tun.c 53.7 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
/*
 *  TUN - Universal TUN/TAP device driver.
 *  Copyright (C) 1999-2002 Maxim Krasnyansky <maxk@qualcomm.com>
 *
 *  This program is free software; you can redistribute it and/or modify
 *  it under the terms of the GNU General Public License as published by
 *  the Free Software Foundation; either version 2 of the License, or
 *  (at your option) any later version.
 *
 *  This program is distributed in the hope that it will be useful,
 *  but WITHOUT ANY WARRANTY; without even the implied warranty of
 *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
 *  GNU General Public License for more details.
 *
 *  $Id: tun.c,v 1.15 2002/03/01 02:44:24 maxk Exp $
 */

/*
 *  Changes:
 *
21 22 23
 *  Mike Kershaw <dragorn@kismetwireless.net> 2005/08/14
 *    Add TUNSETLINK ioctl to set the link encapsulation
 *
L
Linus Torvalds 已提交
24
 *  Mark Smith <markzzzsmith@yahoo.com.au>
J
Joe Perches 已提交
25
 *    Use eth_random_addr() for tap MAC address.
L
Linus Torvalds 已提交
26 27 28 29 30 31 32 33 34 35 36
 *
 *  Harald Roelle <harald.roelle@ifi.lmu.de>  2004/04/20
 *    Fixes in packet dropping, queue length setting and queue wakeup.
 *    Increased default tx queue length.
 *    Added ethtool API.
 *    Minor cleanups
 *
 *  Daniel Podlejski <underley@underley.eu.org>
 *    Modifications for 2.3.99-pre5 kernel.
 */

37 38
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt

L
Linus Torvalds 已提交
39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57
#define DRV_NAME	"tun"
#define DRV_VERSION	"1.6"
#define DRV_DESCRIPTION	"Universal TUN/TAP device driver"
#define DRV_COPYRIGHT	"(C) 1999-2004 Max Krasnyansky <maxk@qualcomm.com>"

#include <linux/module.h>
#include <linux/errno.h>
#include <linux/kernel.h>
#include <linux/major.h>
#include <linux/slab.h>
#include <linux/poll.h>
#include <linux/fcntl.h>
#include <linux/init.h>
#include <linux/skbuff.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/miscdevice.h>
#include <linux/ethtool.h>
#include <linux/rtnetlink.h>
58
#include <linux/compat.h>
L
Linus Torvalds 已提交
59 60 61 62
#include <linux/if.h>
#include <linux/if_arp.h>
#include <linux/if_ether.h>
#include <linux/if_tun.h>
J
Jason Wang 已提交
63
#include <linux/if_vlan.h>
L
Linus Torvalds 已提交
64
#include <linux/crc32.h>
65
#include <linux/nsproxy.h>
66
#include <linux/virtio_net.h>
M
Michael S. Tsirkin 已提交
67
#include <linux/rcupdate.h>
68
#include <net/net_namespace.h>
69
#include <net/netns/generic.h>
70
#include <net/rtnetlink.h>
71
#include <net/sock.h>
L
Linus Torvalds 已提交
72 73 74

#include <asm/uaccess.h>

75 76 77
/* Uncomment to enable debugging */
/* #define TUN_DEBUG 1 */

L
Linus Torvalds 已提交
78 79
#ifdef TUN_DEBUG
static int debug;
80

81 82 83 84 85 86 87 88 89 90
#define tun_debug(level, tun, fmt, args...)			\
do {								\
	if (tun->debug)						\
		netdev_printk(level, tun->dev, fmt, ##args);	\
} while (0)
#define DBG1(level, fmt, args...)				\
do {								\
	if (debug == 2)						\
		printk(level fmt, ##args);			\
} while (0)
91
#else
92 93 94 95 96 97 98 99 100 101
#define tun_debug(level, tun, fmt, args...)			\
do {								\
	if (0)							\
		netdev_printk(level, tun->dev, fmt, ##args);	\
} while (0)
#define DBG1(level, fmt, args...)				\
do {								\
	if (0)							\
		printk(level fmt, ##args);			\
} while (0)
102 103
#endif

104 105
#define GOODCOPY_LEN 128

106 107 108 109 110 111 112
#define FLT_EXACT_COUNT 8
struct tap_filter {
	unsigned int    count;    /* Number of addrs. Zero means disabled */
	u32             mask[2];  /* Mask of the hashed addrs */
	unsigned char	addr[FLT_EXACT_COUNT][ETH_ALEN];
};

113 114 115 116
/* DEFAULT_MAX_NUM_RSS_QUEUES were choosed to let the rx/tx queues allocated for
 * the netdevice to be fit in one page. So we can make sure the success of
 * memory allocation. TODO: increase the limit. */
#define MAX_TAP_QUEUES DEFAULT_MAX_NUM_RSS_QUEUES
117
#define MAX_TAP_FLOWS  4096
J
Jason Wang 已提交
118

J
Jason Wang 已提交
119 120
#define TUN_FLOW_EXPIRE (3 * HZ)

J
Jason Wang 已提交
121 122 123 124
/* A tun_file connects an open character device to a tuntap netdevice. It
 * also contains all socket related strctures (except sock_fprog and tap_filter)
 * to serve as one transmit queue for tuntap device. The sock_fprog and
 * tap_filter were kept in tun_struct since they were used for filtering for the
R
Rami Rosen 已提交
125
 * netdevice not for a specific queue (at least I didn't see the requirement for
J
Jason Wang 已提交
126
 * this).
127 128
 *
 * RCU usage:
R
Rami Rosen 已提交
129
 * The tun_file and tun_struct are loosely coupled, the pointer from one to the
130
 * other can only be read while rcu_read_lock or rtnl_lock is held.
J
Jason Wang 已提交
131
 */
E
Eric W. Biederman 已提交
132
struct tun_file {
J
Jason Wang 已提交
133 134 135
	struct sock sk;
	struct socket socket;
	struct socket_wq wq;
136
	struct tun_struct __rcu *tun;
E
Eric W. Biederman 已提交
137
	struct net *net;
J
Jason Wang 已提交
138 139 140
	struct fasync_struct *fasync;
	/* only used for fasnyc */
	unsigned int flags;
141 142 143 144
	union {
		u16 queue_index;
		unsigned int ifindex;
	};
J
Jason Wang 已提交
145 146
	struct list_head next;
	struct tun_struct *detached;
E
Eric W. Biederman 已提交
147 148
};

J
Jason Wang 已提交
149 150 151 152 153 154 155 156 157 158 159 160
struct tun_flow_entry {
	struct hlist_node hash_link;
	struct rcu_head rcu;
	struct tun_struct *tun;

	u32 rxhash;
	int queue_index;
	unsigned long updated;
};

#define TUN_NUM_FLOW_ENTRIES 1024

J
Jason Wang 已提交
161
/* Since the socket were moved to tun_file, to preserve the behavior of persist
R
Rami Rosen 已提交
162
 * device, socket filter, sndbuf and vnet header size were restore when the
J
Jason Wang 已提交
163 164
 * file were attached to a persist device.
 */
165
struct tun_struct {
J
Jason Wang 已提交
166 167
	struct tun_file __rcu	*tfiles[MAX_TAP_QUEUES];
	unsigned int            numqueues;
168
	unsigned int 		flags;
169 170
	kuid_t			owner;
	kgid_t			group;
171 172

	struct net_device	*dev;
173
	netdev_features_t	set_features;
174 175
#define TUN_USER_FEATURES (NETIF_F_HW_CSUM|NETIF_F_TSO_ECN|NETIF_F_TSO| \
			  NETIF_F_TSO6|NETIF_F_UFO)
176 177

	int			vnet_hdr_sz;
J
Jason Wang 已提交
178 179 180 181 182
	int			sndbuf;
	struct tap_filter	txflt;
	struct sock_fprog	fprog;
	/* protected by rtnl lock */
	bool			filter_attached;
183 184
#ifdef TUN_DEBUG
	int debug;
L
Linus Torvalds 已提交
185
#endif
J
Jason Wang 已提交
186 187 188 189
	spinlock_t lock;
	struct hlist_head flows[TUN_NUM_FLOW_ENTRIES];
	struct timer_list flow_gc_timer;
	unsigned long ageing_time;
J
Jason Wang 已提交
190 191
	unsigned int numdisabled;
	struct list_head disabled;
192
	void *security;
193
	u32 flow_count;
194
};
L
Linus Torvalds 已提交
195

J
Jason Wang 已提交
196 197 198 199 200 201 202 203 204
static inline u32 tun_hashfn(u32 rxhash)
{
	return rxhash & 0x3ff;
}

static struct tun_flow_entry *tun_flow_find(struct hlist_head *head, u32 rxhash)
{
	struct tun_flow_entry *e;

205
	hlist_for_each_entry_rcu(e, head, hash_link) {
J
Jason Wang 已提交
206 207 208 209 210 211 212 213 214 215
		if (e->rxhash == rxhash)
			return e;
	}
	return NULL;
}

static struct tun_flow_entry *tun_flow_create(struct tun_struct *tun,
					      struct hlist_head *head,
					      u32 rxhash, u16 queue_index)
{
216 217
	struct tun_flow_entry *e = kmalloc(sizeof(*e), GFP_ATOMIC);

J
Jason Wang 已提交
218 219 220 221 222 223 224 225
	if (e) {
		tun_debug(KERN_INFO, tun, "create flow: hash %u index %u\n",
			  rxhash, queue_index);
		e->updated = jiffies;
		e->rxhash = rxhash;
		e->queue_index = queue_index;
		e->tun = tun;
		hlist_add_head_rcu(&e->hash_link, head);
226
		++tun->flow_count;
J
Jason Wang 已提交
227 228 229 230 231 232 233 234 235
	}
	return e;
}

static void tun_flow_delete(struct tun_struct *tun, struct tun_flow_entry *e)
{
	tun_debug(KERN_INFO, tun, "delete flow: hash %u index %u\n",
		  e->rxhash, e->queue_index);
	hlist_del_rcu(&e->hash_link);
236
	kfree_rcu(e, rcu);
237
	--tun->flow_count;
J
Jason Wang 已提交
238 239 240 241 242 243 244 245 246
}

static void tun_flow_flush(struct tun_struct *tun)
{
	int i;

	spin_lock_bh(&tun->lock);
	for (i = 0; i < TUN_NUM_FLOW_ENTRIES; i++) {
		struct tun_flow_entry *e;
247
		struct hlist_node *n;
J
Jason Wang 已提交
248

249
		hlist_for_each_entry_safe(e, n, &tun->flows[i], hash_link)
J
Jason Wang 已提交
250 251 252 253 254 255 256 257 258 259 260 261
			tun_flow_delete(tun, e);
	}
	spin_unlock_bh(&tun->lock);
}

static void tun_flow_delete_by_queue(struct tun_struct *tun, u16 queue_index)
{
	int i;

	spin_lock_bh(&tun->lock);
	for (i = 0; i < TUN_NUM_FLOW_ENTRIES; i++) {
		struct tun_flow_entry *e;
262
		struct hlist_node *n;
J
Jason Wang 已提交
263

264
		hlist_for_each_entry_safe(e, n, &tun->flows[i], hash_link) {
J
Jason Wang 已提交
265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284
			if (e->queue_index == queue_index)
				tun_flow_delete(tun, e);
		}
	}
	spin_unlock_bh(&tun->lock);
}

static void tun_flow_cleanup(unsigned long data)
{
	struct tun_struct *tun = (struct tun_struct *)data;
	unsigned long delay = tun->ageing_time;
	unsigned long next_timer = jiffies + delay;
	unsigned long count = 0;
	int i;

	tun_debug(KERN_INFO, tun, "tun_flow_cleanup\n");

	spin_lock_bh(&tun->lock);
	for (i = 0; i < TUN_NUM_FLOW_ENTRIES; i++) {
		struct tun_flow_entry *e;
285
		struct hlist_node *n;
J
Jason Wang 已提交
286

287
		hlist_for_each_entry_safe(e, n, &tun->flows[i], hash_link) {
J
Jason Wang 已提交
288 289 290 291 292 293 294 295 296 297 298 299 300 301 302
			unsigned long this_timer;
			count++;
			this_timer = e->updated + delay;
			if (time_before_eq(this_timer, jiffies))
				tun_flow_delete(tun, e);
			else if (time_before(this_timer, next_timer))
				next_timer = this_timer;
		}
	}

	if (count)
		mod_timer(&tun->flow_gc_timer, round_jiffies_up(next_timer));
	spin_unlock_bh(&tun->lock);
}

303
static void tun_flow_update(struct tun_struct *tun, u32 rxhash,
304
			    struct tun_file *tfile)
J
Jason Wang 已提交
305 306 307 308
{
	struct hlist_head *head;
	struct tun_flow_entry *e;
	unsigned long delay = tun->ageing_time;
309
	u16 queue_index = tfile->queue_index;
J
Jason Wang 已提交
310 311 312 313 314 315 316 317

	if (!rxhash)
		return;
	else
		head = &tun->flows[tun_hashfn(rxhash)];

	rcu_read_lock();

318 319 320
	/* We may get a very small possibility of OOO during switching, not
	 * worth to optimize.*/
	if (tun->numqueues == 1 || tfile->detached)
J
Jason Wang 已提交
321 322 323 324 325 326 327 328 329
		goto unlock;

	e = tun_flow_find(head, rxhash);
	if (likely(e)) {
		/* TODO: keep queueing to old queue until it's empty? */
		e->queue_index = queue_index;
		e->updated = jiffies;
	} else {
		spin_lock_bh(&tun->lock);
330 331
		if (!tun_flow_find(head, rxhash) &&
		    tun->flow_count < MAX_TAP_FLOWS)
J
Jason Wang 已提交
332 333 334 335 336 337 338 339 340 341 342 343
			tun_flow_create(tun, head, rxhash, queue_index);

		if (!timer_pending(&tun->flow_gc_timer))
			mod_timer(&tun->flow_gc_timer,
				  round_jiffies_up(jiffies + delay));
		spin_unlock_bh(&tun->lock);
	}

unlock:
	rcu_read_unlock();
}

J
Jason Wang 已提交
344 345 346 347 348 349 350 351 352 353
/* We try to identify a flow through its rxhash first. The reason that
 * we do not check rxq no. is becuase some cards(e.g 82599), chooses
 * the rxq based on the txq where the last packet of the flow comes. As
 * the userspace application move between processors, we may get a
 * different rxq no. here. If we could not get rxhash, then we would
 * hope the rxq no. may help here.
 */
static u16 tun_select_queue(struct net_device *dev, struct sk_buff *skb)
{
	struct tun_struct *tun = netdev_priv(dev);
J
Jason Wang 已提交
354
	struct tun_flow_entry *e;
J
Jason Wang 已提交
355 356 357 358
	u32 txq = 0;
	u32 numqueues = 0;

	rcu_read_lock();
359
	numqueues = ACCESS_ONCE(tun->numqueues);
J
Jason Wang 已提交
360 361 362

	txq = skb_get_rxhash(skb);
	if (txq) {
J
Jason Wang 已提交
363 364 365 366 367 368
		e = tun_flow_find(&tun->flows[tun_hashfn(txq)], txq);
		if (e)
			txq = e->queue_index;
		else
			/* use multiply and shift instead of expensive divide */
			txq = ((u64)txq * numqueues) >> 32;
J
Jason Wang 已提交
369 370 371 372 373 374 375 376 377 378
	} else if (likely(skb_rx_queue_recorded(skb))) {
		txq = skb_get_rx_queue(skb);
		while (unlikely(txq >= numqueues))
			txq -= numqueues;
	}

	rcu_read_unlock();
	return txq;
}

379 380 381
static inline bool tun_not_capable(struct tun_struct *tun)
{
	const struct cred *cred = current_cred();
382
	struct net *net = dev_net(tun->dev);
383 384 385

	return ((uid_valid(tun->owner) && !uid_eq(cred->euid, tun->owner)) ||
		  (gid_valid(tun->group) && !in_egroup_p(tun->group))) &&
386
		!ns_capable(net->user_ns, CAP_NET_ADMIN);
387 388
}

J
Jason Wang 已提交
389 390 391 392 393 394
static void tun_set_real_num_queues(struct tun_struct *tun)
{
	netif_set_real_num_tx_queues(tun->dev, tun->numqueues);
	netif_set_real_num_rx_queues(tun->dev, tun->numqueues);
}

J
Jason Wang 已提交
395 396 397 398 399 400 401
static void tun_disable_queue(struct tun_struct *tun, struct tun_file *tfile)
{
	tfile->detached = tun;
	list_add_tail(&tfile->next, &tun->disabled);
	++tun->numdisabled;
}

J
Jason Wang 已提交
402
static struct tun_struct *tun_enable_queue(struct tun_file *tfile)
J
Jason Wang 已提交
403 404 405 406 407 408 409 410 411
{
	struct tun_struct *tun = tfile->detached;

	tfile->detached = NULL;
	list_del_init(&tfile->next);
	--tun->numdisabled;
	return tun;
}

J
Jason Wang 已提交
412 413 414 415 416
static void __tun_detach(struct tun_file *tfile, bool clean)
{
	struct tun_file *ntfile;
	struct tun_struct *tun;

417 418
	tun = rtnl_dereference(tfile->tun);

419
	if (tun && !tfile->detached) {
J
Jason Wang 已提交
420 421 422 423 424
		u16 index = tfile->queue_index;
		BUG_ON(index >= tun->numqueues);

		rcu_assign_pointer(tun->tfiles[index],
				   tun->tfiles[tun->numqueues - 1]);
425
		ntfile = rtnl_dereference(tun->tfiles[index]);
J
Jason Wang 已提交
426 427 428
		ntfile->queue_index = index;

		--tun->numqueues;
429 430
		if (clean) {
			rcu_assign_pointer(tfile->tun, NULL);
J
Jason Wang 已提交
431
			sock_put(&tfile->sk);
432
		} else
J
Jason Wang 已提交
433
			tun_disable_queue(tun, tfile);
J
Jason Wang 已提交
434 435

		synchronize_net();
J
Jason Wang 已提交
436
		tun_flow_delete_by_queue(tun, tun->numqueues + 1);
J
Jason Wang 已提交
437 438 439
		/* Drop read queue */
		skb_queue_purge(&tfile->sk.sk_receive_queue);
		tun_set_real_num_queues(tun);
J
Jason Wang 已提交
440
	} else if (tfile->detached && clean) {
J
Jason Wang 已提交
441
		tun = tun_enable_queue(tfile);
J
Jason Wang 已提交
442 443
		sock_put(&tfile->sk);
	}
J
Jason Wang 已提交
444 445

	if (clean) {
446 447 448 449 450
		if (tun && tun->numqueues == 0 && tun->numdisabled == 0) {
			netif_carrier_off(tun->dev);

			if (!(tun->flags & TUN_PERSIST) &&
			    tun->dev->reg_state == NETREG_REGISTERED)
J
Jason Wang 已提交
451
				unregister_netdevice(tun->dev);
452
		}
J
Jason Wang 已提交
453

J
Jason Wang 已提交
454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469
		BUG_ON(!test_bit(SOCK_EXTERNALLY_ALLOCATED,
				 &tfile->socket.flags));
		sk_release_kernel(&tfile->sk);
	}
}

static void tun_detach(struct tun_file *tfile, bool clean)
{
	rtnl_lock();
	__tun_detach(tfile, clean);
	rtnl_unlock();
}

static void tun_detach_all(struct net_device *dev)
{
	struct tun_struct *tun = netdev_priv(dev);
J
Jason Wang 已提交
470
	struct tun_file *tfile, *tmp;
J
Jason Wang 已提交
471 472 473
	int i, n = tun->numqueues;

	for (i = 0; i < n; i++) {
474
		tfile = rtnl_dereference(tun->tfiles[i]);
J
Jason Wang 已提交
475 476 477 478 479
		BUG_ON(!tfile);
		wake_up_all(&tfile->wq.wait);
		rcu_assign_pointer(tfile->tun, NULL);
		--tun->numqueues;
	}
480 481 482 483
	list_for_each_entry(tfile, &tun->disabled, next) {
		wake_up_all(&tfile->wq.wait);
		rcu_assign_pointer(tfile->tun, NULL);
	}
J
Jason Wang 已提交
484 485 486 487
	BUG_ON(tun->numqueues != 0);

	synchronize_net();
	for (i = 0; i < n; i++) {
488
		tfile = rtnl_dereference(tun->tfiles[i]);
J
Jason Wang 已提交
489 490 491 492
		/* Drop read queue */
		skb_queue_purge(&tfile->sk.sk_receive_queue);
		sock_put(&tfile->sk);
	}
J
Jason Wang 已提交
493 494 495 496 497 498
	list_for_each_entry_safe(tfile, tmp, &tun->disabled, next) {
		tun_enable_queue(tfile);
		skb_queue_purge(&tfile->sk.sk_receive_queue);
		sock_put(&tfile->sk);
	}
	BUG_ON(tun->numdisabled != 0);
J
Jason Wang 已提交
499 500 501

	if (tun->flags & TUN_PERSIST)
		module_put(THIS_MODULE);
J
Jason Wang 已提交
502 503
}

504 505
static int tun_attach(struct tun_struct *tun, struct file *file)
{
E
Eric W. Biederman 已提交
506
	struct tun_file *tfile = file->private_data;
507
	int err;
508

509 510 511 512
	err = security_tun_dev_attach(tfile->socket.sk, tun->security);
	if (err < 0)
		goto out;

513
	err = -EINVAL;
514
	if (rtnl_dereference(tfile->tun) && !tfile->detached)
515 516 517
		goto out;

	err = -EBUSY;
J
Jason Wang 已提交
518 519 520 521
	if (!(tun->flags & TUN_TAP_MQ) && tun->numqueues == 1)
		goto out;

	err = -E2BIG;
J
Jason Wang 已提交
522 523
	if (!tfile->detached &&
	    tun->numqueues + tun->numdisabled == MAX_TAP_QUEUES)
524 525 526
		goto out;

	err = 0;
J
Jason Wang 已提交
527

J
Jason Wang 已提交
528
	/* Re-attach the filter to presist device */
J
Jason Wang 已提交
529 530 531 532 533
	if (tun->filter_attached == true) {
		err = sk_attach_filter(&tun->fprog, tfile->socket.sk);
		if (!err)
			goto out;
	}
J
Jason Wang 已提交
534
	tfile->queue_index = tun->numqueues;
535
	rcu_assign_pointer(tfile->tun, tun);
J
Jason Wang 已提交
536 537
	rcu_assign_pointer(tun->tfiles[tun->numqueues], tfile);
	tun->numqueues++;
538

J
Jason Wang 已提交
539 540 541 542 543
	if (tfile->detached)
		tun_enable_queue(tfile);
	else
		sock_hold(&tfile->sk);

J
Jason Wang 已提交
544
	tun_set_real_num_queues(tun);
545

J
Jason Wang 已提交
546 547 548 549 550 551
	/* device is allowed to go away first, so no need to hold extra
	 * refcnt.
	 */

out:
	return err;
E
Eric W. Biederman 已提交
552 553 554 555
}

static struct tun_struct *__tun_get(struct tun_file *tfile)
{
556
	struct tun_struct *tun;
557

558 559 560 561 562
	rcu_read_lock();
	tun = rcu_dereference(tfile->tun);
	if (tun)
		dev_hold(tun->dev);
	rcu_read_unlock();
563 564

	return tun;
E
Eric W. Biederman 已提交
565 566 567 568 569 570 571 572 573
}

static struct tun_struct *tun_get(struct file *file)
{
	return __tun_get(file->private_data);
}

static void tun_put(struct tun_struct *tun)
{
574
	dev_put(tun->dev);
E
Eric W. Biederman 已提交
575 576
}

577
/* TAP filtering */
578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626
static void addr_hash_set(u32 *mask, const u8 *addr)
{
	int n = ether_crc(ETH_ALEN, addr) >> 26;
	mask[n >> 5] |= (1 << (n & 31));
}

static unsigned int addr_hash_test(const u32 *mask, const u8 *addr)
{
	int n = ether_crc(ETH_ALEN, addr) >> 26;
	return mask[n >> 5] & (1 << (n & 31));
}

static int update_filter(struct tap_filter *filter, void __user *arg)
{
	struct { u8 u[ETH_ALEN]; } *addr;
	struct tun_filter uf;
	int err, alen, n, nexact;

	if (copy_from_user(&uf, arg, sizeof(uf)))
		return -EFAULT;

	if (!uf.count) {
		/* Disabled */
		filter->count = 0;
		return 0;
	}

	alen = ETH_ALEN * uf.count;
	addr = kmalloc(alen, GFP_KERNEL);
	if (!addr)
		return -ENOMEM;

	if (copy_from_user(addr, arg + sizeof(uf), alen)) {
		err = -EFAULT;
		goto done;
	}

	/* The filter is updated without holding any locks. Which is
	 * perfectly safe. We disable it first and in the worst
	 * case we'll accept a few undesired packets. */
	filter->count = 0;
	wmb();

	/* Use first set of addresses as an exact filter */
	for (n = 0; n < uf.count && n < FLT_EXACT_COUNT; n++)
		memcpy(filter->addr[n], addr[n].u, ETH_ALEN);

	nexact = n;

627 628
	/* Remaining multicast addresses are hashed,
	 * unicast will leave the filter disabled. */
629
	memset(filter->mask, 0, sizeof(filter->mask));
630 631 632 633 634
	for (; n < uf.count; n++) {
		if (!is_multicast_ether_addr(addr[n].u)) {
			err = 0; /* no filter */
			goto done;
		}
635
		addr_hash_set(filter->mask, addr[n].u);
636
	}
637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664

	/* For ALLMULTI just set the mask to all ones.
	 * This overrides the mask populated above. */
	if ((uf.flags & TUN_FLT_ALLMULTI))
		memset(filter->mask, ~0, sizeof(filter->mask));

	/* Now enable the filter */
	wmb();
	filter->count = nexact;

	/* Return the number of exact filters */
	err = nexact;

done:
	kfree(addr);
	return err;
}

/* Returns: 0 - drop, !=0 - accept */
static int run_filter(struct tap_filter *filter, const struct sk_buff *skb)
{
	/* Cannot use eth_hdr(skb) here because skb_mac_hdr() is incorrect
	 * at this point. */
	struct ethhdr *eh = (struct ethhdr *) skb->data;
	int i;

	/* Exact match */
	for (i = 0; i < filter->count; i++)
665
		if (ether_addr_equal(eh->h_dest, filter->addr[i]))
666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686
			return 1;

	/* Inexact match (multicast only) */
	if (is_multicast_ether_addr(eh->h_dest))
		return addr_hash_test(filter->mask, eh->h_dest);

	return 0;
}

/*
 * Checks whether the packet is accepted or not.
 * Returns: 0 - drop, !=0 - accept
 */
static int check_filter(struct tap_filter *filter, const struct sk_buff *skb)
{
	if (!filter->count)
		return 1;

	return run_filter(filter, skb);
}

L
Linus Torvalds 已提交
687 688
/* Network device part of the driver */

689
static const struct ethtool_ops tun_ethtool_ops;
L
Linus Torvalds 已提交
690

691 692 693
/* Net device detach from fd. */
static void tun_net_uninit(struct net_device *dev)
{
J
Jason Wang 已提交
694
	tun_detach_all(dev);
695 696
}

L
Linus Torvalds 已提交
697 698 699
/* Net device open. */
static int tun_net_open(struct net_device *dev)
{
J
Jason Wang 已提交
700
	netif_tx_start_all_queues(dev);
L
Linus Torvalds 已提交
701 702 703 704 705 706
	return 0;
}

/* Net device close. */
static int tun_net_close(struct net_device *dev)
{
J
Jason Wang 已提交
707
	netif_tx_stop_all_queues(dev);
L
Linus Torvalds 已提交
708 709 710 711
	return 0;
}

/* Net device start xmit */
712
static netdev_tx_t tun_net_xmit(struct sk_buff *skb, struct net_device *dev)
L
Linus Torvalds 已提交
713 714
{
	struct tun_struct *tun = netdev_priv(dev);
J
Jason Wang 已提交
715
	int txq = skb->queue_mapping;
716
	struct tun_file *tfile;
L
Linus Torvalds 已提交
717

718
	rcu_read_lock();
J
Jason Wang 已提交
719 720
	tfile = rcu_dereference(tun->tfiles[txq]);

L
Linus Torvalds 已提交
721
	/* Drop packet if interface is not attached */
J
Jason Wang 已提交
722
	if (txq >= tun->numqueues)
L
Linus Torvalds 已提交
723 724
		goto drop;

725 726
	tun_debug(KERN_INFO, tun, "tun_net_xmit %d\n", skb->len);

J
Jason Wang 已提交
727 728
	BUG_ON(!tfile);

729 730 731 732 733 734
	/* Drop if the filter does not like it.
	 * This is a noop if the filter is disabled.
	 * Filter can be enabled only for the TAP devices. */
	if (!check_filter(&tun->txflt, skb))
		goto drop;

J
Jason Wang 已提交
735 736
	if (tfile->socket.sk->sk_filter &&
	    sk_filter(tfile->socket.sk, skb))
M
Michael S. Tsirkin 已提交
737 738
		goto drop;

R
Rami Rosen 已提交
739
	/* Limit the number of packets queued by dividing txq length with the
J
Jason Wang 已提交
740 741
	 * number of queues.
	 */
J
Jason Wang 已提交
742
	if (skb_queue_len(&tfile->socket.sk->sk_receive_queue)
743 744
			  >= dev->tx_queue_len / tun->numqueues)
		goto drop;
L
Linus Torvalds 已提交
745

746 747 748 749 750
	if (skb->sk) {
		sock_tx_timestamp(skb->sk, &skb_shinfo(skb)->tx_flags);
		sw_tx_timestamp(skb);
	}

M
Michael S. Tsirkin 已提交
751 752
	/* Orphan the skb - required as we might hang on to it
	 * for indefinite time. */
M
Michael S. Tsirkin 已提交
753 754
	if (unlikely(skb_orphan_frags(skb, GFP_ATOMIC)))
		goto drop;
M
Michael S. Tsirkin 已提交
755 756
	skb_orphan(skb);

757 758
	nf_reset(skb);

759
	/* Enqueue packet */
J
Jason Wang 已提交
760
	skb_queue_tail(&tfile->socket.sk->sk_receive_queue, skb);
L
Linus Torvalds 已提交
761 762

	/* Notify and wake up reader process */
J
Jason Wang 已提交
763 764 765
	if (tfile->flags & TUN_FASYNC)
		kill_fasync(&tfile->fasync, SIGIO, POLL_IN);
	wake_up_interruptible_poll(&tfile->wq.wait, POLLIN |
766
				   POLLRDNORM | POLLRDBAND);
767 768

	rcu_read_unlock();
769
	return NETDEV_TX_OK;
L
Linus Torvalds 已提交
770 771

drop:
772
	dev->stats.tx_dropped++;
773
	skb_tx_error(skb);
L
Linus Torvalds 已提交
774
	kfree_skb(skb);
775
	rcu_read_unlock();
776
	return NETDEV_TX_OK;
L
Linus Torvalds 已提交
777 778
}

779
static void tun_net_mclist(struct net_device *dev)
L
Linus Torvalds 已提交
780
{
781 782 783 784 785
	/*
	 * This callback is supposed to deal with mc filter in
	 * _rx_ path and has nothing to do with the _tx_ path.
	 * In rx path we always accept everything userspace gives us.
	 */
L
Linus Torvalds 已提交
786 787
}

E
Ed Swierk 已提交
788 789 790 791 792 793 794 795 796 797 798 799
#define MIN_MTU 68
#define MAX_MTU 65535

static int
tun_net_change_mtu(struct net_device *dev, int new_mtu)
{
	if (new_mtu < MIN_MTU || new_mtu + dev->hard_header_len > MAX_MTU)
		return -EINVAL;
	dev->mtu = new_mtu;
	return 0;
}

800 801
static netdev_features_t tun_net_fix_features(struct net_device *dev,
	netdev_features_t features)
802 803 804 805 806
{
	struct tun_struct *tun = netdev_priv(dev);

	return (features & tun->set_features) | (features & ~TUN_USER_FEATURES);
}
807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823
#ifdef CONFIG_NET_POLL_CONTROLLER
static void tun_poll_controller(struct net_device *dev)
{
	/*
	 * Tun only receives frames when:
	 * 1) the char device endpoint gets data from user space
	 * 2) the tun socket gets a sendmsg call from user space
	 * Since both of those are syncronous operations, we are guaranteed
	 * never to have pending data when we poll for it
	 * so theres nothing to do here but return.
	 * We need this though so netpoll recognizes us as an interface that
	 * supports polling, which enables bridge devices in virt setups to
	 * still use netconsole
	 */
	return;
}
#endif
S
Stephen Hemminger 已提交
824
static const struct net_device_ops tun_netdev_ops = {
825
	.ndo_uninit		= tun_net_uninit,
S
Stephen Hemminger 已提交
826 827
	.ndo_open		= tun_net_open,
	.ndo_stop		= tun_net_close,
828
	.ndo_start_xmit		= tun_net_xmit,
S
Stephen Hemminger 已提交
829
	.ndo_change_mtu		= tun_net_change_mtu,
830
	.ndo_fix_features	= tun_net_fix_features,
J
Jason Wang 已提交
831
	.ndo_select_queue	= tun_select_queue,
832 833 834
#ifdef CONFIG_NET_POLL_CONTROLLER
	.ndo_poll_controller	= tun_poll_controller,
#endif
S
Stephen Hemminger 已提交
835 836 837
};

static const struct net_device_ops tap_netdev_ops = {
838
	.ndo_uninit		= tun_net_uninit,
S
Stephen Hemminger 已提交
839 840
	.ndo_open		= tun_net_open,
	.ndo_stop		= tun_net_close,
841
	.ndo_start_xmit		= tun_net_xmit,
S
Stephen Hemminger 已提交
842
	.ndo_change_mtu		= tun_net_change_mtu,
843
	.ndo_fix_features	= tun_net_fix_features,
844
	.ndo_set_rx_mode	= tun_net_mclist,
S
Stephen Hemminger 已提交
845 846
	.ndo_set_mac_address	= eth_mac_addr,
	.ndo_validate_addr	= eth_validate_addr,
J
Jason Wang 已提交
847
	.ndo_select_queue	= tun_select_queue,
848 849 850
#ifdef CONFIG_NET_POLL_CONTROLLER
	.ndo_poll_controller	= tun_poll_controller,
#endif
S
Stephen Hemminger 已提交
851 852
};

853
static void tun_flow_init(struct tun_struct *tun)
J
Jason Wang 已提交
854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871
{
	int i;

	for (i = 0; i < TUN_NUM_FLOW_ENTRIES; i++)
		INIT_HLIST_HEAD(&tun->flows[i]);

	tun->ageing_time = TUN_FLOW_EXPIRE;
	setup_timer(&tun->flow_gc_timer, tun_flow_cleanup, (unsigned long)tun);
	mod_timer(&tun->flow_gc_timer,
		  round_jiffies_up(jiffies + tun->ageing_time));
}

static void tun_flow_uninit(struct tun_struct *tun)
{
	del_timer_sync(&tun->flow_gc_timer);
	tun_flow_flush(tun);
}

L
Linus Torvalds 已提交
872 873 874 875
/* Initialize net device. */
static void tun_net_init(struct net_device *dev)
{
	struct tun_struct *tun = netdev_priv(dev);
876

L
Linus Torvalds 已提交
877 878
	switch (tun->flags & TUN_TYPE_MASK) {
	case TUN_TUN_DEV:
S
Stephen Hemminger 已提交
879 880
		dev->netdev_ops = &tun_netdev_ops;

L
Linus Torvalds 已提交
881 882 883 884 885 886
		/* Point-to-Point TUN Device */
		dev->hard_header_len = 0;
		dev->addr_len = 0;
		dev->mtu = 1500;

		/* Zero header length */
887
		dev->type = ARPHRD_NONE;
L
Linus Torvalds 已提交
888 889 890 891 892
		dev->flags = IFF_POINTOPOINT | IFF_NOARP | IFF_MULTICAST;
		dev->tx_queue_len = TUN_READQ_SIZE;  /* We prefer our own queue length */
		break;

	case TUN_TAP_DEV:
K
Kusanagi Kouichi 已提交
893
		dev->netdev_ops = &tap_netdev_ops;
L
Linus Torvalds 已提交
894 895
		/* Ethernet TAP Device */
		ether_setup(dev);
896
		dev->priv_flags &= ~IFF_TX_SKB_SHARING;
897
		dev->priv_flags |= IFF_LIVE_ADDR_CHANGE;
898

899
		eth_hw_addr_random(dev);
900

L
Linus Torvalds 已提交
901 902 903 904 905 906 907 908
		dev->tx_queue_len = TUN_READQ_SIZE;  /* We prefer our own queue length */
		break;
	}
}

/* Character device part */

/* Poll */
J
Jason Wang 已提交
909
static unsigned int tun_chr_poll(struct file *file, poll_table *wait)
910
{
911 912
	struct tun_file *tfile = file->private_data;
	struct tun_struct *tun = __tun_get(tfile);
913
	struct sock *sk;
914
	unsigned int mask = 0;
L
Linus Torvalds 已提交
915 916

	if (!tun)
917
		return POLLERR;
L
Linus Torvalds 已提交
918

J
Jason Wang 已提交
919
	sk = tfile->socket.sk;
920

921
	tun_debug(KERN_INFO, tun, "tun_chr_poll\n");
L
Linus Torvalds 已提交
922

J
Jason Wang 已提交
923
	poll_wait(file, &tfile->wq.wait, wait);
924

925
	if (!skb_queue_empty(&sk->sk_receive_queue))
L
Linus Torvalds 已提交
926 927
		mask |= POLLIN | POLLRDNORM;

928 929 930 931 932
	if (sock_writeable(sk) ||
	    (!test_and_set_bit(SOCK_ASYNC_NOSPACE, &sk->sk_socket->flags) &&
	     sock_writeable(sk)))
		mask |= POLLOUT | POLLWRNORM;

933 934 935
	if (tun->dev->reg_state != NETREG_REGISTERED)
		mask = POLLERR;

E
Eric W. Biederman 已提交
936
	tun_put(tun);
L
Linus Torvalds 已提交
937 938 939
	return mask;
}

940 941
/* prepad is the amount to reserve at front.  len is length after that.
 * linear is a hint as to how much to copy (usually headers). */
J
Jason Wang 已提交
942
static struct sk_buff *tun_alloc_skb(struct tun_file *tfile,
943 944
				     size_t prepad, size_t len,
				     size_t linear, int noblock)
945
{
J
Jason Wang 已提交
946
	struct sock *sk = tfile->socket.sk;
947
	struct sk_buff *skb;
948
	int err;
949 950

	/* Under a page?  Don't bother with paged skb. */
H
Herbert Xu 已提交
951
	if (prepad + len < PAGE_SIZE || !linear)
952
		linear = len;
953

954
	skb = sock_alloc_send_pskb(sk, prepad + linear, len - linear, noblock,
955
				   &err, 0);
956
	if (!skb)
957
		return ERR_PTR(err);
958 959 960

	skb_reserve(skb, prepad);
	skb_put(skb, linear);
961 962
	skb->data_len = len - linear;
	skb->len += len - linear;
963 964 965 966

	return skb;
}

L
Linus Torvalds 已提交
967
/* Get packet from user space buffer */
J
Jason Wang 已提交
968 969 970
static ssize_t tun_get_user(struct tun_struct *tun, struct tun_file *tfile,
			    void *msg_control, const struct iovec *iv,
			    size_t total_len, size_t count, int noblock)
L
Linus Torvalds 已提交
971
{
972
	struct tun_pi pi = { 0, cpu_to_be16(ETH_P_IP) };
L
Linus Torvalds 已提交
973
	struct sk_buff *skb;
974
	size_t len = total_len, align = NET_SKB_PAD, linear;
975
	struct virtio_net_hdr gso = { 0 };
976
	int offset = 0;
977 978 979
	int copylen;
	bool zerocopy = false;
	int err;
980
	u32 rxhash;
L
Linus Torvalds 已提交
981 982

	if (!(tun->flags & TUN_NO_PI)) {
983
		if (len < sizeof(pi))
L
Linus Torvalds 已提交
984
			return -EINVAL;
985
		len -= sizeof(pi);
L
Linus Torvalds 已提交
986

987
		if (memcpy_fromiovecend((void *)&pi, iv, 0, sizeof(pi)))
L
Linus Torvalds 已提交
988
			return -EFAULT;
989
		offset += sizeof(pi);
L
Linus Torvalds 已提交
990 991
	}

992
	if (tun->flags & TUN_VNET_HDR) {
993
		if (len < tun->vnet_hdr_sz)
994
			return -EINVAL;
995
		len -= tun->vnet_hdr_sz;
996

997
		if (memcpy_fromiovecend((void *)&gso, iv, offset, sizeof(gso)))
998 999
			return -EFAULT;

1000 1001 1002 1003
		if ((gso.flags & VIRTIO_NET_HDR_F_NEEDS_CSUM) &&
		    gso.csum_start + gso.csum_offset + 2 > gso.hdr_len)
			gso.hdr_len = gso.csum_start + gso.csum_offset + 2;

1004 1005
		if (gso.hdr_len > len)
			return -EINVAL;
1006
		offset += tun->vnet_hdr_sz;
1007 1008
	}

1009
	if ((tun->flags & TUN_TYPE_MASK) == TUN_TAP_DEV) {
1010
		align += NET_IP_ALIGN;
H
Herbert Xu 已提交
1011 1012
		if (unlikely(len < ETH_HLEN ||
			     (gso.hdr_len && gso.hdr_len < ETH_HLEN)))
1013 1014
			return -EINVAL;
	}
1015

1016 1017 1018
	if (msg_control) {
		/* There are 256 bytes to be copied in skb, so there is
		 * enough room for skb expand head in case it is used.
1019 1020
		 * The rest of the buffer is mapped from userspace.
		 */
1021
		copylen = gso.hdr_len ? gso.hdr_len : GOODCOPY_LEN;
1022
		linear = copylen;
1023 1024 1025 1026 1027
		if (iov_pages(iv, offset + copylen, count) <= MAX_SKB_FRAGS)
			zerocopy = true;
	}

	if (!zerocopy) {
1028
		copylen = len;
1029 1030
		linear = gso.hdr_len;
	}
1031

1032
	skb = tun_alloc_skb(tfile, align, copylen, linear, noblock);
1033 1034 1035 1036
	if (IS_ERR(skb)) {
		if (PTR_ERR(skb) != -EAGAIN)
			tun->dev->stats.rx_dropped++;
		return PTR_ERR(skb);
L
Linus Torvalds 已提交
1037 1038
	}

1039 1040
	if (zerocopy)
		err = zerocopy_sg_from_iovec(skb, iv, offset, count);
1041
	else {
1042
		err = skb_copy_datagram_from_iovec(skb, 0, iv, offset, len);
1043 1044 1045 1046 1047
		if (!err && msg_control) {
			struct ubuf_info *uarg = msg_control;
			uarg->callback(uarg, false);
		}
	}
1048 1049

	if (err) {
1050
		tun->dev->stats.rx_dropped++;
D
Dave Jones 已提交
1051
		kfree_skb(skb);
L
Linus Torvalds 已提交
1052
		return -EFAULT;
D
Dave Jones 已提交
1053
	}
L
Linus Torvalds 已提交
1054

1055 1056 1057 1058 1059 1060 1061
	if (gso.flags & VIRTIO_NET_HDR_F_NEEDS_CSUM) {
		if (!skb_partial_csum_set(skb, gso.csum_start,
					  gso.csum_offset)) {
			tun->dev->stats.rx_frame_errors++;
			kfree_skb(skb);
			return -EINVAL;
		}
1062
	}
1063

L
Linus Torvalds 已提交
1064 1065
	switch (tun->flags & TUN_TYPE_MASK) {
	case TUN_TUN_DEV:
1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080
		if (tun->flags & TUN_NO_PI) {
			switch (skb->data[0] & 0xf0) {
			case 0x40:
				pi.proto = htons(ETH_P_IP);
				break;
			case 0x60:
				pi.proto = htons(ETH_P_IPV6);
				break;
			default:
				tun->dev->stats.rx_dropped++;
				kfree_skb(skb);
				return -EINVAL;
			}
		}

1081
		skb_reset_mac_header(skb);
L
Linus Torvalds 已提交
1082
		skb->protocol = pi.proto;
1083
		skb->dev = tun->dev;
L
Linus Torvalds 已提交
1084 1085 1086 1087
		break;
	case TUN_TAP_DEV:
		skb->protocol = eth_type_trans(skb, tun->dev);
		break;
1088
	}
L
Linus Torvalds 已提交
1089

1090 1091 1092 1093
	if (gso.gso_type != VIRTIO_NET_HDR_GSO_NONE) {
		pr_debug("GSO!\n");
		switch (gso.gso_type & ~VIRTIO_NET_HDR_GSO_ECN) {
		case VIRTIO_NET_HDR_GSO_TCPV4:
1094
			skb_shinfo(skb)->gso_type = SKB_GSO_TCPV4;
1095 1096
			break;
		case VIRTIO_NET_HDR_GSO_TCPV6:
1097
			skb_shinfo(skb)->gso_type = SKB_GSO_TCPV6;
1098
			break;
1099
		case VIRTIO_NET_HDR_GSO_UDP:
1100
			skb_shinfo(skb)->gso_type = SKB_GSO_UDP;
1101
			break;
1102 1103 1104 1105 1106 1107 1108
		default:
			tun->dev->stats.rx_frame_errors++;
			kfree_skb(skb);
			return -EINVAL;
		}

		if (gso.gso_type & VIRTIO_NET_HDR_GSO_ECN)
1109
			skb_shinfo(skb)->gso_type |= SKB_GSO_TCP_ECN;
1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121

		skb_shinfo(skb)->gso_size = gso.gso_size;
		if (skb_shinfo(skb)->gso_size == 0) {
			tun->dev->stats.rx_frame_errors++;
			kfree_skb(skb);
			return -EINVAL;
		}

		/* Header must be checked, and gso_segs computed. */
		skb_shinfo(skb)->gso_type |= SKB_GSO_DODGY;
		skb_shinfo(skb)->gso_segs = 0;
	}
1122

1123 1124 1125 1126
	/* copy skb_ubuf_info for callback when skb has no error */
	if (zerocopy) {
		skb_shinfo(skb)->destructor_arg = msg_control;
		skb_shinfo(skb)->tx_flags |= SKBTX_DEV_ZEROCOPY;
1127
		skb_shinfo(skb)->tx_flags |= SKBTX_SHARED_FRAG;
1128 1129
	}

1130
	skb_reset_network_header(skb);
1131
	skb_probe_transport_header(skb, 0);
1132

1133
	rxhash = skb_get_rxhash(skb);
L
Linus Torvalds 已提交
1134
	netif_rx_ni(skb);
1135

1136 1137
	tun->dev->stats.rx_packets++;
	tun->dev->stats.rx_bytes += len;
L
Linus Torvalds 已提交
1138

1139
	tun_flow_update(tun, rxhash, tfile);
1140
	return total_len;
1141
}
L
Linus Torvalds 已提交
1142

1143 1144
static ssize_t tun_chr_aio_write(struct kiocb *iocb, const struct iovec *iv,
			      unsigned long count, loff_t pos)
L
Linus Torvalds 已提交
1145
{
1146
	struct file *file = iocb->ki_filp;
H
Herbert Xu 已提交
1147
	struct tun_struct *tun = tun_get(file);
J
Jason Wang 已提交
1148
	struct tun_file *tfile = file->private_data;
E
Eric W. Biederman 已提交
1149
	ssize_t result;
L
Linus Torvalds 已提交
1150 1151 1152 1153

	if (!tun)
		return -EBADFD;

1154
	tun_debug(KERN_INFO, tun, "tun_chr_write %ld\n", count);
L
Linus Torvalds 已提交
1155

J
Jason Wang 已提交
1156 1157
	result = tun_get_user(tun, tfile, NULL, iv, iov_length(iv, count),
			      count, file->f_flags & O_NONBLOCK);
E
Eric W. Biederman 已提交
1158 1159 1160

	tun_put(tun);
	return result;
L
Linus Torvalds 已提交
1161 1162 1163
}

/* Put packet to the user space buffer */
1164
static ssize_t tun_put_user(struct tun_struct *tun,
J
Jason Wang 已提交
1165
			    struct tun_file *tfile,
1166 1167
			    struct sk_buff *skb,
			    const struct iovec *iv, int len)
L
Linus Torvalds 已提交
1168 1169 1170
{
	struct tun_pi pi = { 0, skb->protocol };
	ssize_t total = 0;
J
Jason Wang 已提交
1171
	int vlan_offset = 0;
L
Linus Torvalds 已提交
1172 1173 1174 1175 1176 1177 1178 1179 1180

	if (!(tun->flags & TUN_NO_PI)) {
		if ((len -= sizeof(pi)) < 0)
			return -EINVAL;

		if (len < skb->len) {
			/* Packet will be striped */
			pi.flags |= TUN_PKT_STRIP;
		}
1181

1182
		if (memcpy_toiovecend(iv, (void *) &pi, 0, sizeof(pi)))
L
Linus Torvalds 已提交
1183 1184
			return -EFAULT;
		total += sizeof(pi);
1185
	}
L
Linus Torvalds 已提交
1186

1187 1188
	if (tun->flags & TUN_VNET_HDR) {
		struct virtio_net_hdr gso = { 0 }; /* no info leak */
1189
		if ((len -= tun->vnet_hdr_sz) < 0)
1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201
			return -EINVAL;

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

			/* This is a hint as to how much should be linear. */
			gso.hdr_len = skb_headlen(skb);
			gso.gso_size = sinfo->gso_size;
			if (sinfo->gso_type & SKB_GSO_TCPV4)
				gso.gso_type = VIRTIO_NET_HDR_GSO_TCPV4;
			else if (sinfo->gso_type & SKB_GSO_TCPV6)
				gso.gso_type = VIRTIO_NET_HDR_GSO_TCPV6;
1202 1203
			else if (sinfo->gso_type & SKB_GSO_UDP)
				gso.gso_type = VIRTIO_NET_HDR_GSO_UDP;
1204
			else {
1205
				pr_err("unexpected GSO type: "
1206 1207 1208 1209 1210 1211 1212 1213 1214 1215
				       "0x%x, gso_size %d, hdr_len %d\n",
				       sinfo->gso_type, gso.gso_size,
				       gso.hdr_len);
				print_hex_dump(KERN_ERR, "tun: ",
					       DUMP_PREFIX_NONE,
					       16, 1, skb->head,
					       min((int)gso.hdr_len, 64), true);
				WARN_ON_ONCE(1);
				return -EINVAL;
			}
1216 1217 1218 1219 1220 1221 1222
			if (sinfo->gso_type & SKB_GSO_TCP_ECN)
				gso.gso_type |= VIRTIO_NET_HDR_GSO_ECN;
		} else
			gso.gso_type = VIRTIO_NET_HDR_GSO_NONE;

		if (skb->ip_summed == CHECKSUM_PARTIAL) {
			gso.flags = VIRTIO_NET_HDR_F_NEEDS_CSUM;
1223
			gso.csum_start = skb_checksum_start_offset(skb);
1224
			gso.csum_offset = skb->csum_offset;
1225 1226
		} else if (skb->ip_summed == CHECKSUM_UNNECESSARY) {
			gso.flags = VIRTIO_NET_HDR_F_DATA_VALID;
1227 1228
		} /* else everything is zero */

1229 1230
		if (unlikely(memcpy_toiovecend(iv, (void *)&gso, total,
					       sizeof(gso))))
1231
			return -EFAULT;
1232
		total += tun->vnet_hdr_sz;
1233 1234
	}

J
Jason Wang 已提交
1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263
	if (!vlan_tx_tag_present(skb)) {
		len = min_t(int, skb->len, len);
	} else {
		int copy, ret;
		struct {
			__be16 h_vlan_proto;
			__be16 h_vlan_TCI;
		} veth;

		veth.h_vlan_proto = skb->vlan_proto;
		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, total, copy);
		len -= copy;
		total += copy;
		if (ret || !len)
			goto done;

		copy = min_t(int, sizeof(veth), len);
		ret = memcpy_toiovecend(iv, (void *)&veth, total, copy);
		len -= copy;
		total += copy;
		if (ret || !len)
			goto done;
	}
L
Linus Torvalds 已提交
1264

J
Jason Wang 已提交
1265 1266
	skb_copy_datagram_const_iovec(skb, vlan_offset, iv, total, len);
	total += len;
L
Linus Torvalds 已提交
1267

J
Jason Wang 已提交
1268
done:
1269 1270
	tun->dev->stats.tx_packets++;
	tun->dev->stats.tx_bytes += len;
L
Linus Torvalds 已提交
1271 1272 1273 1274

	return total;
}

J
Jason Wang 已提交
1275
static ssize_t tun_do_read(struct tun_struct *tun, struct tun_file *tfile,
1276 1277
			   struct kiocb *iocb, const struct iovec *iv,
			   ssize_t len, int noblock)
L
Linus Torvalds 已提交
1278 1279 1280
{
	DECLARE_WAITQUEUE(wait, current);
	struct sk_buff *skb;
1281
	ssize_t ret = 0;
L
Linus Torvalds 已提交
1282

1283
	tun_debug(KERN_INFO, tun, "tun_do_read\n");
L
Linus Torvalds 已提交
1284

1285
	if (unlikely(!noblock))
J
Jason Wang 已提交
1286
		add_wait_queue(&tfile->wq.wait, &wait);
L
Linus Torvalds 已提交
1287 1288 1289 1290
	while (len) {
		current->state = TASK_INTERRUPTIBLE;

		/* Read frames from the queue */
J
Jason Wang 已提交
1291
		if (!(skb = skb_dequeue(&tfile->socket.sk->sk_receive_queue))) {
1292
			if (noblock) {
L
Linus Torvalds 已提交
1293 1294 1295 1296 1297 1298 1299
				ret = -EAGAIN;
				break;
			}
			if (signal_pending(current)) {
				ret = -ERESTARTSYS;
				break;
			}
1300 1301 1302 1303
			if (tun->dev->reg_state != NETREG_REGISTERED) {
				ret = -EIO;
				break;
			}
L
Linus Torvalds 已提交
1304 1305 1306 1307 1308 1309

			/* Nothing to read, let's sleep */
			schedule();
			continue;
		}

J
Jason Wang 已提交
1310
		ret = tun_put_user(tun, tfile, skb, iv, len);
1311 1312
		kfree_skb(skb);
		break;
L
Linus Torvalds 已提交
1313 1314 1315
	}

	current->state = TASK_RUNNING;
1316
	if (unlikely(!noblock))
J
Jason Wang 已提交
1317
		remove_wait_queue(&tfile->wq.wait, &wait);
L
Linus Torvalds 已提交
1318

1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337
	return ret;
}

static ssize_t tun_chr_aio_read(struct kiocb *iocb, const struct iovec *iv,
			    unsigned long count, loff_t pos)
{
	struct file *file = iocb->ki_filp;
	struct tun_file *tfile = file->private_data;
	struct tun_struct *tun = __tun_get(tfile);
	ssize_t len, ret;

	if (!tun)
		return -EBADFD;
	len = iov_length(iv, count);
	if (len < 0) {
		ret = -EINVAL;
		goto out;
	}

J
Jason Wang 已提交
1338 1339
	ret = tun_do_read(tun, tfile, iocb, iv, len,
			  file->f_flags & O_NONBLOCK);
1340
	ret = min_t(ssize_t, ret, len);
E
Eric W. Biederman 已提交
1341 1342
out:
	tun_put(tun);
L
Linus Torvalds 已提交
1343 1344 1345
	return ret;
}

J
Jason Wang 已提交
1346 1347 1348 1349
static void tun_free_netdev(struct net_device *dev)
{
	struct tun_struct *tun = netdev_priv(dev);

J
Jason Wang 已提交
1350
	BUG_ON(!(list_empty(&tun->disabled)));
J
Jason Wang 已提交
1351
	tun_flow_uninit(tun);
1352
	security_tun_dev_free_security(tun->security);
J
Jason Wang 已提交
1353 1354 1355
	free_netdev(dev);
}

L
Linus Torvalds 已提交
1356 1357 1358 1359
static void tun_setup(struct net_device *dev)
{
	struct tun_struct *tun = netdev_priv(dev);

1360 1361
	tun->owner = INVALID_UID;
	tun->group = INVALID_GID;
L
Linus Torvalds 已提交
1362 1363

	dev->ethtool_ops = &tun_ethtool_ops;
J
Jason Wang 已提交
1364
	dev->destructor = tun_free_netdev;
L
Linus Torvalds 已提交
1365 1366
}

1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381
/* Trivial set of netlink ops to allow deleting tun or tap
 * device with netlink.
 */
static int tun_validate(struct nlattr *tb[], struct nlattr *data[])
{
	return -EINVAL;
}

static struct rtnl_link_ops tun_link_ops __read_mostly = {
	.kind		= DRV_NAME,
	.priv_size	= sizeof(struct tun_struct),
	.setup		= tun_setup,
	.validate	= tun_validate,
};

1382 1383
static void tun_sock_write_space(struct sock *sk)
{
J
Jason Wang 已提交
1384
	struct tun_file *tfile;
1385
	wait_queue_head_t *wqueue;
1386 1387 1388 1389 1390 1391 1392

	if (!sock_writeable(sk))
		return;

	if (!test_and_clear_bit(SOCK_ASYNC_NOSPACE, &sk->sk_socket->flags))
		return;

1393 1394 1395
	wqueue = sk_sleep(sk);
	if (wqueue && waitqueue_active(wqueue))
		wake_up_interruptible_sync_poll(wqueue, POLLOUT |
1396
						POLLWRNORM | POLLWRBAND);
H
Herbert Xu 已提交
1397

J
Jason Wang 已提交
1398 1399
	tfile = container_of(sk, struct tun_file, sk);
	kill_fasync(&tfile->fasync, SIGIO, POLL_OUT);
1400 1401
}

1402 1403 1404
static int tun_sendmsg(struct kiocb *iocb, struct socket *sock,
		       struct msghdr *m, size_t total_len)
{
J
Jason Wang 已提交
1405 1406 1407 1408 1409 1410 1411 1412 1413 1414
	int ret;
	struct tun_file *tfile = container_of(sock, struct tun_file, socket);
	struct tun_struct *tun = __tun_get(tfile);

	if (!tun)
		return -EBADFD;
	ret = tun_get_user(tun, tfile, m->msg_control, m->msg_iov, total_len,
			   m->msg_iovlen, m->msg_flags & MSG_DONTWAIT);
	tun_put(tun);
	return ret;
1415 1416 1417 1418 1419 1420
}

static int tun_recvmsg(struct kiocb *iocb, struct socket *sock,
		       struct msghdr *m, size_t total_len,
		       int flags)
{
J
Jason Wang 已提交
1421 1422
	struct tun_file *tfile = container_of(sock, struct tun_file, socket);
	struct tun_struct *tun = __tun_get(tfile);
1423
	int ret;
J
Jason Wang 已提交
1424 1425 1426 1427

	if (!tun)
		return -EBADFD;

1428
	if (flags & ~(MSG_DONTWAIT|MSG_TRUNC|MSG_ERRQUEUE)) {
1429 1430 1431
		ret = -EINVAL;
		goto out;
	}
1432 1433 1434 1435 1436
	if (flags & MSG_ERRQUEUE) {
		ret = sock_recv_errqueue(sock->sk, m, total_len,
					 SOL_PACKET, TUN_TX_TIMESTAMP);
		goto out;
	}
J
Jason Wang 已提交
1437
	ret = tun_do_read(tun, tfile, iocb, m->msg_iov, total_len,
1438 1439 1440 1441 1442
			  flags & MSG_DONTWAIT);
	if (ret > total_len) {
		m->msg_flags |= MSG_TRUNC;
		ret = flags & MSG_TRUNC ? ret : total_len;
	}
1443
out:
J
Jason Wang 已提交
1444
	tun_put(tun);
1445 1446 1447
	return ret;
}

1448 1449 1450 1451 1452 1453 1454
static int tun_release(struct socket *sock)
{
	if (sock->sk)
		sock_put(sock->sk);
	return 0;
}

1455 1456 1457 1458
/* Ops structure to mimic raw sockets with tun */
static const struct proto_ops tun_socket_ops = {
	.sendmsg = tun_sendmsg,
	.recvmsg = tun_recvmsg,
1459
	.release = tun_release,
1460 1461
};

1462 1463 1464
static struct proto tun_proto = {
	.name		= "tun",
	.owner		= THIS_MODULE,
J
Jason Wang 已提交
1465
	.obj_size	= sizeof(struct tun_file),
1466
};
1467

1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479
static int tun_flags(struct tun_struct *tun)
{
	int flags = 0;

	if (tun->flags & TUN_TUN_DEV)
		flags |= IFF_TUN;
	else
		flags |= IFF_TAP;

	if (tun->flags & TUN_NO_PI)
		flags |= IFF_NO_PI;

1480 1481 1482
	/* This flag has no real effect.  We track the value for backwards
	 * compatibility.
	 */
1483 1484 1485 1486 1487 1488
	if (tun->flags & TUN_ONE_QUEUE)
		flags |= IFF_ONE_QUEUE;

	if (tun->flags & TUN_VNET_HDR)
		flags |= IFF_VNET_HDR;

J
Jason Wang 已提交
1489 1490 1491
	if (tun->flags & TUN_TAP_MQ)
		flags |= IFF_MULTI_QUEUE;

1492 1493 1494
	if (tun->flags & TUN_PERSIST)
		flags |= IFF_PERSIST;

1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508
	return flags;
}

static ssize_t tun_show_flags(struct device *dev, struct device_attribute *attr,
			      char *buf)
{
	struct tun_struct *tun = netdev_priv(to_net_dev(dev));
	return sprintf(buf, "0x%x\n", tun_flags(tun));
}

static ssize_t tun_show_owner(struct device *dev, struct device_attribute *attr,
			      char *buf)
{
	struct tun_struct *tun = netdev_priv(to_net_dev(dev));
1509 1510 1511 1512
	return uid_valid(tun->owner)?
		sprintf(buf, "%u\n",
			from_kuid_munged(current_user_ns(), tun->owner)):
		sprintf(buf, "-1\n");
1513 1514 1515 1516 1517 1518
}

static ssize_t tun_show_group(struct device *dev, struct device_attribute *attr,
			      char *buf)
{
	struct tun_struct *tun = netdev_priv(to_net_dev(dev));
1519 1520 1521 1522
	return gid_valid(tun->group) ?
		sprintf(buf, "%u\n",
			from_kgid_munged(current_user_ns(), tun->group)):
		sprintf(buf, "-1\n");
1523 1524 1525 1526 1527 1528
}

static DEVICE_ATTR(tun_flags, 0444, tun_show_flags, NULL);
static DEVICE_ATTR(owner, 0444, tun_show_owner, NULL);
static DEVICE_ATTR(group, 0444, tun_show_group, NULL);

1529
static int tun_set_iff(struct net *net, struct file *file, struct ifreq *ifr)
L
Linus Torvalds 已提交
1530 1531
{
	struct tun_struct *tun;
J
Jason Wang 已提交
1532
	struct tun_file *tfile = file->private_data;
L
Linus Torvalds 已提交
1533 1534 1535
	struct net_device *dev;
	int err;

1536 1537 1538
	if (tfile->detached)
		return -EINVAL;

1539 1540
	dev = __dev_get_by_name(net, ifr->ifr_name);
	if (dev) {
1541 1542
		if (ifr->ifr_flags & IFF_TUN_EXCL)
			return -EBUSY;
1543 1544 1545 1546 1547 1548 1549
		if ((ifr->ifr_flags & IFF_TUN) && dev->netdev_ops == &tun_netdev_ops)
			tun = netdev_priv(dev);
		else if ((ifr->ifr_flags & IFF_TAP) && dev->netdev_ops == &tap_netdev_ops)
			tun = netdev_priv(dev);
		else
			return -EINVAL;

1550 1551 1552 1553
		if (!!(ifr->ifr_flags & IFF_MULTI_QUEUE) !=
		    !!(tun->flags & TUN_TAP_MQ))
			return -EINVAL;

1554
		if (tun_not_capable(tun))
P
Paul Moore 已提交
1555
			return -EPERM;
1556
		err = security_tun_dev_open(tun->security);
P
Paul Moore 已提交
1557 1558 1559
		if (err < 0)
			return err;

1560 1561 1562
		err = tun_attach(tun, file);
		if (err < 0)
			return err;
J
Jason Wang 已提交
1563 1564

		if (tun->flags & TUN_TAP_MQ &&
1565 1566 1567 1568 1569 1570
		    (tun->numqueues + tun->numdisabled > 1)) {
			/* One or more queue has already been attached, no need
			 * to initialize the device again.
			 */
			return 0;
		}
1571
	}
L
Linus Torvalds 已提交
1572 1573 1574
	else {
		char *name;
		unsigned long flags = 0;
1575 1576
		int queues = ifr->ifr_flags & IFF_MULTI_QUEUE ?
			     MAX_TAP_QUEUES : 1;
L
Linus Torvalds 已提交
1577

1578
		if (!ns_capable(net->user_ns, CAP_NET_ADMIN))
1579
			return -EPERM;
P
Paul Moore 已提交
1580 1581 1582
		err = security_tun_dev_create();
		if (err < 0)
			return err;
1583

L
Linus Torvalds 已提交
1584 1585 1586 1587 1588 1589 1590 1591 1592
		/* Set dev type */
		if (ifr->ifr_flags & IFF_TUN) {
			/* TUN device */
			flags |= TUN_TUN_DEV;
			name = "tun%d";
		} else if (ifr->ifr_flags & IFF_TAP) {
			/* TAP device */
			flags |= TUN_TAP_DEV;
			name = "tap%d";
1593
		} else
1594
			return -EINVAL;
1595

L
Linus Torvalds 已提交
1596 1597 1598
		if (*ifr->ifr_name)
			name = ifr->ifr_name;

J
Jason Wang 已提交
1599
		dev = alloc_netdev_mqs(sizeof(struct tun_struct), name,
1600 1601
				       tun_setup, queues, queues);

L
Linus Torvalds 已提交
1602 1603 1604
		if (!dev)
			return -ENOMEM;

1605
		dev_net_set(dev, net);
1606
		dev->rtnl_link_ops = &tun_link_ops;
1607
		dev->ifindex = tfile->ifindex;
S
Stephen Hemminger 已提交
1608

L
Linus Torvalds 已提交
1609 1610 1611
		tun = netdev_priv(dev);
		tun->dev = dev;
		tun->flags = flags;
1612
		tun->txflt.count = 0;
1613
		tun->vnet_hdr_sz = sizeof(struct virtio_net_hdr);
1614

J
Jason Wang 已提交
1615 1616
		tun->filter_attached = false;
		tun->sndbuf = tfile->socket.sk->sk_sndbuf;
1617

J
Jason Wang 已提交
1618 1619
		spin_lock_init(&tun->lock);

1620 1621 1622
		err = security_tun_dev_alloc_security(&tun->security);
		if (err < 0)
			goto err_free_dev;
P
Paul Moore 已提交
1623

L
Linus Torvalds 已提交
1624
		tun_net_init(dev);
1625
		tun_flow_init(tun);
J
Jason Wang 已提交
1626

1627
		dev->hw_features = NETIF_F_SG | NETIF_F_FRAGLIST |
J
Jason Wang 已提交
1628 1629
				   TUN_USER_FEATURES | NETIF_F_HW_VLAN_CTAG_TX |
				   NETIF_F_HW_VLAN_STAG_TX;
1630
		dev->features = dev->hw_features;
J
Jason Wang 已提交
1631
		dev->vlan_features = dev->features;
1632

J
Jason Wang 已提交
1633
		INIT_LIST_HEAD(&tun->disabled);
1634 1635 1636 1637
		err = tun_attach(tun, file);
		if (err < 0)
			goto err_free_dev;

L
Linus Torvalds 已提交
1638 1639
		err = register_netdevice(tun->dev);
		if (err < 0)
J
Jason Wang 已提交
1640
			goto err_free_dev;
1641

1642 1643 1644
		if (device_create_file(&tun->dev->dev, &dev_attr_tun_flags) ||
		    device_create_file(&tun->dev->dev, &dev_attr_owner) ||
		    device_create_file(&tun->dev->dev, &dev_attr_group))
1645
			pr_err("Failed to create tun sysfs files\n");
L
Linus Torvalds 已提交
1646 1647
	}

1648 1649
	netif_carrier_on(tun->dev);

1650
	tun_debug(KERN_INFO, tun, "tun_set_iff\n");
L
Linus Torvalds 已提交
1651 1652 1653

	if (ifr->ifr_flags & IFF_NO_PI)
		tun->flags |= TUN_NO_PI;
1654 1655
	else
		tun->flags &= ~TUN_NO_PI;
L
Linus Torvalds 已提交
1656

1657 1658 1659
	/* This flag has no real effect.  We track the value for backwards
	 * compatibility.
	 */
L
Linus Torvalds 已提交
1660 1661
	if (ifr->ifr_flags & IFF_ONE_QUEUE)
		tun->flags |= TUN_ONE_QUEUE;
1662 1663
	else
		tun->flags &= ~TUN_ONE_QUEUE;
L
Linus Torvalds 已提交
1664

1665 1666 1667 1668 1669
	if (ifr->ifr_flags & IFF_VNET_HDR)
		tun->flags |= TUN_VNET_HDR;
	else
		tun->flags &= ~TUN_VNET_HDR;

J
Jason Wang 已提交
1670 1671 1672 1673 1674
	if (ifr->ifr_flags & IFF_MULTI_QUEUE)
		tun->flags |= TUN_TAP_MQ;
	else
		tun->flags &= ~TUN_TAP_MQ;

1675 1676 1677 1678
	/* Make sure persistent devices do not get stuck in
	 * xoff state.
	 */
	if (netif_running(tun->dev))
J
Jason Wang 已提交
1679
		netif_tx_wake_all_queues(tun->dev);
1680

L
Linus Torvalds 已提交
1681 1682 1683 1684 1685 1686 1687 1688
	strcpy(ifr->ifr_name, tun->dev->name);
	return 0;

 err_free_dev:
	free_netdev(dev);
	return err;
}

R
Rami Rosen 已提交
1689
static void tun_get_iff(struct net *net, struct tun_struct *tun,
1690
		       struct ifreq *ifr)
1691
{
1692
	tun_debug(KERN_INFO, tun, "tun_get_iff\n");
1693 1694 1695

	strcpy(ifr->ifr_name, tun->dev->name);

1696
	ifr->ifr_flags = tun_flags(tun);
1697 1698 1699

}

1700 1701
/* This is like a cut-down ethtool ops, except done via tun fd so no
 * privs required. */
1702
static int set_offload(struct tun_struct *tun, unsigned long arg)
1703
{
1704
	netdev_features_t features = 0;
1705 1706

	if (arg & TUN_F_CSUM) {
1707
		features |= NETIF_F_HW_CSUM;
1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720
		arg &= ~TUN_F_CSUM;

		if (arg & (TUN_F_TSO4|TUN_F_TSO6)) {
			if (arg & TUN_F_TSO_ECN) {
				features |= NETIF_F_TSO_ECN;
				arg &= ~TUN_F_TSO_ECN;
			}
			if (arg & TUN_F_TSO4)
				features |= NETIF_F_TSO;
			if (arg & TUN_F_TSO6)
				features |= NETIF_F_TSO6;
			arg &= ~(TUN_F_TSO4|TUN_F_TSO6);
		}
1721 1722 1723 1724 1725

		if (arg & TUN_F_UFO) {
			features |= NETIF_F_UFO;
			arg &= ~TUN_F_UFO;
		}
1726 1727 1728 1729 1730 1731 1732
	}

	/* This gives the user a way to test for new features in future by
	 * trying to set them. */
	if (arg)
		return -EINVAL;

1733 1734
	tun->set_features = features;
	netdev_update_features(tun->dev);
1735 1736 1737 1738

	return 0;
}

J
Jason Wang 已提交
1739 1740 1741 1742 1743 1744
static void tun_detach_filter(struct tun_struct *tun, int n)
{
	int i;
	struct tun_file *tfile;

	for (i = 0; i < n; i++) {
1745
		tfile = rtnl_dereference(tun->tfiles[i]);
J
Jason Wang 已提交
1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757
		sk_detach_filter(tfile->socket.sk);
	}

	tun->filter_attached = false;
}

static int tun_attach_filter(struct tun_struct *tun)
{
	int i, ret = 0;
	struct tun_file *tfile;

	for (i = 0; i < tun->numqueues; i++) {
1758
		tfile = rtnl_dereference(tun->tfiles[i]);
J
Jason Wang 已提交
1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775
		ret = sk_attach_filter(&tun->fprog, tfile->socket.sk);
		if (ret) {
			tun_detach_filter(tun, i);
			return ret;
		}
	}

	tun->filter_attached = true;
	return ret;
}

static void tun_set_sndbuf(struct tun_struct *tun)
{
	struct tun_file *tfile;
	int i;

	for (i = 0; i < tun->numqueues; i++) {
1776
		tfile = rtnl_dereference(tun->tfiles[i]);
J
Jason Wang 已提交
1777 1778 1779 1780
		tfile->socket.sk->sk_sndbuf = tun->sndbuf;
	}
}

1781 1782 1783 1784 1785 1786 1787 1788 1789
static int tun_set_queue(struct file *file, struct ifreq *ifr)
{
	struct tun_file *tfile = file->private_data;
	struct tun_struct *tun;
	int ret = 0;

	rtnl_lock();

	if (ifr->ifr_flags & IFF_ATTACH_QUEUE) {
J
Jason Wang 已提交
1790
		tun = tfile->detached;
1791
		if (!tun) {
1792
			ret = -EINVAL;
1793 1794 1795 1796 1797 1798
			goto unlock;
		}
		ret = security_tun_dev_attach_queue(tun->security);
		if (ret < 0)
			goto unlock;
		ret = tun_attach(tun, file);
J
Jason Wang 已提交
1799
	} else if (ifr->ifr_flags & IFF_DETACH_QUEUE) {
1800
		tun = rtnl_dereference(tfile->tun);
1801
		if (!tun || !(tun->flags & TUN_TAP_MQ) || tfile->detached)
J
Jason Wang 已提交
1802 1803 1804 1805
			ret = -EINVAL;
		else
			__tun_detach(tfile, false);
	} else
1806 1807
		ret = -EINVAL;

1808
unlock:
1809 1810 1811 1812
	rtnl_unlock();
	return ret;
}

1813 1814
static long __tun_chr_ioctl(struct file *file, unsigned int cmd,
			    unsigned long arg, int ifreq_len)
L
Linus Torvalds 已提交
1815
{
E
Eric W. Biederman 已提交
1816
	struct tun_file *tfile = file->private_data;
E
Eric W. Biederman 已提交
1817
	struct tun_struct *tun;
L
Linus Torvalds 已提交
1818 1819
	void __user* argp = (void __user*)arg;
	struct ifreq ifr;
1820 1821
	kuid_t owner;
	kgid_t group;
1822
	int sndbuf;
1823
	int vnet_hdr_sz;
1824
	unsigned int ifindex;
1825
	int ret;
L
Linus Torvalds 已提交
1826

1827
	if (cmd == TUNSETIFF || cmd == TUNSETQUEUE || _IOC_TYPE(cmd) == 0x89) {
1828
		if (copy_from_user(&ifr, argp, ifreq_len))
L
Linus Torvalds 已提交
1829
			return -EFAULT;
D
David S. Miller 已提交
1830
	} else {
1831
		memset(&ifr, 0, sizeof(ifr));
D
David S. Miller 已提交
1832
	}
E
Eric W. Biederman 已提交
1833 1834 1835 1836 1837
	if (cmd == TUNGETFEATURES) {
		/* Currently this just means: "what IFF flags are valid?".
		 * This is needed because we never checked for invalid flags on
		 * TUNSETIFF. */
		return put_user(IFF_TUN | IFF_TAP | IFF_NO_PI | IFF_ONE_QUEUE |
1838
				IFF_VNET_HDR | IFF_MULTI_QUEUE,
E
Eric W. Biederman 已提交
1839
				(unsigned int __user*)argp);
1840 1841
	} else if (cmd == TUNSETQUEUE)
		return tun_set_queue(file, &ifr);
E
Eric W. Biederman 已提交
1842

J
Jason Wang 已提交
1843
	ret = 0;
1844 1845
	rtnl_lock();

E
Eric W. Biederman 已提交
1846
	tun = __tun_get(tfile);
L
Linus Torvalds 已提交
1847 1848 1849
	if (cmd == TUNSETIFF && !tun) {
		ifr.ifr_name[IFNAMSIZ-1] = '\0';

1850
		ret = tun_set_iff(tfile->net, file, &ifr);
L
Linus Torvalds 已提交
1851

1852 1853
		if (ret)
			goto unlock;
L
Linus Torvalds 已提交
1854

1855
		if (copy_to_user(argp, &ifr, ifreq_len))
1856 1857
			ret = -EFAULT;
		goto unlock;
L
Linus Torvalds 已提交
1858
	}
1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871
	if (cmd == TUNSETIFINDEX) {
		ret = -EPERM;
		if (tun)
			goto unlock;

		ret = -EFAULT;
		if (copy_from_user(&ifindex, argp, sizeof(ifindex)))
			goto unlock;

		ret = 0;
		tfile->ifindex = ifindex;
		goto unlock;
	}
L
Linus Torvalds 已提交
1872

1873
	ret = -EBADFD;
L
Linus Torvalds 已提交
1874
	if (!tun)
1875
		goto unlock;
L
Linus Torvalds 已提交
1876

1877
	tun_debug(KERN_INFO, tun, "tun_chr_ioctl cmd %u\n", cmd);
L
Linus Torvalds 已提交
1878

E
Eric W. Biederman 已提交
1879
	ret = 0;
L
Linus Torvalds 已提交
1880
	switch (cmd) {
1881
	case TUNGETIFF:
R
Rami Rosen 已提交
1882
		tun_get_iff(current->nsproxy->net_ns, tun, &ifr);
1883

1884 1885 1886
		if (tfile->detached)
			ifr.ifr_flags |= IFF_DETACH_QUEUE;

1887
		if (copy_to_user(argp, &ifr, ifreq_len))
E
Eric W. Biederman 已提交
1888
			ret = -EFAULT;
1889 1890
		break;

L
Linus Torvalds 已提交
1891 1892 1893
	case TUNSETNOCSUM:
		/* Disable/Enable checksum */

1894 1895
		/* [unimplemented] */
		tun_debug(KERN_INFO, tun, "ignored: set checksum %s\n",
1896
			  arg ? "disabled" : "enabled");
L
Linus Torvalds 已提交
1897 1898 1899
		break;

	case TUNSETPERSIST:
J
Jason Wang 已提交
1900 1901 1902
		/* Disable/Enable persist mode. Keep an extra reference to the
		 * module to prevent the module being unprobed.
		 */
J
Jason Wang 已提交
1903
		if (arg && !(tun->flags & TUN_PERSIST)) {
L
Linus Torvalds 已提交
1904
			tun->flags |= TUN_PERSIST;
J
Jason Wang 已提交
1905
			__module_get(THIS_MODULE);
J
Jason Wang 已提交
1906 1907
		}
		if (!arg && (tun->flags & TUN_PERSIST)) {
L
Linus Torvalds 已提交
1908
			tun->flags &= ~TUN_PERSIST;
J
Jason Wang 已提交
1909 1910
			module_put(THIS_MODULE);
		}
L
Linus Torvalds 已提交
1911

1912 1913
		tun_debug(KERN_INFO, tun, "persist %s\n",
			  arg ? "enabled" : "disabled");
L
Linus Torvalds 已提交
1914 1915 1916 1917
		break;

	case TUNSETOWNER:
		/* Set owner of the device */
1918 1919 1920 1921 1922 1923
		owner = make_kuid(current_user_ns(), arg);
		if (!uid_valid(owner)) {
			ret = -EINVAL;
			break;
		}
		tun->owner = owner;
1924
		tun_debug(KERN_INFO, tun, "owner set to %u\n",
1925
			  from_kuid(&init_user_ns, tun->owner));
L
Linus Torvalds 已提交
1926 1927
		break;

1928 1929
	case TUNSETGROUP:
		/* Set group of the device */
1930 1931 1932 1933 1934 1935
		group = make_kgid(current_user_ns(), arg);
		if (!gid_valid(group)) {
			ret = -EINVAL;
			break;
		}
		tun->group = group;
1936
		tun_debug(KERN_INFO, tun, "group set to %u\n",
1937
			  from_kgid(&init_user_ns, tun->group));
1938 1939
		break;

1940 1941 1942
	case TUNSETLINK:
		/* Only allow setting the type when the interface is down */
		if (tun->dev->flags & IFF_UP) {
1943 1944
			tun_debug(KERN_INFO, tun,
				  "Linktype set failed because interface is up\n");
1945
			ret = -EBUSY;
1946 1947
		} else {
			tun->dev->type = (int) arg;
1948 1949
			tun_debug(KERN_INFO, tun, "linktype set to %d\n",
				  tun->dev->type);
1950
			ret = 0;
1951
		}
E
Eric W. Biederman 已提交
1952
		break;
1953

L
Linus Torvalds 已提交
1954 1955 1956 1957 1958
#ifdef TUN_DEBUG
	case TUNSETDEBUG:
		tun->debug = arg;
		break;
#endif
1959
	case TUNSETOFFLOAD:
1960
		ret = set_offload(tun, arg);
E
Eric W. Biederman 已提交
1961
		break;
1962

1963 1964
	case TUNSETTXFILTER:
		/* Can be set only for TAPs */
E
Eric W. Biederman 已提交
1965
		ret = -EINVAL;
1966
		if ((tun->flags & TUN_TYPE_MASK) != TUN_TAP_DEV)
E
Eric W. Biederman 已提交
1967
			break;
H
Harvey Harrison 已提交
1968
		ret = update_filter(&tun->txflt, (void __user *)arg);
E
Eric W. Biederman 已提交
1969
		break;
L
Linus Torvalds 已提交
1970 1971

	case SIOCGIFHWADDR:
1972
		/* Get hw address */
1973 1974
		memcpy(ifr.ifr_hwaddr.sa_data, tun->dev->dev_addr, ETH_ALEN);
		ifr.ifr_hwaddr.sa_family = tun->dev->type;
1975
		if (copy_to_user(argp, &ifr, ifreq_len))
E
Eric W. Biederman 已提交
1976 1977
			ret = -EFAULT;
		break;
L
Linus Torvalds 已提交
1978 1979

	case SIOCSIFHWADDR:
1980
		/* Set hw address */
1981 1982
		tun_debug(KERN_DEBUG, tun, "set hw address: %pM\n",
			  ifr.ifr_hwaddr.sa_data);
1983 1984

		ret = dev_set_mac_address(tun->dev, &ifr.ifr_hwaddr);
E
Eric W. Biederman 已提交
1985
		break;
1986 1987

	case TUNGETSNDBUF:
J
Jason Wang 已提交
1988
		sndbuf = tfile->socket.sk->sk_sndbuf;
1989 1990 1991 1992 1993 1994 1995 1996 1997 1998
		if (copy_to_user(argp, &sndbuf, sizeof(sndbuf)))
			ret = -EFAULT;
		break;

	case TUNSETSNDBUF:
		if (copy_from_user(&sndbuf, argp, sizeof(sndbuf))) {
			ret = -EFAULT;
			break;
		}

J
Jason Wang 已提交
1999 2000
		tun->sndbuf = sndbuf;
		tun_set_sndbuf(tun);
2001 2002
		break;

2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021
	case TUNGETVNETHDRSZ:
		vnet_hdr_sz = tun->vnet_hdr_sz;
		if (copy_to_user(argp, &vnet_hdr_sz, sizeof(vnet_hdr_sz)))
			ret = -EFAULT;
		break;

	case TUNSETVNETHDRSZ:
		if (copy_from_user(&vnet_hdr_sz, argp, sizeof(vnet_hdr_sz))) {
			ret = -EFAULT;
			break;
		}
		if (vnet_hdr_sz < (int)sizeof(struct virtio_net_hdr)) {
			ret = -EINVAL;
			break;
		}

		tun->vnet_hdr_sz = vnet_hdr_sz;
		break;

M
Michael S. Tsirkin 已提交
2022 2023 2024 2025 2026 2027
	case TUNATTACHFILTER:
		/* Can be set only for TAPs */
		ret = -EINVAL;
		if ((tun->flags & TUN_TYPE_MASK) != TUN_TAP_DEV)
			break;
		ret = -EFAULT;
J
Jason Wang 已提交
2028
		if (copy_from_user(&tun->fprog, argp, sizeof(tun->fprog)))
M
Michael S. Tsirkin 已提交
2029 2030
			break;

J
Jason Wang 已提交
2031
		ret = tun_attach_filter(tun);
M
Michael S. Tsirkin 已提交
2032 2033 2034 2035 2036 2037 2038
		break;

	case TUNDETACHFILTER:
		/* Can be set only for TAPs */
		ret = -EINVAL;
		if ((tun->flags & TUN_TYPE_MASK) != TUN_TAP_DEV)
			break;
J
Jason Wang 已提交
2039 2040
		ret = 0;
		tun_detach_filter(tun, tun->numqueues);
M
Michael S. Tsirkin 已提交
2041 2042
		break;

L
Linus Torvalds 已提交
2043
	default:
E
Eric W. Biederman 已提交
2044 2045
		ret = -EINVAL;
		break;
2046
	}
L
Linus Torvalds 已提交
2047

2048 2049 2050 2051
unlock:
	rtnl_unlock();
	if (tun)
		tun_put(tun);
E
Eric W. Biederman 已提交
2052
	return ret;
L
Linus Torvalds 已提交
2053 2054
}

2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089
static long tun_chr_ioctl(struct file *file,
			  unsigned int cmd, unsigned long arg)
{
	return __tun_chr_ioctl(file, cmd, arg, sizeof (struct ifreq));
}

#ifdef CONFIG_COMPAT
static long tun_chr_compat_ioctl(struct file *file,
			 unsigned int cmd, unsigned long arg)
{
	switch (cmd) {
	case TUNSETIFF:
	case TUNGETIFF:
	case TUNSETTXFILTER:
	case TUNGETSNDBUF:
	case TUNSETSNDBUF:
	case SIOCGIFHWADDR:
	case SIOCSIFHWADDR:
		arg = (unsigned long)compat_ptr(arg);
		break;
	default:
		arg = (compat_ulong_t)arg;
		break;
	}

	/*
	 * compat_ifreq is shorter than ifreq, so we must not access beyond
	 * the end of that structure. All fields that are used in this
	 * driver are compatible though, we don't need to convert the
	 * contents.
	 */
	return __tun_chr_ioctl(file, cmd, arg, sizeof(struct compat_ifreq));
}
#endif /* CONFIG_COMPAT */

L
Linus Torvalds 已提交
2090 2091
static int tun_chr_fasync(int fd, struct file *file, int on)
{
J
Jason Wang 已提交
2092
	struct tun_file *tfile = file->private_data;
L
Linus Torvalds 已提交
2093 2094
	int ret;

J
Jason Wang 已提交
2095
	if ((ret = fasync_helper(fd, file, on, &tfile->fasync)) < 0)
J
Jonathan Corbet 已提交
2096
		goto out;
2097

L
Linus Torvalds 已提交
2098
	if (on) {
2099
		ret = __f_setown(file, task_pid(current), PIDTYPE_PID, 0);
L
Linus Torvalds 已提交
2100
		if (ret)
J
Jonathan Corbet 已提交
2101
			goto out;
J
Jason Wang 已提交
2102
		tfile->flags |= TUN_FASYNC;
2103
	} else
J
Jason Wang 已提交
2104
		tfile->flags &= ~TUN_FASYNC;
J
Jonathan Corbet 已提交
2105 2106 2107
	ret = 0;
out:
	return ret;
L
Linus Torvalds 已提交
2108 2109 2110 2111
}

static int tun_chr_open(struct inode *inode, struct file * file)
{
E
Eric W. Biederman 已提交
2112
	struct tun_file *tfile;
T
Thomas Gleixner 已提交
2113

2114
	DBG1(KERN_INFO, "tunX: tun_chr_open\n");
E
Eric W. Biederman 已提交
2115

J
Jason Wang 已提交
2116 2117
	tfile = (struct tun_file *)sk_alloc(&init_net, AF_UNSPEC, GFP_KERNEL,
					    &tun_proto);
E
Eric W. Biederman 已提交
2118 2119
	if (!tfile)
		return -ENOMEM;
2120
	rcu_assign_pointer(tfile->tun, NULL);
E
Eric W. Biederman 已提交
2121
	tfile->net = get_net(current->nsproxy->net_ns);
J
Jason Wang 已提交
2122
	tfile->flags = 0;
2123
	tfile->ifindex = 0;
J
Jason Wang 已提交
2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136

	rcu_assign_pointer(tfile->socket.wq, &tfile->wq);
	init_waitqueue_head(&tfile->wq.wait);

	tfile->socket.file = file;
	tfile->socket.ops = &tun_socket_ops;

	sock_init_data(&tfile->socket, &tfile->sk);
	sk_change_net(&tfile->sk, tfile->net);

	tfile->sk.sk_write_space = tun_sock_write_space;
	tfile->sk.sk_sndbuf = INT_MAX;

E
Eric W. Biederman 已提交
2137
	file->private_data = tfile;
J
Jason Wang 已提交
2138
	set_bit(SOCK_EXTERNALLY_ALLOCATED, &tfile->socket.flags);
J
Jason Wang 已提交
2139
	INIT_LIST_HEAD(&tfile->next);
J
Jason Wang 已提交
2140

2141 2142
	sock_set_flag(&tfile->sk, SOCK_ZEROCOPY);

L
Linus Torvalds 已提交
2143 2144 2145 2146 2147
	return 0;
}

static int tun_chr_close(struct inode *inode, struct file *file)
{
E
Eric W. Biederman 已提交
2148
	struct tun_file *tfile = file->private_data;
J
Jason Wang 已提交
2149
	struct net *net = tfile->net;
L
Linus Torvalds 已提交
2150

J
Jason Wang 已提交
2151
	tun_detach(tfile, true);
J
Jason Wang 已提交
2152
	put_net(net);
L
Linus Torvalds 已提交
2153 2154 2155 2156

	return 0;
}

2157
static const struct file_operations tun_fops = {
2158
	.owner	= THIS_MODULE,
L
Linus Torvalds 已提交
2159
	.llseek = no_llseek,
2160 2161 2162 2163
	.read  = do_sync_read,
	.aio_read  = tun_chr_aio_read,
	.write = do_sync_write,
	.aio_write = tun_chr_aio_write,
L
Linus Torvalds 已提交
2164
	.poll	= tun_chr_poll,
2165 2166 2167 2168
	.unlocked_ioctl	= tun_chr_ioctl,
#ifdef CONFIG_COMPAT
	.compat_ioctl = tun_chr_compat_ioctl,
#endif
L
Linus Torvalds 已提交
2169 2170
	.open	= tun_chr_open,
	.release = tun_chr_close,
2171
	.fasync = tun_chr_fasync
L
Linus Torvalds 已提交
2172 2173 2174 2175 2176
};

static struct miscdevice tun_miscdev = {
	.minor = TUN_MINOR,
	.name = "tun",
2177
	.nodename = "net/tun",
L
Linus Torvalds 已提交
2178 2179 2180 2181 2182 2183 2184 2185 2186
	.fops = &tun_fops,
};

/* ethtool interface */

static int tun_get_settings(struct net_device *dev, struct ethtool_cmd *cmd)
{
	cmd->supported		= 0;
	cmd->advertising	= 0;
2187
	ethtool_cmd_speed_set(cmd, SPEED_10);
L
Linus Torvalds 已提交
2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201
	cmd->duplex		= DUPLEX_FULL;
	cmd->port		= PORT_TP;
	cmd->phy_address	= 0;
	cmd->transceiver	= XCVR_INTERNAL;
	cmd->autoneg		= AUTONEG_DISABLE;
	cmd->maxtxpkt		= 0;
	cmd->maxrxpkt		= 0;
	return 0;
}

static void tun_get_drvinfo(struct net_device *dev, struct ethtool_drvinfo *info)
{
	struct tun_struct *tun = netdev_priv(dev);

2202 2203
	strlcpy(info->driver, DRV_NAME, sizeof(info->driver));
	strlcpy(info->version, DRV_VERSION, sizeof(info->version));
L
Linus Torvalds 已提交
2204 2205 2206

	switch (tun->flags & TUN_TYPE_MASK) {
	case TUN_TUN_DEV:
2207
		strlcpy(info->bus_info, "tun", sizeof(info->bus_info));
L
Linus Torvalds 已提交
2208 2209
		break;
	case TUN_TAP_DEV:
2210
		strlcpy(info->bus_info, "tap", sizeof(info->bus_info));
L
Linus Torvalds 已提交
2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232
		break;
	}
}

static u32 tun_get_msglevel(struct net_device *dev)
{
#ifdef TUN_DEBUG
	struct tun_struct *tun = netdev_priv(dev);
	return tun->debug;
#else
	return -EOPNOTSUPP;
#endif
}

static void tun_set_msglevel(struct net_device *dev, u32 value)
{
#ifdef TUN_DEBUG
	struct tun_struct *tun = netdev_priv(dev);
	tun->debug = value;
#endif
}

2233
static const struct ethtool_ops tun_ethtool_ops = {
L
Linus Torvalds 已提交
2234 2235 2236 2237
	.get_settings	= tun_get_settings,
	.get_drvinfo	= tun_get_drvinfo,
	.get_msglevel	= tun_get_msglevel,
	.set_msglevel	= tun_set_msglevel,
2238
	.get_link	= ethtool_op_get_link,
2239
	.get_ts_info	= ethtool_op_get_ts_info,
L
Linus Torvalds 已提交
2240 2241
};

2242

L
Linus Torvalds 已提交
2243 2244 2245 2246
static int __init tun_init(void)
{
	int ret = 0;

2247 2248
	pr_info("%s, %s\n", DRV_DESCRIPTION, DRV_VERSION);
	pr_info("%s\n", DRV_COPYRIGHT);
L
Linus Torvalds 已提交
2249

2250
	ret = rtnl_link_register(&tun_link_ops);
2251
	if (ret) {
2252
		pr_err("Can't register link_ops\n");
2253
		goto err_linkops;
2254 2255
	}

L
Linus Torvalds 已提交
2256
	ret = misc_register(&tun_miscdev);
2257
	if (ret) {
2258
		pr_err("Can't register misc device %d\n", TUN_MINOR);
2259 2260
		goto err_misc;
	}
2261
	return  0;
2262
err_misc:
2263 2264
	rtnl_link_unregister(&tun_link_ops);
err_linkops:
L
Linus Torvalds 已提交
2265 2266 2267 2268 2269
	return ret;
}

static void tun_cleanup(void)
{
2270
	misc_deregister(&tun_miscdev);
2271
	rtnl_link_unregister(&tun_link_ops);
L
Linus Torvalds 已提交
2272 2273
}

2274 2275 2276 2277 2278 2279
/* 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 *tun_get_socket(struct file *file)
{
2280
	struct tun_file *tfile;
2281 2282
	if (file->f_op != &tun_fops)
		return ERR_PTR(-EINVAL);
2283 2284
	tfile = file->private_data;
	if (!tfile)
2285
		return ERR_PTR(-EBADFD);
J
Jason Wang 已提交
2286
	return &tfile->socket;
2287 2288 2289
}
EXPORT_SYMBOL_GPL(tun_get_socket);

L
Linus Torvalds 已提交
2290 2291 2292 2293 2294 2295
module_init(tun_init);
module_exit(tun_cleanup);
MODULE_DESCRIPTION(DRV_DESCRIPTION);
MODULE_AUTHOR(DRV_COPYRIGHT);
MODULE_LICENSE("GPL");
MODULE_ALIAS_MISCDEV(TUN_MINOR);
2296
MODULE_ALIAS("devname:net/tun");