tcp: ensure proper barriers in lockless contexts
Some functions access TCP sockets without holding a lock and
might output non consistent data, depending on compiler and or
architecture.
tcp_diag_get_info(), tcp_get_info(), tcp_poll(), get_tcp4_sock() ...
Introduce sk_state_load() and sk_state_store() to fix the issues,
and more clearly document where this lack of locking is happening.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index 0cfa7c0..c172877 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -451,11 +451,14 @@
unsigned int mask;
struct sock *sk = sock->sk;
const struct tcp_sock *tp = tcp_sk(sk);
+ int state;
sock_rps_record_flow(sk);
sock_poll_wait(file, sk_sleep(sk), wait);
- if (sk->sk_state == TCP_LISTEN)
+
+ state = sk_state_load(sk);
+ if (state == TCP_LISTEN)
return inet_csk_listen_poll(sk);
/* Socket is not locked. We are protected from async events
@@ -492,14 +495,14 @@
* NOTE. Check for TCP_CLOSE is added. The goal is to prevent
* blocking on fresh not-connected or disconnected socket. --ANK
*/
- if (sk->sk_shutdown == SHUTDOWN_MASK || sk->sk_state == TCP_CLOSE)
+ if (sk->sk_shutdown == SHUTDOWN_MASK || state == TCP_CLOSE)
mask |= POLLHUP;
if (sk->sk_shutdown & RCV_SHUTDOWN)
mask |= POLLIN | POLLRDNORM | POLLRDHUP;
/* Connected or passive Fast Open socket? */
- if (sk->sk_state != TCP_SYN_SENT &&
- (sk->sk_state != TCP_SYN_RECV || tp->fastopen_rsk)) {
+ if (state != TCP_SYN_SENT &&
+ (state != TCP_SYN_RECV || tp->fastopen_rsk)) {
int target = sock_rcvlowat(sk, 0, INT_MAX);
if (tp->urg_seq == tp->copied_seq &&
@@ -507,9 +510,6 @@
tp->urg_data)
target++;
- /* Potential race condition. If read of tp below will
- * escape above sk->sk_state, we can be illegally awaken
- * in SYN_* states. */
if (tp->rcv_nxt - tp->copied_seq >= target)
mask |= POLLIN | POLLRDNORM;
@@ -1934,7 +1934,7 @@
/* Change state AFTER socket is unhashed to avoid closed
* socket sitting in hash tables.
*/
- sk->sk_state = state;
+ sk_state_store(sk, state);
#ifdef STATE_TRACE
SOCK_DEBUG(sk, "TCP sk=%p, State %s -> %s\n", sk, statename[oldstate], statename[state]);
@@ -2644,7 +2644,8 @@
if (sk->sk_type != SOCK_STREAM)
return;
- info->tcpi_state = sk->sk_state;
+ info->tcpi_state = sk_state_load(sk);
+
info->tcpi_ca_state = icsk->icsk_ca_state;
info->tcpi_retransmits = icsk->icsk_retransmits;
info->tcpi_probes = icsk->icsk_probes_out;
@@ -2672,7 +2673,7 @@
info->tcpi_snd_mss = tp->mss_cache;
info->tcpi_rcv_mss = icsk->icsk_ack.rcv_mss;
- if (sk->sk_state == TCP_LISTEN) {
+ if (info->tcpi_state == TCP_LISTEN) {
info->tcpi_unacked = sk->sk_ack_backlog;
info->tcpi_sacked = sk->sk_max_ack_backlog;
} else {