netdev - [RFC 4/5] net/tcp-md5: Don't send RST if key (dis)appeared

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

Message-Id: <20230509221608.2569333-5-dima@arista.com>
Date: Tue,  9 May 2023 23:16:07 +0100
From: Dmitry Safonov <dima@...sta.com>
To: linux-kernel@...r.kernel.org,
	David Ahern <dsahern@...nel.org>,
	Eric Dumazet <edumazet@...gle.com>,
	Paolo Abeni <pabeni@...hat.com>,
	Jakub Kicinski <kuba@...nel.org>,
	"David S. Miller" <davem@...emloft.net>
Cc: Dmitry Safonov <dima@...sta.com>,
	Dmitry Safonov <0x7f454c46@...il.com>,
	Hideaki YOSHIFUJI <yoshfuji@...ux-ipv6.org>,
	Leonard Crestez <cdleonard@...il.com>,
	Salam Noureddine <noureddine@...sta.com>,
	netdev@...r.kernel.org
Subject: [RFC 4/5] net/tcp-md5: Don't send RST if key (dis)appeared

Seems cheap at this place as both key and hash_location were looked up
until now.

Signed-off-by: Dmitry Safonov <dima@...sta.com>
---
 net/ipv4/tcp_ipv4.c | 10 ++++++++++
 net/ipv6/tcp_ipv6.c |  8 ++++++++
 2 files changed, 18 insertions(+)

diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index f5b870943dcb..d94cd5e70d58 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -686,6 +686,16 @@ static bool tcp_v4_md5_sign_reset(struct net *net, const struct sock *sk,
 		l3index = tcp_v4_sdif(skb) ? inet_iif(skb) : 0;
 		addr = (union tcp_md5_addr *)&ip_hdr(skb)->saddr;
 		key = tcp_md5_do_lookup(sk, l3index, addr, AF_INET);
+		/* This segment should have been already verified by
+		 * tcp_inbound_md5_hash(). But that might raced with userspace
+		 * adding or deleting keys. So, follow the logic of
+		 * tcp_inbound_md5_hash() and avoid replying with TCP-MD5 sign
+		 * on non-signed segments and vice-versa.
+		 */
+		if (unlikely(!!key != !!hash_location)) {
+			rcu_read_unlock();
+			return true;
+		}
 	} else if (hash_location) {
 		const union tcp_md5_addr *addr;
 		int sdif = tcp_v4_sdif(skb);
diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index 3756a43367a3..498dfa194b8b 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -993,6 +993,14 @@ static int tcp_v6_md5_lookup_reset_key(struct net *net, const struct sock *sk,
 		 */
 		l3index = tcp_v6_sdif(skb) ? tcp_v6_iif_l3_slave(skb) : 0;
 		*key = tcp_v6_md5_do_lookup(sk, &ipv6h->saddr, l3index);
+		/* This segment should have been already verified by
+		 * tcp_inbound_md5_hash(). But that might raced with userspace
+		 * adding or deleting keys. So, follow the logic of
+		 * tcp_inbound_md5_hash() and avoid replying with TCP-MD5 sign
+		 * on non-signed segments and vice-versa.
+		 */
+		if (unlikely(!!*key != !!hash_location))
+			return -ENOKEY;
 	} else if (hash_location) {
 		int dif = tcp_v6_iif_l3_slave(skb);
 		int sdif = tcp_v6_sdif(skb);
-- 
2.40.0