Architecture-specific SMP APIs

Typedefs
typedef void(*	arch_cpustart_t) (void *data)
	Per-cpu entry function.

Functions
void	arch_cpu_start (int cpu_num, k_thread_stack_t *stack, int sz, arch_cpustart_t fn, void *arg)
	Start a numbered CPU on a MP-capable system.
bool	arch_cpu_active (int cpu_num)
	Return CPU power status.
static struct _cpu *	arch_curr_cpu (void)
	Return the CPU struct for the currently executing CPU.
static uint32_t	arch_proc_id (void)
	Processor hardware ID.
void	arch_sched_ipi (void)
	Broadcast an interrupt to all CPUs.
int	arch_smp_init (void)
static unsigned int	arch_num_cpus (void)
	Returns the number of CPUs.

Detailed Description

Typedef Documentation

arch_cpustart_t

typedef void(* arch_cpustart_t) (void *data)

#include <zephyr/arch/arch_interface.h>

Per-cpu entry function.

Parameters

data	context parameter, implementation specific

Function Documentation

arch_cpu_active()

bool arch_cpu_active

(

int

cpu_num

)

#include <zephyr/arch/arch_interface.h>

Return CPU power status.

Parameters

cpu_num

Integer number of the CPU

arch_cpu_start()

void arch_cpu_start	(	int	cpu_num,
		k_thread_stack_t *	stack,
		int	sz,
		arch_cpustart_t	fn,
		void *	arg
	)

#include <zephyr/arch/arch_interface.h>

Start a numbered CPU on a MP-capable system.

This starts and initializes a specific CPU. The main thread on startup is running on CPU zero, other processors are numbered sequentially. On return from this function, the CPU is known to have begun operating and will enter the provided function. Its interrupts will be initialized but disabled such that irq_unlock() with the provided key will work to enable them.

Normally, in SMP mode this function will be called by the kernel initialization and should not be used as a user API. But it is defined here for special-purpose apps which want Zephyr running on one core and to use others for design-specific processing.

Parameters

cpu_num	Integer number of the CPU
stack	Stack memory for the CPU
sz	Stack buffer size, in bytes
fn	Function to begin running on the CPU.
arg	Untyped argument to be passed to "fn"

arch_curr_cpu()

static struct _cpu * arch_curr_cpu ( void  )

inlinestatic

#include <zephyr/arch/arch_interface.h>

Return the CPU struct for the currently executing CPU.

arch_num_cpus()

static unsigned int arch_num_cpus ( void  )

inlinestatic

#include <zephyr/arch/arch_interface.h>

Returns the number of CPUs.

For most systems this will be the same as CONFIG_MP_MAX_NUM_CPUS, however some systems may determine this at runtime instead.

Returns

the number of CPUs

arch_proc_id()

static uint32_t arch_proc_id ( void  )

inlinestatic

#include <zephyr/arch/arch_interface.h>

Processor hardware ID.

Most multiprocessor architectures have a low-level unique ID value associated with the current CPU that can be retrieved rapidly and efficiently in kernel context. Note that while the numbering of the CPUs is guaranteed to be unique, the values are platform-defined. In particular, they are not guaranteed to match Zephyr's own sequential CPU IDs (even though on some platforms they do).

Note

There is an inherent race with this API: the system may preempt the current thread and migrate it to another CPU before the value is used. Safe usage requires knowing the migration is impossible (e.g. because the code is in interrupt context, holds a spinlock, or cannot migrate due to k_cpu_mask state).

Returns

Unique ID for currently-executing CPU

arch_sched_ipi()

void arch_sched_ipi

(

void

)

#include <zephyr/arch/arch_interface.h>

Broadcast an interrupt to all CPUs.

This will invoke z_sched_ipi() on other CPUs in the system.

arch_smp_init()

int arch_smp_init

(

void

)

#include <zephyr/arch/arch_interface.h>