tun.c 55.6 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>
L
Linus Torvalds 已提交
74 75 76

#include <asm/uaccess.h>

77 78 79
/* Uncomment to enable debugging */
/* #define TUN_DEBUG 1 */

L
Linus Torvalds 已提交
80 81
#ifdef TUN_DEBUG
static int debug;
82

83 84 85 86 87 88 89 90 91 92
#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)
93
#else
94 95 96 97 98 99 100 101 102 103
#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)
104 105
#endif

106 107
#define GOODCOPY_LEN 128

108 109 110 111 112 113 114
#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
115
/* DEFAULT_MAX_NUM_RSS_QUEUES were chosen to let the rx/tx queues allocated for
116 117 118
 * 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
119
#define MAX_TAP_FLOWS  4096
J
Jason Wang 已提交
120

J
Jason Wang 已提交
121 122
#define TUN_FLOW_EXPIRE (3 * HZ)

J
Jason Wang 已提交
123
/* A tun_file connects an open character device to a tuntap netdevice. It
S
stephen hemminger 已提交
124
 * also contains all socket related structures (except sock_fprog and tap_filter)
J
Jason Wang 已提交
125 126
 * 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
127
 * netdevice not for a specific queue (at least I didn't see the requirement for
J
Jason Wang 已提交
128
 * this).
129 130
 *
 * RCU usage:
R
Rami Rosen 已提交
131
 * The tun_file and tun_struct are loosely coupled, the pointer from one to the
132
 * other can only be read while rcu_read_lock or rtnl_lock is held.
J
Jason Wang 已提交
133
 */
E
Eric W. Biederman 已提交
134
struct tun_file {
J
Jason Wang 已提交
135 136 137
	struct sock sk;
	struct socket socket;
	struct socket_wq wq;
138
	struct tun_struct __rcu *tun;
E
Eric W. Biederman 已提交
139
	struct net *net;
J
Jason Wang 已提交
140 141 142
	struct fasync_struct *fasync;
	/* only used for fasnyc */
	unsigned int flags;
143 144 145 146
	union {
		u16 queue_index;
		unsigned int ifindex;
	};
J
Jason Wang 已提交
147 148
	struct list_head next;
	struct tun_struct *detached;
E
Eric W. Biederman 已提交
149 150
};

J
Jason Wang 已提交
151 152 153 154 155 156
struct tun_flow_entry {
	struct hlist_node hash_link;
	struct rcu_head rcu;
	struct tun_struct *tun;

	u32 rxhash;
157
	u32 rps_rxhash;
J
Jason Wang 已提交
158 159 160 161 162 163
	int queue_index;
	unsigned long updated;
};

#define TUN_NUM_FLOW_ENTRIES 1024

J
Jason Wang 已提交
164
/* Since the socket were moved to tun_file, to preserve the behavior of persist
R
Rami Rosen 已提交
165
 * device, socket filter, sndbuf and vnet header size were restore when the
J
Jason Wang 已提交
166 167
 * file were attached to a persist device.
 */
168
struct tun_struct {
J
Jason Wang 已提交
169 170
	struct tun_file __rcu	*tfiles[MAX_TAP_QUEUES];
	unsigned int            numqueues;
171
	unsigned int 		flags;
172 173
	kuid_t			owner;
	kgid_t			group;
174 175

	struct net_device	*dev;
176
	netdev_features_t	set_features;
177
#define TUN_USER_FEATURES (NETIF_F_HW_CSUM|NETIF_F_TSO_ECN|NETIF_F_TSO| \
178
			  NETIF_F_TSO6)
179 180

	int			vnet_hdr_sz;
J
Jason Wang 已提交
181 182 183 184 185
	int			sndbuf;
	struct tap_filter	txflt;
	struct sock_fprog	fprog;
	/* protected by rtnl lock */
	bool			filter_attached;
186 187
#ifdef TUN_DEBUG
	int debug;
L
Linus Torvalds 已提交
188
#endif
J
Jason Wang 已提交
189 190 191 192
	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
193 194
	unsigned int numdisabled;
	struct list_head disabled;
195
	void *security;
196
	u32 flow_count;
197
};
L
Linus Torvalds 已提交
198

J
Jason Wang 已提交
199 200 201 202 203 204 205 206 207
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;

208
	hlist_for_each_entry_rcu(e, head, hash_link) {
J
Jason Wang 已提交
209 210 211 212 213 214 215 216 217 218
		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)
{
219 220
	struct tun_flow_entry *e = kmalloc(sizeof(*e), GFP_ATOMIC);

J
Jason Wang 已提交
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;
226
		e->rps_rxhash = 0;
J
Jason Wang 已提交
227 228 229
		e->queue_index = queue_index;
		e->tun = tun;
		hlist_add_head_rcu(&e->hash_link, head);
230
		++tun->flow_count;
J
Jason Wang 已提交
231 232 233 234 235 236 237 238
	}
	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);
239
	sock_rps_reset_flow_hash(e->rps_rxhash);
J
Jason Wang 已提交
240
	hlist_del_rcu(&e->hash_link);
241
	kfree_rcu(e, rcu);
242
	--tun->flow_count;
J
Jason Wang 已提交
243 244 245 246 247 248 249 250 251
}

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;
252
		struct hlist_node *n;
J
Jason Wang 已提交
253

254
		hlist_for_each_entry_safe(e, n, &tun->flows[i], hash_link)
J
Jason Wang 已提交
255 256 257 258 259 260 261 262 263 264 265 266
			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;
267
		struct hlist_node *n;
J
Jason Wang 已提交
268

269
		hlist_for_each_entry_safe(e, n, &tun->flows[i], hash_link) {
J
Jason Wang 已提交
270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289
			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;
290
		struct hlist_node *n;
J
Jason Wang 已提交
291

292
		hlist_for_each_entry_safe(e, n, &tun->flows[i], hash_link) {
J
Jason Wang 已提交
293 294 295 296 297 298 299 300 301 302 303 304 305 306 307
			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);
}

308
static void tun_flow_update(struct tun_struct *tun, u32 rxhash,
309
			    struct tun_file *tfile)
J
Jason Wang 已提交
310 311 312 313
{
	struct hlist_head *head;
	struct tun_flow_entry *e;
	unsigned long delay = tun->ageing_time;
314
	u16 queue_index = tfile->queue_index;
J
Jason Wang 已提交
315 316 317 318 319 320 321 322

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

	rcu_read_lock();

323 324 325
	/* 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
326 327 328 329 330 331 332
		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;
333
		sock_rps_record_flow_hash(e->rps_rxhash);
J
Jason Wang 已提交
334 335
	} else {
		spin_lock_bh(&tun->lock);
336 337
		if (!tun_flow_find(head, rxhash) &&
		    tun->flow_count < MAX_TAP_FLOWS)
J
Jason Wang 已提交
338 339 340 341 342 343 344 345 346 347 348 349
			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();
}

350 351 352 353 354 355 356 357 358 359 360 361
/**
 * 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
362
/* We try to identify a flow through its rxhash first. The reason that
S
stephen hemminger 已提交
363
 * we do not check rxq no. is because some cards(e.g 82599), chooses
J
Jason Wang 已提交
364 365 366 367 368
 * 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.
 */
369
static u16 tun_select_queue(struct net_device *dev, struct sk_buff *skb,
370
			    void *accel_priv, select_queue_fallback_t fallback)
J
Jason Wang 已提交
371 372
{
	struct tun_struct *tun = netdev_priv(dev);
J
Jason Wang 已提交
373
	struct tun_flow_entry *e;
J
Jason Wang 已提交
374 375 376 377
	u32 txq = 0;
	u32 numqueues = 0;

	rcu_read_lock();
378
	numqueues = ACCESS_ONCE(tun->numqueues);
J
Jason Wang 已提交
379

380
	txq = skb_get_hash(skb);
J
Jason Wang 已提交
381
	if (txq) {
J
Jason Wang 已提交
382
		e = tun_flow_find(&tun->flows[tun_hashfn(txq)], txq);
383 384
		if (e) {
			tun_flow_save_rps_rxhash(e, txq);
385
			txq = e->queue_index;
386
		} else
J
Jason Wang 已提交
387 388
			/* use multiply and shift instead of expensive divide */
			txq = ((u64)txq * numqueues) >> 32;
J
Jason Wang 已提交
389 390 391 392 393 394 395 396 397 398
	} 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;
}

399 400 401
static inline bool tun_not_capable(struct tun_struct *tun)
{
	const struct cred *cred = current_cred();
402
	struct net *net = dev_net(tun->dev);
403 404 405

	return ((uid_valid(tun->owner) && !uid_eq(cred->euid, tun->owner)) ||
		  (gid_valid(tun->group) && !in_egroup_p(tun->group))) &&
406
		!ns_capable(net->user_ns, CAP_NET_ADMIN);
407 408
}

J
Jason Wang 已提交
409 410 411 412 413 414
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
415 416 417 418 419 420 421
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
422
static struct tun_struct *tun_enable_queue(struct tun_file *tfile)
J
Jason Wang 已提交
423 424 425 426 427 428 429 430 431
{
	struct tun_struct *tun = tfile->detached;

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

432 433 434 435 436 437
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
438 439 440 441 442
static void __tun_detach(struct tun_file *tfile, bool clean)
{
	struct tun_file *ntfile;
	struct tun_struct *tun;

443 444
	tun = rtnl_dereference(tfile->tun);

445
	if (tun && !tfile->detached) {
J
Jason Wang 已提交
446 447 448 449 450
		u16 index = tfile->queue_index;
		BUG_ON(index >= tun->numqueues);

		rcu_assign_pointer(tun->tfiles[index],
				   tun->tfiles[tun->numqueues - 1]);
451
		ntfile = rtnl_dereference(tun->tfiles[index]);
J
Jason Wang 已提交
452 453 454
		ntfile->queue_index = index;

		--tun->numqueues;
455
		if (clean) {
456
			RCU_INIT_POINTER(tfile->tun, NULL);
J
Jason Wang 已提交
457
			sock_put(&tfile->sk);
458
		} else
J
Jason Wang 已提交
459
			tun_disable_queue(tun, tfile);
J
Jason Wang 已提交
460 461

		synchronize_net();
J
Jason Wang 已提交
462
		tun_flow_delete_by_queue(tun, tun->numqueues + 1);
J
Jason Wang 已提交
463
		/* Drop read queue */
464
		tun_queue_purge(tfile);
J
Jason Wang 已提交
465
		tun_set_real_num_queues(tun);
J
Jason Wang 已提交
466
	} else if (tfile->detached && clean) {
J
Jason Wang 已提交
467
		tun = tun_enable_queue(tfile);
J
Jason Wang 已提交
468 469
		sock_put(&tfile->sk);
	}
J
Jason Wang 已提交
470 471

	if (clean) {
472 473 474 475 476
		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
477
				unregister_netdevice(tun->dev);
478
		}
J
Jason Wang 已提交
479

J
Jason Wang 已提交
480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495
		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
496
	struct tun_file *tfile, *tmp;
J
Jason Wang 已提交
497 498 499
	int i, n = tun->numqueues;

	for (i = 0; i < n; i++) {
500
		tfile = rtnl_dereference(tun->tfiles[i]);
J
Jason Wang 已提交
501
		BUG_ON(!tfile);
502
		tfile->socket.sk->sk_data_ready(tfile->socket.sk);
503
		RCU_INIT_POINTER(tfile->tun, NULL);
J
Jason Wang 已提交
504 505
		--tun->numqueues;
	}
506
	list_for_each_entry(tfile, &tun->disabled, next) {
507
		tfile->socket.sk->sk_data_ready(tfile->socket.sk);
508
		RCU_INIT_POINTER(tfile->tun, NULL);
509
	}
J
Jason Wang 已提交
510 511 512 513
	BUG_ON(tun->numqueues != 0);

	synchronize_net();
	for (i = 0; i < n; i++) {
514
		tfile = rtnl_dereference(tun->tfiles[i]);
J
Jason Wang 已提交
515
		/* Drop read queue */
516
		tun_queue_purge(tfile);
J
Jason Wang 已提交
517 518
		sock_put(&tfile->sk);
	}
J
Jason Wang 已提交
519 520
	list_for_each_entry_safe(tfile, tmp, &tun->disabled, next) {
		tun_enable_queue(tfile);
521
		tun_queue_purge(tfile);
J
Jason Wang 已提交
522 523 524
		sock_put(&tfile->sk);
	}
	BUG_ON(tun->numdisabled != 0);
J
Jason Wang 已提交
525 526 527

	if (tun->flags & TUN_PERSIST)
		module_put(THIS_MODULE);
J
Jason Wang 已提交
528 529
}

530
static int tun_attach(struct tun_struct *tun, struct file *file, bool skip_filter)
531
{
E
Eric W. Biederman 已提交
532
	struct tun_file *tfile = file->private_data;
533
	int err;
534

535 536 537 538
	err = security_tun_dev_attach(tfile->socket.sk, tun->security);
	if (err < 0)
		goto out;

539
	err = -EINVAL;
540
	if (rtnl_dereference(tfile->tun) && !tfile->detached)
541 542 543
		goto out;

	err = -EBUSY;
J
Jason Wang 已提交
544 545 546 547
	if (!(tun->flags & TUN_TAP_MQ) && tun->numqueues == 1)
		goto out;

	err = -E2BIG;
J
Jason Wang 已提交
548 549
	if (!tfile->detached &&
	    tun->numqueues + tun->numdisabled == MAX_TAP_QUEUES)
550 551 552
		goto out;

	err = 0;
J
Jason Wang 已提交
553

S
stephen hemminger 已提交
554
	/* Re-attach the filter to persist device */
555
	if (!skip_filter && (tun->filter_attached == true)) {
J
Jason Wang 已提交
556 557 558 559
		err = sk_attach_filter(&tun->fprog, tfile->socket.sk);
		if (!err)
			goto out;
	}
J
Jason Wang 已提交
560
	tfile->queue_index = tun->numqueues;
561
	rcu_assign_pointer(tfile->tun, tun);
J
Jason Wang 已提交
562 563
	rcu_assign_pointer(tun->tfiles[tun->numqueues], tfile);
	tun->numqueues++;
564

J
Jason Wang 已提交
565 566 567 568 569
	if (tfile->detached)
		tun_enable_queue(tfile);
	else
		sock_hold(&tfile->sk);

J
Jason Wang 已提交
570
	tun_set_real_num_queues(tun);
571

J
Jason Wang 已提交
572 573 574 575 576 577
	/* device is allowed to go away first, so no need to hold extra
	 * refcnt.
	 */

out:
	return err;
E
Eric W. Biederman 已提交
578 579 580 581
}

static struct tun_struct *__tun_get(struct tun_file *tfile)
{
582
	struct tun_struct *tun;
583

584 585 586 587 588
	rcu_read_lock();
	tun = rcu_dereference(tfile->tun);
	if (tun)
		dev_hold(tun->dev);
	rcu_read_unlock();
589 590

	return tun;
E
Eric W. Biederman 已提交
591 592 593 594 595 596 597 598 599
}

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

static void tun_put(struct tun_struct *tun)
{
600
	dev_put(tun->dev);
E
Eric W. Biederman 已提交
601 602
}

603
/* TAP filtering */
604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652
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;

653 654
	/* Remaining multicast addresses are hashed,
	 * unicast will leave the filter disabled. */
655
	memset(filter->mask, 0, sizeof(filter->mask));
656 657 658 659 660
	for (; n < uf.count; n++) {
		if (!is_multicast_ether_addr(addr[n].u)) {
			err = 0; /* no filter */
			goto done;
		}
661
		addr_hash_set(filter->mask, addr[n].u);
662
	}
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

	/* 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++)
691
		if (ether_addr_equal(eh->h_dest, filter->addr[i]))
692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712
			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
713 714
/* Network device part of the driver */

715
static const struct ethtool_ops tun_ethtool_ops;
L
Linus Torvalds 已提交
716

717 718 719
/* Net device detach from fd. */
static void tun_net_uninit(struct net_device *dev)
{
J
Jason Wang 已提交
720
	tun_detach_all(dev);
721 722
}

L
Linus Torvalds 已提交
723 724 725
/* Net device open. */
static int tun_net_open(struct net_device *dev)
{
J
Jason Wang 已提交
726
	netif_tx_start_all_queues(dev);
L
Linus Torvalds 已提交
727 728 729 730 731 732
	return 0;
}

/* Net device close. */
static int tun_net_close(struct net_device *dev)
{
J
Jason Wang 已提交
733
	netif_tx_stop_all_queues(dev);
L
Linus Torvalds 已提交
734 735 736 737
	return 0;
}

/* Net device start xmit */
738
static netdev_tx_t tun_net_xmit(struct sk_buff *skb, struct net_device *dev)
L
Linus Torvalds 已提交
739 740
{
	struct tun_struct *tun = netdev_priv(dev);
J
Jason Wang 已提交
741
	int txq = skb->queue_mapping;
742
	struct tun_file *tfile;
743
	u32 numqueues = 0;
L
Linus Torvalds 已提交
744

745
	rcu_read_lock();
J
Jason Wang 已提交
746
	tfile = rcu_dereference(tun->tfiles[txq]);
747
	numqueues = ACCESS_ONCE(tun->numqueues);
J
Jason Wang 已提交
748

L
Linus Torvalds 已提交
749
	/* Drop packet if interface is not attached */
750
	if (txq >= numqueues)
L
Linus Torvalds 已提交
751 752
		goto drop;

753
	if (numqueues == 1) {
754 755 756 757 758 759 760 761 762 763 764 765 766 767 768
		/* 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);
		}
	}

769 770
	tun_debug(KERN_INFO, tun, "tun_net_xmit %d\n", skb->len);

J
Jason Wang 已提交
771 772
	BUG_ON(!tfile);

773 774 775 776 777 778
	/* 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
779 780
	if (tfile->socket.sk->sk_filter &&
	    sk_filter(tfile->socket.sk, skb))
M
Michael S. Tsirkin 已提交
781 782
		goto drop;

R
Rami Rosen 已提交
783
	/* Limit the number of packets queued by dividing txq length with the
J
Jason Wang 已提交
784 785
	 * number of queues.
	 */
786 787
	if (skb_queue_len(&tfile->socket.sk->sk_receive_queue) * numqueues
			  >= dev->tx_queue_len)
788
		goto drop;
L
Linus Torvalds 已提交
789

790 791 792
	if (unlikely(skb_orphan_frags(skb, GFP_ATOMIC)))
		goto drop;

793 794 795 796 797
	if (skb->sk) {
		sock_tx_timestamp(skb->sk, &skb_shinfo(skb)->tx_flags);
		sw_tx_timestamp(skb);
	}

M
Michael S. Tsirkin 已提交
798
	/* Orphan the skb - required as we might hang on to it
799 800
	 * for indefinite time.
	 */
M
Michael S. Tsirkin 已提交
801 802
	skb_orphan(skb);

803 804
	nf_reset(skb);

805
	/* Enqueue packet */
J
Jason Wang 已提交
806
	skb_queue_tail(&tfile->socket.sk->sk_receive_queue, skb);
L
Linus Torvalds 已提交
807 808

	/* Notify and wake up reader process */
J
Jason Wang 已提交
809 810
	if (tfile->flags & TUN_FASYNC)
		kill_fasync(&tfile->fasync, SIGIO, POLL_IN);
811
	tfile->socket.sk->sk_data_ready(tfile->socket.sk);
812 813

	rcu_read_unlock();
814
	return NETDEV_TX_OK;
L
Linus Torvalds 已提交
815 816

drop:
817
	dev->stats.tx_dropped++;
818
	skb_tx_error(skb);
L
Linus Torvalds 已提交
819
	kfree_skb(skb);
820
	rcu_read_unlock();
821
	return NETDEV_TX_OK;
L
Linus Torvalds 已提交
822 823
}

824
static void tun_net_mclist(struct net_device *dev)
L
Linus Torvalds 已提交
825
{
826 827 828 829 830
	/*
	 * 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
831 832
}

E
Ed Swierk 已提交
833 834 835 836 837 838 839 840 841 842 843 844
#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;
}

845 846
static netdev_features_t tun_net_fix_features(struct net_device *dev,
	netdev_features_t features)
847 848 849 850 851
{
	struct tun_struct *tun = netdev_priv(dev);

	return (features & tun->set_features) | (features & ~TUN_USER_FEATURES);
}
852 853 854 855 856 857 858
#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
859
	 * Since both of those are synchronous operations, we are guaranteed
860
	 * never to have pending data when we poll for it
S
stephen hemminger 已提交
861
	 * so there is nothing to do here but return.
862 863 864 865 866 867 868
	 * 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
869
static const struct net_device_ops tun_netdev_ops = {
870
	.ndo_uninit		= tun_net_uninit,
S
Stephen Hemminger 已提交
871 872
	.ndo_open		= tun_net_open,
	.ndo_stop		= tun_net_close,
873
	.ndo_start_xmit		= tun_net_xmit,
S
Stephen Hemminger 已提交
874
	.ndo_change_mtu		= tun_net_change_mtu,
875
	.ndo_fix_features	= tun_net_fix_features,
J
Jason Wang 已提交
876
	.ndo_select_queue	= tun_select_queue,
877 878 879
#ifdef CONFIG_NET_POLL_CONTROLLER
	.ndo_poll_controller	= tun_poll_controller,
#endif
S
Stephen Hemminger 已提交
880 881 882
};

static const struct net_device_ops tap_netdev_ops = {
883
	.ndo_uninit		= tun_net_uninit,
S
Stephen Hemminger 已提交
884 885
	.ndo_open		= tun_net_open,
	.ndo_stop		= tun_net_close,
886
	.ndo_start_xmit		= tun_net_xmit,
S
Stephen Hemminger 已提交
887
	.ndo_change_mtu		= tun_net_change_mtu,
888
	.ndo_fix_features	= tun_net_fix_features,
889
	.ndo_set_rx_mode	= tun_net_mclist,
S
Stephen Hemminger 已提交
890 891
	.ndo_set_mac_address	= eth_mac_addr,
	.ndo_validate_addr	= eth_validate_addr,
J
Jason Wang 已提交
892
	.ndo_select_queue	= tun_select_queue,
893 894 895
#ifdef CONFIG_NET_POLL_CONTROLLER
	.ndo_poll_controller	= tun_poll_controller,
#endif
S
Stephen Hemminger 已提交
896 897
};

898
static void tun_flow_init(struct tun_struct *tun)
J
Jason Wang 已提交
899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916
{
	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
917 918 919 920
/* Initialize net device. */
static void tun_net_init(struct net_device *dev)
{
	struct tun_struct *tun = netdev_priv(dev);
921

L
Linus Torvalds 已提交
922 923
	switch (tun->flags & TUN_TYPE_MASK) {
	case TUN_TUN_DEV:
S
Stephen Hemminger 已提交
924 925
		dev->netdev_ops = &tun_netdev_ops;

L
Linus Torvalds 已提交
926 927 928 929 930 931
		/* Point-to-Point TUN Device */
		dev->hard_header_len = 0;
		dev->addr_len = 0;
		dev->mtu = 1500;

		/* Zero header length */
932
		dev->type = ARPHRD_NONE;
L
Linus Torvalds 已提交
933 934 935 936 937
		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
938
		dev->netdev_ops = &tap_netdev_ops;
L
Linus Torvalds 已提交
939 940
		/* Ethernet TAP Device */
		ether_setup(dev);
941
		dev->priv_flags &= ~IFF_TX_SKB_SHARING;
942
		dev->priv_flags |= IFF_LIVE_ADDR_CHANGE;
943

944
		eth_hw_addr_random(dev);
945

L
Linus Torvalds 已提交
946 947 948 949 950 951 952 953
		dev->tx_queue_len = TUN_READQ_SIZE;  /* We prefer our own queue length */
		break;
	}
}

/* Character device part */

/* Poll */
J
Jason Wang 已提交
954
static unsigned int tun_chr_poll(struct file *file, poll_table *wait)
955
{
956 957
	struct tun_file *tfile = file->private_data;
	struct tun_struct *tun = __tun_get(tfile);
958
	struct sock *sk;
959
	unsigned int mask = 0;
L
Linus Torvalds 已提交
960 961

	if (!tun)
962
		return POLLERR;
L
Linus Torvalds 已提交
963

J
Jason Wang 已提交
964
	sk = tfile->socket.sk;
965

966
	tun_debug(KERN_INFO, tun, "tun_chr_poll\n");
L
Linus Torvalds 已提交
967

968
	poll_wait(file, sk_sleep(sk), wait);
969

970
	if (!skb_queue_empty(&sk->sk_receive_queue))
L
Linus Torvalds 已提交
971 972
		mask |= POLLIN | POLLRDNORM;

973 974 975 976 977
	if (sock_writeable(sk) ||
	    (!test_and_set_bit(SOCK_ASYNC_NOSPACE, &sk->sk_socket->flags) &&
	     sock_writeable(sk)))
		mask |= POLLOUT | POLLWRNORM;

978 979 980
	if (tun->dev->reg_state != NETREG_REGISTERED)
		mask = POLLERR;

E
Eric W. Biederman 已提交
981
	tun_put(tun);
L
Linus Torvalds 已提交
982 983 984
	return mask;
}

985 986
/* 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
987
static struct sk_buff *tun_alloc_skb(struct tun_file *tfile,
988 989
				     size_t prepad, size_t len,
				     size_t linear, int noblock)
990
{
J
Jason Wang 已提交
991
	struct sock *sk = tfile->socket.sk;
992
	struct sk_buff *skb;
993
	int err;
994 995

	/* Under a page?  Don't bother with paged skb. */
H
Herbert Xu 已提交
996
	if (prepad + len < PAGE_SIZE || !linear)
997
		linear = len;
998

999
	skb = sock_alloc_send_pskb(sk, prepad + linear, len - linear, noblock,
1000
				   &err, 0);
1001
	if (!skb)
1002
		return ERR_PTR(err);
1003 1004 1005

	skb_reserve(skb, prepad);
	skb_put(skb, linear);
1006 1007
	skb->data_len = len - linear;
	skb->len += len - linear;
1008 1009 1010 1011

	return skb;
}

L
Linus Torvalds 已提交
1012
/* Get packet from user space buffer */
J
Jason Wang 已提交
1013 1014 1015
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
1016
{
1017
	struct tun_pi pi = { 0, cpu_to_be16(ETH_P_IP) };
L
Linus Torvalds 已提交
1018
	struct sk_buff *skb;
1019
	size_t len = total_len, align = NET_SKB_PAD, linear;
1020
	struct virtio_net_hdr gso = { 0 };
1021
	int good_linear;
1022
	int offset = 0;
1023 1024 1025
	int copylen;
	bool zerocopy = false;
	int err;
1026
	u32 rxhash;
L
Linus Torvalds 已提交
1027 1028

	if (!(tun->flags & TUN_NO_PI)) {
1029
		if (len < sizeof(pi))
L
Linus Torvalds 已提交
1030
			return -EINVAL;
1031
		len -= sizeof(pi);
L
Linus Torvalds 已提交
1032

1033
		if (memcpy_fromiovecend((void *)&pi, iv, 0, sizeof(pi)))
L
Linus Torvalds 已提交
1034
			return -EFAULT;
1035
		offset += sizeof(pi);
L
Linus Torvalds 已提交
1036 1037
	}

1038
	if (tun->flags & TUN_VNET_HDR) {
1039
		if (len < tun->vnet_hdr_sz)
1040
			return -EINVAL;
1041
		len -= tun->vnet_hdr_sz;
1042

1043
		if (memcpy_fromiovecend((void *)&gso, iv, offset, sizeof(gso)))
1044 1045
			return -EFAULT;

1046 1047 1048 1049
		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;

1050 1051
		if (gso.hdr_len > len)
			return -EINVAL;
1052
		offset += tun->vnet_hdr_sz;
1053 1054
	}

1055
	if ((tun->flags & TUN_TYPE_MASK) == TUN_TAP_DEV) {
1056
		align += NET_IP_ALIGN;
H
Herbert Xu 已提交
1057 1058
		if (unlikely(len < ETH_HLEN ||
			     (gso.hdr_len && gso.hdr_len < ETH_HLEN)))
1059 1060
			return -EINVAL;
	}
1061

1062 1063
	good_linear = SKB_MAX_HEAD(align);

1064 1065 1066
	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.
1067 1068
		 * The rest of the buffer is mapped from userspace.
		 */
1069
		copylen = gso.hdr_len ? gso.hdr_len : GOODCOPY_LEN;
1070 1071
		if (copylen > good_linear)
			copylen = good_linear;
1072
		linear = copylen;
1073 1074 1075 1076 1077
		if (iov_pages(iv, offset + copylen, count) <= MAX_SKB_FRAGS)
			zerocopy = true;
	}

	if (!zerocopy) {
1078
		copylen = len;
1079 1080 1081 1082
		if (gso.hdr_len > good_linear)
			linear = good_linear;
		else
			linear = gso.hdr_len;
1083
	}
1084

1085
	skb = tun_alloc_skb(tfile, align, copylen, linear, noblock);
1086 1087 1088 1089
	if (IS_ERR(skb)) {
		if (PTR_ERR(skb) != -EAGAIN)
			tun->dev->stats.rx_dropped++;
		return PTR_ERR(skb);
L
Linus Torvalds 已提交
1090 1091
	}

1092 1093
	if (zerocopy)
		err = zerocopy_sg_from_iovec(skb, iv, offset, count);
1094
	else {
1095
		err = skb_copy_datagram_from_iovec(skb, 0, iv, offset, len);
1096 1097 1098 1099 1100
		if (!err && msg_control) {
			struct ubuf_info *uarg = msg_control;
			uarg->callback(uarg, false);
		}
	}
1101 1102

	if (err) {
1103
		tun->dev->stats.rx_dropped++;
D
Dave Jones 已提交
1104
		kfree_skb(skb);
L
Linus Torvalds 已提交
1105
		return -EFAULT;
D
Dave Jones 已提交
1106
	}
L
Linus Torvalds 已提交
1107

1108 1109 1110 1111 1112 1113 1114
	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;
		}
1115
	}
1116

L
Linus Torvalds 已提交
1117 1118
	switch (tun->flags & TUN_TYPE_MASK) {
	case TUN_TUN_DEV:
1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133
		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;
			}
		}

1134
		skb_reset_mac_header(skb);
L
Linus Torvalds 已提交
1135
		skb->protocol = pi.proto;
1136
		skb->dev = tun->dev;
L
Linus Torvalds 已提交
1137 1138 1139 1140
		break;
	case TUN_TAP_DEV:
		skb->protocol = eth_type_trans(skb, tun->dev);
		break;
1141
	}
L
Linus Torvalds 已提交
1142

1143 1144
	skb_reset_network_header(skb);

1145 1146 1147 1148
	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:
1149
			skb_shinfo(skb)->gso_type = SKB_GSO_TCPV4;
1150 1151
			break;
		case VIRTIO_NET_HDR_GSO_TCPV6:
1152
			skb_shinfo(skb)->gso_type = SKB_GSO_TCPV6;
1153
			break;
1154
		case VIRTIO_NET_HDR_GSO_UDP:
1155 1156 1157 1158 1159 1160 1161 1162 1163
		{
			static bool warned;

			if (!warned) {
				warned = true;
				netdev_warn(tun->dev,
					    "%s: using disabled UFO feature; please fix this program\n",
					    current->comm);
			}
1164
			skb_shinfo(skb)->gso_type = SKB_GSO_UDP;
1165 1166
			if (skb->protocol == htons(ETH_P_IPV6))
				ipv6_proxy_select_ident(skb);
1167
			break;
1168
		}
1169 1170 1171 1172 1173 1174 1175
		default:
			tun->dev->stats.rx_frame_errors++;
			kfree_skb(skb);
			return -EINVAL;
		}

		if (gso.gso_type & VIRTIO_NET_HDR_GSO_ECN)
1176
			skb_shinfo(skb)->gso_type |= SKB_GSO_TCP_ECN;
1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188

		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;
	}
1189

1190 1191 1192 1193
	/* 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;
1194
		skb_shinfo(skb)->tx_flags |= SKBTX_SHARED_FRAG;
1195 1196
	}

1197
	skb_probe_transport_header(skb, 0);
1198

1199
	rxhash = skb_get_hash(skb);
L
Linus Torvalds 已提交
1200
	netif_rx_ni(skb);
1201

1202 1203
	tun->dev->stats.rx_packets++;
	tun->dev->stats.rx_bytes += len;
L
Linus Torvalds 已提交
1204

1205
	tun_flow_update(tun, rxhash, tfile);
1206
	return total_len;
1207
}
L
Linus Torvalds 已提交
1208

1209 1210
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
1211
{
1212
	struct file *file = iocb->ki_filp;
H
Herbert Xu 已提交
1213
	struct tun_struct *tun = tun_get(file);
J
Jason Wang 已提交
1214
	struct tun_file *tfile = file->private_data;
E
Eric W. Biederman 已提交
1215
	ssize_t result;
L
Linus Torvalds 已提交
1216 1217 1218 1219

	if (!tun)
		return -EBADFD;

1220
	tun_debug(KERN_INFO, tun, "tun_chr_write %ld\n", count);
L
Linus Torvalds 已提交
1221

J
Jason Wang 已提交
1222 1223
	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
1224 1225 1226

	tun_put(tun);
	return result;
L
Linus Torvalds 已提交
1227 1228 1229
}

/* Put packet to the user space buffer */
1230
static ssize_t tun_put_user(struct tun_struct *tun,
J
Jason Wang 已提交
1231
			    struct tun_file *tfile,
1232 1233
			    struct sk_buff *skb,
			    const struct iovec *iv, int len)
L
Linus Torvalds 已提交
1234 1235 1236
{
	struct tun_pi pi = { 0, skb->protocol };
	ssize_t total = 0;
1237
	int vlan_offset = 0, copied;
L
Linus Torvalds 已提交
1238 1239 1240 1241 1242 1243 1244 1245 1246

	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;
		}
1247

1248
		if (memcpy_toiovecend(iv, (void *) &pi, 0, sizeof(pi)))
L
Linus Torvalds 已提交
1249 1250
			return -EFAULT;
		total += sizeof(pi);
1251
	}
L
Linus Torvalds 已提交
1252

1253 1254
	if (tun->flags & TUN_VNET_HDR) {
		struct virtio_net_hdr gso = { 0 }; /* no info leak */
1255
		if ((len -= tun->vnet_hdr_sz) < 0)
1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267
			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;
1268
			else {
1269
				pr_err("unexpected GSO type: "
1270 1271 1272 1273 1274 1275 1276 1277 1278 1279
				       "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;
			}
1280 1281 1282 1283 1284 1285 1286
			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;
1287
			gso.csum_start = skb_checksum_start_offset(skb);
1288
			gso.csum_offset = skb->csum_offset;
1289 1290
		} else if (skb->ip_summed == CHECKSUM_UNNECESSARY) {
			gso.flags = VIRTIO_NET_HDR_F_DATA_VALID;
1291 1292
		} /* else everything is zero */

1293 1294
		if (unlikely(memcpy_toiovecend(iv, (void *)&gso, total,
					       sizeof(gso))))
1295
			return -EFAULT;
1296
		total += tun->vnet_hdr_sz;
1297 1298
	}

1299 1300
	copied = total;
	total += skb->len;
J
Jason Wang 已提交
1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314
	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);
1315
		total += VLAN_HLEN;
J
Jason Wang 已提交
1316 1317

		copy = min_t(int, vlan_offset, len);
1318
		ret = skb_copy_datagram_const_iovec(skb, 0, iv, copied, copy);
J
Jason Wang 已提交
1319
		len -= copy;
1320
		copied += copy;
J
Jason Wang 已提交
1321 1322 1323 1324
		if (ret || !len)
			goto done;

		copy = min_t(int, sizeof(veth), len);
1325
		ret = memcpy_toiovecend(iv, (void *)&veth, copied, copy);
J
Jason Wang 已提交
1326
		len -= copy;
1327
		copied += copy;
J
Jason Wang 已提交
1328 1329 1330
		if (ret || !len)
			goto done;
	}
L
Linus Torvalds 已提交
1331

1332
	skb_copy_datagram_const_iovec(skb, vlan_offset, iv, copied, len);
L
Linus Torvalds 已提交
1333

J
Jason Wang 已提交
1334
done:
1335 1336
	tun->dev->stats.tx_packets++;
	tun->dev->stats.tx_bytes += len;
L
Linus Torvalds 已提交
1337 1338 1339 1340

	return total;
}

J
Jason Wang 已提交
1341
static ssize_t tun_do_read(struct tun_struct *tun, struct tun_file *tfile,
1342
			   const struct iovec *iv, ssize_t len, int noblock)
L
Linus Torvalds 已提交
1343 1344
{
	struct sk_buff *skb;
1345
	ssize_t ret = 0;
1346
	int peeked, err, off = 0;
L
Linus Torvalds 已提交
1347

1348
	tun_debug(KERN_INFO, tun, "tun_do_read\n");
L
Linus Torvalds 已提交
1349

1350 1351
	if (!len)
		return ret;
L
Linus Torvalds 已提交
1352

1353 1354
	if (tun->dev->reg_state != NETREG_REGISTERED)
		return -EIO;
L
Linus Torvalds 已提交
1355

1356 1357 1358 1359
	/* Read frames from queue */
	skb = __skb_recv_datagram(tfile->socket.sk, noblock ? MSG_DONTWAIT : 0,
				  &peeked, &off, &err);
	if (skb) {
J
Jason Wang 已提交
1360
		ret = tun_put_user(tun, tfile, skb, iv, len);
1361
		kfree_skb(skb);
1362 1363
	} else
		ret = err;
L
Linus Torvalds 已提交
1364

1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383
	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;
	}

1384
	ret = tun_do_read(tun, tfile, iv, len,
J
Jason Wang 已提交
1385
			  file->f_flags & O_NONBLOCK);
1386
	ret = min_t(ssize_t, ret, len);
Z
Zhi Yong Wu 已提交
1387 1388
	if (ret > 0)
		iocb->ki_pos = ret;
E
Eric W. Biederman 已提交
1389 1390
out:
	tun_put(tun);
L
Linus Torvalds 已提交
1391 1392 1393
	return ret;
}

J
Jason Wang 已提交
1394 1395 1396 1397
static void tun_free_netdev(struct net_device *dev)
{
	struct tun_struct *tun = netdev_priv(dev);

J
Jason Wang 已提交
1398
	BUG_ON(!(list_empty(&tun->disabled)));
J
Jason Wang 已提交
1399
	tun_flow_uninit(tun);
1400
	security_tun_dev_free_security(tun->security);
J
Jason Wang 已提交
1401 1402 1403
	free_netdev(dev);
}

L
Linus Torvalds 已提交
1404 1405 1406 1407
static void tun_setup(struct net_device *dev)
{
	struct tun_struct *tun = netdev_priv(dev);

1408 1409
	tun->owner = INVALID_UID;
	tun->group = INVALID_GID;
L
Linus Torvalds 已提交
1410 1411

	dev->ethtool_ops = &tun_ethtool_ops;
J
Jason Wang 已提交
1412
	dev->destructor = tun_free_netdev;
L
Linus Torvalds 已提交
1413 1414
}

1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429
/* 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,
};

1430 1431
static void tun_sock_write_space(struct sock *sk)
{
J
Jason Wang 已提交
1432
	struct tun_file *tfile;
1433
	wait_queue_head_t *wqueue;
1434 1435 1436 1437 1438 1439 1440

	if (!sock_writeable(sk))
		return;

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

1441 1442 1443
	wqueue = sk_sleep(sk);
	if (wqueue && waitqueue_active(wqueue))
		wake_up_interruptible_sync_poll(wqueue, POLLOUT |
1444
						POLLWRNORM | POLLWRBAND);
H
Herbert Xu 已提交
1445

J
Jason Wang 已提交
1446 1447
	tfile = container_of(sk, struct tun_file, sk);
	kill_fasync(&tfile->fasync, SIGIO, POLL_OUT);
1448 1449
}

1450 1451 1452
static int tun_sendmsg(struct kiocb *iocb, struct socket *sock,
		       struct msghdr *m, size_t total_len)
{
J
Jason Wang 已提交
1453 1454 1455 1456 1457 1458 1459 1460 1461 1462
	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;
1463 1464 1465 1466 1467 1468
}

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
1469 1470
	struct tun_file *tfile = container_of(sock, struct tun_file, socket);
	struct tun_struct *tun = __tun_get(tfile);
1471
	int ret;
J
Jason Wang 已提交
1472 1473 1474 1475

	if (!tun)
		return -EBADFD;

1476
	if (flags & ~(MSG_DONTWAIT|MSG_TRUNC|MSG_ERRQUEUE)) {
1477 1478 1479
		ret = -EINVAL;
		goto out;
	}
1480 1481 1482 1483 1484
	if (flags & MSG_ERRQUEUE) {
		ret = sock_recv_errqueue(sock->sk, m, total_len,
					 SOL_PACKET, TUN_TX_TIMESTAMP);
		goto out;
	}
1485
	ret = tun_do_read(tun, tfile, m->msg_iov, total_len,
1486
			  flags & MSG_DONTWAIT);
1487 1488 1489 1490
	if (ret > total_len) {
		m->msg_flags |= MSG_TRUNC;
		ret = flags & MSG_TRUNC ? ret : total_len;
	}
1491
out:
J
Jason Wang 已提交
1492
	tun_put(tun);
1493 1494 1495
	return ret;
}

1496 1497 1498 1499 1500 1501 1502
static int tun_release(struct socket *sock)
{
	if (sock->sk)
		sock_put(sock->sk);
	return 0;
}

1503 1504 1505 1506
/* Ops structure to mimic raw sockets with tun */
static const struct proto_ops tun_socket_ops = {
	.sendmsg = tun_sendmsg,
	.recvmsg = tun_recvmsg,
1507
	.release = tun_release,
1508 1509
};

1510 1511 1512
static struct proto tun_proto = {
	.name		= "tun",
	.owner		= THIS_MODULE,
J
Jason Wang 已提交
1513
	.obj_size	= sizeof(struct tun_file),
1514
};
1515

1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527
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;

1528 1529 1530
	/* This flag has no real effect.  We track the value for backwards
	 * compatibility.
	 */
1531 1532 1533 1534 1535 1536
	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
1537 1538 1539
	if (tun->flags & TUN_TAP_MQ)
		flags |= IFF_MULTI_QUEUE;

1540 1541 1542
	if (tun->flags & TUN_PERSIST)
		flags |= IFF_PERSIST;

1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556
	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));
1557 1558 1559 1560
	return uid_valid(tun->owner)?
		sprintf(buf, "%u\n",
			from_kuid_munged(current_user_ns(), tun->owner)):
		sprintf(buf, "-1\n");
1561 1562 1563 1564 1565 1566
}

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));
1567 1568 1569 1570
	return gid_valid(tun->group) ?
		sprintf(buf, "%u\n",
			from_kgid_munged(current_user_ns(), tun->group)):
		sprintf(buf, "-1\n");
1571 1572 1573 1574 1575 1576
}

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);

1577
static int tun_set_iff(struct net *net, struct file *file, struct ifreq *ifr)
L
Linus Torvalds 已提交
1578 1579
{
	struct tun_struct *tun;
J
Jason Wang 已提交
1580
	struct tun_file *tfile = file->private_data;
L
Linus Torvalds 已提交
1581 1582 1583
	struct net_device *dev;
	int err;

1584 1585 1586
	if (tfile->detached)
		return -EINVAL;

1587 1588
	dev = __dev_get_by_name(net, ifr->ifr_name);
	if (dev) {
1589 1590
		if (ifr->ifr_flags & IFF_TUN_EXCL)
			return -EBUSY;
1591 1592 1593 1594 1595 1596 1597
		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;

1598 1599 1600 1601
		if (!!(ifr->ifr_flags & IFF_MULTI_QUEUE) !=
		    !!(tun->flags & TUN_TAP_MQ))
			return -EINVAL;

1602
		if (tun_not_capable(tun))
P
Paul Moore 已提交
1603
			return -EPERM;
1604
		err = security_tun_dev_open(tun->security);
P
Paul Moore 已提交
1605 1606 1607
		if (err < 0)
			return err;

1608
		err = tun_attach(tun, file, ifr->ifr_flags & IFF_NOFILTER);
1609 1610
		if (err < 0)
			return err;
J
Jason Wang 已提交
1611 1612

		if (tun->flags & TUN_TAP_MQ &&
1613 1614 1615 1616 1617 1618
		    (tun->numqueues + tun->numdisabled > 1)) {
			/* One or more queue has already been attached, no need
			 * to initialize the device again.
			 */
			return 0;
		}
1619
	}
L
Linus Torvalds 已提交
1620 1621 1622
	else {
		char *name;
		unsigned long flags = 0;
1623 1624
		int queues = ifr->ifr_flags & IFF_MULTI_QUEUE ?
			     MAX_TAP_QUEUES : 1;
L
Linus Torvalds 已提交
1625

1626
		if (!ns_capable(net->user_ns, CAP_NET_ADMIN))
1627
			return -EPERM;
P
Paul Moore 已提交
1628 1629 1630
		err = security_tun_dev_create();
		if (err < 0)
			return err;
1631

L
Linus Torvalds 已提交
1632 1633 1634 1635 1636 1637 1638 1639 1640
		/* 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";
1641
		} else
1642
			return -EINVAL;
1643

L
Linus Torvalds 已提交
1644 1645 1646
		if (*ifr->ifr_name)
			name = ifr->ifr_name;

J
Jason Wang 已提交
1647
		dev = alloc_netdev_mqs(sizeof(struct tun_struct), name,
1648 1649
				       NET_NAME_UNKNOWN, tun_setup, queues,
				       queues);
1650

L
Linus Torvalds 已提交
1651 1652 1653
		if (!dev)
			return -ENOMEM;

1654
		dev_net_set(dev, net);
1655
		dev->rtnl_link_ops = &tun_link_ops;
1656
		dev->ifindex = tfile->ifindex;
S
Stephen Hemminger 已提交
1657

L
Linus Torvalds 已提交
1658 1659 1660
		tun = netdev_priv(dev);
		tun->dev = dev;
		tun->flags = flags;
1661
		tun->txflt.count = 0;
1662
		tun->vnet_hdr_sz = sizeof(struct virtio_net_hdr);
1663

J
Jason Wang 已提交
1664 1665
		tun->filter_attached = false;
		tun->sndbuf = tfile->socket.sk->sk_sndbuf;
1666

J
Jason Wang 已提交
1667 1668
		spin_lock_init(&tun->lock);

1669 1670 1671
		err = security_tun_dev_alloc_security(&tun->security);
		if (err < 0)
			goto err_free_dev;
P
Paul Moore 已提交
1672

L
Linus Torvalds 已提交
1673
		tun_net_init(dev);
1674
		tun_flow_init(tun);
J
Jason Wang 已提交
1675

1676
		dev->hw_features = NETIF_F_SG | NETIF_F_FRAGLIST |
J
Jason Wang 已提交
1677 1678
				   TUN_USER_FEATURES | NETIF_F_HW_VLAN_CTAG_TX |
				   NETIF_F_HW_VLAN_STAG_TX;
1679
		dev->features = dev->hw_features;
1680 1681 1682
		dev->vlan_features = dev->features &
				     ~(NETIF_F_HW_VLAN_CTAG_TX |
				       NETIF_F_HW_VLAN_STAG_TX);
1683

J
Jason Wang 已提交
1684
		INIT_LIST_HEAD(&tun->disabled);
1685
		err = tun_attach(tun, file, false);
1686
		if (err < 0)
1687
			goto err_free_flow;
1688

L
Linus Torvalds 已提交
1689 1690
		err = register_netdevice(tun->dev);
		if (err < 0)
1691
			goto err_detach;
1692

1693 1694 1695
		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))
1696
			pr_err("Failed to create tun sysfs files\n");
L
Linus Torvalds 已提交
1697 1698
	}

1699 1700
	netif_carrier_on(tun->dev);

1701
	tun_debug(KERN_INFO, tun, "tun_set_iff\n");
L
Linus Torvalds 已提交
1702 1703 1704

	if (ifr->ifr_flags & IFF_NO_PI)
		tun->flags |= TUN_NO_PI;
1705 1706
	else
		tun->flags &= ~TUN_NO_PI;
L
Linus Torvalds 已提交
1707

1708 1709 1710
	/* This flag has no real effect.  We track the value for backwards
	 * compatibility.
	 */
L
Linus Torvalds 已提交
1711 1712
	if (ifr->ifr_flags & IFF_ONE_QUEUE)
		tun->flags |= TUN_ONE_QUEUE;
1713 1714
	else
		tun->flags &= ~TUN_ONE_QUEUE;
L
Linus Torvalds 已提交
1715

1716 1717 1718 1719 1720
	if (ifr->ifr_flags & IFF_VNET_HDR)
		tun->flags |= TUN_VNET_HDR;
	else
		tun->flags &= ~TUN_VNET_HDR;

J
Jason Wang 已提交
1721 1722 1723 1724 1725
	if (ifr->ifr_flags & IFF_MULTI_QUEUE)
		tun->flags |= TUN_TAP_MQ;
	else
		tun->flags &= ~TUN_TAP_MQ;

1726 1727 1728 1729
	/* Make sure persistent devices do not get stuck in
	 * xoff state.
	 */
	if (netif_running(tun->dev))
J
Jason Wang 已提交
1730
		netif_tx_wake_all_queues(tun->dev);
1731

L
Linus Torvalds 已提交
1732 1733 1734
	strcpy(ifr->ifr_name, tun->dev->name);
	return 0;

1735 1736 1737 1738 1739 1740
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
1741 1742 1743 1744
	free_netdev(dev);
	return err;
}

R
Rami Rosen 已提交
1745
static void tun_get_iff(struct net *net, struct tun_struct *tun,
1746
		       struct ifreq *ifr)
1747
{
1748
	tun_debug(KERN_INFO, tun, "tun_get_iff\n");
1749 1750 1751

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

1752
	ifr->ifr_flags = tun_flags(tun);
1753 1754 1755

}

1756 1757
/* This is like a cut-down ethtool ops, except done via tun fd so no
 * privs required. */
1758
static int set_offload(struct tun_struct *tun, unsigned long arg)
1759
{
1760
	netdev_features_t features = 0;
1761 1762

	if (arg & TUN_F_CSUM) {
1763
		features |= NETIF_F_HW_CSUM;
1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783
		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;

1784 1785
	tun->set_features = features;
	netdev_update_features(tun->dev);
1786 1787 1788 1789

	return 0;
}

J
Jason Wang 已提交
1790 1791 1792 1793 1794 1795
static void tun_detach_filter(struct tun_struct *tun, int n)
{
	int i;
	struct tun_file *tfile;

	for (i = 0; i < n; i++) {
1796
		tfile = rtnl_dereference(tun->tfiles[i]);
J
Jason Wang 已提交
1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808
		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++) {
1809
		tfile = rtnl_dereference(tun->tfiles[i]);
J
Jason Wang 已提交
1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826
		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++) {
1827
		tfile = rtnl_dereference(tun->tfiles[i]);
J
Jason Wang 已提交
1828 1829 1830 1831
		tfile->socket.sk->sk_sndbuf = tun->sndbuf;
	}
}

1832 1833 1834 1835 1836 1837 1838 1839 1840
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
1841
		tun = tfile->detached;
1842
		if (!tun) {
1843
			ret = -EINVAL;
1844 1845 1846 1847 1848
			goto unlock;
		}
		ret = security_tun_dev_attach_queue(tun->security);
		if (ret < 0)
			goto unlock;
1849
		ret = tun_attach(tun, file, false);
J
Jason Wang 已提交
1850
	} else if (ifr->ifr_flags & IFF_DETACH_QUEUE) {
1851
		tun = rtnl_dereference(tfile->tun);
1852
		if (!tun || !(tun->flags & TUN_TAP_MQ) || tfile->detached)
J
Jason Wang 已提交
1853 1854 1855 1856
			ret = -EINVAL;
		else
			__tun_detach(tfile, false);
	} else
1857 1858
		ret = -EINVAL;

1859
unlock:
1860 1861 1862 1863
	rtnl_unlock();
	return ret;
}

1864 1865
static long __tun_chr_ioctl(struct file *file, unsigned int cmd,
			    unsigned long arg, int ifreq_len)
L
Linus Torvalds 已提交
1866
{
E
Eric W. Biederman 已提交
1867
	struct tun_file *tfile = file->private_data;
E
Eric W. Biederman 已提交
1868
	struct tun_struct *tun;
L
Linus Torvalds 已提交
1869 1870
	void __user* argp = (void __user*)arg;
	struct ifreq ifr;
1871 1872
	kuid_t owner;
	kgid_t group;
1873
	int sndbuf;
1874
	int vnet_hdr_sz;
1875
	unsigned int ifindex;
1876
	int ret;
L
Linus Torvalds 已提交
1877

1878
	if (cmd == TUNSETIFF || cmd == TUNSETQUEUE || _IOC_TYPE(cmd) == 0x89) {
1879
		if (copy_from_user(&ifr, argp, ifreq_len))
L
Linus Torvalds 已提交
1880
			return -EFAULT;
D
David S. Miller 已提交
1881
	} else {
1882
		memset(&ifr, 0, sizeof(ifr));
D
David S. Miller 已提交
1883
	}
E
Eric W. Biederman 已提交
1884 1885 1886 1887 1888
	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 |
1889
				IFF_VNET_HDR | IFF_MULTI_QUEUE,
E
Eric W. Biederman 已提交
1890
				(unsigned int __user*)argp);
1891 1892
	} else if (cmd == TUNSETQUEUE)
		return tun_set_queue(file, &ifr);
E
Eric W. Biederman 已提交
1893

J
Jason Wang 已提交
1894
	ret = 0;
1895 1896
	rtnl_lock();

E
Eric W. Biederman 已提交
1897
	tun = __tun_get(tfile);
L
Linus Torvalds 已提交
1898 1899 1900
	if (cmd == TUNSETIFF && !tun) {
		ifr.ifr_name[IFNAMSIZ-1] = '\0';

1901
		ret = tun_set_iff(tfile->net, file, &ifr);
L
Linus Torvalds 已提交
1902

1903 1904
		if (ret)
			goto unlock;
L
Linus Torvalds 已提交
1905

1906
		if (copy_to_user(argp, &ifr, ifreq_len))
1907 1908
			ret = -EFAULT;
		goto unlock;
L
Linus Torvalds 已提交
1909
	}
1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922
	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
1923

1924
	ret = -EBADFD;
L
Linus Torvalds 已提交
1925
	if (!tun)
1926
		goto unlock;
L
Linus Torvalds 已提交
1927

1928
	tun_debug(KERN_INFO, tun, "tun_chr_ioctl cmd %u\n", cmd);
L
Linus Torvalds 已提交
1929

E
Eric W. Biederman 已提交
1930
	ret = 0;
L
Linus Torvalds 已提交
1931
	switch (cmd) {
1932
	case TUNGETIFF:
R
Rami Rosen 已提交
1933
		tun_get_iff(current->nsproxy->net_ns, tun, &ifr);
1934

1935 1936
		if (tfile->detached)
			ifr.ifr_flags |= IFF_DETACH_QUEUE;
1937 1938
		if (!tfile->socket.sk->sk_filter)
			ifr.ifr_flags |= IFF_NOFILTER;
1939

1940
		if (copy_to_user(argp, &ifr, ifreq_len))
E
Eric W. Biederman 已提交
1941
			ret = -EFAULT;
1942 1943
		break;

L
Linus Torvalds 已提交
1944 1945 1946
	case TUNSETNOCSUM:
		/* Disable/Enable checksum */

1947 1948
		/* [unimplemented] */
		tun_debug(KERN_INFO, tun, "ignored: set checksum %s\n",
1949
			  arg ? "disabled" : "enabled");
L
Linus Torvalds 已提交
1950 1951 1952
		break;

	case TUNSETPERSIST:
J
Jason Wang 已提交
1953 1954 1955
		/* Disable/Enable persist mode. Keep an extra reference to the
		 * module to prevent the module being unprobed.
		 */
J
Jason Wang 已提交
1956
		if (arg && !(tun->flags & TUN_PERSIST)) {
L
Linus Torvalds 已提交
1957
			tun->flags |= TUN_PERSIST;
J
Jason Wang 已提交
1958
			__module_get(THIS_MODULE);
J
Jason Wang 已提交
1959 1960
		}
		if (!arg && (tun->flags & TUN_PERSIST)) {
L
Linus Torvalds 已提交
1961
			tun->flags &= ~TUN_PERSIST;
J
Jason Wang 已提交
1962 1963
			module_put(THIS_MODULE);
		}
L
Linus Torvalds 已提交
1964

1965 1966
		tun_debug(KERN_INFO, tun, "persist %s\n",
			  arg ? "enabled" : "disabled");
L
Linus Torvalds 已提交
1967 1968 1969 1970
		break;

	case TUNSETOWNER:
		/* Set owner of the device */
1971 1972 1973 1974 1975 1976
		owner = make_kuid(current_user_ns(), arg);
		if (!uid_valid(owner)) {
			ret = -EINVAL;
			break;
		}
		tun->owner = owner;
1977
		tun_debug(KERN_INFO, tun, "owner set to %u\n",
1978
			  from_kuid(&init_user_ns, tun->owner));
L
Linus Torvalds 已提交
1979 1980
		break;

1981 1982
	case TUNSETGROUP:
		/* Set group of the device */
1983 1984 1985 1986 1987 1988
		group = make_kgid(current_user_ns(), arg);
		if (!gid_valid(group)) {
			ret = -EINVAL;
			break;
		}
		tun->group = group;
1989
		tun_debug(KERN_INFO, tun, "group set to %u\n",
1990
			  from_kgid(&init_user_ns, tun->group));
1991 1992
		break;

1993 1994 1995
	case TUNSETLINK:
		/* Only allow setting the type when the interface is down */
		if (tun->dev->flags & IFF_UP) {
1996 1997
			tun_debug(KERN_INFO, tun,
				  "Linktype set failed because interface is up\n");
1998
			ret = -EBUSY;
1999 2000
		} else {
			tun->dev->type = (int) arg;
2001 2002
			tun_debug(KERN_INFO, tun, "linktype set to %d\n",
				  tun->dev->type);
2003
			ret = 0;
2004
		}
E
Eric W. Biederman 已提交
2005
		break;
2006

L
Linus Torvalds 已提交
2007 2008 2009 2010 2011
#ifdef TUN_DEBUG
	case TUNSETDEBUG:
		tun->debug = arg;
		break;
#endif
2012
	case TUNSETOFFLOAD:
2013
		ret = set_offload(tun, arg);
E
Eric W. Biederman 已提交
2014
		break;
2015

2016 2017
	case TUNSETTXFILTER:
		/* Can be set only for TAPs */
E
Eric W. Biederman 已提交
2018
		ret = -EINVAL;
2019
		if ((tun->flags & TUN_TYPE_MASK) != TUN_TAP_DEV)
E
Eric W. Biederman 已提交
2020
			break;
H
Harvey Harrison 已提交
2021
		ret = update_filter(&tun->txflt, (void __user *)arg);
E
Eric W. Biederman 已提交
2022
		break;
L
Linus Torvalds 已提交
2023 2024

	case SIOCGIFHWADDR:
2025
		/* Get hw address */
2026 2027
		memcpy(ifr.ifr_hwaddr.sa_data, tun->dev->dev_addr, ETH_ALEN);
		ifr.ifr_hwaddr.sa_family = tun->dev->type;
2028
		if (copy_to_user(argp, &ifr, ifreq_len))
E
Eric W. Biederman 已提交
2029 2030
			ret = -EFAULT;
		break;
L
Linus Torvalds 已提交
2031 2032

	case SIOCSIFHWADDR:
2033
		/* Set hw address */
2034 2035
		tun_debug(KERN_DEBUG, tun, "set hw address: %pM\n",
			  ifr.ifr_hwaddr.sa_data);
2036 2037

		ret = dev_set_mac_address(tun->dev, &ifr.ifr_hwaddr);
E
Eric W. Biederman 已提交
2038
		break;
2039 2040

	case TUNGETSNDBUF:
J
Jason Wang 已提交
2041
		sndbuf = tfile->socket.sk->sk_sndbuf;
2042 2043 2044 2045 2046 2047 2048 2049 2050 2051
		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
2052 2053
		tun->sndbuf = sndbuf;
		tun_set_sndbuf(tun);
2054 2055
		break;

2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074
	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
2075 2076 2077 2078 2079 2080
	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
2081
		if (copy_from_user(&tun->fprog, argp, sizeof(tun->fprog)))
M
Michael S. Tsirkin 已提交
2082 2083
			break;

J
Jason Wang 已提交
2084
		ret = tun_attach_filter(tun);
M
Michael S. Tsirkin 已提交
2085 2086 2087 2088 2089 2090 2091
		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
2092 2093
		ret = 0;
		tun_detach_filter(tun, tun->numqueues);
M
Michael S. Tsirkin 已提交
2094 2095
		break;

P
Pavel Emelyanov 已提交
2096 2097 2098 2099 2100 2101 2102 2103 2104 2105
	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
2106
	default:
E
Eric W. Biederman 已提交
2107 2108
		ret = -EINVAL;
		break;
2109
	}
L
Linus Torvalds 已提交
2110

2111 2112 2113 2114
unlock:
	rtnl_unlock();
	if (tun)
		tun_put(tun);
E
Eric W. Biederman 已提交
2115
	return ret;
L
Linus Torvalds 已提交
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 2150 2151 2152
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
2153 2154
static int tun_chr_fasync(int fd, struct file *file, int on)
{
J
Jason Wang 已提交
2155
	struct tun_file *tfile = file->private_data;
L
Linus Torvalds 已提交
2156 2157
	int ret;

J
Jason Wang 已提交
2158
	if ((ret = fasync_helper(fd, file, on, &tfile->fasync)) < 0)
J
Jonathan Corbet 已提交
2159
		goto out;
2160

L
Linus Torvalds 已提交
2161
	if (on) {
2162
		__f_setown(file, task_pid(current), PIDTYPE_PID, 0);
J
Jason Wang 已提交
2163
		tfile->flags |= TUN_FASYNC;
2164
	} else
J
Jason Wang 已提交
2165
		tfile->flags &= ~TUN_FASYNC;
J
Jonathan Corbet 已提交
2166 2167 2168
	ret = 0;
out:
	return ret;
L
Linus Torvalds 已提交
2169 2170 2171 2172
}

static int tun_chr_open(struct inode *inode, struct file * file)
{
E
Eric W. Biederman 已提交
2173
	struct tun_file *tfile;
T
Thomas Gleixner 已提交
2174

2175
	DBG1(KERN_INFO, "tunX: tun_chr_open\n");
E
Eric W. Biederman 已提交
2176

J
Jason Wang 已提交
2177 2178
	tfile = (struct tun_file *)sk_alloc(&init_net, AF_UNSPEC, GFP_KERNEL,
					    &tun_proto);
E
Eric W. Biederman 已提交
2179 2180
	if (!tfile)
		return -ENOMEM;
2181
	RCU_INIT_POINTER(tfile->tun, NULL);
E
Eric W. Biederman 已提交
2182
	tfile->net = get_net(current->nsproxy->net_ns);
J
Jason Wang 已提交
2183
	tfile->flags = 0;
2184
	tfile->ifindex = 0;
J
Jason Wang 已提交
2185 2186

	init_waitqueue_head(&tfile->wq.wait);
2187
	RCU_INIT_POINTER(tfile->socket.wq, &tfile->wq);
J
Jason Wang 已提交
2188 2189 2190 2191 2192 2193 2194 2195 2196 2197

	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
2198
	file->private_data = tfile;
J
Jason Wang 已提交
2199
	set_bit(SOCK_EXTERNALLY_ALLOCATED, &tfile->socket.flags);
J
Jason Wang 已提交
2200
	INIT_LIST_HEAD(&tfile->next);
J
Jason Wang 已提交
2201

2202 2203
	sock_set_flag(&tfile->sk, SOCK_ZEROCOPY);

L
Linus Torvalds 已提交
2204 2205 2206 2207 2208
	return 0;
}

static int tun_chr_close(struct inode *inode, struct file *file)
{
E
Eric W. Biederman 已提交
2209
	struct tun_file *tfile = file->private_data;
J
Jason Wang 已提交
2210
	struct net *net = tfile->net;
L
Linus Torvalds 已提交
2211

J
Jason Wang 已提交
2212
	tun_detach(tfile, true);
J
Jason Wang 已提交
2213
	put_net(net);
L
Linus Torvalds 已提交
2214 2215 2216 2217

	return 0;
}

2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238
#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

2239
static const struct file_operations tun_fops = {
2240
	.owner	= THIS_MODULE,
L
Linus Torvalds 已提交
2241
	.llseek = no_llseek,
2242 2243 2244 2245
	.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
2246
	.poll	= tun_chr_poll,
2247 2248 2249 2250
	.unlocked_ioctl	= tun_chr_ioctl,
#ifdef CONFIG_COMPAT
	.compat_ioctl = tun_chr_compat_ioctl,
#endif
L
Linus Torvalds 已提交
2251 2252
	.open	= tun_chr_open,
	.release = tun_chr_close,
2253 2254 2255 2256
	.fasync = tun_chr_fasync,
#ifdef CONFIG_PROC_FS
	.show_fdinfo = tun_chr_show_fdinfo,
#endif
L
Linus Torvalds 已提交
2257 2258 2259 2260 2261
};

static struct miscdevice tun_miscdev = {
	.minor = TUN_MINOR,
	.name = "tun",
2262
	.nodename = "net/tun",
L
Linus Torvalds 已提交
2263 2264 2265 2266 2267 2268 2269 2270 2271
	.fops = &tun_fops,
};

/* ethtool interface */

static int tun_get_settings(struct net_device *dev, struct ethtool_cmd *cmd)
{
	cmd->supported		= 0;
	cmd->advertising	= 0;
2272
	ethtool_cmd_speed_set(cmd, SPEED_10);
L
Linus Torvalds 已提交
2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286
	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);

2287 2288
	strlcpy(info->driver, DRV_NAME, sizeof(info->driver));
	strlcpy(info->version, DRV_VERSION, sizeof(info->version));
L
Linus Torvalds 已提交
2289 2290 2291

	switch (tun->flags & TUN_TYPE_MASK) {
	case TUN_TUN_DEV:
2292
		strlcpy(info->bus_info, "tun", sizeof(info->bus_info));
L
Linus Torvalds 已提交
2293 2294
		break;
	case TUN_TAP_DEV:
2295
		strlcpy(info->bus_info, "tap", sizeof(info->bus_info));
L
Linus Torvalds 已提交
2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317
		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
}

2318
static const struct ethtool_ops tun_ethtool_ops = {
L
Linus Torvalds 已提交
2319 2320 2321 2322
	.get_settings	= tun_get_settings,
	.get_drvinfo	= tun_get_drvinfo,
	.get_msglevel	= tun_get_msglevel,
	.set_msglevel	= tun_set_msglevel,
2323
	.get_link	= ethtool_op_get_link,
2324
	.get_ts_info	= ethtool_op_get_ts_info,
L
Linus Torvalds 已提交
2325 2326
};

2327

L
Linus Torvalds 已提交
2328 2329 2330 2331
static int __init tun_init(void)
{
	int ret = 0;

2332 2333
	pr_info("%s, %s\n", DRV_DESCRIPTION, DRV_VERSION);
	pr_info("%s\n", DRV_COPYRIGHT);
L
Linus Torvalds 已提交
2334

2335
	ret = rtnl_link_register(&tun_link_ops);
2336
	if (ret) {
2337
		pr_err("Can't register link_ops\n");
2338
		goto err_linkops;
2339 2340
	}

L
Linus Torvalds 已提交
2341
	ret = misc_register(&tun_miscdev);
2342
	if (ret) {
2343
		pr_err("Can't register misc device %d\n", TUN_MINOR);
2344 2345
		goto err_misc;
	}
2346
	return  0;
2347
err_misc:
2348 2349
	rtnl_link_unregister(&tun_link_ops);
err_linkops:
L
Linus Torvalds 已提交
2350 2351 2352 2353 2354
	return ret;
}

static void tun_cleanup(void)
{
2355
	misc_deregister(&tun_miscdev);
2356
	rtnl_link_unregister(&tun_link_ops);
L
Linus Torvalds 已提交
2357 2358
}

2359 2360 2361 2362 2363 2364
/* 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)
{
2365
	struct tun_file *tfile;
2366 2367
	if (file->f_op != &tun_fops)
		return ERR_PTR(-EINVAL);
2368 2369
	tfile = file->private_data;
	if (!tfile)
2370
		return ERR_PTR(-EBADFD);
J
Jason Wang 已提交
2371
	return &tfile->socket;
2372 2373 2374
}
EXPORT_SYMBOL_GPL(tun_get_socket);

L
Linus Torvalds 已提交
2375 2376 2377 2378 2379 2380
module_init(tun_init);
module_exit(tun_cleanup);
MODULE_DESCRIPTION(DRV_DESCRIPTION);
MODULE_AUTHOR(DRV_COPYRIGHT);
MODULE_LICENSE("GPL");
MODULE_ALIAS_MISCDEV(TUN_MINOR);
2381
MODULE_ALIAS("devname:net/tun");