On Wed, Jul 31, 2019 at 11:46 AM Mickaël Salaün
[off-list ref] wrote:

quoted

quoted

+    for (i = 0; i < htab->n_buckets; i++) {
+            head = select_bucket(htab, i);
+            hlist_nulls_for_each_entry_safe(l, n, head, hash_node) {
+                    landlock_inode_remove_map(*((struct inode **)l->key), map);
+            }
+    }
+    htab_map_free(map);
+}

user space can delete the map.
that will trigger inode_htab_map_free() which will call
landlock_inode_remove_map().
which will simply itereate the list and delete from the list.

landlock_inode_remove_map() removes the reference to the map (being
freed) from the inode (with an RCU lock).

I'm going to ignore everything else for now and focus only on this bit,
since it's fundamental issue to address before this discussion can
go any further.
rcu_lock is not a spin_lock. I'm pretty sure you know this.
But you're arguing that it's somehow protecting from the race
I mentioned above?