tcp: Fix a data-race around sysctl_tcp_thin_linear_timeouts.

[ Upstream commit 7c6f2a86ca590d5187a073d987e9599985fb1c7c ]

While reading sysctl_tcp_thin_linear_timeouts, it can be changed
concurrently.  Thus, we need to add READ_ONCE() to its reader.

Fixes: 36e31b0af587 ("net: TCP thin linear timeouts")
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c
index e905fff09fea285912331eab687011feb1fc30a9..84069db0423a2af5353056227c8cfbb5af4ea2ef 100644 (file)
--- a/net/ipv4/tcp_timer.c
+++ b/net/ipv4/tcp_timer.c
@@ -545,7 +545,7 @@ void tcp_retransmit_timer(struct sock *sk)
         * linear-timeout retransmissions into a black hole
         */
        if (sk->sk_state == TCP_ESTABLISHED &&
-           (tp->thin_lto || net->ipv4.sysctl_tcp_thin_linear_timeouts) &&
+           (tp->thin_lto || READ_ONCE(net->ipv4.sysctl_tcp_thin_linear_timeouts)) &&
            tcp_stream_is_thin(tp) &&
            icsk->icsk_retransmits <= TCP_THIN_LINEAR_RETRIES) {
                icsk->icsk_backoff = 0;