vsock: support sockmap

This patch adds sockmap support for vsock sockets. It is intended to be
usable by all transports, but only the virtio and loopback transports
are implemented.

SOCK_STREAM, SOCK_DGRAM, and SOCK_SEQPACKET are all supported.

Signed-off-by: Bobby Eshleman <bobby.eshleman@bytedance.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Author: Bobby Eshleman

drivers/vhost/vsock.c

@@ -439,6 +439,7 @@ static struct virtio_transport vhost_transport = {
 		.notify_send_post_enqueue = virtio_transport_notify_send_post_enqueue,
 		.notify_buffer_size       = virtio_transport_notify_buffer_size,
 
+		.read_skb = virtio_transport_read_skb,
 	},
 
 	.send_pkt = vhost_transport_send_pkt,

include/linux/virtio_vsock.h

@@ -245,4 +245,5 @@ u32 virtio_transport_get_credit(struct virtio_vsock_sock *vvs, u32 wanted);
 void virtio_transport_put_credit(struct virtio_vsock_sock *vvs, u32 credit);
 void virtio_transport_deliver_tap_pkt(struct sk_buff *skb);
 int virtio_transport_purge_skbs(void *vsk, struct sk_buff_head *list);
+int virtio_transport_read_skb(struct vsock_sock *vsk, skb_read_actor_t read_actor);
 #endif /* _LINUX_VIRTIO_VSOCK_H */

include/net/af_vsock.h

@@ -75,6 +75,7 @@ struct vsock_sock {
 	void *trans;
 };
 
+s64 vsock_connectible_has_data(struct vsock_sock *vsk);
 s64 vsock_stream_has_data(struct vsock_sock *vsk);
 s64 vsock_stream_has_space(struct vsock_sock *vsk);
 struct sock *vsock_create_connected(struct sock *parent);
@@ -173,6 +174,9 @@ struct vsock_transport {
 
 	/* Addressing. */
 	u32 (*get_local_cid)(void);
+
+	/* Read a single skb */
+	int (*read_skb)(struct vsock_sock *, skb_read_actor_t);
 };
 
 /**** CORE ****/
@@ -225,5 +229,18 @@ int vsock_init_tap(void);
 int vsock_add_tap(struct vsock_tap *vt);
 int vsock_remove_tap(struct vsock_tap *vt);
 void vsock_deliver_tap(struct sk_buff *build_skb(void *opaque), void *opaque);
+int vsock_connectible_recvmsg(struct socket *sock, struct msghdr *msg, size_t len,
+			      int flags);
+int vsock_dgram_recvmsg(struct socket *sock, struct msghdr *msg,
+			size_t len, int flags);
+
+#ifdef CONFIG_BPF_SYSCALL
+extern struct proto vsock_proto;
+int vsock_bpf_update_proto(struct sock *sk, struct sk_psock *psock, bool restore);
+void __init vsock_bpf_build_proto(void);
+#else
+static inline void __init vsock_bpf_build_proto(void)
+{}
+#endif
 
 #endif /* __AF_VSOCK_H__ */

net/vmw_vsock/Makefile

@@ -8,6 +8,7 @@ obj-$(CONFIG_HYPERV_VSOCKETS) += hv_sock.o
 obj-$(CONFIG_VSOCKETS_LOOPBACK) += vsock_loopback.o
 
 vsock-y += af_vsock.o af_vsock_tap.o vsock_addr.o
+vsock-$(CONFIG_BPF_SYSCALL) += vsock_bpf.o
 
 vsock_diag-y += diag.o
 

net/vmw_vsock/af_vsock.c

@@ -116,10 +116,13 @@ static void vsock_sk_destruct(struct sock *sk);
 static int vsock_queue_rcv_skb(struct sock *sk, struct sk_buff *skb);
 
 /* Protocol family. */
-static struct proto vsock_proto = {
+struct proto vsock_proto = {
 	.name = "AF_VSOCK",
 	.owner = THIS_MODULE,
 	.obj_size = sizeof(struct vsock_sock),
+#ifdef CONFIG_BPF_SYSCALL
+	.psock_update_sk_prot = vsock_bpf_update_proto,
+#endif
 };
 
 /* The default peer timeout indicates how long we will wait for a peer response
@@ -865,7 +868,7 @@ s64 vsock_stream_has_data(struct vsock_sock *vsk)
 }
 EXPORT_SYMBOL_GPL(vsock_stream_has_data);
 
-static s64 vsock_connectible_has_data(struct vsock_sock *vsk)
+s64 vsock_connectible_has_data(struct vsock_sock *vsk)
 {
 	struct sock *sk = sk_vsock(vsk);
 
@@ -874,6 +877,7 @@ static s64 vsock_connectible_has_data(struct vsock_sock *vsk)
 	else
 		return vsock_stream_has_data(vsk);
 }
+EXPORT_SYMBOL_GPL(vsock_connectible_has_data);
 
 s64 vsock_stream_has_space(struct vsock_sock *vsk)
 {
@@ -1131,6 +1135,13 @@ static __poll_t vsock_poll(struct file *file, struct socket *sock,
 	return mask;
 }
 
+static int vsock_read_skb(struct sock *sk, skb_read_actor_t read_actor)
+{
+	struct vsock_sock *vsk = vsock_sk(sk);
+
+	return vsk->transport->read_skb(vsk, read_actor);
+}
+
 static int vsock_dgram_sendmsg(struct socket *sock, struct msghdr *msg,
 			       size_t len)
 {
@@ -1242,18 +1253,42 @@ static int vsock_dgram_connect(struct socket *sock,
 	memcpy(&vsk->remote_addr, remote_addr, sizeof(vsk->remote_addr));
 	sock->state = SS_CONNECTED;
 
+	/* sock map disallows redirection of non-TCP sockets with sk_state !=
+	 * TCP_ESTABLISHED (see sock_map_redirect_allowed()), so we set
+	 * TCP_ESTABLISHED here to allow redirection of connected vsock dgrams.
+	 *
+	 * This doesn't seem to be abnormal state for datagram sockets, as the
+	 * same approach can be see in other datagram socket types as well
+	 * (such as unix sockets).
+	 */
+	sk->sk_state = TCP_ESTABLISHED;
+
 out:
 	release_sock(sk);
 	return err;
 }
 
-static int vsock_dgram_recvmsg(struct socket *sock, struct msghdr *msg,
-			       size_t len, int flags)
+int vsock_dgram_recvmsg(struct socket *sock, struct msghdr *msg,
+			size_t len, int flags)
 {
-	struct vsock_sock *vsk = vsock_sk(sock->sk);
+#ifdef CONFIG_BPF_SYSCALL
+	const struct proto *prot;
+#endif
+	struct vsock_sock *vsk;
+	struct sock *sk;
+
+	sk = sock->sk;
+	vsk = vsock_sk(sk);
+
+#ifdef CONFIG_BPF_SYSCALL
+	prot = READ_ONCE(sk->sk_prot);
+	if (prot != &vsock_proto)
+		return prot->recvmsg(sk, msg, len, flags, NULL);
+#endif
 
 	return vsk->transport->dgram_dequeue(vsk, msg, len, flags);
 }
+EXPORT_SYMBOL_GPL(vsock_dgram_recvmsg);
 
 static const struct proto_ops vsock_dgram_ops = {
 	.family = PF_VSOCK,
@@ -1272,6 +1307,7 @@ static const struct proto_ops vsock_dgram_ops = {
 	.recvmsg = vsock_dgram_recvmsg,
 	.mmap = sock_no_mmap,
 	.sendpage = sock_no_sendpage,
+	.read_skb = vsock_read_skb,
 };
 
 static int vsock_transport_cancel_pkt(struct vsock_sock *vsk)
@@ -2086,13 +2122,16 @@ static int __vsock_seqpacket_recvmsg(struct sock *sk, struct msghdr *msg,
 	return err;
 }
 
-static int
+int
 vsock_connectible_recvmsg(struct socket *sock, struct msghdr *msg, size_t len,
 			  int flags)
 {
 	struct sock *sk;
 	struct vsock_sock *vsk;
 	const struct vsock_transport *transport;
+#ifdef CONFIG_BPF_SYSCALL
+	const struct proto *prot;
+#endif
 	int err;
 
 	sk = sock->sk;
@@ -2139,6 +2178,14 @@ vsock_connectible_recvmsg(struct socket *sock, struct msghdr *msg, size_t len,
 		goto out;
 	}
 
+#ifdef CONFIG_BPF_SYSCALL
+	prot = READ_ONCE(sk->sk_prot);
+	if (prot != &vsock_proto) {
+		release_sock(sk);
+		return prot->recvmsg(sk, msg, len, flags, NULL);
+	}
+#endif
+
 	if (sk->sk_type == SOCK_STREAM)
 		err = __vsock_stream_recvmsg(sk, msg, len, flags);
 	else
@@ -2148,6 +2195,7 @@ vsock_connectible_recvmsg(struct socket *sock, struct msghdr *msg, size_t len,
 	release_sock(sk);
 	return err;
 }
+EXPORT_SYMBOL_GPL(vsock_connectible_recvmsg);
 
 static int vsock_set_rcvlowat(struct sock *sk, int val)
 {
@@ -2188,6 +2236,7 @@ static const struct proto_ops vsock_stream_ops = {
 	.mmap = sock_no_mmap,
 	.sendpage = sock_no_sendpage,
 	.set_rcvlowat = vsock_set_rcvlowat,
+	.read_skb = vsock_read_skb,
 };
 
 static const struct proto_ops vsock_seqpacket_ops = {
@@ -2209,6 +2258,7 @@ static const struct proto_ops vsock_seqpacket_ops = {
 	.recvmsg = vsock_connectible_recvmsg,
 	.mmap = sock_no_mmap,
 	.sendpage = sock_no_sendpage,
+	.read_skb = vsock_read_skb,
 };
 
 static int vsock_create(struct net *net, struct socket *sock,
@@ -2348,6 +2398,8 @@ static int __init vsock_init(void)
 		goto err_unregister_proto;
 	}
 
+	vsock_bpf_build_proto();
+
 	return 0;
 
 err_unregister_proto:

net/vmw_vsock/virtio_transport.c

@@ -457,6 +457,8 @@ static struct virtio_transport virtio_transport = {
 		.notify_send_pre_enqueue  = virtio_transport_notify_send_pre_enqueue,
 		.notify_send_post_enqueue = virtio_transport_notify_send_post_enqueue,
 		.notify_buffer_size       = virtio_transport_notify_buffer_size,
+
+		.read_skb = virtio_transport_read_skb,
 	},
 
 	.send_pkt = virtio_transport_send_pkt,

net/vmw_vsock/virtio_transport_common.c

@@ -1418,6 +1418,31 @@ int virtio_transport_purge_skbs(void *vsk, struct sk_buff_head *queue)
 }
 EXPORT_SYMBOL_GPL(virtio_transport_purge_skbs);
 
+int virtio_transport_read_skb(struct vsock_sock *vsk, skb_read_actor_t recv_actor)
+{
+	struct virtio_vsock_sock *vvs = vsk->trans;
+	struct sock *sk = sk_vsock(vsk);
+	struct sk_buff *skb;
+	int off = 0;
+	int copied;
+	int err;
+
+	spin_lock_bh(&vvs->rx_lock);
+	/* Use __skb_recv_datagram() for race-free handling of the receive. It
+	 * works for types other than dgrams.
+	 */
+	skb = __skb_recv_datagram(sk, &vvs->rx_queue, MSG_DONTWAIT, &off, &err);
+	spin_unlock_bh(&vvs->rx_lock);
+
+	if (!skb)
+		return err;
+
+	copied = recv_actor(sk, skb);
+	kfree_skb(skb);
+	return copied;
+}
+EXPORT_SYMBOL_GPL(virtio_transport_read_skb);
+
 MODULE_LICENSE("GPL v2");
 MODULE_AUTHOR("Asias He");
 MODULE_DESCRIPTION("common code for virtio vsock");