tun.c 55.5 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/ipv6.h>
69
#include <net/net_namespace.h>
70
#include <net/netns/generic.h>
71
#include <net/rtnetlink.h>
72
#include <net/sock.h>
73
#include <linux/seq_file.h>
H
Herbert Xu 已提交
74
#include <linux/uio.h>
L
Linus Torvalds 已提交
75 76 77

#include <asm/uaccess.h>

78 79 80
/* Uncomment to enable debugging */
/* #define TUN_DEBUG 1 */

L
Linus Torvalds 已提交
81 82
#ifdef TUN_DEBUG
static int debug;
83

84 85 86 87 88 89 90 91 92 93
#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)
94
#else
95 96 97 98 99 100 101 102 103 104
#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)
105 106
#endif

107 108
#define GOODCOPY_LEN 128

109 110 111 112 113 114 115
#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];
};

S
stephen hemminger 已提交
116
/* DEFAULT_MAX_NUM_RSS_QUEUES were chosen to let the rx/tx queues allocated for
117 118 119
 * 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
120
#define MAX_TAP_FLOWS  4096
J
Jason Wang 已提交
121

J
Jason Wang 已提交
122 123
#define TUN_FLOW_EXPIRE (3 * HZ)

J
Jason Wang 已提交
124
/* A tun_file connects an open character device to a tuntap netdevice. It
S
stephen hemminger 已提交
125
 * also contains all socket related structures (except sock_fprog and tap_filter)
J
Jason Wang 已提交
126 127
 * 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
128
 * netdevice not for a specific queue (at least I didn't see the requirement for
J
Jason Wang 已提交
129
 * this).
130 131
 *
 * RCU usage:
R
Rami Rosen 已提交
132
 * The tun_file and tun_struct are loosely coupled, the pointer from one to the
133
 * other can only be read while rcu_read_lock or rtnl_lock is held.
J
Jason Wang 已提交
134
 */
E
Eric W. Biederman 已提交
135
struct tun_file {
J
Jason Wang 已提交
136 137 138
	struct sock sk;
	struct socket socket;
	struct socket_wq wq;
139
	struct tun_struct __rcu *tun;
E
Eric W. Biederman 已提交
140
	struct net *net;
J
Jason Wang 已提交
141 142 143
	struct fasync_struct *fasync;
	/* only used for fasnyc */
	unsigned int flags;
144 145 146 147
	union {
		u16 queue_index;
		unsigned int ifindex;
	};
J
Jason Wang 已提交
148 149
	struct list_head next;
	struct tun_struct *detached;
E
Eric W. Biederman 已提交
150 151
};

J
Jason Wang 已提交
152 153 154 155 156 157
struct tun_flow_entry {
	struct hlist_node hash_link;
	struct rcu_head rcu;
	struct tun_struct *tun;

	u32 rxhash;
158
	u32 rps_rxhash;
J
Jason Wang 已提交
159 160 161 162 163 164
	int queue_index;
	unsigned long updated;
};

#define TUN_NUM_FLOW_ENTRIES 1024

J
Jason Wang 已提交
165
/* Since the socket were moved to tun_file, to preserve the behavior of persist
R
Rami Rosen 已提交
166
 * device, socket filter, sndbuf and vnet header size were restore when the
J
Jason Wang 已提交
167 168
 * file were attached to a persist device.
 */
169
struct tun_struct {
J
Jason Wang 已提交
170 171
	struct tun_file __rcu	*tfiles[MAX_TAP_QUEUES];
	unsigned int            numqueues;
172
	unsigned int 		flags;
173 174
	kuid_t			owner;
	kgid_t			group;
175 176

	struct net_device	*dev;
177
	netdev_features_t	set_features;
178
#define TUN_USER_FEATURES (NETIF_F_HW_CSUM|NETIF_F_TSO_ECN|NETIF_F_TSO| \
179
			  NETIF_F_TSO6)
180 181

	int			vnet_hdr_sz;
J
Jason Wang 已提交
182 183 184 185 186
	int			sndbuf;
	struct tap_filter	txflt;
	struct sock_fprog	fprog;
	/* protected by rtnl lock */
	bool			filter_attached;
187 188
#ifdef TUN_DEBUG
	int debug;
L
Linus Torvalds 已提交
189
#endif
J
Jason Wang 已提交
190 191 192 193
	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
194 195
	unsigned int numdisabled;
	struct list_head disabled;
196
	void *security;
197
	u32 flow_count;
198
};
L
Linus Torvalds 已提交
199

J
Jason Wang 已提交
200 201 202 203 204 205 206 207 208
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;

209
	hlist_for_each_entry_rcu(e, head, hash_link) {
J
Jason Wang 已提交
210 211 212 213 214 215 216 217 218 219
		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)
{
220 221
	struct tun_flow_entry *e = kmalloc(sizeof(*e), GFP_ATOMIC);

J
Jason Wang 已提交
222 223 224 225 226
	if (e) {
		tun_debug(KERN_INFO, tun, "create flow: hash %u index %u\n",
			  rxhash, queue_index);
		e->updated = jiffies;
		e->rxhash = rxhash;
227
		e->rps_rxhash = 0;
J
Jason Wang 已提交
228 229 230
		e->queue_index = queue_index;
		e->tun = tun;
		hlist_add_head_rcu(&e->hash_link, head);
231
		++tun->flow_count;
J
Jason Wang 已提交
232 233 234 235 236 237 238 239
	}
	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);
240
	sock_rps_reset_flow_hash(e->rps_rxhash);
J
Jason Wang 已提交
241
	hlist_del_rcu(&e->hash_link);
242
	kfree_rcu(e, rcu);
243
	--tun->flow_count;
J
Jason Wang 已提交
244 245 246 247 248 249 250 251 252
}

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;
253
		struct hlist_node *n;
J
Jason Wang 已提交
254

255
		hlist_for_each_entry_safe(e, n, &tun->flows[i], hash_link)
J
Jason Wang 已提交
256 257 258 259 260 261 262 263 264 265 266 267
			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;
268
		struct hlist_node *n;
J
Jason Wang 已提交
269

270
		hlist_for_each_entry_safe(e, n, &tun->flows[i], hash_link) {
J
Jason Wang 已提交
271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290
			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;
291
		struct hlist_node *n;
J
Jason Wang 已提交
292

293
		hlist_for_each_entry_safe(e, n, &tun->flows[i], hash_link) {
J
Jason Wang 已提交
294 295 296 297 298 299 300 301 302 303 304 305 306 307 308
			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);
}

309
static void tun_flow_update(struct tun_struct *tun, u32 rxhash,
310
			    struct tun_file *tfile)
J
Jason Wang 已提交
311 312 313 314
{
	struct hlist_head *head;
	struct tun_flow_entry *e;
	unsigned long delay = tun->ageing_time;
315
	u16 queue_index = tfile->queue_index;
J
Jason Wang 已提交
316 317 318 319 320 321 322 323

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

	rcu_read_lock();

324 325 326
	/* 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
327 328 329 330 331 332 333
		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;
334
		sock_rps_record_flow_hash(e->rps_rxhash);
J
Jason Wang 已提交
335 336
	} else {
		spin_lock_bh(&tun->lock);
337 338
		if (!tun_flow_find(head, rxhash) &&
		    tun->flow_count < MAX_TAP_FLOWS)
J
Jason Wang 已提交
339 340 341 342 343 344 345 346 347 348 349 350
			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();
}

351 352 353 354 355 356 357 358 359 360 361 362
/**
 * Save the hash received in the stack receive path and update the
 * flow_hash table accordingly.
 */
static inline void tun_flow_save_rps_rxhash(struct tun_flow_entry *e, u32 hash)
{
	if (unlikely(e->rps_rxhash != hash)) {
		sock_rps_reset_flow_hash(e->rps_rxhash);
		e->rps_rxhash = hash;
	}
}

J
Jason Wang 已提交
363
/* We try to identify a flow through its rxhash first. The reason that
S
stephen hemminger 已提交
364
 * we do not check rxq no. is because some cards(e.g 82599), chooses
J
Jason Wang 已提交
365 366 367 368 369
 * 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.
 */
370
static u16 tun_select_queue(struct net_device *dev, struct sk_buff *skb,
371
			    void *accel_priv, select_queue_fallback_t fallback)
J
Jason Wang 已提交
372 373
{
	struct tun_struct *tun = netdev_priv(dev);
J
Jason Wang 已提交
374
	struct tun_flow_entry *e;
J
Jason Wang 已提交
375 376 377 378
	u32 txq = 0;
	u32 numqueues = 0;

	rcu_read_lock();
379
	numqueues = ACCESS_ONCE(tun->numqueues);
J
Jason Wang 已提交
380

381
	txq = skb_get_hash(skb);
J
Jason Wang 已提交
382
	if (txq) {
J
Jason Wang 已提交
383
		e = tun_flow_find(&tun->flows[tun_hashfn(txq)], txq);
384 385
		if (e) {
			tun_flow_save_rps_rxhash(e, txq);
386
			txq = e->queue_index;
387
		} else
J
Jason Wang 已提交
388 389
			/* use multiply and shift instead of expensive divide */
			txq = ((u64)txq * numqueues) >> 32;
J
Jason Wang 已提交
390 391 392 393 394 395 396 397 398 399
	} 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;
}

400 401 402
static inline bool tun_not_capable(struct tun_struct *tun)
{
	const struct cred *cred = current_cred();
403
	struct net *net = dev_net(tun->dev);
404 405 406

	return ((uid_valid(tun->owner) && !uid_eq(cred->euid, tun->owner)) ||
		  (gid_valid(tun->group) && !in_egroup_p(tun->group))) &&
407
		!ns_capable(net->user_ns, CAP_NET_ADMIN);
408 409
}

J
Jason Wang 已提交
410 411 412 413 414 415
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
416 417 418 419 420 421 422
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
423
static struct tun_struct *tun_enable_queue(struct tun_file *tfile)
J
Jason Wang 已提交
424 425 426 427 428 429 430 431 432
{
	struct tun_struct *tun = tfile->detached;

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

433 434 435 436 437 438
static void tun_queue_purge(struct tun_file *tfile)
{
	skb_queue_purge(&tfile->sk.sk_receive_queue);
	skb_queue_purge(&tfile->sk.sk_error_queue);
}

J
Jason Wang 已提交
439 440 441 442 443
static void __tun_detach(struct tun_file *tfile, bool clean)
{
	struct tun_file *ntfile;
	struct tun_struct *tun;

444 445
	tun = rtnl_dereference(tfile->tun);

446
	if (tun && !tfile->detached) {
J
Jason Wang 已提交
447 448 449 450 451
		u16 index = tfile->queue_index;
		BUG_ON(index >= tun->numqueues);

		rcu_assign_pointer(tun->tfiles[index],
				   tun->tfiles[tun->numqueues - 1]);
452
		ntfile = rtnl_dereference(tun->tfiles[index]);
J
Jason Wang 已提交
453 454 455
		ntfile->queue_index = index;

		--tun->numqueues;
456
		if (clean) {
457
			RCU_INIT_POINTER(tfile->tun, NULL);
J
Jason Wang 已提交
458
			sock_put(&tfile->sk);
459
		} else
J
Jason Wang 已提交
460
			tun_disable_queue(tun, tfile);
J
Jason Wang 已提交
461 462

		synchronize_net();
J
Jason Wang 已提交
463
		tun_flow_delete_by_queue(tun, tun->numqueues + 1);
J
Jason Wang 已提交
464
		/* Drop read queue */
465
		tun_queue_purge(tfile);
J
Jason Wang 已提交
466
		tun_set_real_num_queues(tun);
J
Jason Wang 已提交
467
	} else if (tfile->detached && clean) {
J
Jason Wang 已提交
468
		tun = tun_enable_queue(tfile);
J
Jason Wang 已提交
469 470
		sock_put(&tfile->sk);
	}
J
Jason Wang 已提交
471 472

	if (clean) {
473 474 475 476 477
		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
478
				unregister_netdevice(tun->dev);
479
		}
J
Jason Wang 已提交
480

J
Jason Wang 已提交
481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496
		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
497
	struct tun_file *tfile, *tmp;
J
Jason Wang 已提交
498 499 500
	int i, n = tun->numqueues;

	for (i = 0; i < n; i++) {
501
		tfile = rtnl_dereference(tun->tfiles[i]);
J
Jason Wang 已提交
502
		BUG_ON(!tfile);
503
		tfile->socket.sk->sk_data_ready(tfile->socket.sk);
504
		RCU_INIT_POINTER(tfile->tun, NULL);
J
Jason Wang 已提交
505 506
		--tun->numqueues;
	}
507
	list_for_each_entry(tfile, &tun->disabled, next) {
508
		tfile->socket.sk->sk_data_ready(tfile->socket.sk);
509
		RCU_INIT_POINTER(tfile->tun, NULL);
510
	}
J
Jason Wang 已提交
511 512 513 514
	BUG_ON(tun->numqueues != 0);

	synchronize_net();
	for (i = 0; i < n; i++) {
515
		tfile = rtnl_dereference(tun->tfiles[i]);
J
Jason Wang 已提交
516
		/* Drop read queue */
517
		tun_queue_purge(tfile);
J
Jason Wang 已提交
518 519
		sock_put(&tfile->sk);
	}
J
Jason Wang 已提交
520 521
	list_for_each_entry_safe(tfile, tmp, &tun->disabled, next) {
		tun_enable_queue(tfile);
522
		tun_queue_purge(tfile);
J
Jason Wang 已提交
523 524 525
		sock_put(&tfile->sk);
	}
	BUG_ON(tun->numdisabled != 0);
J
Jason Wang 已提交
526 527 528

	if (tun->flags & TUN_PERSIST)
		module_put(THIS_MODULE);
J
Jason Wang 已提交
529 530
}

531
static int tun_attach(struct tun_struct *tun, struct file *file, bool skip_filter)
532
{
E
Eric W. Biederman 已提交
533
	struct tun_file *tfile = file->private_data;
534
	int err;
535

536 537 538 539
	err = security_tun_dev_attach(tfile->socket.sk, tun->security);
	if (err < 0)
		goto out;

540
	err = -EINVAL;
541
	if (rtnl_dereference(tfile->tun) && !tfile->detached)
542 543 544
		goto out;

	err = -EBUSY;
J
Jason Wang 已提交
545 546 547 548
	if (!(tun->flags & TUN_TAP_MQ) && tun->numqueues == 1)
		goto out;

	err = -E2BIG;
J
Jason Wang 已提交
549 550
	if (!tfile->detached &&
	    tun->numqueues + tun->numdisabled == MAX_TAP_QUEUES)
551 552 553
		goto out;

	err = 0;
J
Jason Wang 已提交
554

S
stephen hemminger 已提交
555
	/* Re-attach the filter to persist device */
556
	if (!skip_filter && (tun->filter_attached == true)) {
J
Jason Wang 已提交
557 558 559 560
		err = sk_attach_filter(&tun->fprog, tfile->socket.sk);
		if (!err)
			goto out;
	}
J
Jason Wang 已提交
561
	tfile->queue_index = tun->numqueues;
562
	rcu_assign_pointer(tfile->tun, tun);
J
Jason Wang 已提交
563 564
	rcu_assign_pointer(tun->tfiles[tun->numqueues], tfile);
	tun->numqueues++;
565

J
Jason Wang 已提交
566 567 568 569 570
	if (tfile->detached)
		tun_enable_queue(tfile);
	else
		sock_hold(&tfile->sk);

J
Jason Wang 已提交
571
	tun_set_real_num_queues(tun);
572

J
Jason Wang 已提交
573 574 575 576 577 578
	/* device is allowed to go away first, so no need to hold extra
	 * refcnt.
	 */

out:
	return err;
E
Eric W. Biederman 已提交
579 580 581 582
}

static struct tun_struct *__tun_get(struct tun_file *tfile)
{
583
	struct tun_struct *tun;
584

585 586 587 588 589
	rcu_read_lock();
	tun = rcu_dereference(tfile->tun);
	if (tun)
		dev_hold(tun->dev);
	rcu_read_unlock();
590 591

	return tun;
E
Eric W. Biederman 已提交
592 593 594 595 596 597 598 599 600
}

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

static void tun_put(struct tun_struct *tun)
{
601
	dev_put(tun->dev);
E
Eric W. Biederman 已提交
602 603
}

604
/* TAP filtering */
605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653
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;

654 655
	/* Remaining multicast addresses are hashed,
	 * unicast will leave the filter disabled. */
656
	memset(filter->mask, 0, sizeof(filter->mask));
657 658 659 660 661
	for (; n < uf.count; n++) {
		if (!is_multicast_ether_addr(addr[n].u)) {
			err = 0; /* no filter */
			goto done;
		}
662
		addr_hash_set(filter->mask, addr[n].u);
663
	}
664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691

	/* 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++)
692
		if (ether_addr_equal(eh->h_dest, filter->addr[i]))
693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713
			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
714 715
/* Network device part of the driver */

716
static const struct ethtool_ops tun_ethtool_ops;
L
Linus Torvalds 已提交
717

718 719 720
/* Net device detach from fd. */
static void tun_net_uninit(struct net_device *dev)
{
J
Jason Wang 已提交
721
	tun_detach_all(dev);
722 723
}

L
Linus Torvalds 已提交
724 725 726
/* Net device open. */
static int tun_net_open(struct net_device *dev)
{
J
Jason Wang 已提交
727
	netif_tx_start_all_queues(dev);
L
Linus Torvalds 已提交
728 729 730 731 732 733
	return 0;
}

/* Net device close. */
static int tun_net_close(struct net_device *dev)
{
J
Jason Wang 已提交
734
	netif_tx_stop_all_queues(dev);
L
Linus Torvalds 已提交
735 736 737 738
	return 0;
}

/* Net device start xmit */
739
static netdev_tx_t tun_net_xmit(struct sk_buff *skb, struct net_device *dev)
L
Linus Torvalds 已提交
740 741
{
	struct tun_struct *tun = netdev_priv(dev);
J
Jason Wang 已提交
742
	int txq = skb->queue_mapping;
743
	struct tun_file *tfile;
744
	u32 numqueues = 0;
L
Linus Torvalds 已提交
745

746
	rcu_read_lock();
J
Jason Wang 已提交
747
	tfile = rcu_dereference(tun->tfiles[txq]);
748
	numqueues = ACCESS_ONCE(tun->numqueues);
J
Jason Wang 已提交
749

L
Linus Torvalds 已提交
750
	/* Drop packet if interface is not attached */
751
	if (txq >= numqueues)
L
Linus Torvalds 已提交
752 753
		goto drop;

754
	if (numqueues == 1) {
755 756 757 758 759 760 761 762 763 764 765 766 767 768 769
		/* Select queue was not called for the skbuff, so we extract the
		 * RPS hash and save it into the flow_table here.
		 */
		__u32 rxhash;

		rxhash = skb_get_hash(skb);
		if (rxhash) {
			struct tun_flow_entry *e;
			e = tun_flow_find(&tun->flows[tun_hashfn(rxhash)],
					rxhash);
			if (e)
				tun_flow_save_rps_rxhash(e, rxhash);
		}
	}

770 771
	tun_debug(KERN_INFO, tun, "tun_net_xmit %d\n", skb->len);

J
Jason Wang 已提交
772 773
	BUG_ON(!tfile);

774 775 776 777 778 779
	/* 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
780 781
	if (tfile->socket.sk->sk_filter &&
	    sk_filter(tfile->socket.sk, skb))
M
Michael S. Tsirkin 已提交
782 783
		goto drop;

R
Rami Rosen 已提交
784
	/* Limit the number of packets queued by dividing txq length with the
J
Jason Wang 已提交
785 786
	 * number of queues.
	 */
787 788
	if (skb_queue_len(&tfile->socket.sk->sk_receive_queue) * numqueues
			  >= dev->tx_queue_len)
789
		goto drop;
L
Linus Torvalds 已提交
790

791 792 793
	if (unlikely(skb_orphan_frags(skb, GFP_ATOMIC)))
		goto drop;

794 795 796 797 798
	if (skb->sk) {
		sock_tx_timestamp(skb->sk, &skb_shinfo(skb)->tx_flags);
		sw_tx_timestamp(skb);
	}

M
Michael S. Tsirkin 已提交
799
	/* Orphan the skb - required as we might hang on to it
800 801
	 * for indefinite time.
	 */
M
Michael S. Tsirkin 已提交
802 803
	skb_orphan(skb);

804 805
	nf_reset(skb);

806
	/* Enqueue packet */
J
Jason Wang 已提交
807
	skb_queue_tail(&tfile->socket.sk->sk_receive_queue, skb);
L
Linus Torvalds 已提交
808 809

	/* Notify and wake up reader process */
J
Jason Wang 已提交
810 811
	if (tfile->flags & TUN_FASYNC)
		kill_fasync(&tfile->fasync, SIGIO, POLL_IN);
812
	tfile->socket.sk->sk_data_ready(tfile->socket.sk);
813 814

	rcu_read_unlock();
815
	return NETDEV_TX_OK;
L
Linus Torvalds 已提交
816 817

drop:
818
	dev->stats.tx_dropped++;
819
	skb_tx_error(skb);
L
Linus Torvalds 已提交
820
	kfree_skb(skb);
821
	rcu_read_unlock();
822
	return NET_XMIT_DROP;
L
Linus Torvalds 已提交
823 824
}

825
static void tun_net_mclist(struct net_device *dev)
L
Linus Torvalds 已提交
826
{
827 828 829 830 831
	/*
	 * 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
832 833
}

E
Ed Swierk 已提交
834 835 836 837 838 839 840 841 842 843 844 845
#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;
}

846 847
static netdev_features_t tun_net_fix_features(struct net_device *dev,
	netdev_features_t features)
848 849 850 851 852
{
	struct tun_struct *tun = netdev_priv(dev);

	return (features & tun->set_features) | (features & ~TUN_USER_FEATURES);
}
853 854 855 856 857 858 859
#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
S
stephen hemminger 已提交
860
	 * Since both of those are synchronous operations, we are guaranteed
861
	 * never to have pending data when we poll for it
S
stephen hemminger 已提交
862
	 * so there is nothing to do here but return.
863 864 865 866 867 868 869
	 * 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
870
static const struct net_device_ops tun_netdev_ops = {
871
	.ndo_uninit		= tun_net_uninit,
S
Stephen Hemminger 已提交
872 873
	.ndo_open		= tun_net_open,
	.ndo_stop		= tun_net_close,
874
	.ndo_start_xmit		= tun_net_xmit,
S
Stephen Hemminger 已提交
875
	.ndo_change_mtu		= tun_net_change_mtu,
876
	.ndo_fix_features	= tun_net_fix_features,
J
Jason Wang 已提交
877
	.ndo_select_queue	= tun_select_queue,
878 879 880
#ifdef CONFIG_NET_POLL_CONTROLLER
	.ndo_poll_controller	= tun_poll_controller,
#endif
S
Stephen Hemminger 已提交
881 882 883
};

static const struct net_device_ops tap_netdev_ops = {
884
	.ndo_uninit		= tun_net_uninit,
S
Stephen Hemminger 已提交
885 886
	.ndo_open		= tun_net_open,
	.ndo_stop		= tun_net_close,
887
	.ndo_start_xmit		= tun_net_xmit,
S
Stephen Hemminger 已提交
888
	.ndo_change_mtu		= tun_net_change_mtu,
889
	.ndo_fix_features	= tun_net_fix_features,
890
	.ndo_set_rx_mode	= tun_net_mclist,
S
Stephen Hemminger 已提交
891 892
	.ndo_set_mac_address	= eth_mac_addr,
	.ndo_validate_addr	= eth_validate_addr,
J
Jason Wang 已提交
893
	.ndo_select_queue	= tun_select_queue,
894 895 896
#ifdef CONFIG_NET_POLL_CONTROLLER
	.ndo_poll_controller	= tun_poll_controller,
#endif
S
Stephen Hemminger 已提交
897 898
};

899
static void tun_flow_init(struct tun_struct *tun)
J
Jason Wang 已提交
900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917
{
	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
918 919 920 921
/* Initialize net device. */
static void tun_net_init(struct net_device *dev)
{
	struct tun_struct *tun = netdev_priv(dev);
922

L
Linus Torvalds 已提交
923 924
	switch (tun->flags & TUN_TYPE_MASK) {
	case TUN_TUN_DEV:
S
Stephen Hemminger 已提交
925 926
		dev->netdev_ops = &tun_netdev_ops;

L
Linus Torvalds 已提交
927 928 929 930 931 932
		/* Point-to-Point TUN Device */
		dev->hard_header_len = 0;
		dev->addr_len = 0;
		dev->mtu = 1500;

		/* Zero header length */
933
		dev->type = ARPHRD_NONE;
L
Linus Torvalds 已提交
934 935 936 937 938
		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
939
		dev->netdev_ops = &tap_netdev_ops;
L
Linus Torvalds 已提交
940 941
		/* Ethernet TAP Device */
		ether_setup(dev);
942
		dev->priv_flags &= ~IFF_TX_SKB_SHARING;
943
		dev->priv_flags |= IFF_LIVE_ADDR_CHANGE;
944

945
		eth_hw_addr_random(dev);
946

L
Linus Torvalds 已提交
947 948 949 950 951 952 953 954
		dev->tx_queue_len = TUN_READQ_SIZE;  /* We prefer our own queue length */
		break;
	}
}

/* Character device part */

/* Poll */
J
Jason Wang 已提交
955
static unsigned int tun_chr_poll(struct file *file, poll_table *wait)
956
{
957 958
	struct tun_file *tfile = file->private_data;
	struct tun_struct *tun = __tun_get(tfile);
959
	struct sock *sk;
960
	unsigned int mask = 0;
L
Linus Torvalds 已提交
961 962

	if (!tun)
963
		return POLLERR;
L
Linus Torvalds 已提交
964

J
Jason Wang 已提交
965
	sk = tfile->socket.sk;
966

967
	tun_debug(KERN_INFO, tun, "tun_chr_poll\n");
L
Linus Torvalds 已提交
968

969
	poll_wait(file, sk_sleep(sk), wait);
970

971
	if (!skb_queue_empty(&sk->sk_receive_queue))
L
Linus Torvalds 已提交
972 973
		mask |= POLLIN | POLLRDNORM;

974 975 976 977 978
	if (sock_writeable(sk) ||
	    (!test_and_set_bit(SOCK_ASYNC_NOSPACE, &sk->sk_socket->flags) &&
	     sock_writeable(sk)))
		mask |= POLLOUT | POLLWRNORM;

979 980 981
	if (tun->dev->reg_state != NETREG_REGISTERED)
		mask = POLLERR;

E
Eric W. Biederman 已提交
982
	tun_put(tun);
L
Linus Torvalds 已提交
983 984 985
	return mask;
}

986 987
/* 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
988
static struct sk_buff *tun_alloc_skb(struct tun_file *tfile,
989 990
				     size_t prepad, size_t len,
				     size_t linear, int noblock)
991
{
J
Jason Wang 已提交
992
	struct sock *sk = tfile->socket.sk;
993
	struct sk_buff *skb;
994
	int err;
995 996

	/* Under a page?  Don't bother with paged skb. */
H
Herbert Xu 已提交
997
	if (prepad + len < PAGE_SIZE || !linear)
998
		linear = len;
999

1000
	skb = sock_alloc_send_pskb(sk, prepad + linear, len - linear, noblock,
1001
				   &err, 0);
1002
	if (!skb)
1003
		return ERR_PTR(err);
1004 1005 1006

	skb_reserve(skb, prepad);
	skb_put(skb, linear);
1007 1008
	skb->data_len = len - linear;
	skb->len += len - linear;
1009 1010 1011 1012

	return skb;
}

L
Linus Torvalds 已提交
1013
/* Get packet from user space buffer */
J
Jason Wang 已提交
1014
static ssize_t tun_get_user(struct tun_struct *tun, struct tun_file *tfile,
1015 1016
			    void *msg_control, struct iov_iter *from,
			    int noblock)
L
Linus Torvalds 已提交
1017
{
1018
	struct tun_pi pi = { 0, cpu_to_be16(ETH_P_IP) };
L
Linus Torvalds 已提交
1019
	struct sk_buff *skb;
1020
	size_t total_len = iov_iter_count(from);
1021
	size_t len = total_len, align = NET_SKB_PAD, linear;
1022
	struct virtio_net_hdr gso = { 0 };
1023
	int good_linear;
1024 1025 1026
	int copylen;
	bool zerocopy = false;
	int err;
1027
	u32 rxhash;
1028
	ssize_t n;
L
Linus Torvalds 已提交
1029 1030

	if (!(tun->flags & TUN_NO_PI)) {
1031
		if (len < sizeof(pi))
L
Linus Torvalds 已提交
1032
			return -EINVAL;
1033
		len -= sizeof(pi);
L
Linus Torvalds 已提交
1034

1035 1036
		n = copy_from_iter(&pi, sizeof(pi), from);
		if (n != sizeof(pi))
L
Linus Torvalds 已提交
1037 1038 1039
			return -EFAULT;
	}

1040
	if (tun->flags & TUN_VNET_HDR) {
1041
		if (len < tun->vnet_hdr_sz)
1042
			return -EINVAL;
1043
		len -= tun->vnet_hdr_sz;
1044

1045 1046
		n = copy_from_iter(&gso, sizeof(gso), from);
		if (n != sizeof(gso))
1047 1048
			return -EFAULT;

1049 1050 1051 1052
		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;

1053 1054
		if (gso.hdr_len > len)
			return -EINVAL;
1055
		iov_iter_advance(from, tun->vnet_hdr_sz - sizeof(gso));
1056 1057
	}

1058
	if ((tun->flags & TUN_TYPE_MASK) == TUN_TAP_DEV) {
1059
		align += NET_IP_ALIGN;
H
Herbert Xu 已提交
1060 1061
		if (unlikely(len < ETH_HLEN ||
			     (gso.hdr_len && gso.hdr_len < ETH_HLEN)))
1062 1063
			return -EINVAL;
	}
1064

1065 1066
	good_linear = SKB_MAX_HEAD(align);

1067
	if (msg_control) {
1068 1069
		struct iov_iter i = *from;

1070 1071
		/* There are 256 bytes to be copied in skb, so there is
		 * enough room for skb expand head in case it is used.
1072 1073
		 * The rest of the buffer is mapped from userspace.
		 */
1074
		copylen = gso.hdr_len ? gso.hdr_len : GOODCOPY_LEN;
1075 1076
		if (copylen > good_linear)
			copylen = good_linear;
1077
		linear = copylen;
1078 1079
		iov_iter_advance(&i, copylen);
		if (iov_iter_npages(&i, INT_MAX) <= MAX_SKB_FRAGS)
1080 1081 1082 1083
			zerocopy = true;
	}

	if (!zerocopy) {
1084
		copylen = len;
1085 1086 1087 1088
		if (gso.hdr_len > good_linear)
			linear = good_linear;
		else
			linear = gso.hdr_len;
1089
	}
1090

1091
	skb = tun_alloc_skb(tfile, align, copylen, linear, noblock);
1092 1093 1094 1095
	if (IS_ERR(skb)) {
		if (PTR_ERR(skb) != -EAGAIN)
			tun->dev->stats.rx_dropped++;
		return PTR_ERR(skb);
L
Linus Torvalds 已提交
1096 1097
	}

1098
	if (zerocopy)
1099
		err = zerocopy_sg_from_iter(skb, from);
1100
	else {
1101
		err = skb_copy_datagram_from_iter(skb, 0, from, len);
1102 1103 1104 1105 1106
		if (!err && msg_control) {
			struct ubuf_info *uarg = msg_control;
			uarg->callback(uarg, false);
		}
	}
1107 1108

	if (err) {
1109
		tun->dev->stats.rx_dropped++;
D
Dave Jones 已提交
1110
		kfree_skb(skb);
L
Linus Torvalds 已提交
1111
		return -EFAULT;
D
Dave Jones 已提交
1112
	}
L
Linus Torvalds 已提交
1113

1114 1115 1116 1117 1118 1119 1120
	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;
		}
1121
	}
1122

L
Linus Torvalds 已提交
1123 1124
	switch (tun->flags & TUN_TYPE_MASK) {
	case TUN_TUN_DEV:
1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139
		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;
			}
		}

1140
		skb_reset_mac_header(skb);
L
Linus Torvalds 已提交
1141
		skb->protocol = pi.proto;
1142
		skb->dev = tun->dev;
L
Linus Torvalds 已提交
1143 1144 1145 1146
		break;
	case TUN_TAP_DEV:
		skb->protocol = eth_type_trans(skb, tun->dev);
		break;
1147
	}
L
Linus Torvalds 已提交
1148

1149 1150
	skb_reset_network_header(skb);

1151 1152 1153 1154
	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:
1155
			skb_shinfo(skb)->gso_type = SKB_GSO_TCPV4;
1156 1157
			break;
		case VIRTIO_NET_HDR_GSO_TCPV6:
1158
			skb_shinfo(skb)->gso_type = SKB_GSO_TCPV6;
1159
			break;
1160
		case VIRTIO_NET_HDR_GSO_UDP:
1161 1162 1163 1164 1165 1166 1167 1168 1169
		{
			static bool warned;

			if (!warned) {
				warned = true;
				netdev_warn(tun->dev,
					    "%s: using disabled UFO feature; please fix this program\n",
					    current->comm);
			}
1170
			skb_shinfo(skb)->gso_type = SKB_GSO_UDP;
1171 1172
			if (skb->protocol == htons(ETH_P_IPV6))
				ipv6_proxy_select_ident(skb);
1173
			break;
1174
		}
1175 1176 1177 1178 1179 1180 1181
		default:
			tun->dev->stats.rx_frame_errors++;
			kfree_skb(skb);
			return -EINVAL;
		}

		if (gso.gso_type & VIRTIO_NET_HDR_GSO_ECN)
1182
			skb_shinfo(skb)->gso_type |= SKB_GSO_TCP_ECN;
1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194

		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;
	}
1195

1196 1197 1198 1199
	/* 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;
1200
		skb_shinfo(skb)->tx_flags |= SKBTX_SHARED_FRAG;
1201 1202
	}

1203
	skb_probe_transport_header(skb, 0);
1204

1205
	rxhash = skb_get_hash(skb);
L
Linus Torvalds 已提交
1206
	netif_rx_ni(skb);
1207

1208 1209
	tun->dev->stats.rx_packets++;
	tun->dev->stats.rx_bytes += len;
L
Linus Torvalds 已提交
1210

1211
	tun_flow_update(tun, rxhash, tfile);
1212
	return total_len;
1213
}
L
Linus Torvalds 已提交
1214

1215
static ssize_t tun_chr_write_iter(struct kiocb *iocb, struct iov_iter *from)
L
Linus Torvalds 已提交
1216
{
1217
	struct file *file = iocb->ki_filp;
H
Herbert Xu 已提交
1218
	struct tun_struct *tun = tun_get(file);
J
Jason Wang 已提交
1219
	struct tun_file *tfile = file->private_data;
E
Eric W. Biederman 已提交
1220
	ssize_t result;
L
Linus Torvalds 已提交
1221 1222 1223 1224

	if (!tun)
		return -EBADFD;

1225
	result = tun_get_user(tun, tfile, NULL, from, file->f_flags & O_NONBLOCK);
E
Eric W. Biederman 已提交
1226 1227 1228

	tun_put(tun);
	return result;
L
Linus Torvalds 已提交
1229 1230 1231
}

/* Put packet to the user space buffer */
1232
static ssize_t tun_put_user(struct tun_struct *tun,
J
Jason Wang 已提交
1233
			    struct tun_file *tfile,
1234
			    struct sk_buff *skb,
H
Herbert Xu 已提交
1235
			    struct iov_iter *iter)
L
Linus Torvalds 已提交
1236 1237
{
	struct tun_pi pi = { 0, skb->protocol };
H
Herbert Xu 已提交
1238
	ssize_t total;
1239
	int vlan_offset = 0;
1240
	int vlan_hlen = 0;
H
Herbert Xu 已提交
1241
	int vnet_hdr_sz = 0;
1242 1243 1244

	if (vlan_tx_tag_present(skb))
		vlan_hlen = VLAN_HLEN;
L
Linus Torvalds 已提交
1245

H
Herbert Xu 已提交
1246 1247 1248
	if (tun->flags & TUN_VNET_HDR)
		vnet_hdr_sz = tun->vnet_hdr_sz;

H
Herbert Xu 已提交
1249 1250
	total = skb->len + vlan_hlen + vnet_hdr_sz;

L
Linus Torvalds 已提交
1251
	if (!(tun->flags & TUN_NO_PI)) {
H
Herbert Xu 已提交
1252
		if (iov_iter_count(iter) < sizeof(pi))
L
Linus Torvalds 已提交
1253 1254
			return -EINVAL;

H
Herbert Xu 已提交
1255 1256
		total += sizeof(pi);
		if (iov_iter_count(iter) < total) {
L
Linus Torvalds 已提交
1257 1258 1259
			/* Packet will be striped */
			pi.flags |= TUN_PKT_STRIP;
		}
1260

H
Herbert Xu 已提交
1261
		if (copy_to_iter(&pi, sizeof(pi), iter) != sizeof(pi))
L
Linus Torvalds 已提交
1262
			return -EFAULT;
1263
	}
L
Linus Torvalds 已提交
1264

H
Herbert Xu 已提交
1265
	if (vnet_hdr_sz) {
1266
		struct virtio_net_hdr gso = { 0 }; /* no info leak */
H
Herbert Xu 已提交
1267
		if (iov_iter_count(iter) < vnet_hdr_sz)
1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279
			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;
1280
			else {
1281
				pr_err("unexpected GSO type: "
1282 1283 1284 1285 1286 1287 1288 1289 1290 1291
				       "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;
			}
1292 1293 1294 1295 1296 1297 1298
			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;
1299 1300
			gso.csum_start = skb_checksum_start_offset(skb) +
					 vlan_hlen;
1301
			gso.csum_offset = skb->csum_offset;
1302 1303
		} else if (skb->ip_summed == CHECKSUM_UNNECESSARY) {
			gso.flags = VIRTIO_NET_HDR_F_DATA_VALID;
1304 1305
		} /* else everything is zero */

H
Herbert Xu 已提交
1306
		if (copy_to_iter(&gso, sizeof(gso), iter) != sizeof(gso))
1307
			return -EFAULT;
1308 1309

		iov_iter_advance(iter, vnet_hdr_sz - sizeof(gso));
1310 1311
	}

1312
	if (vlan_hlen) {
H
Herbert Xu 已提交
1313
		int ret;
J
Jason Wang 已提交
1314 1315 1316 1317 1318 1319 1320 1321 1322 1323
		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);

H
Herbert Xu 已提交
1324 1325
		ret = skb_copy_datagram_iter(skb, 0, iter, vlan_offset);
		if (ret || !iov_iter_count(iter))
J
Jason Wang 已提交
1326 1327
			goto done;

H
Herbert Xu 已提交
1328 1329
		ret = copy_to_iter(&veth, sizeof(veth), iter);
		if (ret != sizeof(veth) || !iov_iter_count(iter))
J
Jason Wang 已提交
1330 1331
			goto done;
	}
L
Linus Torvalds 已提交
1332

H
Herbert Xu 已提交
1333
	skb_copy_datagram_iter(skb, vlan_offset, iter, skb->len - vlan_offset);
L
Linus Torvalds 已提交
1334

J
Jason Wang 已提交
1335
done:
1336
	tun->dev->stats.tx_packets++;
H
Herbert Xu 已提交
1337
	tun->dev->stats.tx_bytes += skb->len + vlan_hlen;
L
Linus Torvalds 已提交
1338 1339 1340 1341

	return total;
}

J
Jason Wang 已提交
1342
static ssize_t tun_do_read(struct tun_struct *tun, struct tun_file *tfile,
1343 1344
			   struct iov_iter *to,
			   int noblock)
L
Linus Torvalds 已提交
1345 1346
{
	struct sk_buff *skb;
1347
	ssize_t ret;
1348
	int peeked, err, off = 0;
L
Linus Torvalds 已提交
1349

1350
	tun_debug(KERN_INFO, tun, "tun_do_read\n");
L
Linus Torvalds 已提交
1351

1352 1353
	if (!iov_iter_count(to))
		return 0;
L
Linus Torvalds 已提交
1354

1355 1356
	if (tun->dev->reg_state != NETREG_REGISTERED)
		return -EIO;
L
Linus Torvalds 已提交
1357

1358 1359 1360
	/* Read frames from queue */
	skb = __skb_recv_datagram(tfile->socket.sk, noblock ? MSG_DONTWAIT : 0,
				  &peeked, &off, &err);
H
Herbert Xu 已提交
1361
	if (!skb)
1362
		return 0;
H
Herbert Xu 已提交
1363

1364
	ret = tun_put_user(tun, tfile, skb, to);
H
Herbert Xu 已提交
1365
	kfree_skb(skb);
L
Linus Torvalds 已提交
1366

1367 1368 1369
	return ret;
}

1370
static ssize_t tun_chr_read_iter(struct kiocb *iocb, struct iov_iter *to)
1371 1372 1373 1374
{
	struct file *file = iocb->ki_filp;
	struct tun_file *tfile = file->private_data;
	struct tun_struct *tun = __tun_get(tfile);
1375
	ssize_t len = iov_iter_count(to), ret;
1376 1377 1378

	if (!tun)
		return -EBADFD;
1379
	ret = tun_do_read(tun, tfile, to, file->f_flags & O_NONBLOCK);
1380
	ret = min_t(ssize_t, ret, len);
Z
Zhi Yong Wu 已提交
1381 1382
	if (ret > 0)
		iocb->ki_pos = ret;
E
Eric W. Biederman 已提交
1383
	tun_put(tun);
L
Linus Torvalds 已提交
1384 1385 1386
	return ret;
}

J
Jason Wang 已提交
1387 1388 1389 1390
static void tun_free_netdev(struct net_device *dev)
{
	struct tun_struct *tun = netdev_priv(dev);

J
Jason Wang 已提交
1391
	BUG_ON(!(list_empty(&tun->disabled)));
J
Jason Wang 已提交
1392
	tun_flow_uninit(tun);
1393
	security_tun_dev_free_security(tun->security);
J
Jason Wang 已提交
1394 1395 1396
	free_netdev(dev);
}

L
Linus Torvalds 已提交
1397 1398 1399 1400
static void tun_setup(struct net_device *dev)
{
	struct tun_struct *tun = netdev_priv(dev);

1401 1402
	tun->owner = INVALID_UID;
	tun->group = INVALID_GID;
L
Linus Torvalds 已提交
1403 1404

	dev->ethtool_ops = &tun_ethtool_ops;
J
Jason Wang 已提交
1405
	dev->destructor = tun_free_netdev;
L
Linus Torvalds 已提交
1406 1407
}

1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422
/* 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,
};

1423 1424
static void tun_sock_write_space(struct sock *sk)
{
J
Jason Wang 已提交
1425
	struct tun_file *tfile;
1426
	wait_queue_head_t *wqueue;
1427 1428 1429 1430 1431 1432 1433

	if (!sock_writeable(sk))
		return;

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

1434 1435 1436
	wqueue = sk_sleep(sk);
	if (wqueue && waitqueue_active(wqueue))
		wake_up_interruptible_sync_poll(wqueue, POLLOUT |
1437
						POLLWRNORM | POLLWRBAND);
H
Herbert Xu 已提交
1438

J
Jason Wang 已提交
1439 1440
	tfile = container_of(sk, struct tun_file, sk);
	kill_fasync(&tfile->fasync, SIGIO, POLL_OUT);
1441 1442
}

1443 1444 1445
static int tun_sendmsg(struct kiocb *iocb, struct socket *sock,
		       struct msghdr *m, size_t total_len)
{
J
Jason Wang 已提交
1446 1447 1448
	int ret;
	struct tun_file *tfile = container_of(sock, struct tun_file, socket);
	struct tun_struct *tun = __tun_get(tfile);
1449
	struct iov_iter from;
J
Jason Wang 已提交
1450 1451 1452

	if (!tun)
		return -EBADFD;
1453 1454 1455 1456

	iov_iter_init(&from, WRITE, m->msg_iov, m->msg_iovlen, total_len);
	ret = tun_get_user(tun, tfile, m->msg_control, &from,
			   m->msg_flags & MSG_DONTWAIT);
J
Jason Wang 已提交
1457 1458
	tun_put(tun);
	return ret;
1459 1460 1461 1462 1463 1464
}

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
1465 1466
	struct tun_file *tfile = container_of(sock, struct tun_file, socket);
	struct tun_struct *tun = __tun_get(tfile);
1467
	struct iov_iter to;
1468
	int ret;
J
Jason Wang 已提交
1469 1470 1471 1472

	if (!tun)
		return -EBADFD;

1473
	if (flags & ~(MSG_DONTWAIT|MSG_TRUNC|MSG_ERRQUEUE)) {
1474 1475 1476
		ret = -EINVAL;
		goto out;
	}
1477 1478 1479 1480 1481
	if (flags & MSG_ERRQUEUE) {
		ret = sock_recv_errqueue(sock->sk, m, total_len,
					 SOL_PACKET, TUN_TX_TIMESTAMP);
		goto out;
	}
1482 1483
	iov_iter_init(&to, READ, m->msg_iov, m->msg_iovlen, total_len);
	ret = tun_do_read(tun, tfile, &to, flags & MSG_DONTWAIT);
1484 1485 1486 1487
	if (ret > total_len) {
		m->msg_flags |= MSG_TRUNC;
		ret = flags & MSG_TRUNC ? ret : total_len;
	}
1488
out:
J
Jason Wang 已提交
1489
	tun_put(tun);
1490 1491 1492
	return ret;
}

1493 1494 1495 1496 1497 1498 1499
static int tun_release(struct socket *sock)
{
	if (sock->sk)
		sock_put(sock->sk);
	return 0;
}

1500 1501 1502 1503
/* Ops structure to mimic raw sockets with tun */
static const struct proto_ops tun_socket_ops = {
	.sendmsg = tun_sendmsg,
	.recvmsg = tun_recvmsg,
1504
	.release = tun_release,
1505 1506
};

1507 1508 1509
static struct proto tun_proto = {
	.name		= "tun",
	.owner		= THIS_MODULE,
J
Jason Wang 已提交
1510
	.obj_size	= sizeof(struct tun_file),
1511
};
1512

1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524
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;

1525 1526 1527
	/* This flag has no real effect.  We track the value for backwards
	 * compatibility.
	 */
1528 1529 1530 1531 1532 1533
	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
1534 1535 1536
	if (tun->flags & TUN_TAP_MQ)
		flags |= IFF_MULTI_QUEUE;

1537 1538 1539
	if (tun->flags & TUN_PERSIST)
		flags |= IFF_PERSIST;

1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553
	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));
1554 1555 1556 1557
	return uid_valid(tun->owner)?
		sprintf(buf, "%u\n",
			from_kuid_munged(current_user_ns(), tun->owner)):
		sprintf(buf, "-1\n");
1558 1559 1560 1561 1562 1563
}

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));
1564 1565 1566 1567
	return gid_valid(tun->group) ?
		sprintf(buf, "%u\n",
			from_kgid_munged(current_user_ns(), tun->group)):
		sprintf(buf, "-1\n");
1568 1569 1570 1571 1572 1573
}

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);

1574
static int tun_set_iff(struct net *net, struct file *file, struct ifreq *ifr)
L
Linus Torvalds 已提交
1575 1576
{
	struct tun_struct *tun;
J
Jason Wang 已提交
1577
	struct tun_file *tfile = file->private_data;
L
Linus Torvalds 已提交
1578 1579 1580
	struct net_device *dev;
	int err;

1581 1582 1583
	if (tfile->detached)
		return -EINVAL;

1584 1585
	dev = __dev_get_by_name(net, ifr->ifr_name);
	if (dev) {
1586 1587
		if (ifr->ifr_flags & IFF_TUN_EXCL)
			return -EBUSY;
1588 1589 1590 1591 1592 1593 1594
		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;

1595 1596 1597 1598
		if (!!(ifr->ifr_flags & IFF_MULTI_QUEUE) !=
		    !!(tun->flags & TUN_TAP_MQ))
			return -EINVAL;

1599
		if (tun_not_capable(tun))
P
Paul Moore 已提交
1600
			return -EPERM;
1601
		err = security_tun_dev_open(tun->security);
P
Paul Moore 已提交
1602 1603 1604
		if (err < 0)
			return err;

1605
		err = tun_attach(tun, file, ifr->ifr_flags & IFF_NOFILTER);
1606 1607
		if (err < 0)
			return err;
J
Jason Wang 已提交
1608 1609

		if (tun->flags & TUN_TAP_MQ &&
1610 1611 1612 1613 1614 1615
		    (tun->numqueues + tun->numdisabled > 1)) {
			/* One or more queue has already been attached, no need
			 * to initialize the device again.
			 */
			return 0;
		}
1616
	}
L
Linus Torvalds 已提交
1617 1618 1619
	else {
		char *name;
		unsigned long flags = 0;
1620 1621
		int queues = ifr->ifr_flags & IFF_MULTI_QUEUE ?
			     MAX_TAP_QUEUES : 1;
L
Linus Torvalds 已提交
1622

1623
		if (!ns_capable(net->user_ns, CAP_NET_ADMIN))
1624
			return -EPERM;
P
Paul Moore 已提交
1625 1626 1627
		err = security_tun_dev_create();
		if (err < 0)
			return err;
1628

L
Linus Torvalds 已提交
1629 1630 1631 1632 1633 1634 1635 1636 1637
		/* 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";
1638
		} else
1639
			return -EINVAL;
1640

L
Linus Torvalds 已提交
1641 1642 1643
		if (*ifr->ifr_name)
			name = ifr->ifr_name;

J
Jason Wang 已提交
1644
		dev = alloc_netdev_mqs(sizeof(struct tun_struct), name,
1645 1646
				       NET_NAME_UNKNOWN, tun_setup, queues,
				       queues);
1647

L
Linus Torvalds 已提交
1648 1649 1650
		if (!dev)
			return -ENOMEM;

1651
		dev_net_set(dev, net);
1652
		dev->rtnl_link_ops = &tun_link_ops;
1653
		dev->ifindex = tfile->ifindex;
S
Stephen Hemminger 已提交
1654

L
Linus Torvalds 已提交
1655 1656 1657
		tun = netdev_priv(dev);
		tun->dev = dev;
		tun->flags = flags;
1658
		tun->txflt.count = 0;
1659
		tun->vnet_hdr_sz = sizeof(struct virtio_net_hdr);
1660

J
Jason Wang 已提交
1661 1662
		tun->filter_attached = false;
		tun->sndbuf = tfile->socket.sk->sk_sndbuf;
1663

J
Jason Wang 已提交
1664 1665
		spin_lock_init(&tun->lock);

1666 1667 1668
		err = security_tun_dev_alloc_security(&tun->security);
		if (err < 0)
			goto err_free_dev;
P
Paul Moore 已提交
1669

L
Linus Torvalds 已提交
1670
		tun_net_init(dev);
1671
		tun_flow_init(tun);
J
Jason Wang 已提交
1672

1673
		dev->hw_features = NETIF_F_SG | NETIF_F_FRAGLIST |
J
Jason Wang 已提交
1674 1675
				   TUN_USER_FEATURES | NETIF_F_HW_VLAN_CTAG_TX |
				   NETIF_F_HW_VLAN_STAG_TX;
1676
		dev->features = dev->hw_features;
1677 1678 1679
		dev->vlan_features = dev->features &
				     ~(NETIF_F_HW_VLAN_CTAG_TX |
				       NETIF_F_HW_VLAN_STAG_TX);
1680

J
Jason Wang 已提交
1681
		INIT_LIST_HEAD(&tun->disabled);
1682
		err = tun_attach(tun, file, false);
1683
		if (err < 0)
1684
			goto err_free_flow;
1685

L
Linus Torvalds 已提交
1686 1687
		err = register_netdevice(tun->dev);
		if (err < 0)
1688
			goto err_detach;
1689

1690 1691 1692
		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))
1693
			pr_err("Failed to create tun sysfs files\n");
L
Linus Torvalds 已提交
1694 1695
	}

1696 1697
	netif_carrier_on(tun->dev);

1698
	tun_debug(KERN_INFO, tun, "tun_set_iff\n");
L
Linus Torvalds 已提交
1699 1700 1701

	if (ifr->ifr_flags & IFF_NO_PI)
		tun->flags |= TUN_NO_PI;
1702 1703
	else
		tun->flags &= ~TUN_NO_PI;
L
Linus Torvalds 已提交
1704

1705 1706 1707
	/* This flag has no real effect.  We track the value for backwards
	 * compatibility.
	 */
L
Linus Torvalds 已提交
1708 1709
	if (ifr->ifr_flags & IFF_ONE_QUEUE)
		tun->flags |= TUN_ONE_QUEUE;
1710 1711
	else
		tun->flags &= ~TUN_ONE_QUEUE;
L
Linus Torvalds 已提交
1712

1713 1714 1715 1716 1717
	if (ifr->ifr_flags & IFF_VNET_HDR)
		tun->flags |= TUN_VNET_HDR;
	else
		tun->flags &= ~TUN_VNET_HDR;

J
Jason Wang 已提交
1718 1719 1720 1721 1722
	if (ifr->ifr_flags & IFF_MULTI_QUEUE)
		tun->flags |= TUN_TAP_MQ;
	else
		tun->flags &= ~TUN_TAP_MQ;

1723 1724 1725 1726
	/* Make sure persistent devices do not get stuck in
	 * xoff state.
	 */
	if (netif_running(tun->dev))
J
Jason Wang 已提交
1727
		netif_tx_wake_all_queues(tun->dev);
1728

L
Linus Torvalds 已提交
1729 1730 1731
	strcpy(ifr->ifr_name, tun->dev->name);
	return 0;

1732 1733 1734 1735 1736 1737
err_detach:
	tun_detach_all(dev);
err_free_flow:
	tun_flow_uninit(tun);
	security_tun_dev_free_security(tun->security);
err_free_dev:
L
Linus Torvalds 已提交
1738 1739 1740 1741
	free_netdev(dev);
	return err;
}

R
Rami Rosen 已提交
1742
static void tun_get_iff(struct net *net, struct tun_struct *tun,
1743
		       struct ifreq *ifr)
1744
{
1745
	tun_debug(KERN_INFO, tun, "tun_get_iff\n");
1746 1747 1748

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

1749
	ifr->ifr_flags = tun_flags(tun);
1750 1751 1752

}

1753 1754
/* This is like a cut-down ethtool ops, except done via tun fd so no
 * privs required. */
1755
static int set_offload(struct tun_struct *tun, unsigned long arg)
1756
{
1757
	netdev_features_t features = 0;
1758 1759

	if (arg & TUN_F_CSUM) {
1760
		features |= NETIF_F_HW_CSUM;
1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780
		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);
		}
	}

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

1781 1782
	tun->set_features = features;
	netdev_update_features(tun->dev);
1783 1784 1785 1786

	return 0;
}

J
Jason Wang 已提交
1787 1788 1789 1790 1791 1792
static void tun_detach_filter(struct tun_struct *tun, int n)
{
	int i;
	struct tun_file *tfile;

	for (i = 0; i < n; i++) {
1793
		tfile = rtnl_dereference(tun->tfiles[i]);
J
Jason Wang 已提交
1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805
		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++) {
1806
		tfile = rtnl_dereference(tun->tfiles[i]);
J
Jason Wang 已提交
1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823
		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++) {
1824
		tfile = rtnl_dereference(tun->tfiles[i]);
J
Jason Wang 已提交
1825 1826 1827 1828
		tfile->socket.sk->sk_sndbuf = tun->sndbuf;
	}
}

1829 1830 1831 1832 1833 1834 1835 1836 1837
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
1838
		tun = tfile->detached;
1839
		if (!tun) {
1840
			ret = -EINVAL;
1841 1842 1843 1844 1845
			goto unlock;
		}
		ret = security_tun_dev_attach_queue(tun->security);
		if (ret < 0)
			goto unlock;
1846
		ret = tun_attach(tun, file, false);
J
Jason Wang 已提交
1847
	} else if (ifr->ifr_flags & IFF_DETACH_QUEUE) {
1848
		tun = rtnl_dereference(tfile->tun);
1849
		if (!tun || !(tun->flags & TUN_TAP_MQ) || tfile->detached)
J
Jason Wang 已提交
1850 1851 1852 1853
			ret = -EINVAL;
		else
			__tun_detach(tfile, false);
	} else
1854 1855
		ret = -EINVAL;

1856
unlock:
1857 1858 1859 1860
	rtnl_unlock();
	return ret;
}

1861 1862
static long __tun_chr_ioctl(struct file *file, unsigned int cmd,
			    unsigned long arg, int ifreq_len)
L
Linus Torvalds 已提交
1863
{
E
Eric W. Biederman 已提交
1864
	struct tun_file *tfile = file->private_data;
E
Eric W. Biederman 已提交
1865
	struct tun_struct *tun;
L
Linus Torvalds 已提交
1866 1867
	void __user* argp = (void __user*)arg;
	struct ifreq ifr;
1868 1869
	kuid_t owner;
	kgid_t group;
1870
	int sndbuf;
1871
	int vnet_hdr_sz;
1872
	unsigned int ifindex;
1873
	int ret;
L
Linus Torvalds 已提交
1874

1875
	if (cmd == TUNSETIFF || cmd == TUNSETQUEUE || _IOC_TYPE(cmd) == 0x89) {
1876
		if (copy_from_user(&ifr, argp, ifreq_len))
L
Linus Torvalds 已提交
1877
			return -EFAULT;
D
David S. Miller 已提交
1878
	} else {
1879
		memset(&ifr, 0, sizeof(ifr));
D
David S. Miller 已提交
1880
	}
E
Eric W. Biederman 已提交
1881 1882 1883 1884 1885
	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 |
1886
				IFF_VNET_HDR | IFF_MULTI_QUEUE,
E
Eric W. Biederman 已提交
1887
				(unsigned int __user*)argp);
1888 1889
	} else if (cmd == TUNSETQUEUE)
		return tun_set_queue(file, &ifr);
E
Eric W. Biederman 已提交
1890

J
Jason Wang 已提交
1891
	ret = 0;
1892 1893
	rtnl_lock();

E
Eric W. Biederman 已提交
1894
	tun = __tun_get(tfile);
L
Linus Torvalds 已提交
1895 1896 1897
	if (cmd == TUNSETIFF && !tun) {
		ifr.ifr_name[IFNAMSIZ-1] = '\0';

1898
		ret = tun_set_iff(tfile->net, file, &ifr);
L
Linus Torvalds 已提交
1899

1900 1901
		if (ret)
			goto unlock;
L
Linus Torvalds 已提交
1902

1903
		if (copy_to_user(argp, &ifr, ifreq_len))
1904 1905
			ret = -EFAULT;
		goto unlock;
L
Linus Torvalds 已提交
1906
	}
1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919
	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
1920

1921
	ret = -EBADFD;
L
Linus Torvalds 已提交
1922
	if (!tun)
1923
		goto unlock;
L
Linus Torvalds 已提交
1924

1925
	tun_debug(KERN_INFO, tun, "tun_chr_ioctl cmd %u\n", cmd);
L
Linus Torvalds 已提交
1926

E
Eric W. Biederman 已提交
1927
	ret = 0;
L
Linus Torvalds 已提交
1928
	switch (cmd) {
1929
	case TUNGETIFF:
R
Rami Rosen 已提交
1930
		tun_get_iff(current->nsproxy->net_ns, tun, &ifr);
1931

1932 1933
		if (tfile->detached)
			ifr.ifr_flags |= IFF_DETACH_QUEUE;
1934 1935
		if (!tfile->socket.sk->sk_filter)
			ifr.ifr_flags |= IFF_NOFILTER;
1936

1937
		if (copy_to_user(argp, &ifr, ifreq_len))
E
Eric W. Biederman 已提交
1938
			ret = -EFAULT;
1939 1940
		break;

L
Linus Torvalds 已提交
1941 1942 1943
	case TUNSETNOCSUM:
		/* Disable/Enable checksum */

1944 1945
		/* [unimplemented] */
		tun_debug(KERN_INFO, tun, "ignored: set checksum %s\n",
1946
			  arg ? "disabled" : "enabled");
L
Linus Torvalds 已提交
1947 1948 1949
		break;

	case TUNSETPERSIST:
J
Jason Wang 已提交
1950 1951 1952
		/* Disable/Enable persist mode. Keep an extra reference to the
		 * module to prevent the module being unprobed.
		 */
J
Jason Wang 已提交
1953
		if (arg && !(tun->flags & TUN_PERSIST)) {
L
Linus Torvalds 已提交
1954
			tun->flags |= TUN_PERSIST;
J
Jason Wang 已提交
1955
			__module_get(THIS_MODULE);
J
Jason Wang 已提交
1956 1957
		}
		if (!arg && (tun->flags & TUN_PERSIST)) {
L
Linus Torvalds 已提交
1958
			tun->flags &= ~TUN_PERSIST;
J
Jason Wang 已提交
1959 1960
			module_put(THIS_MODULE);
		}
L
Linus Torvalds 已提交
1961

1962 1963
		tun_debug(KERN_INFO, tun, "persist %s\n",
			  arg ? "enabled" : "disabled");
L
Linus Torvalds 已提交
1964 1965 1966 1967
		break;

	case TUNSETOWNER:
		/* Set owner of the device */
1968 1969 1970 1971 1972 1973
		owner = make_kuid(current_user_ns(), arg);
		if (!uid_valid(owner)) {
			ret = -EINVAL;
			break;
		}
		tun->owner = owner;
1974
		tun_debug(KERN_INFO, tun, "owner set to %u\n",
1975
			  from_kuid(&init_user_ns, tun->owner));
L
Linus Torvalds 已提交
1976 1977
		break;

1978 1979
	case TUNSETGROUP:
		/* Set group of the device */
1980 1981 1982 1983 1984 1985
		group = make_kgid(current_user_ns(), arg);
		if (!gid_valid(group)) {
			ret = -EINVAL;
			break;
		}
		tun->group = group;
1986
		tun_debug(KERN_INFO, tun, "group set to %u\n",
1987
			  from_kgid(&init_user_ns, tun->group));
1988 1989
		break;

1990 1991 1992
	case TUNSETLINK:
		/* Only allow setting the type when the interface is down */
		if (tun->dev->flags & IFF_UP) {
1993 1994
			tun_debug(KERN_INFO, tun,
				  "Linktype set failed because interface is up\n");
1995
			ret = -EBUSY;
1996 1997
		} else {
			tun->dev->type = (int) arg;
1998 1999
			tun_debug(KERN_INFO, tun, "linktype set to %d\n",
				  tun->dev->type);
2000
			ret = 0;
2001
		}
E
Eric W. Biederman 已提交
2002
		break;
2003

L
Linus Torvalds 已提交
2004 2005 2006 2007 2008
#ifdef TUN_DEBUG
	case TUNSETDEBUG:
		tun->debug = arg;
		break;
#endif
2009
	case TUNSETOFFLOAD:
2010
		ret = set_offload(tun, arg);
E
Eric W. Biederman 已提交
2011
		break;
2012

2013 2014
	case TUNSETTXFILTER:
		/* Can be set only for TAPs */
E
Eric W. Biederman 已提交
2015
		ret = -EINVAL;
2016
		if ((tun->flags & TUN_TYPE_MASK) != TUN_TAP_DEV)
E
Eric W. Biederman 已提交
2017
			break;
H
Harvey Harrison 已提交
2018
		ret = update_filter(&tun->txflt, (void __user *)arg);
E
Eric W. Biederman 已提交
2019
		break;
L
Linus Torvalds 已提交
2020 2021

	case SIOCGIFHWADDR:
2022
		/* Get hw address */
2023 2024
		memcpy(ifr.ifr_hwaddr.sa_data, tun->dev->dev_addr, ETH_ALEN);
		ifr.ifr_hwaddr.sa_family = tun->dev->type;
2025
		if (copy_to_user(argp, &ifr, ifreq_len))
E
Eric W. Biederman 已提交
2026 2027
			ret = -EFAULT;
		break;
L
Linus Torvalds 已提交
2028 2029

	case SIOCSIFHWADDR:
2030
		/* Set hw address */
2031 2032
		tun_debug(KERN_DEBUG, tun, "set hw address: %pM\n",
			  ifr.ifr_hwaddr.sa_data);
2033 2034

		ret = dev_set_mac_address(tun->dev, &ifr.ifr_hwaddr);
E
Eric W. Biederman 已提交
2035
		break;
2036 2037

	case TUNGETSNDBUF:
J
Jason Wang 已提交
2038
		sndbuf = tfile->socket.sk->sk_sndbuf;
2039 2040 2041 2042 2043 2044 2045 2046 2047 2048
		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
2049 2050
		tun->sndbuf = sndbuf;
		tun_set_sndbuf(tun);
2051 2052
		break;

2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071
	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
2072 2073 2074 2075 2076 2077
	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
2078
		if (copy_from_user(&tun->fprog, argp, sizeof(tun->fprog)))
M
Michael S. Tsirkin 已提交
2079 2080
			break;

J
Jason Wang 已提交
2081
		ret = tun_attach_filter(tun);
M
Michael S. Tsirkin 已提交
2082 2083 2084 2085 2086 2087 2088
		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
2089 2090
		ret = 0;
		tun_detach_filter(tun, tun->numqueues);
M
Michael S. Tsirkin 已提交
2091 2092
		break;

P
Pavel Emelyanov 已提交
2093 2094 2095 2096 2097 2098 2099 2100 2101 2102
	case TUNGETFILTER:
		ret = -EINVAL;
		if ((tun->flags & TUN_TYPE_MASK) != TUN_TAP_DEV)
			break;
		ret = -EFAULT;
		if (copy_to_user(argp, &tun->fprog, sizeof(tun->fprog)))
			break;
		ret = 0;
		break;

L
Linus Torvalds 已提交
2103
	default:
E
Eric W. Biederman 已提交
2104 2105
		ret = -EINVAL;
		break;
2106
	}
L
Linus Torvalds 已提交
2107

2108 2109 2110 2111
unlock:
	rtnl_unlock();
	if (tun)
		tun_put(tun);
E
Eric W. Biederman 已提交
2112
	return ret;
L
Linus Torvalds 已提交
2113 2114
}

2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149
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
2150 2151
static int tun_chr_fasync(int fd, struct file *file, int on)
{
J
Jason Wang 已提交
2152
	struct tun_file *tfile = file->private_data;
L
Linus Torvalds 已提交
2153 2154
	int ret;

J
Jason Wang 已提交
2155
	if ((ret = fasync_helper(fd, file, on, &tfile->fasync)) < 0)
J
Jonathan Corbet 已提交
2156
		goto out;
2157

L
Linus Torvalds 已提交
2158
	if (on) {
2159
		__f_setown(file, task_pid(current), PIDTYPE_PID, 0);
J
Jason Wang 已提交
2160
		tfile->flags |= TUN_FASYNC;
2161
	} else
J
Jason Wang 已提交
2162
		tfile->flags &= ~TUN_FASYNC;
J
Jonathan Corbet 已提交
2163 2164 2165
	ret = 0;
out:
	return ret;
L
Linus Torvalds 已提交
2166 2167 2168 2169
}

static int tun_chr_open(struct inode *inode, struct file * file)
{
E
Eric W. Biederman 已提交
2170
	struct tun_file *tfile;
T
Thomas Gleixner 已提交
2171

2172
	DBG1(KERN_INFO, "tunX: tun_chr_open\n");
E
Eric W. Biederman 已提交
2173

J
Jason Wang 已提交
2174 2175
	tfile = (struct tun_file *)sk_alloc(&init_net, AF_UNSPEC, GFP_KERNEL,
					    &tun_proto);
E
Eric W. Biederman 已提交
2176 2177
	if (!tfile)
		return -ENOMEM;
2178
	RCU_INIT_POINTER(tfile->tun, NULL);
E
Eric W. Biederman 已提交
2179
	tfile->net = get_net(current->nsproxy->net_ns);
J
Jason Wang 已提交
2180
	tfile->flags = 0;
2181
	tfile->ifindex = 0;
J
Jason Wang 已提交
2182 2183

	init_waitqueue_head(&tfile->wq.wait);
2184
	RCU_INIT_POINTER(tfile->socket.wq, &tfile->wq);
J
Jason Wang 已提交
2185 2186 2187 2188 2189 2190 2191 2192 2193 2194

	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
2195
	file->private_data = tfile;
J
Jason Wang 已提交
2196
	set_bit(SOCK_EXTERNALLY_ALLOCATED, &tfile->socket.flags);
J
Jason Wang 已提交
2197
	INIT_LIST_HEAD(&tfile->next);
J
Jason Wang 已提交
2198

2199 2200
	sock_set_flag(&tfile->sk, SOCK_ZEROCOPY);

L
Linus Torvalds 已提交
2201 2202 2203 2204 2205
	return 0;
}

static int tun_chr_close(struct inode *inode, struct file *file)
{
E
Eric W. Biederman 已提交
2206
	struct tun_file *tfile = file->private_data;
J
Jason Wang 已提交
2207
	struct net *net = tfile->net;
L
Linus Torvalds 已提交
2208

J
Jason Wang 已提交
2209
	tun_detach(tfile, true);
J
Jason Wang 已提交
2210
	put_net(net);
L
Linus Torvalds 已提交
2211 2212 2213 2214

	return 0;
}

2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235
#ifdef CONFIG_PROC_FS
static int tun_chr_show_fdinfo(struct seq_file *m, struct file *f)
{
	struct tun_struct *tun;
	struct ifreq ifr;

	memset(&ifr, 0, sizeof(ifr));

	rtnl_lock();
	tun = tun_get(f);
	if (tun)
		tun_get_iff(current->nsproxy->net_ns, tun, &ifr);
	rtnl_unlock();

	if (tun)
		tun_put(tun);

	return seq_printf(m, "iff:\t%s\n", ifr.ifr_name);
}
#endif

2236
static const struct file_operations tun_fops = {
2237
	.owner	= THIS_MODULE,
L
Linus Torvalds 已提交
2238
	.llseek = no_llseek,
2239
	.read  = new_sync_read,
2240
	.write = new_sync_write,
2241
	.read_iter  = tun_chr_read_iter,
2242
	.write_iter = tun_chr_write_iter,
L
Linus Torvalds 已提交
2243
	.poll	= tun_chr_poll,
2244 2245 2246 2247
	.unlocked_ioctl	= tun_chr_ioctl,
#ifdef CONFIG_COMPAT
	.compat_ioctl = tun_chr_compat_ioctl,
#endif
L
Linus Torvalds 已提交
2248 2249
	.open	= tun_chr_open,
	.release = tun_chr_close,
2250 2251 2252 2253
	.fasync = tun_chr_fasync,
#ifdef CONFIG_PROC_FS
	.show_fdinfo = tun_chr_show_fdinfo,
#endif
L
Linus Torvalds 已提交
2254 2255 2256 2257 2258
};

static struct miscdevice tun_miscdev = {
	.minor = TUN_MINOR,
	.name = "tun",
2259
	.nodename = "net/tun",
L
Linus Torvalds 已提交
2260 2261 2262 2263 2264 2265 2266 2267 2268
	.fops = &tun_fops,
};

/* ethtool interface */

static int tun_get_settings(struct net_device *dev, struct ethtool_cmd *cmd)
{
	cmd->supported		= 0;
	cmd->advertising	= 0;
2269
	ethtool_cmd_speed_set(cmd, SPEED_10);
L
Linus Torvalds 已提交
2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283
	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);

2284 2285
	strlcpy(info->driver, DRV_NAME, sizeof(info->driver));
	strlcpy(info->version, DRV_VERSION, sizeof(info->version));
L
Linus Torvalds 已提交
2286 2287 2288

	switch (tun->flags & TUN_TYPE_MASK) {
	case TUN_TUN_DEV:
2289
		strlcpy(info->bus_info, "tun", sizeof(info->bus_info));
L
Linus Torvalds 已提交
2290 2291
		break;
	case TUN_TAP_DEV:
2292
		strlcpy(info->bus_info, "tap", sizeof(info->bus_info));
L
Linus Torvalds 已提交
2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314
		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
}

2315
static const struct ethtool_ops tun_ethtool_ops = {
L
Linus Torvalds 已提交
2316 2317 2318 2319
	.get_settings	= tun_get_settings,
	.get_drvinfo	= tun_get_drvinfo,
	.get_msglevel	= tun_get_msglevel,
	.set_msglevel	= tun_set_msglevel,
2320
	.get_link	= ethtool_op_get_link,
2321
	.get_ts_info	= ethtool_op_get_ts_info,
L
Linus Torvalds 已提交
2322 2323
};

2324

L
Linus Torvalds 已提交
2325 2326 2327 2328
static int __init tun_init(void)
{
	int ret = 0;

2329 2330
	pr_info("%s, %s\n", DRV_DESCRIPTION, DRV_VERSION);
	pr_info("%s\n", DRV_COPYRIGHT);
L
Linus Torvalds 已提交
2331

2332
	ret = rtnl_link_register(&tun_link_ops);
2333
	if (ret) {
2334
		pr_err("Can't register link_ops\n");
2335
		goto err_linkops;
2336 2337
	}

L
Linus Torvalds 已提交
2338
	ret = misc_register(&tun_miscdev);
2339
	if (ret) {
2340
		pr_err("Can't register misc device %d\n", TUN_MINOR);
2341 2342
		goto err_misc;
	}
2343
	return  0;
2344
err_misc:
2345 2346
	rtnl_link_unregister(&tun_link_ops);
err_linkops:
L
Linus Torvalds 已提交
2347 2348 2349 2350 2351
	return ret;
}

static void tun_cleanup(void)
{
2352
	misc_deregister(&tun_miscdev);
2353
	rtnl_link_unregister(&tun_link_ops);
L
Linus Torvalds 已提交
2354 2355
}

2356 2357 2358 2359 2360 2361
/* 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)
{
2362
	struct tun_file *tfile;
2363 2364
	if (file->f_op != &tun_fops)
		return ERR_PTR(-EINVAL);
2365 2366
	tfile = file->private_data;
	if (!tfile)
2367
		return ERR_PTR(-EBADFD);
J
Jason Wang 已提交
2368
	return &tfile->socket;
2369 2370 2371
}
EXPORT_SYMBOL_GPL(tun_get_socket);

L
Linus Torvalds 已提交
2372 2373 2374 2375 2376 2377
module_init(tun_init);
module_exit(tun_cleanup);
MODULE_DESCRIPTION(DRV_DESCRIPTION);
MODULE_AUTHOR(DRV_COPYRIGHT);
MODULE_LICENSE("GPL");
MODULE_ALIAS_MISCDEV(TUN_MINOR);
2378
MODULE_ALIAS("devname:net/tun");