xsk: Replace datagram_poll by sock_poll_wait

[ Upstream commit f5da541 ]

datagram_poll will judge the current socket status (EPOLLIN, EPOLLOUT)
based on the traditional socket information (eg: sk_wmem_alloc), but
this does not apply to xsk. So this patch uses sock_poll_wait instead of
datagram_poll, and the mask is calculated by xsk_poll.

Fixes: c497176 ("xsk: add Rx receive functions and poll support")
Signed-off-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Magnus Karlsson <magnus.karlsson@intel.com>
Link: https://lore.kernel.org/bpf/e82f4697438cd63edbf271ebe1918db8261b7c09.1606555939.git.xuanzhuo@linux.alibaba.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

net/xdp/xsk.c

@@ -429,11 +429,13 @@ static int xsk_sendmsg(struct socket *sock, struct msghdr *m, size_t total_len)
 static __poll_t xsk_poll(struct file *file, struct socket *sock,
 			     struct poll_table_struct *wait)
 {
-	__poll_t mask = datagram_poll(file, sock, wait);
+	__poll_t mask = 0;
 	struct sock *sk = sock->sk;
 	struct xdp_sock *xs = xdp_sk(sk);
 	struct xdp_umem *umem;
 
+	sock_poll_wait(file, sock, wait);
+
 	if (unlikely(!xsk_is_bound(xs)))
 		return mask;