flow_dissector: lookup netns by skb->sk if skb->dev is NULL

BPF flow dissectors are configured per network namespace.
__skb_flow_dissect looks up the netns through dev_net(skb->dev).

In some dissector paths skb->dev is NULL, such as for Unix sockets.
In these cases fall back to looking up the netns by socket.

Analyzing the codepaths leading to __skb_flow_dissect I did not find
a case where both skb->dev and skb->sk are NULL. Warn and fall back to
standard flow dissector if one is found.

Fixes: d58e468b1112 ("flow_dissector: implements flow dissector BPF hook")
Reported-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Willem de Bruijn <willemb@google.com>
Acked-by: Song Liu <songliubraving@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

diff --git a/net/core/flow_dissector.c b/net/core/flow_dissector.c
index 5c5dd74..738c756 100644 (file)
--- a/net/core/flow_dissector.c
+++ b/net/core/flow_dissector.c
@@ -714,7 +714,7 @@ bool __skb_flow_dissect(const struct sk_buff *skb,
        struct flow_dissector_key_vlan *key_vlan;
        enum flow_dissect_ret fdret;
        enum flow_dissector_key_id dissector_vlan = FLOW_DISSECTOR_KEY_MAX;
-       struct bpf_prog *attached;
+       struct bpf_prog *attached = NULL;
        int num_hdrs = 0;
        u8 ip_proto = 0;
        bool ret;
@@ -755,8 +755,14 @@ bool __skb_flow_dissect(const struct sk_buff *skb,
                                              target_container);
 
        rcu_read_lock();
-       attached = skb ? rcu_dereference(dev_net(skb->dev)->flow_dissector_prog)
-                      : NULL;
+       if (skb) {
+               if (skb->dev)
+                       attached = rcu_dereference(dev_net(skb->dev)->flow_dissector_prog);
+               else if (skb->sk)
+                       attached = rcu_dereference(sock_net(skb->sk)->flow_dissector_prog);
+               else
+                       WARN_ON_ONCE(1);
+       }
        if (attached) {
                /* Note that even though the const qualifier is discarded
                 * throughout the execution of the BPF program, all changes(the