diff options
Diffstat (limited to 'Documentation/power/devices.txt')
-rw-r--r-- | Documentation/power/devices.txt | 67 |
1 files changed, 14 insertions, 53 deletions
diff --git a/Documentation/power/devices.txt b/Documentation/power/devices.txt index 88880839ece..64565aac6e4 100644 --- a/Documentation/power/devices.txt +++ b/Documentation/power/devices.txt | |||
@@ -520,59 +520,20 @@ Support for power domains is provided through the pwr_domain field of struct | |||
520 | device. This field is a pointer to an object of type struct dev_power_domain, | 520 | device. This field is a pointer to an object of type struct dev_power_domain, |
521 | defined in include/linux/pm.h, providing a set of power management callbacks | 521 | defined in include/linux/pm.h, providing a set of power management callbacks |
522 | analogous to the subsystem-level and device driver callbacks that are executed | 522 | analogous to the subsystem-level and device driver callbacks that are executed |
523 | for the given device during all power transitions, in addition to the respective | 523 | for the given device during all power transitions, instead of the respective |
524 | subsystem-level callbacks. Specifically, the power domain "suspend" callbacks | 524 | subsystem-level callbacks. Specifically, if a device's pm_domain pointer is |
525 | (i.e. ->runtime_suspend(), ->suspend(), ->freeze(), ->poweroff(), etc.) are | 525 | not NULL, the ->suspend() callback from the object pointed to by it will be |
526 | executed after the analogous subsystem-level callbacks, while the power domain | 526 | executed instead of its subsystem's (e.g. bus type's) ->suspend() callback and |
527 | "resume" callbacks (i.e. ->runtime_resume(), ->resume(), ->thaw(), ->restore, | 527 | anlogously for all of the remaining callbacks. In other words, power management |
528 | etc.) are executed before the analogous subsystem-level callbacks. Error codes | 528 | domain callbacks, if defined for the given device, always take precedence over |
529 | returned by the "suspend" and "resume" power domain callbacks are ignored. | 529 | the callbacks provided by the device's subsystem (e.g. bus type). |
530 | 530 | ||
531 | Power domain ->runtime_idle() callback is executed before the subsystem-level | 531 | The support for device power management domains is only relevant to platforms |
532 | ->runtime_idle() callback and the result returned by it is not ignored. Namely, | 532 | needing to use the same device driver power management callbacks in many |
533 | if it returns error code, the subsystem-level ->runtime_idle() callback will not | 533 | different power domain configurations and wanting to avoid incorporating the |
534 | be called and the helper function rpm_idle() executing it will return error | 534 | support for power domains into subsystem-level callbacks, for example by |
535 | code. This mechanism is intended to help platforms where saving device state | 535 | modifying the platform bus type. Other platforms need not implement it or take |
536 | is a time consuming operation and should only be carried out if all devices | 536 | it into account in any way. |
537 | in the power domain are idle, before turning off the shared power resource(s). | ||
538 | Namely, the power domain ->runtime_idle() callback may return error code until | ||
539 | the pm_runtime_idle() helper (or its asychronous version) has been called for | ||
540 | all devices in the power domain (it is recommended that the returned error code | ||
541 | be -EBUSY in those cases), preventing the subsystem-level ->runtime_idle() | ||
542 | callback from being run prematurely. | ||
543 | |||
544 | The support for device power domains is only relevant to platforms needing to | ||
545 | use the same subsystem-level (e.g. platform bus type) and device driver power | ||
546 | management callbacks in many different power domain configurations and wanting | ||
547 | to avoid incorporating the support for power domains into the subsystem-level | ||
548 | callbacks. The other platforms need not implement it or take it into account | ||
549 | in any way. | ||
550 | |||
551 | |||
552 | System Devices | ||
553 | -------------- | ||
554 | System devices (sysdevs) follow a slightly different API, which can be found in | ||
555 | |||
556 | include/linux/sysdev.h | ||
557 | drivers/base/sys.c | ||
558 | |||
559 | System devices will be suspended with interrupts disabled, and after all other | ||
560 | devices have been suspended. On resume, they will be resumed before any other | ||
561 | devices, and also with interrupts disabled. These things occur in special | ||
562 | "sysdev_driver" phases, which affect only system devices. | ||
563 | |||
564 | Thus, after the suspend_noirq (or freeze_noirq or poweroff_noirq) phase, when | ||
565 | the non-boot CPUs are all offline and IRQs are disabled on the remaining online | ||
566 | CPU, then a sysdev_driver.suspend phase is carried out, and the system enters a | ||
567 | sleep state (or a system image is created). During resume (or after the image | ||
568 | has been created or loaded) a sysdev_driver.resume phase is carried out, IRQs | ||
569 | are enabled on the only online CPU, the non-boot CPUs are enabled, and the | ||
570 | resume_noirq (or thaw_noirq or restore_noirq) phase begins. | ||
571 | |||
572 | Code to actually enter and exit the system-wide low power state sometimes | ||
573 | involves hardware details that are only known to the boot firmware, and | ||
574 | may leave a CPU running software (from SRAM or flash memory) that monitors | ||
575 | the system and manages its wakeup sequence. | ||
576 | 537 | ||
577 | 538 | ||
578 | Device Low Power (suspend) States | 539 | Device Low Power (suspend) States |