linux-kernel - [PATCH] net: tcp: ratelimit warnings in tcp

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Date:   Mon,  9 Nov 2020 08:06:20 -0500
From:   menglong8.dong@...il.com
To:     edumazet@...gle.com
Cc:     davem@...emloft.net, kuznet@....inr.ac.ru, yoshfuji@...ux-ipv6.org,
        kuba@...nel.org, netdev@...r.kernel.org,
        linux-kernel@...r.kernel.org,
        Menglong Dong <dong.menglong@....com.cn>
Subject: [PATCH] net: tcp: ratelimit warnings in tcp_recvmsg

From: Menglong Dong <dong.menglong@....com.cn>

'before(*seq, TCP_SKB_CB(skb)->seq) == true' means that one or more
skbs are lost somehow. Once this happen, it seems that it will
never recover automatically. As a result, a warning will be printed
and a '-EAGAIN' will be returned in non-block mode.

As a general suituation, users call 'poll' on a socket and then receive
skbs with 'recv' in non-block mode. This mode will make every
arriving skb of the socket trigger a warning. Plenty of skbs will cause
high rate of kernel log.

Besides, WARN is for indicating kernel bugs only and should not be
user-triggable. Replace it with 'net_warn_ratelimited' here.

Signed-off-by: Menglong Dong <dong.menglong@....com.cn>
---
 net/ipv4/tcp.c | 17 ++++++++++-------
 1 file changed, 10 insertions(+), 7 deletions(-)

diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index b2bc3d7fe9e8..5e38dfd03036 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -2093,11 +2093,12 @@ int tcp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len, int nonblock,
 			/* Now that we have two receive queues this
 			 * shouldn't happen.
 			 */
-			if (WARN(before(*seq, TCP_SKB_CB(skb)->seq),
-				 "TCP recvmsg seq # bug: copied %X, seq %X, rcvnxt %X, fl %X\n",
-				 *seq, TCP_SKB_CB(skb)->seq, tp->rcv_nxt,
-				 flags))
+			if (unlikely(before(*seq, TCP_SKB_CB(skb)->seq))) {
+				net_warn_ratelimited("TCP recvmsg seq # bug: copied %X, seq %X, rcvnxt %X, fl %X\n",
+						     *seq, TCP_SKB_CB(skb)->seq, tp->rcv_nxt,
+						     flags);
 				break;
+			}
 
 			offset = *seq - TCP_SKB_CB(skb)->seq;
 			if (unlikely(TCP_SKB_CB(skb)->tcp_flags & TCPHDR_SYN)) {
@@ -2108,9 +2109,11 @@ int tcp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len, int nonblock,
 				goto found_ok_skb;
 			if (TCP_SKB_CB(skb)->tcp_flags & TCPHDR_FIN)
 				goto found_fin_ok;
-			WARN(!(flags & MSG_PEEK),
-			     "TCP recvmsg seq # bug 2: copied %X, seq %X, rcvnxt %X, fl %X\n",
-			     *seq, TCP_SKB_CB(skb)->seq, tp->rcv_nxt, flags);
+
+			if (!(flags & MSG_PEEK))
+				net_warn_ratelimited("TCP recvmsg seq # bug 2: copied %X, seq %X, rcvnxt %X, fl %X\n",
+						     *seq, TCP_SKB_CB(skb)->seq, tp->rcv_nxt,
+						     flags);
 		}
 
 		/* Well, if we have backlog, try to process it now yet. */
-- 
2.25.1