From: Eric Dumazet <edumazet@xxxxxxxxxx> [ Upstream commit 8c70521238b7863c2af607e20bcba20f974c969b ] challenge_timestamp can be read an written by concurrent threads. This was expected, but we need to annotate the race to avoid potential issues. Following patch moves challenge_timestamp and challenge_count to per-netns storage to provide better isolation. Fixes: 354e4aa391ed ("tcp: RFC 5961 5.2 Blind Data Injection Attack Mitigation") Reported-by: syzbot <syzkaller@xxxxxxxxxxxxxxxx> Signed-off-by: Eric Dumazet <edumazet@xxxxxxxxxx> Acked-by: Neal Cardwell <ncardwell@xxxxxxxxxx> Signed-off-by: Jakub Kicinski <kuba@xxxxxxxxxx> Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx> --- net/ipv4/tcp_input.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index e1d065ea5a158..0be9d5d3c032f 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -3468,11 +3468,11 @@ static void tcp_send_challenge_ack(struct sock *sk, const struct sk_buff *skb) /* Then check host-wide RFC 5961 rate limit. */ now = jiffies / HZ; - if (now != challenge_timestamp) { + if (now != READ_ONCE(challenge_timestamp)) { u32 ack_limit = READ_ONCE(net->ipv4.sysctl_tcp_challenge_ack_limit); u32 half = (ack_limit + 1) >> 1; - challenge_timestamp = now; + WRITE_ONCE(challenge_timestamp, now); WRITE_ONCE(challenge_count, half + prandom_u32_max(ack_limit)); } count = READ_ONCE(challenge_count); -- 2.35.1