+ if (r < 0 && ret >= 0)
+ ret = r;
+ }
+
+ r = cg_enumerate_subgroups(SYSTEMD_CGROUP_CONTROLLER, path, &d);
+ if (r < 0) {
+ if (ret >= 0)
+ ret = r;
+ } else {
+ char *fn;
+
+ while ((r = cg_read_subgroup(d, &fn)) > 0) {
+ _cleanup_free_ char *p = NULL;
+
+ p = strjoin(path, "/", fn);
+ free(fn);
+
+ if (!p)
+ return -ENOMEM;
+
+ r = unit_watch_pids_in_path(u, p);
+ if (r < 0 && ret >= 0)
+ ret = r;
+ }
+
+ if (r < 0 && ret >= 0)
+ ret = r;
+ }
+
+ return ret;
+}
+
+int unit_watch_all_pids(Unit *u) {
+ int r;
+
+ assert(u);
+
+ /* Adds all PIDs from our cgroup to the set of PIDs we
+ * watch. This is a fallback logic for cases where we do not
+ * get reliable cgroup empty notifications: we try to use
+ * SIGCHLD as replacement. */
+
+ if (!u->cgroup_path)
+ return -ENOENT;
+
+ r = cg_unified_controller(SYSTEMD_CGROUP_CONTROLLER);
+ if (r < 0)
+ return r;
+ if (r > 0) /* On unified we can use proper notifications */
+ return 0;
+
+ return unit_watch_pids_in_path(u, u->cgroup_path);
+}
+
+int unit_notify_cgroup_empty(Unit *u) {
+ int r;
+
+ assert(u);
+
+ if (!u->cgroup_path)
+ return 0;
+
+ r = cg_is_empty_recursive(SYSTEMD_CGROUP_CONTROLLER, u->cgroup_path);
+ if (r <= 0)
+ return r;
+
+ unit_add_to_gc_queue(u);
+
+ if (UNIT_VTABLE(u)->notify_cgroup_empty)
+ UNIT_VTABLE(u)->notify_cgroup_empty(u);
+
+ return 0;
+}
+
+static int on_cgroup_inotify_event(sd_event_source *s, int fd, uint32_t revents, void *userdata) {
+ Manager *m = userdata;
+
+ assert(s);
+ assert(fd >= 0);
+ assert(m);
+
+ for (;;) {
+ union inotify_event_buffer buffer;
+ struct inotify_event *e;
+ ssize_t l;
+
+ l = read(fd, &buffer, sizeof(buffer));
+ if (l < 0) {
+ if (errno == EINTR || errno == EAGAIN)
+ return 0;
+
+ return log_error_errno(errno, "Failed to read control group inotify events: %m");
+ }
+
+ FOREACH_INOTIFY_EVENT(e, buffer, l) {
+ Unit *u;
+
+ if (e->wd < 0)
+ /* Queue overflow has no watch descriptor */
+ continue;
+
+ if (e->mask & IN_IGNORED)
+ /* The watch was just removed */
+ continue;
+
+ u = hashmap_get(m->cgroup_inotify_wd_unit, INT_TO_PTR(e->wd));
+ if (!u) /* Not that inotify might deliver
+ * events for a watch even after it
+ * was removed, because it was queued
+ * before the removal. Let's ignore
+ * this here safely. */
+ continue;
+
+ (void) unit_notify_cgroup_empty(u);
+ }
+ }
+}
+#endif // 0
+
+int manager_setup_cgroup(Manager *m) {
+ _cleanup_free_ char *path = NULL;
+ CGroupController c;
+ int r, all_unified;
+ char *e;
+
+ assert(m);
+
+ /* 1. Determine hierarchy */
+ m->cgroup_root = mfree(m->cgroup_root);
+#if 0 /// elogind is not init and must therefore search for PID 1 instead of self.
+ r = cg_pid_get_path(SYSTEMD_CGROUP_CONTROLLER, 0, &m->cgroup_root);
+#else
+ r = cg_pid_get_path(SYSTEMD_CGROUP_CONTROLLER, 1, &m->cgroup_root);
+#endif // 0
+ if (r < 0)
+ return log_error_errno(r, "Cannot determine cgroup we are running in: %m");
+
+#if 0 /// elogind does not support systemd scopes and slices
+ /* Chop off the init scope, if we are already located in it */
+ e = endswith(m->cgroup_root, "/" SPECIAL_INIT_SCOPE);
+
+ /* LEGACY: Also chop off the system slice if we are in
+ * it. This is to support live upgrades from older systemd
+ * versions where PID 1 was moved there. Also see
+ * cg_get_root_path(). */
+ if (!e && MANAGER_IS_SYSTEM(m)) {
+ e = endswith(m->cgroup_root, "/" SPECIAL_SYSTEM_SLICE);
+ if (!e)
+ e = endswith(m->cgroup_root, "/system"); /* even more legacy */
+ }
+ if (e)
+ *e = 0;
+#endif // 0
+
+ /* And make sure to store away the root value without trailing
+ * slash, even for the root dir, so that we can easily prepend
+ * it everywhere. */
+ while ((e = endswith(m->cgroup_root, "/")))
+ *e = 0;
+ log_debug_elogind("Cgroup Controller \"%s\" -> root \"%s\"",
+ SYSTEMD_CGROUP_CONTROLLER, m->cgroup_root);
+
+ /* 2. Show data */
+ r = cg_get_path(SYSTEMD_CGROUP_CONTROLLER, m->cgroup_root, NULL, &path);
+ if (r < 0)
+ return log_error_errno(r, "Cannot find cgroup mount point: %m");
+
+ r = cg_unified_flush();
+ if (r < 0)
+ return log_error_errno(r, "Couldn't determine if we are running in the unified hierarchy: %m");
+
+ all_unified = cg_all_unified();
+ if (r < 0)
+ return log_error_errno(r, "Couldn't determine whether we are in all unified mode: %m");
+ if (r > 0)
+ log_debug("Unified cgroup hierarchy is located at %s.", path);
+ else {
+ r = cg_unified_controller(SYSTEMD_CGROUP_CONTROLLER);
+ if (r < 0)
+ return log_error_errno(r, "Failed to determine whether systemd's own controller is in unified mode: %m");
+ if (r > 0)
+ log_debug("Unified cgroup hierarchy is located at %s. Controllers are on legacy hierarchies.", path);
+ else
+ log_debug("Using cgroup controller " SYSTEMD_CGROUP_CONTROLLER_LEGACY ". File system hierarchy is at %s.", path);
+ }
+
+ if (!m->test_run) {
+ const char *scope_path;
+
+#if 0 /// elogind is not init, and does not install the agent here.
+ /* 3. Install agent */
+ if (cg_unified_controller(SYSTEMD_CGROUP_CONTROLLER) > 0) {
+
+ /* In the unified hierarchy we can get
+ * cgroup empty notifications via inotify. */
+
+ m->cgroup_inotify_event_source = sd_event_source_unref(m->cgroup_inotify_event_source);
+ safe_close(m->cgroup_inotify_fd);
+
+ m->cgroup_inotify_fd = inotify_init1(IN_NONBLOCK|IN_CLOEXEC);
+ if (m->cgroup_inotify_fd < 0)
+ return log_error_errno(errno, "Failed to create control group inotify object: %m");
+
+ r = sd_event_add_io(m->event, &m->cgroup_inotify_event_source, m->cgroup_inotify_fd, EPOLLIN, on_cgroup_inotify_event, m);
+ if (r < 0)
+ return log_error_errno(r, "Failed to watch control group inotify object: %m");
+
+ /* Process cgroup empty notifications early, but after service notifications and SIGCHLD. Also
+ * see handling of cgroup agent notifications, for the classic cgroup hierarchy support. */
+ r = sd_event_source_set_priority(m->cgroup_inotify_event_source, SD_EVENT_PRIORITY_NORMAL-5);
+ if (r < 0)
+ return log_error_errno(r, "Failed to set priority of inotify event source: %m");
+
+ (void) sd_event_source_set_description(m->cgroup_inotify_event_source, "cgroup-inotify");
+
+ } else if (MANAGER_IS_SYSTEM(m)) {
+
+ /* On the legacy hierarchy we only get
+ * notifications via cgroup agents. (Which
+ * isn't really reliable, since it does not
+ * generate events when control groups with
+ * children run empty. */
+
+ r = cg_install_release_agent(SYSTEMD_CGROUP_CONTROLLER, SYSTEMD_CGROUP_AGENT_PATH);
+ if (r < 0)
+ log_warning_errno(r, "Failed to install release agent, ignoring: %m");
+ else if (r > 0)
+ log_debug("Installed release agent.");
+ else if (r == 0)
+ log_debug("Release agent already installed.");
+ }
+
+ /* 4. Make sure we are in the special "init.scope" unit in the root slice. */
+ scope_path = strjoina(m->cgroup_root, "/" SPECIAL_INIT_SCOPE);
+ r = cg_create_and_attach(SYSTEMD_CGROUP_CONTROLLER, scope_path, 0);
+#else
+ /* Note:
+ * This method is in core, and normally called by systemd
+ * being init. As elogind is never init, we can not install
+ * our agent here. We do so when mounting our cgroup file
+ * system, so only if elogind is its own tiny controller.
+ * Further, elogind is not meant to run in systemd init scope. */
+ if (MANAGER_IS_SYSTEM(m))
+ // we are our own cgroup controller
+ scope_path = strjoina("");
+ else if (streq(m->cgroup_root, "/elogind"))
+ // root already is our cgroup
+ scope_path = strjoina(m->cgroup_root);
+ else
+ // we have to create our own group
+ scope_path = strjoina(m->cgroup_root, "/elogind");
+ r = cg_create_and_attach(SYSTEMD_CGROUP_CONTROLLER, scope_path, 0);
+#endif // 0
+ if (r < 0)
+ return log_error_errno(r, "Failed to create %s control group: %m", scope_path);
+ log_debug_elogind("Created control group \"%s\"", scope_path);
+
+#if 0 /// elogind is not a "sub-controller" like systemd, so migration is not needed.
+ /* also, move all other userspace processes remaining
+ * in the root cgroup into that scope. */
+ r = cg_migrate(SYSTEMD_CGROUP_CONTROLLER, m->cgroup_root, SYSTEMD_CGROUP_CONTROLLER, scope_path, 0);
+ if (r < 0)
+ log_warning_errno(r, "Couldn't move remaining userspace processes, ignoring: %m");
+#endif // 0
+
+ /* 5. And pin it, so that it cannot be unmounted */
+ safe_close(m->pin_cgroupfs_fd);
+ m->pin_cgroupfs_fd = open(path, O_RDONLY|O_CLOEXEC|O_DIRECTORY|O_NOCTTY|O_NONBLOCK);
+ if (m->pin_cgroupfs_fd < 0)
+ return log_error_errno(errno, "Failed to open pin file: %m");
+
+ /* 6. Always enable hierarchical support if it exists... */
+ if (!all_unified)
+ (void) cg_set_attribute("memory", "/", "memory.use_hierarchy", "1");
+ }
+
+ /* 7. Figure out which controllers are supported */
+ r = cg_mask_supported(&m->cgroup_supported);
+ if (r < 0)
+ return log_error_errno(r, "Failed to determine supported controllers: %m");
+
+ for (c = 0; c < _CGROUP_CONTROLLER_MAX; c++)
+ log_debug("Controller '%s' supported: %s", cgroup_controller_to_string(c), yes_no(m->cgroup_supported & CGROUP_CONTROLLER_TO_MASK(c)));
+
+ return 0;
+}
+
+void manager_shutdown_cgroup(Manager *m, bool delete) {
+ assert(m);
+
+ /* We can't really delete the group, since we are in it. But
+ * let's trim it. */
+ if (delete && m->cgroup_root)
+ (void) cg_trim(SYSTEMD_CGROUP_CONTROLLER, m->cgroup_root, false);
+
+#if 0 /// elogind does not support the unified hierarchy, yet.
+ m->cgroup_inotify_wd_unit = hashmap_free(m->cgroup_inotify_wd_unit);
+
+ m->cgroup_inotify_event_source = sd_event_source_unref(m->cgroup_inotify_event_source);
+ m->cgroup_inotify_fd = safe_close(m->cgroup_inotify_fd);
+#endif // 0
+
+ m->pin_cgroupfs_fd = safe_close(m->pin_cgroupfs_fd);
+
+ m->cgroup_root = mfree(m->cgroup_root);
+}
+
+#if 0 /// UNNEEDED by elogind
+Unit* manager_get_unit_by_cgroup(Manager *m, const char *cgroup) {
+ char *p;
+ Unit *u;
+
+ assert(m);
+ assert(cgroup);
+
+ u = hashmap_get(m->cgroup_unit, cgroup);
+ if (u)
+ return u;
+
+ p = strdupa(cgroup);
+ for (;;) {
+ char *e;
+
+ e = strrchr(p, '/');
+ if (!e || e == p)
+ return hashmap_get(m->cgroup_unit, SPECIAL_ROOT_SLICE);
+
+ *e = 0;
+
+ u = hashmap_get(m->cgroup_unit, p);
+ if (u)
+ return u;
+ }
+}
+
+Unit *manager_get_unit_by_pid_cgroup(Manager *m, pid_t pid) {
+ _cleanup_free_ char *cgroup = NULL;
+ int r;
+
+ assert(m);
+
+ if (pid <= 0)
+ return NULL;
+
+ r = cg_pid_get_path(SYSTEMD_CGROUP_CONTROLLER, pid, &cgroup);
+ if (r < 0)
+ return NULL;
+
+ return manager_get_unit_by_cgroup(m, cgroup);
+}
+
+Unit *manager_get_unit_by_pid(Manager *m, pid_t pid) {
+ Unit *u;
+
+ assert(m);
+
+ if (pid <= 0)
+ return NULL;
+
+ if (pid == 1)
+ return hashmap_get(m->units, SPECIAL_INIT_SCOPE);
+
+ u = hashmap_get(m->watch_pids1, PID_TO_PTR(pid));
+ if (u)
+ return u;
+
+ u = hashmap_get(m->watch_pids2, PID_TO_PTR(pid));
+ if (u)
+ return u;
+
+ return manager_get_unit_by_pid_cgroup(m, pid);
+}
+#endif // 0
+
+#if 0 /// elogind must substitute this with its own variant
+int manager_notify_cgroup_empty(Manager *m, const char *cgroup) {
+ Unit *u;
+
+ assert(m);
+ assert(cgroup);
+
+ log_debug("Got cgroup empty notification for: %s", cgroup);
+
+ u = manager_get_unit_by_cgroup(m, cgroup);
+ if (!u)
+ return 0;
+
+ return unit_notify_cgroup_empty(u);
+}
+#else
+int manager_notify_cgroup_empty(Manager *m, const char *cgroup) {
+ Session *s;
+
+ assert(m);
+ assert(cgroup);
+
+ log_debug("Got cgroup empty notification for: %s", cgroup);
+
+ s = hashmap_get(m->sessions, cgroup);
+
+ if (s) {
+ session_finalize(s);
+ session_free(s);
+ } else
+ log_warning("Session not found: %s", cgroup);
+
+ return 0;
+}
+#endif // 0
+#if 0 /// UNNEEDED by elogind
+int unit_get_memory_current(Unit *u, uint64_t *ret) {
+ _cleanup_free_ char *v = NULL;
+ int r;
+
+ assert(u);
+ assert(ret);
+
+ if (!u->cgroup_path)
+ return -ENODATA;
+
+ if ((u->cgroup_realized_mask & CGROUP_MASK_MEMORY) == 0)
+ return -ENODATA;
+
+ r = cg_all_unified();
+ if (r < 0)
+ return r;
+ if (r > 0)
+ r = cg_get_attribute("memory", u->cgroup_path, "memory.current", &v);
+ else
+ r = cg_get_attribute("memory", u->cgroup_path, "memory.usage_in_bytes", &v);
+ if (r == -ENOENT)
+ return -ENODATA;
+ if (r < 0)
+ return r;
+
+ return safe_atou64(v, ret);
+}
+
+int unit_get_tasks_current(Unit *u, uint64_t *ret) {
+ _cleanup_free_ char *v = NULL;
+ int r;
+
+ assert(u);
+ assert(ret);
+
+ if (!u->cgroup_path)
+ return -ENODATA;
+
+ if ((u->cgroup_realized_mask & CGROUP_MASK_PIDS) == 0)
+ return -ENODATA;
+
+ r = cg_get_attribute("pids", u->cgroup_path, "pids.current", &v);
+ if (r == -ENOENT)
+ return -ENODATA;
+ if (r < 0)
+ return r;
+
+ return safe_atou64(v, ret);
+}
+
+static int unit_get_cpu_usage_raw(Unit *u, nsec_t *ret) {
+ _cleanup_free_ char *v = NULL;
+ uint64_t ns;
+ int r;
+
+ assert(u);
+ assert(ret);
+
+ if (!u->cgroup_path)
+ return -ENODATA;
+
+ r = cg_all_unified();
+ if (r < 0)
+ return r;
+ if (r > 0) {
+ const char *keys[] = { "usage_usec", NULL };
+ _cleanup_free_ char *val = NULL;
+ uint64_t us;
+
+ if ((u->cgroup_realized_mask & CGROUP_MASK_CPU) == 0)
+ return -ENODATA;
+
+ r = cg_get_keyed_attribute("cpu", u->cgroup_path, "cpu.stat", keys, &val);
+ if (r < 0)
+ return r;
+
+ r = safe_atou64(val, &us);
+ if (r < 0)
+ return r;
+
+ ns = us * NSEC_PER_USEC;
+ } else {
+ if ((u->cgroup_realized_mask & CGROUP_MASK_CPUACCT) == 0)
+ return -ENODATA;
+
+ r = cg_get_attribute("cpuacct", u->cgroup_path, "cpuacct.usage", &v);
+ if (r == -ENOENT)
+ return -ENODATA;
+ if (r < 0)
+ return r;
+
+ r = safe_atou64(v, &ns);
+ if (r < 0)
+ return r;
+ }
+
+ *ret = ns;
+ return 0;
+}
+
+int unit_get_cpu_usage(Unit *u, nsec_t *ret) {
+ nsec_t ns;
+ int r;
+
+ assert(u);
+
+ /* Retrieve the current CPU usage counter. This will subtract the CPU counter taken when the unit was
+ * started. If the cgroup has been removed already, returns the last cached value. To cache the value, simply
+ * call this function with a NULL return value. */
+
+ r = unit_get_cpu_usage_raw(u, &ns);
+ if (r == -ENODATA && u->cpu_usage_last != NSEC_INFINITY) {
+ /* If we can't get the CPU usage anymore (because the cgroup was already removed, for example), use our
+ * cached value. */
+
+ if (ret)
+ *ret = u->cpu_usage_last;
+ return 0;
+ }
+ if (r < 0)
+ return r;
+
+ if (ns > u->cpu_usage_base)
+ ns -= u->cpu_usage_base;
+ else
+ ns = 0;
+
+ u->cpu_usage_last = ns;
+ if (ret)
+ *ret = ns;
+
+ return 0;
+}
+
+int unit_reset_cpu_usage(Unit *u) {
+ nsec_t ns;
+ int r;
+
+ assert(u);
+
+ u->cpu_usage_last = NSEC_INFINITY;
+
+ r = unit_get_cpu_usage_raw(u, &ns);
+ if (r < 0) {
+ u->cpu_usage_base = 0;
+ return r;
+ }
+
+ u->cpu_usage_base = ns;
+ return 0;
+}
+
+bool unit_cgroup_delegate(Unit *u) {
+ CGroupContext *c;
+
+ assert(u);
+
+ c = unit_get_cgroup_context(u);
+ if (!c)
+ return false;
+
+ return c->delegate;
+}
+
+void unit_invalidate_cgroup(Unit *u, CGroupMask m) {
+ assert(u);
+
+ if (!UNIT_HAS_CGROUP_CONTEXT(u))
+ return;
+
+ if (m == 0)
+ return;
+
+ /* always invalidate compat pairs together */
+ if (m & (CGROUP_MASK_IO | CGROUP_MASK_BLKIO))
+ m |= CGROUP_MASK_IO | CGROUP_MASK_BLKIO;
+
+ if ((u->cgroup_realized_mask & m) == 0)
+ return;
+
+ u->cgroup_realized_mask &= ~m;
+ unit_add_to_cgroup_queue(u);
+}
+
+void manager_invalidate_startup_units(Manager *m) {
+ Iterator i;
+ Unit *u;
+
+ assert(m);
+
+ SET_FOREACH(u, m->startup_units, i)
+ unit_invalidate_cgroup(u, CGROUP_MASK_CPU|CGROUP_MASK_IO|CGROUP_MASK_BLKIO);
+}
+
+static const char* const cgroup_device_policy_table[_CGROUP_DEVICE_POLICY_MAX] = {
+ [CGROUP_AUTO] = "auto",
+ [CGROUP_CLOSED] = "closed",
+ [CGROUP_STRICT] = "strict",
+};
+
+DEFINE_STRING_TABLE_LOOKUP(cgroup_device_policy, CGroupDevicePolicy);
+#endif // 0