core: introduce new Delegate=yes/no property controlling creation of cgroup subhierarchies

For priviliged units this resource control property ensures that the
processes have all controllers systemd manages enabled.

For unpriviliged services (those with User= set) this ensures that
access rights to the service cgroup is granted to the user in question,
to create further subgroups. Note that this only applies to the
name=systemd hierarchy though, as access to other controllers is not
safe for unpriviliged processes.

Delegate=yes should be set for container scopes where a systemd instance
inside the container shall manage the hierarchies below its own cgroup
and have access to all controllers.

Delegate=yes should also be set for user@.service, so that systemd
--user can run, controlling its own cgroup tree.

This commit changes machined, systemd-nspawn@.service and user@.service
to set this boolean, in order to ensure that container management will
just work, and the user systemd instance can run fine.

1 files changed, 19 insertions, 3 deletions

diff --git a/src/core/cgroup.c b/src/core/cgroup.c
index e604c3cbc6..0951a09964 100644
--- a/src/core/cgroup.c
+++ b/src/core/cgroup.c
@@ -105,7 +105,8 @@ void cgroup_context_dump(CGroupContext *c, FILE* f, const char *prefix) {
                 "%sBlockIOWeight=%lu\n"
                 "%sStartupBlockIOWeight=%lu\n"
                 "%sMemoryLimit=%" PRIu64 "\n"
-                "%sDevicePolicy=%s\n",
+                "%sDevicePolicy=%s\n"
+                "%sDelegate=%s\n",
                 prefix, yes_no(c->cpu_accounting),
                 prefix, yes_no(c->blockio_accounting),
                 prefix, yes_no(c->memory_accounting),
@@ -115,7 +116,8 @@ void cgroup_context_dump(CGroupContext *c, FILE* f, const char *prefix) {
                 prefix, c->blockio_weight,
                 prefix, c->startup_blockio_weight,
                 prefix, c->memory_limit,
-                prefix, cgroup_device_policy_to_string(c->device_policy));
+                prefix, cgroup_device_policy_to_string(c->device_policy),
+                prefix, yes_no(c->delegate));
 
         LIST_FOREACH(device_allow, a, c->device_allow)
                 fprintf(f,
@@ -461,7 +463,8 @@ CGroupControllerMask cgroup_context_get_mask(CGroupContext *c) {
             c->memory_limit != (uint64_t) -1)
                 mask |= CGROUP_MEMORY;
 
-        if (c->device_allow || c->device_policy != CGROUP_AUTO)
+        if (c->device_allow ||
+            c->device_policy != CGROUP_AUTO)
                 mask |= CGROUP_DEVICE;
 
         return mask;
@@ -474,6 +477,19 @@ CGroupControllerMask unit_get_cgroup_mask(Unit *u) {
         if (!c)
                 return 0;
 
+        /* If delegation is turned on, then turn on all cgroups,
+         * unless the process we fork into it is known to drop
+         * privileges anyway, and shouldn't get access to the
+         * controllers anyway. */
+
+        if (c->delegate) {
+                ExecContext *e;
+
+                e = unit_get_exec_context(u);
+                if (!e || exec_context_maintains_privileges(e))
+                        return _CGROUP_CONTROLLER_MASK_ALL;
+        }
+
         return cgroup_context_get_mask(c);
 }