diff options
| author | Eric Dumazet <edumazet@google.com> | 2023-03-15 20:57:44 +0000 |
|---|---|---|
| committer | Greg Kroah-Hartman <gregkh@linuxfoundation.org> | 2024-08-03 08:49:04 +0200 |
| commit | a4391e546bcaebc5f9baf0570fe886f2ca5a379c (patch) | |
| tree | 35a6759768e78ea88ac07637038f8d827506bb6a /net/ipv4/tcp_timer.c | |
| parent | 5921e234666dd848da1d08f54801f329d67d8950 (diff) | |
tcp: annotate lockless access to sk->sk_err
[ Upstream commit e13ec3da05d130f0d10da8e1fbe1be26dcdb0e27 ]
tcp_poll() reads sk->sk_err without socket lock held/owned.
We should used READ_ONCE() here, and update writers
to use WRITE_ONCE().
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Stable-dep-of: 853c3bd7b791 ("tcp: fix race in tcp_write_err()")
Signed-off-by: Sasha Levin <sashal@kernel.org>
Diffstat (limited to 'net/ipv4/tcp_timer.c')
| -rw-r--r-- | net/ipv4/tcp_timer.c | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c index fe23a427f6a9d..117fc7031bc10 100644 --- a/net/ipv4/tcp_timer.c +++ b/net/ipv4/tcp_timer.c @@ -67,7 +67,7 @@ u32 tcp_clamp_probe0_to_user_timeout(const struct sock *sk, u32 when) static void tcp_write_err(struct sock *sk) { - sk->sk_err = READ_ONCE(sk->sk_err_soft) ? : ETIMEDOUT; + WRITE_ONCE(sk->sk_err, READ_ONCE(sk->sk_err_soft) ? : ETIMEDOUT); sk_error_report(sk); tcp_write_queue_purge(sk); |