On Mon, Oct 27, 2025 at 05:36:27PM +0100, Thomas Gleixner wrote:

On Fri, Oct 24 2025 at 12:52, Christian Brauner wrote:

quoted

diff --git a/kernel/time/namespace.c b/kernel/time/namespace.c
index ee05cad288da..2e7c110bd13f 100644
--- a/kernel/time/namespace.c
+++ b/kernel/time/namespace.c

@@ -106,6 +106,7 @@ static struct time_namespace *clone_time_ns(struct user_namespace *user_ns,
 	ns->offsets = old_ns->offsets;
 	ns->frozen_offsets = false;
 	ns_tree_add(ns);
+	ns_ref_active_get_owner(ns);

It seems all places where ns_ref_active_get_owner() is added it is
preceeded by a variant of ns_tree_add(). So why don't you stilck that
refcount thing into ns_tree_add()? I'm probably missing something here.

Yes, as usual a very astute observation. I've taken your advice and
folded the parent count into __ns_tree_add_raw(). Thanks!