -0700548Drivers have no way to tell whether their suspend() and resume() calls
549have come through the sysfs power/state file or as part of entering a
550system sleep state, except that when accessed through sysfs the normal
551parent/child sequencing rules are ignored. Drivers (such as bus, bridge,
552or hub drivers) which expose child devices may need to enforce those rules
553on their own.