tcp: fix req->saved_syn race

For the reasons explained in commit ce1050089c ("tcp/dccp: fix
ireq->pktopts race"), we need to make sure we do not access
req->saved_syn unless we own the request sock.

This fixes races for listeners using TCP_SAVE_SYN option.

Fixes: e994b2f0fb ("tcp: do not lock listener to process SYN packets")
Fixes: 079096f103 ("tcp/dccp: install syn_recv requests into ehash table")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Ying Cai <ycai@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This commit is contained in:
Eric Dumazet 2015-11-05 11:07:13 -08:00 committed by David S. Miller
parent 428ad1bc6d
commit 805c4bc057
4 changed files with 21 additions and 11 deletions

View file

@ -397,6 +397,13 @@ static inline void fastopen_queue_tune(struct sock *sk, int backlog)
queue->fastopenq.max_qlen = min_t(unsigned int, backlog, somaxconn);
}
static inline void tcp_move_syn(struct tcp_sock *tp,
struct request_sock *req)
{
tp->saved_syn = req->saved_syn;
req->saved_syn = NULL;
}
static inline void tcp_saved_syn_free(struct tcp_sock *tp)
{
kfree(tp->saved_syn);