+/test-watchdog
/test-journal-send
/systemd-multi-seat-x
/systemd-cgtop
test-cgroup \
test-env-replace \
test-strv \
- test-install
+ test-install \
+ test-watchdog
dist_pkgsysconf_DATA = \
src/system.conf \
src/tcpwrap.h \
src/cgroup-attr.c \
src/cgroup-attr.h \
+ src/watchdog.c \
+ src/watchdog.h \
src/sd-daemon.c \
src/sd-id128.c \
src/macro.h \
test_install_LDADD = \
libsystemd-basic.la
+test_watchdog_SOURCES = \
+ src/test-watchdog.c \
+ src/watchdog.c \
+ src/watchdog.h
+
+test_watchdog_LDADD = \
+ libsystemd-basic.la
+
systemd_initctl_SOURCES = \
src/initctl.c \
src/dbus-common.c
systemd_shutdown_SOURCES = \
src/mount-setup.c \
src/umount.c \
- src/shutdown.c
+ src/shutdown.c \
+ src/watchdog.c \
+ src/watchdog.h
systemd_shutdown_LDADD = \
libsystemd-basic.la \
* There's currently no way to cancel fsck (used to be possible via C-c or c on the console)
-* hook up /dev/watchdog with main event loop for embedded, server uses
-
* when dumping cgroup contents, include main/control PID of a service, explicitly
* keep an eye on https://bugzilla.gnome.org/show_bug.cgi?id=670100
controllers in separate
hierarchies.</para></listitem>
</varlistentry>
+
+ <varlistentry>
+ <term><varname>RuntimeWatchdogSec=</varname></term>
+ <term><varname>ShutdownWatchdogSec=</varname></term>
+
+ <listitem><para>Configure the hardware
+ watchdog at runtime and at
+ reboot. Takes a timeout value in
+ seconds (or in other time units if
+ suffixed with <literal>ms</literal>,
+ <literal>min</literal>,
+ <literal>h</literal>,
+ <literal>d</literal>,
+ <literal>w</literal>). If
+ <varname>RuntimeWatchdogSec=</varname>
+ is set to a non-zero value the
+ watchdog hardware
+ (<filename>/dev/watchdog</filename>)
+ will be programmed to automatically
+ reboot the system if it is not
+ contacted within the specified timeout
+ interval. The system manager will
+ ensure to contact it at least once in
+ half the specified timeout
+ interval. This feature requires a
+ hardware watchdog device to be
+ present, as it is commonly the case in
+ embedded and server systems. Not all
+ hardware watchdogs allow configuration
+ of the reboot timeout, in which case
+ the closest available timeout is
+ picked. <varname>ShutdownWatchdogSec=</varname>
+ may be used to configure the hardware
+ watchdog when the system is asked to
+ reboot. It works as a safety net to
+ ensure that the reboot takes place
+ even if a clean reboot attempt times
+ out. By default
+ <varname>RuntimeWatchdogSec=</varname>
+ defaults to 0 (off), and
+ <varname>ShutdownWatchdogSec=</varname>
+ to 10min. These settings have no
+ effect if a hardware watchdog is not
+ available.</para></listitem>
+ </varlistentry>
</variablelist>
</refsect1>
" <property name=\"SwapAuto\" type=\"b\" access=\"read\"/>\n" \
" <property name=\"DefaultControllers\" type=\"as\" access=\"read\"/>\n" \
" <property name=\"DefaultStandardOutput\" type=\"s\" access=\"read\"/>\n" \
- " <property name=\"DefaultStandardError\" type=\"s\" access=\"read\"/>\n"
+ " <property name=\"DefaultStandardError\" type=\"s\" access=\"read\"/>\n" \
+ " <property name=\"RuntimeWatchdogUSec\" type=\"s\" access=\"read\"/>\n" \
+ " <property name=\"ShutdownWatchdogUSec\" type=\"s\" access=\"read\"/>\n"
#ifdef HAVE_SYSV_COMPAT
#define BUS_MANAGER_INTERFACE_PROPERTIES_SYSV \
return r;
}
-static const char systemd_property_string[] = PACKAGE_STRING "\0" DISTRIBUTION "\0" SYSTEMD_FEATURES;
+static const char systemd_property_string[] =
+ PACKAGE_STRING "\0"
+ DISTRIBUTION "\0"
+ SYSTEMD_FEATURES;
static const BusProperty bus_systemd_properties[] = {
{ "Version", bus_property_append_string, "s", 0 },
static const BusProperty bus_manager_properties[] = {
{ "RunningAs", bus_manager_append_running_as, "s", offsetof(Manager, running_as) },
- { "Tainted", bus_manager_append_tainted, "s", 0 },
+ { "Tainted", bus_manager_append_tainted, "s", 0 },
{ "InitRDTimestamp", bus_property_append_uint64, "t", offsetof(Manager, initrd_timestamp.realtime) },
{ "InitRDTimestampMonotonic", bus_property_append_uint64, "t", offsetof(Manager, initrd_timestamp.monotonic) },
{ "StartupTimestamp", bus_property_append_uint64, "t", offsetof(Manager, startup_timestamp.realtime) },
{ "FinishTimestampMonotonic", bus_property_append_uint64, "t", offsetof(Manager, finish_timestamp.monotonic) },
{ "LogLevel", bus_manager_append_log_level, "s", 0, 0, bus_manager_set_log_level },
{ "LogTarget", bus_manager_append_log_target, "s", 0, 0, bus_manager_set_log_target },
- { "NNames", bus_manager_append_n_names, "u", 0 },
- { "NJobs", bus_manager_append_n_jobs, "u", 0 },
+ { "NNames", bus_manager_append_n_names, "u", 0 },
+ { "NJobs", bus_manager_append_n_jobs, "u", 0 },
{ "NInstalledJobs",bus_property_append_uint32, "u", offsetof(Manager, n_installed_jobs) },
{ "NFailedJobs", bus_property_append_uint32, "u", offsetof(Manager, n_failed_jobs) },
- { "Progress", bus_manager_append_progress, "d", 0 },
+ { "Progress", bus_manager_append_progress, "d", 0 },
{ "Environment", bus_property_append_strv, "as", offsetof(Manager, environment), true },
{ "ConfirmSpawn", bus_property_append_bool, "b", offsetof(Manager, confirm_spawn) },
{ "ShowStatus", bus_property_append_bool, "b", offsetof(Manager, show_status) },
{ "DefaultControllers", bus_property_append_strv, "as", offsetof(Manager, default_controllers), true },
{ "DefaultStandardOutput", bus_manager_append_exec_output, "s", offsetof(Manager, default_std_output) },
{ "DefaultStandardError", bus_manager_append_exec_output, "s", offsetof(Manager, default_std_error) },
+ { "RuntimeWatchdogUSec", bus_property_append_usec, "t", offsetof(Manager, runtime_watchdog), },
+ { "ShutdownWatchdogUSec", bus_property_append_usec, "t", offsetof(Manager, shutdown_watchdog), },
#ifdef HAVE_SYSV_COMPAT
{ "SysVConsole", bus_property_append_bool, "b", offsetof(Manager, sysv_console) },
{ "SysVInitPath", bus_property_append_strv, "as", offsetof(Manager, lookup_paths.sysvinit_path), true },
#include "strv.h"
#include "def.h"
#include "virt.h"
+#include "watchdog.h"
static enum {
ACTION_RUN,
static char ***arg_join_controllers = NULL;
static ExecOutput arg_default_std_output = EXEC_OUTPUT_JOURNAL;
static ExecOutput arg_default_std_error = EXEC_OUTPUT_INHERIT;
+static usec_t arg_runtime_watchdog = 0;
+static usec_t arg_shutdown_watchdog = 10 * USEC_PER_MINUTE;
static FILE* serialization = NULL;
{ "Manager", "DefaultStandardOutput", config_parse_output, 0, &arg_default_std_output },
{ "Manager", "DefaultStandardError", config_parse_output, 0, &arg_default_std_error },
{ "Manager", "JoinControllers", config_parse_join_controllers, 0, &arg_join_controllers },
+ { "Manager", "RuntimeWatchdogSec", config_parse_usec, 0, &arg_runtime_watchdog },
+ { "Manager", "ShutdownWatchdogSec", config_parse_usec, 0, &arg_shutdown_watchdog },
{ NULL, NULL, NULL, 0, NULL }
};
bool is_reexec = false;
int j;
bool loaded_policy = false;
+ bool arm_reboot_watchdog = false;
#ifdef HAVE_SYSV_COMPAT
if (getpid() != 1 && strstr(program_invocation_short_name, "init")) {
test_cgroups();
}
- if ((r = manager_new(arg_running_as, &m)) < 0) {
+ if (arg_running_as == MANAGER_SYSTEM && arg_runtime_watchdog > 0)
+ watchdog_set_timeout(&arg_runtime_watchdog);
+
+ r = manager_new(arg_running_as, &m);
+ if (r < 0) {
log_error("Failed to allocate manager object: %s", strerror(-r));
goto finish;
}
m->swap_auto = arg_swap_auto;
m->default_std_output = arg_default_std_output;
m->default_std_error = arg_default_std_error;
+ m->runtime_watchdog = arg_runtime_watchdog;
+ m->shutdown_watchdog = arg_shutdown_watchdog;
if (dual_timestamp_is_set(&initrd_timestamp))
m->initrd_timestamp = initrd_timestamp;
before_startup = now(CLOCK_MONOTONIC);
- if ((r = manager_startup(m, serialization, fds)) < 0)
+ r = manager_startup(m, serialization, fds);
+ if (r < 0)
log_error("Failed to fully start up daemon: %s", strerror(-r));
if (fds) {
log_debug("Activating default unit: %s", arg_default_unit);
- if ((r = manager_load_unit(m, arg_default_unit, NULL, &error, &target)) < 0) {
+ r = manager_load_unit(m, arg_default_unit, NULL, &error, &target);
+ if (r < 0) {
log_error("Failed to load default target: %s", bus_error(&error, r));
dbus_error_free(&error);
} else if (target->load_state == UNIT_ERROR)
if (!target || target->load_state != UNIT_LOADED) {
log_info("Trying to load rescue target...");
- if ((r = manager_load_unit(m, SPECIAL_RESCUE_TARGET, NULL, &error, &target)) < 0) {
+ r = manager_load_unit(m, SPECIAL_RESCUE_TARGET, NULL, &error, &target);
+ if (r < 0) {
log_error("Failed to load rescue target: %s", bus_error(&error, r));
dbus_error_free(&error);
goto finish;
}
for (;;) {
- if ((r = manager_loop(m)) < 0) {
+ r = manager_loop(m);
+ if (r < 0) {
log_error("Failed to run mainloop: %s", strerror(-r));
goto finish;
}
case MANAGER_RELOAD:
log_info("Reloading.");
- if ((r = manager_reload(m)) < 0)
+ r = manager_reload(m);
+ if (r < 0)
log_error("Failed to reload: %s", strerror(-r));
break;
};
assert_se(shutdown_verb = table[m->exit_code]);
+ arm_reboot_watchdog = m->exit_code == MANAGER_REBOOT;
log_notice("Shutting down.");
goto finish;
fdset_free(fds);
if (shutdown_verb) {
+ char e[32];
+
const char * command_line[] = {
SYSTEMD_SHUTDOWN_BINARY_PATH,
shutdown_verb,
NULL
};
+ const char * env_block[] = {
+ NULL,
+ NULL
+ };
- execv(SYSTEMD_SHUTDOWN_BINARY_PATH, (char **) command_line);
+ if (arm_reboot_watchdog && arg_shutdown_watchdog > 0) {
+ /* If we reboot let's set the shutdown
+ * watchdog and tell the shutdown binary to
+ * repeatedly ping it */
+ watchdog_set_timeout(&arg_shutdown_watchdog);
+ watchdog_close(false);
+
+ /* Tell the binary how often to ping */
+ snprintf(e, sizeof(e), "WATCHDOG_USEC=%llu", (unsigned long long) arg_shutdown_watchdog);
+ char_array_0(e);
+ env_block[0] = e;
+ } else
+ watchdog_close(true);
+
+ execve(SYSTEMD_SHUTDOWN_BINARY_PATH, (char **) command_line, (char**) env_block);
log_error("Failed to execute shutdown binary, freezing: %m");
}
#include "bus-errors.h"
#include "exit-status.h"
#include "virt.h"
+#include "watchdog.h"
/* As soon as 16 units are in our GC queue, make sure to run a gc sweep */
#define GC_QUEUE_ENTRIES_MAX 16
int manager_loop(Manager *m) {
int r;
+ int wait_msec = -1;
RATELIMIT_DEFINE(rl, 1*USEC_PER_SEC, 50000);
/* There might still be some zombies hanging around from
* before we were exec()'ed. Leat's reap them */
- if ((r = manager_dispatch_sigchld(m)) < 0)
+ r = manager_dispatch_sigchld(m);
+ if (r < 0)
return r;
+ /* Sleep for half the watchdog time */
+ if (m->runtime_watchdog > 0 && m->running_as == MANAGER_SYSTEM) {
+ wait_msec = (int) (m->runtime_watchdog / 2 / USEC_PER_MSEC);
+ if (wait_msec <= 0)
+ wait_msec = 1;
+ }
+
while (m->exit_code == MANAGER_RUNNING) {
struct epoll_event event;
int n;
+ if (wait_msec >= 0)
+ watchdog_ping();
+
if (!ratelimit_test(&rl)) {
/* Yay, something is going seriously wrong, pause a little */
log_warning("Looping too fast. Throttling execution a little.");
sleep(1);
+ continue;
}
if (manager_dispatch_load_queue(m) > 0)
if (swap_dispatch_reload(m) > 0)
continue;
- if ((n = epoll_wait(m->epoll_fd, &event, 1, -1)) < 0) {
+ n = epoll_wait(m->epoll_fd, &event, 1, wait_msec);
+ if (n < 0) {
if (errno == EINTR)
continue;
return -errno;
- }
+ } else if (n == 0)
+ continue;
assert(n == 1);
- if ((r = process_event(m, &event)) < 0)
+ r = process_event(m, &event);
+ if (r < 0)
return r;
}
char **environment;
char **default_controllers;
+ usec_t runtime_watchdog;
+ usec_t shutdown_watchdog;
+
dual_timestamp initrd_timestamp;
dual_timestamp startup_timestamp;
dual_timestamp finish_timestamp;
#include "umount.h"
#include "util.h"
#include "virt.h"
+#include "watchdog.h"
#define TIMEOUT_USEC (5 * USEC_PER_SEC)
#define FINALIZE_ATTEMPTS 50
int cmd, r;
unsigned retries;
bool need_umount = true, need_swapoff = true, need_loop_detach = true, need_dm_detach = true;
- bool killed_everbody = false, in_container;
+ bool killed_everbody = false, in_container, use_watchdog = false;
log_parse_environment();
log_set_target(LOG_TARGET_CONSOLE); /* syslog will die if not gone yet */
goto error;
}
+ use_watchdog = !!getenv("WATCHDOG_USEC");
+
/* lock us into memory */
if (mlockall(MCL_CURRENT|MCL_FUTURE) != 0)
log_warning("Cannot lock process memory: %m");
for (retries = 0; retries < FINALIZE_ATTEMPTS; retries++) {
bool changed = false;
+ if (use_watchdog)
+ watchdog_ping();
+
if (need_umount) {
log_info("Unmounting file systems.");
r = umount_all(&changed);
#DefaultStandardOutput=journal
#DefaultStandardError=inherit
#JoinControllers=cpu,cpuacct
+#RuntimeWatchdog=0
+#ShutdownWatchdog=10min
--- /dev/null
+/*-*- Mode: C; c-basic-offset: 8; indent-tabs-mode: nil -*-*/
+
+/***
+ This file is part of systemd.
+
+ Copyright 2012 Lennart Poettering
+
+ systemd is free software; you can redistribute it and/or modify it
+ under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2 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
+ General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with systemd; If not, see <http://www.gnu.org/licenses/>.
+***/
+
+#include <unistd.h>
+#include <string.h>
+
+#include "watchdog.h"
+#include "log.h"
+
+int main(int argc, char *argv[]) {
+ usec_t t = 10 * USEC_PER_SEC;
+ unsigned i;
+ int r;
+
+ log_set_max_level(LOG_DEBUG);
+ log_parse_environment();
+
+ r = watchdog_set_timeout(&t);
+ if (r < 0)
+ log_warning("Failed to open watchdog: %s", strerror(-r));
+
+ for (i = 0; i < 5; i++) {
+ log_info("Pinging...");
+ r = watchdog_ping();
+ if (r < 0)
+ log_warning("Failed to ping watchdog: %s", strerror(-r));
+
+ usleep(t/2);
+ }
+
+ watchdog_close(true);
+ return 0;
+}
--- /dev/null
+/*-*- Mode: C; c-basic-offset: 8; indent-tabs-mode: nil -*-*/
+
+/***
+ This file is part of systemd.
+
+ Copyright 2012 Lennart Poettering
+
+ systemd is free software; you can redistribute it and/or modify it
+ under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2 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
+ General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with systemd; If not, see <http://www.gnu.org/licenses/>.
+***/
+
+#include <sys/ioctl.h>
+#include <errno.h>
+#include <fcntl.h>
+#include <unistd.h>
+#include <linux/watchdog.h>
+
+#include "watchdog.h"
+#include "log.h"
+
+static int watchdog_fd = -1;
+static usec_t watchdog_timeout = (usec_t) -1;
+
+static int update_timeout(void) {
+ int r;
+
+ if (watchdog_fd < 0)
+ return 0;
+
+ if (watchdog_timeout == (usec_t) -1)
+ return 0;
+ else if (watchdog_timeout == 0) {
+ int flags;
+
+ flags = WDIOS_DISABLECARD;
+ r = ioctl(watchdog_fd, WDIOC_SETOPTIONS, &flags);
+ if (r < 0) {
+ log_warning("Failed to disable hardware watchdog: %m");
+ return -errno;
+ }
+ } else {
+ int sec, flags;
+ char buf[FORMAT_TIMESPAN_MAX];
+
+ sec = (int) ((watchdog_timeout + USEC_PER_SEC - 1) / USEC_PER_SEC);
+ r = ioctl(watchdog_fd, WDIOC_SETTIMEOUT, &sec);
+ if (r < 0) {
+ log_warning("Failed to set timeout to %is: %m", sec);
+ return -errno;
+ }
+
+ watchdog_timeout = (usec_t) sec * USEC_PER_SEC;
+ log_info("Set hardware watchdog to %s.", format_timespan(buf, sizeof(buf), watchdog_timeout));
+
+ flags = WDIOS_ENABLECARD;
+ r = ioctl(watchdog_fd, WDIOC_SETOPTIONS, &flags);
+ if (r < 0) {
+ log_warning("Failed to enable hardware watchdog: %m");
+ return -errno;
+ }
+
+ r = ioctl(watchdog_fd, WDIOC_KEEPALIVE, 0);
+ if (r < 0) {
+ log_warning("Failed to ping hardware watchdog: %m");
+ return -errno;
+ }
+ }
+
+ return 0;
+}
+
+static int open_watchdog(void) {
+ struct watchdog_info ident;
+
+ if (watchdog_fd >= 0)
+ return 0;
+
+ watchdog_fd = open("/dev/watchdog", O_WRONLY|O_CLOEXEC);
+ if (watchdog_fd < 0)
+ return -errno;
+
+ if (ioctl(watchdog_fd, WDIOC_GETSUPPORT, &ident) >= 0)
+ log_info("Hardware watchdog '%s', version %x",
+ ident.identity,
+ ident.firmware_version);
+
+ return update_timeout();
+}
+
+int watchdog_set_timeout(usec_t *usec) {
+
+ watchdog_timeout = *usec;
+
+ /* If we didn't open the watchdog yet and didn't get any
+ * explicit timeout value set, don't do anything */
+ if (watchdog_fd < 0 && watchdog_timeout == (usec_t) -1)
+ return 0;
+
+ if (watchdog_fd < 0)
+ return open_watchdog();
+ else
+ return update_timeout();
+
+ *usec = watchdog_timeout;
+}
+
+int watchdog_ping(void) {
+ int r;
+
+ if (watchdog_fd < 0) {
+ r = open_watchdog();
+ if (r < 0)
+ return r;
+ }
+
+ r = ioctl(watchdog_fd, WDIOC_KEEPALIVE, 0);
+ if (r < 0) {
+ log_warning("Failed to ping hardware watchdog: %m");
+ return -errno;
+ }
+
+ return 0;
+}
+
+void watchdog_close(bool disarm) {
+ int r;
+
+ if (watchdog_fd < 0)
+ return;
+
+ if (disarm) {
+ int flags;
+
+ /* Explicitly disarm it */
+ flags = WDIOS_DISABLECARD;
+ r = ioctl(watchdog_fd, WDIOC_SETOPTIONS, &flags);
+ if (r < 0)
+ log_warning("Failed to disable hardware watchdog: %m");
+
+ /* To be sure, use magic close logic, too */
+ for (;;) {
+ static const char v = 'V';
+
+ if (write(watchdog_fd, &v, 1) > 0)
+ break;
+
+ if (errno != EINTR) {
+ log_error("Failed to disarm watchdog timer: %m");
+ break;
+ }
+ }
+ }
+
+ close_nointr_nofail(watchdog_fd);
+ watchdog_fd = -1;
+}
--- /dev/null
+/*-*- Mode: C; c-basic-offset: 8; indent-tabs-mode: nil -*-*/
+
+#ifndef foowatchdoghfoo
+#define foowatchdoghfoo
+
+/***
+ This file is part of systemd.
+
+ Copyright 2012 Lennart Poettering
+
+ systemd is free software; you can redistribute it and/or modify it
+ under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2 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
+ General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with systemd; If not, see <http://www.gnu.org/licenses/>.
+***/
+
+#include "util.h"
+
+int watchdog_set_timeout(usec_t *usec);
+int watchdog_ping(void);
+void watchdog_close(bool disarm);
+
+#endif
projects / systemd / commitdiff