[PATCH 4.19 21/51] netfilter: nf_queue: dont assume sk is full socket

From: Greg Kroah-Hartman
Date: Mon Mar 07 2022 - 04:29:09 EST


From: Florian Westphal <fw@xxxxxxxxx>

commit 747670fd9a2d1b7774030dba65ca022ba442ce71 upstream.

There is no guarantee that state->sk refers to a full socket.

If refcount transitions to 0, sock_put calls sk_free which then ends up
with garbage fields.

I'd like to thank Oleksandr Natalenko and Jiri Benc for considerable
debug work and pointing out state->sk oddities.

Fixes: ca6fb0651883 ("tcp: attach SYNACK messages to request sockets instead of listener")
Tested-by: Oleksandr Natalenko <oleksandr@xxxxxxxxxx>
Signed-off-by: Florian Westphal <fw@xxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
net/netfilter/nf_queue.c | 11 ++++++++++-
1 file changed, 10 insertions(+), 1 deletion(-)

--- a/net/netfilter/nf_queue.c
+++ b/net/netfilter/nf_queue.c
@@ -46,6 +46,15 @@ void nf_unregister_queue_handler(struct
}
EXPORT_SYMBOL(nf_unregister_queue_handler);

+static void nf_queue_sock_put(struct sock *sk)
+{
+#ifdef CONFIG_INET
+ sock_gen_put(sk);
+#else
+ sock_put(sk);
+#endif
+}
+
void nf_queue_entry_release_refs(struct nf_queue_entry *entry)
{
struct nf_hook_state *state = &entry->state;
@@ -56,7 +65,7 @@ void nf_queue_entry_release_refs(struct
if (state->out)
dev_put(state->out);
if (state->sk)
- sock_put(state->sk);
+ nf_queue_sock_put(state->sk);
#if IS_ENABLED(CONFIG_BRIDGE_NETFILTER)
if (entry->skb->nf_bridge) {
struct net_device *physdev;