Linux进程核心调度器之主调度器schedule--Linux进程的管理与调度(十九）

/* Attach to any functions which should be ignored in wchan output. */
#define __sched         __attribute__((__section__(".sched.text")))

void __sched some_function(args, ...)
{
    ......
    schedule();
    ......
}

asmlinkage __visible void __sched schedule(void)
{

    /*  获取当前的进程  */
    struct task_struct *tsk = current;

    /*  避免死锁 */
    sched_submit_work(tsk);
    do {
        preempt_disable();                                  /*  关闭内核抢占  */
        __schedule(false);                                  /*  完成调度  */
        sched_preempt_enable_no_resched();                  /*  开启内核抢占  */
    } while (need_resched());   /*  如果该进程被其他进程设置了TIF_NEED_RESCHED标志，则函数重新执行进行调度    */
}
EXPORT_SYMBOL(schedule);

static inline void sched_submit_work(struct task_struct *tsk)
{
    /*  检测tsk->state是否为0 （runnable）, 若为运行态时则返回，
     *   tsk_is_pi_blocked(tsk),检测tsk的死锁检测器是否为空，若非空的话就return
    if (!tsk->state || tsk_is_pi_blocked(tsk))
        return;
    /*
     * If we are going to sleep and we have plugged IO queued,
     * make sure to submit it to avoid deadlocks.
     */
    if (blk_needs_flush_plug(tsk))  /*  然后检测是否需要刷新plug队列，用来避免死锁  */
        blk_schedule_flush_plug(tsk);
}

#define preempt_disable() \
do { \
    preempt_count_inc(); \
    barrier(); \
} while (0)

#define sched_preempt_enable_no_resched() \
do { \
    barrier(); \
    preempt_count_dec(); \
} while (0)

static void __sched notrace __schedule(bool preempt)
{
    struct task_struct *prev, *next;
    unsigned long *switch_count;
    struct rq *rq;
    int cpu;

    /*  ==1==  
        找到当前cpu上的就绪队列rq
        并将正在运行的进程curr保存到prev中  */
    cpu = smp_processor_id();
    rq = cpu_rq(cpu);
    prev = rq->curr;

    /*
     * do_exit() calls schedule() with preemption disabled as an exception;
     * however we must fix that up, otherwise the next task will see an
     * inconsistent (higher) preempt count.
     *
     * It also avoids the below schedule_debug() test from complaining
     * about this.
     */
    if (unlikely(prev->state == TASK_DEAD))
        preempt_enable_no_resched_notrace();

    /*  如果禁止内核抢占，而又调用了cond_resched就会出错
     *  这里就是用来捕获该错误的  */
    schedule_debug(prev);

    if (sched_feat(HRTICK))
        hrtick_clear(rq);

    /*  关闭本地中断  */
    local_irq_disable();

    /*  更新全局状态，
     *  标识当前CPU发生上下文的切换  */
    rcu_note_context_switch();

    /*
     * Make sure that signal_pending_state()->signal_pending() below
     * can't be reordered with __set_current_state(TASK_INTERRUPTIBLE)
     * done by the caller to avoid the race with signal_wake_up().
     */
    smp_mb__before_spinlock();
    /*  锁住该队列  */
    raw_spin_lock(&rq->lock);
    lockdep_pin_lock(&rq->lock);

    rq->clock_skip_update <<= 1; /* promote REQ to ACT */

    /*  切换次数记录, 默认认为非主动调度计数(抢占)  */
    switch_count = &prev->nivcsw;

    /*
     *  scheduler检查prev的状态state和内核抢占表示
     *  如果prev是不可运行的, 并且在内核态没有被抢占
     *  
     *  此时当前进程不是处于运行态, 并且不是被抢占
     *  此时不能只检查抢占计数
     *  因为可能某个进程(如网卡轮询)直接调用了schedule
     *  如果不判断prev->stat就可能误认为task进程为RUNNING状态
     *  到达这里，有两种可能，一种是主动schedule, 另外一种是被抢占
     *  被抢占有两种情况, 一种是时间片到点, 一种是时间片没到点
     *  时间片到点后, 主要是置当前进程的need_resched标志
     *  接下来在时钟中断结束后, 会preempt_schedule_irq抢占调度
     *  
     *  那么我们正常应该做的是应该将进程prev从就绪队列rq中删除, 
     *  但是如果当前进程prev有非阻塞等待信号, 
     *  并且它的状态是TASK_INTERRUPTIBLE
     *  我们就不应该从就绪队列总删除它 
     *  而是配置其状态为TASK_RUNNING, 并且把他留在rq中

    /*  如果内核态没有被抢占, 并且内核抢占有效
        即是否同时满足以下条件：
        1  该进程处于停止状态
        2  该进程没有在内核态被抢占 */
    if (!preempt && prev->state)
    {

        /*  如果当前进程有非阻塞等待信号，并且它的状态是TASK_INTERRUPTIBLE  */
        if (unlikely(signal_pending_state(prev->state, prev)))
        {
            /*  将当前进程的状态设为：TASK_RUNNING  */
            prev->state = TASK_RUNNING;
        }
        else   /*  否则需要将prev进程从就绪队列中删除*/
        {
            /*  将当前进程从runqueue(运行队列)中删除  */
            deactivate_task(rq, prev, DEQUEUE_SLEEP);

            /*  标识当前进程不在runqueue中  */
            prev->on_rq = 0;

            /*
             * If a worker went to sleep, notify and ask workqueue
             * whether it wants to wake up a task to maintain
             * concurrency.
             */
            if (prev->flags & PF_WQ_WORKER) {
                struct task_struct *to_wakeup;

                to_wakeup = wq_worker_sleeping(prev);
                if (to_wakeup)
                    try_to_wake_up_local(to_wakeup);
            }
        }
        /*  如果不是被抢占的，就累加主动切换次数  */
        switch_count = &prev->nvcsw;
    }

    /*  如果prev进程仍然在就绪队列上没有被删除  */
    if (task_on_rq_queued(prev))
        update_rq_clock(rq);  /*  跟新就绪队列的时钟  */

    /*  挑选一个优先级最高的任务将其排进队列  */
    next = pick_next_task(rq, prev);
    /*  清除pre的TIF_NEED_RESCHED标志  */
    clear_tsk_need_resched(prev);
    /*  清楚内核抢占标识  */
    clear_preempt_need_resched();

    rq->clock_skip_update = 0;

    /*  如果prev和next非同一个进程  */
    if (likely(prev != next))
    {
        rq->nr_switches++;  /*  队列切换次数更新  */
        rq->curr = next;    /*  将next标记为队列的curr进程  */
        ++*switch_count;    /* 进程切换次数更新  */

        trace_sched_switch(preempt, prev, next);
        /*  进程之间上下文切换    */
        rq = context_switch(rq, prev, next); /* unlocks the rq */
    }
    else    /*  如果prev和next为同一进程，则不进行进程切换  */
    {
        lockdep_unpin_lock(&rq->lock);
        raw_spin_unlock_irq(&rq->lock);
    }

    balance_callback(rq);
}
STACK_FRAME_NON_STANDARD(__schedule); /* switch_to() */

next = pick_next_task(rq);

/*
 * Pick up the highest-prio task:
 */
static inline struct task_struct *
pick_next_task(struct rq *rq, struct task_struct *prev)
{
    const struct sched_class *class = &fair_sched_class;
    struct task_struct *p;

    /*
     * Optimization: we know that if all tasks are in
     * the fair class we can call that function directly:
     *
     * 如果待被调度的进程prev是隶属于CFS的普通非实时进程
     * 而当前cpu的全局就绪队列rq中的进程数与cfs_rq的进程数相等
     * 则说明当前cpu上的所有进程都是由cfs调度的普通非实时进程
     *
     * 那么我们选择最优进程的时候
     * 就只需要调用cfs调度器类fair_sched_class的选择函数pick_next_task
     * 就可以找到最优的那个进程p
     */
    /*  如果当前所有的进程都被cfs调度, 没有实时进程  */
    if (likely(prev->sched_class == class &&
           rq->nr_running == rq->cfs.h_nr_running))
    {
        /*  调用cfs的选择函数pick_next_task找到最优的那个进程p*/
        p = fair_sched_class.pick_next_task(rq, prev);
        /*  #define RETRY_TASK ((void *)-1UL)有被其他调度气找到合适的进程  */
        if (unlikely(p == RETRY_TASK))
            goto again; /*  则遍历所有的调度器类找到最优的进程 */

        /* assumes fair_sched_class->next == idle_sched_class */
        if (unlikely(!p))   /*  如果没有进程可被调度  */
            p = idle_sched_class.pick_next_task(rq, prev); /*  则调度idle进程  */

        return p;
    }

/*  进程中所有的调度器类, 是通过next域链接域链接在一起的
 *  调度的顺序为stop -> dl -> rt -> fair -> idle 
 *  again出的循环代码会遍历他们找到一个最优的进程  */
again:
    for_each_class(class)
    {
        p = class->pick_next_task(rq, prev);
        if (p)
        {
            if (unlikely(p == RETRY_TASK))
                goto again;
            return p;
        }
    }

    BUG(); /* the idle class will always have a runnable task */
}

#define sched_class_highest (&stop_sched_class)
#define for_each_class(class) \
   for (class = sched_class_highest; class; class = class->next)

extern const struct sched_class stop_sched_class;
extern const struct sched_class dl_sched_class;
extern const struct sched_class rt_sched_class;
extern const struct sched_class fair_sched_class;
extern const struct sched_class idle_sched_class;

rev->sched_class == class && rq->nr_running == rq->cfs.h_nr_running

//  http://lxr.free-electrons.com/source/tools/virtio/linux/kernel.h?v=4.6#L91
 #ifndef likely
 # define likely(x)     (__builtin_expect(!!(x), 1))
 #endif

 #ifndef unlikely
 # define unlikely(x)   (__builtin_expect(!!(x), 0))
 #endif

/*
 * Saving eflags is important. It switches not only IOPL between tasks,
 * it also protects other tasks from NT leaking through sysenter etc.
*/
#define switch_to(prev, next, last)

// http://lxr.free-electrons.com/source/include/linux/sched.h?v=4.6#L3093
static __always_inline bool need_resched(void)
{
    return unlikely(tif_need_resched());
}

// http://lxr.free-electrons.com/source/include/linux/thread_info.h?v=4.6#L106
#define tif_need_resched() test_thread_flag(TIF_NEED_RESCHED)

do {
        preempt_disable();                                  /*  关闭内核抢占  */
        __schedule(false);                                  /*  完成调度  */
        sched_preempt_enable_no_resched();                  /*  开启内核抢占  */
    } while (need_resched());   /*  如果该进程被其他进程设置了TIF_NEED_RESCHED标志，则函数重新执行进行调度    */