[systemd-devel] [PATCH] Fix RemainAfterExit services keeping a hold on console

[Olivier Brunel]

When a service exits succesfully and has RemainAfterExit set, its hold
on the console (in m->n_on_console) wasn't released since the unit state
didn't change.
---
 TODO               |  2 --
 src/core/service.c | 16 ++++++++++++++++
 src/core/unit.c    |  3 +++
 3 files changed, 19 insertions(+), 2 deletions(-)

diff --git a/TODO b/TODO
index 57e1122..efc7e2a 100644
--- a/TODO
+++ b/TODO
@@ -21,8 +21,6 @@ Bugfixes:
 
   Cannot add dependency job for unit display-manager.service, ignoring: Unit display-manager.service failed to load: No such file or directory. See system logs and 'systemctl status display-manager.service' for details.
 
-* Substract units in SERVICE_EXITED substate from n_on_console.
-
 Fedora 20:
 
 * external: ps should gain colums for slice and machine
diff --git a/src/core/service.c b/src/core/service.c
index 3da32a1..c0ee114 100644
--- a/src/core/service.c
+++ b/src/core/service.c
@@ -1519,6 +1519,22 @@ static void service_set_state(Service *s, ServiceState state) {
         if (state == SERVICE_EXITED && UNIT(s)->manager->n_reloading <= 0)
                 unit_destroy_cgroup(UNIT(s));
 
+        /* For remain_after_exit services, let's see if we can "release" the
+         * hold on the console, since unit_notify() only does that in case of
+         * change of state */
+        if (state == SERVICE_EXITED && s->remain_after_exit &&
+            UNIT(s)->manager->n_on_console > 0) {
+                ExecContext *ec = unit_get_exec_context(UNIT(s));
+                if (ec && exec_context_may_touch_console(ec)) {
+                        Manager *m = UNIT(s)->manager;
+
+                        m->n_on_console --;
+                        if (m->n_on_console == 0)
+                                /* unset no_console_output flag, since the console is free */
+                                m->no_console_output = false;
+                }
+        }
+
         if (old_state != state)
                 log_debug_unit(UNIT(s)->id,
                                "%s changed %s -> %s", UNIT(s)->id,
diff --git a/src/core/unit.c b/src/core/unit.c
index 15e0a82..d41bc90 100644
--- a/src/core/unit.c
+++ b/src/core/unit.c
@@ -1484,6 +1484,9 @@ void unit_notify(Unit *u, UnitActiveState os, UnitActiveState ns, bool reload_su
         if (UNIT_IS_INACTIVE_OR_FAILED(ns))
                 unit_destroy_cgroup(u);
 
+        /* Note that this doesn't apply to RemainAfterExit services exiting
+         * sucessfully, since there's no change of state in that case. Which is
+         * why it is handled in service_set_state() */
         if (UNIT_IS_INACTIVE_OR_FAILED(os) != UNIT_IS_INACTIVE_OR_FAILED(ns)) {
                 ExecContext *ec = unit_get_exec_context(u);
                 if (ec && exec_context_may_touch_console(ec)) {
-- 
1.8.4.2