macvtap.c 27.7 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>
24
#include <net/flow_keys.h>
A
Arnd Bergmann 已提交
25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42

/*
 * 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.
 *
 * TODO: multiqueue support is currently not implemented, even though
 * macvtap is basically prepared for that. We will need to add this
 * here as well as in virtio-net and qemu to get line rate on 10gbit
 * adapters from a guest.
 */
struct macvtap_queue {
	struct sock sk;
	struct socket sock;
43
	struct socket_wq wq;
44
	int vnet_hdr_sz;
45
	struct macvlan_dev __rcu *vlan;
A
Arnd Bergmann 已提交
46
	struct file *file;
47
	unsigned int flags;
A
Arnd Bergmann 已提交
48 49 50 51 52 53 54 55 56
};

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

/*
57
 * Variables for dealing with macvtaps device numbers.
A
Arnd Bergmann 已提交
58
 */
59
static dev_t macvtap_major;
60 61 62 63
#define MACVTAP_NUM_DEVS (1U << MINORBITS)
static DEFINE_MUTEX(minor_lock);
static DEFINE_IDR(minor_idr);

64
#define GOODCOPY_LEN 128
A
Arnd Bergmann 已提交
65 66 67
static struct class *macvtap_class;
static struct cdev macvtap_cdev;

A
Arnd Bergmann 已提交
68 69
static const struct proto_ops macvtap_socket_ops;

A
Arnd Bergmann 已提交
70 71
/*
 * RCU usage:
72 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
 * or macvtap_lock is held.
A
Arnd Bergmann 已提交
75
 *
76 77 78 79
 * 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
80
 *
81 82 83 84
 * 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
85 86 87 88
 */
static DEFINE_SPINLOCK(macvtap_lock);

/*
89 90 91
 * get_slot: return a [unused/occupied] slot in vlan->taps[]:
 *	- if 'q' is NULL, return the first empty slot;
 *	- otherwise, return the slot this pointer occupies.
A
Arnd Bergmann 已提交
92
 */
93 94 95 96 97
static int get_slot(struct macvlan_dev *vlan, struct macvtap_queue *q)
{
	int i;

	for (i = 0; i < MAX_MACVTAP_QUEUES; i++) {
98 99
		if (rcu_dereference_protected(vlan->taps[i],
					      lockdep_is_held(&macvtap_lock)) == q)
100 101 102 103 104 105 106
			return i;
	}

	/* Should never happen */
	BUG_ON(1);
}

A
Arnd Bergmann 已提交
107 108 109 110
static int macvtap_set_queue(struct net_device *dev, struct file *file,
				struct macvtap_queue *q)
{
	struct macvlan_dev *vlan = netdev_priv(dev);
111
	int index;
A
Arnd Bergmann 已提交
112 113 114
	int err = -EBUSY;

	spin_lock(&macvtap_lock);
115
	if (vlan->numvtaps == MAX_MACVTAP_QUEUES)
A
Arnd Bergmann 已提交
116 117 118
		goto out;

	err = 0;
119
	index = get_slot(vlan, NULL);
120
	rcu_assign_pointer(q->vlan, vlan);
121
	rcu_assign_pointer(vlan->taps[index], q);
122
	sock_hold(&q->sk);
A
Arnd Bergmann 已提交
123 124

	q->file = file;
125
	file->private_data = q;
A
Arnd Bergmann 已提交
126

127 128
	vlan->numvtaps++;

A
Arnd Bergmann 已提交
129 130 131 132 133 134
out:
	spin_unlock(&macvtap_lock);
	return err;
}

/*
135 136 137
 * 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
138 139 140 141
 *
 * Using the spinlock makes sure that we don't get
 * to the queue again after destroying it.
 */
142
static void macvtap_put_queue(struct macvtap_queue *q)
A
Arnd Bergmann 已提交
143
{
144
	struct macvlan_dev *vlan;
A
Arnd Bergmann 已提交
145 146

	spin_lock(&macvtap_lock);
147 148
	vlan = rcu_dereference_protected(q->vlan,
					 lockdep_is_held(&macvtap_lock));
149
	if (vlan) {
150 151
		int index = get_slot(vlan, q);

152 153
		RCU_INIT_POINTER(vlan->taps[index], NULL);
		RCU_INIT_POINTER(q->vlan, NULL);
154
		sock_put(&q->sk);
155
		--vlan->numvtaps;
A
Arnd Bergmann 已提交
156 157 158 159 160 161 162 163 164
	}

	spin_unlock(&macvtap_lock);

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

/*
165 166 167 168 169
 * 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
170 171 172 173 174
 */
static struct macvtap_queue *macvtap_get_queue(struct net_device *dev,
					       struct sk_buff *skb)
{
	struct macvlan_dev *vlan = netdev_priv(dev);
175 176 177 178 179 180 181
	struct macvtap_queue *tap = NULL;
	int numvtaps = vlan->numvtaps;
	__u32 rxq;

	if (!numvtaps)
		goto out;

182 183 184 185 186 187 188 189
	/* Check if we can use flow to select a queue */
	rxq = skb_get_rxhash(skb);
	if (rxq) {
		tap = rcu_dereference(vlan->taps[rxq % numvtaps]);
		if (tap)
			goto out;
	}

190 191
	if (likely(skb_rx_queue_recorded(skb))) {
		rxq = skb_get_rx_queue(skb);
A
Arnd Bergmann 已提交
192

193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209
		while (unlikely(rxq >= numvtaps))
			rxq -= numvtaps;

		tap = rcu_dereference(vlan->taps[rxq]);
		if (tap)
			goto out;
	}

	/* Everything failed - find first available queue */
	for (rxq = 0; rxq < MAX_MACVTAP_QUEUES; rxq++) {
		tap = rcu_dereference(vlan->taps[rxq]);
		if (tap)
			break;
	}

out:
	return tap;
A
Arnd Bergmann 已提交
210 211
}

212 213
/*
 * The net_device is going away, give up the reference
214 215
 * that it holds on all queues and safely set the pointer
 * from the queues to NULL.
216
 */
A
Arnd Bergmann 已提交
217 218 219
static void macvtap_del_queues(struct net_device *dev)
{
	struct macvlan_dev *vlan = netdev_priv(dev);
220 221
	struct macvtap_queue *q, *qlist[MAX_MACVTAP_QUEUES];
	int i, j = 0;
222

223
	/* macvtap_put_queue can free some slots, so go through all slots */
224
	spin_lock(&macvtap_lock);
225
	for (i = 0; i < MAX_MACVTAP_QUEUES && vlan->numvtaps; i++) {
226 227
		q = rcu_dereference_protected(vlan->taps[i],
					      lockdep_is_held(&macvtap_lock));
228 229
		if (q) {
			qlist[j++] = q;
230 231
			RCU_INIT_POINTER(vlan->taps[i], NULL);
			RCU_INIT_POINTER(q->vlan, NULL);
232 233
			vlan->numvtaps--;
		}
234
	}
235
	BUG_ON(vlan->numvtaps != 0);
236 237
	/* guarantee that any future macvtap_set_queue will fail */
	vlan->numvtaps = MAX_MACVTAP_QUEUES;
238 239 240
	spin_unlock(&macvtap_lock);

	synchronize_rcu();
241 242 243

	for (--j; j >= 0; j--)
		sock_put(&qlist[j]->sk);
A
Arnd Bergmann 已提交
244 245 246 247 248 249 250 251 252 253 254
}

/*
 * 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
255 256 257 258
		goto drop;

	if (skb_queue_len(&q->sk.sk_receive_queue) >= dev->tx_queue_len)
		goto drop;
A
Arnd Bergmann 已提交
259 260

	skb_queue_tail(&q->sk.sk_receive_queue, skb);
E
Eric Dumazet 已提交
261
	wake_up_interruptible_poll(sk_sleep(&q->sk), POLLIN | POLLRDNORM | POLLRDBAND);
H
Herbert Xu 已提交
262 263 264 265 266
	return NET_RX_SUCCESS;

drop:
	kfree_skb(skb);
	return NET_RX_DROP;
A
Arnd Bergmann 已提交
267 268 269 270 271 272 273 274 275 276 277 278 279
}

/*
 * 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);
}

280 281 282 283 284
static int macvtap_get_minor(struct macvlan_dev *vlan)
{
	int retval = -ENOMEM;

	mutex_lock(&minor_lock);
T
Tejun Heo 已提交
285 286 287 288
	retval = idr_alloc(&minor_idr, vlan, 1, MACVTAP_NUM_DEVS, GFP_KERNEL);
	if (retval >= 0) {
		vlan->minor = retval;
	} else if (retval == -ENOSPC) {
289 290 291 292
		printk(KERN_ERR "too many macvtap devices\n");
		retval = -EINVAL;
	}
	mutex_unlock(&minor_lock);
T
Tejun Heo 已提交
293
	return retval < 0 ? retval : 0;
294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320
}

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
321 322 323 324 325
static int macvtap_newlink(struct net *src_net,
			   struct net_device *dev,
			   struct nlattr *tb[],
			   struct nlattr *data[])
{
326 327 328 329 330
	/* 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
331 332 333 334 335 336 337 338 339
}

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
340 341 342 343 344 345
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
346 347
static struct rtnl_link_ops macvtap_link_ops __read_mostly = {
	.kind		= "macvtap",
H
Herbert Xu 已提交
348
	.setup		= macvtap_setup,
A
Arnd Bergmann 已提交
349 350 351 352 353 354 355
	.newlink	= macvtap_newlink,
	.dellink	= macvtap_dellink,
};


static void macvtap_sock_write_space(struct sock *sk)
{
356 357
	wait_queue_head_t *wqueue;

A
Arnd Bergmann 已提交
358 359 360 361
	if (!sock_writeable(sk) ||
	    !test_and_clear_bit(SOCK_ASYNC_NOSPACE, &sk->sk_socket->flags))
		return;

362 363 364
	wqueue = sk_sleep(sk);
	if (wqueue && waitqueue_active(wqueue))
		wake_up_interruptible_poll(wqueue, POLLOUT | POLLWRNORM | POLLWRBAND);
A
Arnd Bergmann 已提交
365 366
}

367 368 369 370 371
static void macvtap_sock_destruct(struct sock *sk)
{
	skb_queue_purge(&sk->sk_receive_queue);
}

A
Arnd Bergmann 已提交
372 373 374
static int macvtap_open(struct inode *inode, struct file *file)
{
	struct net *net = current->nsproxy->net_ns;
375
	struct net_device *dev = dev_get_by_macvtap_minor(iminor(inode));
A
Arnd Bergmann 已提交
376 377 378 379 380 381 382 383 384 385 386 387 388
	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;

389 390
	q->sock.wq = &q->wq;
	init_waitqueue_head(&q->wq.wait);
A
Arnd Bergmann 已提交
391 392
	q->sock.type = SOCK_RAW;
	q->sock.state = SS_CONNECTED;
A
Arnd Bergmann 已提交
393 394
	q->sock.file = file;
	q->sock.ops = &macvtap_socket_ops;
A
Arnd Bergmann 已提交
395 396
	sock_init_data(&q->sock, &q->sk);
	q->sk.sk_write_space = macvtap_sock_write_space;
397
	q->sk.sk_destruct = macvtap_sock_destruct;
398
	q->flags = IFF_VNET_HDR | IFF_NO_PI | IFF_TAP;
399
	q->vnet_hdr_sz = sizeof(struct virtio_net_hdr);
A
Arnd Bergmann 已提交
400

401 402 403
	/*
	 * so far only KVM virtio_net uses macvtap, enable zero copy between
	 * guest kernel and host kernel when lower device supports zerocopy
404 405 406
	 *
	 * The macvlan supports zerocopy iff the lower device supports zero
	 * copy so we don't have to look at the lower device directly.
407
	 */
408 409
	if ((dev->features & NETIF_F_HIGHDMA) && (dev->features & NETIF_F_SG))
		sock_set_flag(&q->sk, SOCK_ZEROCOPY);
410

A
Arnd Bergmann 已提交
411 412 413 414 415 416 417 418 419 420 421 422 423
	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)
{
424 425
	struct macvtap_queue *q = file->private_data;
	macvtap_put_queue(q);
A
Arnd Bergmann 已提交
426 427 428 429 430
	return 0;
}

static unsigned int macvtap_poll(struct file *file, poll_table * wait)
{
431
	struct macvtap_queue *q = file->private_data;
A
Arnd Bergmann 已提交
432 433 434 435 436 437
	unsigned int mask = POLLERR;

	if (!q)
		goto out;

	mask = 0;
438
	poll_wait(file, &q->wq.wait, wait);
A
Arnd Bergmann 已提交
439 440 441 442 443 444 445 446 447 448 449 450 451

	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;
}

452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474
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;
}

475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499
/* 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;
500 501 502
			offset = 0;
		} else
			offset += size;
503 504 505 506 507 508 509 510 511 512 513
		copy -= size;
		offset1 += size;
	}

	if (len == offset1)
		return 0;

	while (count--) {
		struct page *page[MAX_SKB_FRAGS];
		int num_pages;
		unsigned long base;
514
		unsigned long truesize;
515

516
		len = from->iov_len - offset;
517
		if (!len) {
518
			offset = 0;
519 520 521
			++from;
			continue;
		}
522
		base = (unsigned long)from->iov_base + offset;
523
		size = ((base & ~PAGE_MASK) + len + ~PAGE_MASK) >> PAGE_SHIFT;
524 525
		if (i + size > MAX_SKB_FRAGS)
			return -EMSGSIZE;
526
		num_pages = get_user_pages_fast(base, size, 0, &page[i]);
527
		if (num_pages != size) {
528 529
			for (i = 0; i < num_pages; i++)
				put_page(page[i]);
530
			return -EFAULT;
531
		}
532
		truesize = size * PAGE_SIZE;
533 534
		skb->data_len += len;
		skb->len += len;
535 536
		skb->truesize += truesize;
		atomic_add(truesize, &skb->sk->sk_wmem_alloc);
537
		while (len) {
538 539 540
			int off = base & ~PAGE_MASK;
			int size = min_t(int, len, PAGE_SIZE - off);
			__skb_fill_page_desc(skb, i, page[i], off, size);
541 542
			skb_shinfo(skb)->nr_frags++;
			/* increase sk_wmem_alloc */
543 544
			base += size;
			len -= size;
545 546
			i++;
		}
547
		offset = 0;
548 549 550 551 552
		++from;
	}
	return 0;
}

553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590
/*
 * 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;
591
		skb_shinfo(skb)->gso_type = gso_type;
592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625

		/* 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;
626
		vnet_hdr->csum_start = skb_checksum_start_offset(skb);
627
		vnet_hdr->csum_offset = skb->csum_offset;
628 629
	} else if (skb->ip_summed == CHECKSUM_UNNECESSARY) {
		vnet_hdr->flags = VIRTIO_NET_HDR_F_DATA_VALID;
630 631 632 633 634 635
	} /* else everything is zero */

	return 0;
}


A
Arnd Bergmann 已提交
636
/* Get packet from user space buffer */
637 638 639
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
640 641
{
	struct sk_buff *skb;
642
	struct macvlan_dev *vlan;
643
	unsigned long len = total_len;
A
Arnd Bergmann 已提交
644
	int err;
645 646
	struct virtio_net_hdr vnet_hdr = { 0 };
	int vnet_hdr_len = 0;
647
	int copylen = 0;
648
	bool zerocopy = false;
649
	struct flow_keys keys;
650 651

	if (q->flags & IFF_VNET_HDR) {
652
		vnet_hdr_len = q->vnet_hdr_sz;
653 654

		err = -EINVAL;
655
		if (len < vnet_hdr_len)
656
			goto err;
657
		len -= vnet_hdr_len;
658 659

		err = memcpy_fromiovecend((void *)&vnet_hdr, iv, 0,
660
					   sizeof(vnet_hdr));
661 662 663 664 665 666 667 668 669 670 671
		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
672

673
	err = -EINVAL;
A
Arnd Bergmann 已提交
674
	if (unlikely(len < ETH_HLEN))
675
		goto err;
A
Arnd Bergmann 已提交
676

677 678 679 680
	err = -EMSGSIZE;
	if (unlikely(count > UIO_MAXIOV))
		goto err;

681 682 683 684
	if (m && m->msg_control && sock_flag(&q->sk, SOCK_ZEROCOPY))
		zerocopy = true;

	if (zerocopy) {
685 686 687 688 689 690 691 692 693 694 695
		/* 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;
		}
696 697 698 699
		/* 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.
		 */
700 701
		if (copylen < vnet_hdr.hdr_len)
			copylen = vnet_hdr.hdr_len;
702 703 704 705 706 707 708
		if (!copylen)
			copylen = GOODCOPY_LEN;
	} else
		copylen = len;

	skb = macvtap_alloc_skb(&q->sk, NET_IP_ALIGN, copylen,
				vnet_hdr.hdr_len, noblock, &err);
709 710
	if (!skb)
		goto err;
A
Arnd Bergmann 已提交
711

712
	if (zerocopy)
713
		err = zerocopy_sg_from_iovec(skb, iv, vnet_hdr_len, count);
714
	else
715 716
		err = skb_copy_datagram_from_iovec(skb, 0, iv, vnet_hdr_len,
						   len);
717
	if (err)
718
		goto err_kfree;
A
Arnd Bergmann 已提交
719 720

	skb_set_network_header(skb, ETH_HLEN);
721 722 723 724 725 726 727 728 729
	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;
	}

730 731 732 733 734 735 736
	if (skb->ip_summed == CHECKSUM_PARTIAL)
		skb_set_transport_header(skb, skb_checksum_start_offset(skb));
	else if (skb_flow_dissect(skb, &keys))
		skb_set_transport_header(skb, keys.thoff);
	else
		skb_set_transport_header(skb, ETH_HLEN);

737
	rcu_read_lock_bh();
738
	vlan = rcu_dereference_bh(q->vlan);
739
	/* copy skb_ubuf_info for callback when skb has no error */
740
	if (zerocopy) {
741
		skb_shinfo(skb)->destructor_arg = m->msg_control;
742
		skb_shinfo(skb)->tx_flags |= SKBTX_DEV_ZEROCOPY;
743
		skb_shinfo(skb)->tx_flags |= SKBTX_SHARED_FRAG;
744
	}
745 746 747 748 749
	if (vlan)
		macvlan_start_xmit(skb, vlan->dev);
	else
		kfree_skb(skb);
	rcu_read_unlock_bh();
A
Arnd Bergmann 已提交
750

751
	return total_len;
752

753 754 755
err_kfree:
	kfree_skb(skb);

756 757
err:
	rcu_read_lock_bh();
758
	vlan = rcu_dereference_bh(q->vlan);
759
	if (vlan)
E
Eric Dumazet 已提交
760
		vlan->dev->stats.tx_dropped++;
761 762 763
	rcu_read_unlock_bh();

	return err;
A
Arnd Bergmann 已提交
764 765 766 767 768 769 770
}

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;
771
	struct macvtap_queue *q = file->private_data;
A
Arnd Bergmann 已提交
772

773 774
	result = macvtap_get_user(q, NULL, iv, iov_length(iv, count), count,
				  file->f_flags & O_NONBLOCK);
A
Arnd Bergmann 已提交
775 776 777 778 779 780 781 782
	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)
{
783
	struct macvlan_dev *vlan;
A
Arnd Bergmann 已提交
784
	int ret;
785
	int vnet_hdr_len = 0;
786 787
	int vlan_offset = 0;
	int copied;
788 789 790

	if (q->flags & IFF_VNET_HDR) {
		struct virtio_net_hdr vnet_hdr;
791
		vnet_hdr_len = q->vnet_hdr_sz;
792 793 794 795 796 797 798
		if ((len -= vnet_hdr_len) < 0)
			return -EINVAL;

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

799
		if (memcpy_toiovecend(iv, (void *)&vnet_hdr, 0, sizeof(vnet_hdr)))
800 801
			return -EFAULT;
	}
802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831
	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
832

833 834
	ret = skb_copy_datagram_const_iovec(skb, vlan_offset, iv, copied, len);
	copied += len;
A
Arnd Bergmann 已提交
835

836
done:
837
	rcu_read_lock_bh();
838
	vlan = rcu_dereference_bh(q->vlan);
A
Arnd Bergmann 已提交
839
	if (vlan)
840
		macvlan_count_rx(vlan, copied - vnet_hdr_len, ret == 0, 0);
841
	rcu_read_unlock_bh();
A
Arnd Bergmann 已提交
842

843
	return ret ? ret : copied;
A
Arnd Bergmann 已提交
844 845
}

A
Arnd Bergmann 已提交
846 847 848
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
849
{
850
	DEFINE_WAIT(wait);
A
Arnd Bergmann 已提交
851
	struct sk_buff *skb;
A
Arnd Bergmann 已提交
852
	ssize_t ret = 0;
A
Arnd Bergmann 已提交
853 854

	while (len) {
855
		prepare_to_wait(sk_sleep(&q->sk), &wait, TASK_INTERRUPTIBLE);
A
Arnd Bergmann 已提交
856 857 858 859

		/* Read frames from the queue */
		skb = skb_dequeue(&q->sk.sk_receive_queue);
		if (!skb) {
A
Arnd Bergmann 已提交
860
			if (noblock) {
A
Arnd Bergmann 已提交
861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876
				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;
	}

877
	finish_wait(sk_sleep(&q->sk), &wait);
A
Arnd Bergmann 已提交
878 879 880 881 882 883 884 885 886
	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
887

A
Arnd Bergmann 已提交
888 889 890 891 892 893 894 895
	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
896 897 898 899 900 901 902 903 904 905
out:
	return ret;
}

/*
 * provide compatibility with generic tun/tap interface
 */
static long macvtap_ioctl(struct file *file, unsigned int cmd,
			  unsigned long arg)
{
906 907
	struct macvtap_queue *q = file->private_data;
	struct macvlan_dev *vlan;
A
Arnd Bergmann 已提交
908 909 910 911
	void __user *argp = (void __user *)arg;
	struct ifreq __user *ifr = argp;
	unsigned int __user *up = argp;
	unsigned int u;
912 913
	int __user *sp = argp;
	int s;
914
	int ret;
A
Arnd Bergmann 已提交
915 916 917 918 919 920

	switch (cmd) {
	case TUNSETIFF:
		/* ignore the name, just look at flags */
		if (get_user(u, &ifr->ifr_flags))
			return -EFAULT;
921 922 923 924 925 926 927 928

		ret = 0;
		if ((u & ~IFF_VNET_HDR) != (IFF_NO_PI | IFF_TAP))
			ret = -EINVAL;
		else
			q->flags = u;

		return ret;
A
Arnd Bergmann 已提交
929 930

	case TUNGETIFF:
931
		rcu_read_lock_bh();
932
		vlan = rcu_dereference_bh(q->vlan);
933 934 935 936 937
		if (vlan)
			dev_hold(vlan->dev);
		rcu_read_unlock_bh();

		if (!vlan)
A
Arnd Bergmann 已提交
938 939
			return -ENOLINK;

940
		ret = 0;
941
		if (copy_to_user(&ifr->ifr_name, vlan->dev->name, IFNAMSIZ) ||
942
		    put_user(q->flags, &ifr->ifr_flags))
943 944 945
			ret = -EFAULT;
		dev_put(vlan->dev);
		return ret;
A
Arnd Bergmann 已提交
946 947

	case TUNGETFEATURES:
948
		if (put_user(IFF_TAP | IFF_NO_PI | IFF_VNET_HDR, up))
A
Arnd Bergmann 已提交
949 950 951 952 953 954 955 956 957 958
			return -EFAULT;
		return 0;

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

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

959 960 961 962 963 964 965 966 967 968 969 970 971 972 973
	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
974 975 976
	case TUNSETOFFLOAD:
		/* let the user check for future flags */
		if (arg & ~(TUN_F_CSUM | TUN_F_TSO4 | TUN_F_TSO6 |
977
			    TUN_F_TSO_ECN | TUN_F_UFO))
A
Arnd Bergmann 已提交
978 979
			return -EINVAL;

980 981 982 983
		/* TODO: only accept frames with the features that
			 got enabled for forwarded frames */
		if (!(q->flags & IFF_VNET_HDR))
			return  -EINVAL;
A
Arnd Bergmann 已提交
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
		return 0;

	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
1013 1014 1015 1016
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);
1017
	return macvtap_get_user(q, m, m->msg_iov, total_len, m->msg_iovlen,
A
Arnd Bergmann 已提交
1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059
			    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);

1060 1061 1062 1063
static int macvtap_device_event(struct notifier_block *unused,
				unsigned long event, void *ptr)
{
	struct net_device *dev = ptr;
1064
	struct macvlan_dev *vlan;
1065 1066
	struct device *classdev;
	dev_t devt;
1067
	int err;
1068 1069 1070 1071

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

1072
	vlan = netdev_priv(dev);
1073 1074 1075 1076 1077 1078 1079

	switch (event) {
	case NETDEV_REGISTER:
		/* Create the device node here after the network device has
		 * been registered but before register_netdevice has
		 * finished running.
		 */
1080 1081 1082 1083 1084
		err = macvtap_get_minor(vlan);
		if (err)
			return notifier_from_errno(err);

		devt = MKDEV(MAJOR(macvtap_major), vlan->minor);
1085 1086
		classdev = device_create(macvtap_class, &dev->dev, devt,
					 dev, "tap%d", dev->ifindex);
1087 1088
		if (IS_ERR(classdev)) {
			macvtap_free_minor(vlan);
1089
			return notifier_from_errno(PTR_ERR(classdev));
1090
		}
1091 1092
		break;
	case NETDEV_UNREGISTER:
1093
		devt = MKDEV(MAJOR(macvtap_major), vlan->minor);
1094
		device_destroy(macvtap_class, devt);
1095
		macvtap_free_minor(vlan);
1096 1097 1098 1099 1100 1101 1102 1103 1104 1105
		break;
	}

	return NOTIFY_DONE;
}

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

A
Arnd Bergmann 已提交
1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125
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;
	}

1126
	err = register_netdevice_notifier(&macvtap_notifier_block);
A
Arnd Bergmann 已提交
1127 1128 1129
	if (err)
		goto out4;

1130 1131 1132 1133
	err = macvlan_link_register(&macvtap_link_ops);
	if (err)
		goto out5;

A
Arnd Bergmann 已提交
1134 1135
	return 0;

1136 1137
out5:
	unregister_netdevice_notifier(&macvtap_notifier_block);
A
Arnd Bergmann 已提交
1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151
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);
1152
	unregister_netdevice_notifier(&macvtap_notifier_block);
A
Arnd Bergmann 已提交
1153 1154 1155 1156 1157 1158 1159 1160 1161
	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");