ipc: Store mqueue sysctls in the ipc namespace
Right now, the mqueue sysctls take ipc namespaces into account in a
rather hacky way. This works in most cases, but does not respect the
user namespace.
Within the user namespace, the user cannot change the /proc/sys/fs/mqueue/*
parametres. This poses a problem in the rootless containers.
To solve this I changed the implementation of the mqueue sysctls just
like some other sysctls.
So far, the changes do not provide additional access to files. This will
be done in a future patch.
v3:
* Don't implemenet set_permissions to keep the current behavior.
v2:
* Fixed compilation problem if CONFIG_POSIX_MQUEUE_SYSCTL is not
specified.
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Alexey Gladkov <legion@kernel.org>
Link: https://lkml.kernel.org/r/b0ccbb2489119f1f20c737cf1930c3a9c4e4243a.1644862280.git.legion@kernel.org
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
diff --git a/ipc/namespace.c b/ipc/namespace.c
index ae83f0f..f760243 100644
--- a/ipc/namespace.c
+++ b/ipc/namespace.c
@@ -59,6 +59,10 @@ static struct ipc_namespace *create_ipc_ns(struct user_namespace *user_ns,
if (err)
goto fail_put;
+ err = -ENOMEM;
+ if (!setup_mq_sysctls(ns))
+ goto fail_put;
+
sem_init_ns(ns);
msg_init_ns(ns);
shm_init_ns(ns);
@@ -125,6 +129,8 @@ static void free_ipc_ns(struct ipc_namespace *ns)
msg_exit_ns(ns);
shm_exit_ns(ns);
+ retire_mq_sysctls(ns);
+
dec_ipc_namespaces(ns->ucounts);
put_user_ns(ns->user_ns);
ns_free_inum(&ns->ns);
