sched_4bsd(9) [netbsd man page]

SCHED_4BSD(9)						   BSD Kernel Developer's Manual					     SCHED_4BSD(9)

NAME

     sched_4bsd -- The 4.4BSD thread scheduler

SYNOPSIS

     #include <sys/sched.h>

     void
     resetpriority(lwp_t *l);

     void
     sched_tick(struct cpu_info *ci);

     void
     sched_schedclock(lwp_t *l);

     void
     sched_pstats_hook(struct proc *p, int minslp);

     void
     sched_setrunnable(lwp_t *l);

     void
     updatepri(lwp_t *l);

DESCRIPTION

     The traditional 4.4BSD scheduler employs a ``multilevel feedback queues'' algorithm, favouring interactive, short-running threads to CPU-
     bound ones.

     resetpriority() recomputes the priority of a thread running in user mode.	If the resulting priority is higher than that of the current
     thread, a reschedule is arranged.

     sched_tick() gets called from hardclock(9) every 100ms to force a switch between equal priority threads.

     The priority of the current thread is adjusted through sched_schedclock().  The priority of a thread gets worse as it accumulates CPU time.

     sched_pstats_hook() gets called from sched_pstats() every Hz ticks in order to recompute the priorities of all threads.

     sched_setrunnable() checks if an LWP has slept for more than one second.  If so, its priority is updated by updatepri().

CODE REFERENCES

     The 4.4BSD scheduler subsystem is implemented within the file sys/kern/sched_4bsd.c.

SEE ALSO

     csf(9), hardclock(9), mi_switch(9), sched_m2(9), userret(9)

     Marshall Kirk McKusick, Keith Bostic, Michael J. Karels, and John S. Quarterman, The Design and Implementation of the 4.4BSD Operating
     System, Addison Wesley, 1996.

BSD
								   July 14, 2007							       BSD

Check Out this Related Man Page

SCHED_4BSD(4)						   BSD Kernel Interfaces Manual 					     SCHED_4BSD(4)

NAME

     sched_4bsd -- 4.4BSD scheduler

SYNOPSIS

     options SCHED_4BSD

DESCRIPTION

     The sched_4bsd scheduler is the traditional system scheduler, providing both high throughput and solid interactive response in the presence
     of load.

     The following sysctls are relevant to the operation of sched_4bsd:

     kern.sched.name
	     This read-only sysctl reports the name of the active scheduler.

     kern.sched.quantum
	     This read-write sysctl reports or sets the length of the quantum (in micro-seconds) granted to a thread.

     kern.sched.ipiwakeup.enabled
	     This read-write sysctl sets whether or not the scheduler will generate an inter-processor interrupt (IPI) to an idle CPU when a
	     thread is woken up.  Otherwise, idle CPUs will wait until the next clock tick before looking for new work.

     kern.sched.preemption
	     This read-only sysctl reports whether or not the kernel is configured to support preemption, which reduces the latency to run lower
	     priority threads on wakeup.

     Some sysctls will be available only on systems supporting SMP.

SEE ALSO

     sched_ule(4), sysctl(8)

HISTORY

     The sched_4bsd scheduler has been present, in various forms, since the inception of BSD.

BUGS

     While a highly robust and time-tested scheduler, sched_4bsd lacks specific knowledge of how to schedule advantageously in non-symmetric pro-
     cessor configurations, such as hyper-threading.

BSD
								 January 21, 2008							       BSD

Linux and UNIX Man Pages

sched_4bsd(9) [netbsd man page]

Check Out this Related Man Page