check_asym_packing(9) [centos man page]

CHECK_ASYM_PACKING(9)						   Driver Basics					     CHECK_ASYM_PACKING(9)

NAME

       check_asym_packing - Check to see if the group is packed into the sched doman.

SYNOPSIS

       int check_asym_packing(struct lb_env * env, struct sd_lb_stats * sds);

ARGUMENTS

       env
	   The load balancing environment.

       sds
	   Statistics of the sched_domain which is to be packed

DESCRIPTION

       This is primarily intended to used at the sibling level. Some cores like POWER7 prefer to use lower numbered SMT threads. In the case of
       POWER7, it can move to lower SMT modes only when higher threads are idle. When in lower SMT modes, the threads will perform better since
       they share less core resources. Hence when we have idle threads, we want them to be the higher ones.

       This packing function is run on idle threads. It checks to see if the busiest CPU in this domain (core in the P7 case) has a higher CPU
       number than the packing function is being run on. Here we are assuming lower CPU number will be equivalent to lower a SMT thread number.

RETURN

       1 when packing is required and a task should be moved to this CPU. The amount of the imbalance is returned in *imbalance.

COPYRIGHT

Kernel Hackers Manual 3.10					     June 2014						     CHECK_ASYM_PACKING(9)

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

NAME

     choosethread, procrunnable, remrunqueue, setrunqueue -- manage the queue of runnable processes

SYNOPSIS

     #include <sys/param.h>
     #include <sys/proc.h>

     extern struct rq itqueues[];
     extern struct rq rtqueues[];
     extern struct rq queues[];
     extern struct rq idqueues[];

     struct thread *
     choosethread(void);

     int
     procrunnable(void);

     void
     remrunqueue(struct thread *td);

     void
     setrunqueue(struct thread *td);

DESCRIPTION

     The run queue consists of four priority queues: itqueues for interrupt threads, rtqueues for realtime priority processes, queues for time
     sharing processes, and idqueues for idle priority processes.  Each priority queue consists of an array of NQS queue header structures.  Each
     queue header identifies a list of runnable processes of equal priority.  Each queue also has a single word that contains a bit mask identify-
     ing non-empty queues to assist in selecting a process quickly.  These are named itqueuebits, rtqueuebits, queuebits, and idqueuebits.  The
     run queues are protected by the sched_lock mutex.

     procrunnable() returns zero if there are no runnable processes other than the idle process.  If there is at least one runnable process other
     than the idle process, it will return a non-zero value.  Note that the sched_lock mutex does not need to be held when this function is
     called.  There is a small race window where one CPU may place a process on the run queue when there are currently no other runnable processes
     while another CPU is calling this function.  In that case the second CPU will simply travel through the idle loop one additional time before
     noticing that there is a runnable process.  This works because idle CPUs are not halted in SMP systems.  If idle CPUs are halted in SMP sys-
     tems, then this race condition might have more serious repercussions in the losing case, and procrunnable() may have to require that the
     sched_lock mutex be acquired.

     choosethread() returns the highest priority runnable thread.  If there are no runnable threads, then the idle thread is returned.	This func-
     tion is called by cpu_switch() and cpu_throw() to determine which thread to switch to.  choosethread() must be called with the sched_lock
     mutex held.

     setrunqueue() adds the thread td to the tail of the appropriate queue in the proper priority queue.  The thread must be runnable, i.e. p_stat
     must be set to SRUN.  This function must be called with the sched_lock mutex held.

     remrunqueue() removes thread td from its run queue.  If td is not on a run queue, then the kernel will panic(9).  This function must be
     called with the sched_lock mutex held.

SEE ALSO

     cpu_switch(9), scheduler(9), sleepqueue(9)

BSD
								  August 15, 2010							       BSD