Hooks

System Power Management Hooks. More...

Functions
void	pm_state_set (enum pm_state state, uint8_t substate_id)
	Put processor into a power state.
void	pm_state_exit_post_ops (enum pm_state state, uint8_t substate_id)
	Do any SoC or architecture specific post ops after sleep state exits.

Detailed Description

System Power Management Hooks.

Function Documentation

pm_state_exit_post_ops()

void pm_state_exit_post_ops	(	enum pm_state	state,
		uint8_t	substate_id )

#include <zephyr/pm/pm.h>

Do any SoC or architecture specific post ops after sleep state exits.

This function is a place holder to do any operations that may be needed after a sleep state exits. By default, it is called after system-managed devices have been resumed, before PM exit notifications and system clock idle-exit accounting, and while interrupts are still locked.

With CONFIG_PM_EARLY_SYSTEM_TIMER_RESUME, this hook runs before system clock idle-exit accounting and system-managed device resume. Platforms enabling this experimental sequence must ensure this hook does not depend on devices having been resumed, for example when releasing I/O retention or isolation.

Implementations that do not select CONFIG_PM_STATE_SET_IRQ_UNLOCKED must use this hook for hardware resume operations only. They must not unmask interrupts or otherwise dispatch pending wake-source ISRs from this hook; the kernel idle path restores the original interrupt state after PM resume housekeeping is complete.

Note

The system timer has not been restarted when this hook runs in either sequence. If the low-power state stopped it, code that waits on that timer, such as a timer-based k_busy_wait(), may never return.

As with pm_state_set, when system PM keeps interrupts locked across resume, this ordering covers only interrupts that arch_irq_lock() can mask. A zero-latency interrupt (IRQ_ZERO_LATENCY) may run during the resume window before this hook completes and must be PM-wake-safe (see pm_state_set).

Parameters

state	Power state.
substate_id	Power substate id.

pm_state_set()

void pm_state_set	(	enum pm_state	state,
		uint8_t	substate_id )

#include <zephyr/pm/pm.h>

Put processor into a power state.

This function implements the SoC specific details necessary to put the processor into available power states. Implementations that select CONFIG_PM_STATE_SET_IRQ_UNLOCKED may still follow the legacy behavior of unmasking interrupts from this hook. Other implementations must not unmask interrupts or otherwise dispatch pending wake-source ISRs from this hook. Architecture helpers may adjust interrupt state immediately around the low-power instruction, but the kernel idle path restores the original interrupt state after PM resume housekeeping is complete.

Note

When system PM keeps interrupts locked across resume, the locked-resume ordering guarantee covers only interrupts that arch_irq_lock() can mask. A zero-latency interrupt (IRQ_ZERO_LATENCY) is outside this ordering and must be PM-wake-safe, or its interrupt source must be masked or disabled while the system state does not allow the ISR to execute.

Parameters

state	Power state.
substate_id	Power substate id.