init进程【1】——init启动过程

http://blog.csdn.net/zhgxhuaa/article/details/22994095

欢迎转载，转载请注明：http://blog.csdn.net/zhgxhuaa

init启动过程

众所周知，Linux中的所有进程都是有init进程创建并运行的。首先Linux内核启动，然后在用户空间中启动init进程，再启动其他系统进程。在系统启动完成完成后，init将变为守护进程监视系统其他进程。Android是基于Linux的操作系统，所以init也是Android系统中用户空间的第一个进程，它的进程号是1。下面先简单的看一下init进程的启动过程。

@/kernel/goodfish/init/main.c

 static int __init kernel_init(void * unused)
 {
     /*
      * Wait until kthreadd is all set-up.
      */
     wait_for_completion(&kthreadd_done);
     /*
      * init can allocate pages on any node
      */
     set_mems_allowed(node_states[N_HIGH_MEMORY]);
     /*
      * init can run on any cpu.
      */
     set_cpus_allowed_ptr(current, cpu_all_mask);

     cad_pid = task_pid(current);

     smp_prepare_cpus(setup_max_cpus);

     do_pre_smp_initcalls();
     lockup_detector_init();

     smp_init();
     sched_init_smp();

     do_basic_setup();

     /* Open the /dev/console on the rootfs, this should never fail */
     if (sys_open((const char __user *) "/dev/console", O_RDWR, 0) < 0)
         printk(KERN_WARNING "Warning: unable to open an initial console.\n");

     (void) sys_dup(0);
     (void) sys_dup(0);
     /*
      * check if there is an early userspace init.  If yes, let it do all
      * the work
      */

     if (!ramdisk_execute_command)
         ramdisk_execute_command = "/init";

     if (sys_access((const char __user *) ramdisk_execute_command, 0) != 0) {
         ramdisk_execute_command = NULL;
         prepare_namespace();
     }

     /*
      * Ok, we have completed the initial bootup, and
      * we're essentially up and running. Get rid of the
      * initmem segments and start the user-mode stuff..
      */

     init_post();
     return 0;
 }

 /* This is a non __init function. Force it to be noinline otherwise gcc
  * makes it inline to init() and it becomes part of init.text section
  */
 static noinline int init_post(void)
 {
     /* need to finish all async __init code before freeing the memory */
     async_synchronize_full();
     free_initmem();
     mark_rodata_ro();
     system_state = SYSTEM_RUNNING;
     numa_default_policy();


     current->signal->flags |= SIGNAL_UNKILLABLE;

     if (ramdisk_execute_command) {
         run_init_process(ramdisk_execute_command);
         printk(KERN_WARNING "Failed to execute %s\n",
                 ramdisk_execute_command);
     }

     /*
      * We try each of these until one succeeds.
      *
      * The Bourne shell can be used instead of init if we are
      * trying to recover a really broken machine.
      */
     if (execute_command) {
         run_init_process(execute_command);
         printk(KERN_WARNING "Failed to execute %s.  Attempting "
                     "defaults...\n", execute_command);
     }
     run_init_process("/sbin/init");
     run_init_process("/etc/init");
     run_init_process("/bin/init");
     run_init_process("/bin/sh");

     panic("No init found.  Try passing init= option to kernel. "
           "See Linux Documentation/init.txt for guidance.");
 }

 static void run_init_process(const char *init_filename)
 {
     argv_init[0] = init_filename;
     kernel_execve(init_filename, argv_init, envp_init);
 }

在init_post()中会判断execute_command是否为空，如果不为空则执行run_init_process调用。execute_command的赋值在init_setup()中，所以这里应该注意在设置内核启动选项时，应设置为“ init=/init”，以便正常启动init进程，因为编译完Android后生成的文件系统中，init位于最顶层目录。

 <span style="font-size:14px;">static const char * argv_init[MAX_INIT_ARGS+2] = { "init", NULL, };</span>

 static int __init init_setup(char *str)
 {
     unsigned int i;

     execute_command = str;
     /*
      * In case LILO is going to boot us with default command line,
      * it prepends "auto" before the whole cmdline which makes
      * the shell think it should execute a script with such name.
      * So we ignore all arguments entered _before_ init=... [MJ]
      */
     for (i = 1; i < MAX_INIT_ARGS; i++)
         argv_init[i] = NULL;
     return 1;
 }
 __setup("init=", init_setup);

当根目录中不存在init时，或者未指定启动项“init=”时，内核会到/sbin、/etc、/bin目录下查找init。

了解了init进程的启动过程后，接下来看一下init进程都干了些什么？Android中的init进程与Linux不同，其职责可以归结如下：

• 作为守护进程
• 解析和执行init.rc文件
• 生成设备驱动节点
• 属性服务

init源码分析

init进程的入口函数是main，它的代码如下：

@/system/core/init/init.c

 int main(int argc, char **argv)
 {
     int fd_count = 0;
     struct pollfd ufds[4];
     char *tmpdev;
     char* debuggable;
     char tmp[32];
     int property_set_fd_init = 0;
     int signal_fd_init = 0;
     int keychord_fd_init = 0;
     bool is_charger = false;


     //启动ueventd
     if (!strcmp(basename(argv[0]), "ueventd"))
         return ueventd_main(argc, argv);


     //启动watchdogd
     if (!strcmp(basename(argv[0]), "watchdogd"))
         return watchdogd_main(argc, argv);


     /* clear the umask */
     umask(0);


         /* Get the basic filesystem setup we need put
          * together in the initramdisk on / and then we'll
          * let the rc file figure out the rest.
          */
     //创建并挂在启动所需的文件目录
     mkdir("/dev", 0755);
     mkdir("/proc", 0755);
     mkdir("/sys", 0755);


     mount("tmpfs", "/dev", "tmpfs", MS_NOSUID, "mode=0755");
     mkdir("/dev/pts", 0755);
     mkdir("/dev/socket", 0755);
     mount("devpts", "/dev/pts", "devpts", 0, NULL);
     mount("proc", "/proc", "proc", 0, NULL);
     mount("sysfs", "/sys", "sysfs", 0, NULL);


         /* indicate that booting is in progress to background fw loaders, etc */
     close(open("/dev/.booting", O_WRONLY | O_CREAT, 0000));//检测/dev/.booting文件是否可读写和创建


         /* We must have some place other than / to create the
          * device nodes for kmsg and null, otherwise we won't
          * be able to remount / read-only later on.
          * Now that tmpfs is mounted on /dev, we can actually
          * talk to the outside world.
          */
     open_devnull_stdio();//重定向标准输入/输出/错误输出到/dev/_null_
     klog_init();//log初始化
     property_init();//属性服务初始化


     //从/proc/cpuinfo中读取Hardware名，在后面的mix_hwrng_into_linux_rng_action函数中会将hardware的值设置给属性ro.hardware
     get_hardware_name(hardware, &revision);


     //导入并设置内核变量
     process_kernel_cmdline();


     //selinux相关，暂不分析
     union selinux_callback cb;
     cb.func_log = klog_write;
     selinux_set_callback(SELINUX_CB_LOG, cb);


     cb.func_audit = audit_callback;
     selinux_set_callback(SELINUX_CB_AUDIT, cb);


     selinux_initialize();
     /* These directories were necessarily created before initial policy load
      * and therefore need their security context restored to the proper value.
      * This must happen before /dev is populated by ueventd.
      */
     restorecon("/dev");
     restorecon("/dev/socket");
     restorecon("/dev/__properties__");
     restorecon_recursive("/sys");


     is_charger = !strcmp(bootmode, "charger");//关机充电相关，暂不做分析


     INFO("property init\n");
     if (!is_charger)
         property_load_boot_defaults();


     INFO("reading config file\n");
     init_parse_config_file("/init.rc");//解析init.rc配置文件


     /*
      * 解析完init.rc后会得到一系列的action等，下面的代码将执行处于early-init阶段的action。
      * init将action按照执行时间段的不同分为early-init、init、early-boot、boot。
      * 进行这样的划分是由于有些动作之间具有依赖关系，某些动作只有在其他动作完成后才能执行，所以就有了先后的区别。
      * 具体哪些动作属于哪个阶段是在init.rc中的配置决定的
      */
     action_for_each_trigger("early-init", action_add_queue_tail);


     queue_builtin_action(wait_for_coldboot_done_action, "wait_for_coldboot_done");
     queue_builtin_action(mix_hwrng_into_linux_rng_action, "mix_hwrng_into_linux_rng");
     queue_builtin_action(keychord_init_action, "keychord_init");
     queue_builtin_action(console_init_action, "console_init");


     /* execute all the boot actions to get us started */
     action_for_each_trigger("init", action_add_queue_tail);


     /* skip mounting filesystems in charger mode */
     if (!is_charger) {
         action_for_each_trigger("early-fs", action_add_queue_tail);
         action_for_each_trigger("fs", action_add_queue_tail);
         action_for_each_trigger("post-fs", action_add_queue_tail);
         action_for_each_trigger("post-fs-data", action_add_queue_tail);
     }


     /* Repeat mix_hwrng_into_linux_rng in case /dev/hw_random or /dev/random
      * wasn't ready immediately after wait_for_coldboot_done
      */
     queue_builtin_action(mix_hwrng_into_linux_rng_action, "mix_hwrng_into_linux_rng");


     queue_builtin_action(property_service_init_action, "property_service_init");
     queue_builtin_action(signal_init_action, "signal_init");
     queue_builtin_action(check_startup_action, "check_startup");


     if (is_charger) {
         action_for_each_trigger("charger", action_add_queue_tail);
     } else {
         action_for_each_trigger("early-boot", action_add_queue_tail);
         action_for_each_trigger("boot", action_add_queue_tail);
     }


         /* run all property triggers based on current state of the properties */
     queue_builtin_action(queue_property_triggers_action, "queue_property_triggers");




 #if BOOTCHART
     queue_builtin_action(bootchart_init_action, "bootchart_init");
 #endif


     for(;;) {//init进入无限循环
         int nr, i, timeout = -1;
         //检查action_queue列表是否为空。如果不为空则移除并执行列表头中的action
         execute_one_command();
         restart_processes();//重启已经死去的进程


         if (!property_set_fd_init && get_property_set_fd() > 0) {
             ufds[fd_count].fd = get_property_set_fd();
             ufds[fd_count].events = POLLIN;
             ufds[fd_count].revents = 0;
             fd_count++;
             property_set_fd_init = 1;
         }
         if (!signal_fd_init && get_signal_fd() > 0) {
             ufds[fd_count].fd = get_signal_fd();
             ufds[fd_count].events = POLLIN;
             ufds[fd_count].revents = 0;
             fd_count++;
             signal_fd_init = 1;
         }
         if (!keychord_fd_init && get_keychord_fd() > 0) {
             ufds[fd_count].fd = get_keychord_fd();
             ufds[fd_count].events = POLLIN;
             ufds[fd_count].revents = 0;
             fd_count++;
             keychord_fd_init = 1;
         }


         if (process_needs_restart) {
             timeout = (process_needs_restart - gettime()) * 1000;
             if (timeout < 0)
                 timeout = 0;
         }


         if (!action_queue_empty() || cur_action)
             timeout = 0;


 #if BOOTCHART
         if (bootchart_count > 0) {
             if (timeout < 0 || timeout > BOOTCHART_POLLING_MS)
                 timeout = BOOTCHART_POLLING_MS;
             if (bootchart_step() < 0 || --bootchart_count == 0) {
                 bootchart_finish();
                 bootchart_count = 0;
             }
         }
 #endif
         //等待事件发生
         nr = poll(ufds, fd_count, timeout);
         if (nr <= 0)
             continue;


         for (i = 0; i < fd_count; i++) {
             if (ufds[i].revents == POLLIN) {
                 if (ufds[i].fd == get_property_set_fd())//处理属性服务事件
                     handle_property_set_fd();
                 else if (ufds[i].fd == get_keychord_fd())//处理keychord事件
                     handle_keychord();
                 else if (ufds[i].fd == get_signal_fd())//处理
                     handle_signal();//处理SIGCHLD信号
             }
         }
     }


     return 0;
 }

main函数分析：

   if (!strcmp(basename(argv[0]), "ueventd"))
         return ueventd_main(argc, argv);

main函数一开始就会判断参数argv[0]的值是否等于“ueventd”，如果是就调用ueventd进程的入口函数ueventd_main()启动ueventd进程。这是怎么回事呢？当前正在启动的进程不是init吗？它的名称怎么可能会等于“ueventd”？所以这里有必要看一下ueventd的启动过程，ueventd是在init.rc中被启动的。

 on boot
 service ueventd /sbin/ueventd
     class core
     critical
     seclabel u:r:ueventd:s0

可以看出ueventd可执行文件位于/sbin/ueventd，在观察了/sbin/ueventd后我们发现，它只不过是 是可执行文件/init的一个符号链接文件，即应用程序ueventd和init运行的是同一个可执行文件。

所以，整个过程是这样的：内核启动完成之后，可执行文件/init首先会被执行，即init进程会首先被启动。init进程在启动的过程中，会对启动脚本/init.rc进行解析。在启动脚本/init.rc中，配置了一个ueventd进程，它对应的可执行文件为/sbin/ueventd，即ueventd进程加载的可执行文件也为/init（此时init中main函数的参数argv[0] = “/sbin/ueventd”）。因此，通过判断参数argv[0]的值，就可以知道当前正在启动的是init进程还是ueventd进程。      

PS：ueventd是一个守护进程，主要作用是接收uevent来创建或删除/dev/xxx(设备节点)，其实现位于@system/core/init/ueventd.c中。ueventd进程会通过一个socket接口来和内核通信，以便可以监控系统设备事件。

在开始所有的工作之前，main进程首先做的是创建并挂载启动所需的（其他的会在解析init.rc时创建）文件目录，如下所示：

     /* Get the basic filesystem setup we need put
      * together in the initramdisk on / and then we'll
      * let the rc file figure out the rest.
      */
 //创建并挂在启动所需的文件目录
 mkdir("/dev", 0755);
 mkdir("/proc", 0755);
 mkdir("/sys", 0755);

 mount("tmpfs", "/dev", "tmpfs", MS_NOSUID, "mode=0755");
 mkdir("/dev/pts", 0755);
 mkdir("/dev/socket", 0755);
 mount("devpts", "/dev/pts", "devpts", 0, NULL);
 mount("proc", "/proc", "proc", 0, NULL);
 mount("sysfs", "/sys", "sysfs", 0, NULL);

说明：

tmpfs是一种虚拟内存的文件系统，典型的tmpfs文件系统完全驻留在RAM中，读写速度远快于内存或硬盘文件系统。

/dev目录保存着硬件设备访问所需要的设备驱动程序。在Android中，将相关目录作用于tmpfs，可以大幅度提高设备访问的速度。

devpts是一种虚拟终端文件系统。

proc是一种虚拟文件系统，只存在于内存中，不占用外存空间。借助此文件系统，应用程序可以与内核内部数据结构进行交互。

sysfs是一种特殊的文件系统，在Linux 2.6中引入，用于将系统中的设备组织成层次结构，并向用户模式程序提供详细的内核数据结构信息，将proc、devpts、devfs三种文件系统统一起来。
编译Android系统源码时，在生成的根文件系统中，并不存在/dev、/proc、/sys这类目录，它们是系统运行时的目录，有init进程在运行中生成，当系统终止时，它们就会消失。上面的代码所形成的的文件层次结构为：

 <span style="font-size:14px;">        /* We must have some place other than / to create the
          * device nodes for kmsg and null, otherwise we won't be able to remount / read-only later on.
          * Now that tmpfs is mounted on /dev, we can actually talk to the outside world.
          */
     open_devnull_stdio();//重定向标准输入/输出/错误输出到/dev/_null_</span>

open_devnull_stdio()函数的作用是重定向标准输入/输出/错误输出到/dev/_null_，至于为什么要重定向的原因在注释中已经写明。open_devnull_stdio()的实现如下：

@system/core/init/util.c

 void open_devnull_stdio(void)
 {
     int fd;
     static const char *name = "/dev/__null__";
     if (mknod(name, S_IFCHR | 0600, (1 << 8) | 3) == 0) {
         fd = open(name, O_RDWR);
         unlink(name);
         if (fd >= 0) {
             dup2(fd, 0);
             dup2(fd, 1);
             dup2(fd, 2);
             if (fd > 2) {
                 close(fd);
             }
             return;
         }
     }

     exit(1);
 }

 <span style="font-size:14px;">klog_init();//log初始化</span>

klog_init()用于初始化log，通过其实现可以看出log被打印到/dev/__kmsg__文件中。主要在代码中最后通过fcntl和unlink使得/dev/__kmsg__不可被访问，这就保证了只有log程序才可以访问。

 void klog_init(void)
 {
     static const char *name = "/dev/__kmsg__";

     if (klog_fd >= 0) return; /* Already initialized */

     if (mknod(name, S_IFCHR | 0600, (1 << 8) | 11) == 0) {
         klog_fd = open(name, O_WRONLY);
         if (klog_fd < 0)
                 return;
         fcntl(klog_fd, F_SETFD, FD_CLOEXEC);
         unlink(name);
     }
 }

 property_init

属性服务初始化，这里先不深究，接下来会单独分析。

 <span style="font-size:14px;">    </span>//从/proc/cpuinfo中读取Hardware名，在后面的mix_hwrng_into_linux_rng_action函数中会将hardware的值设置给属性ro.hardware
     get_hardware_name(hardware, &revision);

get_hardware_name()函数的作用是从/proc/cpuinfo中获取Hardware和Revision的值，并保持到全局变量hardware和revision中。

下面的截图是在我的手机上的CPU info信息：

这里获取hardware信息有什么用呢？在main()函数后面的代码中，我们可以看见这样一句：

     //导入并设置内核变量
     process_kernel_cmdline();

下面看一下process_kernel_cmdline的实现：

@system/core/init/init.c

 static void process_kernel_cmdline(void)
 {
     /* don't expose the raw commandline to nonpriv processes */
     chmod("/proc/cmdline", 0440);

     /* first pass does the common stuff, and finds if we are in qemu.
      * second pass is only necessary for qemu to export all kernel params
      * as props.
      */
     import_kernel_cmdline(0, import_kernel_nv);
     if (qemu[0])
         import_kernel_cmdline(1, import_kernel_nv);

     /* now propogate the info given on command line to internal variables
      * used by init as well as the current required properties
      */
     export_kernel_boot_props();
 }

 static void export_kernel_boot_props(void)
 {
     char tmp[PROP_VALUE_MAX];

     ......

     /* if this was given on kernel command line, override what we read
      * before (e.g. from /proc/cpuinfo), if anything */
     ret = property_get("ro.boot.hardware", tmp);
     if (ret)
         strlcpy(hardware, tmp, sizeof(hardware));
     property_set("ro.hardware", hardware);

     snprintf(tmp, PROP_VALUE_MAX, "%d", revision);
     property_set("ro.revision", tmp);

     ......
 }

process_kernel_cmdline()函数用于导入和设置一些内核变量，在export_kernel_boot_props()中我们看见将hardware的值赋值给了属性"ro.hardware"。那这个赋值又是干什么的呢？我们再看一下main()函数，在解析init.rc配置文件的时候，有没有发现少了点什么？

 INFO("reading config file\n");
 init_parse_config_file("/init.rc");//解析init.rc配置文件

是的，在以前比较老的代码中（例如2.3和4.0）这里除了init.rc以外还会有一个与硬件相关的rc脚本，如下：

 snprintf(tmp, sizeof(tmp), "/init.%s.rc", hardware);
 init_parse_config_file(tmp);

那现在这段代码跑去哪里了呢？我们在init.rc中找到了它：

所以，之前设置的ro.hardware的值是在这里用的，在init.rc中用来导入init.${ro.hardware}.rc脚本，然后一起进行解析。与之前相比，这里只是方式变了，本质上还是一样的。

 INFO("reading config file\n");
 init_parse_config_file("/init.rc");//解析init.rc配置文件

 action_for_each_trigger("early-init", action_add_queue_tail);

 queue_builtin_action(wait_for_coldboot_done_action, "wait_for_coldboot_done");
 queue_builtin_action(mix_hwrng_into_linux_rng_action, "mix_hwrng_into_linux_rng");
 queue_builtin_action(keychord_init_action, "keychord_init");
 queue_builtin_action(console_init_action, "console_init");

 /* execute all the boot actions to get us started */
 action_for_each_trigger("init", action_add_queue_tail);

 /* skip mounting filesystems in charger mode */
 if (!is_charger) {
     action_for_each_trigger("early-fs", action_add_queue_tail);
     action_for_each_trigger("fs", action_add_queue_tail);
     action_for_each_trigger("post-fs", action_add_queue_tail);
     action_for_each_trigger("post-fs-data", action_add_queue_tail);
 }

 /* Repeat mix_hwrng_into_linux_rng in case /dev/hw_random or /dev/random
  * wasn't ready immediately after wait_for_coldboot_done
  */
 queue_builtin_action(mix_hwrng_into_linux_rng_action, "mix_hwrng_into_linux_rng");

 queue_builtin_action(property_service_init_action, "property_service_init");
 queue_builtin_action(signal_init_action, "signal_init");
 queue_builtin_action(check_startup_action, "check_startup");

 if (is_charger) {
     action_for_each_trigger("charger", action_add_queue_tail);
 } else {
     action_for_each_trigger("early-boot", action_add_queue_tail);
     action_for_each_trigger("boot", action_add_queue_tail);
 }

     /* run all property triggers based on current state of the properties */
 queue_builtin_action(queue_property_triggers_action, "queue_property_triggers");


 BOOTCHART
 queue_builtin_action(bootchart_init_action, "bootchart_init");
 if

这部分代码用于解析init.rc脚本，并触发执行解析生成的action。这部分后面单独进行分析。

在main()函数的最后，init进入了一个无限循环，并等待一些事情的发生。即：在执行完前面的初始化工作以后，init变为一个守护进程。init所关心的事件有三类：属性服务事件、keychord事件和SIGNAL，当有这三类事件发生时，init进程会调用相应的handle函数进行处理。