The opinfo pointer obtained via rcu_dereference(fp->f_opinfo) is
dereferenced after rcu_read_unlock(), creating a use-after-free
window. A concurrent opinfo_put() can free the opinfo between the
unlock and the subsequent access to opinfo->is_lease,
opinfo->o_lease->state, and opinfo->level.
Fix this by deferring rcu_read_unlock() until after all opinfo
field accesses are complete. The values needed (const_names, count,
level) are copied into local variables under the RCU read lock,
and the potentially-sleeping seq_printf calls happen after the
lock is released.
Found by AI-assisted code review (Claude Opus 4.6, Anthropic)
in collaboration with Ali Khaledi.
Cc: stable@vger.kernel.org
Fixes: b38f99c1217a ("ksmbd: add procfs interface for runtime monitoring and statistics")
Signed-off-by: Ali Khaledi <ali.khaledi1989@gmail.com>
Acked-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
rcu_read_lock();
opinfo = rcu_dereference(fp->f_opinfo);
- rcu_read_unlock();
-
- if (!opinfo) {
- seq_printf(m, " %-15s", " ");
- } else {
+ if (opinfo) {
const struct ksmbd_const_name *const_names;
int count;
unsigned int level;
count = ARRAY_SIZE(ksmbd_oplock_const_names);
level = opinfo->level;
}
+ rcu_read_unlock();
ksmbd_proc_show_const_name(m, " %-15s",
const_names, count, level);
+ } else {
+ rcu_read_unlock();
+ seq_printf(m, " %-15s", " ");
}
seq_printf(m, " %#010x %#010x %s\n",