Re: BUG: IPv4: Attempt to release TCP socket in state 1
From: Eric Dumazet <hidden>
Date: 2013-03-16 17:36:11
Also in:
lkml
Subsystem:
networking [general], networking [sockets], networking [tcp], the rest · Maintainers:
"David S. Miller", Eric Dumazet, Jakub Kicinski, Paolo Abeni, Kuniyuki Iwashima, Willem de Bruijn, Neal Cardwell, Linus Torvalds
On Fri, 2013-03-15 at 00:19 +0100, Eric Dumazet wrote:
Thanks thats really useful, we might miss to increment socket refcount in a timer setup.
Hmm, please add following debugging patch as well
diff --git a/include/net/sock.h b/include/net/sock.h
index 14f6e9d..fe7c8a6 100644
--- a/include/net/sock.h
+++ b/include/net/sock.h@@ -530,7 +530,9 @@ static inline void sock_hold(struct sock *sk) */ static inline void __sock_put(struct sock *sk) { - atomic_dec(&sk->sk_refcnt); + int newref = atomic_dec_return(&sk->sk_refcnt); + + BUG_ON(newref <= 0); } static inline bool sk_del_node_init(struct sock *sk)
diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c
index 786d97a..a445e15 100644
--- a/net/ipv4/inet_connection_sock.c
+++ b/net/ipv4/inet_connection_sock.c@@ -739,7 +739,7 @@ void inet_csk_prepare_forced_close(struct sock *sk) { /* sk_clone_lock locked the socket and set refcnt to 2 */ bh_unlock_sock(sk); - sock_put(sk); + __sock_put(sk); /* The below has to be done to allow calling inet_csk_destroy_sock */ sock_set_flag(sk, SOCK_DEAD);
@@ -835,13 +835,13 @@ void inet_csk_listen_stop(struct sock *sk) * tcp_v4_destroy_sock(). */ tcp_sk(child)->fastopen_rsk = NULL; - sock_put(sk); + __sock_put(sk); } inet_csk_destroy_sock(child); bh_unlock_sock(child); local_bh_enable(); - sock_put(child); + __sock_put(child); sk_acceptq_removed(sk); __reqsk_free(req);