diff options
author | Kay Sievers <kay@vrfy.org> | 2012-04-12 17:15:18 +0200 |
---|---|---|
committer | Kay Sievers <kay@vrfy.org> | 2012-04-12 17:54:42 +0200 |
commit | dce818b390a857a11f7dd634684500675cf79833 (patch) | |
tree | e71acb0f2e50f5a6b77c8af90af9a973c8cea6cb /src/shutdownd | |
parent | 4fc5568804ef50d9cc857eee86012d388620edd3 (diff) | |
download | systemd-dce818b390a857a11f7dd634684500675cf79833.tar.gz |
move all tools to subdirs
Diffstat (limited to 'src/shutdownd')
-rw-r--r-- | src/shutdownd/shutdownd.c | 476 |
1 files changed, 476 insertions, 0 deletions
diff --git a/src/shutdownd/shutdownd.c b/src/shutdownd/shutdownd.c new file mode 100644 index 0000000000..0497cd41a0 --- /dev/null +++ b/src/shutdownd/shutdownd.c @@ -0,0 +1,476 @@ +/*-*- Mode: C; c-basic-offset: 8; indent-tabs-mode: nil -*-*/ + +/*** + This file is part of systemd. + + Copyright 2010 Lennart Poettering + + systemd is free software; you can redistribute it and/or modify it + under the terms of the GNU Lesser General Public License as published by + the Free Software Foundation; either version 2.1 of the License, or + (at your option) any later version. + + systemd is distributed in the hope that it will be useful, but + WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public License + along with systemd; If not, see <http://www.gnu.org/licenses/>. +***/ + +#include <sys/socket.h> +#include <sys/poll.h> +#include <sys/types.h> +#include <sys/timerfd.h> +#include <assert.h> +#include <string.h> +#include <errno.h> +#include <unistd.h> +#include <fcntl.h> +#include <stddef.h> + +#include <systemd/sd-daemon.h> +#include <systemd/sd-shutdown.h> + +#include "log.h" +#include "macro.h" +#include "util.h" +#include "utmp-wtmp.h" +#include "mkdir.h" + +union shutdown_buffer { + struct sd_shutdown_command command; + char space[offsetof(struct sd_shutdown_command, wall_message) + LINE_MAX]; +}; + +static int read_packet(int fd, union shutdown_buffer *_b) { + struct msghdr msghdr; + struct iovec iovec; + struct ucred *ucred; + union { + struct cmsghdr cmsghdr; + uint8_t buf[CMSG_SPACE(sizeof(struct ucred))]; + } control; + ssize_t n; + union shutdown_buffer b; /* We maintain our own copy here, in order not to corrupt the last message */ + + assert(fd >= 0); + assert(_b); + + zero(iovec); + iovec.iov_base = &b; + iovec.iov_len = sizeof(b) - 1; + + zero(control); + zero(msghdr); + msghdr.msg_iov = &iovec; + msghdr.msg_iovlen = 1; + msghdr.msg_control = &control; + msghdr.msg_controllen = sizeof(control); + + n = recvmsg(fd, &msghdr, MSG_DONTWAIT); + if (n <= 0) { + if (n == 0) { + log_error("Short read"); + return -EIO; + } + + if (errno == EAGAIN || errno == EINTR) + return 0; + + log_error("recvmsg(): %m"); + return -errno; + } + + if (msghdr.msg_controllen < CMSG_LEN(sizeof(struct ucred)) || + control.cmsghdr.cmsg_level != SOL_SOCKET || + control.cmsghdr.cmsg_type != SCM_CREDENTIALS || + control.cmsghdr.cmsg_len != CMSG_LEN(sizeof(struct ucred))) { + log_warning("Received message without credentials. Ignoring."); + return 0; + } + + ucred = (struct ucred*) CMSG_DATA(&control.cmsghdr); + if (ucred->uid != 0) { + log_warning("Got request from unprivileged user. Ignoring."); + return 0; + } + + if ((size_t) n < offsetof(struct sd_shutdown_command, wall_message)) { + log_warning("Message has invalid size. Ignoring."); + return 0; + } + + if (b.command.mode != SD_SHUTDOWN_NONE && + b.command.mode != SD_SHUTDOWN_REBOOT && + b.command.mode != SD_SHUTDOWN_POWEROFF && + b.command.mode != SD_SHUTDOWN_HALT && + b.command.mode != SD_SHUTDOWN_KEXEC) { + log_warning("Message has invalid mode. Ignoring."); + return 0; + } + + b.space[n] = 0; + + *_b = b; + return 1; +} + +static void warn_wall(usec_t n, struct sd_shutdown_command *c) { + char date[FORMAT_TIMESTAMP_MAX]; + const char *prefix; + char *l = NULL; + + assert(c); + assert(c->warn_wall); + + if (n >= c->usec) + return; + + if (c->mode == SD_SHUTDOWN_HALT) + prefix = "The system is going down for system halt at "; + else if (c->mode == SD_SHUTDOWN_POWEROFF) + prefix = "The system is going down for power-off at "; + else if (c->mode == SD_SHUTDOWN_REBOOT) + prefix = "The system is going down for reboot at "; + else if (c->mode == SD_SHUTDOWN_KEXEC) + prefix = "The system is going down for kexec reboot at "; + else + assert_not_reached("Unknown mode!"); + + if (asprintf(&l, "%s%s%s%s!", c->wall_message, c->wall_message[0] ? "\n" : "", + prefix, format_timestamp(date, sizeof(date), c->usec)) < 0) + log_error("Failed to allocate wall message"); + else { + utmp_wall(l, NULL); + free(l); + } +} + +static usec_t when_wall(usec_t n, usec_t elapse) { + + static const struct { + usec_t delay; + usec_t interval; + } table[] = { + { 10 * USEC_PER_MINUTE, USEC_PER_MINUTE }, + { USEC_PER_HOUR, 15 * USEC_PER_MINUTE }, + { 3 * USEC_PER_HOUR, 30 * USEC_PER_MINUTE } + }; + + usec_t left, sub; + unsigned i; + + /* If the time is already passed, then don't announce */ + if (n >= elapse) + return 0; + + left = elapse - n; + for (i = 0; i < ELEMENTSOF(table); i++) + if (n + table[i].delay >= elapse) { + sub = ((left / table[i].interval) * table[i].interval); + break; + } + + if (i >= ELEMENTSOF(table)) + sub = ((left / USEC_PER_HOUR) * USEC_PER_HOUR); + + return elapse > sub ? elapse - sub : 1; +} + +static usec_t when_nologin(usec_t elapse) { + return elapse > 5*USEC_PER_MINUTE ? elapse - 5*USEC_PER_MINUTE : 1; +} + +static const char *mode_to_string(enum sd_shutdown_mode m) { + switch (m) { + case SD_SHUTDOWN_REBOOT: + return "reboot"; + case SD_SHUTDOWN_POWEROFF: + return "poweroff"; + case SD_SHUTDOWN_HALT: + return "halt"; + case SD_SHUTDOWN_KEXEC: + return "kexec"; + default: + return NULL; + } +} + +static int update_schedule_file(struct sd_shutdown_command *c) { + int r; + FILE *f; + char *temp_path, *t; + + assert(c); + + r = safe_mkdir("/run/systemd/shutdown", 0755, 0, 0); + if (r < 0) { + log_error("Failed to create shutdown subdirectory: %s", strerror(-r)); + return r; + } + + t = cescape(c->wall_message); + if (!t) { + log_error("Out of memory"); + return -ENOMEM; + } + + r = fopen_temporary("/run/systemd/shutdown/scheduled", &f, &temp_path); + if (r < 0) { + log_error("Failed to save information about scheduled shutdowns: %s", strerror(-r)); + free(t); + return r; + } + + fchmod(fileno(f), 0644); + + fprintf(f, + "USEC=%llu\n" + "WARN_WALL=%i\n" + "MODE=%s\n", + (unsigned long long) c->usec, + c->warn_wall, + mode_to_string(c->mode)); + + if (c->dry_run) + fputs("DRY_RUN=1\n", f); + + if (!isempty(t)) + fprintf(f, "WALL_MESSAGE=%s\n", t); + + free(t); + + fflush(f); + + if (ferror(f) || rename(temp_path, "/run/systemd/shutdown/scheduled") < 0) { + log_error("Failed to write information about scheduled shutdowns: %m"); + r = -errno; + + unlink(temp_path); + unlink("/run/systemd/shutdown/scheduled"); + } + + fclose(f); + free(temp_path); + + return r; +} + +static bool scheduled(struct sd_shutdown_command *c) { + return c->usec > 0 && c->mode != SD_SHUTDOWN_NONE; +} + +int main(int argc, char *argv[]) { + enum { + FD_SOCKET, + FD_WALL_TIMER, + FD_NOLOGIN_TIMER, + FD_SHUTDOWN_TIMER, + _FD_MAX + }; + + int r = EXIT_FAILURE, n_fds; + union shutdown_buffer b; + struct pollfd pollfd[_FD_MAX]; + bool exec_shutdown = false, unlink_nologin = false; + unsigned i; + + if (getppid() != 1) { + log_error("This program should be invoked by init only."); + return EXIT_FAILURE; + } + + if (argc > 1) { + log_error("This program does not take arguments."); + return EXIT_FAILURE; + } + + log_set_target(LOG_TARGET_AUTO); + log_parse_environment(); + log_open(); + + umask(0022); + + n_fds = sd_listen_fds(true); + if (n_fds < 0) { + log_error("Failed to read listening file descriptors from environment: %s", strerror(-r)); + return EXIT_FAILURE; + } + + if (n_fds != 1) { + log_error("Need exactly one file descriptor."); + return EXIT_FAILURE; + } + + zero(b); + zero(pollfd); + + pollfd[FD_SOCKET].fd = SD_LISTEN_FDS_START; + pollfd[FD_SOCKET].events = POLLIN; + + for (i = FD_WALL_TIMER; i < _FD_MAX; i++) { + pollfd[i].events = POLLIN; + pollfd[i].fd = timerfd_create(CLOCK_REALTIME, TFD_NONBLOCK|TFD_CLOEXEC); + if (pollfd[i].fd < 0) { + log_error("timerfd_create(): %m"); + goto finish; + } + } + + log_debug("systemd-shutdownd running as pid %lu", (unsigned long) getpid()); + + sd_notify(false, + "READY=1\n" + "STATUS=Processing requests..."); + + for (;;) { + int k; + usec_t n; + + k = poll(pollfd, _FD_MAX, scheduled(&b.command) ? -1 : 0); + if (k < 0) { + + if (errno == EAGAIN || errno == EINTR) + continue; + + log_error("poll(): %m"); + goto finish; + } + + /* Exit on idle */ + if (k == 0) + break; + + n = now(CLOCK_REALTIME); + + if (pollfd[FD_SOCKET].revents) { + + k = read_packet(pollfd[FD_SOCKET].fd, &b); + if (k < 0) + goto finish; + else if (k > 0) { + struct itimerspec its; + char date[FORMAT_TIMESTAMP_MAX]; + + if (!scheduled(&b.command)) { + log_info("Shutdown canceled."); + break; + } + + zero(its); + if (b.command.warn_wall) { + /* Send wall messages every so often */ + timespec_store(&its.it_value, when_wall(n, b.command.usec)); + + /* Warn immediately if less than 15 minutes are left */ + if (n < b.command.usec && + n + 15*USEC_PER_MINUTE >= b.command.usec) + warn_wall(n, &b.command); + } + if (timerfd_settime(pollfd[FD_WALL_TIMER].fd, TFD_TIMER_ABSTIME, &its, NULL) < 0) { + log_error("timerfd_settime(): %m"); + goto finish; + } + + /* Disallow logins 5 minutes prior to shutdown */ + zero(its); + timespec_store(&its.it_value, when_nologin(b.command.usec)); + if (timerfd_settime(pollfd[FD_NOLOGIN_TIMER].fd, TFD_TIMER_ABSTIME, &its, NULL) < 0) { + log_error("timerfd_settime(): %m"); + goto finish; + } + + /* Shutdown after the specified time is reached */ + zero(its); + timespec_store(&its.it_value, b.command.usec); + if (timerfd_settime(pollfd[FD_SHUTDOWN_TIMER].fd, TFD_TIMER_ABSTIME, &its, NULL) < 0) { + log_error("timerfd_settime(): %m"); + goto finish; + } + + update_schedule_file(&b.command); + + sd_notifyf(false, + "STATUS=Shutting down at %s (%s)...", + format_timestamp(date, sizeof(date), b.command.usec), + mode_to_string(b.command.mode)); + + log_info("Shutting down at %s (%s)...", date, mode_to_string(b.command.mode)); + } + } + + if (pollfd[FD_WALL_TIMER].revents) { + struct itimerspec its; + + warn_wall(n, &b.command); + flush_fd(pollfd[FD_WALL_TIMER].fd); + + /* Restart timer */ + zero(its); + timespec_store(&its.it_value, when_wall(n, b.command.usec)); + if (timerfd_settime(pollfd[FD_WALL_TIMER].fd, TFD_TIMER_ABSTIME, &its, NULL) < 0) { + log_error("timerfd_settime(): %m"); + goto finish; + } + } + + if (pollfd[FD_NOLOGIN_TIMER].revents) { + int e; + + log_info("Creating /run/nologin, blocking further logins..."); + + e = write_one_line_file_atomic("/run/nologin", "System is going down."); + if (e < 0) + log_error("Failed to create /run/nologin: %s", strerror(-e)); + else + unlink_nologin = true; + + flush_fd(pollfd[FD_NOLOGIN_TIMER].fd); + } + + if (pollfd[FD_SHUTDOWN_TIMER].revents) { + exec_shutdown = true; + goto finish; + } + } + + r = EXIT_SUCCESS; + + log_debug("systemd-shutdownd stopped as pid %lu", (unsigned long) getpid()); + +finish: + + for (i = 0; i < _FD_MAX; i++) + if (pollfd[i].fd >= 0) + close_nointr_nofail(pollfd[i].fd); + + if (unlink_nologin) + unlink("/run/nologin"); + + unlink("/run/systemd/shutdown/scheduled"); + + if (exec_shutdown && !b.command.dry_run) { + char sw[3]; + + sw[0] = '-'; + sw[1] = b.command.mode; + sw[2] = 0; + + execl(SYSTEMCTL_BINARY_PATH, + "shutdown", + sw, + "now", + (b.command.warn_wall && b.command.wall_message[0]) ? b.command.wall_message : + (b.command.warn_wall ? NULL : "--no-wall"), + NULL); + + log_error("Failed to execute /sbin/shutdown: %m"); + } + + sd_notify(false, + "STATUS=Exiting..."); + + return r; +} |