Home  >  Article  >  System Tutorial  >  Wake and sleep of Linux processes

Wake and sleep of Linux processes

PHPz
PHPzforward
2024-02-10 16:40:121093browse

In Linux systems, only processes waiting for CPU time are called ready processes. They are placed in the run queue with the status flag TASK_RUNNING. When a running process uses up its time slice, the Linux kernel scheduler takes away control of the CPU and selects an appropriate process from the run queue to run.

Linux 进程的唤醒和睡眠

Of course, the process can also actively give up control of the CPU. schedule() The function is a scheduling function that can be actively called by a process to schedule other processes to occupy the CPU. Once the process that actively gives up the CPU is rescheduled to occupy the CPU, it will start execution from the position where it last stopped, that is, from the next line of code that calls schedule().

Sometimes, a process needs to wait for a specific event to occur, such as the completion of device initialization, the completion of an I/O operation, or the expiration of a timer. In this case, the process must be removed from the run queue and added to the wait queue, at which point the process goes to sleep.

There are two process sleep states in Linux:

  • One is an interruptible sleep state, with its status flag TASK_INTERRUPTIBLE.
  • The other is an uninterruptible sleep state, and its status flag is TASK_UNINTERRUPTIBLE.

A process in interruptible sleep state will sleep until a certain condition becomes true. For example, generating a hardware interrupt, releasing the system resources that the process is waiting for, or passing a signal can be the conditions to wake up the process. The uninterruptible sleep state is similar to the interruptible sleep state, but it has one exception, that is, the process that delivers the signal to this sleep state cannot change its state, that is, it does not respond to the signal to wake up. The uninterruptible sleep state is generally less used, but it is still very useful in some specific situations. For example, the process must wait and cannot be interrupted until a specific event occurs.

In modern Linux operating systems, processes generally enter the sleep state by calling schedule(). The following code demonstrates how to put a running process into the sleep state.

sleeping_task = current;
set_current_state(TASK_INTERRUPTIBLE);
schedule();
func1();
/* Rest of the code ... */

In the first statement, the program stores a process structure pointer sleeping_task, current is a macro, which points to the executing process structure.

set_current_state() Change the state of the process from execution state TASK_RUNNING to sleep state TASK_INTERRUPTIBLE. If schedule() is scheduled by a process with status TASK_RUNNING, then schedule() will schedule another process to occupy the CPU.

如果 schedule() 是被一个状态为 TASK_INTERRUPTIBLETASK_UNINTERRUPTIBLE 的进程调度,那么还有一个附加的步骤将被执行:当前执行的进程在另外一个进程被调度之前会被从运行队列中移出,这将导致正在运行的那个进程进入睡眠,因为它已经不在运行队列中了。

我们可以使用下面的这个函数将刚才那个进入睡眠的进程唤醒。

wake_up_process(sleeping_task);

在调用了 wake_up_process() 以后,这个睡眠进程的状态会被设置为 TASK_RUNNING,而且调度器会把它加入到运行队列中去。当然,这个进程只有在下次被调度器调度到的时候才能真正地投入运行。

无效唤醒

几乎在所有的情况下,进程都会在检查了某些条件之后,发现条件不满足才进入睡眠。可是有的时候进程却会在判定条件为真后开始睡眠,如果这样的话进程就会无限期地休眠下去,这就是所谓的无效唤醒问题。

在操作系统中,当多个进程都企图对共享数据进行某种处理,而 最后的结果又取决于进程运行的顺序时,就会发生竞争条件,这是操作系统中一个典型的问题,无效唤醒恰恰就是由于竞争条件导致的。

设想有两个进程A 和B,A 进程正在处理一个链表,它需要检查这个链表是否为空,如果不空就对链表里面的数据进行一些操作,同时B进程也在往这个链表添加节点。当这个链表是空的时候,由于无数据可操作,这时A进程就进入睡眠,当B进程向链表里面添加了节点之后它就唤醒A 进程,其代码如下:

A进程:
1 spin_lock(&list_lock);
2 if (list_empty(&list_head)) {
3     spin_unlock(&list_lock);
4     set_current_state(TASK_INTERRUPTIBLE);
5     schedule();
6     spin_lock(&list_lock);
7 }
8
9 /* Rest of the code ... */
10 spin_unlock(&list_lock);
B进程:
100 spin_lock(&list_lock);
101 list_add_tail(&list_head, new_node);
102 spin_unlock(&list_lock);
103 wake_up_process(processa_task);

这里会出现一个问题,假如当A进程执行到第3行后第4行前的时候,B进程被另外一个处理器调度投入运行。在这个时间片内,B进程执行完了它所有的指令,因此它试图唤醒A进程,而此时的A进程还没有进入睡眠,所以唤醒操作无效。

在这之后,A 进程继续执行,它会错误地认为这个时候链表仍然是空的,于是将自己的状态设置为 TASK_INTERRUPTIBLE 然后调用 schedule() 进入睡 眠。由于错过了B进程唤醒,它将会无限期的睡眠下去,这就是无效唤醒问题,因为即使链表中有数据需要处理,A 进程也还是睡眠了。

避免无效唤醒

如何避免无效唤醒问题呢?

我们发现无效唤醒主要发生在检查条件之后和进程状态被设置为睡眠状态之前,本来B进程的 wake_up_process() 提供了一次将A进程状态置为 TASK_RUNNING 的机会,可惜这个时候A进程的状态仍然是 TASK_RUNNING,所以 wake_up_process() 将A进程状态从睡眠状态转变为运行状态的努力 没有起到预期的作用。

要解决这个问题,必须使用一种保障机制使得判断链表为空和设置进程状态为睡眠状态成为一个不可分割的步骤才行,也就是必须消除竞争条 件产生的根源,这样在这之后出现的 wake_up_process() 就可以起到唤醒状态是睡眠状态的进程的作用了。

找到了原因后,重新设计一下A进程的代码结构,就可以避免上面例子中的无效唤醒问题了。

A进程:
1 set_current_state(TASK_INTERRUPTIBLE);
2 spin_lock(&list_lock);
3 if (list_empty(&list_head)) {
4     spin_unlock(&list_lock);
5     schedule();
6     spin_lock(&list_lock);
7 }
8 set_current_state(TASK_RUNNING);
9
10 /* Rest of the code ... */
11 spin_unlock(&list_lock);

可以看到,这段代码在测试条件之前就将当前执行进程状态转设置成 TASK_INTERRUPTIBLE 了,并且在链表不为空的情况下又将自己置为 TASK_RUNNING 状态。

这样一来如果B进程在A进程进程检查了链表为空以后调用 wake_up_process(),那么A进程的状态就会自动由原来 TASK_INTERRUPTIBLE 变成 TASK_RUNNING,此后即使进程又调用了 schedule(),由于它现在的状态是 TASK_RUNNING,所以仍然不会被从运行队列中移出,因而不会错误的进入睡眠,当然也就避免了无效唤醒问题。

Linux内核的例子

在Linux操作系统中,内核的稳定性至关重要,为了避免在Linux操作系统内核中出现无效唤醒问题,Linux内核在需要进程睡眠的时候应该使用类似如下的操作:

/* q 是我们希望睡眠的等待队列 */
DECLARE_WAITQUEUE(wait, current);
add_wait_queue(q, &wait);
set_current_state(TASK_INTERRUPTIBLE);
/* condition 是等待的条件 */
while (!condition) {
    schedule();
}
set_current_state(TASK_RUNNING);
remove_wait_queue(q, &wait);

上面的操作,使得进程通过下面的一系列步骤安全地将自己加入到一个等待队列中进行睡眠:首先调用 DECLARE_WAITQUEUE() 创建一个等待队列的项,然后调用 add_wait_queue() 把自己加入到等待队列中,并且将进程的状态设置为 TASK_INTERRUPTIBLE 或者 TASK_INTERRUPTIBLE

然后循环检查条件是否为真:如果是的话就没有必要睡眠,如果条件不为真,就调用 schedule()。当进程检查的条件满足后,进程又将自己设置为 TASK_RUNNING 并调用 remove_wait_queue() 将自己移出等待队列。

从上面可以看到,Linux的内核代码维护者也是在进程检查条件之前就设置进程的状态为睡眠状态,然后才循环检查条件。如果在进程开始睡眠之前条件就已经达成了,那么循环会退出并用 set_current_state() 将自己的状态设置为就绪,这样同样保证了进程不会存在错误的进入睡眠的倾向,当然也就不会导致出现无效唤醒问题。

下面让我们用 Linux 内核中的实例来看看其是如何避免无效睡眠的,这段代码出自 Linux2.6 的内核 (/kernel/sched.c):

/* Wait for kthread_stop */
set_current_state(TASK_INTERRUPTIBLE);
while (!kthread_should_stop()) {
    schedule();
    set_current_state(TASK_INTERRUPTIBLE);
}
__set_current_state(TASK_RUNNING);
return 0;

上面的这些代码属于迁移服务线程 migration_thread,这个线程不断地检查 kthread_should_stop(),直到 kthread_should_stop() 返回 1 它才可以退出循环,也就是说只要 kthread_should_stop() 返回 0 该进程就会一直睡眠。

从代码中我们可以看出,检查 kthread_should_stop() 确实是在进程的状态被置为 TASK_INTERRUPTIBLE 后才开始执行的。因此,如果在条件检查之后但是在 schedule() 之前有其他进程试图唤醒它,那么该进程的唤醒操作不会失效。

summary

Through the above discussion, it can be found that the key to avoid invalid wake-up of the process in Linux is to set the status of the process to TASK_INTERRUPTIBLE or TASK_UNINTERRUPTIBLE before the process checks the condition, and If the checked conditions are met, its status should be reset to TASK_RUNNING.

In this way, no matter whether the waiting conditions of the process are met or not, the process will not enter the sleep state by mistake because it is removed from the ready queue, thereby avoiding the problem of invalid wake-up.

The above is the detailed content of Wake and sleep of Linux processes. For more information, please follow other related articles on the PHP Chinese website!

Statement:
This article is reproduced at:lxlinux.net. If there is any infringement, please contact admin@php.cn delete