Systemd/src/core/service.c

3980 lines
144 KiB
C
Raw Normal View History

/* SPDX-License-Identifier: LGPL-2.1+ */
/***
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 <errno.h>
#include <signal.h>
2010-02-14 01:09:01 +01:00
#include <unistd.h>
#include "sd-messages.h"
#include "alloc-util.h"
#include "async.h"
#include "bus-error.h"
#include "bus-kernel.h"
#include "bus-util.h"
2010-04-18 03:08:16 +02:00
#include "dbus-service.h"
#include "def.h"
#include "env-util.h"
#include "escape.h"
#include "exit-status.h"
#include "fd-util.h"
#include "fileio.h"
#include "format-util.h"
#include "fs-util.h"
#include "load-dropin.h"
#include "load-fragment.h"
#include "log.h"
#include "manager.h"
#include "parse-util.h"
#include "path-util.h"
2015-04-10 19:10:00 +02:00
#include "process-util.h"
#include "service.h"
#include "signal-util.h"
#include "special.h"
service: serialize information about currently executing command Stored information will help us to resume execution after the daemon-reload. This commit implements following scheme, * On serialization: - we count rank of the currently executing command - we store command type, its rank and command line arguments * On deserialization: - configuration is parsed and loaded - we deserialize stored data, command type, rank and arguments - we look at the given rank in the list and if command there has same arguments then we restore execution at that point - otherwise we search respective command list and we look for command that has the same arguments - if both methods fail we do not do not resume execution at all To better illustrate how does above scheme works, please consider following cases (<<< denotes position where we resume execution after reload) ; Original unit file [Service] ExecStart=/bin/true <<< ExecStart=/bin/false ; Swapped commands ; Second command is not going to be executed [Service] ExecStart=/bin/false ExecStart=/bin/true <<< ; Commands added before ; Same commands are problematic and execution could be restarted at wrong place [Service] ExecStart=/bin/foo ExecStart=/bin/bar ExecStart=/bin/true <<< ExecStart=/bin/false ; Commands added after ; Same commands are not an issue in this case [Service] ExecStart=/bin/true <<< ExecStart=/bin/false ExecStart=/bin/foo ExecStart=/bin/bar ; New commands interleaved with old commands ; Some new commands will be executed while others won't ExecStart=/bin/foo ExecStart=/bin/true <<< ExecStart=/bin/bar ExecStart=/bin/false As you can see, above scheme has some drawbacks. However, in most cases (we assume that in most common case unit file command list is not changed while some other command is running for the same unit) it should cause that systemd does the right thing, which is restoring execution exactly at the point we were before daemon-reload. Fixes #518
2017-01-23 17:12:35 +01:00
#include "stdio-util.h"
#include "string-table.h"
#include "string-util.h"
#include "strv.h"
#include "unit-name.h"
#include "unit.h"
#include "utf8.h"
#include "util.h"
2010-01-27 04:31:52 +01:00
static const UnitActiveState state_translation_table[_SERVICE_STATE_MAX] = {
2010-01-26 21:39:06 +01:00
[SERVICE_DEAD] = UNIT_INACTIVE,
[SERVICE_START_PRE] = UNIT_ACTIVATING,
[SERVICE_START] = UNIT_ACTIVATING,
[SERVICE_START_POST] = UNIT_ACTIVATING,
[SERVICE_RUNNING] = UNIT_ACTIVE,
[SERVICE_EXITED] = UNIT_ACTIVE,
[SERVICE_RELOAD] = UNIT_RELOADING,
2010-01-26 21:39:06 +01:00
[SERVICE_STOP] = UNIT_DEACTIVATING,
[SERVICE_STOP_SIGABRT] = UNIT_DEACTIVATING,
2010-01-26 21:39:06 +01:00
[SERVICE_STOP_SIGTERM] = UNIT_DEACTIVATING,
[SERVICE_STOP_SIGKILL] = UNIT_DEACTIVATING,
[SERVICE_STOP_POST] = UNIT_DEACTIVATING,
[SERVICE_FINAL_SIGTERM] = UNIT_DEACTIVATING,
[SERVICE_FINAL_SIGKILL] = UNIT_DEACTIVATING,
[SERVICE_FAILED] = UNIT_FAILED,
[SERVICE_AUTO_RESTART] = UNIT_ACTIVATING
};
/* For Type=idle we never want to delay any other jobs, hence we
* consider idle jobs active as soon as we start working on them */
static const UnitActiveState state_translation_table_idle[_SERVICE_STATE_MAX] = {
[SERVICE_DEAD] = UNIT_INACTIVE,
[SERVICE_START_PRE] = UNIT_ACTIVE,
[SERVICE_START] = UNIT_ACTIVE,
[SERVICE_START_POST] = UNIT_ACTIVE,
[SERVICE_RUNNING] = UNIT_ACTIVE,
[SERVICE_EXITED] = UNIT_ACTIVE,
[SERVICE_RELOAD] = UNIT_RELOADING,
[SERVICE_STOP] = UNIT_DEACTIVATING,
[SERVICE_STOP_SIGABRT] = UNIT_DEACTIVATING,
[SERVICE_STOP_SIGTERM] = UNIT_DEACTIVATING,
[SERVICE_STOP_SIGKILL] = UNIT_DEACTIVATING,
[SERVICE_STOP_POST] = UNIT_DEACTIVATING,
[SERVICE_FINAL_SIGTERM] = UNIT_DEACTIVATING,
[SERVICE_FINAL_SIGKILL] = UNIT_DEACTIVATING,
[SERVICE_FAILED] = UNIT_FAILED,
[SERVICE_AUTO_RESTART] = UNIT_ACTIVATING
};
static int service_dispatch_io(sd_event_source *source, int fd, uint32_t events, void *userdata);
static int service_dispatch_timer(sd_event_source *source, usec_t usec, void *userdata);
static int service_dispatch_watchdog(sd_event_source *source, usec_t usec, void *userdata);
2013-12-11 20:49:43 +01:00
static void service_enter_signal(Service *s, ServiceState state, ServiceResult f);
static void service_enter_reload_by_notify(Service *s);
2013-12-11 20:49:43 +01:00
2010-04-21 03:27:44 +02:00
static void service_init(Unit *u) {
Service *s = SERVICE(u);
assert(u);
assert(u->load_state == UNIT_STUB);
2010-04-21 03:27:44 +02:00
s->timeout_start_usec = u->manager->default_timeout_start_usec;
s->timeout_stop_usec = u->manager->default_timeout_stop_usec;
s->restart_usec = u->manager->default_restart_usec;
s->runtime_max_usec = USEC_INFINITY;
s->type = _SERVICE_TYPE_INVALID;
2010-04-21 03:27:44 +02:00
s->socket_fd = -1;
s->stdin_fd = s->stdout_fd = s->stderr_fd = -1;
s->guess_main_pid = true;
2010-04-21 03:27:44 +02:00
s->control_command_id = _SERVICE_EXEC_COMMAND_INVALID;
s->exec_context.keyring_mode = MANAGER_IS_SYSTEM(u->manager) ?
EXEC_KEYRING_PRIVATE : EXEC_KEYRING_INHERIT;
2010-04-21 03:27:44 +02:00
}
static void service_unwatch_control_pid(Service *s) {
assert(s);
if (s->control_pid <= 0)
return;
unit_unwatch_pid(UNIT(s), s->control_pid);
s->control_pid = 0;
}
static void service_unwatch_main_pid(Service *s) {
assert(s);
if (s->main_pid <= 0)
return;
unit_unwatch_pid(UNIT(s), s->main_pid);
s->main_pid = 0;
}
static void service_unwatch_pid_file(Service *s) {
if (!s->pid_file_pathspec)
return;
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_debug(UNIT(s), "Stopping watch for PID file %s", s->pid_file_pathspec->path);
path_spec_unwatch(s->pid_file_pathspec);
path_spec_done(s->pid_file_pathspec);
s->pid_file_pathspec = mfree(s->pid_file_pathspec);
}
static int service_set_main_pid(Service *s, pid_t pid) {
pid_t ppid;
assert(s);
if (pid <= 1)
return -EINVAL;
if (pid == getpid_cached())
return -EINVAL;
if (s->main_pid == pid && s->main_pid_known)
return 0;
if (s->main_pid != pid) {
service_unwatch_main_pid(s);
exec_status_start(&s->main_exec_status, pid);
}
s->main_pid = pid;
s->main_pid_known = true;
if (get_process_ppid(pid, &ppid) >= 0 && ppid != getpid_cached()) {
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning(UNIT(s), "Supervising process "PID_FMT" which is not our child. We'll most likely not notice when it exits.", pid);
s->main_pid_alien = true;
} else
s->main_pid_alien = false;
return 0;
}
void service_close_socket_fd(Service *s) {
assert(s);
/* Undo the effect of service_set_socket_fd(). */
s->socket_fd = asynchronous_close(s->socket_fd);
if (UNIT_ISSET(s->accept_socket)) {
socket_connection_unref(SOCKET(UNIT_DEREF(s->accept_socket)));
unit_ref_unset(&s->accept_socket);
}
}
static void service_stop_watchdog(Service *s) {
assert(s);
s->watchdog_event_source = sd_event_source_unref(s->watchdog_event_source);
2013-12-11 20:49:43 +01:00
s->watchdog_timestamp = DUAL_TIMESTAMP_NULL;
}
static usec_t service_get_watchdog_usec(Service *s) {
assert(s);
if (s->watchdog_override_enable)
return s->watchdog_override_usec;
else
return s->watchdog_usec;
}
2013-12-11 20:49:43 +01:00
static void service_start_watchdog(Service *s) {
int r;
usec_t watchdog_usec;
assert(s);
watchdog_usec = service_get_watchdog_usec(s);
2017-10-04 16:01:32 +02:00
if (IN_SET(watchdog_usec, 0, USEC_INFINITY))
return;
if (s->watchdog_event_source) {
r = sd_event_source_set_time(s->watchdog_event_source, usec_add(s->watchdog_timestamp.monotonic, watchdog_usec));
if (r < 0) {
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning_errno(UNIT(s), r, "Failed to reset watchdog timer: %m");
return;
}
2013-12-11 20:49:43 +01:00
r = sd_event_source_set_enabled(s->watchdog_event_source, SD_EVENT_ONESHOT);
} else {
r = sd_event_add_time(
UNIT(s)->manager->event,
&s->watchdog_event_source,
CLOCK_MONOTONIC,
usec_add(s->watchdog_timestamp.monotonic, watchdog_usec), 0,
service_dispatch_watchdog, s);
if (r < 0) {
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning_errno(UNIT(s), r, "Failed to add watchdog timer: %m");
return;
}
2015-04-29 16:05:32 +02:00
(void) sd_event_source_set_description(s->watchdog_event_source, "service-watchdog");
/* Let's process everything else which might be a sign
* of living before we consider a service died. */
r = sd_event_source_set_priority(s->watchdog_event_source, SD_EVENT_PRIORITY_IDLE);
}
if (r < 0)
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning_errno(UNIT(s), r, "Failed to install watchdog timer: %m");
}
static void service_extend_timeout(Service *s, usec_t extend_timeout_usec) {
assert(s);
if (s->timer_event_source) {
uint64_t current = 0, extended = 0;
int r;
if (IN_SET(extend_timeout_usec, 0, USEC_INFINITY))
return;
extended = usec_add(now(CLOCK_MONOTONIC), extend_timeout_usec);
r = sd_event_source_get_time(s->timer_event_source, &current);
if (r < 0)
log_unit_error_errno(UNIT(s), r, "Failed to retrieve timeout timer: %m");
else if (extended > current) {
r = sd_event_source_set_time(s->timer_event_source, extended);
if (r < 0)
log_unit_warning_errno(UNIT(s), r, "Failed to set timeout timer: %m");
}
if (s->watchdog_event_source) {
/* extend watchdog if necessary. We've asked for an extended timeout so we
* shouldn't expect a watchdog timeout in the interval in between */
r = sd_event_source_get_time(s->watchdog_event_source, &current);
if (r < 0) {
log_unit_error_errno(UNIT(s), r, "Failed to retrieve watchdog timer: %m");
return;
}
if (extended > current) {
r = sd_event_source_set_time(s->watchdog_event_source, extended);
if (r < 0)
log_unit_warning_errno(UNIT(s), r, "Failed to set watchdog timer: %m");
}
}
}
}
static void service_reset_watchdog(Service *s) {
assert(s);
dual_timestamp_get(&s->watchdog_timestamp);
2013-12-11 20:49:43 +01:00
service_start_watchdog(s);
}
static void service_reset_watchdog_timeout(Service *s, usec_t watchdog_override_usec) {
assert(s);
s->watchdog_override_enable = true;
s->watchdog_override_usec = watchdog_override_usec;
service_reset_watchdog(s);
log_unit_debug(UNIT(s), "watchdog_usec="USEC_FMT, s->watchdog_usec);
log_unit_debug(UNIT(s), "watchdog_override_usec="USEC_FMT, s->watchdog_override_usec);
}
static void service_fd_store_unlink(ServiceFDStore *fs) {
if (!fs)
return;
if (fs->service) {
assert(fs->service->n_fd_store > 0);
LIST_REMOVE(fd_store, fs->service->fd_store, fs);
fs->service->n_fd_store--;
}
if (fs->event_source) {
sd_event_source_set_enabled(fs->event_source, SD_EVENT_OFF);
sd_event_source_unref(fs->event_source);
}
free(fs->fdname);
safe_close(fs->fd);
free(fs);
}
static void service_release_fd_store(Service *s) {
assert(s);
if (s->n_keep_fd_store > 0)
return;
log_unit_debug(UNIT(s), "Releasing all stored fds");
while (s->fd_store)
service_fd_store_unlink(s->fd_store);
assert(s->n_fd_store == 0);
}
static void service_release_resources(Unit *u) {
Service *s = SERVICE(u);
assert(s);
if (!s->fd_store && s->stdin_fd < 0 && s->stdout_fd < 0 && s->stderr_fd < 0)
return;
log_unit_debug(u, "Releasing resources.");
s->stdin_fd = safe_close(s->stdin_fd);
s->stdout_fd = safe_close(s->stdout_fd);
s->stderr_fd = safe_close(s->stderr_fd);
service_release_fd_store(s);
}
2010-01-26 21:39:06 +01:00
static void service_done(Unit *u) {
Service *s = SERVICE(u);
2010-01-26 07:02:51 +01:00
assert(s);
s->pid_file = mfree(s->pid_file);
s->status_text = mfree(s->status_text);
s->exec_runtime = exec_runtime_unref(s->exec_runtime, false);
exec_command_free_array(s->exec_command, _SERVICE_EXEC_COMMAND_MAX);
2010-01-26 07:02:51 +01:00
s->control_command = NULL;
s->main_command = NULL;
2010-01-26 07:02:51 +01:00
dynamic_creds_unref(&s->dynamic_creds);
exit_status_set_free(&s->restart_prevent_status);
exit_status_set_free(&s->restart_force_status);
exit_status_set_free(&s->success_status);
2010-01-26 07:02:51 +01:00
/* This will leak a process, but at least no memory or any of
* our resources */
service_unwatch_main_pid(s);
service_unwatch_control_pid(s);
service_unwatch_pid_file(s);
2010-01-26 07:02:51 +01:00
if (s->bus_name) {
unit_unwatch_bus_name(u, s->bus_name);
s->bus_name = mfree(s->bus_name);
}
s->bus_name_owner = mfree(s->bus_name_owner);
s->usb_function_descriptors = mfree(s->usb_function_descriptors);
s->usb_function_strings = mfree(s->usb_function_strings);
service_close_socket_fd(s);
s->peer = socket_peer_unref(s->peer);
unit_ref_unset(&s->accept_socket);
service_stop_watchdog(s);
s->timer_event_source = sd_event_source_unref(s->timer_event_source);
service_release_resources(u);
}
static int on_fd_store_io(sd_event_source *e, int fd, uint32_t revents, void *userdata) {
ServiceFDStore *fs = userdata;
assert(e);
assert(fs);
/* If we get either EPOLLHUP or EPOLLERR, it's time to remove this entry from the fd store */
log_unit_debug(UNIT(fs->service),
"Received %s on stored fd %d (%s), closing.",
revents & EPOLLERR ? "EPOLLERR" : "EPOLLHUP",
fs->fd, strna(fs->fdname));
service_fd_store_unlink(fs);
return 0;
}
static int service_add_fd_store(Service *s, int fd, const char *name) {
ServiceFDStore *fs;
int r;
/* fd is always consumed if we return >= 0 */
assert(s);
assert(fd >= 0);
if (s->n_fd_store >= s->n_fd_store_max)
return -EXFULL; /* Our store is full.
* Use this errno rather than E[NM]FILE to distinguish from
* the case where systemd itself hits the file limit. */
LIST_FOREACH(fd_store, fs, s->fd_store) {
r = same_fd(fs->fd, fd);
if (r < 0)
return r;
if (r > 0) {
safe_close(fd);
return 0; /* fd already included */
}
}
fs = new0(ServiceFDStore, 1);
if (!fs)
return -ENOMEM;
fs->fd = fd;
fs->service = s;
fs->fdname = strdup(name ?: "stored");
if (!fs->fdname) {
free(fs);
return -ENOMEM;
}
r = sd_event_add_io(UNIT(s)->manager->event, &fs->event_source, fd, 0, on_fd_store_io, fs);
if (r < 0 && r != -EPERM) { /* EPERM indicates fds that aren't pollable, which is OK */
free(fs->fdname);
free(fs);
return r;
} else if (r >= 0)
(void) sd_event_source_set_description(fs->event_source, "service-fd-store");
2015-04-29 16:05:32 +02:00
LIST_PREPEND(fd_store, s->fd_store, fs);
s->n_fd_store++;
return 1; /* fd newly stored */
}
static int service_add_fd_store_set(Service *s, FDSet *fds, const char *name) {
int r;
assert(s);
while (fdset_size(fds) > 0) {
_cleanup_close_ int fd = -1;
fd = fdset_steal_first(fds);
if (fd < 0)
break;
r = service_add_fd_store(s, fd, name);
if (r == -EXFULL)
return log_unit_warning_errno(UNIT(s), r,
"Cannot store more fds than FileDescriptorStoreMax=%u, closing remaining.",
s->n_fd_store_max);
if (r < 0)
return log_unit_error_errno(UNIT(s), r, "Failed to add fd to store: %m");
if (r > 0)
log_unit_debug(UNIT(s), "Added fd %u (%s) to fd store.", fd, strna(name));
fd = -1;
}
return 0;
}
static void service_remove_fd_store(Service *s, const char *name) {
ServiceFDStore *fs, *n;
assert(s);
assert(name);
LIST_FOREACH_SAFE(fd_store, fs, n, s->fd_store) {
if (!streq(fs->fdname, name))
continue;
log_unit_debug(UNIT(s), "Got explicit request to remove fd %i (%s), closing.", fs->fd, name);
service_fd_store_unlink(fs);
}
}
static int service_arm_timer(Service *s, usec_t usec) {
int r;
assert(s);
if (s->timer_event_source) {
r = sd_event_source_set_time(s->timer_event_source, usec);
if (r < 0)
return r;
return sd_event_source_set_enabled(s->timer_event_source, SD_EVENT_ONESHOT);
}
if (usec == USEC_INFINITY)
return 0;
2015-04-29 16:05:32 +02:00
r = sd_event_add_time(
UNIT(s)->manager->event,
&s->timer_event_source,
CLOCK_MONOTONIC,
usec, 0,
service_dispatch_timer, s);
2015-04-29 16:05:32 +02:00
if (r < 0)
return r;
(void) sd_event_source_set_description(s->timer_event_source, "service-timer");
return 0;
2010-01-26 07:02:51 +01:00
}
2010-04-13 04:00:03 +02:00
static int service_verify(Service *s) {
assert(s);
if (UNIT(s)->load_state != UNIT_LOADED)
2010-04-13 04:00:03 +02:00
return 0;
if (!s->exec_command[SERVICE_EXEC_START] && !s->exec_command[SERVICE_EXEC_STOP]) {
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_error(UNIT(s), "Service lacks both ExecStart= and ExecStop= setting. Refusing.");
2010-04-13 04:00:03 +02:00
return -EINVAL;
}
if (s->type != SERVICE_ONESHOT && !s->exec_command[SERVICE_EXEC_START]) {
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_error(UNIT(s), "Service has no ExecStart= setting, which is only allowed for Type=oneshot services. Refusing.");
return -EINVAL;
}
if (!s->remain_after_exit && !s->exec_command[SERVICE_EXEC_START]) {
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_error(UNIT(s), "Service has no ExecStart= setting, which is only allowed for RemainAfterExit=yes services. Refusing.");
return -EINVAL;
}
if (s->type != SERVICE_ONESHOT && s->exec_command[SERVICE_EXEC_START]->command_next) {
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_error(UNIT(s), "Service has more than one ExecStart= setting, which is only allowed for Type=oneshot services. Refusing.");
return -EINVAL;
}
if (s->type == SERVICE_ONESHOT && s->restart != SERVICE_RESTART_NO) {
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_error(UNIT(s), "Service has Restart= setting other than no, which isn't allowed for Type=oneshot services. Refusing.");
return -EINVAL;
}
if (s->type == SERVICE_ONESHOT && !exit_status_set_is_empty(&s->restart_force_status)) {
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_error(UNIT(s), "Service has RestartForceStatus= set, which isn't allowed for Type=oneshot services. Refusing.");
return -EINVAL;
}
if (s->type == SERVICE_DBUS && !s->bus_name) {
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_error(UNIT(s), "Service is of type D-Bus but no D-Bus service name has been specified. Refusing.");
return -EINVAL;
}
if (s->bus_name && s->type != SERVICE_DBUS)
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning(UNIT(s), "Service has a D-Bus service name specified, but is not of type dbus. Ignoring.");
if (s->exec_context.pam_name && !IN_SET(s->kill_context.kill_mode, KILL_CONTROL_GROUP, KILL_MIXED)) {
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_error(UNIT(s), "Service has PAM enabled. Kill mode must be set to 'control-group' or 'mixed'. Refusing.");
return -EINVAL;
}
if (s->usb_function_descriptors && !s->usb_function_strings)
log_unit_warning(UNIT(s), "Service has USBFunctionDescriptors= setting, but no USBFunctionStrings=. Ignoring.");
if (!s->usb_function_descriptors && s->usb_function_strings)
log_unit_warning(UNIT(s), "Service has USBFunctionStrings= setting, but no USBFunctionDescriptors=. Ignoring.");
if (s->runtime_max_usec != USEC_INFINITY && s->type == SERVICE_ONESHOT)
log_unit_warning(UNIT(s), "MaxRuntimeSec= has no effect in combination with Type=oneshot. Ignoring.");
2010-04-13 04:00:03 +02:00
return 0;
}
static int service_add_default_dependencies(Service *s) {
int r;
assert(s);
if (!UNIT(s)->default_dependencies)
return 0;
/* Add a number of automatic dependencies useful for the
* majority of services. */
if (MANAGER_IS_SYSTEM(UNIT(s)->manager)) {
/* First, pull in the really early boot stuff, and
* require it, so that we fail if we can't acquire
* it. */
r = unit_add_two_dependencies_by_name(UNIT(s), UNIT_AFTER, UNIT_REQUIRES, SPECIAL_SYSINIT_TARGET, NULL, true, UNIT_DEPENDENCY_DEFAULT);
if (r < 0)
return r;
} else {
/* In the --user instance there's no sysinit.target,
* in that case require basic.target instead. */
r = unit_add_dependency_by_name(UNIT(s), UNIT_REQUIRES, SPECIAL_BASIC_TARGET, NULL, true, UNIT_DEPENDENCY_DEFAULT);
if (r < 0)
return r;
}
/* Second, if the rest of the base system is in the same
* transaction, order us after it, but do not pull it in or
* even require it. */
r = unit_add_dependency_by_name(UNIT(s), UNIT_AFTER, SPECIAL_BASIC_TARGET, NULL, true, UNIT_DEPENDENCY_DEFAULT);
if (r < 0)
return r;
/* Third, add us in for normal shutdown. */
return unit_add_two_dependencies_by_name(UNIT(s), UNIT_BEFORE, UNIT_CONFLICTS, SPECIAL_SHUTDOWN_TARGET, NULL, true, UNIT_DEPENDENCY_DEFAULT);
}
static void service_fix_output(Service *s) {
assert(s);
/* If nothing has been explicitly configured, patch default output in. If input is socket/tty we avoid this
* however, since in that case we want output to default to the same place as we read input from. */
if (s->exec_context.std_error == EXEC_OUTPUT_INHERIT &&
s->exec_context.std_output == EXEC_OUTPUT_INHERIT &&
s->exec_context.std_input == EXEC_INPUT_NULL)
s->exec_context.std_error = UNIT(s)->manager->default_std_error;
if (s->exec_context.std_output == EXEC_OUTPUT_INHERIT &&
s->exec_context.std_input == EXEC_INPUT_NULL)
s->exec_context.std_output = UNIT(s)->manager->default_std_output;
if (s->exec_context.std_input == EXEC_INPUT_NULL &&
s->exec_context.stdin_data_size > 0)
s->exec_context.std_input = EXEC_INPUT_DATA;
}
static int service_setup_bus_name(Service *s) {
int r;
assert(s);
if (!s->bus_name)
return 0;
r = unit_add_dependency_by_name(UNIT(s), UNIT_REQUIRES, SPECIAL_DBUS_SOCKET, NULL, true, UNIT_DEPENDENCY_FILE);
if (r < 0)
return log_unit_error_errno(UNIT(s), r, "Failed to add dependency on " SPECIAL_DBUS_SOCKET ": %m");
/* We always want to be ordered against dbus.socket if both are in the transaction. */
r = unit_add_dependency_by_name(UNIT(s), UNIT_AFTER, SPECIAL_DBUS_SOCKET, NULL, true, UNIT_DEPENDENCY_FILE);
if (r < 0)
2015-11-12 08:23:21 +01:00
return log_unit_error_errno(UNIT(s), r, "Failed to add dependency on " SPECIAL_DBUS_SOCKET ": %m");
r = unit_watch_bus_name(UNIT(s), s->bus_name);
if (r == -EEXIST)
return log_unit_error_errno(UNIT(s), r, "Two services allocated for the same bus name %s, refusing operation.", s->bus_name);
if (r < 0)
return log_unit_error_errno(UNIT(s), r, "Cannot watch bus name %s: %m", s->bus_name);
return 0;
}
static int service_add_extras(Service *s) {
int r;
assert(s);
if (s->type == _SERVICE_TYPE_INVALID) {
/* Figure out a type automatically */
if (s->bus_name)
s->type = SERVICE_DBUS;
else if (s->exec_command[SERVICE_EXEC_START])
s->type = SERVICE_SIMPLE;
else
s->type = SERVICE_ONESHOT;
}
/* Oneshot services have disabled start timeout by default */
if (s->type == SERVICE_ONESHOT && !s->start_timeout_defined)
s->timeout_start_usec = USEC_INFINITY;
service_fix_output(s);
r = unit_patch_contexts(UNIT(s));
if (r < 0)
return r;
r = unit_add_exec_dependencies(UNIT(s), &s->exec_context);
if (r < 0)
return r;
r = unit_set_default_slice(UNIT(s));
if (r < 0)
return r;
if (s->type == SERVICE_NOTIFY && s->notify_access == NOTIFY_NONE)
s->notify_access = NOTIFY_MAIN;
if (s->watchdog_usec > 0 && s->notify_access == NOTIFY_NONE)
s->notify_access = NOTIFY_MAIN;
r = service_add_default_dependencies(s);
if (r < 0)
return r;
r = service_setup_bus_name(s);
if (r < 0)
return r;
return 0;
}
static int service_load(Unit *u) {
Service *s = SERVICE(u);
int r;
assert(s);
2010-01-29 04:42:57 +01:00
/* Load a .service file */
r = unit_load_fragment(u);
if (r < 0)
return r;
2010-04-06 02:43:58 +02:00
/* Still nothing found? Then let's give up */
if (u->load_state == UNIT_STUB)
2010-04-06 02:43:58 +02:00
return -ENOENT;
2010-04-06 02:43:58 +02:00
/* This is a new unit? Then let's add in some extras */
if (u->load_state == UNIT_LOADED) {
/* We were able to load something, then let's add in
* the dropin directories. */
r = unit_load_dropin(u);
if (r < 0)
return r;
/* This is a new unit? Then let's add in some
* extras */
r = service_add_extras(s);
if (r < 0)
return r;
2010-03-31 16:29:55 +02:00
}
2010-04-13 04:00:03 +02:00
return service_verify(s);
}
2010-01-26 21:39:06 +01:00
static void service_dump(Unit *u, FILE *f, const char *prefix) {
char buf_restart[FORMAT_TIMESPAN_MAX], buf_start[FORMAT_TIMESPAN_MAX], buf_stop[FORMAT_TIMESPAN_MAX];
char buf_runtime[FORMAT_TIMESPAN_MAX], buf_watchdog[FORMAT_TIMESPAN_MAX];
ServiceExecCommand c;
2010-01-26 21:39:06 +01:00
Service *s = SERVICE(u);
2010-02-03 14:21:48 +01:00
const char *prefix2;
assert(s);
prefix = strempty(prefix);
prefix2 = strjoina(prefix, "\t");
2010-01-26 07:02:51 +01:00
fprintf(f,
"%sService State: %s\n"
"%sResult: %s\n"
"%sReload Result: %s\n"
"%sPermissionsStartOnly: %s\n"
2010-03-31 16:29:55 +02:00
"%sRootDirectoryStartOnly: %s\n"
"%sRemainAfterExit: %s\n"
"%sGuessMainPID: %s\n"
"%sType: %s\n"
2010-08-09 17:03:46 +02:00
"%sRestart: %s\n"
"%sNotifyAccess: %s\n"
"%sNotifyState: %s\n",
prefix, service_state_to_string(s->state),
prefix, service_result_to_string(s->result),
prefix, service_result_to_string(s->reload_result),
prefix, yes_no(s->permissions_start_only),
2010-03-31 16:29:55 +02:00
prefix, yes_no(s->root_directory_start_only),
prefix, yes_no(s->remain_after_exit),
prefix, yes_no(s->guess_main_pid),
prefix, service_type_to_string(s->type),
2010-08-09 17:03:46 +02:00
prefix, service_restart_to_string(s->restart),
prefix, notify_access_to_string(s->notify_access),
prefix, notify_state_to_string(s->notify_state));
if (s->control_pid > 0)
fprintf(f,
"%sControl PID: "PID_FMT"\n",
prefix, s->control_pid);
if (s->main_pid > 0)
fprintf(f,
"%sMain PID: "PID_FMT"\n"
"%sMain PID Known: %s\n"
"%sMain PID Alien: %s\n",
prefix, s->main_pid,
prefix, yes_no(s->main_pid_known),
prefix, yes_no(s->main_pid_alien));
if (s->pid_file)
fprintf(f,
"%sPIDFile: %s\n",
prefix, s->pid_file);
if (s->bus_name)
fprintf(f,
"%sBusName: %s\n"
"%sBus Name Good: %s\n",
prefix, s->bus_name,
prefix, yes_no(s->bus_name_good));
if (UNIT_ISSET(s->accept_socket))
fprintf(f,
"%sAccept Socket: %s\n",
prefix, UNIT_DEREF(s->accept_socket)->id);
2018-01-30 09:10:47 +01:00
fprintf(f,
"%sRestartSec: %s\n"
"%sTimeoutStartSec: %s\n"
"%sTimeoutStopSec: %s\n"
"%sRuntimeMaxSec: %s\n"
"%sWatchdogSec: %s\n",
prefix, format_timespan(buf_restart, sizeof(buf_restart), s->restart_usec, USEC_PER_SEC),
prefix, format_timespan(buf_start, sizeof(buf_start), s->timeout_start_usec, USEC_PER_SEC),
prefix, format_timespan(buf_stop, sizeof(buf_stop), s->timeout_stop_usec, USEC_PER_SEC),
prefix, format_timespan(buf_runtime, sizeof(buf_runtime), s->runtime_max_usec, USEC_PER_SEC),
prefix, format_timespan(buf_watchdog, sizeof(buf_watchdog), s->watchdog_usec, USEC_PER_SEC));
kill_context_dump(&s->kill_context, f, prefix);
exec_context_dump(&s->exec_context, f, prefix);
for (c = 0; c < _SERVICE_EXEC_COMMAND_MAX; c++) {
2010-01-26 07:02:51 +01:00
if (!s->exec_command[c])
continue;
fprintf(f, "%s-> %s:\n",
prefix, service_exec_command_to_string(c));
2010-01-26 07:02:51 +01:00
exec_command_dump_list(s->exec_command[c], f, prefix2);
}
2010-01-26 07:02:51 +01:00
if (s->status_text)
fprintf(f, "%sStatus Text: %s\n",
prefix, s->status_text);
if (s->n_fd_store_max > 0)
fprintf(f,
"%sFile Descriptor Store Max: %u\n"
"%sFile Descriptor Store Current: %u\n",
prefix, s->n_fd_store_max,
prefix, s->n_fd_store);
cgroup_context_dump(&s->cgroup_context, f, prefix);
}
static int service_is_suitable_main_pid(Service *s, pid_t pid, int prio) {
Unit *owner;
assert(s);
assert(pid_is_valid(pid));
/* Checks whether the specified PID is suitable as main PID for this service. returns negative if not, 0 if the
* PID is questionnable but should be accepted if the source of configuration is trusted. > 0 if the PID is
* good */
if (pid == getpid_cached() || pid == 1) {
log_unit_full(UNIT(s), prio, 0, "New main PID "PID_FMT" is the manager, refusing.", pid);
return -EPERM;
}
if (pid == s->control_pid) {
log_unit_full(UNIT(s), prio, 0, "New main PID "PID_FMT" is the control process, refusing.", pid);
return -EPERM;
}
if (!pid_is_alive(pid)) {
log_unit_full(UNIT(s), prio, 0, "New main PID "PID_FMT" does not exist or is a zombie.", pid);
return -ESRCH;
}
owner = manager_get_unit_by_pid(UNIT(s)->manager, pid);
if (owner == UNIT(s)) {
log_unit_debug(UNIT(s), "New main PID "PID_FMT" belongs to service, we are happy.", pid);
return 1; /* Yay, it's definitely a good PID */
}
return 0; /* Hmm it's a suspicious PID, let's accept it if configuration source is trusted */
}
static int service_load_pid_file(Service *s, bool may_warn) {
char procfs[STRLEN("/proc/self/fd/") + DECIMAL_STR_MAX(int)];
bool questionable_pid_file = false;
_cleanup_free_ char *k = NULL;
_cleanup_close_ int fd = -1;
int r, prio;
pid_t pid;
assert(s);
if (!s->pid_file)
return -ENOENT;
prio = may_warn ? LOG_INFO : LOG_DEBUG;
fd = chase_symlinks(s->pid_file, NULL, CHASE_OPEN|CHASE_SAFE, NULL);
if (fd == -EPERM) {
log_unit_full(UNIT(s), LOG_DEBUG, fd, "Permission denied while opening PID file or potentially unsafe symlink chain, will now retry with relaxed checks: %s", s->pid_file);
questionable_pid_file = true;
fd = chase_symlinks(s->pid_file, NULL, CHASE_OPEN, NULL);
}
if (fd < 0)
return log_unit_full(UNIT(s), prio, fd, "Can't open PID file %s (yet?) after %s: %m", s->pid_file, service_state_to_string(s->state));
/* Let's read the PID file now that we chased it down. But we need to convert the O_PATH fd chase_symlinks() returned us into a proper fd first. */
xsprintf(procfs, "/proc/self/fd/%i", fd);
r = read_one_line_file(procfs, &k);
if (r < 0)
return log_unit_error_errno(UNIT(s), r, "Can't convert PID files %s O_PATH file descriptor to proper file descriptor: %m", s->pid_file);
r = parse_pid(k, &pid);
if (r < 0)
return log_unit_full(UNIT(s), prio, r, "Failed to parse PID from file %s: %m", s->pid_file);
if (s->main_pid_known && pid == s->main_pid)
return 0;
r = service_is_suitable_main_pid(s, pid, prio);
if (r < 0)
return r;
if (r == 0) {
struct stat st;
2010-06-15 19:47:13 +02:00
if (questionable_pid_file) {
log_unit_error(UNIT(s), "Refusing to accept PID outside of service control group, acquired through unsafe symlink chain: %s", s->pid_file);
return -EPERM;
}
/* Hmm, it's not clear if the new main PID is safe. Let's allow this if the PID file is owned by root */
if (fstat(fd, &st) < 0)
return log_unit_error_errno(UNIT(s), errno, "Failed to fstat() PID file O_PATH fd: %m");
if (st.st_uid != 0) {
log_unit_error(UNIT(s), "New main PID "PID_FMT" does not belong to service, and PID file is not owned by root. Refusing.", pid);
return -EPERM;
}
log_unit_debug(UNIT(s), "New main PID "PID_FMT" does not belong to service, but we'll accept it since PID file is owned by root.", pid);
}
if (s->main_pid_known) {
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_debug(UNIT(s), "Main PID changing: "PID_FMT" -> "PID_FMT, s->main_pid, pid);
service_unwatch_main_pid(s);
s->main_pid_known = false;
} else
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_debug(UNIT(s), "Main PID loaded: "PID_FMT, pid);
r = service_set_main_pid(s, pid);
if (r < 0)
return r;
r = unit_watch_pid(UNIT(s), pid);
if (r < 0) /* FIXME: we need to do something here */
return log_unit_warning_errno(UNIT(s), r, "Failed to watch PID "PID_FMT" for service: %m", pid);
return 1;
}
static void service_search_main_pid(Service *s) {
core: unified cgroup hierarchy support This patch set adds full support the new unified cgroup hierarchy logic of modern kernels. A new kernel command line option "systemd.unified_cgroup_hierarchy=1" is added. If specified the unified hierarchy is mounted to /sys/fs/cgroup instead of a tmpfs. No further hierarchies are mounted. The kernel command line option defaults to off. We can turn it on by default as soon as the kernel's APIs regarding this are stabilized (but even then downstream distros might want to turn this off, as this will break any tools that access cgroupfs directly). It is possibly to choose for each boot individually whether the unified or the legacy hierarchy is used. nspawn will by default provide the legacy hierarchy to containers if the host is using it, and the unified otherwise. However it is possible to run containers with the unified hierarchy on a legacy host and vice versa, by setting the $UNIFIED_CGROUP_HIERARCHY environment variable for nspawn to 1 or 0, respectively. The unified hierarchy provides reliable cgroup empty notifications for the first time, via inotify. To make use of this we maintain one manager-wide inotify fd, and each cgroup to it. This patch also removes cg_delete() which is unused now. On kernel 4.2 only the "memory" controller is compatible with the unified hierarchy, hence that's the only controller systemd exposes when booted in unified heirarchy mode. This introduces a new enum for enumerating supported controllers, plus a related enum for the mask bits mapping to it. The core is changed to make use of this everywhere. This moves PID 1 into a new "init.scope" implicit scope unit in the root slice. This is necessary since on the unified hierarchy cgroups may either contain subgroups or processes but not both. PID 1 hence has to move out of the root cgroup (strictly speaking the root cgroup is the only one where processes and subgroups are still allowed, but in order to support containers nicey, we move PID 1 into the new scope in all cases.) This new unit is also used on legacy hierarchy setups. It's actually pretty useful on all systems, as it can then be used to filter journal messages coming from PID 1, and so on. The root slice ("-.slice") is now implicitly created and started (and does not require a unit file on disk anymore), since that's where "init.scope" is located and the slice needs to be started before the scope can. To check whether we are in unified or legacy hierarchy mode we use statfs() on /sys/fs/cgroup. If the .f_type field reports tmpfs we are in legacy mode, if it reports cgroupfs we are in unified mode. This patch set carefuly makes sure that cgls and cgtop continue to work as desired. When invoking nspawn as a service it will implicitly create two subcgroups in the cgroup it is using, one to move the nspawn process into, the other to move the actual container processes into. This is done because of the requirement that cgroups may either contain processes or other subgroups.
2015-09-01 19:22:36 +02:00
pid_t pid = 0;
int r;
assert(s);
/* If we know it anyway, don't ever fallback to unreliable
* heuristics */
if (s->main_pid_known)
return;
if (!s->guess_main_pid)
return;
assert(s->main_pid <= 0);
if (unit_search_main_pid(UNIT(s), &pid) < 0)
return;
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_debug(UNIT(s), "Main PID guessed: "PID_FMT, pid);
if (service_set_main_pid(s, pid) < 0)
return;
r = unit_watch_pid(UNIT(s), pid);
if (r < 0)
/* FIXME: we need to do something here */
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning_errno(UNIT(s), r, "Failed to watch PID "PID_FMT" from: %m", pid);
}
static void service_set_state(Service *s, ServiceState state) {
ServiceState old_state;
const UnitActiveState *table;
2013-12-11 20:49:43 +01:00
assert(s);
table = s->type == SERVICE_IDLE ? state_translation_table_idle : state_translation_table;
old_state = s->state;
s->state = state;
service_unwatch_pid_file(s);
2013-12-11 20:49:43 +01:00
if (!IN_SET(state,
SERVICE_START_PRE, SERVICE_START, SERVICE_START_POST,
SERVICE_RUNNING,
2013-12-11 20:49:43 +01:00
SERVICE_RELOAD,
SERVICE_STOP, SERVICE_STOP_SIGABRT, SERVICE_STOP_SIGTERM, SERVICE_STOP_SIGKILL, SERVICE_STOP_POST,
2013-12-11 20:49:43 +01:00
SERVICE_FINAL_SIGTERM, SERVICE_FINAL_SIGKILL,
SERVICE_AUTO_RESTART))
s->timer_event_source = sd_event_source_unref(s->timer_event_source);
2013-12-11 20:49:43 +01:00
if (!IN_SET(state,
SERVICE_START, SERVICE_START_POST,
SERVICE_RUNNING, SERVICE_RELOAD,
SERVICE_STOP, SERVICE_STOP_SIGABRT, SERVICE_STOP_SIGTERM, SERVICE_STOP_SIGKILL, SERVICE_STOP_POST,
SERVICE_FINAL_SIGTERM, SERVICE_FINAL_SIGKILL)) {
service_unwatch_main_pid(s);
s->main_command = NULL;
}
2013-12-11 20:49:43 +01:00
if (!IN_SET(state,
SERVICE_START_PRE, SERVICE_START, SERVICE_START_POST,
SERVICE_RELOAD,
SERVICE_STOP, SERVICE_STOP_SIGABRT, SERVICE_STOP_SIGTERM, SERVICE_STOP_SIGKILL, SERVICE_STOP_POST,
2013-12-11 20:49:43 +01:00
SERVICE_FINAL_SIGTERM, SERVICE_FINAL_SIGKILL)) {
service_unwatch_control_pid(s);
s->control_command = NULL;
2010-04-21 03:27:44 +02:00
s->control_command_id = _SERVICE_EXEC_COMMAND_INVALID;
}
if (IN_SET(state, SERVICE_DEAD, SERVICE_FAILED, SERVICE_AUTO_RESTART))
unit_unwatch_all_pids(UNIT(s));
2013-12-11 20:49:43 +01:00
if (!IN_SET(state,
SERVICE_START_PRE, SERVICE_START, SERVICE_START_POST,
SERVICE_RUNNING, SERVICE_RELOAD,
SERVICE_STOP, SERVICE_STOP_SIGABRT, SERVICE_STOP_SIGTERM, SERVICE_STOP_SIGKILL, SERVICE_STOP_POST,
SERVICE_FINAL_SIGTERM, SERVICE_FINAL_SIGKILL) &&
!(state == SERVICE_DEAD && UNIT(s)->job))
service_close_socket_fd(s);
if (!IN_SET(state, SERVICE_START_POST, SERVICE_RUNNING, SERVICE_RELOAD))
service_stop_watchdog(s);
/* For the inactive states unit_notify() will trim the cgroup,
* but for exit we have to do that ourselves... */
if (state == SERVICE_EXITED && !MANAGER_IS_RELOADING(UNIT(s)->manager))
core: unified cgroup hierarchy support This patch set adds full support the new unified cgroup hierarchy logic of modern kernels. A new kernel command line option "systemd.unified_cgroup_hierarchy=1" is added. If specified the unified hierarchy is mounted to /sys/fs/cgroup instead of a tmpfs. No further hierarchies are mounted. The kernel command line option defaults to off. We can turn it on by default as soon as the kernel's APIs regarding this are stabilized (but even then downstream distros might want to turn this off, as this will break any tools that access cgroupfs directly). It is possibly to choose for each boot individually whether the unified or the legacy hierarchy is used. nspawn will by default provide the legacy hierarchy to containers if the host is using it, and the unified otherwise. However it is possible to run containers with the unified hierarchy on a legacy host and vice versa, by setting the $UNIFIED_CGROUP_HIERARCHY environment variable for nspawn to 1 or 0, respectively. The unified hierarchy provides reliable cgroup empty notifications for the first time, via inotify. To make use of this we maintain one manager-wide inotify fd, and each cgroup to it. This patch also removes cg_delete() which is unused now. On kernel 4.2 only the "memory" controller is compatible with the unified hierarchy, hence that's the only controller systemd exposes when booted in unified heirarchy mode. This introduces a new enum for enumerating supported controllers, plus a related enum for the mask bits mapping to it. The core is changed to make use of this everywhere. This moves PID 1 into a new "init.scope" implicit scope unit in the root slice. This is necessary since on the unified hierarchy cgroups may either contain subgroups or processes but not both. PID 1 hence has to move out of the root cgroup (strictly speaking the root cgroup is the only one where processes and subgroups are still allowed, but in order to support containers nicey, we move PID 1 into the new scope in all cases.) This new unit is also used on legacy hierarchy setups. It's actually pretty useful on all systems, as it can then be used to filter journal messages coming from PID 1, and so on. The root slice ("-.slice") is now implicitly created and started (and does not require a unit file on disk anymore), since that's where "init.scope" is located and the slice needs to be started before the scope can. To check whether we are in unified or legacy hierarchy mode we use statfs() on /sys/fs/cgroup. If the .f_type field reports tmpfs we are in legacy mode, if it reports cgroupfs we are in unified mode. This patch set carefuly makes sure that cgls and cgtop continue to work as desired. When invoking nspawn as a service it will implicitly create two subcgroups in the cgroup it is using, one to move the nspawn process into, the other to move the actual container processes into. This is done because of the requirement that cgroups may either contain processes or other subgroups.
2015-09-01 19:22:36 +02:00
unit_prune_cgroup(UNIT(s));
if (old_state != state)
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_debug(UNIT(s), "Changed %s -> %s", service_state_to_string(old_state), service_state_to_string(state));
2010-01-27 04:31:52 +01:00
unit_notify(UNIT(s), table[old_state], table[state], s->reload_result == SERVICE_SUCCESS);
}
static usec_t service_coldplug_timeout(Service *s) {
assert(s);
switch (s->deserialized_state) {
case SERVICE_START_PRE:
case SERVICE_START:
case SERVICE_START_POST:
case SERVICE_RELOAD:
return usec_add(UNIT(s)->state_change_timestamp.monotonic, s->timeout_start_usec);
case SERVICE_RUNNING:
return usec_add(UNIT(s)->active_enter_timestamp.monotonic, s->runtime_max_usec);
case SERVICE_STOP:
case SERVICE_STOP_SIGABRT:
case SERVICE_STOP_SIGTERM:
case SERVICE_STOP_SIGKILL:
case SERVICE_STOP_POST:
case SERVICE_FINAL_SIGTERM:
case SERVICE_FINAL_SIGKILL:
return usec_add(UNIT(s)->state_change_timestamp.monotonic, s->timeout_stop_usec);
case SERVICE_AUTO_RESTART:
return usec_add(UNIT(s)->inactive_enter_timestamp.monotonic, s->restart_usec);
default:
return USEC_INFINITY;
}
}
static int service_coldplug(Unit *u) {
2010-04-21 03:27:44 +02:00
Service *s = SERVICE(u);
int r;
assert(s);
assert(s->state == SERVICE_DEAD);
if (s->deserialized_state == s->state)
return 0;
r = service_arm_timer(s, service_coldplug_timeout(s));
if (r < 0)
return r;
if (s->main_pid > 0 &&
pid_is_unwaited(s->main_pid) &&
(IN_SET(s->deserialized_state,
SERVICE_START, SERVICE_START_POST,
SERVICE_RUNNING, SERVICE_RELOAD,
SERVICE_STOP, SERVICE_STOP_SIGABRT, SERVICE_STOP_SIGTERM, SERVICE_STOP_SIGKILL, SERVICE_STOP_POST,
SERVICE_FINAL_SIGTERM, SERVICE_FINAL_SIGKILL))) {
r = unit_watch_pid(UNIT(s), s->main_pid);
if (r < 0)
return r;
}
if (s->control_pid > 0 &&
pid_is_unwaited(s->control_pid) &&
IN_SET(s->deserialized_state,
SERVICE_START_PRE, SERVICE_START, SERVICE_START_POST,
SERVICE_RELOAD,
SERVICE_STOP, SERVICE_STOP_SIGABRT, SERVICE_STOP_SIGTERM, SERVICE_STOP_SIGKILL, SERVICE_STOP_POST,
SERVICE_FINAL_SIGTERM, SERVICE_FINAL_SIGKILL)) {
r = unit_watch_pid(UNIT(s), s->control_pid);
if (r < 0)
return r;
2010-04-21 03:27:44 +02:00
}
if (!IN_SET(s->deserialized_state, SERVICE_DEAD, SERVICE_FAILED, SERVICE_AUTO_RESTART))
unit_watch_all_pids(UNIT(s));
if (IN_SET(s->deserialized_state, SERVICE_START_POST, SERVICE_RUNNING, SERVICE_RELOAD))
service_start_watchdog(s);
if (!IN_SET(s->deserialized_state, SERVICE_DEAD, SERVICE_FAILED, SERVICE_AUTO_RESTART)) {
(void) unit_setup_dynamic_creds(u);
(void) unit_setup_exec_runtime(u);
}
if (UNIT_ISSET(s->accept_socket)) {
Socket* socket = SOCKET(UNIT_DEREF(s->accept_socket));
if (socket->max_connections_per_source > 0) {
SocketPeer *peer;
/* Make a best-effort attempt at bumping the connection count */
if (socket_acquire_peer(socket, s->socket_fd, &peer) > 0) {
socket_peer_unref(s->peer);
s->peer = peer;
}
}
}
service_set_state(s, s->deserialized_state);
2010-04-21 03:27:44 +02:00
return 0;
}
static int service_collect_fds(Service *s,
int **fds,
char ***fd_names,
unsigned *n_storage_fds,
unsigned *n_socket_fds) {
_cleanup_strv_free_ char **rfd_names = NULL;
_cleanup_free_ int *rfds = NULL;
unsigned rn_socket_fds = 0, rn_storage_fds = 0;
int r;
2010-01-26 07:02:51 +01:00
assert(s);
assert(fds);
assert(fd_names);
assert(n_socket_fds);
2010-01-26 07:02:51 +01:00
if (s->socket_fd >= 0) {
/* Pass the per-connection socket */
rfds = new(int, 1);
if (!rfds)
return -ENOMEM;
rfds[0] = s->socket_fd;
rfd_names = strv_new("connection", NULL);
if (!rfd_names)
return -ENOMEM;
2010-01-26 07:02:51 +01:00
rn_socket_fds = 1;
} else {
Iterator i;
void *v;
Unit *u;
2010-01-26 07:02:51 +01:00
/* Pass all our configured sockets for singleton services */
2010-01-26 07:02:51 +01:00
HASHMAP_FOREACH_KEY(v, u, UNIT(s)->dependencies[UNIT_TRIGGERED_BY], i) {
_cleanup_free_ int *cfds = NULL;
Socket *sock;
int cn_fds;
2010-01-26 07:02:51 +01:00
if (u->type != UNIT_SOCKET)
continue;
2010-01-26 07:02:51 +01:00
sock = SOCKET(u);
cn_fds = socket_collect_fds(sock, &cfds);
if (cn_fds < 0)
return cn_fds;
2010-01-26 07:02:51 +01:00
if (cn_fds <= 0)
continue;
if (!rfds) {
rfds = cfds;
rn_socket_fds = cn_fds;
cfds = NULL;
} else {
int *t;
t = reallocarray(rfds, rn_socket_fds + cn_fds, sizeof(int));
if (!t)
return -ENOMEM;
memcpy(t + rn_socket_fds, cfds, cn_fds * sizeof(int));
rfds = t;
rn_socket_fds += cn_fds;
}
r = strv_extend_n(&rfd_names, socket_fdname(sock), cn_fds);
if (r < 0)
return r;
2010-01-26 07:02:51 +01:00
}
}
if (s->n_fd_store > 0) {
ServiceFDStore *fs;
unsigned n_fds;
char **nl;
int *t;
t = reallocarray(rfds, rn_socket_fds + s->n_fd_store, sizeof(int));
if (!t)
return -ENOMEM;
rfds = t;
nl = reallocarray(rfd_names, rn_socket_fds + s->n_fd_store + 1, sizeof(char *));
if (!nl)
return -ENOMEM;
rfd_names = nl;
n_fds = rn_socket_fds;
LIST_FOREACH(fd_store, fs, s->fd_store) {
rfds[n_fds] = fs->fd;
rfd_names[n_fds] = strdup(strempty(fs->fdname));
if (!rfd_names[n_fds])
return -ENOMEM;
rn_storage_fds++;
n_fds++;
}
rfd_names[n_fds] = NULL;
}
2010-01-26 07:02:51 +01:00
*fds = rfds;
*fd_names = rfd_names;
*n_socket_fds = rn_socket_fds;
*n_storage_fds = rn_storage_fds;
rfds = NULL;
rfd_names = NULL;
return 0;
2010-01-26 07:02:51 +01:00
}
static bool service_exec_needs_notify_socket(Service *s, ExecFlags flags) {
assert(s);
/* Notifications are accepted depending on the process and
* the access setting of the service:
* process: \ access: NONE MAIN EXEC ALL
* main no yes yes yes
* control no no yes yes
* other (forked) no no no yes */
if (flags & EXEC_IS_CONTROL)
/* A control process */
return IN_SET(s->notify_access, NOTIFY_EXEC, NOTIFY_ALL);
/* We only spawn main processes and control processes, so any
* process that is not a control process is a main process */
return s->notify_access != NOTIFY_NONE;
}
static int service_spawn(
Service *s,
ExecCommand *c,
usec_t timeout,
ExecFlags flags,
pid_t *_pid) {
ExecParameters exec_params = {
.flags = flags,
.stdin_fd = -1,
.stdout_fd = -1,
.stderr_fd = -1,
};
_cleanup_strv_free_ char **final_env = NULL, **our_env = NULL, **fd_names = NULL;
unsigned n_storage_fds = 0, n_socket_fds = 0, n_env = 0;
_cleanup_free_ int *fds = NULL;
pid_t pid;
int r;
assert(s);
assert(c);
assert(_pid);
r = unit_prepare_exec(UNIT(s));
if (r < 0)
return r;
if (flags & EXEC_IS_CONTROL) {
/* If this is a control process, mask the permissions/chroot application if this is requested. */
if (s->permissions_start_only)
exec_params.flags &= ~EXEC_APPLY_SANDBOXING;
if (s->root_directory_start_only)
exec_params.flags &= ~EXEC_APPLY_CHROOT;
}
if ((flags & EXEC_PASS_FDS) ||
s->exec_context.std_input == EXEC_INPUT_SOCKET ||
s->exec_context.std_output == EXEC_OUTPUT_SOCKET ||
s->exec_context.std_error == EXEC_OUTPUT_SOCKET) {
r = service_collect_fds(s, &fds, &fd_names, &n_storage_fds, &n_socket_fds);
if (r < 0)
return r;
log_unit_debug(UNIT(s), "Passing %i fds to service", n_storage_fds + n_socket_fds);
}
2010-01-26 07:02:51 +01:00
r = service_arm_timer(s, usec_add(now(CLOCK_MONOTONIC), timeout));
if (r < 0)
return r;
our_env = new0(char*, 9);
if (!our_env)
return -ENOMEM;
if (service_exec_needs_notify_socket(s, flags))
if (asprintf(our_env + n_env++, "NOTIFY_SOCKET=%s", UNIT(s)->manager->notify_socket) < 0)
return -ENOMEM;
if (s->main_pid > 0)
if (asprintf(our_env + n_env++, "MAINPID="PID_FMT, s->main_pid) < 0)
return -ENOMEM;
if (MANAGER_IS_USER(UNIT(s)->manager))
if (asprintf(our_env + n_env++, "MANAGERPID="PID_FMT, getpid_cached()) < 0)
return -ENOMEM;
if (s->socket_fd >= 0) {
union sockaddr_union sa;
socklen_t salen = sizeof(sa);
/* If this is a per-connection service instance, let's set $REMOTE_ADDR and $REMOTE_PORT to something
* useful. Note that we do this only when we are still connected at this point in time, which we might
* very well not be. Hence we ignore all errors when retrieving peer information (as that might result
* in ENOTCONN), and just use whate we can use. */
if (getpeername(s->socket_fd, &sa.sa, &salen) >= 0 &&
IN_SET(sa.sa.sa_family, AF_INET, AF_INET6, AF_VSOCK)) {
_cleanup_free_ char *addr = NULL;
char *t;
unsigned port;
r = sockaddr_pretty(&sa.sa, salen, true, false, &addr);
if (r < 0)
return r;
t = strappend("REMOTE_ADDR=", addr);
if (!t)
return -ENOMEM;
our_env[n_env++] = t;
r = sockaddr_port(&sa.sa, &port);
if (r < 0)
return r;
if (asprintf(&t, "REMOTE_PORT=%u", port) < 0)
return -ENOMEM;
our_env[n_env++] = t;
}
}
if (flags & EXEC_SETENV_RESULT) {
if (asprintf(our_env + n_env++, "SERVICE_RESULT=%s", service_result_to_string(s->result)) < 0)
return -ENOMEM;
if (s->main_exec_status.pid > 0 &&
dual_timestamp_is_set(&s->main_exec_status.exit_timestamp)) {
if (asprintf(our_env + n_env++, "EXIT_CODE=%s", sigchld_code_to_string(s->main_exec_status.code)) < 0)
return -ENOMEM;
if (s->main_exec_status.code == CLD_EXITED)
r = asprintf(our_env + n_env++, "EXIT_STATUS=%i", s->main_exec_status.status);
else
r = asprintf(our_env + n_env++, "EXIT_STATUS=%s", signal_to_string(s->main_exec_status.status));
if (r < 0)
return -ENOMEM;
}
}
unit_set_exec_params(UNIT(s), &exec_params);
final_env = strv_env_merge(2, exec_params.environment, our_env, NULL);
if (!final_env)
return -ENOMEM;
/* System services should get a new keyring by default. */
SET_FLAG(exec_params.flags, EXEC_NEW_KEYRING, MANAGER_IS_SYSTEM(UNIT(s)->manager));
/* System D-Bus needs nss-systemd disabled, so that we don't deadlock */
SET_FLAG(exec_params.flags, EXEC_NSS_BYPASS_BUS,
MANAGER_IS_SYSTEM(UNIT(s)->manager) && unit_has_name(UNIT(s), SPECIAL_DBUS_SERVICE));
exec_params.argv = c->argv;
exec_params.environment = final_env;
exec_params.fds = fds;
exec_params.fd_names = fd_names;
exec_params.n_storage_fds = n_storage_fds;
exec_params.n_socket_fds = n_socket_fds;
exec_params.watchdog_usec = s->watchdog_usec;
exec_params.selinux_context_net = s->socket_fd_selinux_context_net;
if (s->type == SERVICE_IDLE)
exec_params.idle_pipe = UNIT(s)->manager->idle_pipe;
exec_params.stdin_fd = s->stdin_fd;
exec_params.stdout_fd = s->stdout_fd;
exec_params.stderr_fd = s->stderr_fd;
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
r = exec_spawn(UNIT(s),
c,
2010-04-15 03:11:11 +02:00
&s->exec_context,
&exec_params,
s->exec_runtime,
&s->dynamic_creds,
2010-04-15 03:11:11 +02:00
&pid);
if (r < 0)
return r;
r = unit_watch_pid(UNIT(s), pid);
if (r < 0) /* FIXME: we need to do something here */
return r;
*_pid = pid;
return 0;
}
static int main_pid_good(Service *s) {
assert(s);
/* Returns 0 if the pid is dead, > 0 if it is good, < 0 if we don't know */
2015-07-07 02:09:54 +02:00
/* If we know the pid file, then let's just check if it is
* still valid */
if (s->main_pid_known) {
/* If it's an alien child let's check if it is still
* alive ... */
if (s->main_pid_alien && s->main_pid > 0)
return pid_is_alive(s->main_pid);
/* .. otherwise assume we'll get a SIGCHLD for it,
* which we really should wait for to collect exit
* status and code */
return s->main_pid > 0;
}
/* We don't know the pid */
return -EAGAIN;
}
static int control_pid_good(Service *s) {
assert(s);
/* Returns 0 if the control PID is dead, > 0 if it is good. We never actually return < 0 here, but in order to
* make this function as similar as possible to main_pid_good() and cgroup_good(), we pretend that < 0 also
* means: we can't figure it out. */
return s->control_pid > 0;
}
static int cgroup_good(Service *s) {
int r;
assert(s);
/* Returns 0 if the cgroup is empty or doesn't exist, > 0 if it is exists and is populated, < 0 if we can't
* figure it out */
if (!UNIT(s)->cgroup_path)
return 0;
r = cg_is_empty_recursive(SYSTEMD_CGROUP_CONTROLLER, UNIT(s)->cgroup_path);
if (r < 0)
return r;
return r == 0;
}
static bool service_shall_restart(Service *s) {
assert(s);
/* Don't restart after manual stops */
if (s->forbid_restart)
return false;
/* Never restart if this is configured as special exception */
if (exit_status_set_test(&s->restart_prevent_status, s->main_exec_status.code, s->main_exec_status.status))
return false;
/* Restart if the exit code/status are configured as restart triggers */
if (exit_status_set_test(&s->restart_force_status, s->main_exec_status.code, s->main_exec_status.status))
return true;
switch (s->restart) {
case SERVICE_RESTART_NO:
return false;
case SERVICE_RESTART_ALWAYS:
return true;
case SERVICE_RESTART_ON_SUCCESS:
return s->result == SERVICE_SUCCESS;
case SERVICE_RESTART_ON_FAILURE:
return s->result != SERVICE_SUCCESS;
case SERVICE_RESTART_ON_ABNORMAL:
return !IN_SET(s->result, SERVICE_SUCCESS, SERVICE_FAILURE_EXIT_CODE);
case SERVICE_RESTART_ON_WATCHDOG:
return s->result == SERVICE_FAILURE_WATCHDOG;
case SERVICE_RESTART_ON_ABORT:
return IN_SET(s->result, SERVICE_FAILURE_SIGNAL, SERVICE_FAILURE_CORE_DUMP);
default:
assert_not_reached("unknown restart setting");
}
}
static bool service_will_restart(Unit *u) {
Service *s = SERVICE(u);
assert(s);
if (s->will_auto_restart)
return true;
if (s->state == SERVICE_AUTO_RESTART)
return true;
if (!UNIT(s)->job)
return false;
if (UNIT(s)->job->type == JOB_START)
return true;
return false;
}
static void service_enter_dead(Service *s, ServiceResult f, bool allow_restart) {
int r;
assert(s);
/* If there's a stop job queued before we enter the DEAD state, we shouldn't act on Restart=, in order to not
* undo what has already been enqueued. */
if (unit_stop_pending(UNIT(s)))
allow_restart = false;
if (s->result == SERVICE_SUCCESS)
s->result = f;
if (s->result != SERVICE_SUCCESS)
log_unit_warning(UNIT(s), "Failed with result '%s'.", service_result_to_string(s->result));
if (allow_restart && service_shall_restart(s))
s->will_auto_restart = true;
/* Make sure service_release_resources() doesn't destroy our FD store, while we are changing through
* SERVICE_FAILED/SERVICE_DEAD before entering into SERVICE_AUTO_RESTART. */
s->n_keep_fd_store ++;
service_set_state(s, s->result != SERVICE_SUCCESS ? SERVICE_FAILED : SERVICE_DEAD);
if (s->will_auto_restart) {
s->will_auto_restart = false;
r = service_arm_timer(s, usec_add(now(CLOCK_MONOTONIC), s->restart_usec));
if (r < 0) {
s->n_keep_fd_store--;
goto fail;
}
service_set_state(s, SERVICE_AUTO_RESTART);
} else
/* If we shan't restart, then flush out the restart counter. But don't do that immediately, so that the
* user can still introspect the counter. Do so on the next start. */
s->flush_n_restarts = true;
/* The new state is in effect, let's decrease the fd store ref counter again. Let's also readd us to the GC
* queue, so that the fd store is possibly gc'ed again */
s->n_keep_fd_store--;
unit_add_to_gc_queue(UNIT(s));
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
/* The next restart might not be a manual stop, hence reset the flag indicating manual stops */
s->forbid_restart = false;
/* We want fresh tmpdirs in case service is started again immediately */
s->exec_runtime = exec_runtime_unref(s->exec_runtime, true);
if (s->exec_context.runtime_directory_preserve_mode == EXEC_PRESERVE_NO ||
(s->exec_context.runtime_directory_preserve_mode == EXEC_PRESERVE_RESTART && !service_will_restart(UNIT(s))))
/* Also, remove the runtime directory */
exec_context_destroy_runtime_directory(&s->exec_context, UNIT(s)->manager->prefix[EXEC_DIRECTORY_RUNTIME]);
/* Get rid of the IPC bits of the user */
unit_unref_uid_gid(UNIT(s), true);
/* Release the user, and destroy it if we are the only remaining owner */
dynamic_creds_destroy(&s->dynamic_creds);
/* Try to delete the pid file. At this point it will be
* out-of-date, and some software might be confused by it, so
* let's remove it. */
if (s->pid_file)
2016-02-01 21:55:39 +01:00
(void) unlink(s->pid_file);
return;
fail:
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning_errno(UNIT(s), r, "Failed to run install restart timer: %m");
service_enter_dead(s, SERVICE_FAILURE_RESOURCES, false);
}
static void service_enter_stop_post(Service *s, ServiceResult f) {
int r;
assert(s);
if (s->result == SERVICE_SUCCESS)
s->result = f;
service_unwatch_control_pid(s);
unit_watch_all_pids(UNIT(s));
s->control_command = s->exec_command[SERVICE_EXEC_STOP_POST];
if (s->control_command) {
s->control_command_id = SERVICE_EXEC_STOP_POST;
r = service_spawn(s,
s->control_command,
s->timeout_stop_usec,
EXEC_APPLY_SANDBOXING|EXEC_APPLY_CHROOT|EXEC_APPLY_TTY_STDIN|EXEC_IS_CONTROL|EXEC_SETENV_RESULT,
&s->control_pid);
if (r < 0)
goto fail;
service_set_state(s, SERVICE_STOP_POST);
} else
service_enter_signal(s, SERVICE_FINAL_SIGTERM, SERVICE_SUCCESS);
return;
fail:
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning_errno(UNIT(s), r, "Failed to run 'stop-post' task: %m");
service_enter_signal(s, SERVICE_FINAL_SIGTERM, SERVICE_FAILURE_RESOURCES);
}
static int state_to_kill_operation(ServiceState state) {
switch (state) {
case SERVICE_STOP_SIGABRT:
return KILL_ABORT;
case SERVICE_STOP_SIGTERM:
case SERVICE_FINAL_SIGTERM:
return KILL_TERMINATE;
case SERVICE_STOP_SIGKILL:
case SERVICE_FINAL_SIGKILL:
return KILL_KILL;
default:
return _KILL_OPERATION_INVALID;
}
}
static void service_enter_signal(Service *s, ServiceState state, ServiceResult f) {
int r;
assert(s);
if (s->result == SERVICE_SUCCESS)
s->result = f;
unit_watch_all_pids(UNIT(s));
r = unit_kill_context(
UNIT(s),
&s->kill_context,
state_to_kill_operation(state),
s->main_pid,
s->control_pid,
s->main_pid_alien);
if (r < 0)
goto fail;
if (r > 0) {
r = service_arm_timer(s, usec_add(now(CLOCK_MONOTONIC), s->timeout_stop_usec));
if (r < 0)
goto fail;
service_set_state(s, state);
} else if (IN_SET(state, SERVICE_STOP_SIGABRT, SERVICE_STOP_SIGTERM) && s->kill_context.send_sigkill)
service_enter_signal(s, SERVICE_STOP_SIGKILL, SERVICE_SUCCESS);
else if (IN_SET(state, SERVICE_STOP_SIGABRT, SERVICE_STOP_SIGTERM, SERVICE_STOP_SIGKILL))
service_enter_stop_post(s, SERVICE_SUCCESS);
else if (state == SERVICE_FINAL_SIGTERM && s->kill_context.send_sigkill)
service_enter_signal(s, SERVICE_FINAL_SIGKILL, SERVICE_SUCCESS);
else
service_enter_dead(s, SERVICE_SUCCESS, true);
return;
fail:
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning_errno(UNIT(s), r, "Failed to kill processes: %m");
if (IN_SET(state, SERVICE_STOP_SIGABRT, SERVICE_STOP_SIGTERM, SERVICE_STOP_SIGKILL))
service_enter_stop_post(s, SERVICE_FAILURE_RESOURCES);
else
service_enter_dead(s, SERVICE_FAILURE_RESOURCES, true);
}
static void service_enter_stop_by_notify(Service *s) {
assert(s);
unit_watch_all_pids(UNIT(s));
service_arm_timer(s, usec_add(now(CLOCK_MONOTONIC), s->timeout_stop_usec));
/* The service told us it's stopping, so it's as if we SIGTERM'd it. */
service_set_state(s, SERVICE_STOP_SIGTERM);
}
static void service_enter_stop(Service *s, ServiceResult f) {
int r;
assert(s);
if (s->result == SERVICE_SUCCESS)
s->result = f;
service_unwatch_control_pid(s);
unit_watch_all_pids(UNIT(s));
s->control_command = s->exec_command[SERVICE_EXEC_STOP];
if (s->control_command) {
s->control_command_id = SERVICE_EXEC_STOP;
r = service_spawn(s,
s->control_command,
s->timeout_stop_usec,
EXEC_APPLY_SANDBOXING|EXEC_APPLY_CHROOT|EXEC_IS_CONTROL|EXEC_SETENV_RESULT,
&s->control_pid);
if (r < 0)
goto fail;
service_set_state(s, SERVICE_STOP);
} else
service_enter_signal(s, SERVICE_STOP_SIGTERM, SERVICE_SUCCESS);
return;
fail:
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning_errno(UNIT(s), r, "Failed to run 'stop' task: %m");
service_enter_signal(s, SERVICE_STOP_SIGTERM, SERVICE_FAILURE_RESOURCES);
}
static bool service_good(Service *s) {
int main_pid_ok;
assert(s);
if (s->type == SERVICE_DBUS && !s->bus_name_good)
return false;
main_pid_ok = main_pid_good(s);
if (main_pid_ok > 0) /* It's alive */
return true;
if (main_pid_ok == 0) /* It's dead */
return false;
/* OK, we don't know anything about the main PID, maybe
* because there is none. Let's check the control group
* instead. */
return cgroup_good(s) != 0;
}
static void service_enter_running(Service *s, ServiceResult f) {
assert(s);
if (s->result == SERVICE_SUCCESS)
s->result = f;
service_unwatch_control_pid(s);
if (service_good(s)) {
/* If there are any queued up sd_notify()
* notifications, process them now */
if (s->notify_state == NOTIFY_RELOADING)
service_enter_reload_by_notify(s);
else if (s->notify_state == NOTIFY_STOPPING)
service_enter_stop_by_notify(s);
else {
service_set_state(s, SERVICE_RUNNING);
service_arm_timer(s, usec_add(UNIT(s)->active_enter_timestamp.monotonic, s->runtime_max_usec));
}
} else if (f != SERVICE_SUCCESS)
service_enter_signal(s, SERVICE_STOP_SIGTERM, f);
else if (s->remain_after_exit)
service_set_state(s, SERVICE_EXITED);
else
service_enter_stop(s, SERVICE_SUCCESS);
}
static void service_enter_start_post(Service *s) {
int r;
assert(s);
service_unwatch_control_pid(s);
2013-12-11 20:49:43 +01:00
service_reset_watchdog(s);
s->control_command = s->exec_command[SERVICE_EXEC_START_POST];
if (s->control_command) {
s->control_command_id = SERVICE_EXEC_START_POST;
r = service_spawn(s,
s->control_command,
s->timeout_start_usec,
EXEC_APPLY_SANDBOXING|EXEC_APPLY_CHROOT|EXEC_IS_CONTROL,
&s->control_pid);
if (r < 0)
goto fail;
service_set_state(s, SERVICE_START_POST);
} else
service_enter_running(s, SERVICE_SUCCESS);
return;
fail:
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning_errno(UNIT(s), r, "Failed to run 'start-post' task: %m");
service_enter_stop(s, SERVICE_FAILURE_RESOURCES);
}
static void service_kill_control_process(Service *s) {
int r;
assert(s);
if (s->control_pid <= 0)
return;
r = kill_and_sigcont(s->control_pid, SIGKILL);
if (r < 0) {
_cleanup_free_ char *comm = NULL;
(void) get_process_comm(s->control_pid, &comm);
log_unit_debug_errno(UNIT(s), r, "Failed to kill control process " PID_FMT " (%s), ignoring: %m",
s->control_pid, strna(comm));
}
}
static void service_enter_start(Service *s) {
ExecCommand *c;
usec_t timeout;
pid_t pid;
int r;
assert(s);
service_unwatch_control_pid(s);
service_unwatch_main_pid(s);
unit_warn_leftover_processes(UNIT(s));
if (s->type == SERVICE_FORKING) {
s->control_command_id = SERVICE_EXEC_START;
c = s->control_command = s->exec_command[SERVICE_EXEC_START];
s->main_command = NULL;
} else {
s->control_command_id = _SERVICE_EXEC_COMMAND_INVALID;
s->control_command = NULL;
c = s->main_command = s->exec_command[SERVICE_EXEC_START];
}
if (!c) {
if (s->type != SERVICE_ONESHOT) {
/* There's no command line configured for the main command? Hmm, that is strange. This can only
* happen if the configuration changes at runtime. In this case, let's enter a failure
* state. */
log_unit_error(UNIT(s), "There's no 'start' task anymore we could start: %m");
r = -ENXIO;
goto fail;
}
service_enter_start_post(s);
return;
}
if (IN_SET(s->type, SERVICE_SIMPLE, SERVICE_IDLE))
/* For simple + idle this is the main process. We don't apply any timeout here, but
* service_enter_running() will later apply the .runtime_max_usec timeout. */
timeout = USEC_INFINITY;
else
timeout = s->timeout_start_usec;
r = service_spawn(s,
c,
timeout,
EXEC_PASS_FDS|EXEC_APPLY_SANDBOXING|EXEC_APPLY_CHROOT|EXEC_APPLY_TTY_STDIN|EXEC_SET_WATCHDOG,
&pid);
if (r < 0)
goto fail;
if (IN_SET(s->type, SERVICE_SIMPLE, SERVICE_IDLE)) {
/* For simple services we immediately start
* the START_POST binaries. */
service_set_main_pid(s, pid);
service_enter_start_post(s);
} else if (s->type == SERVICE_FORKING) {
/* For forking services we wait until the start
* process exited. */
s->control_pid = pid;
service_set_state(s, SERVICE_START);
} else if (IN_SET(s->type, SERVICE_ONESHOT, SERVICE_DBUS, SERVICE_NOTIFY)) {
2010-01-30 02:07:35 +01:00
/* For oneshot services we wait until the start
2010-01-30 02:07:35 +01:00
* process exited, too, but it is our main process. */
/* For D-Bus services we know the main pid right away,
* but wait for the bus name to appear on the
* bus. Notify services are similar. */
service_set_main_pid(s, pid);
service_set_state(s, SERVICE_START);
} else
assert_not_reached("Unknown service type");
return;
fail:
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning_errno(UNIT(s), r, "Failed to run 'start' task: %m");
service_enter_signal(s, SERVICE_STOP_SIGTERM, SERVICE_FAILURE_RESOURCES);
}
static void service_enter_start_pre(Service *s) {
int r;
assert(s);
service_unwatch_control_pid(s);
s->control_command = s->exec_command[SERVICE_EXEC_START_PRE];
if (s->control_command) {
unit_warn_leftover_processes(UNIT(s));
s->control_command_id = SERVICE_EXEC_START_PRE;
r = service_spawn(s,
s->control_command,
s->timeout_start_usec,
EXEC_APPLY_SANDBOXING|EXEC_APPLY_CHROOT|EXEC_IS_CONTROL|EXEC_APPLY_TTY_STDIN,
&s->control_pid);
if (r < 0)
goto fail;
service_set_state(s, SERVICE_START_PRE);
} else
service_enter_start(s);
return;
fail:
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning_errno(UNIT(s), r, "Failed to run 'start-pre' task: %m");
service_enter_dead(s, SERVICE_FAILURE_RESOURCES, true);
}
static void service_enter_restart(Service *s) {
_cleanup_(sd_bus_error_free) sd_bus_error error = SD_BUS_ERROR_NULL;
int r;
assert(s);
if (UNIT(s)->job && UNIT(s)->job->type == JOB_STOP) {
/* Don't restart things if we are going down anyway */
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_info(UNIT(s), "Stop job pending for unit, delaying automatic restart.");
r = service_arm_timer(s, usec_add(now(CLOCK_MONOTONIC), s->restart_usec));
if (r < 0)
goto fail;
return;
}
/* Any units that are bound to this service must also be
* restarted. We use JOB_RESTART (instead of the more obvious
* JOB_START) here so that those dependency jobs will be added
* as well. */
r = manager_add_job(UNIT(s)->manager, JOB_RESTART, UNIT(s), JOB_FAIL, &error, NULL);
if (r < 0)
goto fail;
/* Count the jobs we enqueue for restarting. This counter is maintained as long as the unit isn't fully
* stopped, i.e. as long as it remains up or remains in auto-start states. The use can reset the counter
* explicitly however via the usual "systemctl reset-failure" logic. */
s->n_restarts ++;
s->flush_n_restarts = false;
log_struct(LOG_INFO,
"MESSAGE_ID=" SD_MESSAGE_UNIT_RESTART_SCHEDULED_STR,
LOG_UNIT_ID(UNIT(s)),
LOG_UNIT_INVOCATION_ID(UNIT(s)),
LOG_UNIT_MESSAGE(UNIT(s), "Scheduled restart job, restart counter is at %u.", s->n_restarts),
"N_RESTARTS=%u", s->n_restarts,
NULL);
/* Notify clients about changed restart counter */
unit_add_to_dbus_queue(UNIT(s));
/* Note that we stay in the SERVICE_AUTO_RESTART state here,
* it will be canceled as part of the service_stop() call that
* is executed as part of JOB_RESTART. */
return;
fail:
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning(UNIT(s), "Failed to schedule restart job: %s", bus_error_message(&error, -r));
service_enter_dead(s, SERVICE_FAILURE_RESOURCES, false);
}
static void service_enter_reload_by_notify(Service *s) {
_cleanup_(sd_bus_error_free) sd_bus_error error = SD_BUS_ERROR_NULL;
int r;
assert(s);
service_arm_timer(s, usec_add(now(CLOCK_MONOTONIC), s->timeout_start_usec));
service_set_state(s, SERVICE_RELOAD);
/* service_enter_reload_by_notify is never called during a reload, thus no loops are possible. */
r = manager_propagate_reload(UNIT(s)->manager, UNIT(s), JOB_FAIL, &error);
if (r < 0)
log_unit_warning(UNIT(s), "Failed to schedule propagation of reload: %s", bus_error_message(&error, -r));
}
static void service_enter_reload(Service *s) {
int r;
assert(s);
service_unwatch_control_pid(s);
s->reload_result = SERVICE_SUCCESS;
s->control_command = s->exec_command[SERVICE_EXEC_RELOAD];
if (s->control_command) {
s->control_command_id = SERVICE_EXEC_RELOAD;
r = service_spawn(s,
s->control_command,
s->timeout_start_usec,
EXEC_APPLY_SANDBOXING|EXEC_APPLY_CHROOT|EXEC_IS_CONTROL,
&s->control_pid);
if (r < 0)
goto fail;
service_set_state(s, SERVICE_RELOAD);
} else
service_enter_running(s, SERVICE_SUCCESS);
return;
fail:
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning_errno(UNIT(s), r, "Failed to run 'reload' task: %m");
s->reload_result = SERVICE_FAILURE_RESOURCES;
service_enter_running(s, SERVICE_SUCCESS);
}
static void service_run_next_control(Service *s) {
usec_t timeout;
int r;
assert(s);
assert(s->control_command);
assert(s->control_command->command_next);
assert(s->control_command_id != SERVICE_EXEC_START);
s->control_command = s->control_command->command_next;
service_unwatch_control_pid(s);
if (IN_SET(s->state, SERVICE_START_PRE, SERVICE_START, SERVICE_START_POST, SERVICE_RUNNING, SERVICE_RELOAD))
timeout = s->timeout_start_usec;
else
timeout = s->timeout_stop_usec;
r = service_spawn(s,
s->control_command,
timeout,
EXEC_APPLY_SANDBOXING|EXEC_APPLY_CHROOT|EXEC_IS_CONTROL|
(IN_SET(s->control_command_id, SERVICE_EXEC_START_PRE, SERVICE_EXEC_STOP_POST) ? EXEC_APPLY_TTY_STDIN : 0)|
(IN_SET(s->control_command_id, SERVICE_EXEC_STOP, SERVICE_EXEC_STOP_POST) ? EXEC_SETENV_RESULT : 0),
&s->control_pid);
if (r < 0)
goto fail;
return;
fail:
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning_errno(UNIT(s), r, "Failed to run next control task: %m");
if (IN_SET(s->state, SERVICE_START_PRE, SERVICE_START_POST, SERVICE_STOP))
service_enter_signal(s, SERVICE_STOP_SIGTERM, SERVICE_FAILURE_RESOURCES);
else if (s->state == SERVICE_STOP_POST)
service_enter_dead(s, SERVICE_FAILURE_RESOURCES, true);
else if (s->state == SERVICE_RELOAD) {
s->reload_result = SERVICE_FAILURE_RESOURCES;
service_enter_running(s, SERVICE_SUCCESS);
} else
service_enter_stop(s, SERVICE_FAILURE_RESOURCES);
}
static void service_run_next_main(Service *s) {
pid_t pid;
int r;
assert(s);
assert(s->main_command);
assert(s->main_command->command_next);
assert(s->type == SERVICE_ONESHOT);
s->main_command = s->main_command->command_next;
service_unwatch_main_pid(s);
r = service_spawn(s,
s->main_command,
s->timeout_start_usec,
EXEC_PASS_FDS|EXEC_APPLY_SANDBOXING|EXEC_APPLY_CHROOT|EXEC_APPLY_TTY_STDIN|EXEC_SET_WATCHDOG,
&pid);
if (r < 0)
goto fail;
service_set_main_pid(s, pid);
return;
fail:
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning_errno(UNIT(s), r, "Failed to run next main task: %m");
service_enter_stop(s, SERVICE_FAILURE_RESOURCES);
}
2010-01-26 21:39:06 +01:00
static int service_start(Unit *u) {
Service *s = SERVICE(u);
int r;
assert(s);
/* We cannot fulfill this request right now, try again later
* please! */
if (IN_SET(s->state,
SERVICE_STOP, SERVICE_STOP_SIGABRT, SERVICE_STOP_SIGTERM, SERVICE_STOP_SIGKILL, SERVICE_STOP_POST,
SERVICE_FINAL_SIGTERM, SERVICE_FINAL_SIGKILL))
return -EAGAIN;
/* Already on it! */
if (IN_SET(s->state, SERVICE_START_PRE, SERVICE_START, SERVICE_START_POST))
return 0;
/* A service that will be restarted must be stopped first to
* trigger BindsTo and/or OnFailure dependencies. If a user
* does not want to wait for the holdoff time to elapse, the
* service should be manually restarted, not started. We
* simply return EAGAIN here, so that any start jobs stay
* queued, and assume that the auto restart timer will
* eventually trigger the restart. */
if (s->state == SERVICE_AUTO_RESTART)
return -EAGAIN;
assert(IN_SET(s->state, SERVICE_DEAD, SERVICE_FAILED));
/* Make sure we don't enter a busy loop of some kind. */
r = unit_start_limit_test(u);
if (r < 0) {
service_enter_dead(s, SERVICE_FAILURE_START_LIMIT_HIT, false);
return r;
}
core: add "invocation ID" concept to service manager This adds a new invocation ID concept to the service manager. The invocation ID identifies each runtime cycle of a unit uniquely. A new randomized 128bit ID is generated each time a unit moves from and inactive to an activating or active state. The primary usecase for this concept is to connect the runtime data PID 1 maintains about a service with the offline data the journal stores about it. Previously we'd use the unit name plus start/stop times, which however is highly racy since the journal will generally process log data after the service already ended. The "invocation ID" kinda matches the "boot ID" concept of the Linux kernel, except that it applies to an individual unit instead of the whole system. The invocation ID is passed to the activated processes as environment variable. It is additionally stored as extended attribute on the cgroup of the unit. The latter is used by journald to automatically retrieve it for each log logged message and attach it to the log entry. The environment variable is very easily accessible, even for unprivileged services. OTOH the extended attribute is only accessible to privileged processes (this is because cgroupfs only supports the "trusted." xattr namespace, not "user."). The environment variable may be altered by services, the extended attribute may not be, hence is the better choice for the journal. Note that reading the invocation ID off the extended attribute from journald is racy, similar to the way reading the unit name for a logging process is. This patch adds APIs to read the invocation ID to sd-id128: sd_id128_get_invocation() may be used in a similar fashion to sd_id128_get_boot(). PID1's own logging is updated to always include the invocation ID when it logs information about a unit. A new bus call GetUnitByInvocationID() is added that allows retrieving a bus path to a unit by its invocation ID. The bus path is built using the invocation ID, thus providing a path for referring to a unit that is valid only for the current runtime cycleof it. Outlook for the future: should the kernel eventually allow passing of cgroup information along AF_UNIX/SOCK_DGRAM messages via a unique cgroup id, then we can alter the invocation ID to be generated as hash from that rather than entirely randomly. This way we can derive the invocation race-freely from the messages.
2016-08-30 23:18:46 +02:00
r = unit_acquire_invocation_id(u);
if (r < 0)
return r;
s->result = SERVICE_SUCCESS;
s->reload_result = SERVICE_SUCCESS;
s->main_pid_known = false;
s->main_pid_alien = false;
s->forbid_restart = false;
u->reset_accounting = true;
s->status_text = mfree(s->status_text);
s->status_errno = 0;
s->notify_state = NOTIFY_UNKNOWN;
s->watchdog_override_enable = false;
s->watchdog_override_usec = 0;
/* This is not an automatic restart? Flush the restart counter then */
if (s->flush_n_restarts) {
s->n_restarts = 0;
s->flush_n_restarts = false;
}
service_enter_start_pre(s);
return 1;
}
2010-01-26 21:39:06 +01:00
static int service_stop(Unit *u) {
Service *s = SERVICE(u);
assert(s);
/* Don't create restart jobs from manual stops. */
s->forbid_restart = true;
/* Already on it */
if (IN_SET(s->state,
SERVICE_STOP, SERVICE_STOP_SIGABRT, SERVICE_STOP_SIGTERM, SERVICE_STOP_SIGKILL, SERVICE_STOP_POST,
SERVICE_FINAL_SIGTERM, SERVICE_FINAL_SIGKILL))
return 0;
/* A restart will be scheduled or is in progress. */
if (s->state == SERVICE_AUTO_RESTART) {
service_set_state(s, SERVICE_DEAD);
return 0;
}
/* If there's already something running we go directly into
* kill mode. */
if (IN_SET(s->state, SERVICE_START_PRE, SERVICE_START, SERVICE_START_POST, SERVICE_RELOAD)) {
service_enter_signal(s, SERVICE_STOP_SIGTERM, SERVICE_SUCCESS);
return 0;
}
assert(IN_SET(s->state, SERVICE_RUNNING, SERVICE_EXITED));
service_enter_stop(s, SERVICE_SUCCESS);
return 1;
}
2010-01-26 21:39:06 +01:00
static int service_reload(Unit *u) {
Service *s = SERVICE(u);
assert(s);
assert(IN_SET(s->state, SERVICE_RUNNING, SERVICE_EXITED));
service_enter_reload(s);
return 1;
}
_pure_ static bool service_can_reload(Unit *u) {
2010-01-26 21:39:06 +01:00
Service *s = SERVICE(u);
assert(s);
return !!s->exec_command[SERVICE_EXEC_RELOAD];
}
service: serialize information about currently executing command Stored information will help us to resume execution after the daemon-reload. This commit implements following scheme, * On serialization: - we count rank of the currently executing command - we store command type, its rank and command line arguments * On deserialization: - configuration is parsed and loaded - we deserialize stored data, command type, rank and arguments - we look at the given rank in the list and if command there has same arguments then we restore execution at that point - otherwise we search respective command list and we look for command that has the same arguments - if both methods fail we do not do not resume execution at all To better illustrate how does above scheme works, please consider following cases (<<< denotes position where we resume execution after reload) ; Original unit file [Service] ExecStart=/bin/true <<< ExecStart=/bin/false ; Swapped commands ; Second command is not going to be executed [Service] ExecStart=/bin/false ExecStart=/bin/true <<< ; Commands added before ; Same commands are problematic and execution could be restarted at wrong place [Service] ExecStart=/bin/foo ExecStart=/bin/bar ExecStart=/bin/true <<< ExecStart=/bin/false ; Commands added after ; Same commands are not an issue in this case [Service] ExecStart=/bin/true <<< ExecStart=/bin/false ExecStart=/bin/foo ExecStart=/bin/bar ; New commands interleaved with old commands ; Some new commands will be executed while others won't ExecStart=/bin/foo ExecStart=/bin/true <<< ExecStart=/bin/bar ExecStart=/bin/false As you can see, above scheme has some drawbacks. However, in most cases (we assume that in most common case unit file command list is not changed while some other command is running for the same unit) it should cause that systemd does the right thing, which is restoring execution exactly at the point we were before daemon-reload. Fixes #518
2017-01-23 17:12:35 +01:00
static unsigned service_exec_command_index(Unit *u, ServiceExecCommand id, ExecCommand *current) {
Service *s = SERVICE(u);
unsigned idx = 0;
ExecCommand *first, *c;
assert(s);
first = s->exec_command[id];
/* Figure out where we are in the list by walking back to the beginning */
for (c = current; c != first; c = c->command_prev)
idx++;
return idx;
}
static int service_serialize_exec_command(Unit *u, FILE *f, ExecCommand *command) {
Service *s = SERVICE(u);
ServiceExecCommand id;
unsigned idx;
const char *type;
char **arg;
_cleanup_free_ char *args = NULL, *p = NULL;
size_t allocated = 0, length = 0;
assert(s);
assert(f);
if (!command)
return 0;
if (command == s->control_command) {
type = "control";
id = s->control_command_id;
} else {
type = "main";
id = SERVICE_EXEC_START;
}
idx = service_exec_command_index(u, id, command);
STRV_FOREACH(arg, command->argv) {
size_t n;
_cleanup_free_ char *e = NULL;
e = xescape(*arg, WHITESPACE);
if (!e)
return -ENOMEM;
n = strlen(e);
if (!GREEDY_REALLOC(args, allocated, length + 1 + n + 1))
return -ENOMEM;
if (length > 0)
args[length++] = ' ';
memcpy(args + length, e, n);
length += n;
}
if (!GREEDY_REALLOC(args, allocated, length + 1))
return -ENOMEM;
args[length++] = 0;
p = xescape(command->path, WHITESPACE);
if (!p)
return -ENOMEM;
fprintf(f, "%s-command=%s %u %s %s\n", type, service_exec_command_to_string(id), idx, p, args);
return 0;
}
2010-04-21 03:27:44 +02:00
static int service_serialize(Unit *u, FILE *f, FDSet *fds) {
Service *s = SERVICE(u);
ServiceFDStore *fs;
int r;
2010-04-21 03:27:44 +02:00
assert(u);
assert(f);
assert(fds);
unit_serialize_item(u, f, "state", service_state_to_string(s->state));
unit_serialize_item(u, f, "result", service_result_to_string(s->result));
unit_serialize_item(u, f, "reload-result", service_result_to_string(s->reload_result));
2010-04-21 03:27:44 +02:00
if (s->control_pid > 0)
unit_serialize_item_format(u, f, "control-pid", PID_FMT, s->control_pid);
2010-04-21 03:27:44 +02:00
if (s->main_pid_known && s->main_pid > 0)
unit_serialize_item_format(u, f, "main-pid", PID_FMT, s->main_pid);
2010-04-21 03:27:44 +02:00
unit_serialize_item(u, f, "main-pid-known", yes_no(s->main_pid_known));
unit_serialize_item(u, f, "bus-name-good", yes_no(s->bus_name_good));
unit_serialize_item(u, f, "bus-name-owner", s->bus_name_owner);
2010-04-21 03:27:44 +02:00
unit_serialize_item_format(u, f, "n-restarts", "%u", s->n_restarts);
unit_serialize_item(u, f, "flush-n-restarts", yes_no(s->flush_n_restarts));
r = unit_serialize_item_escaped(u, f, "status-text", s->status_text);
if (r < 0)
return r;
service: serialize information about currently executing command Stored information will help us to resume execution after the daemon-reload. This commit implements following scheme, * On serialization: - we count rank of the currently executing command - we store command type, its rank and command line arguments * On deserialization: - configuration is parsed and loaded - we deserialize stored data, command type, rank and arguments - we look at the given rank in the list and if command there has same arguments then we restore execution at that point - otherwise we search respective command list and we look for command that has the same arguments - if both methods fail we do not do not resume execution at all To better illustrate how does above scheme works, please consider following cases (<<< denotes position where we resume execution after reload) ; Original unit file [Service] ExecStart=/bin/true <<< ExecStart=/bin/false ; Swapped commands ; Second command is not going to be executed [Service] ExecStart=/bin/false ExecStart=/bin/true <<< ; Commands added before ; Same commands are problematic and execution could be restarted at wrong place [Service] ExecStart=/bin/foo ExecStart=/bin/bar ExecStart=/bin/true <<< ExecStart=/bin/false ; Commands added after ; Same commands are not an issue in this case [Service] ExecStart=/bin/true <<< ExecStart=/bin/false ExecStart=/bin/foo ExecStart=/bin/bar ; New commands interleaved with old commands ; Some new commands will be executed while others won't ExecStart=/bin/foo ExecStart=/bin/true <<< ExecStart=/bin/bar ExecStart=/bin/false As you can see, above scheme has some drawbacks. However, in most cases (we assume that in most common case unit file command list is not changed while some other command is running for the same unit) it should cause that systemd does the right thing, which is restoring execution exactly at the point we were before daemon-reload. Fixes #518
2017-01-23 17:12:35 +01:00
service_serialize_exec_command(u, f, s->control_command);
service_serialize_exec_command(u, f, s->main_command);
2010-04-21 03:27:44 +02:00
r = unit_serialize_item_fd(u, f, fds, "stdin-fd", s->stdin_fd);
if (r < 0)
return r;
r = unit_serialize_item_fd(u, f, fds, "stdout-fd", s->stdout_fd);
if (r < 0)
return r;
r = unit_serialize_item_fd(u, f, fds, "stderr-fd", s->stderr_fd);
if (r < 0)
return r;
if (UNIT_ISSET(s->accept_socket)) {
r = unit_serialize_item(u, f, "accept-socket", UNIT_DEREF(s->accept_socket)->id);
if (r < 0)
return r;
}
r = unit_serialize_item_fd(u, f, fds, "socket-fd", s->socket_fd);
if (r < 0)
return r;
LIST_FOREACH(fd_store, fs, s->fd_store) {
_cleanup_free_ char *c = NULL;
int copy;
copy = fdset_put_dup(fds, fs->fd);
if (copy < 0)
return copy;
c = cescape(fs->fdname);
unit_serialize_item_format(u, f, "fd-store-fd", "%i %s", copy, strempty(c));
}
2010-07-07 04:12:59 +02:00
if (s->main_exec_status.pid > 0) {
unit_serialize_item_format(u, f, "main-exec-status-pid", PID_FMT, s->main_exec_status.pid);
dual_timestamp_serialize(f, "main-exec-status-start", &s->main_exec_status.start_timestamp);
dual_timestamp_serialize(f, "main-exec-status-exit", &s->main_exec_status.exit_timestamp);
2010-07-07 04:12:59 +02:00
if (dual_timestamp_is_set(&s->main_exec_status.exit_timestamp)) {
unit_serialize_item_format(u, f, "main-exec-status-code", "%i", s->main_exec_status.code);
unit_serialize_item_format(u, f, "main-exec-status-status", "%i", s->main_exec_status.status);
2010-07-07 04:12:59 +02:00
}
}
dual_timestamp_serialize(f, "watchdog-timestamp", &s->watchdog_timestamp);
2010-07-07 04:12:59 +02:00
unit_serialize_item(u, f, "forbid-restart", yes_no(s->forbid_restart));
if (s->watchdog_override_enable)
unit_serialize_item_format(u, f, "watchdog-override-usec", USEC_FMT, s->watchdog_override_usec);
2010-04-21 03:27:44 +02:00
return 0;
}
service: serialize information about currently executing command Stored information will help us to resume execution after the daemon-reload. This commit implements following scheme, * On serialization: - we count rank of the currently executing command - we store command type, its rank and command line arguments * On deserialization: - configuration is parsed and loaded - we deserialize stored data, command type, rank and arguments - we look at the given rank in the list and if command there has same arguments then we restore execution at that point - otherwise we search respective command list and we look for command that has the same arguments - if both methods fail we do not do not resume execution at all To better illustrate how does above scheme works, please consider following cases (<<< denotes position where we resume execution after reload) ; Original unit file [Service] ExecStart=/bin/true <<< ExecStart=/bin/false ; Swapped commands ; Second command is not going to be executed [Service] ExecStart=/bin/false ExecStart=/bin/true <<< ; Commands added before ; Same commands are problematic and execution could be restarted at wrong place [Service] ExecStart=/bin/foo ExecStart=/bin/bar ExecStart=/bin/true <<< ExecStart=/bin/false ; Commands added after ; Same commands are not an issue in this case [Service] ExecStart=/bin/true <<< ExecStart=/bin/false ExecStart=/bin/foo ExecStart=/bin/bar ; New commands interleaved with old commands ; Some new commands will be executed while others won't ExecStart=/bin/foo ExecStart=/bin/true <<< ExecStart=/bin/bar ExecStart=/bin/false As you can see, above scheme has some drawbacks. However, in most cases (we assume that in most common case unit file command list is not changed while some other command is running for the same unit) it should cause that systemd does the right thing, which is restoring execution exactly at the point we were before daemon-reload. Fixes #518
2017-01-23 17:12:35 +01:00
static int service_deserialize_exec_command(Unit *u, const char *key, const char *value) {
Service *s = SERVICE(u);
int r;
unsigned idx = 0, i;
bool control, found = false;
ServiceExecCommand id = _SERVICE_EXEC_COMMAND_INVALID;
ExecCommand *command = NULL;
2017-04-28 23:49:50 +02:00
_cleanup_free_ char *path = NULL;
service: serialize information about currently executing command Stored information will help us to resume execution after the daemon-reload. This commit implements following scheme, * On serialization: - we count rank of the currently executing command - we store command type, its rank and command line arguments * On deserialization: - configuration is parsed and loaded - we deserialize stored data, command type, rank and arguments - we look at the given rank in the list and if command there has same arguments then we restore execution at that point - otherwise we search respective command list and we look for command that has the same arguments - if both methods fail we do not do not resume execution at all To better illustrate how does above scheme works, please consider following cases (<<< denotes position where we resume execution after reload) ; Original unit file [Service] ExecStart=/bin/true <<< ExecStart=/bin/false ; Swapped commands ; Second command is not going to be executed [Service] ExecStart=/bin/false ExecStart=/bin/true <<< ; Commands added before ; Same commands are problematic and execution could be restarted at wrong place [Service] ExecStart=/bin/foo ExecStart=/bin/bar ExecStart=/bin/true <<< ExecStart=/bin/false ; Commands added after ; Same commands are not an issue in this case [Service] ExecStart=/bin/true <<< ExecStart=/bin/false ExecStart=/bin/foo ExecStart=/bin/bar ; New commands interleaved with old commands ; Some new commands will be executed while others won't ExecStart=/bin/foo ExecStart=/bin/true <<< ExecStart=/bin/bar ExecStart=/bin/false As you can see, above scheme has some drawbacks. However, in most cases (we assume that in most common case unit file command list is not changed while some other command is running for the same unit) it should cause that systemd does the right thing, which is restoring execution exactly at the point we were before daemon-reload. Fixes #518
2017-01-23 17:12:35 +01:00
_cleanup_strv_free_ char **argv = NULL;
enum ExecCommandState {
STATE_EXEC_COMMAND_TYPE,
STATE_EXEC_COMMAND_INDEX,
STATE_EXEC_COMMAND_PATH,
STATE_EXEC_COMMAND_ARGS,
_STATE_EXEC_COMMAND_MAX,
_STATE_EXEC_COMMAND_INVALID = -1,
} state;
assert(s);
assert(key);
assert(value);
control = streq(key, "control-command");
state = STATE_EXEC_COMMAND_TYPE;
for (;;) {
_cleanup_free_ char *arg = NULL;
r = extract_first_word(&value, &arg, NULL, EXTRACT_CUNESCAPE);
if (r == 0)
break;
else if (r < 0)
return r;
switch (state) {
case STATE_EXEC_COMMAND_TYPE:
id = service_exec_command_from_string(arg);
if (id < 0)
return -EINVAL;
state = STATE_EXEC_COMMAND_INDEX;
break;
case STATE_EXEC_COMMAND_INDEX:
r = safe_atou(arg, &idx);
if (r < 0)
return -EINVAL;
state = STATE_EXEC_COMMAND_PATH;
break;
case STATE_EXEC_COMMAND_PATH:
path = TAKE_PTR(arg);
service: serialize information about currently executing command Stored information will help us to resume execution after the daemon-reload. This commit implements following scheme, * On serialization: - we count rank of the currently executing command - we store command type, its rank and command line arguments * On deserialization: - configuration is parsed and loaded - we deserialize stored data, command type, rank and arguments - we look at the given rank in the list and if command there has same arguments then we restore execution at that point - otherwise we search respective command list and we look for command that has the same arguments - if both methods fail we do not do not resume execution at all To better illustrate how does above scheme works, please consider following cases (<<< denotes position where we resume execution after reload) ; Original unit file [Service] ExecStart=/bin/true <<< ExecStart=/bin/false ; Swapped commands ; Second command is not going to be executed [Service] ExecStart=/bin/false ExecStart=/bin/true <<< ; Commands added before ; Same commands are problematic and execution could be restarted at wrong place [Service] ExecStart=/bin/foo ExecStart=/bin/bar ExecStart=/bin/true <<< ExecStart=/bin/false ; Commands added after ; Same commands are not an issue in this case [Service] ExecStart=/bin/true <<< ExecStart=/bin/false ExecStart=/bin/foo ExecStart=/bin/bar ; New commands interleaved with old commands ; Some new commands will be executed while others won't ExecStart=/bin/foo ExecStart=/bin/true <<< ExecStart=/bin/bar ExecStart=/bin/false As you can see, above scheme has some drawbacks. However, in most cases (we assume that in most common case unit file command list is not changed while some other command is running for the same unit) it should cause that systemd does the right thing, which is restoring execution exactly at the point we were before daemon-reload. Fixes #518
2017-01-23 17:12:35 +01:00
state = STATE_EXEC_COMMAND_ARGS;
if (!path_is_absolute(path))
return -EINVAL;
break;
case STATE_EXEC_COMMAND_ARGS:
r = strv_extend(&argv, arg);
if (r < 0)
return -ENOMEM;
break;
default:
assert_not_reached("Unknown error at deserialization of exec command");
break;
}
}
if (state != STATE_EXEC_COMMAND_ARGS)
return -EINVAL;
/* Let's check whether exec command on given offset matches data that we just deserialized */
for (command = s->exec_command[id], i = 0; command; command = command->command_next, i++) {
if (i != idx)
continue;
found = strv_equal(argv, command->argv) && streq(command->path, path);
break;
}
if (!found) {
/* Command at the index we serialized is different, let's look for command that exactly
* matches but is on different index. If there is no such command we will not resume execution. */
for (command = s->exec_command[id]; command; command = command->command_next)
if (strv_equal(command->argv, argv) && streq(command->path, path))
break;
}
if (command && control)
s->control_command = command;
else if (command)
s->main_command = command;
else
log_unit_warning(u, "Current command vanished from the unit file, execution of the command list won't be resumed.");
return 0;
}
2010-04-21 03:27:44 +02:00
static int service_deserialize_item(Unit *u, const char *key, const char *value, FDSet *fds) {
Service *s = SERVICE(u);
int r;
2010-04-21 03:27:44 +02:00
assert(u);
assert(key);
assert(value);
assert(fds);
if (streq(key, "state")) {
ServiceState state;
state = service_state_from_string(value);
if (state < 0)
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_debug(u, "Failed to parse state value: %s", value);
2010-04-21 03:27:44 +02:00
else
s->deserialized_state = state;
} else if (streq(key, "result")) {
ServiceResult f;
f = service_result_from_string(value);
if (f < 0)
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_debug(u, "Failed to parse result value: %s", value);
else if (f != SERVICE_SUCCESS)
s->result = f;
} else if (streq(key, "reload-result")) {
ServiceResult f;
f = service_result_from_string(value);
if (f < 0)
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_debug(u, "Failed to parse reload result value: %s", value);
else if (f != SERVICE_SUCCESS)
s->reload_result = f;
2010-04-21 03:27:44 +02:00
} else if (streq(key, "control-pid")) {
pid_t pid;
2010-04-21 03:27:44 +02:00
if (parse_pid(value, &pid) < 0)
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_debug(u, "Failed to parse control-pid value: %s", value);
2010-04-21 03:27:44 +02:00
else
s->control_pid = pid;
2010-04-21 03:27:44 +02:00
} else if (streq(key, "main-pid")) {
pid_t pid;
2010-04-21 03:27:44 +02:00
if (parse_pid(value, &pid) < 0)
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_debug(u, "Failed to parse main-pid value: %s", value);
else
(void) service_set_main_pid(s, pid);
2010-04-21 03:27:44 +02:00
} else if (streq(key, "main-pid-known")) {
int b;
b = parse_boolean(value);
if (b < 0)
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_debug(u, "Failed to parse main-pid-known value: %s", value);
2010-04-21 03:27:44 +02:00
else
s->main_pid_known = b;
} else if (streq(key, "bus-name-good")) {
int b;
b = parse_boolean(value);
if (b < 0)
log_unit_debug(u, "Failed to parse bus-name-good value: %s", value);
else
s->bus_name_good = b;
} else if (streq(key, "bus-name-owner")) {
r = free_and_strdup(&s->bus_name_owner, value);
if (r < 0)
log_unit_error_errno(u, r, "Unable to deserialize current bus owner %s: %m", value);
} else if (streq(key, "status-text")) {
char *t;
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
r = cunescape(value, 0, &t);
if (r < 0)
log_unit_debug_errno(u, r, "Failed to unescape status text: %s", value);
else {
free(s->status_text);
s->status_text = t;
}
} else if (streq(key, "accept-socket")) {
Unit *socket;
r = manager_load_unit(u->manager, value, NULL, NULL, &socket);
if (r < 0)
log_unit_debug_errno(u, r, "Failed to load accept-socket unit: %s", value);
else {
unit_ref_set(&s->accept_socket, u, socket);
SOCKET(socket)->n_connections++;
}
2010-04-21 03:27:44 +02:00
} else if (streq(key, "socket-fd")) {
int fd;
if (safe_atoi(value, &fd) < 0 || fd < 0 || !fdset_contains(fds, fd))
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_debug(u, "Failed to parse socket-fd value: %s", value);
2010-04-21 03:27:44 +02:00
else {
asynchronous_close(s->socket_fd);
2010-04-21 03:27:44 +02:00
s->socket_fd = fdset_remove(fds, fd);
}
} else if (streq(key, "fd-store-fd")) {
const char *fdv;
size_t pf;
int fd;
pf = strcspn(value, WHITESPACE);
fdv = strndupa(value, pf);
if (safe_atoi(fdv, &fd) < 0 || fd < 0 || !fdset_contains(fds, fd))
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_debug(u, "Failed to parse fd-store-fd value: %s", value);
else {
_cleanup_free_ char *t = NULL;
const char *fdn;
fdn = value + pf;
fdn += strspn(fdn, WHITESPACE);
(void) cunescape(fdn, 0, &t);
r = service_add_fd_store(s, fd, t);
if (r < 0)
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_error_errno(u, r, "Failed to add fd to store: %m");
else
fdset_remove(fds, fd);
}
2010-07-07 04:12:59 +02:00
} else if (streq(key, "main-exec-status-pid")) {
pid_t pid;
if (parse_pid(value, &pid) < 0)
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_debug(u, "Failed to parse main-exec-status-pid value: %s", value);
2010-07-07 04:12:59 +02:00
else
s->main_exec_status.pid = pid;
} else if (streq(key, "main-exec-status-code")) {
int i;
if (safe_atoi(value, &i) < 0)
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_debug(u, "Failed to parse main-exec-status-code value: %s", value);
2010-07-07 04:12:59 +02:00
else
s->main_exec_status.code = i;
} else if (streq(key, "main-exec-status-status")) {
int i;
if (safe_atoi(value, &i) < 0)
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_debug(u, "Failed to parse main-exec-status-status value: %s", value);
2010-07-07 04:12:59 +02:00
else
s->main_exec_status.status = i;
} else if (streq(key, "main-exec-status-start"))
dual_timestamp_deserialize(value, &s->main_exec_status.start_timestamp);
else if (streq(key, "main-exec-status-exit"))
dual_timestamp_deserialize(value, &s->main_exec_status.exit_timestamp);
else if (streq(key, "watchdog-timestamp"))
dual_timestamp_deserialize(value, &s->watchdog_timestamp);
else if (streq(key, "forbid-restart")) {
int b;
b = parse_boolean(value);
if (b < 0)
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_debug(u, "Failed to parse forbid-restart value: %s", value);
else
s->forbid_restart = b;
} else if (streq(key, "stdin-fd")) {
int fd;
if (safe_atoi(value, &fd) < 0 || fd < 0 || !fdset_contains(fds, fd))
log_unit_debug(u, "Failed to parse stdin-fd value: %s", value);
else {
asynchronous_close(s->stdin_fd);
s->stdin_fd = fdset_remove(fds, fd);
s->exec_context.stdio_as_fds = true;
}
} else if (streq(key, "stdout-fd")) {
int fd;
if (safe_atoi(value, &fd) < 0 || fd < 0 || !fdset_contains(fds, fd))
log_unit_debug(u, "Failed to parse stdout-fd value: %s", value);
else {
asynchronous_close(s->stdout_fd);
s->stdout_fd = fdset_remove(fds, fd);
s->exec_context.stdio_as_fds = true;
}
} else if (streq(key, "stderr-fd")) {
int fd;
if (safe_atoi(value, &fd) < 0 || fd < 0 || !fdset_contains(fds, fd))
log_unit_debug(u, "Failed to parse stderr-fd value: %s", value);
else {
asynchronous_close(s->stderr_fd);
s->stderr_fd = fdset_remove(fds, fd);
s->exec_context.stdio_as_fds = true;
}
} else if (streq(key, "watchdog-override-usec")) {
usec_t watchdog_override_usec;
if (timestamp_deserialize(value, &watchdog_override_usec) < 0)
log_unit_debug(u, "Failed to parse watchdog_override_usec value: %s", value);
else {
s->watchdog_override_enable = true;
s->watchdog_override_usec = watchdog_override_usec;
}
service: serialize information about currently executing command Stored information will help us to resume execution after the daemon-reload. This commit implements following scheme, * On serialization: - we count rank of the currently executing command - we store command type, its rank and command line arguments * On deserialization: - configuration is parsed and loaded - we deserialize stored data, command type, rank and arguments - we look at the given rank in the list and if command there has same arguments then we restore execution at that point - otherwise we search respective command list and we look for command that has the same arguments - if both methods fail we do not do not resume execution at all To better illustrate how does above scheme works, please consider following cases (<<< denotes position where we resume execution after reload) ; Original unit file [Service] ExecStart=/bin/true <<< ExecStart=/bin/false ; Swapped commands ; Second command is not going to be executed [Service] ExecStart=/bin/false ExecStart=/bin/true <<< ; Commands added before ; Same commands are problematic and execution could be restarted at wrong place [Service] ExecStart=/bin/foo ExecStart=/bin/bar ExecStart=/bin/true <<< ExecStart=/bin/false ; Commands added after ; Same commands are not an issue in this case [Service] ExecStart=/bin/true <<< ExecStart=/bin/false ExecStart=/bin/foo ExecStart=/bin/bar ; New commands interleaved with old commands ; Some new commands will be executed while others won't ExecStart=/bin/foo ExecStart=/bin/true <<< ExecStart=/bin/bar ExecStart=/bin/false As you can see, above scheme has some drawbacks. However, in most cases (we assume that in most common case unit file command list is not changed while some other command is running for the same unit) it should cause that systemd does the right thing, which is restoring execution exactly at the point we were before daemon-reload. Fixes #518
2017-01-23 17:12:35 +01:00
} else if (STR_IN_SET(key, "main-command", "control-command")) {
r = service_deserialize_exec_command(u, key, value);
if (r < 0)
log_unit_debug_errno(u, r, "Failed to parse serialized command \"%s\": %m", value);
} else if (streq(key, "n-restarts")) {
r = safe_atou(value, &s->n_restarts);
if (r < 0)
log_unit_debug_errno(u, r, "Failed to parse serialized restart counter '%s': %m", value);
} else if (streq(key, "flush-n-restarts")) {
r = parse_boolean(value);
if (r < 0)
log_unit_debug_errno(u, r, "Failed to parse serialized flush restart counter setting '%s': %m", value);
else
s->flush_n_restarts = r;
} else
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_debug(u, "Unknown serialization key: %s", key);
2010-04-21 03:27:44 +02:00
return 0;
}
_pure_ static UnitActiveState service_active_state(Unit *u) {
const UnitActiveState *table;
2010-01-26 21:39:06 +01:00
assert(u);
table = SERVICE(u)->type == SERVICE_IDLE ? state_translation_table_idle : state_translation_table;
return table[SERVICE(u)->state];
}
static const char *service_sub_state_to_string(Unit *u) {
assert(u);
return service_state_to_string(SERVICE(u)->state);
}
static bool service_may_gc(Unit *u) {
Service *s = SERVICE(u);
assert(s);
/* Never clean up services that still have a process around, even if the service is formally dead. Note that
* unit_may_gc() already checked our cgroup for us, we just check our two additional PIDs, too, in case they
* have moved outside of the cgroup. */
if (main_pid_good(s) > 0 ||
control_pid_good(s) > 0)
return false;
return true;
}
static int service_retry_pid_file(Service *s) {
int r;
assert(s->pid_file);
assert(IN_SET(s->state, SERVICE_START, SERVICE_START_POST));
r = service_load_pid_file(s, false);
if (r < 0)
return r;
service_unwatch_pid_file(s);
service_enter_running(s, SERVICE_SUCCESS);
return 0;
}
static int service_watch_pid_file(Service *s) {
int r;
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_debug(UNIT(s), "Setting watch for PID file %s", s->pid_file_pathspec->path);
r = path_spec_watch(s->pid_file_pathspec, service_dispatch_io);
if (r < 0)
goto fail;
/* the pidfile might have appeared just before we set the watch */
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_debug(UNIT(s), "Trying to read PID file %s in case it changed", s->pid_file_pathspec->path);
service_retry_pid_file(s);
return 0;
fail:
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_error_errno(UNIT(s), r, "Failed to set a watch for PID file %s: %m", s->pid_file_pathspec->path);
service_unwatch_pid_file(s);
return r;
}
static int service_demand_pid_file(Service *s) {
PathSpec *ps;
assert(s->pid_file);
assert(!s->pid_file_pathspec);
ps = new0(PathSpec, 1);
if (!ps)
return -ENOMEM;
ps->unit = UNIT(s);
ps->path = strdup(s->pid_file);
if (!ps->path) {
free(ps);
return -ENOMEM;
}
path_kill_slashes(ps->path);
/* PATH_CHANGED would not be enough. There are daemons (sendmail) that
* keep their PID file open all the time. */
ps->type = PATH_MODIFIED;
ps->inotify_fd = -1;
s->pid_file_pathspec = ps;
return service_watch_pid_file(s);
}
static int service_dispatch_io(sd_event_source *source, int fd, uint32_t events, void *userdata) {
PathSpec *p = userdata;
Service *s;
assert(p);
s = SERVICE(p->unit);
assert(s);
assert(fd >= 0);
assert(IN_SET(s->state, SERVICE_START, SERVICE_START_POST));
assert(s->pid_file_pathspec);
assert(path_spec_owns_inotify_fd(s->pid_file_pathspec, fd));
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_debug(UNIT(s), "inotify event");
if (path_spec_fd_event(p, events) < 0)
goto fail;
if (service_retry_pid_file(s) == 0)
return 0;
if (service_watch_pid_file(s) < 0)
goto fail;
return 0;
fail:
service_unwatch_pid_file(s);
service_enter_signal(s, SERVICE_STOP_SIGTERM, SERVICE_FAILURE_RESOURCES);
return 0;
}
static void service_notify_cgroup_empty_event(Unit *u) {
Service *s = SERVICE(u);
assert(u);
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_debug(u, "cgroup is empty");
switch (s->state) {
/* Waiting for SIGCHLD is usually more interesting,
* because it includes return codes/signals. Which is
* why we ignore the cgroup events for most cases,
* except when we don't know pid which to expect the
* SIGCHLD for. */
case SERVICE_START:
if (s->type == SERVICE_NOTIFY &&
main_pid_good(s) == 0 &&
control_pid_good(s) == 0) {
/* No chance of getting a ready notification anymore */
service_enter_stop_post(s, SERVICE_FAILURE_PROTOCOL);
break;
}
_fallthrough_;
case SERVICE_START_POST:
if (s->pid_file_pathspec &&
main_pid_good(s) == 0 &&
control_pid_good(s) == 0) {
/* Give up hoping for the daemon to write its PID file */
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning(u, "Daemon never wrote its PID file. Failing.");
service_unwatch_pid_file(s);
if (s->state == SERVICE_START)
service_enter_stop_post(s, SERVICE_FAILURE_PROTOCOL);
else
service_enter_stop(s, SERVICE_FAILURE_PROTOCOL);
}
break;
case SERVICE_RUNNING:
/* service_enter_running() will figure out what to do */
service_enter_running(s, SERVICE_SUCCESS);
break;
case SERVICE_STOP_SIGABRT:
case SERVICE_STOP_SIGTERM:
case SERVICE_STOP_SIGKILL:
if (main_pid_good(s) <= 0 && control_pid_good(s) <= 0)
service_enter_stop_post(s, SERVICE_SUCCESS);
break;
case SERVICE_STOP_POST:
case SERVICE_FINAL_SIGTERM:
case SERVICE_FINAL_SIGKILL:
if (main_pid_good(s) <= 0 && control_pid_good(s) <= 0)
service_enter_dead(s, SERVICE_SUCCESS, true);
break;
default:
;
}
}
2010-01-26 21:39:06 +01:00
static void service_sigchld_event(Unit *u, pid_t pid, int code, int status) {
bool notify_dbus = true;
2010-01-26 21:39:06 +01:00
Service *s = SERVICE(u);
ServiceResult f;
assert(s);
assert(pid >= 0);
if (is_clean_exit(code, status, s->type == SERVICE_ONESHOT ? EXIT_CLEAN_COMMAND : EXIT_CLEAN_DAEMON, &s->success_status))
f = SERVICE_SUCCESS;
else if (code == CLD_EXITED)
f = SERVICE_FAILURE_EXIT_CODE;
else if (code == CLD_KILLED)
f = SERVICE_FAILURE_SIGNAL;
else if (code == CLD_DUMPED)
f = SERVICE_FAILURE_CORE_DUMP;
else
assert_not_reached("Unknown code");
if (s->main_pid == pid) {
/* Forking services may occasionally move to a new PID.
* As long as they update the PID file before exiting the old
* PID, they're fine. */
if (service_load_pid_file(s, false) > 0)
return;
s->main_pid = 0;
exec_status_exit(&s->main_exec_status, &s->exec_context, pid, code, status);
if (s->main_command) {
/* If this is not a forking service than the
* main process got started and hence we copy
* the exit status so that it is recorded both
* as main and as control process exit
* status */
s->main_command->exec_status = s->main_exec_status;
if (s->main_command->flags & EXEC_COMMAND_IGNORE_FAILURE)
f = SERVICE_SUCCESS;
} else if (s->exec_command[SERVICE_EXEC_START]) {
/* If this is a forked process, then we should
* ignore the return value if this was
* configured for the starter process */
if (s->exec_command[SERVICE_EXEC_START]->flags & EXEC_COMMAND_IGNORE_FAILURE)
f = SERVICE_SUCCESS;
}
/* When this is a successful exit, let's log about the exit code on DEBUG level. If this is a failure
* and the process exited on its own via exit(), then let's make this a NOTICE, under the assumption
* that the service already logged the reason at a higher log level on its own. However, if the service
* died due to a signal, then it most likely didn't say anything about any reason, hence let's raise
* our log level to WARNING then. */
log_struct(f == SERVICE_SUCCESS ? LOG_DEBUG :
(code == CLD_EXITED ? LOG_NOTICE : LOG_WARNING),
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
LOG_UNIT_MESSAGE(u, "Main process exited, code=%s, status=%i/%s",
sigchld_code_to_string(code), status,
strna(code == CLD_EXITED
? exit_status_to_string(status, EXIT_STATUS_FULL)
: signal_to_string(status))),
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
"EXIT_CODE=%s", sigchld_code_to_string(code),
"EXIT_STATUS=%i", status,
LOG_UNIT_ID(u),
LOG_UNIT_INVOCATION_ID(u),
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
NULL);
if (s->result == SERVICE_SUCCESS)
s->result = f;
if (s->main_command &&
s->main_command->command_next &&
s->type == SERVICE_ONESHOT &&
f == SERVICE_SUCCESS) {
/* There is another command to *
* execute, so let's do that. */
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_debug(u, "Running next main command for state %s.", service_state_to_string(s->state));
service_run_next_main(s);
} else {
/* The service exited, so the service is officially
* gone. */
s->main_command = NULL;
switch (s->state) {
case SERVICE_START_POST:
case SERVICE_RELOAD:
case SERVICE_STOP:
/* Need to wait until the operation is
* done */
break;
2010-01-30 02:07:35 +01:00
case SERVICE_START:
if (s->type == SERVICE_ONESHOT) {
/* This was our main goal, so let's go on */
if (f == SERVICE_SUCCESS)
service_enter_start_post(s);
else
service_enter_signal(s, SERVICE_STOP_SIGTERM, f);
break;
} else if (s->type == SERVICE_NOTIFY) {
/* Only enter running through a notification, so that the
* SERVICE_START state signifies that no ready notification
* has been received */
if (f != SERVICE_SUCCESS)
service_enter_signal(s, SERVICE_STOP_SIGTERM, f);
else if (!s->remain_after_exit || s->notify_access == NOTIFY_MAIN)
/* The service has never been and will never be active */
service_enter_signal(s, SERVICE_STOP_SIGTERM, SERVICE_FAILURE_PROTOCOL);
break;
}
_fallthrough_;
case SERVICE_RUNNING:
service_enter_running(s, f);
break;
case SERVICE_STOP_SIGABRT:
case SERVICE_STOP_SIGTERM:
case SERVICE_STOP_SIGKILL:
if (control_pid_good(s) <= 0)
service_enter_stop_post(s, f);
/* If there is still a control process, wait for that first */
break;
case SERVICE_STOP_POST:
case SERVICE_FINAL_SIGTERM:
case SERVICE_FINAL_SIGKILL:
if (control_pid_good(s) <= 0)
service_enter_dead(s, f, true);
break;
default:
assert_not_reached("Uh, main process died at wrong time.");
}
}
} else if (s->control_pid == pid) {
s->control_pid = 0;
if (s->control_command) {
exec_status_exit(&s->control_command->exec_status, &s->exec_context, pid, code, status);
2010-04-21 03:27:44 +02:00
if (s->control_command->flags & EXEC_COMMAND_IGNORE_FAILURE)
f = SERVICE_SUCCESS;
}
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_full(u, f == SERVICE_SUCCESS ? LOG_DEBUG : LOG_NOTICE, 0,
"Control process exited, code=%s status=%i",
sigchld_code_to_string(code), status);
if (s->result == SERVICE_SUCCESS)
s->result = f;
if (s->control_command &&
s->control_command->command_next &&
f == SERVICE_SUCCESS) {
/* There is another command to *
* execute, so let's do that. */
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_debug(u, "Running next control command for state %s.", service_state_to_string(s->state));
service_run_next_control(s);
} else {
/* No further commands for this step, so let's
* figure out what to do next */
2010-04-21 03:27:44 +02:00
s->control_command = NULL;
s->control_command_id = _SERVICE_EXEC_COMMAND_INVALID;
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_debug(u, "Got final SIGCHLD for state %s.", service_state_to_string(s->state));
2010-01-27 05:32:31 +01:00
switch (s->state) {
case SERVICE_START_PRE:
if (f == SERVICE_SUCCESS)
service_enter_start(s);
else
service_enter_signal(s, SERVICE_STOP_SIGTERM, f);
break;
case SERVICE_START:
if (s->type != SERVICE_FORKING)
/* Maybe spurious event due to a reload that changed the type? */
break;
if (f != SERVICE_SUCCESS) {
service_enter_signal(s, SERVICE_STOP_SIGTERM, f);
break;
}
if (s->pid_file) {
bool has_start_post;
int r;
/* Let's try to load the pid file here if we can.
* The PID file might actually be created by a START_POST
* script. In that case don't worry if the loading fails. */
has_start_post = !!s->exec_command[SERVICE_EXEC_START_POST];
r = service_load_pid_file(s, !has_start_post);
if (!has_start_post && r < 0) {
r = service_demand_pid_file(s);
if (r < 0 || cgroup_good(s) == 0)
service_enter_signal(s, SERVICE_STOP_SIGTERM, SERVICE_FAILURE_PROTOCOL);
break;
}
} else
service_search_main_pid(s);
service_enter_start_post(s);
break;
case SERVICE_START_POST:
if (f != SERVICE_SUCCESS) {
service_enter_signal(s, SERVICE_STOP_SIGTERM, f);
break;
}
if (s->pid_file) {
int r;
r = service_load_pid_file(s, true);
if (r < 0) {
r = service_demand_pid_file(s);
if (r < 0 || cgroup_good(s) == 0)
service_enter_stop(s, SERVICE_FAILURE_PROTOCOL);
break;
}
} else
service_search_main_pid(s);
service_enter_running(s, SERVICE_SUCCESS);
break;
case SERVICE_RELOAD:
if (f == SERVICE_SUCCESS)
if (service_load_pid_file(s, true) < 0)
service_search_main_pid(s);
s->reload_result = f;
service_enter_running(s, SERVICE_SUCCESS);
break;
case SERVICE_STOP:
service_enter_signal(s, SERVICE_STOP_SIGTERM, f);
break;
case SERVICE_STOP_SIGABRT:
case SERVICE_STOP_SIGTERM:
case SERVICE_STOP_SIGKILL:
if (main_pid_good(s) <= 0)
service_enter_stop_post(s, f);
/* If there is still a service
* process around, wait until
* that one quit, too */
break;
case SERVICE_STOP_POST:
case SERVICE_FINAL_SIGTERM:
case SERVICE_FINAL_SIGKILL:
if (main_pid_good(s) <= 0)
service_enter_dead(s, f, true);
break;
default:
assert_not_reached("Uh, control process died at wrong time.");
}
}
} else /* Neither control nor main PID? If so, don't notify about anything */
notify_dbus = false;
/* Notify clients about changed exit status */
if (notify_dbus)
unit_add_to_dbus_queue(u);
/* If we get a SIGCHLD event for one of the processes we were interested in, then we look for others to watch,
* under the assumption that we'll sooner or later get a SIGCHLD for them, as the original process we watched
* was probably the parent of them, and they are hence now our children. */
unit_tidy_watch_pids(u, s->main_pid, s->control_pid);
unit_watch_all_pids(u);
/* If the PID set is empty now, then let's check if the cgroup is empty too and finish off the unit. */
unit_synthesize_cgroup_empty_event(u);
}
static int service_dispatch_timer(sd_event_source *source, usec_t usec, void *userdata) {
Service *s = SERVICE(userdata);
assert(s);
assert(source == s->timer_event_source);
switch (s->state) {
case SERVICE_START_PRE:
case SERVICE_START:
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning(UNIT(s), "%s operation timed out. Terminating.", s->state == SERVICE_START ? "Start" : "Start-pre");
service_enter_signal(s, SERVICE_STOP_SIGTERM, SERVICE_FAILURE_TIMEOUT);
break;
case SERVICE_START_POST:
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning(UNIT(s), "Start-post operation timed out. Stopping.");
service_enter_signal(s, SERVICE_STOP_SIGTERM, SERVICE_FAILURE_TIMEOUT);
break;
case SERVICE_RUNNING:
log_unit_warning(UNIT(s), "Service reached runtime time limit. Stopping.");
service_enter_stop(s, SERVICE_FAILURE_TIMEOUT);
break;
case SERVICE_RELOAD:
log_unit_warning(UNIT(s), "Reload operation timed out. Killing reload process.");
service_kill_control_process(s);
s->reload_result = SERVICE_FAILURE_TIMEOUT;
service_enter_running(s, SERVICE_SUCCESS);
break;
case SERVICE_STOP:
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning(UNIT(s), "Stopping timed out. Terminating.");
service_enter_signal(s, SERVICE_STOP_SIGTERM, SERVICE_FAILURE_TIMEOUT);
break;
case SERVICE_STOP_SIGABRT:
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning(UNIT(s), "State 'stop-sigabrt' timed out. Terminating.");
service_enter_signal(s, SERVICE_STOP_SIGTERM, SERVICE_FAILURE_TIMEOUT);
break;
case SERVICE_STOP_SIGTERM:
if (s->kill_context.send_sigkill) {
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning(UNIT(s), "State 'stop-sigterm' timed out. Killing.");
service_enter_signal(s, SERVICE_STOP_SIGKILL, SERVICE_FAILURE_TIMEOUT);
} else {
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning(UNIT(s), "State 'stop-sigterm' timed out. Skipping SIGKILL.");
service_enter_stop_post(s, SERVICE_FAILURE_TIMEOUT);
}
break;
case SERVICE_STOP_SIGKILL:
/* Uh, we sent a SIGKILL and it is still not gone?
* Must be something we cannot kill, so let's just be
* weirded out and continue */
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning(UNIT(s), "Processes still around after SIGKILL. Ignoring.");
service_enter_stop_post(s, SERVICE_FAILURE_TIMEOUT);
break;
case SERVICE_STOP_POST:
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning(UNIT(s), "State 'stop-post' timed out. Terminating.");
service_enter_signal(s, SERVICE_FINAL_SIGTERM, SERVICE_FAILURE_TIMEOUT);
break;
case SERVICE_FINAL_SIGTERM:
if (s->kill_context.send_sigkill) {
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning(UNIT(s), "State 'stop-final-sigterm' timed out. Killing.");
service_enter_signal(s, SERVICE_FINAL_SIGKILL, SERVICE_FAILURE_TIMEOUT);
} else {
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning(UNIT(s), "State 'stop-final-sigterm' timed out. Skipping SIGKILL. Entering failed mode.");
service_enter_dead(s, SERVICE_FAILURE_TIMEOUT, false);
}
break;
case SERVICE_FINAL_SIGKILL:
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning(UNIT(s), "Processes still around after final SIGKILL. Entering failed mode.");
service_enter_dead(s, SERVICE_FAILURE_TIMEOUT, true);
break;
case SERVICE_AUTO_RESTART:
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_info(UNIT(s),
s->restart_usec > 0 ?
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
"Service hold-off time over, scheduling restart." :
"Service has no hold-off time, scheduling restart.");
service_enter_restart(s);
break;
default:
assert_not_reached("Timeout at wrong time.");
}
return 0;
}
static int service_dispatch_watchdog(sd_event_source *source, usec_t usec, void *userdata) {
Service *s = SERVICE(userdata);
char t[FORMAT_TIMESPAN_MAX];
usec_t watchdog_usec;
assert(s);
assert(source == s->watchdog_event_source);
watchdog_usec = service_get_watchdog_usec(s);
if (UNIT(s)->manager->service_watchdogs) {
log_unit_error(UNIT(s), "Watchdog timeout (limit %s)!",
format_timespan(t, sizeof(t), watchdog_usec, 1));
service_enter_signal(s, SERVICE_STOP_SIGABRT, SERVICE_FAILURE_WATCHDOG);
} else
log_unit_warning(UNIT(s), "Watchdog disabled! Ignoring watchdog timeout (limit %s)!",
format_timespan(t, sizeof(t), watchdog_usec, 1));
2013-12-11 20:49:43 +01:00
return 0;
}
static bool service_notify_message_authorized(Service *s, pid_t pid, char **tags, FDSet *fds) {
assert(s);
if (s->notify_access == NOTIFY_NONE) {
log_unit_warning(UNIT(s), "Got notification message from PID "PID_FMT", but reception is disabled.", pid);
return false;
}
if (s->notify_access == NOTIFY_MAIN && pid != s->main_pid) {
if (s->main_pid != 0)
log_unit_warning(UNIT(s), "Got notification message from PID "PID_FMT", but reception only permitted for main PID "PID_FMT, pid, s->main_pid);
else
log_unit_warning(UNIT(s), "Got notification message from PID "PID_FMT", but reception only permitted for main PID which is currently not known", pid);
return false;
}
if (s->notify_access == NOTIFY_EXEC && pid != s->main_pid && pid != s->control_pid) {
if (s->main_pid != 0 && s->control_pid != 0)
log_unit_warning(UNIT(s), "Got notification message from PID "PID_FMT", but reception only permitted for main PID "PID_FMT" and control PID "PID_FMT,
pid, s->main_pid, s->control_pid);
else if (s->main_pid != 0)
log_unit_warning(UNIT(s), "Got notification message from PID "PID_FMT", but reception only permitted for main PID "PID_FMT, pid, s->main_pid);
else if (s->control_pid != 0)
log_unit_warning(UNIT(s), "Got notification message from PID "PID_FMT", but reception only permitted for control PID "PID_FMT, pid, s->control_pid);
else
log_unit_warning(UNIT(s), "Got notification message from PID "PID_FMT", but reception only permitted for main PID and control PID which are currently not known", pid);
return false;
}
return true;
}
static void service_notify_message(
Unit *u,
const struct ucred *ucred,
char **tags,
FDSet *fds) {
Service *s = SERVICE(u);
bool notify_dbus = false;
const char *e;
char **i;
int r;
assert(u);
assert(ucred);
if (!service_notify_message_authorized(SERVICE(u), ucred->pid, tags, fds))
return;
if (DEBUG_LOGGING) {
_cleanup_free_ char *cc = NULL;
cc = strv_join(tags, ", ");
log_unit_debug(u, "Got notification message from PID "PID_FMT" (%s)", ucred->pid, isempty(cc) ? "n/a" : cc);
}
/* Interpret MAINPID= */
e = strv_find_startswith(tags, "MAINPID=");
2014-07-07 17:03:34 +02:00
if (e && IN_SET(s->state, SERVICE_START, SERVICE_START_POST, SERVICE_RUNNING, SERVICE_RELOAD)) {
pid_t new_main_pid;
if (parse_pid(e, &new_main_pid) < 0)
log_unit_warning(u, "Failed to parse MAINPID= field in notification message, ignoring: %s", e);
else if (!s->main_pid_known || new_main_pid != s->main_pid) {
r = service_is_suitable_main_pid(s, new_main_pid, LOG_WARNING);
if (r == 0) {
/* The new main PID is a bit suspicous, which is OK if the sender is privileged. */
if (ucred->uid == 0) {
log_unit_debug(u, "New main PID "PID_FMT" does not belong to service, but we'll accept it as the request to change it came from a privileged process.", new_main_pid);
r = 1;
} else
log_unit_debug(u, "New main PID "PID_FMT" does not belong to service, refusing.", new_main_pid);
}
if (r > 0) {
service_set_main_pid(s, new_main_pid);
unit_watch_pid(UNIT(s), new_main_pid);
notify_dbus = true;
}
}
}
/* Interpret READY=/STOPPING=/RELOADING=. Last one wins. */
STRV_FOREACH_BACKWARDS(i, tags) {
if (streq(*i, "READY=1")) {
s->notify_state = NOTIFY_READY;
/* Type=notify services inform us about completed
* initialization with READY=1 */
if (s->type == SERVICE_NOTIFY && s->state == SERVICE_START)
service_enter_start_post(s);
/* Sending READY=1 while we are reloading informs us
* that the reloading is complete */
if (s->state == SERVICE_RELOAD && s->control_pid == 0)
service_enter_running(s, SERVICE_SUCCESS);
notify_dbus = true;
break;
} else if (streq(*i, "RELOADING=1")) {
s->notify_state = NOTIFY_RELOADING;
if (s->state == SERVICE_RUNNING)
service_enter_reload_by_notify(s);
notify_dbus = true;
break;
} else if (streq(*i, "STOPPING=1")) {
s->notify_state = NOTIFY_STOPPING;
if (s->state == SERVICE_RUNNING)
service_enter_stop_by_notify(s);
notify_dbus = true;
break;
}
}
/* Interpret STATUS= */
e = strv_find_startswith(tags, "STATUS=");
if (e) {
_cleanup_free_ char *t = NULL;
if (!isempty(e)) {
if (!utf8_is_valid(e))
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_warning(u, "Status message in notification message is not UTF-8 clean.");
else {
t = strdup(e);
if (!t)
log_oom();
}
}
if (!streq_ptr(s->status_text, t)) {
free_and_replace(s->status_text, t);
notify_dbus = true;
}
}
2013-12-11 20:49:43 +01:00
/* Interpret ERRNO= */
e = strv_find_startswith(tags, "ERRNO=");
if (e) {
int status_errno;
status_errno = parse_errno(e);
if (status_errno < 0)
log_unit_warning_errno(u, status_errno,
"Failed to parse ERRNO= field in notification message: %s", e);
else if (s->status_errno != status_errno) {
s->status_errno = status_errno;
notify_dbus = true;
}
}
/* Interpret EXTEND_TIMEOUT= */
e = strv_find_startswith(tags, "EXTEND_TIMEOUT_USEC=");
if (e) {
usec_t extend_timeout_usec;
if (safe_atou64(e, &extend_timeout_usec) < 0)
log_unit_warning(u, "Failed to parse EXTEND_TIMEOUT_USEC=%s", e);
else
service_extend_timeout(s, extend_timeout_usec);
}
/* Interpret WATCHDOG= */
if (strv_find(tags, "WATCHDOG=1"))
2013-12-11 20:49:43 +01:00
service_reset_watchdog(s);
e = strv_find_startswith(tags, "WATCHDOG_USEC=");
if (e) {
usec_t watchdog_override_usec;
if (safe_atou64(e, &watchdog_override_usec) < 0)
log_unit_warning(u, "Failed to parse WATCHDOG_USEC=%s", e);
else
service_reset_watchdog_timeout(s, watchdog_override_usec);
}
/* Process FD store messages. Either FDSTOREREMOVE=1 for removal, or FDSTORE=1 for addition. In both cases,
* process FDNAME= for picking the file descriptor name to use. Note that FDNAME= is required when removing
* fds, but optional when pushing in new fds, for compatibility reasons. */
if (strv_find(tags, "FDSTOREREMOVE=1")) {
const char *name;
name = strv_find_startswith(tags, "FDNAME=");
if (!name || !fdname_is_valid(name))
log_unit_warning(u, "FDSTOREREMOVE=1 requested, but no valid file descriptor name passed, ignoring.");
else
service_remove_fd_store(s, name);
} else if (strv_find(tags, "FDSTORE=1")) {
const char *name;
name = strv_find_startswith(tags, "FDNAME=");
if (name && !fdname_is_valid(name)) {
log_unit_warning(u, "Passed FDNAME= name is invalid, ignoring.");
name = NULL;
}
(void) service_add_fd_store_set(s, fds, name);
}
/* Notify clients about changed status or main pid */
if (notify_dbus)
unit_add_to_dbus_queue(u);
}
static int service_get_timeout(Unit *u, usec_t *timeout) {
Service *s = SERVICE(u);
uint64_t t;
int r;
if (!s->timer_event_source)
return 0;
r = sd_event_source_get_time(s->timer_event_source, &t);
if (r < 0)
return r;
if (t == USEC_INFINITY)
return 0;
*timeout = t;
return 1;
}
static void service_bus_name_owner_change(
Unit *u,
const char *name,
const char *old_owner,
const char *new_owner) {
Service *s = SERVICE(u);
int r;
assert(s);
assert(name);
assert(streq(s->bus_name, name));
assert(old_owner || new_owner);
if (old_owner && new_owner)
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_debug(u, "D-Bus name %s changed owner from %s to %s", name, old_owner, new_owner);
else if (old_owner)
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_debug(u, "D-Bus name %s no longer registered by %s", name, old_owner);
else
core,network: major per-object logging rework This changes log_unit_info() (and friends) to take a real Unit* object insted of just a unit name as parameter. The call will now prefix all logged messages with the unit name, thus allowing the unit name to be dropped from the various passed romat strings, simplifying invocations drastically, and unifying log output across messages. Also, UNIT= vs. USER_UNIT= is now derived from the Manager object attached to the Unit object, instead of getpid(). This has the benefit of correcting the field for --test runs. Also contains a couple of other logging improvements: - Drops a couple of strerror() invocations in favour of using %m. - Not only .mount units now warn if a symlinks exist for the mount point already, .automount units do that too, now. - A few invocations of log_struct() that didn't actually pass any additional structured data have been replaced by simpler invocations of log_unit_info() and friends. - For structured data a new LOG_UNIT_MESSAGE() macro has been added, that works like LOG_MESSAGE() but prefixes the message with the unit name. Similar, there's now LOG_LINK_MESSAGE() and LOG_NETDEV_MESSAGE(). - For structured data new LOG_UNIT_ID(), LOG_LINK_INTERFACE(), LOG_NETDEV_INTERFACE() macros have been added that generate the necessary per object fields. The old log_unit_struct() call has been removed in favour of these new macros used in raw log_struct() invocations. In addition to removing one more function call this allows generated structured log messages that contain two object fields, as necessary for example for network interfaces that are joined into another network interface, and whose messages shall be indexed by both. - The LOG_ERRNO() macro has been removed, in favour of log_struct_errno(). The latter has the benefit of ensuring that %m in format strings is properly resolved to the specified error number. - A number of logging messages have been converted to use log_unit_info() instead of log_info() - The client code in sysv-generator no longer #includes core code from src/core/. - log_unit_full_errno() has been removed, log_unit_full() instead takes an errno now, too. - log_unit_info(), log_link_info(), log_netdev_info() and friends, now avoid double evaluation of their parameters
2015-05-11 20:38:21 +02:00
log_unit_debug(u, "D-Bus name %s now registered by %s", name, new_owner);
s->bus_name_good = !!new_owner;
/* Track the current owner, so we can reconstruct changes after a daemon reload */
r = free_and_strdup(&s->bus_name_owner, new_owner);
if (r < 0) {
log_unit_error_errno(u, r, "Unable to set new bus name owner %s: %m", new_owner);
return;
}
if (s->type == SERVICE_DBUS) {
/* service_enter_running() will figure out what to
* do */
if (s->state == SERVICE_RUNNING)
service_enter_running(s, SERVICE_SUCCESS);
else if (s->state == SERVICE_START && new_owner)
service_enter_start_post(s);
} else if (new_owner &&
s->main_pid <= 0 &&
IN_SET(s->state,
SERVICE_START,
SERVICE_START_POST,
SERVICE_RUNNING,
SERVICE_RELOAD)) {
_cleanup_(sd_bus_creds_unrefp) sd_bus_creds *creds = NULL;
pid_t pid;
/* Try to acquire PID from bus service */
r = sd_bus_get_name_creds(u->manager->api_bus, name, SD_BUS_CREDS_PID, &creds);
if (r >= 0)
r = sd_bus_creds_get_pid(creds, &pid);
if (r >= 0) {
2017-02-28 16:07:56 +01:00
log_unit_debug(u, "D-Bus name %s is now owned by process " PID_FMT, name, pid);
service_set_main_pid(s, pid);
unit_watch_pid(UNIT(s), pid);
}
}
}
int service_set_socket_fd(Service *s, int fd, Socket *sock, bool selinux_context_net) {
_cleanup_free_ char *peer = NULL;
int r;
assert(s);
assert(fd >= 0);
/* This is called by the socket code when instantiating a new service for a stream socket and the socket needs
* to be configured. We take ownership of the passed fd on success. */
if (UNIT(s)->load_state != UNIT_LOADED)
return -EINVAL;
if (s->socket_fd >= 0)
return -EBUSY;
if (s->state != SERVICE_DEAD)
return -EAGAIN;
if (getpeername_pretty(fd, true, &peer) >= 0) {
if (UNIT(s)->description) {
_cleanup_free_ char *a;
a = strjoin(UNIT(s)->description, " (", peer, ")");
if (!a)
return -ENOMEM;
r = unit_set_description(UNIT(s), a);
} else
r = unit_set_description(UNIT(s), peer);
if (r < 0)
return r;
}
r = unit_add_two_dependencies(UNIT(sock), UNIT_BEFORE, UNIT_TRIGGERS, UNIT(s), false, UNIT_DEPENDENCY_IMPLICIT);
if (r < 0)
return r;
s->socket_fd = fd;
s->socket_fd_selinux_context_net = selinux_context_net;
unit_ref_set(&s->accept_socket, UNIT(s), UNIT(sock));
return 0;
}
static void service_reset_failed(Unit *u) {
Service *s = SERVICE(u);
assert(s);
if (s->state == SERVICE_FAILED)
service_set_state(s, SERVICE_DEAD);
s->result = SERVICE_SUCCESS;
s->reload_result = SERVICE_SUCCESS;
s->n_restarts = 0;
s->flush_n_restarts = false;
}
static int service_kill(Unit *u, KillWho who, int signo, sd_bus_error *error) {
2010-10-22 16:11:50 +02:00
Service *s = SERVICE(u);
assert(s);
return unit_kill_common(u, who, signo, s->main_pid, s->control_pid, error);
2010-10-22 16:11:50 +02:00
}
static int service_main_pid(Unit *u) {
Service *s = SERVICE(u);
assert(s);
return s->main_pid;
}
static int service_control_pid(Unit *u) {
Service *s = SERVICE(u);
assert(s);
return s->control_pid;
}
static bool service_needs_console(Unit *u) {
Service *s = SERVICE(u);
assert(s);
/* We provide our own implementation of this here, instead of relying of the generic implementation
* unit_needs_console() provides, since we want to return false if we are in SERVICE_EXITED state. */
if (!exec_context_may_touch_console(&s->exec_context))
return false;
return IN_SET(s->state,
SERVICE_START_PRE,
SERVICE_START,
SERVICE_START_POST,
SERVICE_RUNNING,
SERVICE_RELOAD,
SERVICE_STOP,
SERVICE_STOP_SIGABRT,
SERVICE_STOP_SIGTERM,
SERVICE_STOP_SIGKILL,
SERVICE_STOP_POST,
SERVICE_FINAL_SIGTERM,
SERVICE_FINAL_SIGKILL);
}
static const char* const service_restart_table[_SERVICE_RESTART_MAX] = {
[SERVICE_RESTART_NO] = "no",
[SERVICE_RESTART_ON_SUCCESS] = "on-success",
[SERVICE_RESTART_ON_FAILURE] = "on-failure",
[SERVICE_RESTART_ON_ABNORMAL] = "on-abnormal",
[SERVICE_RESTART_ON_WATCHDOG] = "on-watchdog",
[SERVICE_RESTART_ON_ABORT] = "on-abort",
[SERVICE_RESTART_ALWAYS] = "always",
};
DEFINE_STRING_TABLE_LOOKUP(service_restart, ServiceRestart);
static const char* const service_type_table[_SERVICE_TYPE_MAX] = {
[SERVICE_SIMPLE] = "simple",
2010-07-01 19:39:35 +02:00
[SERVICE_FORKING] = "forking",
[SERVICE_ONESHOT] = "oneshot",
[SERVICE_DBUS] = "dbus",
[SERVICE_NOTIFY] = "notify",
[SERVICE_IDLE] = "idle"
};
DEFINE_STRING_TABLE_LOOKUP(service_type, ServiceType);
static const char* const service_exec_command_table[_SERVICE_EXEC_COMMAND_MAX] = {
[SERVICE_EXEC_START_PRE] = "ExecStartPre",
[SERVICE_EXEC_START] = "ExecStart",
[SERVICE_EXEC_START_POST] = "ExecStartPost",
[SERVICE_EXEC_RELOAD] = "ExecReload",
[SERVICE_EXEC_STOP] = "ExecStop",
[SERVICE_EXEC_STOP_POST] = "ExecStopPost",
};
DEFINE_STRING_TABLE_LOOKUP(service_exec_command, ServiceExecCommand);
static const char* const notify_state_table[_NOTIFY_STATE_MAX] = {
[NOTIFY_UNKNOWN] = "unknown",
[NOTIFY_READY] = "ready",
[NOTIFY_RELOADING] = "reloading",
[NOTIFY_STOPPING] = "stopping",
};
DEFINE_STRING_TABLE_LOOKUP(notify_state, NotifyState);
static const char* const service_result_table[_SERVICE_RESULT_MAX] = {
[SERVICE_SUCCESS] = "success",
[SERVICE_FAILURE_RESOURCES] = "resources",
[SERVICE_FAILURE_PROTOCOL] = "protocol",
[SERVICE_FAILURE_TIMEOUT] = "timeout",
[SERVICE_FAILURE_EXIT_CODE] = "exit-code",
[SERVICE_FAILURE_SIGNAL] = "signal",
[SERVICE_FAILURE_CORE_DUMP] = "core-dump",
[SERVICE_FAILURE_WATCHDOG] = "watchdog",
[SERVICE_FAILURE_START_LIMIT_HIT] = "start-limit-hit",
};
DEFINE_STRING_TABLE_LOOKUP(service_result, ServiceResult);
2010-01-26 21:39:06 +01:00
const UnitVTable service_vtable = {
.object_size = sizeof(Service),
.exec_context_offset = offsetof(Service, exec_context),
.cgroup_context_offset = offsetof(Service, cgroup_context),
.kill_context_offset = offsetof(Service, kill_context),
.exec_runtime_offset = offsetof(Service, exec_runtime),
.dynamic_creds_offset = offsetof(Service, dynamic_creds),
.sections =
"Unit\0"
"Service\0"
"Install\0",
.private_section = "Service",
.can_transient = true,
.can_delegate = true,
.init = service_init,
.done = service_done,
2010-04-21 03:27:44 +02:00
.load = service_load,
.release_resources = service_release_resources,
2010-04-21 03:27:44 +02:00
.coldplug = service_coldplug,
.dump = service_dump,
.start = service_start,
.stop = service_stop,
.reload = service_reload,
.can_reload = service_can_reload,
2010-10-22 16:11:50 +02:00
.kill = service_kill,
2010-04-21 03:27:44 +02:00
.serialize = service_serialize,
.deserialize_item = service_deserialize_item,
.active_state = service_active_state,
.sub_state_to_string = service_sub_state_to_string,
.will_restart = service_will_restart,
.may_gc = service_may_gc,
.sigchld_event = service_sigchld_event,
2010-02-14 01:09:01 +01:00
.reset_failed = service_reset_failed,
.notify_cgroup_empty = service_notify_cgroup_empty_event,
.notify_message = service_notify_message,
2010-03-31 16:29:55 +02:00
.main_pid = service_main_pid,
.control_pid = service_control_pid,
.bus_name_owner_change = service_bus_name_owner_change,
.bus_vtable = bus_service_vtable,
.bus_set_property = bus_service_set_property,
.bus_commit_properties = bus_service_commit_properties,
2010-04-18 03:08:16 +02:00
.get_timeout = service_get_timeout,
.needs_console = service_needs_console,
.status_message_formats = {
.starting_stopping = {
[0] = "Starting %s...",
[1] = "Stopping %s...",
},
.finished_start_job = {
[JOB_DONE] = "Started %s.",
[JOB_FAILED] = "Failed to start %s.",
},
.finished_stop_job = {
[JOB_DONE] = "Stopped %s.",
[JOB_FAILED] = "Stopped (with error) %s.",
},
},
};