/*-*- Mode: C; c-basic-offset: 8 -*-*/
/***
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 General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
systemd is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
General Public License for more details.
You should have received a copy of the GNU General Public License
along with systemd; If not, see .
***/
#include
#include
#include
#include
#include
#include
#include
#include
#include "unit.h"
#include "socket.h"
#include "log.h"
#include "load-dropin.h"
#include "load-fragment.h"
#include "strv.h"
#include "unit-name.h"
#include "dbus-socket.h"
#include "missing.h"
#include "special.h"
#include "bus-errors.h"
static const UnitActiveState state_translation_table[_SOCKET_STATE_MAX] = {
[SOCKET_DEAD] = UNIT_INACTIVE,
[SOCKET_START_PRE] = UNIT_ACTIVATING,
[SOCKET_START_POST] = UNIT_ACTIVATING,
[SOCKET_LISTENING] = UNIT_ACTIVE,
[SOCKET_RUNNING] = UNIT_ACTIVE,
[SOCKET_STOP_PRE] = UNIT_DEACTIVATING,
[SOCKET_STOP_PRE_SIGTERM] = UNIT_DEACTIVATING,
[SOCKET_STOP_PRE_SIGKILL] = UNIT_DEACTIVATING,
[SOCKET_STOP_POST] = UNIT_DEACTIVATING,
[SOCKET_FINAL_SIGTERM] = UNIT_DEACTIVATING,
[SOCKET_FINAL_SIGKILL] = UNIT_DEACTIVATING,
[SOCKET_MAINTENANCE] = UNIT_MAINTENANCE
};
static void socket_init(Unit *u) {
Socket *s = SOCKET(u);
assert(u);
assert(u->meta.load_state == UNIT_STUB);
s->backlog = SOMAXCONN;
s->timeout_usec = DEFAULT_TIMEOUT_USEC;
s->directory_mode = 0755;
s->socket_mode = 0666;
s->max_connections = 64;
s->priority = -1;
s->ip_tos = -1;
s->ip_ttl = -1;
s->mark = -1;
exec_context_init(&s->exec_context);
s->control_command_id = _SOCKET_EXEC_COMMAND_INVALID;
}
static void socket_unwatch_control_pid(Socket *s) {
assert(s);
if (s->control_pid <= 0)
return;
unit_unwatch_pid(UNIT(s), s->control_pid);
s->control_pid = 0;
}
static void socket_done(Unit *u) {
Socket *s = SOCKET(u);
SocketPort *p;
Meta *i;
assert(s);
while ((p = s->ports)) {
LIST_REMOVE(SocketPort, port, s->ports, p);
if (p->fd >= 0) {
unit_unwatch_fd(UNIT(s), &p->fd_watch);
close_nointr_nofail(p->fd);
}
free(p->path);
free(p);
}
exec_context_done(&s->exec_context);
exec_command_free_array(s->exec_command, _SOCKET_EXEC_COMMAND_MAX);
s->control_command = NULL;
socket_unwatch_control_pid(s);
s->service = NULL;
free(s->bind_to_device);
s->bind_to_device = NULL;
unit_unwatch_timer(u, &s->timer_watch);
/* Make sure no service instance refers to us anymore. */
LIST_FOREACH(units_per_type, i, u->meta.manager->units_per_type[UNIT_SERVICE]) {
Service *service = (Service *) i;
if (service->socket == s)
service->socket = NULL;
}
}
static int socket_instantiate_service(Socket *s) {
char *prefix, *name;
int r;
Unit *u;
assert(s);
/* This fills in s->service if it isn't filled in yet. For
* Accept=yes sockets we create the next connection service
* here. For Accept=no this is mostly a NOP since the service
* is figured out at load time anyway. */
if (s->service)
return 0;
assert(s->accept);
if (!(prefix = unit_name_to_prefix(s->meta.id)))
return -ENOMEM;
r = asprintf(&name, "%s@%u.service", prefix, s->n_accepted);
free(prefix);
if (r < 0)
return -ENOMEM;
r = manager_load_unit(s->meta.manager, name, NULL, NULL, &u);
free(name);
if (r < 0)
return r;
s->service = SERVICE(u);
return 0;
}
static bool have_non_accept_socket(Socket *s) {
SocketPort *p;
assert(s);
if (!s->accept)
return true;
LIST_FOREACH(port, p, s->ports) {
if (p->type != SOCKET_SOCKET)
return true;
if (!socket_address_can_accept(&p->address))
return true;
}
return false;
}
static int socket_verify(Socket *s) {
assert(s);
if (s->meta.load_state != UNIT_LOADED)
return 0;
if (!s->ports) {
log_error("%s lacks Listen setting. Refusing.", s->meta.id);
return -EINVAL;
}
if (s->accept && have_non_accept_socket(s)) {
log_error("%s configured for accepting sockets, but sockets are non-accepting. Refusing.", s->meta.id);
return -EINVAL;
}
if (s->accept && s->max_connections <= 0) {
log_error("%s's MaxConnection setting too small. Refusing.", s->meta.id);
return -EINVAL;
}
if (s->exec_context.pam_name && s->exec_context.kill_mode != KILL_CONTROL_GROUP) {
log_error("%s has PAM enabled. Kill mode must be set to 'control-group'. Refusing.", s->meta.id);
return -EINVAL;
}
return 0;
}
static bool socket_needs_mount(Socket *s, const char *prefix) {
SocketPort *p;
assert(s);
LIST_FOREACH(port, p, s->ports) {
if (p->type == SOCKET_SOCKET) {
if (socket_address_needs_mount(&p->address, prefix))
return true;
} else {
assert(p->type == SOCKET_FIFO);
if (path_startswith(p->path, prefix))
return true;
}
}
return false;
}
int socket_add_one_mount_link(Socket *s, Mount *m) {
int r;
assert(s);
assert(m);
if (s->meta.load_state != UNIT_LOADED ||
m->meta.load_state != UNIT_LOADED)
return 0;
if (!socket_needs_mount(s, m->where))
return 0;
if ((r = unit_add_two_dependencies(UNIT(s), UNIT_AFTER, UNIT_REQUIRES, UNIT(m), true)) < 0)
return r;
return 0;
}
static int socket_add_mount_links(Socket *s) {
Meta *other;
int r;
assert(s);
LIST_FOREACH(units_per_type, other, s->meta.manager->units_per_type[UNIT_MOUNT])
if ((r = socket_add_one_mount_link(s, (Mount*) other)) < 0)
return r;
return 0;
}
static int socket_add_device_link(Socket *s) {
char *t;
int r;
assert(s);
if (!s->bind_to_device)
return 0;
if (asprintf(&t, "/sys/subsystem/net/devices/%s", s->bind_to_device) < 0)
return -ENOMEM;
r = unit_add_node_link(UNIT(s), t, false);
free(t);
return r;
}
static int socket_add_default_dependencies(Socket *s) {
int r;
assert(s);
if (s->meta.manager->running_as == MANAGER_SYSTEM)
if ((r = unit_add_two_dependencies_by_name(UNIT(s), UNIT_AFTER, UNIT_REQUIRES, SPECIAL_SYSINIT_TARGET, NULL, true)) < 0)
return r;
return unit_add_two_dependencies_by_name(UNIT(s), UNIT_BEFORE, UNIT_CONFLICTS, SPECIAL_SHUTDOWN_TARGET, NULL, true);
}
static int socket_load(Unit *u) {
Socket *s = SOCKET(u);
int r;
assert(u);
assert(u->meta.load_state == UNIT_STUB);
if ((r = unit_load_fragment_and_dropin(u)) < 0)
return r;
/* This is a new unit? Then let's add in some extras */
if (u->meta.load_state == UNIT_LOADED) {
if (have_non_accept_socket(s)) {
if ((r = unit_load_related_unit(u, ".service", (Unit**) &s->service)) < 0)
return r;
if ((r = unit_add_dependency(u, UNIT_BEFORE, UNIT(s->service), true)) < 0)
return r;
}
if ((r = socket_add_mount_links(s)) < 0)
return r;
if ((r = socket_add_device_link(s)) < 0)
return r;
if ((r = unit_add_exec_dependencies(u, &s->exec_context)) < 0)
return r;
if ((r = unit_add_default_cgroup(u)) < 0)
return r;
if (s->meta.default_dependencies)
if ((r = socket_add_default_dependencies(s)) < 0)
return r;
}
return socket_verify(s);
}
static const char* listen_lookup(int type) {
if (type == SOCK_STREAM)
return "ListenStream";
else if (type == SOCK_DGRAM)
return "ListenDatagram";
else if (type == SOCK_SEQPACKET)
return "ListenSequentialPacket";
assert_not_reached("Unknown socket type");
return NULL;
}
static void socket_dump(Unit *u, FILE *f, const char *prefix) {
SocketExecCommand c;
Socket *s = SOCKET(u);
SocketPort *p;
const char *prefix2;
char *p2;
assert(s);
assert(f);
p2 = strappend(prefix, "\t");
prefix2 = p2 ? p2 : prefix;
fprintf(f,
"%sSocket State: %s\n"
"%sBindIPv6Only: %s\n"
"%sBacklog: %u\n"
"%sSocketMode: %04o\n"
"%sDirectoryMode: %04o\n"
"%sKeepAlive: %s\n"
"%sFreeBind: %s\n",
prefix, socket_state_to_string(s->state),
prefix, socket_address_bind_ipv6_only_to_string(s->bind_ipv6_only),
prefix, s->backlog,
prefix, s->socket_mode,
prefix, s->directory_mode,
prefix, yes_no(s->keep_alive),
prefix, yes_no(s->free_bind));
if (s->control_pid > 0)
fprintf(f,
"%sControl PID: %lu\n",
prefix, (unsigned long) s->control_pid);
if (s->bind_to_device)
fprintf(f,
"%sBindToDevice: %s\n",
prefix, s->bind_to_device);
if (s->accept)
fprintf(f,
"%sAccepted: %u\n"
"%sNConnections: %u\n"
"%sMaxConnections: %u\n",
prefix, s->n_accepted,
prefix, s->n_connections,
prefix, s->max_connections);
if (s->priority >= 0)
fprintf(f,
"%sPriority: %i\n",
prefix, s->priority);
if (s->receive_buffer > 0)
fprintf(f,
"%sReceiveBuffer: %zu\n",
prefix, s->receive_buffer);
if (s->send_buffer > 0)
fprintf(f,
"%sSendBuffer: %zu\n",
prefix, s->send_buffer);
if (s->ip_tos >= 0)
fprintf(f,
"%sIPTOS: %i\n",
prefix, s->ip_tos);
if (s->ip_ttl >= 0)
fprintf(f,
"%sIPTTL: %i\n",
prefix, s->ip_ttl);
if (s->pipe_size > 0)
fprintf(f,
"%sPipeSize: %zu\n",
prefix, s->pipe_size);
if (s->mark >= 0)
fprintf(f,
"%sMark: %i\n",
prefix, s->mark);
LIST_FOREACH(port, p, s->ports) {
if (p->type == SOCKET_SOCKET) {
const char *t;
int r;
char *k;
if ((r = socket_address_print(&p->address, &k)) < 0)
t = strerror(-r);
else
t = k;
fprintf(f, "%s%s: %s\n", prefix, listen_lookup(p->address.type), k);
free(k);
} else
fprintf(f, "%sListenFIFO: %s\n", prefix, p->path);
}
exec_context_dump(&s->exec_context, f, prefix);
for (c = 0; c < _SOCKET_EXEC_COMMAND_MAX; c++) {
if (!s->exec_command[c])
continue;
fprintf(f, "%s-> %s:\n",
prefix, socket_exec_command_to_string(c));
exec_command_dump_list(s->exec_command[c], f, prefix2);
}
free(p2);
}
static int instance_from_socket(int fd, unsigned nr, char **instance) {
socklen_t l;
char *r;
union {
struct sockaddr sa;
struct sockaddr_un un;
struct sockaddr_in in;
struct sockaddr_in6 in6;
struct sockaddr_storage storage;
} local, remote;
assert(fd >= 0);
assert(instance);
l = sizeof(local);
if (getsockname(fd, &local.sa, &l) < 0)
return -errno;
l = sizeof(remote);
if (getpeername(fd, &remote.sa, &l) < 0)
return -errno;
switch (local.sa.sa_family) {
case AF_INET: {
uint32_t
a = ntohl(local.in.sin_addr.s_addr),
b = ntohl(remote.in.sin_addr.s_addr);
if (asprintf(&r,
"%u.%u.%u.%u:%u-%u.%u.%u.%u:%u",
a >> 24, (a >> 16) & 0xFF, (a >> 8) & 0xFF, a & 0xFF,
ntohs(local.in.sin_port),
b >> 24, (b >> 16) & 0xFF, (b >> 8) & 0xFF, b & 0xFF,
ntohs(remote.in.sin_port)) < 0)
return -ENOMEM;
break;
}
case AF_INET6: {
static const char ipv4_prefix[] = {
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0xFF, 0xFF
};
if (memcmp(&local.in6.sin6_addr, ipv4_prefix, sizeof(ipv4_prefix)) == 0 &&
memcmp(&remote.in6.sin6_addr, ipv4_prefix, sizeof(ipv4_prefix)) == 0) {
const uint8_t
*a = local.in6.sin6_addr.s6_addr+12,
*b = remote.in6.sin6_addr.s6_addr+12;
if (asprintf(&r,
"%u.%u.%u.%u:%u-%u.%u.%u.%u:%u",
a[0], a[1], a[2], a[3],
ntohs(local.in6.sin6_port),
b[0], b[1], b[2], b[3],
ntohs(remote.in6.sin6_port)) < 0)
return -ENOMEM;
} else {
char a[INET6_ADDRSTRLEN], b[INET6_ADDRSTRLEN];
if (asprintf(&r,
"%s:%u-%s:%u",
inet_ntop(AF_INET6, &local.in6.sin6_addr, a, sizeof(a)),
ntohs(local.in6.sin6_port),
inet_ntop(AF_INET6, &remote.in6.sin6_addr, b, sizeof(b)),
ntohs(remote.in6.sin6_port)) < 0)
return -ENOMEM;
}
break;
}
case AF_UNIX: {
struct ucred ucred;
l = sizeof(ucred);
if (getsockopt(fd, SOL_SOCKET, SO_PEERCRED, &ucred, &l) < 0)
return -errno;
if (asprintf(&r,
"%u-%lu-%lu",
nr,
(unsigned long) ucred.pid,
(unsigned long) ucred.uid) < 0)
return -ENOMEM;
break;
}
default:
assert_not_reached("Unhandled socket type.");
}
*instance = r;
return 0;
}
static void socket_close_fds(Socket *s) {
SocketPort *p;
assert(s);
LIST_FOREACH(port, p, s->ports) {
if (p->fd < 0)
continue;
unit_unwatch_fd(UNIT(s), &p->fd_watch);
close_nointr_nofail(p->fd);
/* One little note: we should never delete any sockets
* in the file system here! After all some other
* process we spawned might still have a reference of
* this fd and wants to continue to use it. Therefore
* we delete sockets in the file system before we
* create a new one, not after we stopped using
* one! */
p->fd = -1;
}
}
static void socket_apply_socket_options(Socket *s, int fd) {
assert(s);
assert(fd >= 0);
if (s->keep_alive) {
int b = s->keep_alive;
if (setsockopt(fd, SOL_SOCKET, SO_KEEPALIVE, &b, sizeof(b)) < 0)
log_warning("SO_KEEPALIVE failed: %m");
}
if (s->priority >= 0)
if (setsockopt(fd, SOL_SOCKET, SO_PRIORITY, &s->priority, sizeof(s->priority)) < 0)
log_warning("SO_PRIORITY failed: %m");
if (s->receive_buffer > 0) {
int value = (int) s->receive_buffer;
if (setsockopt(fd, SOL_SOCKET, SO_RCVBUF, &value, sizeof(value)) < 0)
log_warning("SO_RCVBUF failed: %m");
}
if (s->send_buffer > 0) {
int value = (int) s->send_buffer;
if (setsockopt(fd, SOL_SOCKET, SO_SNDBUF, &value, sizeof(value)) < 0)
log_warning("SO_SNDBUF failed: %m");
}
if (s->mark >= 0)
if (setsockopt(fd, SOL_SOCKET, SO_MARK, &s->mark, sizeof(s->mark)) < 0)
log_warning("SO_MARK failed: %m");
if (s->ip_tos >= 0)
if (setsockopt(fd, IPPROTO_IP, IP_TOS, &s->ip_tos, sizeof(s->ip_tos)) < 0)
log_warning("IP_TOS failed: %m");
if (s->ip_ttl >= 0) {
int r, x;
r = setsockopt(fd, IPPROTO_IP, IP_TTL, &s->ip_ttl, sizeof(s->ip_ttl));
x = setsockopt(fd, IPPROTO_IPV6, IPV6_UNICAST_HOPS, &s->ip_ttl, sizeof(s->ip_ttl));
if (r < 0 && x < 0)
log_warning("IP_TTL/IPV6_UNICAST_HOPS failed: %m");
}
}
static void socket_apply_fifo_options(Socket *s, int fd) {
assert(s);
assert(fd >= 0);
if (s->pipe_size > 0)
if (fcntl(fd, F_SETPIPE_SZ, s->pipe_size) < 0)
log_warning("F_SETPIPE_SZ: %m");
}
static int fifo_address_create(
const char *path,
mode_t directory_mode,
mode_t socket_mode,
/* FIXME SELINUX: pass SELinux context object here */
int *_fd) {
int fd = -1, r;
struct stat st;
mode_t old_mask;
assert(path);
assert(_fd);
mkdir_parents(path, directory_mode);
/* FIXME SELINUX: The mkfifo here should be done with
* the right SELinux context set */
/* Enforce the right access mode for the fifo */
old_mask = umask(~ socket_mode);
/* Include the original umask in our mask */
umask(~socket_mode | old_mask);
r = mkfifo(path, socket_mode);
umask(old_mask);
if (r < 0) {
r = -errno;
goto fail;
}
if ((fd = open(path, O_RDWR|O_CLOEXEC|O_NOCTTY|O_NONBLOCK|O_NOFOLLOW)) < 0) {
r = -errno;
goto fail;
}
if (fstat(fd, &st) < 0) {
r = -errno;
goto fail;
}
if (!S_ISFIFO(st.st_mode) ||
st.st_mode != (socket_mode & ~old_mask) ||
st.st_uid != getuid() ||
st.st_gid != getgid()) {
r = -EEXIST;
goto fail;
}
*_fd = fd;
return 0;
fail:
if (fd >= 0)
close_nointr_nofail(fd);
return r;
}
static int socket_open_fds(Socket *s) {
SocketPort *p;
int r;
assert(s);
/* FIXME SELINUX: Somewhere here we must set the the SELinux
context for the created sockets and FIFOs. To figure out
the executable name for this, use
socket_instantiate_service() and then access the executable
path name via
s->service->exec_command[SERVICE_EXEC_START]->path. Example:
if ((r = socket_instantiate_service(s)) < 0)
return r;
log_debug("Socket unit %s will spawn service unit %s with executable path %s.",
s->meta.id,
s->service->meta.id,
s->service->exec_command[SERVICE_EXEC_START]->path);
*/
LIST_FOREACH(port, p, s->ports) {
if (p->fd >= 0)
continue;
if (p->type == SOCKET_SOCKET) {
if ((r = socket_address_listen(
&p->address,
s->backlog,
s->bind_ipv6_only,
s->bind_to_device,
s->free_bind,
s->directory_mode,
s->socket_mode,
/* FIXME SELINUX: Pass the SELinux context object here */
&p->fd)) < 0)
goto rollback;
socket_apply_socket_options(s, p->fd);
} else if (p->type == SOCKET_FIFO) {
if ((r = fifo_address_create(
p->path,
s->directory_mode,
s->socket_mode,
/* FIXME SELINUX: Pass the SELinux context object here */
&p->fd)) < 0)
goto rollback;
socket_apply_fifo_options(s, p->fd);
} else
assert_not_reached("Unknown port type");
}
return 0;
rollback:
socket_close_fds(s);
return r;
}
static void socket_unwatch_fds(Socket *s) {
SocketPort *p;
assert(s);
LIST_FOREACH(port, p, s->ports) {
if (p->fd < 0)
continue;
unit_unwatch_fd(UNIT(s), &p->fd_watch);
}
}
static int socket_watch_fds(Socket *s) {
SocketPort *p;
int r;
assert(s);
LIST_FOREACH(port, p, s->ports) {
if (p->fd < 0)
continue;
p->fd_watch.socket_accept =
s->accept &&
p->type == SOCKET_SOCKET &&
socket_address_can_accept(&p->address);
if ((r = unit_watch_fd(UNIT(s), p->fd, EPOLLIN, &p->fd_watch)) < 0)
goto fail;
}
return 0;
fail:
socket_unwatch_fds(s);
return r;
}
static void socket_set_state(Socket *s, SocketState state) {
SocketState old_state;
assert(s);
old_state = s->state;
s->state = state;
if (state != SOCKET_START_PRE &&
state != SOCKET_START_POST &&
state != SOCKET_STOP_PRE &&
state != SOCKET_STOP_PRE_SIGTERM &&
state != SOCKET_STOP_PRE_SIGKILL &&
state != SOCKET_STOP_POST &&
state != SOCKET_FINAL_SIGTERM &&
state != SOCKET_FINAL_SIGKILL) {
unit_unwatch_timer(UNIT(s), &s->timer_watch);
socket_unwatch_control_pid(s);
s->control_command = NULL;
s->control_command_id = _SOCKET_EXEC_COMMAND_INVALID;
}
if (state != SOCKET_LISTENING)
socket_unwatch_fds(s);
if (state != SOCKET_START_POST &&
state != SOCKET_LISTENING &&
state != SOCKET_RUNNING &&
state != SOCKET_STOP_PRE &&
state != SOCKET_STOP_PRE_SIGTERM &&
state != SOCKET_STOP_PRE_SIGKILL)
socket_close_fds(s);
if (state != old_state)
log_debug("%s changed %s -> %s",
s->meta.id,
socket_state_to_string(old_state),
socket_state_to_string(state));
unit_notify(UNIT(s), state_translation_table[old_state], state_translation_table[state]);
}
static int socket_coldplug(Unit *u) {
Socket *s = SOCKET(u);
int r;
assert(s);
assert(s->state == SOCKET_DEAD);
if (s->deserialized_state != s->state) {
if (s->deserialized_state == SOCKET_START_PRE ||
s->deserialized_state == SOCKET_START_POST ||
s->deserialized_state == SOCKET_STOP_PRE ||
s->deserialized_state == SOCKET_STOP_PRE_SIGTERM ||
s->deserialized_state == SOCKET_STOP_PRE_SIGKILL ||
s->deserialized_state == SOCKET_STOP_POST ||
s->deserialized_state == SOCKET_FINAL_SIGTERM ||
s->deserialized_state == SOCKET_FINAL_SIGKILL) {
if (s->control_pid <= 0)
return -EBADMSG;
if ((r = unit_watch_pid(UNIT(s), s->control_pid)) < 0)
return r;
if ((r = unit_watch_timer(UNIT(s), s->timeout_usec, &s->timer_watch)) < 0)
return r;
}
if (s->deserialized_state == SOCKET_START_POST ||
s->deserialized_state == SOCKET_LISTENING ||
s->deserialized_state == SOCKET_RUNNING ||
s->deserialized_state == SOCKET_STOP_PRE ||
s->deserialized_state == SOCKET_STOP_PRE_SIGTERM ||
s->deserialized_state == SOCKET_STOP_PRE_SIGKILL)
if ((r = socket_open_fds(s)) < 0)
return r;
if (s->deserialized_state == SOCKET_LISTENING)
if ((r = socket_watch_fds(s)) < 0)
return r;
socket_set_state(s, s->deserialized_state);
}
return 0;
}
static int socket_spawn(Socket *s, ExecCommand *c, pid_t *_pid) {
pid_t pid;
int r;
char **argv;
assert(s);
assert(c);
assert(_pid);
if ((r = unit_watch_timer(UNIT(s), s->timeout_usec, &s->timer_watch)) < 0)
goto fail;
if (!(argv = unit_full_printf_strv(UNIT(s), c->argv))) {
r = -ENOMEM;
goto fail;
}
r = exec_spawn(c,
argv,
&s->exec_context,
NULL, 0,
s->meta.manager->environment,
true,
true,
true,
s->meta.manager->confirm_spawn,
s->meta.cgroup_bondings,
&pid);
strv_free(argv);
if (r < 0)
goto fail;
if ((r = unit_watch_pid(UNIT(s), pid)) < 0)
/* FIXME: we need to do something here */
goto fail;
*_pid = pid;
return 0;
fail:
unit_unwatch_timer(UNIT(s), &s->timer_watch);
return r;
}
static void socket_enter_dead(Socket *s, bool success) {
assert(s);
if (!success)
s->failure = true;
socket_set_state(s, s->failure ? SOCKET_MAINTENANCE : SOCKET_DEAD);
}
static void socket_enter_signal(Socket *s, SocketState state, bool success);
static void socket_enter_stop_post(Socket *s, bool success) {
int r;
assert(s);
if (!success)
s->failure = true;
socket_unwatch_control_pid(s);
s->control_command_id = SOCKET_EXEC_STOP_POST;
if ((s->control_command = s->exec_command[SOCKET_EXEC_STOP_POST])) {
if ((r = socket_spawn(s, s->control_command, &s->control_pid)) < 0)
goto fail;
socket_set_state(s, SOCKET_STOP_POST);
} else
socket_enter_signal(s, SOCKET_FINAL_SIGTERM, true);
return;
fail:
log_warning("%s failed to run 'stop-post' task: %s", s->meta.id, strerror(-r));
socket_enter_signal(s, SOCKET_FINAL_SIGTERM, false);
}
static void socket_enter_signal(Socket *s, SocketState state, bool success) {
int r;
bool sent = false;
assert(s);
if (!success)
s->failure = true;
if (s->exec_context.kill_mode != KILL_NONE) {
int sig = (state == SOCKET_STOP_PRE_SIGTERM || state == SOCKET_FINAL_SIGTERM) ? s->exec_context.kill_signal : SIGKILL;
if (s->exec_context.kill_mode == KILL_CONTROL_GROUP) {
if ((r = cgroup_bonding_kill_list(s->meta.cgroup_bondings, sig)) < 0) {
if (r != -EAGAIN && r != -ESRCH)
goto fail;
} else
sent = true;
}
if (!sent && s->control_pid > 0)
if (kill(s->exec_context.kill_mode == KILL_PROCESS ? s->control_pid : -s->control_pid, sig) < 0 && errno != ESRCH) {
r = -errno;
goto fail;
}
}
if (sent && s->control_pid > 0) {
if ((r = unit_watch_timer(UNIT(s), s->timeout_usec, &s->timer_watch)) < 0)
goto fail;
socket_set_state(s, state);
} else if (state == SOCKET_STOP_PRE_SIGTERM || state == SOCKET_STOP_PRE_SIGKILL)
socket_enter_stop_post(s, true);
else
socket_enter_dead(s, true);
return;
fail:
log_warning("%s failed to kill processes: %s", s->meta.id, strerror(-r));
if (state == SOCKET_STOP_PRE_SIGTERM || state == SOCKET_STOP_PRE_SIGKILL)
socket_enter_stop_post(s, false);
else
socket_enter_dead(s, false);
}
static void socket_enter_stop_pre(Socket *s, bool success) {
int r;
assert(s);
if (!success)
s->failure = true;
socket_unwatch_control_pid(s);
s->control_command_id = SOCKET_EXEC_STOP_PRE;
if ((s->control_command = s->exec_command[SOCKET_EXEC_STOP_PRE])) {
if ((r = socket_spawn(s, s->control_command, &s->control_pid)) < 0)
goto fail;
socket_set_state(s, SOCKET_STOP_PRE);
} else
socket_enter_stop_post(s, true);
return;
fail:
log_warning("%s failed to run 'stop-pre' task: %s", s->meta.id, strerror(-r));
socket_enter_stop_post(s, false);
}
static void socket_enter_listening(Socket *s) {
int r;
assert(s);
if ((r = socket_watch_fds(s)) < 0) {
log_warning("%s failed to watch sockets: %s", s->meta.id, strerror(-r));
goto fail;
}
socket_set_state(s, SOCKET_LISTENING);
return;
fail:
socket_enter_stop_pre(s, false);
}
static void socket_enter_start_post(Socket *s) {
int r;
assert(s);
if ((r = socket_open_fds(s)) < 0) {
log_warning("%s failed to listen on sockets: %s", s->meta.id, strerror(-r));
goto fail;
}
socket_unwatch_control_pid(s);
s->control_command_id = SOCKET_EXEC_START_POST;
if ((s->control_command = s->exec_command[SOCKET_EXEC_START_POST])) {
if ((r = socket_spawn(s, s->control_command, &s->control_pid)) < 0) {
log_warning("%s failed to run 'start-post' task: %s", s->meta.id, strerror(-r));
goto fail;
}
socket_set_state(s, SOCKET_START_POST);
} else
socket_enter_listening(s);
return;
fail:
socket_enter_stop_pre(s, false);
}
static void socket_enter_start_pre(Socket *s) {
int r;
assert(s);
socket_unwatch_control_pid(s);
s->control_command_id = SOCKET_EXEC_START_PRE;
if ((s->control_command = s->exec_command[SOCKET_EXEC_START_PRE])) {
if ((r = socket_spawn(s, s->control_command, &s->control_pid)) < 0)
goto fail;
socket_set_state(s, SOCKET_START_PRE);
} else
socket_enter_start_post(s);
return;
fail:
log_warning("%s failed to run 'start-pre' task: %s", s->meta.id, strerror(-r));
socket_enter_dead(s, false);
}
static void socket_enter_running(Socket *s, int cfd) {
int r;
DBusError error;
assert(s);
dbus_error_init(&error);
/* We don't take connections anymore if we are supposed to
* shut down anyway */
if (s->meta.job && s->meta.job->type == JOB_STOP) {
if (cfd >= 0)
close_nointr_nofail(cfd);
else {
/* Flush all sockets by closing and reopening them */
socket_close_fds(s);
if ((r = socket_watch_fds(s)) < 0) {
log_warning("%s failed to watch sockets: %s", s->meta.id, strerror(-r));
socket_enter_stop_pre(s, false);
}
}
return;
}
if (cfd < 0) {
if ((r = manager_add_job(s->meta.manager, JOB_START, UNIT(s->service), JOB_REPLACE, true, &error, NULL)) < 0)
goto fail;
socket_set_state(s, SOCKET_RUNNING);
} else {
char *prefix, *instance = NULL, *name;
Service *service;
if (s->n_connections >= s->max_connections) {
log_warning("Too many incoming connections (%u)", s->n_connections);
close_nointr_nofail(cfd);
return;
}
if ((r = socket_instantiate_service(s)) < 0)
goto fail;
if ((r = instance_from_socket(cfd, s->n_accepted, &instance)) < 0)
goto fail;
if (!(prefix = unit_name_to_prefix(s->meta.id))) {
free(instance);
r = -ENOMEM;
goto fail;
}
name = unit_name_build(prefix, instance, ".service");
free(prefix);
free(instance);
if (!name) {
r = -ENOMEM;
goto fail;
}
if ((r = unit_add_name(UNIT(s->service), name)) < 0) {
free(name);
goto fail;
}
service = s->service;
s->service = NULL;
s->n_accepted ++;
unit_choose_id(UNIT(service), name);
free(name);
if ((r = service_set_socket_fd(service, cfd, s)) < 0)
goto fail;
cfd = -1;
s->n_connections ++;
if ((r = manager_add_job(s->meta.manager, JOB_START, UNIT(service), JOB_REPLACE, true, &error, NULL)) < 0)
goto fail;
}
return;
fail:
log_warning("%s failed to queue socket startup job: %s", s->meta.id, bus_error(&error, r));
socket_enter_stop_pre(s, false);
if (cfd >= 0)
close_nointr_nofail(cfd);
dbus_error_free(&error);
}
static void socket_run_next(Socket *s, bool success) {
int r;
assert(s);
assert(s->control_command);
assert(s->control_command->command_next);
if (!success)
s->failure = true;
socket_unwatch_control_pid(s);
s->control_command = s->control_command->command_next;
if ((r = socket_spawn(s, s->control_command, &s->control_pid)) < 0)
goto fail;
return;
fail:
log_warning("%s failed to run next task: %s", s->meta.id, strerror(-r));
if (s->state == SOCKET_START_POST)
socket_enter_stop_pre(s, false);
else if (s->state == SOCKET_STOP_POST)
socket_enter_dead(s, false);
else
socket_enter_signal(s, SOCKET_FINAL_SIGTERM, false);
}
static int socket_start(Unit *u) {
Socket *s = SOCKET(u);
assert(s);
/* We cannot fulfill this request right now, try again later
* please! */
if (s->state == SOCKET_STOP_PRE ||
s->state == SOCKET_STOP_PRE_SIGKILL ||
s->state == SOCKET_STOP_PRE_SIGTERM ||
s->state == SOCKET_STOP_POST ||
s->state == SOCKET_FINAL_SIGTERM ||
s->state == SOCKET_FINAL_SIGKILL)
return -EAGAIN;
if (s->state == SOCKET_START_PRE ||
s->state == SOCKET_START_POST)
return 0;
/* Cannot run this without the service being around */
if (s->service) {
if (s->service->meta.load_state != UNIT_LOADED)
return -ENOENT;
/* If the service is alredy actvie we cannot start the
* socket */
if (s->service->state != SERVICE_DEAD &&
s->service->state != SERVICE_MAINTENANCE &&
s->service->state != SERVICE_AUTO_RESTART)
return -EBUSY;
}
assert(s->state == SOCKET_DEAD || s->state == SOCKET_MAINTENANCE);
s->failure = false;
socket_enter_start_pre(s);
return 0;
}
static int socket_stop(Unit *u) {
Socket *s = SOCKET(u);
assert(s);
/* Already on it */
if (s->state == SOCKET_STOP_PRE ||
s->state == SOCKET_STOP_PRE_SIGTERM ||
s->state == SOCKET_STOP_PRE_SIGKILL ||
s->state == SOCKET_STOP_POST ||
s->state == SOCKET_FINAL_SIGTERM ||
s->state == SOCKET_FINAL_SIGKILL)
return 0;
/* If there's already something running we go directly into
* kill mode. */
if (s->state == SOCKET_START_PRE ||
s->state == SOCKET_START_POST) {
socket_enter_signal(s, SOCKET_STOP_PRE_SIGTERM, true);
return -EAGAIN;
}
assert(s->state == SOCKET_LISTENING || s->state == SOCKET_RUNNING);
socket_enter_stop_pre(s, true);
return 0;
}
static int socket_serialize(Unit *u, FILE *f, FDSet *fds) {
Socket *s = SOCKET(u);
SocketPort *p;
int r;
assert(u);
assert(f);
assert(fds);
unit_serialize_item(u, f, "state", socket_state_to_string(s->state));
unit_serialize_item(u, f, "failure", yes_no(s->failure));
unit_serialize_item_format(u, f, "n-accepted", "%u", s->n_accepted);
if (s->control_pid > 0)
unit_serialize_item_format(u, f, "control-pid", "%lu", (unsigned long) s->control_pid);
if (s->control_command_id >= 0)
unit_serialize_item(u, f, "control-command", socket_exec_command_to_string(s->control_command_id));
LIST_FOREACH(port, p, s->ports) {
int copy;
if (p->fd < 0)
continue;
if ((copy = fdset_put_dup(fds, p->fd)) < 0)
return copy;
if (p->type == SOCKET_SOCKET) {
char *t;
if ((r = socket_address_print(&p->address, &t)) < 0)
return r;
unit_serialize_item_format(u, f, "socket", "%i %i %s", copy, p->address.type, t);
free(t);
} else {
assert(p->type == SOCKET_FIFO);
unit_serialize_item_format(u, f, "fifo", "%i %s", copy, p->path);
}
}
return 0;
}
static int socket_deserialize_item(Unit *u, const char *key, const char *value, FDSet *fds) {
Socket *s = SOCKET(u);
int r;
assert(u);
assert(key);
assert(value);
assert(fds);
if (streq(key, "state")) {
SocketState state;
if ((state = socket_state_from_string(value)) < 0)
log_debug("Failed to parse state value %s", value);
else
s->deserialized_state = state;
} else if (streq(key, "failure")) {
int b;
if ((b = parse_boolean(value)) < 0)
log_debug("Failed to parse failure value %s", value);
else
s->failure = b || s->failure;
} else if (streq(key, "n-accepted")) {
unsigned k;
if ((r = safe_atou(value, &k)) < 0)
log_debug("Failed to parse n-accepted value %s", value);
else
s->n_accepted += k;
} else if (streq(key, "control-pid")) {
pid_t pid;
if ((r = parse_pid(value, &pid)) < 0)
log_debug("Failed to parse control-pid value %s", value);
else
s->control_pid = pid;
} else if (streq(key, "control-command")) {
SocketExecCommand id;
if ((id = socket_exec_command_from_string(value)) < 0)
log_debug("Failed to parse exec-command value %s", value);
else {
s->control_command_id = id;
s->control_command = s->exec_command[id];
}
} else if (streq(key, "fifo")) {
int fd, skip = 0;
SocketPort *p;
if (sscanf(value, "%i %n", &fd, &skip) < 1 || fd < 0 || !fdset_contains(fds, fd))
log_debug("Failed to parse fifo value %s", value);
else {
LIST_FOREACH(port, p, s->ports)
if (streq(p->path, value+skip))
break;
if (p) {
if (p->fd >= 0)
close_nointr_nofail(p->fd);
p->fd = fdset_remove(fds, fd);
}
}
} else if (streq(key, "socket")) {
int fd, type, skip = 0;
SocketPort *p;
if (sscanf(value, "%i %i %n", &fd, &type, &skip) < 2 || fd < 0 || type < 0 || !fdset_contains(fds, fd))
log_debug("Failed to parse socket value %s", value);
else {
LIST_FOREACH(port, p, s->ports)
if (socket_address_is(&p->address, value+skip, type))
break;
if (p) {
if (p->fd >= 0)
close_nointr_nofail(p->fd);
p->fd = fdset_remove(fds, fd);
}
}
} else
log_debug("Unknown serialization key '%s'", key);
return 0;
}
static UnitActiveState socket_active_state(Unit *u) {
assert(u);
return state_translation_table[SOCKET(u)->state];
}
static const char *socket_sub_state_to_string(Unit *u) {
assert(u);
return socket_state_to_string(SOCKET(u)->state);
}
static bool socket_check_gc(Unit *u) {
Socket *s = SOCKET(u);
assert(u);
return s->n_connections > 0;
}
static void socket_fd_event(Unit *u, int fd, uint32_t events, Watch *w) {
Socket *s = SOCKET(u);
int cfd = -1;
assert(s);
assert(fd >= 0);
if (s->state != SOCKET_LISTENING)
return;
log_debug("Incoming traffic on %s", u->meta.id);
if (events != EPOLLIN) {
log_error("Got invalid poll event on socket.");
goto fail;
}
if (w->socket_accept) {
for (;;) {
if ((cfd = accept4(fd, NULL, NULL, SOCK_NONBLOCK)) < 0) {
if (errno == EINTR)
continue;
log_error("Failed to accept socket: %m");
goto fail;
}
break;
}
socket_apply_socket_options(s, cfd);
}
socket_enter_running(s, cfd);
return;
fail:
socket_enter_stop_pre(s, false);
}
static void socket_sigchld_event(Unit *u, pid_t pid, int code, int status) {
Socket *s = SOCKET(u);
bool success;
assert(s);
assert(pid >= 0);
if (pid != s->control_pid)
return;
s->control_pid = 0;
success = is_clean_exit(code, status);
if (s->control_command) {
exec_status_exit(&s->control_command->exec_status, pid, code, status);
if (s->control_command->ignore)
success = true;
}
log_full(success ? LOG_DEBUG : LOG_NOTICE,
"%s control process exited, code=%s status=%i", u->meta.id, sigchld_code_to_string(code), status);
s->failure = s->failure || !success;
if (s->control_command && s->control_command->command_next && success) {
log_debug("%s running next command for state %s", u->meta.id, socket_state_to_string(s->state));
socket_run_next(s, success);
} else {
s->control_command = NULL;
s->control_command_id = _SOCKET_EXEC_COMMAND_INVALID;
/* No further commands for this step, so let's figure
* out what to do next */
log_debug("%s got final SIGCHLD for state %s", u->meta.id, socket_state_to_string(s->state));
switch (s->state) {
case SOCKET_START_PRE:
if (success)
socket_enter_start_post(s);
else
socket_enter_signal(s, SOCKET_FINAL_SIGTERM, false);
break;
case SOCKET_START_POST:
if (success)
socket_enter_listening(s);
else
socket_enter_stop_pre(s, false);
break;
case SOCKET_STOP_PRE:
case SOCKET_STOP_PRE_SIGTERM:
case SOCKET_STOP_PRE_SIGKILL:
socket_enter_stop_post(s, success);
break;
case SOCKET_STOP_POST:
case SOCKET_FINAL_SIGTERM:
case SOCKET_FINAL_SIGKILL:
socket_enter_dead(s, success);
break;
default:
assert_not_reached("Uh, control process died at wrong time.");
}
}
}
static void socket_timer_event(Unit *u, uint64_t elapsed, Watch *w) {
Socket *s = SOCKET(u);
assert(s);
assert(elapsed == 1);
assert(w == &s->timer_watch);
switch (s->state) {
case SOCKET_START_PRE:
log_warning("%s starting timed out. Terminating.", u->meta.id);
socket_enter_signal(s, SOCKET_FINAL_SIGTERM, false);
case SOCKET_START_POST:
log_warning("%s starting timed out. Stopping.", u->meta.id);
socket_enter_stop_pre(s, false);
break;
case SOCKET_STOP_PRE:
log_warning("%s stopping timed out. Terminating.", u->meta.id);
socket_enter_signal(s, SOCKET_STOP_PRE_SIGTERM, false);
break;
case SOCKET_STOP_PRE_SIGTERM:
log_warning("%s stopping timed out. Killing.", u->meta.id);
socket_enter_signal(s, SOCKET_STOP_PRE_SIGKILL, false);
break;
case SOCKET_STOP_PRE_SIGKILL:
log_warning("%s still around after SIGKILL. Ignoring.", u->meta.id);
socket_enter_stop_post(s, false);
break;
case SOCKET_STOP_POST:
log_warning("%s stopping timed out (2). Terminating.", u->meta.id);
socket_enter_signal(s, SOCKET_FINAL_SIGTERM, false);
break;
case SOCKET_FINAL_SIGTERM:
log_warning("%s stopping timed out (2). Killing.", u->meta.id);
socket_enter_signal(s, SOCKET_FINAL_SIGKILL, false);
break;
case SOCKET_FINAL_SIGKILL:
log_warning("%s still around after SIGKILL (2). Entering maintenance mode.", u->meta.id);
socket_enter_dead(s, false);
break;
default:
assert_not_reached("Timeout at wrong time.");
}
}
int socket_collect_fds(Socket *s, int **fds, unsigned *n_fds) {
int *rfds;
unsigned rn_fds, k;
SocketPort *p;
assert(s);
assert(fds);
assert(n_fds);
/* Called from the service code for requesting our fds */
rn_fds = 0;
LIST_FOREACH(port, p, s->ports)
if (p->fd >= 0)
rn_fds++;
if (!(rfds = new(int, rn_fds)) < 0)
return -ENOMEM;
k = 0;
LIST_FOREACH(port, p, s->ports)
if (p->fd >= 0)
rfds[k++] = p->fd;
assert(k == rn_fds);
*fds = rfds;
*n_fds = rn_fds;
return 0;
}
void socket_notify_service_dead(Socket *s) {
assert(s);
/* The service is dead. Dang!
*
* This is strictly for one-instance-for-all-connections
* services. */
if (s->state == SOCKET_RUNNING) {
log_debug("%s got notified about service death.", s->meta.id);
socket_enter_listening(s);
}
}
void socket_connection_unref(Socket *s) {
assert(s);
/* The service is dead. Yay!
*
* This is strictly for one-onstance-per-connection
* services. */
assert(s->n_connections > 0);
s->n_connections--;
log_debug("%s: One connection closed, %u left.", s->meta.id, s->n_connections);
}
static void socket_reset_maintenance(Unit *u) {
Socket *s = SOCKET(u);
assert(s);
if (s->state == SOCKET_MAINTENANCE)
socket_set_state(s, SOCKET_DEAD);
s->failure = false;
}
static const char* const socket_state_table[_SOCKET_STATE_MAX] = {
[SOCKET_DEAD] = "dead",
[SOCKET_START_PRE] = "start-pre",
[SOCKET_START_POST] = "start-post",
[SOCKET_LISTENING] = "listening",
[SOCKET_RUNNING] = "running",
[SOCKET_STOP_PRE] = "stop-pre",
[SOCKET_STOP_PRE_SIGTERM] = "stop-pre-sigterm",
[SOCKET_STOP_PRE_SIGKILL] = "stop-pre-sigkill",
[SOCKET_STOP_POST] = "stop-post",
[SOCKET_FINAL_SIGTERM] = "final-sigterm",
[SOCKET_FINAL_SIGKILL] = "final-sigkill",
[SOCKET_MAINTENANCE] = "maintenance"
};
DEFINE_STRING_TABLE_LOOKUP(socket_state, SocketState);
static const char* const socket_exec_command_table[_SOCKET_EXEC_COMMAND_MAX] = {
[SOCKET_EXEC_START_PRE] = "StartPre",
[SOCKET_EXEC_START_POST] = "StartPost",
[SOCKET_EXEC_STOP_PRE] = "StopPre",
[SOCKET_EXEC_STOP_POST] = "StopPost"
};
DEFINE_STRING_TABLE_LOOKUP(socket_exec_command, SocketExecCommand);
const UnitVTable socket_vtable = {
.suffix = ".socket",
.init = socket_init,
.done = socket_done,
.load = socket_load,
.coldplug = socket_coldplug,
.dump = socket_dump,
.start = socket_start,
.stop = socket_stop,
.serialize = socket_serialize,
.deserialize_item = socket_deserialize_item,
.active_state = socket_active_state,
.sub_state_to_string = socket_sub_state_to_string,
.check_gc = socket_check_gc,
.fd_event = socket_fd_event,
.sigchld_event = socket_sigchld_event,
.timer_event = socket_timer_event,
.reset_maintenance = socket_reset_maintenance,
.bus_message_handler = bus_socket_message_handler
};