android recovery 主系统代码分析

阅读完上一篇文章： http://blog.csdn.net/andyhuabing/article/details/9226569 

我们已经清楚了如何进入正常模式和Recovery模式已有深刻理解了，假设进入了Recovery模式，那么其核心代码是怎么做的呢？

代码路径在 android 源码的根路径： bootable ecovery 其入口文件就是 recovery.c 中 main函数

下面就开始逐步了解其Recovery的设计思想：

static const char *COMMAND_FILE = "/cache/recovery/command";
static const char *INTENT_FILE = "/cache/recovery/intent";
static const char *LOG_FILE = "/cache/recovery/log";

注解里面描述的相当清楚：

 * The recovery tool communicates with the main system through /cache files.
 *   /cache/recovery/command - INPUT - command line for tool, one arg per line
 *   /cache/recovery/log - OUTPUT - combined log file from recovery run(s)
 *   /cache/recovery/intent - OUTPUT - intent that was passed in

static const char *LAST_LOG_FILE = "/cache/recovery/last_log";

static const char *LAST_INSTALL_FILE = "/cache/recovery/last_install";
static const char *CACHE_ROOT = "/cache";
static const char *SDCARD_ROOT = "/sdcard";

下面的描述针对写入的 command 有大致的介绍：

 * The arguments which may be supplied in the recovery.command file:
 *   --send_intent=anystring - write the text out to recovery.intent
 *   --update_package=path - verify install an OTA package file
 *   --wipe_data - erase user data (and cache), then reboot
 *   --wipe_cache - wipe cache (but not user data), then reboot
 *   --set_encrypted_filesystem=on|off - enables / diasables encrypted fs

两种升级模式步骤说明：

 * After completing, we remove /cache/recovery/command and reboot.
 * Arguments may also be supplied in the bootloader control block (BCB).
 * These important scenarios must be safely restartable at any point:
 *
 * FACTORY RESET
 * 1. user selects "factory reset"
 * 2. main system writes "--wipe_data" to /cache/recovery/command
 * 3. main system reboots into recovery
 * 4. get_args() writes BCB with "boot-recovery" and "--wipe_data"
 *    -- after this, rebooting will restart the erase --
 * 5. erase_volume() reformats /data
 * 6. erase_volume() reformats /cache
 * 7. finish_recovery() erases BCB
 *    -- after this, rebooting will restart the main system --
 * 8. main() calls reboot() to boot main system
 *
 * OTA INSTALL
 * 1. main system downloads OTA package to /cache/some-filename.zip
 * 2. main system writes "--update_package=/cache/some-filename.zip"
 * 3. main system reboots into recovery
 * 4. get_args() writes BCB with "boot-recovery" and "--update_package=..."
 *    -- after this, rebooting will attempt to reinstall the update --
 * 5. install_package() attempts to install the update
 *    NOTE: the package install must itself be restartable from any point
 * 6. finish_recovery() erases BCB
 *    -- after this, rebooting will (try to) restart the main system --
 * 7. ** if install failed **
 *    7a. prompt_and_wait() shows an error icon and waits for the user
 *    7b; the user reboots (pulling the battery, etc) into the main system
 * 8. main() calls maybe_install_firmware_update()
 *    ** if the update contained radio/hboot firmware **:
 *    8a. m_i_f_u() writes BCB with "boot-recovery" and "--wipe_cache"
 *        -- after this, rebooting will reformat cache & restart main system --
 *    8b. m_i_f_u() writes firmware image into raw cache partition
 *    8c. m_i_f_u() writes BCB with "update-radio/hboot" and "--wipe_cache"
 *        -- after this, rebooting will attempt to reinstall firmware --
 *    8d. bootloader tries to flash firmware
 *    8e. bootloader writes BCB with "boot-recovery" (keeping "--wipe_cache")
 *        -- after this, rebooting will reformat cache & restart main system --
 *    8f. erase_volume() reformats /cache
 *    8g. finish_recovery() erases BCB
 *        -- after this, rebooting will (try to) restart the main system --
 * 9. main() calls reboot() to boot main system

从上面的几段注解中，基本上就明白的 Recovery 是如何工作的啦。下面就从具体代码开始一步步分析。

1、recovery main 函数

int
main(int argc, char **argv) {
    time_t start = time(NULL);

    // If these fail, there's not really anywhere to complain...
    freopen(TEMPORARY_LOG_FILE, "a", stdout); setbuf(stdout, NULL);
    freopen(TEMPORARY_LOG_FILE, "a", stderr); setbuf(stderr, NULL);
    printf("Starting recovery on %s", ctime(&start));

将标准输出和标准错误输出重定位到 "/tmp/recovery.log", 如果是 eng 模式，就可以通过 adb pull /tmp/recovery.log,  看到当前的 log 信息，这为我们提供了有效的调试手段。

ui_init();

一个简单的基于framebuffer的ui系统，叫miniui 主要建立了图像部分（gglInit、gr_init_font、framebuffer）及进度条和事件处理（input_callback）

load_volume_table();

根据 /etc/recovery.fstab 建立分区表

// command line args come from, in decreasing precedence:
//   - the actual command line
//   - the bootloader control block (one per line, after "recovery")
//   - the contents of COMMAND_FILE (one per line)

get_args(&argc, &argv);

从misc 分区以及 CACHE:recovery/command 文件中读入参数，写入到argc, argv （get_bootloader_message） 并有可能写回 misc 分区（set_bootloader_message）

做完以上事情后就开始解析具体参数：

    while ((arg = getopt_long(argc, argv, "", OPTIONS, NULL)) != -1) {
        switch (arg) {
        case 'p': previous_runs = atoi(optarg); break;
        case 's': send_intent = optarg; break;
        case 'u': update_package = optarg; break;
        case 'w': wipe_data = wipe_cache = 1; break;
        case 'c': wipe_cache = 1; break;
        case 't': ui_show_text(1); break;
        case '?':
            LOGE("Invalid command argument ");
            continue;
        }
    }

    printf("Command:");
    for (arg = 0; arg < argc; arg++) {
        printf(" "%s"", argv[arg]);
    }
    printf(" ");

以上仅仅是打印表明进入到哪一步，方便调试情况的掌握

下面的代码就是具体干的事情了：

    if (update_package != NULL) {
        status = install_package(update_package, &wipe_cache, TEMPORARY_INSTALL_FILE);
        if (status == INSTALL_SUCCESS && wipe_cache) {
            if (erase_volume("/cache")) {
                LOGE("Cache wipe (requested by package) failed.");
            }
        }
        if (status != INSTALL_SUCCESS) ui_print("Installation aborted. ");
    } else if (wipe_data) {
        if (device_wipe_data()) status = INSTALL_ERROR;
        if (erase_volume("/data")) status = INSTALL_ERROR;
        if (wipe_cache && erase_volume("/cache")) status = INSTALL_ERROR;
        if (status != INSTALL_SUCCESS) ui_print("Data wipe failed. ");
        clear_sdcard_update_bootloader_message();
    } else if (wipe_cache) {
        if (wipe_cache && erase_volume("/cache")) status = INSTALL_ERROR;
        if (status != INSTALL_SUCCESS) ui_print("Cache wipe failed. ");
        clear_sdcard_update_bootloader_message();
    } else {
        status = update_by_key();  // No command specified
    }

根据用户提供参数，调用各项功能，比如，安装一个升级包，擦除cache分区, 擦除user data分区等等，后在会将继续详细分解。

if (status != INSTALL_SUCCESS) prompt_and_wait();

如果前面做的操作成功则进入重启流程，否则由用户操作，可选操作为： reboot, 安装update.zip，除cache分区, 擦除user data分区

    // Otherwise, get ready to boot the main system...
    finish_recovery(send_intent);

先看函数注解：

// clear the recovery command and prepare to boot a (hopefully working) system,
// copy our log file to cache as well (for the system to read), and
// record any intent we were asked to communicate back to the system.
// this function is idempotent: call it as many times as you like.

其实主要的就是如下函数操作：

    // Remove the command file, so recovery won't repeat indefinitely.
    if (ensure_path_mounted(COMMAND_FILE) != 0 || 
        (unlink(COMMAND_FILE) && errno != ENOENT)) {
        LOGW("Can't unlink %s ", COMMAND_FILE);
    }
    

    将指定分区mounted 成功并 unlink 删除一个文件的目录项并减少它的链接数

    ensure_path_unmounted(CACHE_ROOT);

    将指定分区 unmounted 
    sync();  // For good measure.

对于上面的代码总结：

它的功能如下：

1、将前面定义的intent字符串写入（如果有的话）：CACHE:recovery/command

2、将 /tmp/recovery.log 复制到 "CACHE:recovery/log";

3、清空 misc 分区，这样重启就不会进入recovery模式

4、删除command 文件：CACHE:recovery/command;

最后重启机器

    ui_print("Rebooting... ");
    android_reboot(ANDROID_RB_RESTART, 0, 0);

2、factory reset 核心代码实现

按照前面所列的8条步骤，其中1-6及7-8都与 main 通用流程一样，不再复述。

 * 5. erase_volume() reformats /data
 * 6. erase_volume() reformats /cache

这两个操作是如何做到的呢？

if (erase_volume("/data")) status = INSTALL_ERROR;

if (erase_volume("/cache")) status = INSTALL_ERROR;

最后就是

clear_sdcard_update_bootloader_message();

看看 erase_volume() 函数先：

static int
erase_volume(const char *volume) {
    ui_set_background(BACKGROUND_ICON_INSTALLING);
    ui_show_indeterminate_progress();
    ui_print("Formatting %s...
", volume);

    ensure_path_unmounted(volume);

    if (strcmp(volume, "/cache") == 0) {
        // Any part of the log we'd copied to cache is now gone.
        // Reset the pointer so we copy from the beginning of the temp
        // log.
        tmplog_offset = 0;
    }

    return format_volume(volume);
}

上面红字标明的是重要函数调用

int ensure_path_unmounted(const char* path) {

Volume* v = volume_for_path(path);

 result = scan_mounted_volumes();

return unmount_mounted_volume(mv);

}

就是将指定的path中径mount point进行卸载掉，而 format_volume的主要功能就是：

MtdWriteContext *write = mtd_write_partition(partition);

mtd_erase_blocks(write, -1);

mtd_write_close(write);

不要细说了吧，就是将整个分区数据全清掉。

最后一个函数：

void
clear_sdcard_update_bootloader_message() {
    struct bootloader_message boot;
    memset(&boot, 0, sizeof(boot));
    set_bootloader_message(&boot);
}

就是将misc分区数据重置清0

这样子就完成的恢复出厂设置的情况了。将 data/cache分区erase擦掉就好了。

3、OTA 安装 核心代码实现

主要函数就是如何安装 Package ：

 * 5. install_package() attempts to install the update
 *    NOTE: the package install must itself be restartable from any point

int
install_package(const char* path, int* wipe_cache, const char* install_file)

-->

static int
really_install_package(const char *path, int* wipe_cache){

clear_sdcard_update_bootloader_message();

    ui_set_background(BACKGROUND_ICON_INSTALLING);
    ui_print("Finding update package... ");
    ui_show_indeterminate_progress();
    LOGI("Update location: %s ", path);

    更新 ui 显示

    for(;((i < 5)&&(ensure_path_mounted(path) != 0));i++){
        LOGE("Can't mount %s ",path);
        sleep(1);
    }
    if((i >= 5)&&(ensure_path_mounted(path) != 0)){
        return INSTALL_CORRUPT;
    }

    

  确保升级包所在分区已经mount,通常为 cache 分区或者 SD 分区

  RSAPublicKey* loadedKeys = load_keys(PUBLIC_KEYS_FILE, &numKeys);

// Look for an RSA signature embedded in the .ZIP file comment given
// the path to the zip.  Verify it matches one of the given public
// keys.
//
// Return VERIFY_SUCCESS, VERIFY_FAILURE (if any error is encountered
// or no key matches the signature).

  err = verify_file(path, loadedKeys, numKeys);

从/res/keys中装载公钥，并进行确认文件的合法性

    /* Try to open the package.
     */
    ZipArchive zip;
    err = mzOpenZipArchive(path, &zip);

    打开升级包，将相关信息存到ZipArchive数据机构中，便于后面处理。

    /* Verify and install the contents of the package.
     */
    ui_print("Installing update...
");
    return try_update_binary(path, &zip, wipe_cache);    

     进行最后的安装包文件

}

// If the package contains an update binary, extract it and run it.
static int
try_update_binary(const char *path, ZipArchive *zip, int* wipe_cache) {
    const ZipEntry* binary_entry =
            mzFindZipEntry(zip, ASSUMED_UPDATE_BINARY_NAME);

    char* binary = "/tmp/update_binary";
    unlink(binary);
    int fd = creat(binary, 0755);

    bool ok = mzExtractZipEntryToFile(zip, binary_entry, fd);
    close(fd);
    mzCloseZipArchive(zip);

   

将升级包内文件META-INF/com/google/android/update-binary 复制为/tmp/update_binary

    // When executing the update binary contained in the package, the
    // arguments passed are:
    //
    //   - the version number for this interface
    //
    //   - an fd to which the program can write in order to update the
    //     progress bar.  The program can write single-line commands:

    int pipefd[2];
    pipe(pipefd);

    char** args = malloc(sizeof(char*) * 5);

    args[0] = binary;
    args[1] = EXPAND(RECOVERY_API_VERSION);   // defined in Android.mk
    args[2] = malloc(10);
    sprintf(args[2], "%d", pipefd[1]);
    args[3] = (char*)path;
    args[4] = buf_uuid;
    args[5] = NULL;

    组装新的进程参数    

    pid_t pid = fork();
    if (pid == 0) { // child process
        close(pipefd[0]);
        execv(binary, args);

    } 

    // parent process
    close(pipefd[1]);

    ui_show_progress

    ui_set_progress

    ui_print

总结一下代码主要行为功能：

1、将会创建新的进程，执行：/tmp/update_binary

2、同时，会给该进程传入一些参数，其中最重要的就是一个管道fd，供新进程与原进程通信。

3、新进程诞生后，原进程就变成了一个服务进程，它提供若干UI更新服务：

a)   progress

b)   set_progress

c)   ui_print

这样，新进程就可以通过老进程的UI系统完成显示任务。而其他功能就靠它自己了。