[PATCH net v2 2/2] tipc: make conn_idr teardown safe

[Yuqi Xu <xuyuqiabc@gmail.com>]

The teardown walk iterated conn_idr by incrementing a numeric ID while
holding idr_lock, so it could scan a large range of unused IDs without
letting a connection's final reference release make progress. An entry
whose last reference had already been dropped could also be resurrected
by the unconditional conn_get() while its release callback was blocked
on the same lock.

Walk conn_idr with idr_get_next(), release the lock and reschedule when
no entry can be taken, and use kref_get_unless_zero() so a connection
that is already being released cannot be revived.

Fixes: 35e22e49a5d6 ("tipc: fix cleanup at module unload")
Fixes: 667eeab4999e ("tipc: Fix use-after-free in tipc_conn_close().")
Cc: stable@vger.kernel.org
Reported-by: Vega <vega@nebusec.ai>
Assisted-by: LLM
Signed-off-by: Yuqi Xu <xuyuqiabc@gmail.com>
Reviewed-by: Ren Wei <weir@nebusec.ai>
---
Changes in v2:
 - Rebased onto the current net/main tip; patch content unchanged.

 net/tipc/topsrv.c | 17 +++++++++++------
 1 file changed, 11 insertions(+), 6 deletions(-)

diff --git a/net/tipc/topsrv.c b/net/tipc/topsrv.c
index 908622a3d0fc..9333e36a74de 100644
--- a/net/tipc/topsrv.c
+++ b/net/tipc/topsrv.c
@@ -710,15 +710,20 @@ static void tipc_topsrv_stop(struct net *net)
 	cancel_work_sync(&srv->awork);
 
 	spin_lock_bh(&srv->idr_lock);
-	for (id = 0; srv->idr_in_use; id++) {
-		con = idr_find(&srv->conn_idr, id);
-		if (con) {
-			conn_get(con);
+	for (id = 0; srv->idr_in_use;) {
+		con = idr_get_next(&srv->conn_idr, &id);
+		if (!con || !kref_get_unless_zero(&con->kref)) {
 			spin_unlock_bh(&srv->idr_lock);
-			tipc_conn_close(con);
-			conn_put(con);
+			cond_resched();
 			spin_lock_bh(&srv->idr_lock);
+			id = 0;
+			continue;
 		}
+		id++;
+		spin_unlock_bh(&srv->idr_lock);
+		tipc_conn_close(con);
+		conn_put(con);
+		spin_lock_bh(&srv->idr_lock);
 	}
 	__module_get(lsock->ops->owner);
 	__module_get(lsock->sk->sk_prot_creator->owner);
-- 
2.55.0