linux 进程调度

kernel/sched/core.c

一、主调度器

inlude/linux/sched.h中定义了struct task_struct {}。

进程调度最核心的两个问题是，什么时间放弃cpu，接下来由谁获得cpu。放弃cpu有两种方式，一种是主动调用schedule()，一种是周期调度器强行调度scheduler_tick()。

------------------------------------------------------------------------------------------------

pick_next_task();

判断当前cpu就绪队列rq-running中的进程数目是否与普通进程的就绪队列中的进程数目相同，如果相同就说明系统中全是普通进程，直接通过cfs算法的调度类的pick_next_task_fair（kernel/sched/fair.c）函数来从普通进程的就绪队列中寻找进程即可。  

否则，遍历调度类的链表，并从中选择一个优先级最高的进程。调度器类已经按如下顺序排序：rt_sched_class → fair_sched_class → idle_sched_class → NULL。

__schedule()调用了pick_next_task(); __schedule()是主调度器。

schedule()，do_task_dead()，schedule_idle()，preempt_schedule_common()，preemt_schedule_lock()，preempt_schedule_notrace()，preempt_schedule_irq()都调用了__schedule(); __schedule()函数主要是对当前task做善后工作，对新task做准备工作。

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

	cpu = smp_processor_id();//返回当前处理器的id号
	rq = cpu_rq(cpu);// 取该处理器上的运行队列
	prev = rq->curr;//获取当前正在执行的task

	schedule_debug(prev, preempt);//检查当前task并输出debug信息

	if (sched_feat(HRTICK) || sched_feat(HRTICK_DL))//取消为当前进程运行的hrtimer
		hrtick_clear(rq);

	local_irq_disable();//关中断
	rcu_note_context_switch(preempt);//更新全局状态，标识当前CPU发生上下文的切换。

	/*
	 * 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():
	 *
	 * __set_current_state(@state)		signal_wake_up()
	 * schedule()				  set_tsk_thread_flag(p, TIF_SIGPENDING)
	 *					  wake_up_state(p, state)
	 *   LOCK rq->lock			    LOCK p->pi_state
	 *   smp_mb__after_spinlock()		    smp_mb__after_spinlock()
	 *     if (signal_pending_state())	    if (p->state & @state)
	 *
	 * Also, the membarrier system call requires a full memory barrier
	 * after coming from user-space, before storing to rq->curr.
	 */
	rq_lock(rq, &rf);
	smp_mb__after_spinlock();//内存屏障

	/* Promote REQ to ACT */
	rq->clock_update_flags <<= 1;
	update_rq_clock(rq);//更新可运行队列时间

	switch_count = &prev->nivcsw;//记录当前进程切换的次数

	/*
	 * We must load prev->state once (task_struct::state is volatile), such
	 * that:
	 *
	 *  - we form a control dependency vs deactivate_task() below.
	 *  - ptrace_{,un}freeze_traced() can change ->state underneath us.
	 */
	prev_state = prev->state;//更新当前task状态
	if (!preempt && prev_state) {
		if (signal_pending_state(prev_state, prev)) {//当前进程有信号待处理，设置进程为就绪态
			prev->state = TASK_RUNNING;
		} else {
			prev->sched_contributes_to_load =
				(prev_state & TASK_UNINTERRUPTIBLE) &&
				!(prev_state & TASK_NOLOAD) &&
				!(prev->flags & PF_FROZEN);

			if (prev->sched_contributes_to_load)
				rq->nr_uninterruptible++; //uninterruptible 队列长度+1

			/*
			 * __schedule()			ttwu()
			 *   prev_state = prev->state;    if (p->on_rq && ...)
			 *   if (prev_state)		    goto out;
			 *     p->on_rq = 0;		  smp_acquire__after_ctrl_dep();
			 *				  p->state = TASK_WAKING
			 *
			 * Where __schedule() and ttwu() have matching control dependencies.
			 *
			 * After this, schedule() must not care about p->state any more.
			 */
			deactivate_task(rq, prev, DEQUEUE_SLEEP | DEQUEUE_NOCLOCK);//将其从运行队列中删除

			if (prev->in_iowait) {//如果正在进行IO操作,进行原子操作，delay IO操作
				atomic_inc(&rq->nr_iowait);
				delayacct_blkio_start();
			}
		}
		switch_count = &prev->nvcsw;//记录当前进程切换的次数
	}

	next = pick_next_task(rq, prev, &rf);//选择下一个task
	clear_tsk_need_resched(prev);//清除新task中待schedule标志
	clear_preempt_need_resched();
#ifdef CONFIG_SCHED_DEBUG
	rq->last_seen_need_resched_ns = 0;
#endif

	if (likely(prev != next)) {
		rq->nr_switches++;//就绪队列的切换计数
		/*
		 * RCU users of rcu_dereference(rq->curr) may not see
		 * changes to task_struct made by pick_next_task().
		 */
		RCU_INIT_POINTER(rq->curr, next);//大概是通过读写锁将next赋值给curr
		/*
		 * The membarrier system call requires each architecture
		 * to have a full memory barrier after updating
		 * rq->curr, before returning to user-space.
		 *
		 * Here are the schemes providing that barrier on the
		 * various architectures:
		 * - mm ? switch_mm() : mmdrop() for x86, s390, sparc, PowerPC.
		 *   switch_mm() rely on membarrier_arch_switch_mm() on PowerPC.
		 * - finish_lock_switch() for weakly-ordered
		 *   architectures where spin_unlock is a full barrier,
		 * - switch_to() for arm64 (weakly-ordered, spin_unlock
		 *   is a RELEASE barrier),
		 */
		++*switch_count;

		migrate_disable_switch(rq, prev);
		psi_sched_switch(prev, next, !task_on_rq_queued(prev));

		trace_sched_switch(preempt, prev, next);

		/* Also unlocks the rq: */
		rq = context_switch(rq, prev, next, &rf);//切换上下文
	} else {
		rq->clock_update_flags &= ~(RQCF_ACT_SKIP|RQCF_REQ_SKIP);

		rq_unpin_lock(rq, &rf);
		__balance_callbacks(rq);
		raw_spin_unlock_irq(&rq->lock);
	}
}

二、调度器类

每一种调度器类都必须实现struct sched_class. 其成员包括task_tick等。