bpf: Fix use of sk->sk_reuseport from sk_assign

In testing, we found that for request sockets the sk->sk_reuseport field
may yet be uninitialized, which caused bpf_sk_assign() to randomly
succeed or return -ESOCKTNOSUPPORT when handling the forward ACK in a
three-way handshake.

Fix it by only applying the reuseport check for full sockets.

Fixes: cf7fbe660f ("bpf: Add socket assign support")
Signed-off-by: Joe Stringer <[email protected]>
Signed-off-by: Daniel Borkmann <[email protected]>
Acked-by: Martin KaFai Lau <[email protected]>
Link: https://lore.kernel.org/bpf/[email protected]
This commit is contained in:
Joe Stringer authored and Daniel Borkmann committed 2020-04-09 01:02:14 +02:00
1 parent eb203f4b89
commit 8e368dc72e
1 file changed
+1 -1
+1 -1
View File
@@ -5925,7 +5925,7 @@ BPF_CALL_3(bpf_sk_assign, struct sk_buff *, skb, struct sock *, sk, u64, flags)
return -EOPNOTSUPP;
if (unlikely(dev_net(skb->dev) != sock_net(sk)))
return -ENETUNREACH;
if (unlikely(sk->sk_reuseport))
if (unlikely(sk_fullsock(sk) && sk->sk_reuseport))
return -ESOCKTNOSUPPORT;
if (sk_is_refcounted(sk) &&
unlikely(!refcount_inc_not_zero(&sk->sk_refcnt)))