Caution: This version of this document is no longer maintained. For the latest documentation, see


Structure that describes scheduling parameters


#include <sched.h>

struct sched_param { 
    int32_t  sched_priority; 
    int32_t  sched_curpriority; 
    union { 
        int32_t  reserved[8]; 
        struct {    
            int32_t  __ss_low_priority;  
            int32_t  __ss_max_repl;  
            struct timespec     __ss_repl_period;   
            struct timespec     __ss_init_budget;   
        }           __ss;   
    }           __ss_un;    

#define sched_ss_low_priority   __ss_un.__ss.__ss_low_priority
#define sched_ss_max_repl       __ss_un.__ss.__ss_max_repl
#define sched_ss_repl_period    __ss_un.__ss.__ss_repl_period
#define sched_ss_init_budget    __ss_un.__ss.__ss_init_budget


You'll use the sched_param structure when you get or set the scheduling parameters for a thread or process.

You can use these functions to get the scheduling parameters:

You can use these functions to set the scheduling parameters:

The members of sched_param include:

When you get the scheduling parameters, this member reflects the priority that was assigned to the thread or process. It doesn't reflect any temporary adjustments due to priority inheritance.

When you set the scheduling parameters, set this member to the priority that you want to use. The priority must be between the minimum and maximum values returned by sched_get_priority_min() and sched_get_priority_max() for the scheduling policy.

When you get the scheduling parameters, this member is set to the priority that the thread or process is currently running at. This is the value that the kernel uses when making scheduling decisions.

When you set the scheduling parameters, this member is ignored.

The other members are used with sporadic scheduling. The following #define directives create the POSIX names that correspond to those members and should be used instead of accessing members directly.

The background or low priority for the thread that's executing.
The maximum number of times a replenishment will be scheduled, typically because of a blocking operation. After a thread has blocked this many times, it automatically drops to the low-priority level for the remainder of its execution until its execution budget is replenished.
The time that should be used for scheduling the replenishment of an execution budget after being blocked or having overrun the maximum number of replenishments. This time is used as an offset against the time that a thread is made READY.
The time that should be used for the thread's execution budget. As the thread runs at its high-priority level, its execution time is carved out of this budget. Once the budget is entirely depleted, the thread drops to its low-priority level, where, if possible because of priority arrangements, it can continue to run until the execution budget is replenished.

  • The sched_priority must always be higher than sched_ss_low_priority.
  • The sched_ss_max_repl must be smaller than SS_REPL_MAX.
  • The sched_ss_init_budget must be larger than sched_ss_repl_period.

For more information, see Scheduling algorithms in the Neutrino Microkernel chapter of the System Architecture guide.


This code shows a duty-cycle usage of the sporadic server thread:

#include <stdio.h>
#include <errno.h>
#include <sched.h>
#include <pthread.h>
#include <inttypes.h>
#include <sys/syspage.h>
#include <sys/neutrino.h>

/* 50 % duty cycle of 5 secs on 5 secs off */
struct timespec g_init_budget = { 5, 0 };
struct timespec g_repl_period = { 10, 0 };

#define MY_REPL_PERIOD g_repl_period
#define MY_INIT_BUDGET g_init_budget
#define MY_MAX_REPL 10


 Run a compute-bound thread (minimal blocking) to show the
 duty cycle.
void *st_duty_check(void *arg) {
    struct sched_param  params;
    uint64_t        stime, etime, cps;
    double          secs;
    int         ret, prio;
    int         prevprio, iterations;

    stime = ClockCycles();
    cps = SYSPAGE_ENTRY(qtime)->cycles_per_sec;
    iterations = 0;


    prevprio = -1;
    while(iterations < DUTY_CYCLE_LOOPS) {
        etime = ClockCycles();
        ret = pthread_getschedparam(pthread_self(), &prio,

        if(ret != 0) {
            printf("pthread_getschedparam() failed %d \n",
        } else if (prevprio != -1 && prevprio !=
                   params.sched_priority) {
            stime = etime - stime;
            secs = (double)stime / (double)cps;
            printf("pri %d (cur %d) %lld cycles %g secs\n", 
                stime, secs);
            stime = etime;
        prevprio = params.sched_priority;

    return NULL;

int main(int argc, char **argv) {
    struct sched_param params;
    pthread_attr_t     attr;
    pthread_t      thr;
    int        ret;

    /* Set the attribute structure with the sporadic values */
    printf("# Set sporadic attributes ...");
    ret = pthread_attr_setinheritsched(&attr,
    if(ret != 0) {
        printf("pthread_attr_setinheritsched() failed %d \n",
        return 1;

    ret = pthread_attr_setschedpolicy(&attr, SCHED_SPORADIC);
    if(ret != 0) {
        printf("pthread_attr_setschedpolicy() failed %d %d\n",
               ret, errno);
        return 1;

    params.sched_priority = MY_HIGH_PRIORITY;
    params.sched_ss_low_priority = MY_LOW_PRIORITY;
    memcpy(&params.sched_ss_init_budget, &MY_INIT_BUDGET,
    memcpy(&params.sched_ss_repl_period, &MY_REPL_PERIOD,
    params.sched_ss_max_repl = MY_MAX_REPL; 
    ret = pthread_attr_setschedparam(&attr, &params);
    if(ret != 0) {
        printf("pthread_attr_setschedparam() failed %d \n", errno);
        return 1;

    /* Create a sporadic thread to check the duty cycle */
    printf("# Creating thread duty cycle thread (%d changes) ... ",
    ret = pthread_create(&thr, &attr, st_duty_check, NULL);
    if(ret != 0) {
        printf("pthread_create() failed %d \n", errno);
        return 1;
    pthread_join(thr, NULL);

    return 0;

See also sched_getparam().


POSIX 1003.1 PS

See also:

pthread_attr_getschedparam(), pthread_attr_setschedparam(), pthread_getschedparam(), pthread_setschedparam(), sched_getparam(), sched_setparam(), sched_setscheduler(), SchedGet(), SchedSet(), ThreadCreate()

Scheduling algorithms in the Neutrino Microkernel chapter of the System Architecture guide