bpf: sockmap, sk_wait_event needed to handle blocking cases

In the recvmsg handler we need to add a wait event to support the
blocking use cases. Without this we return zero and may confuse
user applications. In the wait event any data received on the
sk either via sk_receive_queue or the psock ingress list will
wake up the sock.

Fixes: fa24669 ("bpf: sockmap, BPF_F_INGRESS flag for BPF_SK_SKB_STREAM_VERDICT")
Signed-off-by: John Fastabend <[email protected]>
Signed-off-by: Daniel Borkmann <[email protected]>

kernel/bpf/sockmap.c

@@ -43,6 +43,7 @@
 #include <net/tcp.h>
 #include <linux/ptr_ring.h>
 #include <net/inet_common.h>
+#include <linux/sched/signal.h>
 
 #define SOCK_CREATE_FLAG_MASK \
 	(BPF_F_NUMA_NODE | BPF_F_RDONLY | BPF_F_WRONLY)
@@ -732,6 +733,26 @@ static int bpf_exec_tx_verdict(struct smap_psock *psock,
 	return err;
 }
 
+static int bpf_wait_data(struct sock *sk,
+			 struct smap_psock *psk, int flags,
+			 long timeo, int *err)
+{
+	int rc;
+
+	DEFINE_WAIT_FUNC(wait, woken_wake_function);
+
+	add_wait_queue(sk_sleep(sk), &wait);
+	sk_set_bit(SOCKWQ_ASYNC_WAITDATA, sk);
+	rc = sk_wait_event(sk, &timeo,
+			   !list_empty(&psk->ingress) ||
+			   !skb_queue_empty(&sk->sk_receive_queue),
+			   &wait);
+	sk_clear_bit(SOCKWQ_ASYNC_WAITDATA, sk);
+	remove_wait_queue(sk_sleep(sk), &wait);
+
+	return rc;
+}
+
 static int bpf_tcp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len,
 			   int nonblock, int flags, int *addr_len)
 {
@@ -755,6 +776,7 @@ static int bpf_tcp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len,
 		return tcp_recvmsg(sk, msg, len, nonblock, flags, addr_len);
 
 	lock_sock(sk);
+bytes_ready:
 	while (copied != len) {
 		struct scatterlist *sg;
 		struct sk_msg_buff *md;
@@ -809,6 +831,28 @@ static int bpf_tcp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len,
 		}
 	}
 
+	if (!copied) {
+		long timeo;
+		int data;
+		int err = 0;
+
+		timeo = sock_rcvtimeo(sk, nonblock);
+		data = bpf_wait_data(sk, psock, flags, timeo, &err);
+
+		if (data) {
+			if (!skb_queue_empty(&sk->sk_receive_queue)) {
+				release_sock(sk);
+				smap_release_sock(psock, sk);
+				copied = tcp_recvmsg(sk, msg, len, nonblock, flags, addr_len);
+				return copied;
+			}
+			goto bytes_ready;
+		}
+
+		if (err)
+			copied = err;
+	}
+
 	release_sock(sk);
 	smap_release_sock(psock, sk);
 	return copied;