X-Git-Url: https://www.chiark.greenend.org.uk/ucgi/~ianmdlvl/git?p=elogind.git;a=blobdiff_plain;f=src%2Fcore%2Fshutdown.c;h=4cbdf1258751b8ffb8d1429f415dc572b8bb2a10;hp=4709746de40b616fde8b8fbcb17260b9b5d34c21;hb=4968105790c65af58d4ab42bffa2a4bedc0be8ee;hpb=7801356442578ff6e1c65844eb9e65c819af4660 diff --git a/src/core/shutdown.c b/src/core/shutdown.c index 4709746de..4cbdf1258 100644 --- a/src/core/shutdown.c +++ b/src/core/shutdown.c @@ -24,7 +24,6 @@ #include #include #include -#include #include #include #include @@ -36,6 +35,7 @@ #include #include #include +#include #include "missing.h" #include "log.h" @@ -46,162 +46,172 @@ #include "virt.h" #include "watchdog.h" #include "killall.h" +#include "cgroup-util.h" +#include "def.h" +#include "switch-root.h" #define FINALIZE_ATTEMPTS 50 -static int prepare_new_root(void) { - static const char dirs[] = - "/run/initramfs/oldroot\0" - "/run/initramfs/proc\0" - "/run/initramfs/sys\0" - "/run/initramfs/dev\0" - "/run/initramfs/run\0"; +static char* arg_verb; - const char *dir; +static int parse_argv(int argc, char *argv[]) { + enum { + ARG_LOG_LEVEL = 0x100, + ARG_LOG_TARGET, + ARG_LOG_COLOR, + ARG_LOG_LOCATION, + }; - if (mount("/run/initramfs", "/run/initramfs", NULL, MS_BIND, NULL) < 0) { - log_error("Failed to mount bind /run/initramfs on /run/initramfs: %m"); - return -errno; - } + static const struct option options[] = { + { "log-level", required_argument, NULL, ARG_LOG_LEVEL }, + { "log-target", required_argument, NULL, ARG_LOG_TARGET }, + { "log-color", optional_argument, NULL, ARG_LOG_COLOR }, + { "log-location", optional_argument, NULL, ARG_LOG_LOCATION }, + {} + }; - if (mount(NULL, "/run/initramfs", NULL, MS_PRIVATE, NULL) < 0) { - log_error("Failed to make /run/initramfs private mount: %m"); - return -errno; - } + int c, r; - NULSTR_FOREACH(dir, dirs) - if (mkdir_p_label(dir, 0755) < 0 && errno != EEXIST) { - log_error("Failed to mkdir %s: %m", dir); - return -errno; - } + assert(argc >= 1); + assert(argv); - if (mount("/sys", "/run/initramfs/sys", NULL, MS_BIND, NULL) < 0) { - log_error("Failed to mount bind /sys on /run/initramfs/sys: %m"); - return -errno; - } + /* "-" prevents getopt from permuting argv[] and moving the verb away + * from argv[1]. Our interface to initrd promises it'll be there. */ + while ((c = getopt_long(argc, argv, "-", options, NULL)) >= 0) + switch (c) { - if (mount("/proc", "/run/initramfs/proc", NULL, MS_BIND, NULL) < 0) { - log_error("Failed to mount bind /proc on /run/initramfs/proc: %m"); - return -errno; - } + case ARG_LOG_LEVEL: + r = log_set_max_level_from_string(optarg); + if (r < 0) + log_error("Failed to parse log level %s, ignoring.", optarg); - if (mount("/dev", "/run/initramfs/dev", NULL, MS_BIND, NULL) < 0) { - log_error("Failed to mount bind /dev on /run/initramfs/dev: %m"); - return -errno; - } + break; - if (mount("/run", "/run/initramfs/run", NULL, MS_BIND, NULL) < 0) { - log_error("Failed to mount bind /run on /run/initramfs/run: %m"); - return -errno; - } + case ARG_LOG_TARGET: + r = log_set_target_from_string(optarg); + if (r < 0) + log_error("Failed to parse log target %s, ignoring", optarg); - return 0; -} + break; -static int pivot_to_new_root(void) { + case ARG_LOG_COLOR: - if (chdir("/run/initramfs") < 0) { - log_error("Failed to change directory to /run/initramfs: %m"); - return -errno; - } + if (optarg) { + r = log_show_color_from_string(optarg); + if (r < 0) + log_error("Failed to parse log color setting %s, ignoring", optarg); + } else + log_show_color(true); - /* Work-around for a kernel bug: for some reason the kernel - * refuses switching root if any file systems are mounted - * MS_SHARED. Hence remount them MS_PRIVATE here as a - * work-around. - * - * https://bugzilla.redhat.com/show_bug.cgi?id=847418 */ - if (mount(NULL, "/", NULL, MS_REC|MS_PRIVATE, NULL) < 0) - log_warning("Failed to make \"/\" private mount: %m"); - - if (pivot_root(".", "oldroot") < 0) { - log_error("pivot failed: %m"); - /* only chroot if pivot root succeeded */ - return -errno; - } + break; + + case ARG_LOG_LOCATION: + if (optarg) { + r = log_show_location_from_string(optarg); + if (r < 0) + log_error("Failed to parse log location setting %s, ignoring", optarg); + } else + log_show_location(true); + + break; - chroot("."); + case '\001': + if (!arg_verb) + arg_verb = optarg; + else + log_error("Excess arguments, ignoring"); + break; - setsid(); - make_console_stdio(); + case '?': + return -EINVAL; - log_info("Successfully changed into root pivot."); + default: + assert_not_reached("Unhandled option code."); + } + + if (!arg_verb) { + log_error("Verb argument missing."); + return -EINVAL; + } return 0; } +static int switch_root_initramfs(void) { + if (mount("/run/initramfs", "/run/initramfs", NULL, MS_BIND, NULL) < 0) + return log_error_errno(errno, "Failed to mount bind /run/initramfs on /run/initramfs: %m"); + + if (mount(NULL, "/run/initramfs", NULL, MS_PRIVATE, NULL) < 0) + return log_error_errno(errno, "Failed to make /run/initramfs private mount: %m"); + + /* switch_root with MS_BIND, because there might still be processes lurking around, which have open file descriptors. + * /run/initramfs/shutdown will take care of these. + * Also do not detach the old root, because /run/initramfs/shutdown needs to access it. + */ + return switch_root("/run/initramfs", "/oldroot", false, MS_BIND); +} + + int main(int argc, char *argv[]) { - _cleanup_free_ char *line = NULL; - int cmd, r; - unsigned retries; - bool need_umount = true, need_swapoff = true, need_loop_detach = true, need_dm_detach = true; + bool need_umount, need_swapoff, need_loop_detach, need_dm_detach; bool in_container, use_watchdog = false; + _cleanup_free_ char *cgroup = NULL; char *arguments[3]; + unsigned retries; + int cmd, r; - /* suppress shutdown status output if 'quiet' is used */ - r = read_one_line_file("/proc/cmdline", &line); - if (r >= 0) { - char *w, *state; - size_t l; + log_parse_environment(); + r = parse_argv(argc, argv); + if (r < 0) + goto error; - FOREACH_WORD_QUOTED(w, l, line, state) { - if (l == 5 && memcmp(w, "quiet", 5) == 0) { - log_set_max_level(LOG_WARNING); - break; - } - } - } + /* journald will die if not gone yet. The log target defaults + * to console, but may have been changed by command line options. */ - log_parse_environment(); - log_set_target(LOG_TARGET_CONSOLE); /* syslog will die if not gone yet */ + log_close_console(); /* force reopen of /dev/console */ log_open(); umask(0022); if (getpid() != 1) { - log_error("Not executed by init (pid 1)."); + log_error("Not executed by init (PID 1)."); r = -EPERM; goto error; } - if (argc != 2) { - log_error("Invalid number of arguments."); - r = -EINVAL; - goto error; - } - - in_container = detect_container(NULL) > 0; - - if (streq(argv[1], "reboot")) + if (streq(arg_verb, "reboot")) cmd = RB_AUTOBOOT; - else if (streq(argv[1], "poweroff")) + else if (streq(arg_verb, "poweroff")) cmd = RB_POWER_OFF; - else if (streq(argv[1], "halt")) + else if (streq(arg_verb, "halt")) cmd = RB_HALT_SYSTEM; - else if (streq(argv[1], "kexec")) + else if (streq(arg_verb, "kexec")) cmd = LINUX_REBOOT_CMD_KEXEC; else { - log_error("Unknown action '%s'.", argv[1]); r = -EINVAL; + log_error("Unknown action '%s'.", arg_verb); goto error; } + cg_get_root_path(&cgroup); + use_watchdog = !!getenv("WATCHDOG_USEC"); /* lock us into memory */ mlockall(MCL_CURRENT|MCL_FUTURE); log_info("Sending SIGTERM to remaining processes..."); - broadcast_signal(SIGTERM, true); + broadcast_signal(SIGTERM, true, true); log_info("Sending SIGKILL to remaining processes..."); - broadcast_signal(SIGKILL, true); + broadcast_signal(SIGKILL, true, false); - if (in_container) { - need_swapoff = false; - need_dm_detach = false; - need_loop_detach = false; - } + in_container = detect_container(NULL) > 0; + + need_umount = !in_container; + need_swapoff = !in_container; + need_loop_detach = !in_container; + need_dm_detach = !in_container; /* Unmount all mountpoints, swaps, and loopback devices */ for (retries = 0; retries < FINALIZE_ATTEMPTS; retries++) { @@ -210,6 +220,13 @@ int main(int argc, char *argv[]) { if (use_watchdog) watchdog_ping(); + /* Let's trim the cgroup tree on each iteration so + that we leave an empty cgroup tree around, so that + container managers get a nice notify event when we + are down */ + if (cgroup) + cg_trim(SYSTEMD_CGROUP_CONTROLLER, cgroup, false); + if (need_umount) { log_info("Unmounting file systems."); r = umount_all(&changed); @@ -219,7 +236,7 @@ int main(int argc, char *argv[]) { } else if (r > 0) log_info("Not all file systems unmounted, %d left.", r); else - log_error("Failed to unmount file systems: %s", strerror(-r)); + log_error_errno(r, "Failed to unmount file systems: %m"); } if (need_swapoff) { @@ -231,7 +248,7 @@ int main(int argc, char *argv[]) { } else if (r > 0) log_info("Not all swaps deactivated, %d left.", r); else - log_error("Failed to deactivate swaps: %s", strerror(-r)); + log_error_errno(r, "Failed to deactivate swaps: %m"); } if (need_loop_detach) { @@ -243,7 +260,7 @@ int main(int argc, char *argv[]) { } else if (r > 0) log_info("Not all loop devices detached, %d left.", r); else - log_error("Failed to detach loop devices: %s", strerror(-r)); + log_error_errno(r, "Failed to detach loop devices: %m"); } if (need_dm_detach) { @@ -255,49 +272,70 @@ int main(int argc, char *argv[]) { } else if (r > 0) log_info("Not all DM devices detached, %d left.", r); else - log_error("Failed to detach DM devices: %s", strerror(-r)); + log_error_errno(r, "Failed to detach DM devices: %m"); } if (!need_umount && !need_swapoff && !need_loop_detach && !need_dm_detach) { if (retries > 0) log_info("All filesystems, swaps, loop devices, DM devices detached."); /* Yay, done */ - break; + goto initrd_jump; } /* If in this iteration we didn't manage to * unmount/deactivate anything, we simply give up */ if (!changed) { - log_error("Cannot finalize remaining file systems and devices, giving up."); - break; + log_info("Cannot finalize remaining%s%s%s%s continuing.", + need_umount ? " file systems," : "", + need_swapoff ? " swap devices," : "", + need_loop_detach ? " loop devices," : "", + need_dm_detach ? " DM devices," : ""); + goto initrd_jump; } - log_debug("Couldn't finalize remaining file systems and devices after %u retries, trying again.", retries+1); + log_debug("After %u retries, couldn't finalize remaining %s%s%s%s trying again.", + retries + 1, + need_umount ? " file systems," : "", + need_swapoff ? " swap devices," : "", + need_loop_detach ? " loop devices," : "", + need_dm_detach ? " DM devices," : ""); } - if (retries >= FINALIZE_ATTEMPTS) - log_error("Too many iterations, giving up."); - else - log_info("Storage is finalized."); + log_error("Too many iterations, giving up."); + + initrd_jump: arguments[0] = NULL; - arguments[1] = argv[1]; + arguments[1] = arg_verb; arguments[2] = NULL; - execute_directory(SYSTEM_SHUTDOWN_PATH, NULL, arguments); + execute_directory(SYSTEM_SHUTDOWN_PATH, DEFAULT_TIMEOUT_USEC, arguments); if (!in_container && !in_initrd() && access("/run/initramfs/shutdown", X_OK) == 0) { + r = switch_root_initramfs(); + if (r >= 0) { + argv[0] = (char*) "/shutdown"; - if (prepare_new_root() >= 0 && - pivot_to_new_root() >= 0) { + setsid(); + make_console_stdio(); - log_info("Returning to initrd..."); + log_info("Successfully changed into root pivot.\n" + "Returning to initrd..."); execv("/shutdown", argv); - log_error("Failed to execute shutdown binary: %m"); - } + log_error_errno(errno, "Failed to execute shutdown binary: %m"); + } else + log_error_errno(r, "Failed to switch root to \"/run/initramfs\": %m"); + } + if (need_umount || need_swapoff || need_loop_detach || need_dm_detach) + log_error("Failed to finalize %s%s%s%s ignoring", + need_umount ? " file systems," : "", + need_swapoff ? " swap devices," : "", + need_loop_detach ? " loop devices," : "", + need_dm_detach ? " DM devices," : ""); + /* The kernel will automaticall flush ATA disks and suchlike * on reboot(), but the file systems need to be synce'd * explicitly in advance. So let's do this here, but not @@ -305,44 +343,76 @@ int main(int argc, char *argv[]) { if (!in_container) sync(); - if (cmd == LINUX_REBOOT_CMD_KEXEC) { + switch (cmd) { + + case LINUX_REBOOT_CMD_KEXEC: if (!in_container) { /* We cheat and exec kexec to avoid doing all its work */ - pid_t pid = fork(); + pid_t pid; + + log_info("Rebooting with kexec."); + pid = fork(); if (pid < 0) - log_error("Could not fork: %m. Falling back to normal reboot."); - else if (pid > 0) { - wait_for_terminate_and_warn("kexec", pid); - log_warning("kexec failed. Falling back to normal reboot."); - } else { + log_error_errno(errno, "Failed to fork: %m"); + else if (pid == 0) { + + const char * const args[] = { + KEXEC, "-e", NULL + }; + /* Child */ - const char *args[3] = { KEXEC, "-e", NULL }; + execv(args[0], (char * const *) args); - return EXIT_FAILURE; - } + _exit(EXIT_FAILURE); + } else + wait_for_terminate_and_warn("kexec", pid, true); } cmd = RB_AUTOBOOT; + /* Fall through */ + + case RB_AUTOBOOT: + + if (!in_container) { + _cleanup_free_ char *param = NULL; + + if (read_one_line_file(REBOOT_PARAM_FILE, ¶m) >= 0) { + log_info("Rebooting with argument '%s'.", param); + syscall(SYS_reboot, LINUX_REBOOT_MAGIC1, LINUX_REBOOT_MAGIC2, LINUX_REBOOT_CMD_RESTART2, param); + } + } + + log_info("Rebooting."); + break; + + case RB_POWER_OFF: + log_info("Powering off."); + break; + + case RB_HALT_SYSTEM: + log_info("Halting system."); + break; + + default: + assert_not_reached("Unknown magic"); } reboot(cmd); - if (errno == EPERM && in_container) { /* If we are in a container, and we lacked * CAP_SYS_BOOT just exit, this will kill our * container for good. */ - log_error("Exiting container."); + log_info("Exiting container."); exit(0); } - log_error("Failed to invoke reboot(): %m"); + log_error_errno(errno, "Failed to invoke reboot(): %m"); r = -errno; error: - log_error("Critical error while doing system shutdown: %s", strerror(-r)); + log_emergency_errno(r, "Critical error while doing system shutdown: %m"); freeze(); - return EXIT_FAILURE; }