Netdev List
 help / color / mirror / Atom feed
From: David Lee <david.lee@trailofbits.com>
To: edumazet@google.com, ncardwell@google.com, davem@davemloft.net,
	kuba@kernel.org, pabeni@redhat.com
Cc: David Lee <david.lee@trailofbits.com>,
	Kyle Zeng <kylebot@openai.com>,
	Dominik 'Disconnect3d' Czarnota
	<dominik.czarnota@trailofbits.com>,
	kuniyu@google.com, horms@kernel.org, netdev@vger.kernel.org,
	linux-kernel@vger.kernel.org
Subject: [PATCH] net: inet: prevent lhash2 lookups from escaping into ehash
Date: Fri, 31 Jul 2026 14:05:12 +0000	[thread overview]
Message-ID: <20260731140512.566464-1-david.lee@trailofbits.com> (raw)

The lhash2 and ehash tables share sk_nulls_node. When a listening
socket is unhashed and rehashed as a connected socket while an RCU
listener lookup is walking it, the reader can follow the node's new
next pointer into ehash. inet_lhash2_lookup() and
inet6_lhash2_lookup() do not validate the terminal nulls marker, so
they can return an ehash entry under the listener lookup's
unreferenced ownership contract.

In particular, returning a TIME_WAIT socket makes the TCP receive
path consume a reference that the lookup never acquired. Repeated
races can free the object while it remains linked in the hash tables.

Skip non-listening sockets before scoring them. Also restart the walk
when its terminal nulls marker does not match the expected lhash2 slot,
as the established lookup already does.

Fixes: cae3873c5b3a ("net: inet: Retire port only listening_hash")
Bug found and triaged by OpenAI Security Research and 
validated by Trail of Bits.

Assisted-by: Codex:gpt-5.6-sol gpt-5.5-cyber
Signed-off-by: Kyle Zeng <kylebot@openai.com>
---
Trail of Bits has a reproducer for this bug that triggers a
KASAN use-after-free and can share if needed.

 net/ipv4/inet_hashtables.c  | 20 ++++++++++++++++----
 net/ipv6/inet6_hashtables.c | 20 ++++++++++++++++----
 2 files changed, 32 insertions(+), 8 deletions(-)

diff --git a/net/ipv4/inet_hashtables.c b/net/ipv4/inet_hashtables.c
index ba0faa9ae..563c8af44 100644
--- a/net/ipv4/inet_hashtables.c
+++ b/net/ipv4/inet_hashtables.c
@@ -415,16 +415,23 @@ EXPORT_SYMBOL_GPL(inet_lookup_reuseport);
 /* called with rcu_read_lock() : No refcount taken on the socket */
 static struct sock *inet_lhash2_lookup(const struct net *net,
 				struct inet_listen_hashbucket *ilb2,
+				unsigned int slot,
 				struct sk_buff *skb, int doff,
 				const __be32 saddr, __be16 sport,
 				const __be32 daddr, const unsigned short hnum,
 				const int dif, const int sdif)
 {
-	struct sock *sk, *result = NULL;
+	struct sock *sk, *result;
 	struct hlist_nulls_node *node;
-	int score, hiscore = 0;
+	int score, hiscore;
 
+begin:
+	result = NULL;
+	hiscore = 0;
 	sk_nulls_for_each_rcu(sk, node, &ilb2->nulls_head) {
+		if (READ_ONCE(sk->sk_state) != TCP_LISTEN)
+			continue;
+
 		score = compute_score(sk, net, hnum, daddr, dif, sdif);
 		if (score > hiscore) {
 			result = inet_lookup_reuseport(net, sk, skb, doff,
@@ -437,6 +444,9 @@ static struct sock *inet_lhash2_lookup(const struct net *net,
 		}
 	}
 
+	if (get_nulls_value(node) != slot + LISTENING_NULLS_BASE)
+		goto begin;
+
 	return result;
 }
 
@@ -486,7 +496,8 @@ struct sock *__inet_lookup_listener(const struct net *net,
 	hash2 = ipv4_portaddr_hash(net, daddr, hnum);
 	ilb2 = inet_lhash2_bucket(hashinfo, hash2);
 
-	result = inet_lhash2_lookup(net, ilb2, skb, doff,
+	result = inet_lhash2_lookup(net, ilb2,
+				    hash2 & hashinfo->lhash2_mask, skb, doff,
 				    saddr, sport, daddr, hnum,
 				    dif, sdif);
 	if (result)
@@ -496,7 +507,8 @@ struct sock *__inet_lookup_listener(const struct net *net,
 	hash2 = ipv4_portaddr_hash(net, htonl(INADDR_ANY), hnum);
 	ilb2 = inet_lhash2_bucket(hashinfo, hash2);
 
-	result = inet_lhash2_lookup(net, ilb2, skb, doff,
+	result = inet_lhash2_lookup(net, ilb2,
+				    hash2 & hashinfo->lhash2_mask, skb, doff,
 				    saddr, sport, htonl(INADDR_ANY), hnum,
 				    dif, sdif);
 done:
diff --git a/net/ipv6/inet6_hashtables.c b/net/ipv6/inet6_hashtables.c
index b111b51d6..a9dfd47dd 100644
--- a/net/ipv6/inet6_hashtables.c
+++ b/net/ipv6/inet6_hashtables.c
@@ -184,16 +184,23 @@ EXPORT_SYMBOL_GPL(inet6_lookup_reuseport);
 /* called with rcu_read_lock() */
 static struct sock *inet6_lhash2_lookup(const struct net *net,
 		struct inet_listen_hashbucket *ilb2,
+		unsigned int slot,
 		struct sk_buff *skb, int doff,
 		const struct in6_addr *saddr,
 		const __be16 sport, const struct in6_addr *daddr,
 		const unsigned short hnum, const int dif, const int sdif)
 {
-	struct sock *sk, *result = NULL;
+	struct sock *sk, *result;
 	struct hlist_nulls_node *node;
-	int score, hiscore = 0;
+	int score, hiscore;
 
+begin:
+	result = NULL;
+	hiscore = 0;
 	sk_nulls_for_each_rcu(sk, node, &ilb2->nulls_head) {
+		if (READ_ONCE(sk->sk_state) != TCP_LISTEN)
+			continue;
+
 		score = compute_score(sk, net, hnum, daddr, dif, sdif);
 		if (score > hiscore) {
 			result = inet6_lookup_reuseport(net, sk, skb, doff,
@@ -206,6 +213,9 @@ static struct sock *inet6_lhash2_lookup(const struct net *net,
 		}
 	}
 
+	if (get_nulls_value(node) != slot + LISTENING_NULLS_BASE)
+		goto begin;
+
 	return result;
 }
 
@@ -260,7 +270,8 @@ struct sock *inet6_lookup_listener(const struct net *net,
 	hash2 = ipv6_portaddr_hash(net, daddr, hnum);
 	ilb2 = inet_lhash2_bucket(hashinfo, hash2);
 
-	result = inet6_lhash2_lookup(net, ilb2, skb, doff,
+	result = inet6_lhash2_lookup(net, ilb2,
+				     hash2 & hashinfo->lhash2_mask, skb, doff,
 				     saddr, sport, daddr, hnum,
 				     dif, sdif);
 	if (result)
@@ -270,7 +281,8 @@ struct sock *inet6_lookup_listener(const struct net *net,
 	hash2 = ipv6_portaddr_hash(net, &in6addr_any, hnum);
 	ilb2 = inet_lhash2_bucket(hashinfo, hash2);
 
-	result = inet6_lhash2_lookup(net, ilb2, skb, doff,
+	result = inet6_lhash2_lookup(net, ilb2,
+				     hash2 & hashinfo->lhash2_mask, skb, doff,
 				     saddr, sport, &in6addr_any, hnum,
 				     dif, sdif);
 done:

                 reply	other threads:[~2026-07-31 14:05 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20260731140512.566464-1-david.lee@trailofbits.com \
    --to=david.lee@trailofbits.com \
    --cc=davem@davemloft.net \
    --cc=dominik.czarnota@trailofbits.com \
    --cc=edumazet@google.com \
    --cc=horms@kernel.org \
    --cc=kuba@kernel.org \
    --cc=kuniyu@google.com \
    --cc=kylebot@openai.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=ncardwell@google.com \
    --cc=netdev@vger.kernel.org \
    --cc=pabeni@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox