macvtap.c 30.4 KB
Newer Older
A
Arnd Bergmann 已提交
1 2
#include <linux/etherdevice.h>
#include <linux/if_macvlan.h>
3
#include <linux/if_vlan.h>
A
Arnd Bergmann 已提交
4 5 6 7 8 9 10 11 12
#include <linux/interrupt.h>
#include <linux/nsproxy.h>
#include <linux/compat.h>
#include <linux/if_tun.h>
#include <linux/module.h>
#include <linux/skbuff.h>
#include <linux/cache.h>
#include <linux/sched.h>
#include <linux/types.h>
13
#include <linux/slab.h>
A
Arnd Bergmann 已提交
14 15 16
#include <linux/init.h>
#include <linux/wait.h>
#include <linux/cdev.h>
A
Al Viro 已提交
17
#include <linux/idr.h>
A
Arnd Bergmann 已提交
18 19 20 21 22
#include <linux/fs.h>

#include <net/net_namespace.h>
#include <net/rtnetlink.h>
#include <net/sock.h>
23
#include <linux/virtio_net.h>
A
Arnd Bergmann 已提交
24 25 26 27 28 29 30 31 32 33 34 35 36 37

/*
 * A macvtap queue is the central object of this driver, it connects
 * an open character device to a macvlan interface. There can be
 * multiple queues on one interface, which map back to queues
 * implemented in hardware on the underlying device.
 *
 * macvtap_proto is used to allocate queues through the sock allocation
 * mechanism.
 *
 */
struct macvtap_queue {
	struct sock sk;
	struct socket sock;
38
	struct socket_wq wq;
39
	int vnet_hdr_sz;
40
	struct macvlan_dev __rcu *vlan;
A
Arnd Bergmann 已提交
41
	struct file *file;
42
	unsigned int flags;
J
Jason Wang 已提交
43
	u16 queue_index;
J
Jason Wang 已提交
44 45
	bool enabled;
	struct list_head next;
A
Arnd Bergmann 已提交
46 47 48 49 50 51 52 53 54
};

static struct proto macvtap_proto = {
	.name = "macvtap",
	.owner = THIS_MODULE,
	.obj_size = sizeof (struct macvtap_queue),
};

/*
55
 * Variables for dealing with macvtaps device numbers.
A
Arnd Bergmann 已提交
56
 */
57
static dev_t macvtap_major;
58 59 60 61
#define MACVTAP_NUM_DEVS (1U << MINORBITS)
static DEFINE_MUTEX(minor_lock);
static DEFINE_IDR(minor_idr);

62
#define GOODCOPY_LEN 128
A
Arnd Bergmann 已提交
63 64 65
static struct class *macvtap_class;
static struct cdev macvtap_cdev;

A
Arnd Bergmann 已提交
66 67
static const struct proto_ops macvtap_socket_ops;

68 69 70
#define TUN_OFFLOADS (NETIF_F_HW_CSUM | NETIF_F_TSO_ECN | NETIF_F_TSO | \
		      NETIF_F_TSO6 | NETIF_F_UFO)
#define RX_OFFLOADS (NETIF_F_GRO | NETIF_F_LRO)
A
Arnd Bergmann 已提交
71 72
/*
 * RCU usage:
73 74
 * The macvtap_queue and the macvlan_dev are loosely coupled, the
 * pointers from one to the other can only be read while rcu_read_lock
75
 * or rtnl is held.
A
Arnd Bergmann 已提交
76
 *
77 78 79 80
 * Both the file and the macvlan_dev hold a reference on the macvtap_queue
 * through sock_hold(&q->sk). When the macvlan_dev goes away first,
 * q->vlan becomes inaccessible. When the files gets closed,
 * macvtap_get_queue() fails.
A
Arnd Bergmann 已提交
81
 *
82 83 84 85
 * There may still be references to the struct sock inside of the
 * queue from outbound SKBs, but these never reference back to the
 * file or the dev. The data structure is freed through __sk_free
 * when both our references and any pending SKBs are gone.
A
Arnd Bergmann 已提交
86 87
 */

J
Jason Wang 已提交
88
static int macvtap_enable_queue(struct net_device *dev, struct file *file,
A
Arnd Bergmann 已提交
89
				struct macvtap_queue *q)
J
Jason Wang 已提交
90 91 92 93
{
	struct macvlan_dev *vlan = netdev_priv(dev);
	int err = -EINVAL;

94
	ASSERT_RTNL();
J
Jason Wang 已提交
95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110

	if (q->enabled)
		goto out;

	err = 0;
	rcu_assign_pointer(vlan->taps[vlan->numvtaps], q);
	q->queue_index = vlan->numvtaps;
	q->enabled = true;

	vlan->numvtaps++;
out:
	return err;
}

static int macvtap_set_queue(struct net_device *dev, struct file *file,
			     struct macvtap_queue *q)
A
Arnd Bergmann 已提交
111 112 113 114
{
	struct macvlan_dev *vlan = netdev_priv(dev);
	int err = -EBUSY;

115
	rtnl_lock();
J
Jason Wang 已提交
116
	if (vlan->numqueues == MAX_MACVTAP_QUEUES)
A
Arnd Bergmann 已提交
117 118 119
		goto out;

	err = 0;
120
	rcu_assign_pointer(q->vlan, vlan);
J
Jason Wang 已提交
121
	rcu_assign_pointer(vlan->taps[vlan->numvtaps], q);
122
	sock_hold(&q->sk);
A
Arnd Bergmann 已提交
123 124

	q->file = file;
J
Jason Wang 已提交
125
	q->queue_index = vlan->numvtaps;
J
Jason Wang 已提交
126
	q->enabled = true;
127
	file->private_data = q;
J
Jason Wang 已提交
128
	list_add_tail(&q->next, &vlan->queue_list);
A
Arnd Bergmann 已提交
129

130
	vlan->numvtaps++;
J
Jason Wang 已提交
131
	vlan->numqueues++;
132

A
Arnd Bergmann 已提交
133
out:
134
	rtnl_unlock();
A
Arnd Bergmann 已提交
135 136 137
	return err;
}

138
static int macvtap_disable_queue(struct macvtap_queue *q)
J
Jason Wang 已提交
139 140 141 142
{
	struct macvlan_dev *vlan;
	struct macvtap_queue *nq;

143
	ASSERT_RTNL();
J
Jason Wang 已提交
144 145 146
	if (!q->enabled)
		return -EINVAL;

147 148
	vlan = rtnl_dereference(q->vlan);

J
Jason Wang 已提交
149 150 151
	if (vlan) {
		int index = q->queue_index;
		BUG_ON(index >= vlan->numvtaps);
152
		nq = rtnl_dereference(vlan->taps[vlan->numvtaps - 1]);
J
Jason Wang 已提交
153 154 155 156 157 158 159 160 161 162 163 164
		nq->queue_index = index;

		rcu_assign_pointer(vlan->taps[index], nq);
		RCU_INIT_POINTER(vlan->taps[vlan->numvtaps - 1], NULL);
		q->enabled = false;

		vlan->numvtaps--;
	}

	return 0;
}

A
Arnd Bergmann 已提交
165
/*
166 167 168
 * The file owning the queue got closed, give up both
 * the reference that the files holds as well as the
 * one from the macvlan_dev if that still exists.
A
Arnd Bergmann 已提交
169 170 171 172
 *
 * Using the spinlock makes sure that we don't get
 * to the queue again after destroying it.
 */
173
static void macvtap_put_queue(struct macvtap_queue *q)
A
Arnd Bergmann 已提交
174
{
175
	struct macvlan_dev *vlan;
A
Arnd Bergmann 已提交
176

177 178 179
	rtnl_lock();
	vlan = rtnl_dereference(q->vlan);

180
	if (vlan) {
J
Jason Wang 已提交
181
		if (q->enabled)
182
			BUG_ON(macvtap_disable_queue(q));
J
Jason Wang 已提交
183

J
Jason Wang 已提交
184
		vlan->numqueues--;
185
		RCU_INIT_POINTER(q->vlan, NULL);
186
		sock_put(&q->sk);
J
Jason Wang 已提交
187
		list_del_init(&q->next);
A
Arnd Bergmann 已提交
188 189
	}

190
	rtnl_unlock();
A
Arnd Bergmann 已提交
191 192 193 194 195 196

	synchronize_rcu();
	sock_put(&q->sk);
}

/*
197 198 199 200 201
 * Select a queue based on the rxq of the device on which this packet
 * arrived. If the incoming device is not mq, calculate a flow hash
 * to select a queue. If all fails, find the first available queue.
 * Cache vlan->numvtaps since it can become zero during the execution
 * of this function.
A
Arnd Bergmann 已提交
202 203 204 205 206
 */
static struct macvtap_queue *macvtap_get_queue(struct net_device *dev,
					       struct sk_buff *skb)
{
	struct macvlan_dev *vlan = netdev_priv(dev);
207
	struct macvtap_queue *tap = NULL;
J
Jason Wang 已提交
208 209 210 211 212
	/* Access to taps array is protected by rcu, but access to numvtaps
	 * isn't. Below we use it to lookup a queue, but treat it as a hint
	 * and validate that the result isn't NULL - in case we are
	 * racing against queue removal.
	 */
213
	int numvtaps = ACCESS_ONCE(vlan->numvtaps);
214 215 216 217 218
	__u32 rxq;

	if (!numvtaps)
		goto out;

219 220 221 222
	/* Check if we can use flow to select a queue */
	rxq = skb_get_rxhash(skb);
	if (rxq) {
		tap = rcu_dereference(vlan->taps[rxq % numvtaps]);
J
Jason Wang 已提交
223
		goto out;
224 225
	}

226 227
	if (likely(skb_rx_queue_recorded(skb))) {
		rxq = skb_get_rx_queue(skb);
A
Arnd Bergmann 已提交
228

229 230 231 232
		while (unlikely(rxq >= numvtaps))
			rxq -= numvtaps;

		tap = rcu_dereference(vlan->taps[rxq]);
J
Jason Wang 已提交
233
		goto out;
234 235
	}

J
Jason Wang 已提交
236
	tap = rcu_dereference(vlan->taps[0]);
237 238
out:
	return tap;
A
Arnd Bergmann 已提交
239 240
}

241 242
/*
 * The net_device is going away, give up the reference
243 244
 * that it holds on all queues and safely set the pointer
 * from the queues to NULL.
245
 */
A
Arnd Bergmann 已提交
246 247 248
static void macvtap_del_queues(struct net_device *dev)
{
	struct macvlan_dev *vlan = netdev_priv(dev);
J
Jason Wang 已提交
249
	struct macvtap_queue *q, *tmp, *qlist[MAX_MACVTAP_QUEUES];
250
	int i, j = 0;
251

252
	ASSERT_RTNL();
J
Jason Wang 已提交
253 254
	list_for_each_entry_safe(q, tmp, &vlan->queue_list, next) {
		list_del_init(&q->next);
J
Jason Wang 已提交
255 256
		qlist[j++] = q;
		RCU_INIT_POINTER(q->vlan, NULL);
J
Jason Wang 已提交
257 258 259
		if (q->enabled)
			vlan->numvtaps--;
		vlan->numqueues--;
260
	}
J
Jason Wang 已提交
261 262 263 264
	for (i = 0; i < vlan->numvtaps; i++)
		RCU_INIT_POINTER(vlan->taps[i], NULL);
	BUG_ON(vlan->numvtaps);
	BUG_ON(vlan->numqueues);
265 266
	/* guarantee that any future macvtap_set_queue will fail */
	vlan->numvtaps = MAX_MACVTAP_QUEUES;
267 268 269

	for (--j; j >= 0; j--)
		sock_put(&qlist[j]->sk);
A
Arnd Bergmann 已提交
270 271 272 273 274 275 276 277 278 279 280
}

/*
 * Forward happens for data that gets sent from one macvlan
 * endpoint to another one in bridge mode. We just take
 * the skb and put it into the receive queue.
 */
static int macvtap_forward(struct net_device *dev, struct sk_buff *skb)
{
	struct macvtap_queue *q = macvtap_get_queue(dev, skb);
	if (!q)
H
Herbert Xu 已提交
281 282 283 284
		goto drop;

	if (skb_queue_len(&q->sk.sk_receive_queue) >= dev->tx_queue_len)
		goto drop;
A
Arnd Bergmann 已提交
285 286

	skb_queue_tail(&q->sk.sk_receive_queue, skb);
E
Eric Dumazet 已提交
287
	wake_up_interruptible_poll(sk_sleep(&q->sk), POLLIN | POLLRDNORM | POLLRDBAND);
H
Herbert Xu 已提交
288 289 290 291 292
	return NET_RX_SUCCESS;

drop:
	kfree_skb(skb);
	return NET_RX_DROP;
A
Arnd Bergmann 已提交
293 294 295 296 297 298 299 300 301 302 303 304 305
}

/*
 * Receive is for data from the external interface (lowerdev),
 * in case of macvtap, we can treat that the same way as
 * forward, which macvlan cannot.
 */
static int macvtap_receive(struct sk_buff *skb)
{
	skb_push(skb, ETH_HLEN);
	return macvtap_forward(skb->dev, skb);
}

306 307 308 309 310
static int macvtap_get_minor(struct macvlan_dev *vlan)
{
	int retval = -ENOMEM;

	mutex_lock(&minor_lock);
T
Tejun Heo 已提交
311 312 313 314
	retval = idr_alloc(&minor_idr, vlan, 1, MACVTAP_NUM_DEVS, GFP_KERNEL);
	if (retval >= 0) {
		vlan->minor = retval;
	} else if (retval == -ENOSPC) {
315 316 317 318
		printk(KERN_ERR "too many macvtap devices\n");
		retval = -EINVAL;
	}
	mutex_unlock(&minor_lock);
T
Tejun Heo 已提交
319
	return retval < 0 ? retval : 0;
320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346
}

static void macvtap_free_minor(struct macvlan_dev *vlan)
{
	mutex_lock(&minor_lock);
	if (vlan->minor) {
		idr_remove(&minor_idr, vlan->minor);
		vlan->minor = 0;
	}
	mutex_unlock(&minor_lock);
}

static struct net_device *dev_get_by_macvtap_minor(int minor)
{
	struct net_device *dev = NULL;
	struct macvlan_dev *vlan;

	mutex_lock(&minor_lock);
	vlan = idr_find(&minor_idr, minor);
	if (vlan) {
		dev = vlan->dev;
		dev_hold(dev);
	}
	mutex_unlock(&minor_lock);
	return dev;
}

A
Arnd Bergmann 已提交
347 348 349 350 351
static int macvtap_newlink(struct net *src_net,
			   struct net_device *dev,
			   struct nlattr *tb[],
			   struct nlattr *data[])
{
J
Jason Wang 已提交
352 353 354
	struct macvlan_dev *vlan = netdev_priv(dev);
	INIT_LIST_HEAD(&vlan->queue_list);

355 356 357 358 359
	/* Since macvlan supports all offloads by default, make
	 * tap support all offloads also.
	 */
	vlan->tap_features = TUN_OFFLOADS;

360 361 362 363 364
	/* Don't put anything that may fail after macvlan_common_newlink
	 * because we can't undo what it does.
	 */
	return macvlan_common_newlink(src_net, dev, tb, data,
				      macvtap_receive, macvtap_forward);
A
Arnd Bergmann 已提交
365 366 367 368 369 370 371 372 373
}

static void macvtap_dellink(struct net_device *dev,
			    struct list_head *head)
{
	macvtap_del_queues(dev);
	macvlan_dellink(dev, head);
}

H
Herbert Xu 已提交
374 375 376 377 378 379
static void macvtap_setup(struct net_device *dev)
{
	macvlan_common_setup(dev);
	dev->tx_queue_len = TUN_READQ_SIZE;
}

A
Arnd Bergmann 已提交
380 381
static struct rtnl_link_ops macvtap_link_ops __read_mostly = {
	.kind		= "macvtap",
H
Herbert Xu 已提交
382
	.setup		= macvtap_setup,
A
Arnd Bergmann 已提交
383 384 385 386 387 388 389
	.newlink	= macvtap_newlink,
	.dellink	= macvtap_dellink,
};


static void macvtap_sock_write_space(struct sock *sk)
{
390 391
	wait_queue_head_t *wqueue;

A
Arnd Bergmann 已提交
392 393 394 395
	if (!sock_writeable(sk) ||
	    !test_and_clear_bit(SOCK_ASYNC_NOSPACE, &sk->sk_socket->flags))
		return;

396 397 398
	wqueue = sk_sleep(sk);
	if (wqueue && waitqueue_active(wqueue))
		wake_up_interruptible_poll(wqueue, POLLOUT | POLLWRNORM | POLLWRBAND);
A
Arnd Bergmann 已提交
399 400
}

401 402 403 404 405
static void macvtap_sock_destruct(struct sock *sk)
{
	skb_queue_purge(&sk->sk_receive_queue);
}

A
Arnd Bergmann 已提交
406 407 408
static int macvtap_open(struct inode *inode, struct file *file)
{
	struct net *net = current->nsproxy->net_ns;
409
	struct net_device *dev = dev_get_by_macvtap_minor(iminor(inode));
A
Arnd Bergmann 已提交
410 411 412 413 414 415 416 417 418 419 420 421 422
	struct macvtap_queue *q;
	int err;

	err = -ENODEV;
	if (!dev)
		goto out;

	err = -ENOMEM;
	q = (struct macvtap_queue *)sk_alloc(net, AF_UNSPEC, GFP_KERNEL,
					     &macvtap_proto);
	if (!q)
		goto out;

J
Jason Wang 已提交
423
	RCU_INIT_POINTER(q->sock.wq, &q->wq);
424
	init_waitqueue_head(&q->wq.wait);
A
Arnd Bergmann 已提交
425 426
	q->sock.type = SOCK_RAW;
	q->sock.state = SS_CONNECTED;
A
Arnd Bergmann 已提交
427 428
	q->sock.file = file;
	q->sock.ops = &macvtap_socket_ops;
A
Arnd Bergmann 已提交
429 430
	sock_init_data(&q->sock, &q->sk);
	q->sk.sk_write_space = macvtap_sock_write_space;
431
	q->sk.sk_destruct = macvtap_sock_destruct;
432
	q->flags = IFF_VNET_HDR | IFF_NO_PI | IFF_TAP;
433
	q->vnet_hdr_sz = sizeof(struct virtio_net_hdr);
A
Arnd Bergmann 已提交
434

435 436 437
	/*
	 * so far only KVM virtio_net uses macvtap, enable zero copy between
	 * guest kernel and host kernel when lower device supports zerocopy
438 439 440
	 *
	 * The macvlan supports zerocopy iff the lower device supports zero
	 * copy so we don't have to look at the lower device directly.
441
	 */
442 443
	if ((dev->features & NETIF_F_HIGHDMA) && (dev->features & NETIF_F_SG))
		sock_set_flag(&q->sk, SOCK_ZEROCOPY);
444

A
Arnd Bergmann 已提交
445 446 447 448 449 450 451 452 453 454 455 456 457
	err = macvtap_set_queue(dev, file, q);
	if (err)
		sock_put(&q->sk);

out:
	if (dev)
		dev_put(dev);

	return err;
}

static int macvtap_release(struct inode *inode, struct file *file)
{
458 459
	struct macvtap_queue *q = file->private_data;
	macvtap_put_queue(q);
A
Arnd Bergmann 已提交
460 461 462 463 464
	return 0;
}

static unsigned int macvtap_poll(struct file *file, poll_table * wait)
{
465
	struct macvtap_queue *q = file->private_data;
A
Arnd Bergmann 已提交
466 467 468 469 470 471
	unsigned int mask = POLLERR;

	if (!q)
		goto out;

	mask = 0;
472
	poll_wait(file, &q->wq.wait, wait);
A
Arnd Bergmann 已提交
473 474 475 476 477 478 479 480 481 482 483 484 485

	if (!skb_queue_empty(&q->sk.sk_receive_queue))
		mask |= POLLIN | POLLRDNORM;

	if (sock_writeable(&q->sk) ||
	    (!test_and_set_bit(SOCK_ASYNC_NOSPACE, &q->sock.flags) &&
	     sock_writeable(&q->sk)))
		mask |= POLLOUT | POLLWRNORM;

out:
	return mask;
}

486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508
static inline struct sk_buff *macvtap_alloc_skb(struct sock *sk, size_t prepad,
						size_t len, size_t linear,
						int noblock, int *err)
{
	struct sk_buff *skb;

	/* Under a page?  Don't bother with paged skb. */
	if (prepad + len < PAGE_SIZE || !linear)
		linear = len;

	skb = sock_alloc_send_pskb(sk, prepad + linear, len - linear, noblock,
				   err);
	if (!skb)
		return NULL;

	skb_reserve(skb, prepad);
	skb_put(skb, linear);
	skb->data_len = len - linear;
	skb->len += len - linear;

	return skb;
}

509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533
/* set skb frags from iovec, this can move to core network code for reuse */
static int zerocopy_sg_from_iovec(struct sk_buff *skb, const struct iovec *from,
				  int offset, size_t count)
{
	int len = iov_length(from, count) - offset;
	int copy = skb_headlen(skb);
	int size, offset1 = 0;
	int i = 0;

	/* Skip over from offset */
	while (count && (offset >= from->iov_len)) {
		offset -= from->iov_len;
		++from;
		--count;
	}

	/* copy up to skb headlen */
	while (count && (copy > 0)) {
		size = min_t(unsigned int, copy, from->iov_len - offset);
		if (copy_from_user(skb->data + offset1, from->iov_base + offset,
				   size))
			return -EFAULT;
		if (copy > size) {
			++from;
			--count;
534 535 536
			offset = 0;
		} else
			offset += size;
537 538 539 540 541 542 543 544 545 546 547
		copy -= size;
		offset1 += size;
	}

	if (len == offset1)
		return 0;

	while (count--) {
		struct page *page[MAX_SKB_FRAGS];
		int num_pages;
		unsigned long base;
548
		unsigned long truesize;
549

550
		len = from->iov_len - offset;
551
		if (!len) {
552
			offset = 0;
553 554 555
			++from;
			continue;
		}
556
		base = (unsigned long)from->iov_base + offset;
557
		size = ((base & ~PAGE_MASK) + len + ~PAGE_MASK) >> PAGE_SHIFT;
558 559
		if (i + size > MAX_SKB_FRAGS)
			return -EMSGSIZE;
560
		num_pages = get_user_pages_fast(base, size, 0, &page[i]);
561
		if (num_pages != size) {
562 563
			for (i = 0; i < num_pages; i++)
				put_page(page[i]);
564
			return -EFAULT;
565
		}
566
		truesize = size * PAGE_SIZE;
567 568
		skb->data_len += len;
		skb->len += len;
569 570
		skb->truesize += truesize;
		atomic_add(truesize, &skb->sk->sk_wmem_alloc);
571
		while (len) {
572 573 574
			int off = base & ~PAGE_MASK;
			int size = min_t(int, len, PAGE_SIZE - off);
			__skb_fill_page_desc(skb, i, page[i], off, size);
575 576
			skb_shinfo(skb)->nr_frags++;
			/* increase sk_wmem_alloc */
577 578
			base += size;
			len -= size;
579 580
			i++;
		}
581
		offset = 0;
582 583 584 585 586
		++from;
	}
	return 0;
}

587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624
/*
 * macvtap_skb_from_vnet_hdr and macvtap_skb_to_vnet_hdr should
 * be shared with the tun/tap driver.
 */
static int macvtap_skb_from_vnet_hdr(struct sk_buff *skb,
				     struct virtio_net_hdr *vnet_hdr)
{
	unsigned short gso_type = 0;
	if (vnet_hdr->gso_type != VIRTIO_NET_HDR_GSO_NONE) {
		switch (vnet_hdr->gso_type & ~VIRTIO_NET_HDR_GSO_ECN) {
		case VIRTIO_NET_HDR_GSO_TCPV4:
			gso_type = SKB_GSO_TCPV4;
			break;
		case VIRTIO_NET_HDR_GSO_TCPV6:
			gso_type = SKB_GSO_TCPV6;
			break;
		case VIRTIO_NET_HDR_GSO_UDP:
			gso_type = SKB_GSO_UDP;
			break;
		default:
			return -EINVAL;
		}

		if (vnet_hdr->gso_type & VIRTIO_NET_HDR_GSO_ECN)
			gso_type |= SKB_GSO_TCP_ECN;

		if (vnet_hdr->gso_size == 0)
			return -EINVAL;
	}

	if (vnet_hdr->flags & VIRTIO_NET_HDR_F_NEEDS_CSUM) {
		if (!skb_partial_csum_set(skb, vnet_hdr->csum_start,
					  vnet_hdr->csum_offset))
			return -EINVAL;
	}

	if (vnet_hdr->gso_type != VIRTIO_NET_HDR_GSO_NONE) {
		skb_shinfo(skb)->gso_size = vnet_hdr->gso_size;
625
		skb_shinfo(skb)->gso_type = gso_type;
626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659

		/* Header must be checked, and gso_segs computed. */
		skb_shinfo(skb)->gso_type |= SKB_GSO_DODGY;
		skb_shinfo(skb)->gso_segs = 0;
	}
	return 0;
}

static int macvtap_skb_to_vnet_hdr(const struct sk_buff *skb,
				   struct virtio_net_hdr *vnet_hdr)
{
	memset(vnet_hdr, 0, sizeof(*vnet_hdr));

	if (skb_is_gso(skb)) {
		struct skb_shared_info *sinfo = skb_shinfo(skb);

		/* This is a hint as to how much should be linear. */
		vnet_hdr->hdr_len = skb_headlen(skb);
		vnet_hdr->gso_size = sinfo->gso_size;
		if (sinfo->gso_type & SKB_GSO_TCPV4)
			vnet_hdr->gso_type = VIRTIO_NET_HDR_GSO_TCPV4;
		else if (sinfo->gso_type & SKB_GSO_TCPV6)
			vnet_hdr->gso_type = VIRTIO_NET_HDR_GSO_TCPV6;
		else if (sinfo->gso_type & SKB_GSO_UDP)
			vnet_hdr->gso_type = VIRTIO_NET_HDR_GSO_UDP;
		else
			BUG();
		if (sinfo->gso_type & SKB_GSO_TCP_ECN)
			vnet_hdr->gso_type |= VIRTIO_NET_HDR_GSO_ECN;
	} else
		vnet_hdr->gso_type = VIRTIO_NET_HDR_GSO_NONE;

	if (skb->ip_summed == CHECKSUM_PARTIAL) {
		vnet_hdr->flags = VIRTIO_NET_HDR_F_NEEDS_CSUM;
660
		vnet_hdr->csum_start = skb_checksum_start_offset(skb);
661
		vnet_hdr->csum_offset = skb->csum_offset;
662 663
	} else if (skb->ip_summed == CHECKSUM_UNNECESSARY) {
		vnet_hdr->flags = VIRTIO_NET_HDR_F_DATA_VALID;
664 665 666 667 668 669
	} /* else everything is zero */

	return 0;
}


A
Arnd Bergmann 已提交
670
/* Get packet from user space buffer */
671 672 673
static ssize_t macvtap_get_user(struct macvtap_queue *q, struct msghdr *m,
				const struct iovec *iv, unsigned long total_len,
				size_t count, int noblock)
A
Arnd Bergmann 已提交
674 675
{
	struct sk_buff *skb;
676
	struct macvlan_dev *vlan;
677
	unsigned long len = total_len;
A
Arnd Bergmann 已提交
678
	int err;
679 680
	struct virtio_net_hdr vnet_hdr = { 0 };
	int vnet_hdr_len = 0;
681
	int copylen = 0;
682
	bool zerocopy = false;
683 684

	if (q->flags & IFF_VNET_HDR) {
685
		vnet_hdr_len = q->vnet_hdr_sz;
686 687

		err = -EINVAL;
688
		if (len < vnet_hdr_len)
689
			goto err;
690
		len -= vnet_hdr_len;
691 692

		err = memcpy_fromiovecend((void *)&vnet_hdr, iv, 0,
693
					   sizeof(vnet_hdr));
694 695 696 697 698 699 700 701 702 703 704
		if (err < 0)
			goto err;
		if ((vnet_hdr.flags & VIRTIO_NET_HDR_F_NEEDS_CSUM) &&
		     vnet_hdr.csum_start + vnet_hdr.csum_offset + 2 >
							vnet_hdr.hdr_len)
			vnet_hdr.hdr_len = vnet_hdr.csum_start +
						vnet_hdr.csum_offset + 2;
		err = -EINVAL;
		if (vnet_hdr.hdr_len > len)
			goto err;
	}
A
Arnd Bergmann 已提交
705

706
	err = -EINVAL;
A
Arnd Bergmann 已提交
707
	if (unlikely(len < ETH_HLEN))
708
		goto err;
A
Arnd Bergmann 已提交
709

710 711 712 713
	err = -EMSGSIZE;
	if (unlikely(count > UIO_MAXIOV))
		goto err;

714 715 716 717
	if (m && m->msg_control && sock_flag(&q->sk, SOCK_ZEROCOPY))
		zerocopy = true;

	if (zerocopy) {
718 719 720 721 722 723 724 725 726 727 728
		/* Userspace may produce vectors with count greater than
		 * MAX_SKB_FRAGS, so we need to linearize parts of the skb
		 * to let the rest of data to be fit in the frags.
		 */
		if (count > MAX_SKB_FRAGS) {
			copylen = iov_length(iv, count - MAX_SKB_FRAGS);
			if (copylen < vnet_hdr_len)
				copylen = 0;
			else
				copylen -= vnet_hdr_len;
		}
729 730 731 732
		/* There are 256 bytes to be copied in skb, so there is enough
		 * room for skb expand head in case it is used.
		 * The rest buffer is mapped from userspace.
		 */
733 734
		if (copylen < vnet_hdr.hdr_len)
			copylen = vnet_hdr.hdr_len;
735 736 737 738 739 740 741
		if (!copylen)
			copylen = GOODCOPY_LEN;
	} else
		copylen = len;

	skb = macvtap_alloc_skb(&q->sk, NET_IP_ALIGN, copylen,
				vnet_hdr.hdr_len, noblock, &err);
742 743
	if (!skb)
		goto err;
A
Arnd Bergmann 已提交
744

745
	if (zerocopy)
746
		err = zerocopy_sg_from_iovec(skb, iv, vnet_hdr_len, count);
747
	else
748 749
		err = skb_copy_datagram_from_iovec(skb, 0, iv, vnet_hdr_len,
						   len);
750
	if (err)
751
		goto err_kfree;
A
Arnd Bergmann 已提交
752 753

	skb_set_network_header(skb, ETH_HLEN);
754 755 756 757 758 759 760 761 762
	skb_reset_mac_header(skb);
	skb->protocol = eth_hdr(skb)->h_proto;

	if (vnet_hdr_len) {
		err = macvtap_skb_from_vnet_hdr(skb, &vnet_hdr);
		if (err)
			goto err_kfree;
	}

763
	skb_probe_transport_header(skb, ETH_HLEN);
764

765 766
	rcu_read_lock();
	vlan = rcu_dereference(q->vlan);
767
	/* copy skb_ubuf_info for callback when skb has no error */
768
	if (zerocopy) {
769
		skb_shinfo(skb)->destructor_arg = m->msg_control;
770
		skb_shinfo(skb)->tx_flags |= SKBTX_DEV_ZEROCOPY;
771
		skb_shinfo(skb)->tx_flags |= SKBTX_SHARED_FRAG;
772
	}
773 774 775 776
	if (vlan)
		macvlan_start_xmit(skb, vlan->dev);
	else
		kfree_skb(skb);
777
	rcu_read_unlock();
A
Arnd Bergmann 已提交
778

779
	return total_len;
780

781 782 783
err_kfree:
	kfree_skb(skb);

784
err:
785 786
	rcu_read_lock();
	vlan = rcu_dereference(q->vlan);
787
	if (vlan)
E
Eric Dumazet 已提交
788
		vlan->dev->stats.tx_dropped++;
789
	rcu_read_unlock();
790 791

	return err;
A
Arnd Bergmann 已提交
792 793 794 795 796 797 798
}

static ssize_t macvtap_aio_write(struct kiocb *iocb, const struct iovec *iv,
				 unsigned long count, loff_t pos)
{
	struct file *file = iocb->ki_filp;
	ssize_t result = -ENOLINK;
799
	struct macvtap_queue *q = file->private_data;
A
Arnd Bergmann 已提交
800

801 802
	result = macvtap_get_user(q, NULL, iv, iov_length(iv, count), count,
				  file->f_flags & O_NONBLOCK);
A
Arnd Bergmann 已提交
803 804 805 806 807 808 809 810
	return result;
}

/* Put packet to the user space buffer */
static ssize_t macvtap_put_user(struct macvtap_queue *q,
				const struct sk_buff *skb,
				const struct iovec *iv, int len)
{
811
	struct macvlan_dev *vlan;
A
Arnd Bergmann 已提交
812
	int ret;
813
	int vnet_hdr_len = 0;
814 815
	int vlan_offset = 0;
	int copied;
816 817 818

	if (q->flags & IFF_VNET_HDR) {
		struct virtio_net_hdr vnet_hdr;
819
		vnet_hdr_len = q->vnet_hdr_sz;
820 821 822 823 824 825 826
		if ((len -= vnet_hdr_len) < 0)
			return -EINVAL;

		ret = macvtap_skb_to_vnet_hdr(skb, &vnet_hdr);
		if (ret)
			return ret;

827
		if (memcpy_toiovecend(iv, (void *)&vnet_hdr, 0, sizeof(vnet_hdr)))
828 829
			return -EFAULT;
	}
830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859
	copied = vnet_hdr_len;

	if (!vlan_tx_tag_present(skb))
		len = min_t(int, skb->len, len);
	else {
		int copy;
		struct {
			__be16 h_vlan_proto;
			__be16 h_vlan_TCI;
		} veth;
		veth.h_vlan_proto = htons(ETH_P_8021Q);
		veth.h_vlan_TCI = htons(vlan_tx_tag_get(skb));

		vlan_offset = offsetof(struct vlan_ethhdr, h_vlan_proto);
		len = min_t(int, skb->len + VLAN_HLEN, len);

		copy = min_t(int, vlan_offset, len);
		ret = skb_copy_datagram_const_iovec(skb, 0, iv, copied, copy);
		len -= copy;
		copied += copy;
		if (ret || !len)
			goto done;

		copy = min_t(int, sizeof(veth), len);
		ret = memcpy_toiovecend(iv, (void *)&veth, copied, copy);
		len -= copy;
		copied += copy;
		if (ret || !len)
			goto done;
	}
A
Arnd Bergmann 已提交
860

861 862
	ret = skb_copy_datagram_const_iovec(skb, vlan_offset, iv, copied, len);
	copied += len;
A
Arnd Bergmann 已提交
863

864
done:
865 866
	rcu_read_lock();
	vlan = rcu_dereference(q->vlan);
A
Arnd Bergmann 已提交
867
	if (vlan)
868
		macvlan_count_rx(vlan, copied - vnet_hdr_len, ret == 0, 0);
869
	rcu_read_unlock();
A
Arnd Bergmann 已提交
870

871
	return ret ? ret : copied;
A
Arnd Bergmann 已提交
872 873
}

A
Arnd Bergmann 已提交
874 875 876
static ssize_t macvtap_do_read(struct macvtap_queue *q, struct kiocb *iocb,
			       const struct iovec *iv, unsigned long len,
			       int noblock)
A
Arnd Bergmann 已提交
877
{
878
	DEFINE_WAIT(wait);
A
Arnd Bergmann 已提交
879
	struct sk_buff *skb;
A
Arnd Bergmann 已提交
880
	ssize_t ret = 0;
A
Arnd Bergmann 已提交
881 882

	while (len) {
883 884 885
		if (!noblock)
			prepare_to_wait(sk_sleep(&q->sk), &wait,
					TASK_INTERRUPTIBLE);
A
Arnd Bergmann 已提交
886 887 888 889

		/* Read frames from the queue */
		skb = skb_dequeue(&q->sk.sk_receive_queue);
		if (!skb) {
A
Arnd Bergmann 已提交
890
			if (noblock) {
A
Arnd Bergmann 已提交
891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906
				ret = -EAGAIN;
				break;
			}
			if (signal_pending(current)) {
				ret = -ERESTARTSYS;
				break;
			}
			/* Nothing to read, let's sleep */
			schedule();
			continue;
		}
		ret = macvtap_put_user(q, skb, iv, len);
		kfree_skb(skb);
		break;
	}

907 908
	if (!noblock)
		finish_wait(sk_sleep(&q->sk), &wait);
A
Arnd Bergmann 已提交
909 910 911 912 913 914 915 916 917
	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;
A
Arnd Bergmann 已提交
918

A
Arnd Bergmann 已提交
919 920 921 922 923 924 925 926
	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? */
A
Arnd Bergmann 已提交
927 928 929 930
out:
	return ret;
}

931 932 933 934
static struct macvlan_dev *macvtap_get_vlan(struct macvtap_queue *q)
{
	struct macvlan_dev *vlan;

935 936
	ASSERT_RTNL();
	vlan = rtnl_dereference(q->vlan);
937 938 939 940 941 942 943 944 945 946 947
	if (vlan)
		dev_hold(vlan->dev);

	return vlan;
}

static void macvtap_put_vlan(struct macvlan_dev *vlan)
{
	dev_put(vlan->dev);
}

J
Jason Wang 已提交
948 949 950 951 952 953 954 955 956 957 958 959 960 961
static int macvtap_ioctl_set_queue(struct file *file, unsigned int flags)
{
	struct macvtap_queue *q = file->private_data;
	struct macvlan_dev *vlan;
	int ret;

	vlan = macvtap_get_vlan(q);
	if (!vlan)
		return -EINVAL;

	if (flags & IFF_ATTACH_QUEUE)
		ret = macvtap_enable_queue(vlan->dev, file, q);
	else if (flags & IFF_DETACH_QUEUE)
		ret = macvtap_disable_queue(q);
962 963
	else
		ret = -EINVAL;
J
Jason Wang 已提交
964 965 966 967 968

	macvtap_put_vlan(vlan);
	return ret;
}

969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020
static int set_offload(struct macvtap_queue *q, unsigned long arg)
{
	struct macvlan_dev *vlan;
	netdev_features_t features;
	netdev_features_t feature_mask = 0;

	vlan = rtnl_dereference(q->vlan);
	if (!vlan)
		return -ENOLINK;

	features = vlan->dev->features;

	if (arg & TUN_F_CSUM) {
		feature_mask = NETIF_F_HW_CSUM;

		if (arg & (TUN_F_TSO4 | TUN_F_TSO6)) {
			if (arg & TUN_F_TSO_ECN)
				feature_mask |= NETIF_F_TSO_ECN;
			if (arg & TUN_F_TSO4)
				feature_mask |= NETIF_F_TSO;
			if (arg & TUN_F_TSO6)
				feature_mask |= NETIF_F_TSO6;
		}

		if (arg & TUN_F_UFO)
			feature_mask |= NETIF_F_UFO;
	}

	/* tun/tap driver inverts the usage for TSO offloads, where
	 * setting the TSO bit means that the userspace wants to
	 * accept TSO frames and turning it off means that user space
	 * does not support TSO.
	 * For macvtap, we have to invert it to mean the same thing.
	 * When user space turns off TSO, we turn off GSO/LRO so that
	 * user-space will not receive TSO frames.
	 */
	if (feature_mask & (NETIF_F_TSO | NETIF_F_TSO6 | NETIF_F_UFO))
		features |= RX_OFFLOADS;
	else
		features &= ~RX_OFFLOADS;

	/* tap_features are the same as features on tun/tap and
	 * reflect user expectations.
	 */
	vlan->tap_features = vlan->dev->features &
			    (feature_mask | ~TUN_OFFLOADS);
	vlan->set_features = features;
	netdev_update_features(vlan->dev);

	return 0;
}

A
Arnd Bergmann 已提交
1021 1022 1023 1024 1025 1026
/*
 * provide compatibility with generic tun/tap interface
 */
static long macvtap_ioctl(struct file *file, unsigned int cmd,
			  unsigned long arg)
{
1027 1028
	struct macvtap_queue *q = file->private_data;
	struct macvlan_dev *vlan;
A
Arnd Bergmann 已提交
1029 1030 1031 1032
	void __user *argp = (void __user *)arg;
	struct ifreq __user *ifr = argp;
	unsigned int __user *up = argp;
	unsigned int u;
1033 1034
	int __user *sp = argp;
	int s;
1035
	int ret;
A
Arnd Bergmann 已提交
1036 1037 1038 1039 1040 1041

	switch (cmd) {
	case TUNSETIFF:
		/* ignore the name, just look at flags */
		if (get_user(u, &ifr->ifr_flags))
			return -EFAULT;
1042 1043

		ret = 0;
J
Jason Wang 已提交
1044 1045
		if ((u & ~(IFF_VNET_HDR | IFF_MULTI_QUEUE)) !=
		    (IFF_NO_PI | IFF_TAP))
1046 1047 1048 1049 1050
			ret = -EINVAL;
		else
			q->flags = u;

		return ret;
A
Arnd Bergmann 已提交
1051 1052

	case TUNGETIFF:
1053
		rtnl_lock();
1054
		vlan = macvtap_get_vlan(q);
1055 1056
		if (!vlan) {
			rtnl_unlock();
A
Arnd Bergmann 已提交
1057
			return -ENOLINK;
1058
		}
A
Arnd Bergmann 已提交
1059

1060
		ret = 0;
1061
		if (copy_to_user(&ifr->ifr_name, vlan->dev->name, IFNAMSIZ) ||
1062
		    put_user(q->flags, &ifr->ifr_flags))
1063
			ret = -EFAULT;
1064
		macvtap_put_vlan(vlan);
1065
		rtnl_unlock();
1066
		return ret;
A
Arnd Bergmann 已提交
1067

J
Jason Wang 已提交
1068 1069 1070
	case TUNSETQUEUE:
		if (get_user(u, &ifr->ifr_flags))
			return -EFAULT;
1071 1072 1073
		rtnl_lock();
		ret = macvtap_ioctl_set_queue(file, u);
		rtnl_unlock();
J
Jason Wang 已提交
1074

A
Arnd Bergmann 已提交
1075
	case TUNGETFEATURES:
J
Jason Wang 已提交
1076 1077
		if (put_user(IFF_TAP | IFF_NO_PI | IFF_VNET_HDR |
			     IFF_MULTI_QUEUE, up))
A
Arnd Bergmann 已提交
1078 1079 1080 1081 1082 1083 1084 1085 1086 1087
			return -EFAULT;
		return 0;

	case TUNSETSNDBUF:
		if (get_user(u, up))
			return -EFAULT;

		q->sk.sk_sndbuf = u;
		return 0;

1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102
	case TUNGETVNETHDRSZ:
		s = q->vnet_hdr_sz;
		if (put_user(s, sp))
			return -EFAULT;
		return 0;

	case TUNSETVNETHDRSZ:
		if (get_user(s, sp))
			return -EFAULT;
		if (s < (int)sizeof(struct virtio_net_hdr))
			return -EINVAL;

		q->vnet_hdr_sz = s;
		return 0;

A
Arnd Bergmann 已提交
1103 1104 1105
	case TUNSETOFFLOAD:
		/* let the user check for future flags */
		if (arg & ~(TUN_F_CSUM | TUN_F_TSO4 | TUN_F_TSO6 |
1106
			    TUN_F_TSO_ECN | TUN_F_UFO))
A
Arnd Bergmann 已提交
1107 1108
			return -EINVAL;

1109 1110 1111 1112
		/* TODO: only accept frames with the features that
			 got enabled for forwarded frames */
		if (!(q->flags & IFF_VNET_HDR))
			return  -EINVAL;
1113 1114 1115 1116
		rtnl_lock();
		ret = set_offload(q, arg);
		rtnl_unlock();
		return ret;
A
Arnd Bergmann 已提交
1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144

	default:
		return -EINVAL;
	}
}

#ifdef CONFIG_COMPAT
static long macvtap_compat_ioctl(struct file *file, unsigned int cmd,
				 unsigned long arg)
{
	return macvtap_ioctl(file, cmd, (unsigned long)compat_ptr(arg));
}
#endif

static const struct file_operations macvtap_fops = {
	.owner		= THIS_MODULE,
	.open		= macvtap_open,
	.release	= macvtap_release,
	.aio_read	= macvtap_aio_read,
	.aio_write	= macvtap_aio_write,
	.poll		= macvtap_poll,
	.llseek		= no_llseek,
	.unlocked_ioctl	= macvtap_ioctl,
#ifdef CONFIG_COMPAT
	.compat_ioctl	= macvtap_compat_ioctl,
#endif
};

A
Arnd Bergmann 已提交
1145 1146 1147 1148
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);
1149
	return macvtap_get_user(q, m, m->msg_iov, total_len, m->msg_iovlen,
A
Arnd Bergmann 已提交
1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191
			    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);

1192 1193 1194
static int macvtap_device_event(struct notifier_block *unused,
				unsigned long event, void *ptr)
{
1195
	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
1196
	struct macvlan_dev *vlan;
1197 1198
	struct device *classdev;
	dev_t devt;
1199
	int err;
1200 1201 1202 1203

	if (dev->rtnl_link_ops != &macvtap_link_ops)
		return NOTIFY_DONE;

1204
	vlan = netdev_priv(dev);
1205 1206 1207 1208 1209 1210 1211

	switch (event) {
	case NETDEV_REGISTER:
		/* Create the device node here after the network device has
		 * been registered but before register_netdevice has
		 * finished running.
		 */
1212 1213 1214 1215 1216
		err = macvtap_get_minor(vlan);
		if (err)
			return notifier_from_errno(err);

		devt = MKDEV(MAJOR(macvtap_major), vlan->minor);
1217 1218
		classdev = device_create(macvtap_class, &dev->dev, devt,
					 dev, "tap%d", dev->ifindex);
1219 1220
		if (IS_ERR(classdev)) {
			macvtap_free_minor(vlan);
1221
			return notifier_from_errno(PTR_ERR(classdev));
1222
		}
1223 1224
		break;
	case NETDEV_UNREGISTER:
1225
		devt = MKDEV(MAJOR(macvtap_major), vlan->minor);
1226
		device_destroy(macvtap_class, devt);
1227
		macvtap_free_minor(vlan);
1228 1229 1230 1231 1232 1233 1234 1235 1236 1237
		break;
	}

	return NOTIFY_DONE;
}

static struct notifier_block macvtap_notifier_block __read_mostly = {
	.notifier_call	= macvtap_device_event,
};

A
Arnd Bergmann 已提交
1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257
static int macvtap_init(void)
{
	int err;

	err = alloc_chrdev_region(&macvtap_major, 0,
				MACVTAP_NUM_DEVS, "macvtap");
	if (err)
		goto out1;

	cdev_init(&macvtap_cdev, &macvtap_fops);
	err = cdev_add(&macvtap_cdev, macvtap_major, MACVTAP_NUM_DEVS);
	if (err)
		goto out2;

	macvtap_class = class_create(THIS_MODULE, "macvtap");
	if (IS_ERR(macvtap_class)) {
		err = PTR_ERR(macvtap_class);
		goto out3;
	}

1258
	err = register_netdevice_notifier(&macvtap_notifier_block);
A
Arnd Bergmann 已提交
1259 1260 1261
	if (err)
		goto out4;

1262 1263 1264 1265
	err = macvlan_link_register(&macvtap_link_ops);
	if (err)
		goto out5;

A
Arnd Bergmann 已提交
1266 1267
	return 0;

1268 1269
out5:
	unregister_netdevice_notifier(&macvtap_notifier_block);
A
Arnd Bergmann 已提交
1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283
out4:
	class_unregister(macvtap_class);
out3:
	cdev_del(&macvtap_cdev);
out2:
	unregister_chrdev_region(macvtap_major, MACVTAP_NUM_DEVS);
out1:
	return err;
}
module_init(macvtap_init);

static void macvtap_exit(void)
{
	rtnl_link_unregister(&macvtap_link_ops);
1284
	unregister_netdevice_notifier(&macvtap_notifier_block);
A
Arnd Bergmann 已提交
1285 1286 1287 1288 1289 1290 1291 1292 1293
	class_unregister(macvtap_class);
	cdev_del(&macvtap_cdev);
	unregister_chrdev_region(macvtap_major, MACVTAP_NUM_DEVS);
}
module_exit(macvtap_exit);

MODULE_ALIAS_RTNL_LINK("macvtap");
MODULE_AUTHOR("Arnd Bergmann <arnd@arndb.de>");
MODULE_LICENSE("GPL");