net/macvtap: add vhost support

This adds support for passing a macvtap file descriptor into vhost-net, much like we already do for tun/tap. Most of the new code is taken from the respective patch in the tun driver and may get consolidated in the future. Signed-off-by: Arnd Bergmann <arnd@arndb.de> Acked-by: Sridhar Samudrala <sri@us.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>

net/macvtap: add vhost support
This adds support for passing a macvtap file descriptor into vhost-net, much like we already do for tun/tap. Most of the new code is taken from the respective patch in the tun driver and may get consolidated in the future. Signed-off-by: Arnd Bergmann <arnd@arndb.de> Acked-by: Sridhar Samudrala <sri@us.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Arnd Bergmann · David S. Miller
1 parent 02df55d28c
Showing 4 changed files with 97 additions and 24 deletions Side-by-side Diff
drivers/net/macvtap.c
drivers/vhost/Kconfig
drivers/vhost/net.c
include/linux/if_macvlan.h
@@ -58,6 +58,8 @@
 static struct class *macvtap_class;
 static struct cdev macvtap_cdev;
  
+static const struct proto_ops macvtap_socket_ops;
+
 /*
  * RCU usage:
  * The macvtap_queue and the macvlan_dev are loosely coupled, the
@@ -176,7 +178,7 @@
 		return -ENOLINK;
  
 	skb_queue_tail(&q->sk.sk_receive_queue, skb);
-	wake_up(q->sk.sk_sleep);
+	wake_up_interruptible_poll(q->sk.sk_sleep, POLLIN | POLLRDNORM | POLLRDBAND);
 	return 0;
 }
  
@@ -242,7 +244,7 @@
 		return;
  
 	if (sk->sk_sleep && waitqueue_active(sk->sk_sleep))
-		wake_up_interruptible_sync(sk->sk_sleep);
+		wake_up_interruptible_poll(sk->sk_sleep, POLLOUT | POLLWRNORM | POLLWRBAND);
 }
  
 static int macvtap_open(struct inode *inode, struct file *file)
@@ -270,6 +272,8 @@
 	init_waitqueue_head(&q->sock.wait);
 	q->sock.type = SOCK_RAW;
 	q->sock.state = SS_CONNECTED;
+	q->sock.file = file;
+	q->sock.ops = &macvtap_socket_ops;
 	sock_init_data(&q->sock, &q->sk);
 	q->sk.sk_write_space = macvtap_sock_write_space;
  
  
  
  
  
@@ -387,33 +391,21 @@
  
 	rcu_read_lock_bh();
 	vlan = rcu_dereference(q->vlan);
-	macvlan_count_rx(vlan, len, ret == 0, 0);
+	if (vlan)
+		macvlan_count_rx(vlan, len, ret == 0, 0);
 	rcu_read_unlock_bh();
  
 	return ret ? ret : len;
 }
  
-static ssize_t macvtap_aio_read(struct kiocb *iocb, const struct iovec *iv,
-				unsigned long count, loff_t pos)
+static ssize_t macvtap_do_read(struct macvtap_queue *q, struct kiocb *iocb,
+			       const struct iovec *iv, unsigned long len,
+			       int noblock)
 {
-	struct file *file = iocb->ki_filp;
-	struct macvtap_queue *q = file->private_data;
-
 	DECLARE_WAITQUEUE(wait, current);
 	struct sk_buff *skb;
-	ssize_t len, ret = 0;
+	ssize_t ret = 0;
  
-	if (!q) {
-		ret = -ENOLINK;
-		goto out;
-	}
-
-	len = iov_length(iv, count);
-	if (len < 0) {
-		ret = -EINVAL;
-		goto out;
-	}
-
 	add_wait_queue(q->sk.sk_sleep, &wait);
 	while (len) {
 		current->state = TASK_INTERRUPTIBLE;
@@ -421,7 +413,7 @@
 		/* Read frames from the queue */
 		skb = skb_dequeue(&q->sk.sk_receive_queue);
 		if (!skb) {
-			if (file->f_flags & O_NONBLOCK) {
+			if (noblock) {
 				ret = -EAGAIN;
 				break;
 			}
  
@@ -440,7 +432,24 @@
  
 	current->state = TASK_RUNNING;
 	remove_wait_queue(q->sk.sk_sleep, &wait);
+	return ret;
+}
  
+static ssize_t macvtap_aio_read(struct kiocb *iocb, const struct iovec *iv,
+				unsigned long count, loff_t pos)
+{
+	struct file *file = iocb->ki_filp;
+	struct macvtap_queue *q = file->private_data;
+	ssize_t len, ret = 0;
+
+	len = iov_length(iv, count);
+	if (len < 0) {
+		ret = -EINVAL;
+		goto out;
+	}
+
+	ret = macvtap_do_read(q, iocb, iv, len, file->f_flags & O_NONBLOCK);
+	ret = min_t(ssize_t, ret, len); /* XXX copied from tun.c. Why? */
 out:
 	return ret;
 }
@@ -537,6 +546,53 @@
 	.compat_ioctl	= macvtap_compat_ioctl,
 #endif
 };
+
+static int macvtap_sendmsg(struct kiocb *iocb, struct socket *sock,
+			   struct msghdr *m, size_t total_len)
+{
+	struct macvtap_queue *q = container_of(sock, struct macvtap_queue, sock);
+	return macvtap_get_user(q, m->msg_iov, total_len,
+			    m->msg_flags & MSG_DONTWAIT);
+}
+
+static int macvtap_recvmsg(struct kiocb *iocb, struct socket *sock,
+			   struct msghdr *m, size_t total_len,
+			   int flags)
+{
+	struct macvtap_queue *q = container_of(sock, struct macvtap_queue, sock);
+	int ret;
+	if (flags & ~(MSG_DONTWAIT|MSG_TRUNC))
+		return -EINVAL;
+	ret = macvtap_do_read(q, iocb, m->msg_iov, total_len,
+			  flags & MSG_DONTWAIT);
+	if (ret > total_len) {
+		m->msg_flags |= MSG_TRUNC;
+		ret = flags & MSG_TRUNC ? ret : total_len;
+	}
+	return ret;
+}
+
+/* Ops structure to mimic raw sockets with tun */
+static const struct proto_ops macvtap_socket_ops = {
+	.sendmsg = macvtap_sendmsg,
+	.recvmsg = macvtap_recvmsg,
+};
+
+/* Get an underlying socket object from tun file.  Returns error unless file is
+ * attached to a device.  The returned object works like a packet socket, it
+ * can be used for sock_sendmsg/sock_recvmsg.  The caller is responsible for
+ * holding a reference to the file for as long as the socket is in use. */
+struct socket *macvtap_get_socket(struct file *file)
+{
+	struct macvtap_queue *q;
+	if (file->f_op != &macvtap_fops)
+		return ERR_PTR(-EINVAL);
+	q = file->private_data;
+	if (!q)
+		return ERR_PTR(-EBADFD);
+	return &q->sock;
+}
+EXPORT_SYMBOL_GPL(macvtap_get_socket);
  
 static int macvtap_init(void)
 {
 config VHOST_NET
 	tristate "Host kernel accelerator for virtio net (EXPERIMENTAL)"
-	depends on NET && EVENTFD && (TUN || !TUN) && EXPERIMENTAL
+	depends on NET && EVENTFD && (TUN || !TUN) && (MACVTAP || !MACVTAP) && EXPERIMENTAL
 	---help---
 	  This kernel module can be loaded in host kernel to accelerate
 	  guest networking with virtio_net. Not to be confused with virtio_net
@@ -22,6 +22,7 @@
 #include <linux/if_packet.h>
 #include <linux/if_arp.h>
 #include <linux/if_tun.h>
+#include <linux/if_macvlan.h>
  
 #include <net/sock.h>
  
  
@@ -452,13 +453,16 @@
 	return ERR_PTR(r);
 }
  
-static struct socket *get_tun_socket(int fd)
+static struct socket *get_tap_socket(int fd)
 {
 	struct file *file = fget(fd);
 	struct socket *sock;
 	if (!file)
 		return ERR_PTR(-EBADF);
 	sock = tun_get_socket(file);
+	if (!IS_ERR(sock))
+		return sock;
+	sock = macvtap_get_socket(file);
 	if (IS_ERR(sock))
 		fput(file);
 	return sock;
@@ -473,7 +477,7 @@
 	sock = get_raw_socket(fd);
 	if (!IS_ERR(sock))
 		return sock;
-	sock = get_tun_socket(fd);
+	sock = get_tap_socket(fd);
 	if (!IS_ERR(sock))
 		return sock;
 	return ERR_PTR(-ENOTSOCK);
@@ -7,6 +7,19 @@
 #include <linux/netlink.h>
 #include <net/netlink.h>
  
+#if defined(CONFIG_MACVTAP) || defined(CONFIG_MACVTAP_MODULE)
+struct socket *macvtap_get_socket(struct file *);
+#else
+#include <linux/err.h>
+#include <linux/errno.h>
+struct file;
+struct socket;
+static inline struct socket *macvtap_get_socket(struct file *f)
+{
+	return ERR_PTR(-EINVAL);
+}
+#endif /* CONFIG_MACVTAP */
+
 struct macvlan_port;
 struct macvtap_queue;
...	...	@@ -58,6 +58,8 @@
58	58	static struct class *macvtap_class;
59	59	static struct cdev macvtap_cdev;
60	60
	61	+static const struct proto_ops macvtap_socket_ops;
	62	+
61	63	/*
62	64	* RCU usage:
63	65	* The macvtap_queue and the macvlan_dev are loosely coupled, the
...	...	@@ -176,7 +178,7 @@
176	178	return -ENOLINK;
177	179
178	180	skb_queue_tail(&q->sk.sk_receive_queue, skb);
179		- wake_up(q->sk.sk_sleep);
	181	+ wake_up_interruptible_poll(q->sk.sk_sleep, POLLIN \| POLLRDNORM \| POLLRDBAND);
180	182	return 0;
181	183	}
182	184
...	...	@@ -242,7 +244,7 @@
242	244	return;
243	245
244	246	if (sk->sk_sleep && waitqueue_active(sk->sk_sleep))
245		- wake_up_interruptible_sync(sk->sk_sleep);
	247	+ wake_up_interruptible_poll(sk->sk_sleep, POLLOUT \| POLLWRNORM \| POLLWRBAND);
246	248	}
247	249
248	250	static int macvtap_open(struct inode inode, struct file file)
...	...	@@ -270,6 +272,8 @@
270	272	init_waitqueue_head(&q->sock.wait);
271	273	q->sock.type = SOCK_RAW;
272	274	q->sock.state = SS_CONNECTED;
	275	+ q->sock.file = file;
	276	+ q->sock.ops = &macvtap_socket_ops;
273	277	sock_init_data(&q->sock, &q->sk);
274	278	q->sk.sk_write_space = macvtap_sock_write_space;
275	279
276	280
277	281
278	282
279	283
...	...	@@ -387,33 +391,21 @@
387	391
388	392	rcu_read_lock_bh();
389	393	vlan = rcu_dereference(q->vlan);
390		- macvlan_count_rx(vlan, len, ret == 0, 0);
	394	+ if (vlan)
	395	+ macvlan_count_rx(vlan, len, ret == 0, 0);
391	396	rcu_read_unlock_bh();
392	397
393	398	return ret ? ret : len;
394	399	}
395	400
396		-static ssize_t macvtap_aio_read(struct kiocb iocb, const struct iovec iv,
397		- unsigned long count, loff_t pos)
	401	+static ssize_t macvtap_do_read(struct macvtap_queue q, struct kiocb iocb,
	402	+ const struct iovec *iv, unsigned long len,
	403	+ int noblock)
398	404	{
399		- struct file *file = iocb->ki_filp;
400		- struct macvtap_queue *q = file->private_data;
401		-
402	405	DECLARE_WAITQUEUE(wait, current);
403	406	struct sk_buff *skb;
404		- ssize_t len, ret = 0;
	407	+ ssize_t ret = 0;
405	408
406		- if (!q) {
407		- ret = -ENOLINK;
408		- goto out;
409		- }
410		-
411		- len = iov_length(iv, count);
412		- if (len < 0) {
413		- ret = -EINVAL;
414		- goto out;
415		- }
416		-
417	409	add_wait_queue(q->sk.sk_sleep, &wait);
418	410	while (len) {
419	411	current->state = TASK_INTERRUPTIBLE;
...	...	@@ -421,7 +413,7 @@
421	413	/* Read frames from the queue */
422	414	skb = skb_dequeue(&q->sk.sk_receive_queue);
423	415	if (!skb) {
424		- if (file->f_flags & O_NONBLOCK) {
	416	+ if (noblock) {
425	417	ret = -EAGAIN;
426	418	break;
427	419	}
428	420
...	...	@@ -440,7 +432,24 @@
440	432
441	433	current->state = TASK_RUNNING;
442	434	remove_wait_queue(q->sk.sk_sleep, &wait);
	435	+ return ret;
	436	+}
443	437
	438	+static ssize_t macvtap_aio_read(struct kiocb iocb, const struct iovec iv,
	439	+ unsigned long count, loff_t pos)
	440	+{
	441	+ struct file *file = iocb->ki_filp;
	442	+ struct macvtap_queue *q = file->private_data;
	443	+ ssize_t len, ret = 0;
	444	+
	445	+ len = iov_length(iv, count);
	446	+ if (len < 0) {
	447	+ ret = -EINVAL;
	448	+ goto out;
	449	+ }
	450	+
	451	+ ret = macvtap_do_read(q, iocb, iv, len, file->f_flags & O_NONBLOCK);
	452	+ ret = min_t(ssize_t, ret, len); /* XXX copied from tun.c. Why? */
444	453	out:
445	454	return ret;
446	455	}
...	...	@@ -537,6 +546,53 @@
537	546	.compat_ioctl = macvtap_compat_ioctl,
538	547	#endif
539	548	};
	549	+
	550	+static int macvtap_sendmsg(struct kiocb iocb, struct socket sock,
	551	+ struct msghdr *m, size_t total_len)
	552	+{
	553	+ struct macvtap_queue *q = container_of(sock, struct macvtap_queue, sock);
	554	+ return macvtap_get_user(q, m->msg_iov, total_len,
	555	+ m->msg_flags & MSG_DONTWAIT);
	556	+}
	557	+
	558	+static int macvtap_recvmsg(struct kiocb iocb, struct socket sock,
	559	+ struct msghdr *m, size_t total_len,
	560	+ int flags)
	561	+{
	562	+ struct macvtap_queue *q = container_of(sock, struct macvtap_queue, sock);
	563	+ int ret;
	564	+ if (flags & ~(MSG_DONTWAIT\|MSG_TRUNC))
	565	+ return -EINVAL;
	566	+ ret = macvtap_do_read(q, iocb, m->msg_iov, total_len,
	567	+ flags & MSG_DONTWAIT);
	568	+ if (ret > total_len) {
	569	+ m->msg_flags \|= MSG_TRUNC;
	570	+ ret = flags & MSG_TRUNC ? ret : total_len;
	571	+ }
	572	+ return ret;
	573	+}
	574	+
	575	+/* Ops structure to mimic raw sockets with tun */
	576	+static const struct proto_ops macvtap_socket_ops = {
	577	+ .sendmsg = macvtap_sendmsg,
	578	+ .recvmsg = macvtap_recvmsg,
	579	+};
	580	+
	581	+/* Get an underlying socket object from tun file. Returns error unless file is
	582	+ * attached to a device. The returned object works like a packet socket, it
	583	+ * can be used for sock_sendmsg/sock_recvmsg. The caller is responsible for
	584	+ * holding a reference to the file for as long as the socket is in use. */
	585	+struct socket macvtap_get_socket(struct file file)
	586	+{
	587	+ struct macvtap_queue *q;
	588	+ if (file->f_op != &macvtap_fops)
	589	+ return ERR_PTR(-EINVAL);
	590	+ q = file->private_data;
	591	+ if (!q)
	592	+ return ERR_PTR(-EBADFD);
	593	+ return &q->sock;
	594	+}
	595	+EXPORT_SYMBOL_GPL(macvtap_get_socket);
540	596
541	597	static int macvtap_init(void)
542	598	{
1	1	config VHOST_NET
2	2	tristate "Host kernel accelerator for virtio net (EXPERIMENTAL)"
3		- depends on NET && EVENTFD && (TUN \|\| !TUN) && EXPERIMENTAL
	3	+ depends on NET && EVENTFD && (TUN \|\| !TUN) && (MACVTAP \|\| !MACVTAP) && EXPERIMENTAL
4	4	---help---
5	5	This kernel module can be loaded in host kernel to accelerate
6	6	guest networking with virtio_net. Not to be confused with virtio_net
...	...	@@ -22,6 +22,7 @@
22	22	#include <linux/if_packet.h>
23	23	#include <linux/if_arp.h>
24	24	#include <linux/if_tun.h>
	25	+#include <linux/if_macvlan.h>
25	26
26	27	#include <net/sock.h>
27	28
28	29
...	...	@@ -452,13 +453,16 @@
452	453	return ERR_PTR(r);
453	454	}
454	455
455		-static struct socket *get_tun_socket(int fd)
	456	+static struct socket *get_tap_socket(int fd)
456	457	{
457	458	struct file *file = fget(fd);
458	459	struct socket *sock;
459	460	if (!file)
460	461	return ERR_PTR(-EBADF);
461	462	sock = tun_get_socket(file);
	463	+ if (!IS_ERR(sock))
	464	+ return sock;
	465	+ sock = macvtap_get_socket(file);
462	466	if (IS_ERR(sock))
463	467	fput(file);
464	468	return sock;
...	...	@@ -473,7 +477,7 @@
473	477	sock = get_raw_socket(fd);
474	478	if (!IS_ERR(sock))
475	479	return sock;
476		- sock = get_tun_socket(fd);
	480	+ sock = get_tap_socket(fd);
477	481	if (!IS_ERR(sock))
478	482	return sock;
479	483	return ERR_PTR(-ENOTSOCK);
...	...	@@ -7,6 +7,19 @@
7	7	#include <linux/netlink.h>
8	8	#include <net/netlink.h>
9	9
	10	+#if defined(CONFIG_MACVTAP) \|\| defined(CONFIG_MACVTAP_MODULE)
	11	+struct socket macvtap_get_socket(struct file );
	12	+#else
	13	+#include <linux/err.h>
	14	+#include <linux/errno.h>
	15	+struct file;
	16	+struct socket;
	17	+static inline struct socket macvtap_get_socket(struct file f)
	18	+{
	19	+ return ERR_PTR(-EINVAL);
	20	+}
	21	+#endif /* CONFIG_MACVTAP */
	22	+
10	23	struct macvlan_port;
11	24	struct macvtap_queue;
12	25