tipc_exit_net() is very slow and is abused by syzbot.
tipc_nametbl_stop() is called for each netns being dismantled.
Calling synchronize_net() right before freeing tn->nametbl
is a big hammer.
Replace this with kfree_rcu().
Note that RCU is not properly used here, otherwise
tn->nametbl should be cleared before the synchronize_net()
or kfree_rcu(), or even before the cleanup loop.
We might need to fix this at some point.
Also note tipc uses other synchronize_rcu() calls,
more work is needed to make tipc_exit_net() much faster.
List of remaining calls to synchronize_rcu()
tipc_detach_loopback() (dev_remove_pack())
tipc_bcast_stop()
tipc_sk_rht_destroy()
Signed-off-by: Eric Dumazet <edumazet@google.com>
Link: https://patch.msgid.link/20241204210234.319484-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
}
spin_unlock_bh(&tn->nametbl_lock);
- synchronize_net();
- kfree(nt);
+ /* TODO: clear tn->nametbl, implement proper RCU rules ? */
+ kfree_rcu(nt, rcu);
}
static int __tipc_nl_add_nametable_publ(struct tipc_nl_msg *msg,
/**
* struct name_table - table containing all existing port name publications
+ * @rcu: RCU callback head used for deferred freeing
* @services: name sequence hash lists
* @node_scope: all local publications with node scope
* - used by name_distr during re-init of name table
* @snd_nxt: next sequence number to be used
*/
struct name_table {
+ struct rcu_head rcu;
struct hlist_head services[TIPC_NAMETBL_SIZE];
struct list_head node_scope;
struct list_head cluster_scope;