Linux System Programming 学习笔记(五) 进程管理

1. 进程是unix系统中两个最重要的基础抽象之一(另一个是文件)

A process is a running program

A thread is the unit of activity inside of a process

 

the virtualization of memory is associated with the process, the threads all share the same memory address space

 

2. pid

The idle process has the pid 0

The first process that the kernel executes after booting the system, called the init process, has the pid 1

 

默认情况下，内核规定一个最大的进程ID值：32768（16位有符号数）

内核使用严格的线性方式为进程分配PID，如果当前所有分配的PID值最大为17，那么下一次分配的PID值为18

进程层级：每一个进程(init进程除外)都是从另一个进程引生的

Children normally belong to the same process groups as their parents

 

3. 运行新进程

In Unix, the act of loading into memory and executing a program image is separate from the act of creating a new process

exec system call：loads a binary program into memory, replacing the previous contents of the address space, and begins execution of  the new program

fork system call：create a new process

to execute a new program in a new process ：first a fork to create a new process, and then an exec to load a new binary into that process

 

exec系列函数中，只有execve是系统调用，其余函数都是根据execve进行封装的C语言库函数

 

fork系统调用之后，子进程与父进程几乎完全相同，除了以下：

a.进程ID不同

b.挂起的信号被清除，并不被子进程继承

c.父进程的文件锁不被子进程继承

 

This “fork plus exec” combination is frequent and simple：

pid_t pid = fork();
if (pid == -) {
    fprintf(stderr, "fork error\n");
}

/* the child */
if (pid == ) {
    const char* args[] = {"grep", NULL};
    int ret = execv("bin/grep", args);
    if (ret == -) {
        fprintf(stderr, "execv error\n");
        exit(EXIT_FAILURE);
    }
}

4. copy-on-write

早期Unix中，实现fork非常简单，the kernel created copies of all internal data structures, duplicated the process's page table entries, and then performed a

page-by-page copy of the parent's address space into the child's new address space, 这样拷贝实现无疑是费时的

 

copy-on-write是一种减小资源复制开销的延迟策略。

如果多个用户请求对它所拥有的资源进行读操作，资源副本的拷贝是不需要进行的，每个用户只需要操作一个指向共同资源的指针即可。只要没有用户试图修改这个资源，那么副本的拷贝开销就可以避免。如果有一个用户确实需要修改这个资源，那么直到这个时候，资源才复制一份，复制的一份资源被交给需要修改的用户，其余所有用户继续共享原初的资源

 

在虚拟内存的具体示例中，copy-on-write是以内存页为单位进行的。

进程的内存页首先被标记为 read-only and copy-on-write, 如果有进程需要修改这个内存页，就产生页错误，就在这时，内核复制该内存页，并且内存页的copy-on-write标记被清除。

现代机器体系结构都硬件层面上支持copy-on-write

 

5. exit

当一个进程终止时，内核清除它包含的所有资源：分配的内存、打开的文件、信号量等

结束一个程序最经典的方式不是显式调用exit，而是 "falling off the end" of the program

 

当一个进程终止时，内核向其父进程发送 SIGCHLD 信号。默认情况下，该信号被忽略，父进程也可以调用 sigaction 捕捉该信号

 

6. wait

#include <sys/types.h>
#include <sys/wait.h>

/* wait() returns the pid of a terminated child or −1 on error. */
pid_t wait (int *status);

pid_t waitpid (pid_t pid, int *status, int options);

当pid==-1时，waitpid等待任一子进程，等同于 wait，即 wait (&status) 等同于 waitpid (−1, &status, 0)

 

7. system

#define _XOPEN_SOURCE    /* if we want WEXITSTATUS, etc. */
#include <stdlib.h>
int system (const char *command);

It is common to use system() to run a simple utility or shell script, often with the explicit goal of simply obtaining its return value

If command is NULL, system() returns a nonzero value if the shell /bin/sh is available, and 0 otherwise

 

here is a sample implementation for system():

/* my_system - sync spwans and waits for the command
 * /bin/sh -c <cmd> .
 *
 * Return -1 on error of any sort, or the exit code from
 * the launched process. Does not block or ignore any signals.
 */

int my_system(const char* cmd)
{
    int status;
    pid_t pid = fork();
    if (pid == -) {
        fprintf(stderr, "fork error\n");
        return -;
    } else if (pid == ) {
        const char* argv[];
        argv[] = "sh";
        argv[] = "-c";
        argv[] = "cmd";
        argv[] = NULL;
        execv("bin/sh", argv);
       exit(-);
    }
    if (waitpid(pid, &status, ) == -) {
        return -;
    } else if (WIFEXITED(status)) {
        return WEXITSTATUS(status);
    }
    return -;
}

8. 僵尸进程

a process that has terminated but has not yet been waited upon by its parent is called a “zombie.”

Zombie processes continue to consume system resources, although only a small percentage

如果一个父进程fork产生了子进程，那么该父进程就必须负责wait其子进程(回收僵尸子进程的所占资源)

如果一个父进程在终止之前没有wait其子进程呢？ 内核会视察其所有子进程，并将这些子进程挂在init进程之下(确保任何进程在任何时候都有一个父进程)，init进程会周期性wait其所有子进程，这些子进程最终还是会被回收资源，而不会长期以僵尸进程形式存在

 

9. Users and Groups

#include <sys/types.h>
#include <unistd.h>

int setuid (uid_t uid);
int setgid (gid_t gid);

uid_t getuid (void);
gid_t getgid (void);

Each process is a member of a process group，Process groups that contain more than one process are generally implementing job control

 

A command on the shell such as this one:

/*  one process group containing three processes */
$ cat ship-inventory.txt | grep booty | sort

 

10. 守护进程

A daemon is a process that runs in the background, not connecting to any controlling terminal

通常在系统启动时开始运行，并且以root权限运行，处理系统级任务，名字后缀一般为d

 

一个守护进程必须满足：a. 是init进程的子进程   b. 不允许与终端连接