From: Jon Maxwell <jmaxwell37@gmail.com>
Date: Wed, 15 Jun 2022 01:15:40 +0000 (+1000)
Subject: bpf: Fix request_sock leak in sk lookup helpers
X-Git-Tag: baikal/aarch64/sdk6.1~3680^2~22^2~7
X-Git-Url: https://git.baikalelectronics.ru/sdk/?a=commitdiff_plain;h=b82debc39444354e2efc809c3e6190ca98844bf8;p=kernel.git

bpf: Fix request_sock leak in sk lookup helpers

A customer reported a request_socket leak in a Calico cloud environment. We
found that a BPF program was doing a socket lookup with takes a refcnt on
the socket and that it was finding the request_socket but returning the parent
LISTEN socket via sk_to_full_sk() without decrementing the child request socket
1st, resulting in request_sock slab object leak. This patch retains the
existing behaviour of returning full socks to the caller but it also decrements
the child request_socket if one is present before doing so to prevent the leak.

Thanks to Curtis Taylor for all the help in diagnosing and testing this. And
thanks to Antoine Tenart for the reproducer and patch input.

v2 of this patch contains, refactor as per Daniel Borkmann's suggestions to
validate RCU flags on the listen socket so that it balances with bpf_sk_release()
and update comments as per Martin KaFai Lau's suggestion. One small change to
Daniels suggestion, put "sk = sk2" under "if (sk2 != sk)" to avoid an extra
instruction.

Fixes: 212f702a0b0a ("bpf: Check sk_fullsock() before returning from bpf_sk_lookup()")
Fixes: 44343a49bc65 ("bpf: add skc_lookup_tcp helper")
Co-developed-by: Antoine Tenart <atenart@kernel.org>
Signed-off-by: Antoine Tenart <atenart@kernel.org>
Signed-off-by: Jon Maxwell <jmaxwell37@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Tested-by: Curtis Taylor <cutaylor-pub@yahoo.com>
Cc: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/56d6f898-bde0-bb25-3427-12a330b29fb8@iogearbox.net
Link: https://lore.kernel.org/bpf/20220615011540.813025-1-jmaxwell37@gmail.com
---

diff --git a/net/core/filter.c b/net/core/filter.c
index 5af58eb485875..5d16d66727fc8 100644
--- a/net/core/filter.c
+++ b/net/core/filter.c
@@ -6516,10 +6516,21 @@ __bpf_sk_lookup(struct sk_buff *skb, struct bpf_sock_tuple *tuple, u32 len,
 					   ifindex, proto, netns_id, flags);
 
 	if (sk) {
-		sk = sk_to_full_sk(sk);
-		if (!sk_fullsock(sk)) {
+		struct sock *sk2 = sk_to_full_sk(sk);
+
+		/* sk_to_full_sk() may return (sk)->rsk_listener, so make sure the original sk
+		 * sock refcnt is decremented to prevent a request_sock leak.
+		 */
+		if (!sk_fullsock(sk2))
+			sk2 = NULL;
+		if (sk2 != sk) {
 			sock_gen_put(sk);
-			return NULL;
+			/* Ensure there is no need to bump sk2 refcnt */
+			if (unlikely(sk2 && !sock_flag(sk2, SOCK_RCU_FREE))) {
+				WARN_ONCE(1, "Found non-RCU, unreferenced socket!");
+				return NULL;
+			}
+			sk = sk2;
 		}
 	}
 
@@ -6553,10 +6564,21 @@ bpf_sk_lookup(struct sk_buff *skb, struct bpf_sock_tuple *tuple, u32 len,
 					 flags);
 
 	if (sk) {
-		sk = sk_to_full_sk(sk);
-		if (!sk_fullsock(sk)) {
+		struct sock *sk2 = sk_to_full_sk(sk);
+
+		/* sk_to_full_sk() may return (sk)->rsk_listener, so make sure the original sk
+		 * sock refcnt is decremented to prevent a request_sock leak.
+		 */
+		if (!sk_fullsock(sk2))
+			sk2 = NULL;
+		if (sk2 != sk) {
 			sock_gen_put(sk);
-			return NULL;
+			/* Ensure there is no need to bump sk2 refcnt */
+			if (unlikely(sk2 && !sock_flag(sk2, SOCK_RCU_FREE))) {
+				WARN_ONCE(1, "Found non-RCU, unreferenced socket!");
+				return NULL;
+			}
+			sk = sk2;
 		}
 	}