diff mbox series

[net-next,7/9] ipv4: annotate data races arount inet->min_ttl

Message ID 20211021162253.333616-10-eric.dumazet@gmail.com (mailing list archive)
State Superseded
Delegated to: Netdev Maintainers
Headers show
Series tcp: receive path optimizations | expand

Checks

Context Check Description
netdev/apply fail Patch does not apply to net-next
netdev/tree_selection success Clearly marked for net-next

Commit Message

Eric Dumazet Oct. 21, 2021, 4:22 p.m. UTC 
From: Eric Dumazet <edumazet@google.com>

No report yet from KCSAN, yet worth documenting the races.

Signed-off-by: Eric Dumazet <edumazet@google.com>
---
 net/ipv4/ip_sockglue.c | 5 ++++-
 net/ipv4/tcp_ipv4.c    | 7 +++++--
 2 files changed, 9 insertions(+), 3 deletions(-)
diff mbox series

Patch

diff --git a/net/ipv4/ip_sockglue.c b/net/ipv4/ip_sockglue.c
index b297bb28556ec5cf383068f67ee910af38591cc3..d5487c8580674a01df8c7d8ce88f97c9add846b6 100644
--- a/net/ipv4/ip_sockglue.c
+++ b/net/ipv4/ip_sockglue.c
@@ -1352,7 +1352,10 @@  static int do_ip_setsockopt(struct sock *sk, int level, int optname,
 			goto e_inval;
 		if (val < 0 || val > 255)
 			goto e_inval;
-		inet->min_ttl = val;
+		/* tcp_v4_err() and tcp_v4_rcv() might read min_ttl
+		 * while we are changint it.
+		 */
+		WRITE_ONCE(inet->min_ttl, val);
 		break;
 
 	default:
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index e8ca8539b436cf8a8af5b53645a25923003afc41..97b8acf726d0cdcb6b87b6ef45e366591d997a2b 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -508,7 +508,8 @@  int tcp_v4_err(struct sk_buff *skb, u32 info)
 	if (sk->sk_state == TCP_CLOSE)
 		goto out;
 
-	if (unlikely(iph->ttl < inet_sk(sk)->min_ttl)) {
+	/* min_ttl can be changed concurrently from do_ip_setsockopt() */
+	if (unlikely(iph->ttl < READ_ONCE(inet_sk(sk)->min_ttl))) {
 		__NET_INC_STATS(net, LINUX_MIB_TCPMINTTLDROP);
 		goto out;
 	}
@@ -2049,7 +2050,9 @@  int tcp_v4_rcv(struct sk_buff *skb)
 			return 0;
 		}
 	}
-	if (unlikely(iph->ttl < inet_sk(sk)->min_ttl)) {
+
+	/* min_ttl can be changed concurrently from do_ip_setsockopt() */
+	if (unlikely(iph->ttl < READ_ONCE(inet_sk(sk)->min_ttl))) {
 		__NET_INC_STATS(net, LINUX_MIB_TCPMINTTLDROP);
 		goto discard_and_relse;
 	}