Android6.0系统启动流程分析
Android6.0系统启动流程分析一：init进程
阳光玻璃杯
https://blog.****.net/u011913612/article/details/53204253
到了Android6.0,Init进程使用c++来写了，不过没有关系，它和c写的init没有太大的区别。
Init进程的入口代码是：system\core\init\init.cpp
main函数：
int main(int argc, char** argv) {
if (!strcmp(basename(argv[0]), "ueventd")) {
return ueventd_main(argc, argv);
}
if (!strcmp(basename(argv[0]), "watchdogd")) {
return watchdogd_main(argc, argv);
}
// Clear the umask.
umask(0);
add_environment("PATH", _PATH_DEFPATH);
bool is_first_stage = (argc == 1) || (strcmp(argv[1], "--second-stage") != 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.
if (is_first_stage) {
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);
}
// 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();
klog_init();
klog_set_level(KLOG_NOTICE_LEVEL);
NOTICE("init%s started!\n", is_first_stage ? "" : " second stage");
if (!is_first_stage) {
// Indicate that booting is in progress to background fw loaders, etc.
close(open("/dev/.booting", O_WRONLY | O_CREAT | O_CLOEXEC, 0000));
property_init();
// If arguments are passed both on the command line and in DT,
// properties set in DT always have priority over the command-line ones.
process_kernel_dt();
process_kernel_cmdline();
// Propogate the kernel variables to internal variables
// used by init as well as the current required properties.
export_kernel_boot_props();
}
// Set up SELinux, including loading the SELinux policy if we're in the kernel domain.
selinux_initialize(is_first_stage);
// If we're in the kernel domain, re-exec init to transition to the init domain now
// that the SELinux policy has been loaded.
if (is_first_stage) {
if (restorecon("/init") == -1) {
ERROR("restorecon failed: %s\n", strerror(errno));
security_failure();
}
char* path = argv[0];
char* args[] = { path, const_cast<char*>("--second-stage"), nullptr };
if (execv(path, args) == -1) {
ERROR("execv(\"%s\") failed: %s\n", path, strerror(errno));
security_failure();
}
}
// 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.
INFO("Running restorecon...\n");
restorecon("/dev");
restorecon("/dev/socket");
restorecon("/dev/__properties__");
restorecon_recursive("/sys");
epoll_fd = epoll_create1(EPOLL_CLOEXEC);
if (epoll_fd == -1) {
ERROR("epoll_create1 failed: %s\n", strerror(errno));
exit(1);
}
signal_handler_init();
property_load_boot_defaults();
start_property_service();
init_parse_config_file("/init.rc");
action_for_each_trigger("early-init", action_add_queue_tail);
// Queue an action that waits for coldboot done so we know ueventd has set up all of /dev...
queue_builtin_action(wait_for_coldboot_done_action, "wait_for_coldboot_done");
// ... so that we can start queuing up actions that require stuff from /dev.
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");
// Trigger all the boot actions to get us started.
action_for_each_trigger("init", 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");
// Don't mount filesystems or start core system services in charger mode.
char bootmode[PROP_VALUE_MAX];
if (property_get("ro.bootmode", bootmode) > 0 && strcmp(bootmode, "charger") == 0) {
action_for_each_trigger("charger", action_add_queue_tail);
} else {
action_for_each_trigger("late-init", 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");
while (true) {
if (!waiting_for_exec) {
execute_one_command();
restart_processes();
}
int timeout = -1;
if (process_needs_restart) {
timeout = (process_needs_restart - gettime()) * 1000;
if (timeout < 0)
timeout = 0;
}
if (!action_queue_empty() || cur_action) {
timeout = 0;
}
bootchart_sample(&timeout);
epoll_event ev;
int nr = TEMP_FAILURE_RETRY(epoll_wait(epoll_fd, &ev, 1, timeout));
if (nr == -1) {
ERROR("epoll_wait failed: %s\n", strerror(errno));
} else if (nr == 1) {
((void (*)()) ev.data.ptr)();
}
}
return 0;
}
1.这个函数是否往下执行取决于传入的参数，如果第0个参数的basename为ueventd，则执行ueventd_main(argc, argv);如果basename为watchdogd_main，则执行watchdogd_main(argc, argv);只有basename不为这二者时，才会继续往下执行。
2.如果argv[1]不为”–second-stage”或者只有一个参数的话，那么is_first_stage就为true,就会创建/dev/pts和”/dev/socket”两个设备文件节点，并挂载一个文件系统。可以看出来init进程分两个阶段，不同的阶段有不同的行为。具体的内涵鄙人还没搞明白。
3.启动属性服务。创建一个socket，并在之后的死循环中监听这个socket返回的文件描述符。
3.解析init.rc。这个过程也是我最感兴趣的，也是最重要的复杂的。
4.对各个阶段的action排序。
5.进入死循环。
6.第一次进入死循环后，action_queue里面有很多时间，因此需要不断调用execute_one_command来执行命令。此时，action_queue_empty为假，timeout 为0，init线程不会在epoll_wait方法中休眠，因为设置的timeout=0哦，这一点曾一度困扰了我。
7.所有的命令执行完后，init进程进入休眠，监听property_set_fd和signal_read_fd两个文件描述符，一点他们有事件过来，立刻被唤醒，进而做事件处理。
init.rc梳理
在我们分析init.rc的解析过程之前，我们还需要先对init.rc有个基本的认识。
先看一张我根据理解绘制的图：
从图来看，init.rc主要有section组成，section由on,import,section三个关键字标示。其中on标示的section叫做action。
import就不用说了，和c语言中的include功能有点类似。
service格式如下
service <name> <pathname> [ <argument> ]*
<option>
<option>
...
action后面会跟一个触发器，然后另起一行开始放置命令（command）,格式如下：
on <trigger>
<command>
<command>
<command>
跟在service后面的是option,跟在action后面的是command.command都会对应一个处理函数，定义在keywords.h中：
...
KEYWORD(loglevel, COMMAND, 1, do_loglevel)
KEYWORD(mkdir, COMMAND, 1, do_mkdir)
KEYWORD(mount_all, COMMAND, 1, do_mount_all)
KEYWORD(mount, COMMAND, 3, do_mount)
...
命名也是很有规则的。比如mkdir,对应的函数就是do_mkdir。我们看看do_mkdir做了什么：
int do_mkdir(int nargs, char **args)
{
mode_t mode = 0755;
int ret;
/* mkdir <path> [mode] [owner] [group] */
if (nargs >= 3) {
mode = strtoul(args[2], 0, 8);
}
ret = make_dir(args[1], mode);
/* chmod in case the directory already exists */
if (ret == -1 && errno == EEXIST) {
ret = fchmodat(AT_FDCWD, args[1], mode, AT_SYMLINK_NOFOLLOW);
}
if (ret == -1) {
return -errno;
}
if (nargs >= 4) {
uid_t uid = decode_uid(args[3]);
gid_t gid = -1;
if (nargs == 5) {
gid = decode_uid(args[4]);
}
if (lchown(args[1], uid, gid) == -1) {
return -errno;
}
/* chown may have cleared S_ISUID and S_ISGID, chmod again */
if (mode & (S_ISUID | S_ISGID)) {
ret = fchmodat(AT_FDCWD, args[1], mode, AT_SYMLINK_NOFOLLOW);
if (ret == -1) {
return -errno;
}
}
}
return e4crypt_set_directory_policy(args[1]);
}
其实就是调用了make_dir并做了一些权限等方面的操作。所以，跟在action后面的命令并不能随随便便乱加，而是要确保这个命令被定义了，不然就会出错。
init.rc的解析过程（以import为例）
因为init.rc的第一行代码就是Import语句。万事开头难，只要我们理清了第一行的解析过程，后面行的解析分析起来就不怎么费劲了。所以下面我们主要看看init.rc中第一行的解析过程。
init.tc的解析函数为：init_parse_config_file
int init_parse_config_file(const char* path) {
INFO("Parsing %s...\n", path);
Timer t;
std::string data;
if (!read_file(path, &data)) {
return -1;
}
data.push_back('\n'); // TODO: fix parse_config.
parse_config(path, data);
dump_parser_state();
// MStar Android Patch Begin
INFO("(Parsing %s took %.2fs.)\n", path, t.duration());
// MStar Android Patch End
return 0;
}
这个函数把/init.rc中的内容读出来，并让data这个string类型的变量指向它。
把读出来的data传递给parse_config函数做真正的解析工作。parse_config函数如下：
static void parse_config(const char *fn, const std::string& data)
{
char *args[UEVENTD_PARSER_MAXARGS];
int nargs = 0;
parse_state state;
state.filename = fn;
state.line = 1;
state.ptr = strdup(data.c_str()); // TODO: fix this code!
state.nexttoken = 0;
state.parse_line = parse_line_no_op;
for (;;) {
int token = next_token(&state);
switch (token) {
case T_EOF:
parse_line(&state, args, nargs);
return;
case T_NEWLINE:
if (nargs) {
parse_line(&state, args, nargs);
nargs = 0;
}
state.line++;
break;
case T_TEXT:
if (nargs < UEVENTD_PARSER_MAXARGS) {
args[nargs++] = state.text;
}
break;
}
}
}
我看到这个函数的时候，我想起了xml解析方法之一的pull解析，感觉挺像的。每次循环都会找到一个token，token就是一个特定的符号，然后根据这个toke做不同的处理。这里使用到了parse_state结构，启动以如下：
struct parse_state
{
char *ptr;
char *text;
int line;
int nexttoken;
void *context;
void (*parse_line)(struct parse_state *state, int nargs, char **args);
const char *filename;
void *priv;
};
这个就够中：ptr执行init.rc字符流的，text后面会用到，用来保存参数，line当然就是行数了，nexttoken保存下一个token,filename保存init.rc的文件描述符，filename当然是/init.rc了.parse_line是一个函数指针。context暂时没明白…state.priv 指向Import的一个文件链表。
我们打开Init.rc看看，从头分析它的解析过程。
import /init.environ.rc
import /init.usb.rc
import /init.${ro.hardware}.rc
import /init.${ro.zygote}.rc
import /init.trace.rc
...
init.rc前面几行都是import语句，我们看看一开始的解析流程。
这个时候，parse_satate的状态为：
state.filename = fn;
state.line = 1;
state.ptr = strdup(data.c_str()); // TODO: fix this code!
state.nexttoken = 0;
state.parse_line = parse_line_no_op;
list_init(&import_list);
state.priv = &import_list;
step 1.第一次循环
然后进入死循环，第一次调用next_token函数：
int next_token(struct parse_state *state)
{
char *x = state->ptr;
char *s;
if (state->nexttoken) {
int t = state->nexttoken;
state->nexttoken = 0;
return t;
}
for (;;) {
switch (*x) {
case 0:
state->ptr = x;
return T_EOF;
case '\n':
x++;
state->ptr = x;
return T_NEWLINE;
case ' ':
case '\t':
case '\r':
x++;
continue;
case '#':
while (*x && (*x != '\n')) x++;
if (*x == '\n') {
state->ptr = x+1;
return T_NEWLINE;
} else {
state->ptr = x;
return T_EOF;
}
default:
goto text;
}
}
textdone:
state->ptr = x;
*s = 0;
return T_TEXT;
text:
state->text = s = x;
textresume:
for (;;) {
switch (*x) {
case 0:
goto textdone;
case ' ':
case '\t':
case '\r':
x++;
goto textdone;
case '\n':
state->nexttoken = T_NEWLINE;
x++;
goto textdone;
case '"':
x++;
for (;;) {
switch (*x) {
case 0:
/* unterminated quoted thing */
state->ptr = x;
return T_EOF;
case '"':
x++;
goto textresume;
default:
*s++ = *x++;
}
}
break;
case '\\':
x++;
switch (*x) {
case 0:
goto textdone;
case 'n':
*s++ = '\n';
break;
case 'r':
*s++ = '\r';
break;
case 't':
*s++ = '\t';
break;
case '\\':
*s++ = '\\';
break;
case '\r':
/* \ <cr> <lf> -> line continuation */
if (x[1] != '\n') {
x++;
continue;
}
case '\n':
/* \ <lf> -> line continuation */
state->line++;
x++;
/* eat any extra whitespace */
while((*x == ' ') || (*x == '\t')) x++;
continue;
default:
/* unknown escape -- just copy */
*s++ = *x++;
}
continue;
default:
*s++ = *x++;
}
}
return T_EOF;
}
这时候，init.rc中的第一个符号应该是i(impor,省去空格)，所以next_token直接进入到text:标签执行，执行的结果是state->text = s = ‘i’;然后继续执行textresume:标签后面的内容：
标签后面的for死循环中，发现第一个字符是i,于是执行default分支：*s++ = *x++;这样直到import的t被检测完以后，在下一次循环变得到一个空格，于是执行x++,并goto textdone.。textdown执行完后函数返回，返回后，state->ptr 指向’/’符号 。*s = 0;意味着state.text就是字符串“import”,因为0就是字符串结束符了。注意返回值为T_TEXT。这个时候执行parse_config函数中的case T_TEXT:分支。
case T_TEXT:
if (nargs < INIT_PARSER_MAXARGS) {
args[nargs++] = state.text;
}
这个时候，nargs为0，state.text位import,于是args数组的第0项就存了”import”字符串了。然后nargs++,也就是等于1了。然后进入下次循环。
step 2.第二次循环
第二次循环再次调用next_token函数，这次state->ptr=’\’，这我们分析过了。因此next_token函数不断执行defaulty分支，最终state.text = “/init.environ.rc”，返回类型还是T_TEXT。于是和之前一样，args[1]=”/init.environ.rc”,nargs=2。
step 3.第三次循环
这个时候一行结束，parse_config函数进入case T_NEWLINE:分支。
这个分支中，首先执行lookup_keyword函数，从名字来看是查找关键字。肯定就是import了，它肯定是关键字。不信请看代码：
static int lookup_keyword(const char *s)
{
switch (*s++) {
case 'b':
if (!strcmp(s, "ootchart_init")) return K_bootchart_init;
break;
case 'c':
if (!strcmp(s, "opy")) return K_copy;
if (!strcmp(s, "lass")) return K_class;
if (!strcmp(s, "lass_start")) return K_class_start;
if (!strcmp(s, "lass_stop")) return K_class_stop;
if (!strcmp(s, "lass_reset")) return K_class_reset;
if (!strcmp(s, "onsole")) return K_console;
if (!strcmp(s, "hown")) return K_chown;
if (!strcmp(s, "hmod")) return K_chmod;
if (!strcmp(s, "ritical")) return K_critical;
break;
case 'd':
if (!strcmp(s, "isabled")) return K_disabled;
if (!strcmp(s, "omainname")) return K_domainname;
break;
case 'e':
if (!strcmp(s, "nable")) return K_enable;
if (!strcmp(s, "xec")) return K_exec;
if (!strcmp(s, "xport")) return K_export;
break;
case 'g':
if (!strcmp(s, "roup")) return K_group;
break;
case 'h':
if (!strcmp(s, "ostname")) return K_hostname;
break;
case 'i':
if (!strcmp(s, "oprio")) return K_ioprio;
if (!strcmp(s, "fup")) return K_ifup;
if (!strcmp(s, "nsmod")) return K_insmod;
if (!strcmp(s, "mport")) return K_import;
if (!strcmp(s, "nstallkey")) return K_installkey;
break;
case 'k':
if (!strcmp(s, "eycodes")) return K_keycodes;
break;
case 'l':
if (!strcmp(s, "oglevel")) return K_loglevel;
if (!strcmp(s, "oad_persist_props")) return K_load_persist_props;
if (!strcmp(s, "oad_all_props")) return K_load_all_props;
break;
case 'm':
if (!strcmp(s, "kdir")) return K_mkdir;
if (!strcmp(s, "ount_all")) return K_mount_all;
if (!strcmp(s, "ount")) return K_mount;
break;
case 'o':
if (!strcmp(s, "n")) return K_on;
if (!strcmp(s, "neshot")) return K_oneshot;
if (!strcmp(s, "nrestart")) return K_onrestart;
break;
case 'p':
if (!strcmp(s, "owerctl")) return K_powerctl;
break;
case 'r':
if (!strcmp(s, "estart")) return K_restart;
if (!strcmp(s, "estorecon")) return K_restorecon;
if (!strcmp(s, "estorecon_recursive")) return K_restorecon_recursive;
if (!strcmp(s, "mdir")) return K_rmdir;
if (!strcmp(s, "m")) return K_rm;
break;
case 's':
if (!strcmp(s, "eclabel")) return K_seclabel;
if (!strcmp(s, "ervice")) return K_service;
if (!strcmp(s, "etenv")) return K_setenv;
if (!strcmp(s, "etprop")) return K_setprop;
if (!strcmp(s, "etrlimit")) return K_setrlimit;
if (!strcmp(s, "ocket")) return K_socket;
if (!strcmp(s, "tart")) return K_start;
if (!strcmp(s, "top")) return K_stop;
if (!strcmp(s, "wapon_all")) return K_swapon_all;
if (!strcmp(s, "ymlink")) return K_symlink;
if (!strcmp(s, "ysclktz")) return K_sysclktz;
break;
case 't':
if (!strcmp(s, "rigger")) return K_trigger;
break;
case 'u':
if (!strcmp(s, "ser")) return K_user;
break;
case 'v':
if (!strcmp(s, "erity_load_state")) return K_verity_load_state;
if (!strcmp(s, "erity_update_state")) return K_verity_update_state;
break;
case 'w':
if (!strcmp(s, "rite")) return K_write;
if (!strcmp(s, "ritepid")) return K_writepid;
if (!strcmp(s, "ait")) return K_wait;
break;
}
return K_UNKNOWN;
}
调用这个函数的时候，我们传入的参数args[0]=”import”.显而易见该函数返回K_import。它是一个整数。返回以后使用kw_is函数看他是不是一个Section。当然是一个section了，import也是一个section。不信看代码：
#define kw_is(kw, type) (keyword_info[kw].flags & (type))
1
keyword_info定义在system/core/init/keywords.h中：
...
KEYWORD(group, OPTION, 0, 0)
KEYWORD(hostname, COMMAND, 1, do_hostname)
KEYWORD(ifup, COMMAND, 1, do_ifup)
KEYWORD(import, SECTION, 1, 0)
...
截取含有import的一部分代码，后面SECTION已经表明它是个Section了。KEYWORD自后一个参数是这个关键字对应的处理函数。比如这其中的hostname。如果你在init.rc中使用hostname 关键字，那么最终会调用do_hostname函数来处理。
既然import是一个section。那么parce_config就会调用state.parse_line函数，这里是一个函数指针，其实调用的是parse_line_no_op，不记得回去看下state的初始就知道了。
static void parse_line_no_op(struct parse_state*, int, char**) {
}
这个函数是空的。接下来调用parse_new_section函数：
static void parse_new_section(struct parse_state *state, int kw,
int nargs, char **args)
{
printf("[ %s %s ]\n", args[0],
nargs > 1 ? args[1] : "");
switch(kw) {
case K_service:
state->context = parse_service(state, nargs, args);
if (state->context) {
state->parse_line = parse_line_service;
return;
}
break;
case K_on:
state->context = parse_action(state, nargs, args);
if (state->context) {
state->parse_line = parse_line_action;
return;
}
break;
case K_import:
parse_import(state, nargs, args);
break;
}
state->parse_line = parse_line_no_op;
}
我们当然是执行case K_import:分支了，想都不用想。所以接下来执行parse_import方法：
static void parse_import(struct parse_state *state, int nargs, char **args)
{
struct listnode *import_list = (listnode*) state->priv;
char conf_file[PATH_MAX];
int ret;
if (nargs != 2) {
ERROR("single argument needed for import\n");
return;
}
ret = expand_props(conf_file, args[1], sizeof(conf_file));
if (ret) {
ERROR("error while handling import on line '%d' in '%s'\n",
state->line, state->filename);
return;
}
struct import* import = (struct import*) calloc(1, sizeof(struct import));
import->filename = strdup(conf_file);
list_add_tail(import_list, &import->list);
INFO("Added '%s' to import list\n", import->filename);
}
这个函数首先使用expand_props方法对args[1]也就是“/init.environ.rc”做进一步处理。这个函数如下：
int expand_props(char *dst, const char *src, int dst_size)
{
char *dst_ptr = dst;
const char *src_ptr = src;
int ret = 0;
int left = dst_size - 1;
if (!src || !dst || dst_size == 0)
return -1;
/* - variables can either be $x.y or ${x.y}, in case they are only part
* of the string.
* - will accept $$ as a literal $.
* - no nested property expansion, i.e. ${foo.${bar}} is not supported,
* bad things will happen
*/
while (*src_ptr && left > 0) {
char *c;
char prop[PROP_NAME_MAX + 1];
char prop_val[PROP_VALUE_MAX];
int prop_len = 0;
int prop_val_len;
c = strchr(src_ptr, '$');
if (!c) {
while (left-- > 0 && *src_ptr)
*(dst_ptr++) = *(src_ptr++);
break;
}
...
可以看出，这个函数的作用是拓展args[1].这里不需要拓展，因为我们的args[1]=”/init.environ.rc”没有$符号，所以直接就跳出循环了。这里应该是对那些有包含变量的字符串，把变量的内容展开。
然后构建了一个import结构体。import中的filename项赋值为”/init.environ.rc”.并把它加入到import_list链表中。
import定义如下：
struct import {
struct listnode list;
const char *filename;
};
这样，第一行就分析完了，从而我们也彻底明白了怎么解析一个import 的section。
只要能看懂一个，其他的就简单了。因为，他们都是类似的。
service的解析与启动
service的解析
和import解析过程类似，遇到service关键字后，service关键字和后面的参数会保存在args[]数组中。然后通过对args[0]提取关键字，发现args[0]=”service”,于是开始执行parse_new_section函数。此时这个函数必然会进入 case K_service:分支执行：
case K_service:
state->context = parse_service(state, nargs, args);
if (state->context) {
state->parse_line = parse_line_service;
return;
}
break;
这里做了两件事情非常重要，一件是调用parse_service解析service这个section。另一件事情是给state->parse_line赋值为parse_line_service。也就是service 关键字所在的行后面的那些options行都是使用parse_line_service函数来解析的。我们从parse_service看起：
static void *parse_service(struct parse_state *state, int nargs, char **args)
{
if (nargs < 3) {
parse_error(state, "services must have a name and a program\n");
return 0;
}
if (!valid_name(args[1])) {
parse_error(state, "invalid service name '%s'\n", args[1]);
return 0;
}
service* svc = (service*) service_find_by_name(args[1]);
if (svc) {
parse_error(state, "ignored duplicate definition of service '%s'\n", args[1]);
return 0;
}
nargs -= 2;
svc = (service*) calloc(1, sizeof(*svc) + sizeof(char*) * nargs);
if (!svc) {
parse_error(state, "out of memory\n");
return 0;
}
svc->name = strdup(args[1]);
svc->classname = "default";
memcpy(svc->args, args + 2, sizeof(char*) * nargs);
trigger* cur_trigger = (trigger*) calloc(1, sizeof(*cur_trigger));
svc->args[nargs] = 0;
svc->nargs = nargs;
list_init(&svc->onrestart.triggers);
cur_trigger->name = "onrestart";
list_add_tail(&svc->onrestart.triggers, &cur_trigger->nlist);
list_init(&svc->onrestart.commands);
list_add_tail(&service_list, &svc->slist);
return svc;
}
可以看到和import做的事情差不多。import解析的最后，会创建一个import结构体，并把它添加到import_list双向链表中。service解析从这里看，也是构建一个service结构体，然后把service结构体添加到service_list链表中。
我们看下service结构体：
struct service {
void NotifyStateChange(const char* new_state);
/* list of all services */
struct listnode slist;
char *name;
const char *classname;
unsigned flags;
pid_t pid;
time_t time_started; /* time of last start */
time_t time_crashed; /* first crash within inspection window */
int nr_crashed; /* number of times crashed within window */
uid_t uid;
gid_t gid;
gid_t supp_gids[NR_SVC_SUPP_GIDS];
size_t nr_supp_gids;
const char* seclabel;
struct socketinfo *sockets;
struct svcenvinfo *envvars;
struct action onrestart; /* Actions to execute on restart. */
std::vector<std::string>* writepid_files_;
/* keycodes for triggering this service via /dev/keychord */
int *keycodes;
int nkeycodes;
int keychord_id;
IoSchedClass ioprio_class;
int ioprio_pri;
int nargs;
/* "MUST BE AT THE END OF THE STRUCT" */
char *args[1];
}; /*
socketinfo 用来保存socket option的相关信息。
classname 给service定义一个类名，如果多个service使用相同的类型，可以方便进行批量操作。
nargs 保存参数的个数。
很多字段不理解，没关系，我们看看parse_service函数给service做了那些初始化：
1. svc->name = strdup(args[1]);名字就是service 关键字后面的第一个参数
2. svc->classname = “default”; 类别名是default
3. memcpy(svc->args, args + 2, sizeof(char*) * nargs); svc->args[nargs] = 0;
把所有参数保存在args数组中，并把最有一个成员只为0。
4. svc->nargs = nargs; nargs保存参数个数
5. trigger* cur_trigger = (trigger*) calloc(1, sizeof(*cur_trigger));
list_init(&svc->onrestart.triggers);
cur_trigger->name = “onrestart”;
list_add_tail(&svc->onrestart.triggers, &cur_trigger->nlist);
构建一个触发器，并把它添加到service中的onrestart.triger列表中。
因此，我们可以知道么一个service都会有一个onrestart的action,这个action有一个触发器。这个action用来重启service。
解析完service后，就会解析service后面的option了，这个时候会调用parse_line_service。
static void parse_line_service(struct parse_state *state, int nargs, char **args)
{
struct service *svc = (service*) state->context;
struct command *cmd;
int i, kw, kw_nargs;
if (nargs == 0) {
return;
}
svc->ioprio_class = IoSchedClass_NONE;
kw = lookup_keyword(args[0]);
switch (kw) {
case K_class:
if (nargs != 2) {
parse_error(state, "class option requires a classname\n");
} else {
svc->classname = args[1];
}
break;
case K_console:
svc->flags |= SVC_CONSOLE;
break;
case K_disabled:
svc->flags |= SVC_DISABLED;
svc->flags |= SVC_RC_DISABLED;
break;
case K_ioprio:
if (nargs != 3) {
parse_error(state, "ioprio optin usage: ioprio <rt|be|idle> <ioprio 0-7>\n");
} else {
svc->ioprio_pri = strtoul(args[2], 0, 8);
if (svc->ioprio_pri < 0 || svc->ioprio_pri > 7) {
parse_error(state, "priority value must be range 0 - 7\n");
break;
}
if (!strcmp(args[1], "rt")) {
svc->ioprio_class = IoSchedClass_RT;
} else if (!strcmp(args[1], "be")) {
svc->ioprio_class = IoSchedClass_BE;
} else if (!strcmp(args[1], "idle")) {
svc->ioprio_class = IoSchedClass_IDLE;
} else {
parse_error(state, "ioprio option usage: ioprio <rt|be|idle> <0-7>\n");
}
}
break;
case K_group:
if (nargs < 2) {
parse_error(state, "group option requires a group id\n");
} else if (nargs > NR_SVC_SUPP_GIDS + 2) {
parse_error(state, "group option accepts at most %d supp. groups\n",
NR_SVC_SUPP_GIDS);
} else {
int n;
svc->gid = decode_uid(args[1]);
for (n = 2; n < nargs; n++) {
svc->supp_gids[n-2] = decode_uid(args[n]);
}
svc->nr_supp_gids = n - 2;
}
break;
case K_keycodes:
if (nargs < 2) {
parse_error(state, "keycodes option requires atleast one keycode\n");
} else {
svc->keycodes = (int*) malloc((nargs - 1) * sizeof(svc->keycodes[0]));
if (!svc->keycodes) {
parse_error(state, "could not allocate keycodes\n");
} else {
svc->nkeycodes = nargs - 1;
for (i = 1; i < nargs; i++) {
svc->keycodes[i - 1] = atoi(args[i]);
}
}
}
break;
case K_oneshot:
svc->flags |= SVC_ONESHOT;
break;
case K_onrestart:
nargs--;
args++;
kw = lookup_keyword(args[0]);
if (!kw_is(kw, COMMAND)) {
parse_error(state, "invalid command '%s'\n", args[0]);
break;
}
kw_nargs = kw_nargs(kw);
if (nargs < kw_nargs) {
parse_error(state, "%s requires %d %s\n", args[0], kw_nargs - 1,
kw_nargs > 2 ? "arguments" : "argument");
break;
}
cmd = (command*) malloc(sizeof(*cmd) + sizeof(char*) * nargs);
cmd->func = kw_func(kw);
cmd->nargs = nargs;
memcpy(cmd->args, args, sizeof(char*) * nargs);
list_add_tail(&svc->onrestart.commands, &cmd->clist);
break;
case K_critical:
svc->flags |= SVC_CRITICAL;
break;
case K_setenv: { /* name value */
if (nargs < 3) {
parse_error(state, "setenv option requires name and value arguments\n");
break;
}
svcenvinfo* ei = (svcenvinfo*) calloc(1, sizeof(*ei));
if (!ei) {
parse_error(state, "out of memory\n");
break;
}
ei->name = args[1];
ei->value = args[2];
ei->next = svc->envvars;
svc->envvars = ei;
break;
}
case K_socket: {/* name type perm [ uid gid context ] */
if (nargs < 4) {
parse_error(state, "socket option requires name, type, perm arguments\n");
break;
}
if (strcmp(args[2],"dgram") && strcmp(args[2],"stream")
&& strcmp(args[2],"seqpacket")) {
parse_error(state, "socket type must be 'dgram', 'stream' or 'seqpacket'\n");
break;
}
socketinfo* si = (socketinfo*) calloc(1, sizeof(*si));
if (!si) {
parse_error(state, "out of memory\n");
break;
}
si->name = args[1];
si->type = args[2];
si->perm = strtoul(args[3], 0, 8);
if (nargs > 4)
si->uid = decode_uid(args[4]);
if (nargs > 5)
si->gid = decode_uid(args[5]);
if (nargs > 6)
si->socketcon = args[6];
si->next = svc->sockets;
svc->sockets = si;
break;
}
case K_user:
if (nargs != 2) {
parse_error(state, "user option requires a user id\n");
} else {
svc->uid = decode_uid(args[1]);
}
break;
case K_seclabel:
if (nargs != 2) {
parse_error(state, "seclabel option requires a label string\n");
} else {
svc->seclabel = args[1];
}
break;
case K_writepid:
if (nargs < 2) {
parse_error(state, "writepid option requires at least one filename\n");
break;
}
svc->writepid_files_ = new std::vector<std::string>;
for (int i = 1; i < nargs; ++i) {
svc->writepid_files_->push_back(args[i]);
}
break;
default:
parse_error(state, "invalid option '%s'\n", args[0]);
}
}
这个函数中定义了所有的option，每一个option处理方法都不相同，大家遇到感兴趣的option可以自行分析。
service解析完成后，是时候看看service的启动了。
service的启动
service解析完成以后，有了一个service_list的链表，可以service是在是么地方启动的呢？当然是在action中了，action中有个命名叫start,它对应的处理函数是do_start,着我们在前面已经说过了，do_start函数如下：
int do_start(int nargs, char **args)
{
struct service *svc;
svc = service_find_by_name(args[1]);
if (svc) {
service_start(svc, NULL);
}
return 0;
}
这个函数非常见到，找到服务，启动它。
查找service的过程：
struct service *service_find_by_name(const char *name)
{
struct listnode *node;
struct service *svc;
list_for_each(node, &service_list) {
svc = node_to_item(node, struct service, slist);
if (!strcmp(svc->name, name)) {
return svc;
}
}
return 0;
}
遍历service_list，对比名字，相同就返回。
启动service过程：
void service_start(struct service *svc, const char *dynamic_args)
{
// Starting a service removes it from the disabled or reset state and
// immediately takes it out of the restarting state if it was in there.
svc->flags &= (~(SVC_DISABLED|SVC_RESTARTING|SVC_RESET|SVC_RESTART|SVC_DISABLED_START));
svc->time_started = 0;
// Running processes require no additional work --- if they're in the
// process of exiting, we've ensured that they will immediately restart
// on exit, unless they are ONESHOT.
if (svc->flags & SVC_RUNNING) {
return;
}
bool needs_console = (svc->flags & SVC_CONSOLE);
if (needs_console && !have_console) {
ERROR("service '%s' requires console\n", svc->name);
svc->flags |= SVC_DISABLED;
return;
}
struct stat s;
if (stat(svc->args[0], &s) != 0) {
ERROR("cannot find '%s', disabling '%s'\n", svc->args[0], svc->name);
svc->flags |= SVC_DISABLED;
return;
}
if ((!(svc->flags & SVC_ONESHOT)) && dynamic_args) {
ERROR("service '%s' must be one-shot to use dynamic args, disabling\n",
svc->args[0]);
svc->flags |= SVC_DISABLED;
return;
}
char* scon = NULL;
if (is_selinux_enabled() > 0) {
if (svc->seclabel) {
scon = strdup(svc->seclabel);
if (!scon) {
ERROR("Out of memory while starting '%s'\n", svc->name);
return;
}
} else {
char *mycon = NULL, *fcon = NULL;
INFO("computing context for service '%s'\n", svc->args[0]);
int rc = getcon(&mycon);
if (rc < 0) {
ERROR("could not get context while starting '%s'\n", svc->name);
return;
}
rc = getfilecon(svc->args[0], &fcon);
if (rc < 0) {
ERROR("could not get context while starting '%s'\n", svc->name);
freecon(mycon);
return;
}
rc = security_compute_create(mycon, fcon, string_to_security_class("process"), &scon);
if (rc == 0 && !strcmp(scon, mycon)) {
ERROR("Warning! Service %s needs a SELinux domain defined; please fix!\n", svc->name);
}
freecon(mycon);
freecon(fcon);
if (rc < 0) {
ERROR("could not get context while starting '%s'\n", svc->name);
return;
}
}
}
NOTICE("Starting service '%s'...\n", svc->name);
pid_t pid = fork();
if (pid == 0) {
struct socketinfo *si;
struct svcenvinfo *ei;
char tmp[32];
int fd, sz;
umask(077);
if (properties_initialized()) {
get_property_workspace(&fd, &sz);
snprintf(tmp, sizeof(tmp), "%d,%d", dup(fd), sz);
add_environment("ANDROID_PROPERTY_WORKSPACE", tmp);
}
for (ei = svc->envvars; ei; ei = ei->next)
add_environment(ei->name, ei->value);
for (si = svc->sockets; si; si = si->next) {
int socket_type = (
!strcmp(si->type, "stream") ? SOCK_STREAM :
(!strcmp(si->type, "dgram") ? SOCK_DGRAM : SOCK_SEQPACKET));
int s = create_socket(si->name, socket_type,
si->perm, si->uid, si->gid, si->socketcon ?: scon);
if (s >= 0) {
publish_socket(si->name, s);
}
}
freecon(scon);
scon = NULL;
if (svc->writepid_files_) {
std::string pid_str = android::base::StringPrintf("%d", pid);
for (auto& file : *svc->writepid_files_) {
if (!android::base::WriteStringToFile(pid_str, file)) {
ERROR("couldn't write %s to %s: %s\n",
pid_str.c_str(), file.c_str(), strerror(errno));
}
}
}
if (svc->ioprio_class != IoSchedClass_NONE) {
if (android_set_ioprio(getpid(), svc->ioprio_class, svc->ioprio_pri)) {
ERROR("Failed to set pid %d ioprio = %d,%d: %s\n",
getpid(), svc->ioprio_class, svc->ioprio_pri, strerror(errno));
}
}
if (needs_console) {
setsid();
open_console();
} else {
zap_stdio();
}
if (false) {
for (size_t n = 0; svc->args[n]; n++) {
INFO("args[%zu] = '%s'\n", n, svc->args[n]);
}
for (size_t n = 0; ENV[n]; n++) {
INFO("env[%zu] = '%s'\n", n, ENV[n]);
}
}
setpgid(0, getpid());
// As requested, set our gid, supplemental gids, and uid.
if (svc->gid) {
if (setgid(svc->gid) != 0) {
ERROR("setgid failed: %s\n", strerror(errno));
_exit(127);
}
}
if (svc->nr_supp_gids) {
if (setgroups(svc->nr_supp_gids, svc->supp_gids) != 0) {
ERROR("setgroups failed: %s\n", strerror(errno));
_exit(127);
}
}
if (svc->uid) {
if (setuid(svc->uid) != 0) {
ERROR("setuid failed: %s\n", strerror(errno));
_exit(127);
}
}
if (svc->seclabel) {
if (is_selinux_enabled() > 0 && setexeccon(svc->seclabel) < 0) {
ERROR("cannot setexeccon('%s'): %s\n", svc->seclabel, strerror(errno));
_exit(127);
}
}
if (!dynamic_args) {
if (execve(svc->args[0], (char**) svc->args, (char**) ENV) < 0) {
ERROR("cannot execve('%s'): %s\n", svc->args[0], strerror(errno));
}
} else {
char *arg_ptrs[INIT_PARSER_MAXARGS+1];
int arg_idx = svc->nargs;
char *tmp = strdup(dynamic_args);
char *next = tmp;
char *bword;
/* Copy the static arguments */
memcpy(arg_ptrs, svc->args, (svc->nargs * sizeof(char *)));
while((bword = strsep(&next, " "))) {
arg_ptrs[arg_idx++] = bword;
if (arg_idx == INIT_PARSER_MAXARGS)
break;
}
arg_ptrs[arg_idx] = NULL;
execve(svc->args[0], (char**) arg_ptrs, (char**) ENV);
}
_exit(127);
}
freecon(scon);
if (pid < 0) {
ERROR("failed to start '%s'\n", svc->name);
svc->pid = 0;
return;
}
svc->time_started = gettime();
svc->pid = pid;
svc->flags |= SVC_RUNNING;
if ((svc->flags & SVC_EXEC) != 0) {
INFO("SVC_EXEC pid %d (uid %d gid %d+%zu context %s) started; waiting...\n",
svc->pid, svc->uid, svc->gid, svc->nr_supp_gids,
svc->seclabel ? : "default");
waiting_for_exec = true;
}
svc->NotifyStateChange("running");
}
这个函数虽然长，但总结起来无非就做了这些事情：
1.解析参数
2.fork一个进程。
3.初始化子进程，主要是根据service结构体中的信息创建一个写东西。比如，根据sockets创建socket等。
4.执行execve，也就是加载可执行文件了。
也就是说，当init.rc等rc配置脚本解析完成后，开始执行action中的命令，并通过start命令来启动service。
action的解析与命令的执行
action的解析
对比service的解析来看，action的解析应该是调用parse_action函数：
static void *parse_action(struct parse_state *state, int nargs, char **args)
{
struct trigger *cur_trigger;
int i;
if (nargs < 2) {
parse_error(state, "actions must have a trigger\n");
return 0;
}
action* act = (action*) calloc(1, sizeof(*act));
list_init(&act->triggers);
for (i = 1; i < nargs; i++) {
if (!(i % 2)) {
if (strcmp(args[i], "&&")) {
struct listnode *node;
struct listnode *node2;
parse_error(state, "& is the only symbol allowed to concatenate actions\n");
list_for_each_safe(node, node2, &act->triggers) {
struct trigger *trigger = node_to_item(node, struct trigger, nlist);
free(trigger);
}
free(act);
return 0;
} else
continue;
}
cur_trigger = (trigger*) calloc(1, sizeof(*cur_trigger));
cur_trigger->name = args[i];
list_add_tail(&act->triggers, &cur_trigger->nlist);
}
list_init(&act->commands);
list_init(&act->qlist);
list_add_tail(&action_list, &act->alist);
/* XXX add to hash */
return act;
}
action中命令的执行
可以看到所有的section解析都是类似的，构建一个结构体并添加到对应的链表中。
这里就不继续展开分析了。我们回到init.cpp的main函数中，看看action_list中存放的action是如何被执行的。
在解析init.rc结束后，会有如下函数：
action_for_each_trigger("early-init", action_add_queue_tail);
action_for_each_trigger("init", action_add_queue_tail);
action_for_each_trigger("charger", action_add_queue_tail);
action_for_each_trigger("late-init", action_add_queue_tail);
action_for_each_trigger函数如下：
void action_for_each_trigger(const char *trigger,
void (*func)(struct action *act))
{
struct listnode *node, *node2;
struct action *act;
struct trigger *cur_trigger;
list_for_each(node, &action_list) {
act = node_to_item(node, struct action, alist);
list_for_each(node2, &act->triggers) {
cur_trigger = node_to_item(node2, struct trigger, nlist);
if (!strcmp(cur_trigger->name, trigger)) {
func(act);
}
}
}
}
也就是说，这个函数的作用是遍历action_list链表，找到对应名字的触发器，然后盗用传入的func函数，也就是action_add_queue_tail函数，这个函数如下：
void action_add_queue_tail(struct action *act)
{
if (list_empty(&act->qlist)) {
list_add_tail(&action_queue, &act->qlist);
}
}
再把action添加到action_queuw中。这里是不是可以理解为给action排队呢？按照action的名字（early-init,init…）把action排好顺序。这个时候还是没有执行action中的命令。
继续往下看，看到execute_one_command函数，从名字来看是执行一个命令，我们看看是怎么执行的，execute_one_command函数如下：
void execute_one_command() {
Timer t;
char cmd_str[256] = "";
char name_str[256] = "";
if (!cur_action || !cur_command || is_last_command(cur_action, cur_command)) {
cur_action = action_remove_queue_head();
cur_command = NULL;
if (!cur_action) {
return;
}
build_triggers_string(name_str, sizeof(name_str), cur_action);
INFO("processing action %p (%s)\n", cur_action, name_str);
cur_command = get_first_command(cur_action);
} else {
cur_command = get_next_command(cur_action, cur_command);
}
if (!cur_command) {
return;
}
int result = cur_command->func(cur_command->nargs, cur_command->args);
if (klog_get_level() >= KLOG_INFO_LEVEL) {
for (int i = 0; i < cur_command->nargs; i++) {
strlcat(cmd_str, cur_command->args[i], sizeof(cmd_str));
if (i < cur_command->nargs - 1) {
strlcat(cmd_str, " ", sizeof(cmd_str));
}
}
char source[256];
if (cur_command->filename) {
snprintf(source, sizeof(source), " (%s:%d)", cur_command->filename, cur_command->line);
} else {
*source = '\0';
}
INFO("Command '%s' action=%s%s returned %d took %.2fs\n",
cmd_str, cur_action ? name_str : "", source, result, t.duration());
}
}
这个函数就是从aciton_queue中取出头部的action，然后执行command中的函数。那这个函数是什么呢？我们在文章一开始就说过了，每一个command都对应一个do_xxxx的函数来处理该命令。是不是这样呢？
我们需要从command的解析说起。和解析service的option一样。解析command使用的是parse_line_action函数。这里不明白的可以返回去看看service的解析过程。
parse_line_action定义在system/core/init/init_parser.cpp中，该函数如下：
static void parse_line_action(struct parse_state* state, int nargs, char **args)
{
struct action *act = (action*) state->context;
int kw, n;
if (nargs == 0) {
return;
}
kw = lookup_keyword(args[0]);
if (!kw_is(kw, COMMAND)) {
parse_error(state, "invalid command '%s'\n", args[0]);
return;
}
n = kw_nargs(kw);
if (nargs < n) {
parse_error(state, "%s requires %d %s\n", args[0], n - 1,
n > 2 ? "arguments" : "argument");
return;
}
command* cmd = (command*) malloc(sizeof(*cmd) + sizeof(char*) * nargs);
cmd->func = kw_func(kw);
cmd->line = state->line;
cmd->filename = state->filename;
cmd->nargs = nargs;
memcpy(cmd->args, args, sizeof(char*) * nargs);
list_add_tail(&act->commands, &cmd->clist);
}
关键的代码只有一行cmd->func = kw_func(kw)，kw_func函数定义在相同文件下，是一个宏：
#define kw_func(kw) (keyword_info[kw].func)
1
这里keyword_info数组又出现了吧，这个数组也定义在system/core/init/init_parser.cpp中：
static struct {
const char *name;
int (*func)(int nargs, char **args);
unsigned char nargs;
unsigned char flags;
} keyword_info[KEYWORD_COUNT] = {
[ K_UNKNOWN ] = { "unknown", 0, 0, 0 },
#include "keywords.h"
};
可以看到数组的初始化使用的是#include “keywords.h”。这中写法我还是第一次见到。
keywords.h：
#ifndef KEYWORD
int do_bootchart_init(int nargs, char **args);
int do_class_start(int nargs, char **args);
int do_class_stop(int nargs, char **args);
int do_class_reset(int nargs, char **args);
int do_domainname(int nargs, char **args);
int do_enable(int nargs, char **args);
int do_exec(int nargs, char **args);
int do_export(int nargs, char **args);
int do_hostname(int nargs, char **args);
int do_ifup(int nargs, char **args);
int do_insmod(int nargs, char **args);
int do_installkey(int nargs, char **args);
int do_mkdir(int nargs, char **args);
int do_mount_all(int nargs, char **args);
int do_mount(int nargs, char **args);
int do_powerctl(int nargs, char **args);
int do_restart(int nargs, char **args);
int do_restorecon(int nargs, char **args);
int do_restorecon_recursive(int nargs, char **args);
int do_rm(int nargs, char **args);
int do_rmdir(int nargs, char **args);
int do_setprop(int nargs, char **args);
int do_setrlimit(int nargs, char **args);
int do_start(int nargs, char **args);
int do_stop(int nargs, char **args);
int do_swapon_all(int nargs, char **args);
int do_trigger(int nargs, char **args);
int do_symlink(int nargs, char **args);
int do_sysclktz(int nargs, char **args);
int do_write(int nargs, char **args);
int do_copy(int nargs, char **args);
int do_chown(int nargs, char **args);
int do_chmod(int nargs, char **args);
int do_loglevel(int nargs, char **args);
int do_load_persist_props(int nargs, char **args);
int do_load_all_props(int nargs, char **args);
int do_verity_load_state(int nargs, char **args);
int do_verity_update_state(int nargs, char **args);
int do_wait(int nargs, char **args);
#define __MAKE_KEYWORD_ENUM__
#define KEYWORD(symbol, flags, nargs, func) K_##symbol,
enum {
K_UNKNOWN,
#endif
KEYWORD(bootchart_init, COMMAND, 0, do_bootchart_init)
KEYWORD(chmod, COMMAND, 2, do_chmod)
KEYWORD(chown, COMMAND, 2, do_chown)
KEYWORD(class, OPTION, 0, 0)
KEYWORD(class_reset, COMMAND, 1, do_class_reset)
KEYWORD(class_start, COMMAND, 1, do_class_start)
KEYWORD(class_stop, COMMAND, 1, do_class_stop)
KEYWORD(console, OPTION, 0, 0)
KEYWORD(copy, COMMAND, 2, do_copy)
KEYWORD(critical, OPTION, 0, 0)
KEYWORD(disabled, OPTION, 0, 0)
KEYWORD(domainname, COMMAND, 1, do_domainname)
KEYWORD(enable, COMMAND, 1, do_enable)
KEYWORD(exec, COMMAND, 1, do_exec)
KEYWORD(export, COMMAND, 2, do_export)
KEYWORD(group, OPTION, 0, 0)
KEYWORD(hostname, COMMAND, 1, do_hostname)
KEYWORD(ifup, COMMAND, 1, do_ifup)
KEYWORD(import, SECTION, 1, 0)
KEYWORD(insmod, COMMAND, 1, do_insmod)
KEYWORD(installkey, COMMAND, 1, do_installkey)
KEYWORD(ioprio, OPTION, 0, 0)
KEYWORD(keycodes, OPTION, 0, 0)
KEYWORD(load_all_props, COMMAND, 0, do_load_all_props)
KEYWORD(load_persist_props, COMMAND, 0, do_load_persist_props)
KEYWORD(loglevel, COMMAND, 1, do_loglevel)
KEYWORD(mkdir, COMMAND, 1, do_mkdir)
KEYWORD(mount_all, COMMAND, 1, do_mount_all)
KEYWORD(mount, COMMAND, 3, do_mount)
KEYWORD(oneshot, OPTION, 0, 0)
KEYWORD(onrestart, OPTION, 0, 0)
KEYWORD(on, SECTION, 0, 0)
KEYWORD(powerctl, COMMAND, 1, do_powerctl)
KEYWORD(restart, COMMAND, 1, do_restart)
KEYWORD(restorecon, COMMAND, 1, do_restorecon)
KEYWORD(restorecon_recursive, COMMAND, 1, do_restorecon_recursive)
KEYWORD(rm, COMMAND, 1, do_rm)
KEYWORD(rmdir, COMMAND, 1, do_rmdir)
KEYWORD(seclabel, OPTION, 0, 0)
KEYWORD(service, SECTION, 0, 0)
KEYWORD(setenv, OPTION, 2, 0)
KEYWORD(setprop, COMMAND, 2, do_setprop)
KEYWORD(setrlimit, COMMAND, 3, do_setrlimit)
KEYWORD(socket, OPTION, 0, 0)
KEYWORD(start, COMMAND, 1, do_start)
KEYWORD(stop, COMMAND, 1, do_stop)
KEYWORD(swapon_all, COMMAND, 1, do_swapon_all)
KEYWORD(symlink, COMMAND, 1, do_symlink)
KEYWORD(sysclktz, COMMAND, 1, do_sysclktz)
KEYWORD(trigger, COMMAND, 1, do_trigger)
KEYWORD(user, OPTION, 0, 0)
KEYWORD(verity_load_state, COMMAND, 0, do_verity_load_state)
KEYWORD(verity_update_state, COMMAND, 0, do_verity_update_state)
KEYWORD(wait, COMMAND, 1, do_wait)
KEYWORD(write, COMMAND, 2, do_write)
KEYWORD(writepid, OPTION, 0, 0)
#ifdef __MAKE_KEYWORD_ENUM__
KEYWORD_COUNT,
};
#undef __MAKE_KEYWORD_ENUM__
#undef KEYWORD
#endif
这其中定义了所有的init.rc中需要的关键字。从中可以知道哪些是命令，哪些是option,哪些是sercion。此外，也验证我们说的命令（command）都对应这一个do_xxxx的函数，执行这些命令其实就是执行这些函数。
Android6.0系统启动流程分析二：zygote进程
上一篇ndroid6.0系统启动流程分析一：init进程博客我们分析了init进程，主要分析了init.rc的解析过程和command的执行流程。我们说Init.rc中配置的service都是在init.rc中的action中使用start命令启动的，start命令对应的处理函数是do_start。我们也分析了这个函数。那么这一节，我们分析zygote进程，zygote进程在init.rc中也被配置为一个服务，那么它是不是使用start命令来启动的呢？我在Init.rc中搜索发现并没有，难道我错了？
zygote进程的启动
看下zygote在rc文件中的定义：
service zygote /system/bin/app_process -Xzygote /system/bin --zygote --start-system-server
class main
socket zygote stream 660 root system
onrestart write /sys/android_power/request_state wake
onrestart write /sys/power/state on
onrestart restart media
onrestart restart netd
这个时候，大家不要忘了class main这行。我们在上片博客中不是说过class的作用吗？用于批量管理service。搜索main果然发现有如下语句：
on nonencrypted
class_start main
class_start late_start
on property:vold.decrypt=trigger_restart_min_framework
class_start main
on property:vold.decrypt=trigger_restart_framework
class_start main
class_start late_start
但是发现有三个地方出现 class_start main，所以本菜就迷糊了，不知道是哪里启动的zygote，希望路过的大神能指点下迷津…
class_start 对应的处理方法是do_class_start,该方法定义在system\core\init\builtins.cpp中：
int do_class_start(int nargs, char **args)
{
/* Starting a class does not start services
* which are explicitly disabled. They must
* be started individually.
*/
service_for_each_class(args[1], service_start_if_not_disabled);
return 0;
}
调用service_for_each_class方法进一步处理：
void service_for_each_class(const char *classname,
void (*func)(struct service *svc))
{
struct listnode *node;
struct service *svc;
list_for_each(node, &service_list) {
svc = node_to_item(node, struct service, slist);
if (!strcmp(svc->classname, classname)) {
func(svc);
}
}
}
遍历service_list链表，没找到一个classname为main的service,就调用service_start_if_not_disabled来进一步处理，service_start_if_not_disabled方法如下：
static void service_start_if_not_disabled(struct service *svc)
{
if (!(svc->flags & SVC_DISABLED)) {
service_start(svc, NULL);
} else {
svc->flags |= SVC_DISABLED_START;
}
}
可以看到最终还是调用了service_start方法来启动service,这个启动单个service一样了。这份函数我们在上节已经分析过了，这里就不再啰嗦了。
Zygote进程的native部分
zygote进程的可执行文件就是/system/bin/app_process，源码在\frameworks\base\cmds\app_process/app_main.cpp中。我们从它的main函数看起：
int main(int argc, char* const argv[])
{
if (prctl(PR_SET_NO_NEW_PRIVS, 1, 0, 0, 0) < 0) {
// Older kernels don't understand PR_SET_NO_NEW_PRIVS and return
// EINVAL. Don't die on such kernels.
if (errno != EINVAL) {
LOG_ALWAYS_FATAL("PR_SET_NO_NEW_PRIVS failed: %s", strerror(errno));
return 12;
}
}
AppRuntime runtime(argv[0], computeArgBlockSize(argc, argv));
// Process command line arguments
// ignore argv[0]
argc--;
argv++;
// Everything up to '--' or first non '-' arg goes to the vm.
//
// The first argument after the VM args is the "parent dir", which
// is currently unused.
//
// After the parent dir, we expect one or more the following internal
// arguments :
//
// --zygote : Start in zygote mode
// --start-system-server : Start the system server.
// --application : Start in application (stand alone, non zygote) mode.
// --nice-name : The nice name for this process.
//
// For non zygote starts, these arguments will be followed by
// the main class name. All remaining arguments are passed to
// the main method of this class.
//
// For zygote starts, all remaining arguments are passed to the zygote.
// main function.
//
// Note that we must copy argument string values since we will rewrite the
// entire argument block when we apply the nice name to argv0.
int i;
for (i = 0; i < argc; i++) {
if (argv[i][0] != '-') {
break;
}
if (argv[i][1] == '-' && argv[i][2] == 0) {
++i; // Skip --.
break;
}
runtime.addOption(strdup(argv[i]));
}
// Parse runtime arguments. Stop at first unrecognized option.
bool zygote = false;
bool startSystemServer = false;
bool application = false;
String8 niceName;
String8 className;
++i; // Skip unused "parent dir" argument.
while (i < argc) {
const char* arg = argv[i++];
if (strcmp(arg, "--zygote") == 0) {
zygote = true;
niceName = ZYGOTE_NICE_NAME;
} else if (strcmp(arg, "--start-system-server") == 0) {
startSystemServer = true;
} else if (strcmp(arg, "--application") == 0) {
application = true;
} else if (strncmp(arg, "--nice-name=", 12) == 0) {
niceName.setTo(arg + 12);
} else if (strncmp(arg, "--", 2) != 0) {
className.setTo(arg);
break;
} else {
--i;
break;
}
}
Vector<String8> args;
if (!className.isEmpty()) {
// We're not in zygote mode, the only argument we need to pass
// to RuntimeInit is the application argument.
//
// The Remainder of args get passed to startup class main(). Make
// copies of them before we overwrite them with the process name.
args.add(application ? String8("application") : String8("tool"));
runtime.setClassNameAndArgs(className, argc - i, argv + i);
} else {
// We're in zygote mode.
maybeCreateDalvikCache();
if (startSystemServer) {
args.add(String8("start-system-server"));
}
char prop[PROP_VALUE_MAX];
if (property_get(ABI_LIST_PROPERTY, prop, NULL) == 0) {
LOG_ALWAYS_FATAL("app_process: Unable to determine ABI list from property %s.",
ABI_LIST_PROPERTY);
return 11;
}
String8 abiFlag("--abi-list=");
abiFlag.append(prop);
args.add(abiFlag);
// In zygote mode, pass all remaining arguments to the zygote
// main() method.
for (; i < argc; ++i) {
args.add(String8(argv[i]));
}
}
if (!niceName.isEmpty()) {
runtime.setArgv0(niceName.string());
set_process_name(niceName.string());
}
if (zygote) {
runtime.start("com.android.internal.os.ZygoteInit", args, zygote);
} else if (className) {
runtime.start("com.android.internal.os.RuntimeInit", args, zygote);
} else {
fprintf(stderr, "Error: no class name or --zygote supplied.\n");
app_usage();
LOG_ALWAYS_FATAL("app_process: no class name or --zygote supplied.");
return 10;
}
}
因为zygote是在init进程中启动的，我们的启动它的时候传了一些参数上来，我们看看这些参数有什么：
在system/core/init/init.cpp中的service_start函数中，使用如下代码启动zygote。
if (execve(svc->args[0], (char**) svc->args, (char**) ENV) < 0) {
ERROR("cannot execve('%s'): %s\n", svc->args[0], strerror(errno));
}
svc->args[0]=“/system/bin/app_process”，因为在.rc中就是这么配置的。svc->args=“/system/bin/app_process -Xzygote /system/bin –zygote –start-system-server”,因为svc->args的构造是在service的解析中完成的，具体过程可以参考 Android6.0系统启动流程分析一：init进程。
理清了参数，我们再来回到main函数。main函数主要做的事情有：
1. 构造一个AppRuntime实例。
2. 解析参数，通过解析参数，我们知道：
zygote = true;
startSystemServer = true;
application = false;
3. 调用AppRuntime的start方法。
前面没什么好说的，我们看看AppRuntime的start做了什么。
调用这个方法时，传入了如下参数：
参数一：”com.android.internal.os.ZygoteInit”
参数二：/system/bin/app_process -Xzygote /system/bin –zygote –start-system-server”
参数三：true
下面看看这个start方法：
/*
* Start the Android runtime. This involves starting the virtual machine
* and calling the "static void main(String[] args)" method in the class
* named by "className".
*
* Passes the main function two arguments, the class name and the specified
* options string.
*/
void AndroidRuntime::start(const char* className, const Vector<String8>& options, bool zygote)
{
ALOGD(">>>>>> START %s uid %d <<<<<<\n",
className != NULL ? className : "(unknown)", getuid());
static const String8 startSystemServer("start-system-server");
/*
* 'startSystemServer == true' means runtime is obsolete and not run from
* init.rc anymore, so we print out the boot start event here.
*/
for (size_t i = 0; i < options.size(); ++i) {
if (options[i] == startSystemServer) {
/* track our progress through the boot sequence */
const int LOG_BOOT_PROGRESS_START = 3000;
LOG_EVENT_LONG(LOG_BOOT_PROGRESS_START, ns2ms(systemTime(SYSTEM_TIME_MONOTONIC)));
}
}
const char* rootDir = getenv("ANDROID_ROOT");
if (rootDir == NULL) {
rootDir = "/system";
if (!hasDir("/system")) {
LOG_FATAL("No root directory specified, and /android does not exist.");
return;
}
setenv("ANDROID_ROOT", rootDir, 1);
}
//const char* kernelHack = getenv("LD_ASSUME_KERNEL");
//ALOGD("Found LD_ASSUME_KERNEL='%s'\n", kernelHack);
/* start the virtual machine */
JniInvocation jni_invocation;
jni_invocation.Init(NULL);
JNIEnv* env;
if (startVm(&mJavaVM, &env, zygote) != 0) {
return;
}
onVmCreated(env);
/*
* Register android functions.
*/
if (startReg(env) < 0) {
ALOGE("Unable to register all android natives\n");
return;
}
/*
* We want to call main() with a String array with arguments in it.
* At present we have two arguments, the class name and an option string.
* Create an array to hold them.
*/
jclass stringClass;
jobjectArray strArray;
jstring classNameStr;
stringClass = env->FindClass("java/lang/String");
assert(stringClass != NULL);
strArray = env->NewObjectArray(options.size() + 1, stringClass, NULL);
assert(strArray != NULL);
classNameStr = env->NewStringUTF(className);
assert(classNameStr != NULL);
env->SetObjectArrayElement(strArray, 0, classNameStr);
for (size_t i = 0; i < options.size(); ++i) {
jstring optionsStr = env->NewStringUTF(options.itemAt(i).string());
assert(optionsStr != NULL);
env->SetObjectArrayElement(strArray, i + 1, optionsStr);
}
/*
* Start VM. This thread becomes the main thread of the VM, and will
* not return until the VM exits.
*/
char* slashClassName = toSlashClassName(className);
jclass startClass = env->FindClass(slashClassName);
if (startClass == NULL) {
ALOGE("JavaVM unable to locate class '%s'\n", slashClassName);
/* keep going */
} else {
jmethodID startMeth = env->GetStaticMethodID(startClass, "main",
"([Ljava/lang/String;)V");
if (startMeth == NULL) {
ALOGE("JavaVM unable to find main() in '%s'\n", className);
/* keep going */
} else {
env->CallStaticVoidMethod(startClass, startMeth, strArray);
#if 0
if (env->ExceptionCheck())
threadExitUncaughtException(env);
#endif
}
}
free(slashClassName);
ALOGD("Shutting down VM\n");
if (mJavaVM->DetachCurrentThread() != JNI_OK)
ALOGW("Warning: unable to detach main thread\n");
if (mJavaVM->DestroyJavaVM() != 0)
ALOGW("Warning: VM did not shut down cleanly\n");
}
这份函数作如下事情：
1. 启动java虚拟机。对虚拟机感兴趣的，这里或许是一个重要的突破口。我对虚拟机了解不多，这里就不展开了。
2. 注册本地方法。注册本地方式调用startReg方法，这个方法可以看一看：
/*static*/ int AndroidRuntime::startReg(JNIEnv* env)
{
/*
* This hook causes all future threads created in this process to be
* attached to the JavaVM. (This needs to go away in favor of JNI
* Attach calls.)
*/
androidSetCreateThreadFunc((android_create_thread_fn) javaCreateThreadEtc);
ALOGV("--- registering native functions ---\n");
/*
* Every "register" function calls one or more things that return
* a local reference (e.g. FindClass). Because we haven't really
* started the VM yet, they're all getting stored in the base frame
* and never released. Use Push/Pop to manage the storage.
*/
env->PushLocalFrame(200);
if (register_jni_procs(gRegJNI, NELEM(gRegJNI), env) < 0) {
env->PopLocalFrame(NULL);
return -1;
}
env->PopLocalFrame(NULL);
//createJavaThread("fubar", quickTest, (void*) "hello");
return 0;
}
这个方法中，使用register_jni_procs方法完成注册，该方法如下：
static int register_jni_procs(const RegJNIRec array[], size_t count, JNIEnv* env)
{
for (size_t i = 0; i < count; i++) {
if (array[i].mProc(env) < 0) {
#ifndef NDEBUG
ALOGD("----------!!! %s failed to load\n", array[i].mName);
#endif
return -1;
}
}
return 0;
}
这个方法传入的一个数组作为参数，这个方法的作用就是遍历这个数组，调用这个数组的mProc方法。这个数组就是gRegJNI：
static const RegJNIRec gRegJNI[] = {
REG_JNI(register_com_android_internal_os_RuntimeInit),
REG_JNI(register_android_os_SystemClock),
REG_JNI(register_android_util_EventLog),
REG_JNI(register_android_util_Log),
REG_JNI(register_android_content_AssetManager),
REG_JNI(register_android_content_StringBlock),
REG_JNI(register_android_content_XmlBlock),
REG_JNI(register_android_emoji_EmojiFactory),
...
};
也就是说这个方法会调用上述数组中的每一项REG_JNI中声明的函数来实现各个模块的jni函数的注册的。
回到start方法，接下来就是用jni规范，从c++中调用java中的静态方法了，这个方法就是ZygoteInit中的main方法。
Zygote进程的java部分
现在代码从c++转入到java部分了，我们来看ZygoteInit中的main方法：
public static void main(String argv[]) {
try {
RuntimeInit.enableDdms();
// Start profiling the zygote initialization.
SamplingProfilerIntegration.start();
boolean startSystemServer = false;
String socketName = "zygote";
String abiList = null;
for (int i = 1; i < argv.length; i++) {
if ("start-system-server".equals(argv[i])) {
startSystemServer = true;
} else if (argv[i].startsWith(ABI_LIST_ARG)) {
abiList = argv[i].substring(ABI_LIST_ARG.length());
} else if (argv[i].startsWith(SOCKET_NAME_ARG)) {
socketName = argv[i].substring(SOCKET_NAME_ARG.length());
} else {
throw new RuntimeException("Unknown command line argument: " + argv[i]);
}
}
if (abiList == null) {
throw new RuntimeException("No ABI list supplied.");
}
registerZygoteSocket(socketName);
EventLog.writeEvent(LOG_BOOT_PROGRESS_PRELOAD_START,
SystemClock.uptimeMillis());
preload();
EventLog.writeEvent(LOG_BOOT_PROGRESS_PRELOAD_END,
SystemClock.uptimeMillis());
// Finish profiling the zygote initialization.
SamplingProfilerIntegration.writeZygoteSnapshot();
// Do an initial gc to clean up after startup
gcAndFinalize();
// Disable tracing so that forked processes do not inherit stale tracing tags from
// Zygote.
Trace.setTracingEnabled(false);
if (startSystemServer) {
startSystemServer(abiList, socketName);
}
Log.i(TAG, "Accepting command socket connections");
runSelectLoop(abiList);
closeServerSocket();
} catch (MethodAndArgsCaller caller) {
caller.run();
} catch (RuntimeException ex) {
Log.e(TAG, "Zygote died with exception", ex);
closeServerSocket();
throw ex;
}
}
这个方法做了如下事情：
1. 注册Zytote套接字。这个套接字是在system/core/init/init.cpp中的service_start方法中创建的，不清楚的请看上一篇博客。service_start方法创建完套接字以后会发布这个套接字，使用publish_socket方法。这个方法也是定义在system/core/init/init.cpp中:
static void publish_socket(const char *name, int fd)
{
char key[64] = ANDROID_SOCKET_ENV_PREFIX;
char val[64];
strlcpy(key + sizeof(ANDROID_SOCKET_ENV_PREFIX) - 1,
name,
sizeof(key) - sizeof(ANDROID_SOCKET_ENV_PREFIX));
snprintf(val, sizeof(val), "%d", fd);
add_environment(key, val);
/* make sure we don't close-on-exec */
fcntl(fd, F_SETFD, 0);
}
这里主要使用add_environment方法把这个套接字添加到环境变量中，添加的结果是一对键值对：(ANDROID_SOCKET_zygote,fd)。那么我们在这里是不是就可以从环境变量中获取呢？
看看registerZygoteSocket方法：
private static void registerZygoteSocket(String socketName) {
if (sServerSocket == null) {
int fileDesc;
final String fullSocketName = ANDROID_SOCKET_PREFIX + socketName;
try {
String env = System.getenv(fullSocketName);
fileDesc = Integer.parseInt(env);
} catch (RuntimeException ex) {
throw new RuntimeException(fullSocketName + " unset or invalid", ex);
}
try {
FileDescriptor fd = new FileDescriptor();
fd.setInt$(fileDesc);
sServerSocket = new LocalServerSocket(fd);
} catch (IOException ex) {
throw new RuntimeException(
"Error binding to local socket '" + fileDesc + "'", ex);
}
}
}
果不其然吧，就是通过System.getenv方法获取到我们在service_start方法中创建的套接字，然后对其进一步封装。并把封装的结果保存在sServerSocket 变量中。
第一件事情就做完了，回到main方法。
2.初始化gc:gcAndFinalize();
3.启动SystemServer。
SystemServer是Android系统非常和核心的服务，它会在创建后启动系统中的其他服务，然后成为所有服务的管理者，向应用程序和其他服务提供服务。这一部分我们放在下一节来分析。
4.进入监听状态：runSelectLoop
Zygote监听套接字
private static void runSelectLoop(String abiList) throws MethodAndArgsCaller {
ArrayList<FileDescriptor> fds = new ArrayList<FileDescriptor>();
ArrayList<ZygoteConnection> peers = new ArrayList<ZygoteConnection>();
fds.add(sServerSocket.getFileDescriptor());
peers.add(null);
while (true) {
StructPollfd[] pollFds = new StructPollfd[fds.size()];
for (int i = 0; i < pollFds.length; ++i) {
pollFds[i] = new StructPollfd();
pollFds[i].fd = fds.get(i);
pollFds[i].events = (short) POLLIN;
}
try {
Os.poll(pollFds, -1);
} catch (ErrnoException ex) {
throw new RuntimeException("poll failed", ex);
}
for (int i = pollFds.length - 1; i >= 0; --i) {
if ((pollFds[i].revents & POLLIN) == 0) {
continue;
}
if (i == 0) {
ZygoteConnection newPeer = acceptCommandPeer(abiList);
peers.add(newPeer);
fds.add(newPeer.getFileDesciptor());
} else {
boolean done = peers.get(i).runOnce();
if (done) {
peers.remove(i);
fds.remove(i);
}
}
}
}
}
这个方法开始进入zygote套接字的监听状态了。当zygote套接字接受到写入操作，Zygote进程唤醒，执行ZygoteConnection的runOnce方法。
我们知道Zygote俗称“受精卵”，它是所有java进程的祖先进程。这里的runOnce便是创建进程的开端了。虽然暂时不知道上层在什么情况下，如何向zytoge套接字写入数据，从而开启进程创建的，但是没有关系，我们依然可以顺着代码的思路，看看从runOnce开始，是如何创建出一个进程的。
Zygote创建子进程的过程
runOnce方法如下：
/**
* Reads one start command from the command socket. If successful,
* a child is forked and a {@link ZygoteInit.MethodAndArgsCaller}
* exception is thrown in that child while in the parent process,
* the method returns normally. On failure, the child is not
* spawned and messages are printed to the log and stderr. Returns
* a boolean status value indicating whether an end-of-file on the command
* socket has been encountered.
*
* @return false if command socket should continue to be read from, or
* true if an end-of-file has been encountered.
* @throws ZygoteInit.MethodAndArgsCaller trampoline to invoke main()
* method in child process
*/
boolean runOnce() throws ZygoteInit.MethodAndArgsCaller {
String args[];
Arguments parsedArgs = null;
FileDescriptor[] descriptors;
try {
args = readArgumentList();
descriptors = mSocket.getAncillaryFileDescriptors();
} catch (IOException ex) {
Log.w(TAG, "IOException on command socket " + ex.getMessage());
closeSocket();
return true;
}
if (args == null) {
// EOF reached.
closeSocket();
return true;
}
/** the stderr of the most recent request, if avail */
PrintStream newStderr = null;
if (descriptors != null && descriptors.length >= 3) {
newStderr = new PrintStream(
new FileOutputStream(descriptors[2]));
}
int pid = -1;
FileDescriptor childPipeFd = null;
FileDescriptor serverPipeFd = null;
try {
parsedArgs = new Arguments(args);
if (parsedArgs.abiListQuery) {
return handleAbiListQuery();
}
if (parsedArgs.permittedCapabilities != 0 || parsedArgs.effectiveCapabilities != 0) {
throw new ZygoteSecurityException("Client may not specify capabilities: " +
"permitted=0x" + Long.toHexString(parsedArgs.permittedCapabilities) +
", effective=0x" + Long.toHexString(parsedArgs.effectiveCapabilities));
}
applyUidSecurityPolicy(parsedArgs, peer);
applyInvokeWithSecurityPolicy(parsedArgs, peer);
applyDebuggerSystemProperty(parsedArgs);
applyInvokeWithSystemProperty(parsedArgs);
int[][] rlimits = null;
if (parsedArgs.rlimits != null) {
rlimits = parsedArgs.rlimits.toArray(intArray2d);
}
if (parsedArgs.invokeWith != null) {
FileDescriptor[] pipeFds = Os.pipe2(O_CLOEXEC);
childPipeFd = pipeFds[1];
serverPipeFd = pipeFds[0];
Os.fcntlInt(childPipeFd, F_SETFD, 0);
}
/**
* In order to avoid leaking descriptors to the Zygote child,
* the native code must close the two Zygote socket descriptors
* in the child process before it switches from Zygote-root to
* the UID and privileges of the application being launched.
*
* In order to avoid "bad file descriptor" errors when the
* two LocalSocket objects are closed, the Posix file
* descriptors are released via a dup2() call which closes
* the socket and substitutes an open descriptor to /dev/null.
*/
int [] fdsToClose = { -1, -1 };
FileDescriptor fd = mSocket.getFileDescriptor();
if (fd != null) {
fdsToClose[0] = fd.getInt$();
}
fd = ZygoteInit.getServerSocketFileDescriptor();
if (fd != null) {
fdsToClose[1] = fd.getInt$();
}
fd = null;
pid = Zygote.forkAndSpecialize(parsedArgs.uid, parsedArgs.gid, parsedArgs.gids,
parsedArgs.debugFlags, rlimits, parsedArgs.mountExternal, parsedArgs.seInfo,
parsedArgs.niceName, fdsToClose, parsedArgs.instructionSet,
parsedArgs.appDataDir);
} catch (ErrnoException ex) {
logAndPrintError(newStderr, "Exception creating pipe", ex);
} catch (IllegalArgumentException ex) {
logAndPrintError(newStderr, "Invalid zygote arguments", ex);
} catch (ZygoteSecurityException ex) {
logAndPrintError(newStderr,
"Zygote security policy prevents request: ", ex);
}
try {
if (pid == 0) {
// in child
IoUtils.closeQuietly(serverPipeFd);
serverPipeFd = null;
handleChildProc(parsedArgs, descriptors, childPipeFd, newStderr);
// should never get here, the child is expected to either
// throw ZygoteInit.MethodAndArgsCaller or exec().
return true;
} else {
// in parent...pid of < 0 means failure
IoUtils.closeQuietly(childPipeFd);
childPipeFd = null;
return handleParentProc(pid, descriptors, serverPipeFd, parsedArgs);
}
} finally {
IoUtils.closeQuietly(childPipeFd);
IoUtils.closeQuietly(serverPipeFd);
}
}
这个方法比较长，我们先看看它的注释，大概意思如下：
这个方法会从socket中读取一个命令，成功的话就会创建一个子进程出来，并且会抛出一个异常：ZygoteInit.MethodAndArgsCaller 。
也就是说即使创建成功了也会抛出异常，然后会在Zygote的main方法中接受检测到异常，进而调用caller.run()方法作进一步处理。
整个过程有很多地方看不懂，这里主要梳理的是整个流程。大神请绕道…
下面对这个过程做简要分析。
runOnce方法主要调用了 Zygote.forkAndSpecialize方法进一步处理，这个方法执行完成以后 ，子进程就已经创建好了，这个时候pid=0也就是在子进程中执行，在执行handleChildProc时，子进程会抛出异常，异常被捕获后执行MethodAndArgsCaller类中的run方法被执行，这个方法最终会调用到ActivityThread的main方法。这个过程是在子进程中实现的。pid!=0就意味着后面的代码是在Zygote进程中执行，这个时候执行的handleParentProc方法，这个方法会做一些清理工作（从注释了解到的，具体code没能理解）。
接下来看看forkAndSpecialize方法，这个方法定义下Zygote类中，看看它怎么一步步创建子进程的。
public static int forkAndSpecialize(int uid, int gid, int[] gids, int debugFlags,
int[][] rlimits, int mountExternal, String seInfo, String niceName, int[] fdsToClose,
String instructionSet, String appDataDir) {
VM_HOOKS.preFork();
int pid = nativeForkAndSpecialize(
uid, gid, gids, debugFlags, rlimits, mountExternal, seInfo, niceName, fdsToClose,
instructionSet, appDataDir);
// Enable tracing as soon as possible for the child process.
if (pid == 0) {
Trace.setTracingEnabled(true);
// Note that this event ends at the end of handleChildProc,
Trace.traceBegin(Trace.TRACE_TAG_ACTIVITY_MANAGER, "PostFork");
}
VM_HOOKS.postForkCommon();
return pid;
}
三件事情：
1. VM_HOOKS.preFork(),做准备工作。
2. nativeForkAndSpecialize，创建子进程
3. VM_HOOKS.postForkCommon();启动Zygote的4个Daemon线程，java堆整理，引用队列，以及析构线程。
step 1
首先看看preFork方法，这个方法定义在libcore\dalvik\src\main\java\dalvik\system\ZygoteHooks类中。
public void preFork() {
Daemons.stop();
waitUntilAllThreadsStopped();
token = nativePreFork();
}
Daemons.stop():
public static void stop() {
HeapTaskDaemon.INSTANCE.stop();
ReferenceQueueDaemon.INSTANCE.stop();
FinalizerDaemon.INSTANCE.stop();
FinalizerWatchdogDaemon.INSTANCE.stop();
}
停止四个线程：Daemon线程，java堆整理，引用队列，析构线程
也就是创建子进程的时候，不能有这几个线程搅和。
waitUntilAllThreadsStopped：
private static void waitUntilAllThreadsStopped() {
File tasks = new File("/proc/self/task");
// All Java daemons are stopped already. We're just waiting for their OS counterparts to
// finish as well. This shouldn't take much time so spinning is ok here.
while (tasks.list().length > 1) {
Thread.yield();
}
}
}
当/proc/self/task文件中记录的线程数大于1，就不断的让出cpu，直到只剩下一个线程。
nativePreFork：
这个方法定义在art\runtime\native\dalvik_system_ZygoteHooks.cc中：
static jlong ZygoteHooks_nativePreFork(JNIEnv* env, jclass) {
Runtime* runtime = Runtime::Current();
CHECK(runtime->IsZygote()) << "runtime instance not started with -Xzygote";
runtime->PreZygoteFork();
if (Trace::GetMethodTracingMode() != TracingMode::kTracingInactive) {
// Tracing active, pause it.
Trace::Pause();
}
// Grab thread before fork potentially makes Thread::pthread_key_self_ unusable.
return reinterpret_cast<jlong>(ThreadForEnv(env));
}
调用PreZygoteFork方法，这个方法定义在art\runtime\runtime.cc中：
void Runtime::PreZygoteFork() {
heap_->PreZygoteFork();
}
这个函数用来初始化堆。
step 2
nativeForkAndSpecialize方法定义在frameworks\base\core\jni\com_android_internal_os_Zygote.cpp中：
static jint com_android_internal_os_Zygote_nativeForkAndSpecialize(
JNIEnv* env, jclass, jint uid, jint gid, jintArray gids,
jint debug_flags, jobjectArray rlimits,
jint mount_external, jstring se_info, jstring se_name,
jintArray fdsToClose, jstring instructionSet, jstring appDataDir) {
// Grant CAP_WAKE_ALARM to the Bluetooth process.
jlong capabilities = 0;
if (uid == AID_BLUETOOTH) {
capabilities |= (1LL << CAP_WAKE_ALARM);
}
return ForkAndSpecializeCommon(env, uid, gid, gids, debug_flags,
rlimits, capabilities, capabilities, mount_external, se_info,
se_name, false, fdsToClose, instructionSet, appDataDir);
}
调用ForkAndSpecializeCommon方法进一步处理：
// Utility routine to fork zygote and specialize the child process.
static pid_t ForkAndSpecializeCommon(JNIEnv* env, uid_t uid, gid_t gid, jintArray javaGids,
jint debug_flags, jobjectArray javaRlimits,
jlong permittedCapabilities, jlong effectiveCapabilities,
jint mount_external,
jstring java_se_info, jstring java_se_name,
bool is_system_server, jintArray fdsToClose,
jstring instructionSet, jstring dataDir) {
SetSigChldHandler();
pid_t pid = fork();
if (pid == 0) {
// The child process.
gMallocLeakZygoteChild = 1;
// Clean up any descriptors which must be closed immediately
DetachDescriptors(env, fdsToClose);
// Keep capabilities across UID change, unless we're staying root.
if (uid != 0) {
EnableKeepCapabilities(env);
}
DropCapabilitiesBoundingSet(env);
bool use_native_bridge = !is_system_server && (instructionSet != NULL)
&& android::NativeBridgeAvailable();
if (use_native_bridge) {
ScopedUtfChars isa_string(env, instructionSet);
use_native_bridge = android::NeedsNativeBridge(isa_string.c_str());
}
if (use_native_bridge && dataDir == NULL) {
// dataDir should never be null if we need to use a native bridge.
// In general, dataDir will never be null for normal applications. It can only happen in
// special cases (for isolated processes which are not associated with any app). These are
// launched by the framework and should not be emulated anyway.
use_native_bridge = false;
ALOGW("Native bridge will not be used because dataDir == NULL.");
}
if (!MountEmulatedStorage(uid, mount_external, use_native_bridge)) {
ALOGW("Failed to mount emulated storage: %s", strerror(errno));
if (errno == ENOTCONN || errno == EROFS) {
// When device is actively encrypting, we get ENOTCONN here
// since FUSE was mounted before the framework restarted.
// When encrypted device is booting, we get EROFS since
// FUSE hasn't been created yet by init.
// In either case, continue without external storage.
} else {
ALOGE("Cannot continue without emulated storage");
RuntimeAbort(env);
}
}
if (!is_system_server) {
int rc = createProcessGroup(uid, getpid());
if (rc != 0) {
if (rc == -EROFS) {
ALOGW("createProcessGroup failed, kernel missing CONFIG_CGROUP_CPUACCT?");
} else {
ALOGE("createProcessGroup(%d, %d) failed: %s", uid, pid, strerror(-rc));
}
}
}
SetGids(env, javaGids);
SetRLimits(env, javaRlimits);
if (use_native_bridge) {
ScopedUtfChars isa_string(env, instructionSet);
ScopedUtfChars data_dir(env, dataDir);
android::PreInitializeNativeBridge(data_dir.c_str(), isa_string.c_str());
}
int rc = setresgid(gid, gid, gid);
if (rc == -1) {
ALOGE("setresgid(%d) failed: %s", gid, strerror(errno));
RuntimeAbort(env);
}
rc = setresuid(uid, uid, uid);
if (rc == -1) {
ALOGE("setresuid(%d) failed: %s", uid, strerror(errno));
RuntimeAbort(env);
}
if (NeedsNoRandomizeWorkaround()) {
// Work around ARM kernel ASLR lossage (http://b/5817320).
int old_personality = personality(0xffffffff);
int new_personality = personality(old_personality | ADDR_NO_RANDOMIZE);
if (new_personality == -1) {
ALOGW("personality(%d) failed: %s", new_personality, strerror(errno));
}
}
SetCapabilities(env, permittedCapabilities, effectiveCapabilities);
SetSchedulerPolicy(env);
const char* se_info_c_str = NULL;
ScopedUtfChars* se_info = NULL;
if (java_se_info != NULL) {
se_info = new ScopedUtfChars(env, java_se_info);
se_info_c_str = se_info->c_str();
if (se_info_c_str == NULL) {
ALOGE("se_info_c_str == NULL");
RuntimeAbort(env);
}
}
const char* se_name_c_str = NULL;
ScopedUtfChars* se_name = NULL;
if (java_se_name != NULL) {
se_name = new ScopedUtfChars(env, java_se_name);
se_name_c_str = se_name->c_str();
if (se_name_c_str == NULL) {
ALOGE("se_name_c_str == NULL");
RuntimeAbort(env);
}
}
rc = selinux_android_setcontext(uid, is_system_server, se_info_c_str, se_name_c_str);
if (rc == -1) {
ALOGE("selinux_android_setcontext(%d, %d, \"%s\", \"%s\") failed", uid,
is_system_server, se_info_c_str, se_name_c_str);
RuntimeAbort(env);
}
// Make it easier to debug audit logs by setting the main thread's name to the
// nice name rather than "app_process".
if (se_info_c_str == NULL && is_system_server) {
se_name_c_str = "system_server";
}
if (se_info_c_str != NULL) {
SetThreadName(se_name_c_str);
}
delete se_info;
delete se_name;
UnsetSigChldHandler();
env->CallStaticVoidMethod(gZygoteClass, gCallPostForkChildHooks, debug_flags,
is_system_server ? NULL : instructionSet);
if (env->ExceptionCheck()) {
ALOGE("Error calling post fork hooks.");
RuntimeAbort(env);
}
} else if (pid > 0) {
// the parent process
}
return pid;
}
这个方法出现了我们非常熟悉的fork方法，然后在子进程中调用Zygote类中的callPostForkChildHooks方法：
private static void callPostForkChildHooks(int debugFlags, String instructionSet) {
VM_HOOKS.postForkChild(debugFlags, instructionSet);
}
这里就不进一步追踪下去了，总之ForkAndSpecializeCommon方法调用fork系统调用创建了子进程，在进程中做了些初始化工作，然后返回了pid。这里会返回两次，子进程返回一次，父进程返回一次。
step 3
返回到nativeForkAndSpecialize方法后进一步返回，最终回到Zygote类的forkAndSpecialize方法，这个方法接下来会调用postForkCommon方法。postForkCommon方法定义在libcore\dalvik\src\main\java\dalvik\system\ZygoteHooks类中：
/**
* Called by the zygote in both the parent and child processes after
* every fork. In the child process, this method is called after
* {@code postForkChild}.
*/
public void postForkCommon() {
Daemons.start();
}
注意这里子进程和父进程都会执行，也就是都会启动4个线程:Daemon线程，java堆整理，引用队列，以及析构线程。
父进程的线程在fork之前停止了，这里也要重新启动。
子进程和父进程都继续返回。
返回到ZygoteConnect的runOnce方法中，继续往下执行，则子进程执行handleChildProc方法，父进程执行handleParentProc方法。
父进程不是我们关注的，我们关注的是子进程，所以接下来看看handleChildProc方法，这个方法定义在frameworks\base\core\java\com\android\internal\os\ZygoteConnection中：
private void handleChildProc(Arguments parsedArgs,
FileDescriptor[] descriptors, FileDescriptor pipeFd, PrintStream newStderr)
throws ZygoteInit.MethodAndArgsCaller {
/**
* By the time we get here, the native code has closed the two actual Zygote
* socket connections, and substituted /dev/null in their place. The LocalSocket
* objects still need to be closed properly.
*/
closeSocket();
ZygoteInit.closeServerSocket();
if (descriptors != null) {
try {
Os.dup2(descriptors[0], STDIN_FILENO);
Os.dup2(descriptors[1], STDOUT_FILENO);
Os.dup2(descriptors[2], STDERR_FILENO);
for (FileDescriptor fd: descriptors) {
IoUtils.closeQuietly(fd);
}
newStderr = System.err;
} catch (ErrnoException ex) {
Log.e(TAG, "Error reopening stdio", ex);
}
}
if (parsedArgs.niceName != null) {
Process.setArgV0(parsedArgs.niceName);
}
// End of the postFork event.
Trace.traceEnd(Trace.TRACE_TAG_ACTIVITY_MANAGER);
if (parsedArgs.invokeWith != null) {
WrapperInit.execApplication(parsedArgs.invokeWith,
parsedArgs.niceName, parsedArgs.targetSdkVersion,
VMRuntime.getCurrentInstructionSet(),
pipeFd, parsedArgs.remainingArgs);
} else {
RuntimeInit.zygoteInit(parsedArgs.targetSdkVersion,
parsedArgs.remainingArgs, null /* classLoader */);
}
}
子进程会继承父进程打开的文件描述符，所以子进程中有zygote套接字描述符，这里需要把它关掉。然后重要的是调用RuntimeInit.zygoteInit方法。这个方法定义在frameworks\base\core\java\com\android\internal\os\RuntimeInit中：
public static final void zygoteInit(int targetSdkVersion, String[] argv, ClassLoader classLoader)
throws ZygoteInit.MethodAndArgsCaller {
if (DEBUG) Slog.d(TAG, "RuntimeInit: Starting application from zygote");
Trace.traceBegin(Trace.TRACE_TAG_ACTIVITY_MANAGER, "RuntimeInit");
redirectLogStreams();
commonInit();
nativeZygoteInit();
applicationInit(targetSdkVersion, argv, classLoader);
}
从名字看也是做初始化工作。重点来看看applicationInit方法吧，这个方法定义在frameworks\base\core\java\com\android\internal\os\RuntimeInit中：
private static void applicationInit(int targetSdkVersion, String[] argv, ClassLoader classLoader)
throws ZygoteInit.MethodAndArgsCaller {
// If the application calls System.exit(), terminate the process
// immediately without running any shutdown hooks. It is not possible to
// shutdown an Android application gracefully. Among other things, the
// Android runtime shutdown hooks close the Binder driver, which can cause
// leftover running threads to crash before the process actually exits.
nativeSetExitWithoutCleanup(true);
// We want to be fairly aggressive about heap utilization, to avoid
// holding on to a lot of memory that isn't needed.
VMRuntime.getRuntime().setTargetHeapUtilization(0.75f);
VMRuntime.getRuntime().setTargetSdkVersion(targetSdkVersion);
final Arguments args;
try {
args = new Arguments(argv);
} catch (IllegalArgumentException ex) {
Slog.e(TAG, ex.getMessage());
// let the process exit
return;
}
// The end of of the RuntimeInit event (see #zygoteInit).
Trace.traceEnd(Trace.TRACE_TAG_ACTIVITY_MANAGER);
// Remaining arguments are passed to the start class's static main
invokeStaticMain(args.startClass, args.startArgs, classLoader);
}
调用invokeStaticMain方法进一步处理，这个方法还是在RuntimeInit中：
private static void invokeStaticMain(String className, String[] argv, ClassLoader classLoader)
throws ZygoteInit.MethodAndArgsCaller {
Class<?> cl;
try {
cl = Class.forName(className, true, classLoader);
} catch (ClassNotFoundException ex) {
throw new RuntimeException(
"Missing class when invoking static main " + className,
ex);
}
Method m;
try {
m = cl.getMethod("main", new Class[] { String[].class });
} catch (NoSuchMethodException ex) {
throw new RuntimeException(
"Missing static main on " + className, ex);
} catch (SecurityException ex) {
throw new RuntimeException(
"Problem getting static main on " + className, ex);
}
int modifiers = m.getModifiers();
if (! (Modifier.isStatic(modifiers) && Modifier.isPublic(modifiers))) {
throw new RuntimeException(
"Main method is not public and static on " + className);
}
/*
* This throw gets caught in ZygoteInit.main(), which responds
* by invoking the exception's run() method. This arrangement
* clears up all the stack frames that were required in setting
* up the process.
*/
throw new ZygoteInit.MethodAndArgsCaller(m, argv);
}
这里会抛出异常，也就重新回到了ZygoteInit的main方法中：
Log.i(TAG, "Accepting command socket connections");
runSelectLoop(abiList);
closeServerSocket();
} catch (MethodAndArgsCaller caller) {
caller.run();
} catch (RuntimeException ex) {
Log.e(TAG, "Zygote died with exception", ex);
closeServerSocket();
throw ex;
}
这个时候会调用caller.run();也就是MethodAndArgsCaller中的run方法：
public void run() {
try {
mMethod.invoke(null, new Object[] { mArgs });
} catch (IllegalAccessException ex) {
throw new RuntimeException(ex);
} catch (InvocationTargetException ex) {
Throwable cause = ex.getCause();
if (cause instanceof RuntimeException) {
throw (RuntimeException) cause;
} else if (cause instanceof Error) {
throw (Error) cause;
}
throw new RuntimeException(ex);
}
}
}
接着调用mMethod.invoke方法，也就是Method类中的Invoke方法，这个方法是一本地方法,这个方法就展看了。根据Activity的启动流程来看，发起创建进程的请求的地方在ActivityManagerService中的startProcessLocked方法中，有如下代码：
Process.ProcessStartResult startResult = Process.start(entryPoint,
app.processName, uid, uid, gids, debugFlags, mountExternal,
app.info.targetSdkVersion, app.info.seinfo, requiredAbi, instructionSet,
app.info.dataDir, entryPointArgs);
mMethod.invoke方法中的参数mArgs 是有这里传下去的，感兴趣可以追踪下，总之mMethod.invoke会调用ActivityThread中的main方法。
至此，Zygote进程创建子进程的流程便分析结束。