af_packet.c 61.3 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7
/*
 * INET		An implementation of the TCP/IP protocol suite for the LINUX
 *		operating system.  INET is implemented using the  BSD Socket
 *		interface as the means of communication with the user level.
 *
 *		PACKET - implements raw packet sockets.
 *
8
 * Authors:	Ross Biro
L
Linus Torvalds 已提交
9 10 11
 *		Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
 *		Alan Cox, <gw4pts@gw4pts.ampr.org>
 *
12
 * Fixes:
L
Linus Torvalds 已提交
13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
 *		Alan Cox	:	verify_area() now used correctly
 *		Alan Cox	:	new skbuff lists, look ma no backlogs!
 *		Alan Cox	:	tidied skbuff lists.
 *		Alan Cox	:	Now uses generic datagram routines I
 *					added. Also fixed the peek/read crash
 *					from all old Linux datagram code.
 *		Alan Cox	:	Uses the improved datagram code.
 *		Alan Cox	:	Added NULL's for socket options.
 *		Alan Cox	:	Re-commented the code.
 *		Alan Cox	:	Use new kernel side addressing
 *		Rob Janssen	:	Correct MTU usage.
 *		Dave Platt	:	Counter leaks caused by incorrect
 *					interrupt locking and some slightly
 *					dubious gcc output. Can you read
 *					compiler: it said _VOLATILE_
 *	Richard Kooijman	:	Timestamp fixes.
 *		Alan Cox	:	New buffers. Use sk->mac.raw.
 *		Alan Cox	:	sendmsg/recvmsg support.
 *		Alan Cox	:	Protocol setting support
 *	Alexey Kuznetsov	:	Untied from IPv4 stack.
 *	Cyrus Durgin		:	Fixed kerneld for kmod.
 *	Michal Ostrowski        :       Module initialization cleanup.
35
 *         Ulises Alonso        :       Frame number limit removal and
L
Linus Torvalds 已提交
36
 *                                      packet_set_ring memory leak.
37 38 39
 *		Eric Biederman	:	Allow for > 8 byte hardware addresses.
 *					The convention is that longer addresses
 *					will simply extend the hardware address
40
 *					byte arrays at the end of sockaddr_ll
41
 *					and packet_mreq.
J
Johann Baudy 已提交
42
 *		Johann Baudy	:	Added TX RING.
L
Linus Torvalds 已提交
43 44 45 46 47 48 49
 *
 *		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.
 *
 */
50

L
Linus Torvalds 已提交
51 52
#include <linux/types.h>
#include <linux/mm.h>
53
#include <linux/capability.h>
L
Linus Torvalds 已提交
54 55 56 57 58 59 60
#include <linux/fcntl.h>
#include <linux/socket.h>
#include <linux/in.h>
#include <linux/inet.h>
#include <linux/netdevice.h>
#include <linux/if_packet.h>
#include <linux/wireless.h>
61
#include <linux/kernel.h>
L
Linus Torvalds 已提交
62
#include <linux/kmod.h>
63
#include <linux/slab.h>
64
#include <linux/vmalloc.h>
65
#include <net/net_namespace.h>
L
Linus Torvalds 已提交
66 67 68 69 70 71 72 73 74 75
#include <net/ip.h>
#include <net/protocol.h>
#include <linux/skbuff.h>
#include <net/sock.h>
#include <linux/errno.h>
#include <linux/timer.h>
#include <asm/system.h>
#include <asm/uaccess.h>
#include <asm/ioctls.h>
#include <asm/page.h>
A
Al Viro 已提交
76
#include <asm/cacheflush.h>
L
Linus Torvalds 已提交
77 78 79 80 81 82
#include <asm/io.h>
#include <linux/proc_fs.h>
#include <linux/seq_file.h>
#include <linux/poll.h>
#include <linux/module.h>
#include <linux/init.h>
83
#include <linux/mutex.h>
E
Eric Dumazet 已提交
84
#include <linux/if_vlan.h>
85
#include <linux/virtio_net.h>
86
#include <linux/errqueue.h>
87
#include <linux/net_tstamp.h>
L
Linus Torvalds 已提交
88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107

#ifdef CONFIG_INET
#include <net/inet_common.h>
#endif

/*
   Assumptions:
   - if device has no dev->hard_header routine, it adds and removes ll header
     inside itself. In this case ll header is invisible outside of device,
     but higher levels still should reserve dev->hard_header_len.
     Some devices are enough clever to reallocate skb, when header
     will not fit to reserved space (tunnel), another ones are silly
     (PPP).
   - packet socket receives packets with pulled ll header,
     so that SOCK_RAW should push it back.

On receive:
-----------

Incoming, dev->hard_header!=NULL
108 109
   mac_header -> ll header
   data       -> data
L
Linus Torvalds 已提交
110 111

Outgoing, dev->hard_header!=NULL
112 113
   mac_header -> ll header
   data       -> ll header
L
Linus Torvalds 已提交
114 115

Incoming, dev->hard_header==NULL
116 117
   mac_header -> UNKNOWN position. It is very likely, that it points to ll
		 header.  PPP makes it, that is wrong, because introduce
118
		 assymetry between rx and tx paths.
119
   data       -> data
L
Linus Torvalds 已提交
120 121

Outgoing, dev->hard_header==NULL
122 123
   mac_header -> data. ll header is still not built!
   data       -> data
L
Linus Torvalds 已提交
124 125 126 127 128 129 130 131 132

Resume
  If dev->hard_header==NULL we are unlikely to restore sensible ll header.


On transmit:
------------

dev->hard_header != NULL
133 134
   mac_header -> ll header
   data       -> ll header
L
Linus Torvalds 已提交
135 136

dev->hard_header == NULL (ll header is added by device, we cannot control it)
137 138
   mac_header -> data
   data       -> data
L
Linus Torvalds 已提交
139 140 141 142 143 144 145

   We should set nh.raw on output to correct posistion,
   packet classifier depends on it.
 */

/* Private packet socket structures. */

E
Eric Dumazet 已提交
146
struct packet_mclist {
L
Linus Torvalds 已提交
147 148 149 150 151
	struct packet_mclist	*next;
	int			ifindex;
	int			count;
	unsigned short		type;
	unsigned short		alen;
152 153 154 155 156
	unsigned char		addr[MAX_ADDR_LEN];
};
/* identical to struct packet_mreq except it has
 * a longer address field.
 */
E
Eric Dumazet 已提交
157
struct packet_mreq_max {
158 159 160 161
	int		mr_ifindex;
	unsigned short	mr_type;
	unsigned short	mr_alen;
	unsigned char	mr_address[MAX_ADDR_LEN];
L
Linus Torvalds 已提交
162
};
163

J
Johann Baudy 已提交
164 165 166
static int packet_set_ring(struct sock *sk, struct tpacket_req *req,
		int closing, int tx_ring);

167 168 169 170 171
#define PGV_FROM_VMALLOC 1
struct pgv {
	char *buffer;
};

J
Johann Baudy 已提交
172
struct packet_ring_buffer {
173
	struct pgv		*pg_vec;
J
Johann Baudy 已提交
174 175 176 177 178 179 180 181 182 183 184 185 186 187
	unsigned int		head;
	unsigned int		frames_per_block;
	unsigned int		frame_size;
	unsigned int		frame_max;

	unsigned int		pg_vec_order;
	unsigned int		pg_vec_pages;
	unsigned int		pg_vec_len;

	atomic_t		pending;
};

struct packet_sock;
static int tpacket_snd(struct packet_sock *po, struct msghdr *msg);
L
Linus Torvalds 已提交
188 189 190 191 192 193 194

static void packet_flush_mclist(struct sock *sk);

struct packet_sock {
	/* struct sock has to be the first member of packet_sock */
	struct sock		sk;
	struct tpacket_stats	stats;
J
Johann Baudy 已提交
195 196
	struct packet_ring_buffer	rx_ring;
	struct packet_ring_buffer	tx_ring;
L
Linus Torvalds 已提交
197 198
	int			copy_thresh;
	spinlock_t		bind_lock;
199
	struct mutex		pg_vec_lock;
200
	unsigned int		running:1,	/* prot_hook is attached*/
201
				auxdata:1,
202 203
				origdev:1,
				has_vnet_hdr:1;
L
Linus Torvalds 已提交
204
	int			ifindex;	/* bound device		*/
A
Al Viro 已提交
205
	__be16			num;
L
Linus Torvalds 已提交
206 207
	struct packet_mclist	*mclist;
	atomic_t		mapped;
208 209
	enum tpacket_versions	tp_version;
	unsigned int		tp_hdrlen;
210
	unsigned int		tp_reserve;
J
Johann Baudy 已提交
211
	unsigned int		tp_loss:1;
212
	unsigned int		tp_tstamp;
213
	struct packet_type	prot_hook ____cacheline_aligned_in_smp;
L
Linus Torvalds 已提交
214 215
};

216 217 218 219 220 221 222 223 224
struct packet_skb_cb {
	unsigned int origlen;
	union {
		struct sockaddr_pkt pkt;
		struct sockaddr_ll ll;
	} sa;
};

#define PACKET_SKB_CB(__skb)	((struct packet_skb_cb *)((__skb)->cb))
225

226 227 228 229 230 231 232
static inline struct page *pgv_to_page(void *addr)
{
	if (is_vmalloc_addr(addr))
		return vmalloc_to_page(addr);
	return virt_to_page(addr);
}

J
Johann Baudy 已提交
233
static void __packet_set_status(struct packet_sock *po, void *frame, int status)
L
Linus Torvalds 已提交
234
{
235 236 237 238 239
	union {
		struct tpacket_hdr *h1;
		struct tpacket2_hdr *h2;
		void *raw;
	} h;
L
Linus Torvalds 已提交
240

J
Johann Baudy 已提交
241
	h.raw = frame;
242 243
	switch (po->tp_version) {
	case TPACKET_V1:
J
Johann Baudy 已提交
244
		h.h1->tp_status = status;
245
		flush_dcache_page(pgv_to_page(&h.h1->tp_status));
246 247
		break;
	case TPACKET_V2:
J
Johann Baudy 已提交
248
		h.h2->tp_status = status;
249
		flush_dcache_page(pgv_to_page(&h.h2->tp_status));
250
		break;
J
Johann Baudy 已提交
251
	default:
E
Eric Dumazet 已提交
252
		pr_err("TPACKET version not supported\n");
J
Johann Baudy 已提交
253
		BUG();
254
	}
J
Johann Baudy 已提交
255 256

	smp_wmb();
257 258
}

J
Johann Baudy 已提交
259
static int __packet_get_status(struct packet_sock *po, void *frame)
260 261 262 263 264 265 266
{
	union {
		struct tpacket_hdr *h1;
		struct tpacket2_hdr *h2;
		void *raw;
	} h;

J
Johann Baudy 已提交
267 268
	smp_rmb();

269 270 271
	h.raw = frame;
	switch (po->tp_version) {
	case TPACKET_V1:
272
		flush_dcache_page(pgv_to_page(&h.h1->tp_status));
J
Johann Baudy 已提交
273
		return h.h1->tp_status;
274
	case TPACKET_V2:
275
		flush_dcache_page(pgv_to_page(&h.h2->tp_status));
J
Johann Baudy 已提交
276 277
		return h.h2->tp_status;
	default:
E
Eric Dumazet 已提交
278
		pr_err("TPACKET version not supported\n");
J
Johann Baudy 已提交
279 280
		BUG();
		return 0;
281
	}
L
Linus Torvalds 已提交
282
}
J
Johann Baudy 已提交
283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298

static void *packet_lookup_frame(struct packet_sock *po,
		struct packet_ring_buffer *rb,
		unsigned int position,
		int status)
{
	unsigned int pg_vec_pos, frame_offset;
	union {
		struct tpacket_hdr *h1;
		struct tpacket2_hdr *h2;
		void *raw;
	} h;

	pg_vec_pos = position / rb->frames_per_block;
	frame_offset = position % rb->frames_per_block;

299 300
	h.raw = rb->pg_vec[pg_vec_pos].buffer +
		(frame_offset * rb->frame_size);
J
Johann Baudy 已提交
301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327

	if (status != __packet_get_status(po, h.raw))
		return NULL;

	return h.raw;
}

static inline void *packet_current_frame(struct packet_sock *po,
		struct packet_ring_buffer *rb,
		int status)
{
	return packet_lookup_frame(po, rb, rb->head, status);
}

static inline void *packet_previous_frame(struct packet_sock *po,
		struct packet_ring_buffer *rb,
		int status)
{
	unsigned int previous = rb->head ? rb->head - 1 : rb->frame_max;
	return packet_lookup_frame(po, rb, previous, status);
}

static inline void packet_increment_head(struct packet_ring_buffer *buff)
{
	buff->head = buff->head != buff->frame_max ? buff->head+1 : 0;
}

L
Linus Torvalds 已提交
328 329 330 331 332 333 334
static inline struct packet_sock *pkt_sk(struct sock *sk)
{
	return (struct packet_sock *)sk;
}

static void packet_sock_destruct(struct sock *sk)
{
335 336
	skb_queue_purge(&sk->sk_error_queue);

337 338
	WARN_ON(atomic_read(&sk->sk_rmem_alloc));
	WARN_ON(atomic_read(&sk->sk_wmem_alloc));
L
Linus Torvalds 已提交
339 340

	if (!sock_flag(sk, SOCK_DEAD)) {
E
Eric Dumazet 已提交
341
		pr_err("Attempt to release alive packet socket: %p\n", sk);
L
Linus Torvalds 已提交
342 343 344
		return;
	}

345
	sk_refcnt_debug_dec(sk);
L
Linus Torvalds 已提交
346 347 348
}


349
static const struct proto_ops packet_ops;
L
Linus Torvalds 已提交
350

351
static const struct proto_ops packet_ops_spkt;
L
Linus Torvalds 已提交
352

E
Eric Dumazet 已提交
353 354
static int packet_rcv_spkt(struct sk_buff *skb, struct net_device *dev,
			   struct packet_type *pt, struct net_device *orig_dev)
L
Linus Torvalds 已提交
355 356 357 358 359 360 361 362 363 364
{
	struct sock *sk;
	struct sockaddr_pkt *spkt;

	/*
	 *	When we registered the protocol we saved the socket in the data
	 *	field for just this event.
	 */

	sk = pt->af_packet_priv;
365

L
Linus Torvalds 已提交
366 367 368 369 370 371 372
	/*
	 *	Yank back the headers [hope the device set this
	 *	right or kerboom...]
	 *
	 *	Incoming packets have ll header pulled,
	 *	push it back.
	 *
373
	 *	For outgoing ones skb->data == skb_mac_header(skb)
L
Linus Torvalds 已提交
374 375 376 377 378 379
	 *	so that this procedure is noop.
	 */

	if (skb->pkt_type == PACKET_LOOPBACK)
		goto out;

O
Octavian Purdila 已提交
380
	if (!net_eq(dev_net(dev), sock_net(sk)))
381 382
		goto out;

E
Eric Dumazet 已提交
383 384
	skb = skb_share_check(skb, GFP_ATOMIC);
	if (skb == NULL)
L
Linus Torvalds 已提交
385 386 387
		goto oom;

	/* drop any routing info */
E
Eric Dumazet 已提交
388
	skb_dst_drop(skb);
L
Linus Torvalds 已提交
389

390 391 392
	/* drop conntrack reference */
	nf_reset(skb);

393
	spkt = &PACKET_SKB_CB(skb)->sa.pkt;
L
Linus Torvalds 已提交
394

395
	skb_push(skb, skb->data - skb_mac_header(skb));
L
Linus Torvalds 已提交
396 397 398 399 400 401 402 403 404 405 406 407 408 409

	/*
	 *	The SOCK_PACKET socket receives _all_ frames.
	 */

	spkt->spkt_family = dev->type;
	strlcpy(spkt->spkt_device, dev->name, sizeof(spkt->spkt_device));
	spkt->spkt_protocol = skb->protocol;

	/*
	 *	Charge the memory to the socket. This is done specifically
	 *	to prevent sockets using all the memory up.
	 */

E
Eric Dumazet 已提交
410
	if (sock_queue_rcv_skb(sk, skb) == 0)
L
Linus Torvalds 已提交
411 412 413 414 415 416 417 418 419 420 421 422 423
		return 0;

out:
	kfree_skb(skb);
oom:
	return 0;
}


/*
 *	Output a raw packet to a device layer. This bypasses all the other
 *	protocol layers and you must therefore supply it with a complete frame
 */
424

L
Linus Torvalds 已提交
425 426 427 428
static int packet_sendmsg_spkt(struct kiocb *iocb, struct socket *sock,
			       struct msghdr *msg, size_t len)
{
	struct sock *sk = sock->sk;
E
Eric Dumazet 已提交
429
	struct sockaddr_pkt *saddr = (struct sockaddr_pkt *)msg->msg_name;
430
	struct sk_buff *skb = NULL;
L
Linus Torvalds 已提交
431
	struct net_device *dev;
E
Eric Dumazet 已提交
432
	__be16 proto = 0;
L
Linus Torvalds 已提交
433
	int err;
434

L
Linus Torvalds 已提交
435
	/*
436
	 *	Get and verify the address.
L
Linus Torvalds 已提交
437 438
	 */

E
Eric Dumazet 已提交
439
	if (saddr) {
L
Linus Torvalds 已提交
440
		if (msg->msg_namelen < sizeof(struct sockaddr))
E
Eric Dumazet 已提交
441 442 443 444 445
			return -EINVAL;
		if (msg->msg_namelen == sizeof(struct sockaddr_pkt))
			proto = saddr->spkt_protocol;
	} else
		return -ENOTCONN;	/* SOCK_PACKET must be sent giving an address */
L
Linus Torvalds 已提交
446 447

	/*
448
	 *	Find the device first to size check it
L
Linus Torvalds 已提交
449 450 451
	 */

	saddr->spkt_device[13] = 0;
452
retry:
E
Eric Dumazet 已提交
453 454
	rcu_read_lock();
	dev = dev_get_by_name_rcu(sock_net(sk), saddr->spkt_device);
L
Linus Torvalds 已提交
455 456 457
	err = -ENODEV;
	if (dev == NULL)
		goto out_unlock;
458

459 460 461 462
	err = -ENETDOWN;
	if (!(dev->flags & IFF_UP))
		goto out_unlock;

L
Linus Torvalds 已提交
463
	/*
E
Eric Dumazet 已提交
464 465
	 * You may not queue a frame bigger than the mtu. This is the lowest level
	 * raw protocol and you must do your own fragmentation at this level.
L
Linus Torvalds 已提交
466
	 */
467

L
Linus Torvalds 已提交
468
	err = -EMSGSIZE;
469
	if (len > dev->mtu + dev->hard_header_len)
L
Linus Torvalds 已提交
470 471
		goto out_unlock;

472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497
	if (!skb) {
		size_t reserved = LL_RESERVED_SPACE(dev);
		unsigned int hhlen = dev->header_ops ? dev->hard_header_len : 0;

		rcu_read_unlock();
		skb = sock_wmalloc(sk, len + reserved, 0, GFP_KERNEL);
		if (skb == NULL)
			return -ENOBUFS;
		/* FIXME: Save some space for broken drivers that write a hard
		 * header at transmission time by themselves. PPP is the notable
		 * one here. This should really be fixed at the driver level.
		 */
		skb_reserve(skb, reserved);
		skb_reset_network_header(skb);

		/* Try to align data part correctly */
		if (hhlen) {
			skb->data -= hhlen;
			skb->tail -= hhlen;
			if (len < hhlen)
				skb_reset_network_header(skb);
		}
		err = memcpy_fromiovec(skb_put(skb, len), msg->msg_iov, len);
		if (err)
			goto out_free;
		goto retry;
L
Linus Torvalds 已提交
498 499
	}

500

L
Linus Torvalds 已提交
501 502 503
	skb->protocol = proto;
	skb->dev = dev;
	skb->priority = sk->sk_priority;
504
	skb->mark = sk->sk_mark;
505
	err = sock_tx_timestamp(sk, &skb_shinfo(skb)->tx_flags);
506 507
	if (err < 0)
		goto out_unlock;
L
Linus Torvalds 已提交
508 509

	dev_queue_xmit(skb);
E
Eric Dumazet 已提交
510
	rcu_read_unlock();
E
Eric Dumazet 已提交
511
	return len;
L
Linus Torvalds 已提交
512 513

out_unlock:
E
Eric Dumazet 已提交
514
	rcu_read_unlock();
515 516
out_free:
	kfree_skb(skb);
L
Linus Torvalds 已提交
517 518 519
	return err;
}

E
Eric Dumazet 已提交
520 521
static inline unsigned int run_filter(const struct sk_buff *skb,
				      const struct sock *sk,
D
David S. Miller 已提交
522
				      unsigned int res)
L
Linus Torvalds 已提交
523 524
{
	struct sk_filter *filter;
525 526

	rcu_read_lock_bh();
527
	filter = rcu_dereference_bh(sk->sk_filter);
D
David S. Miller 已提交
528
	if (filter != NULL)
E
Eric Dumazet 已提交
529
		res = sk_run_filter(skb, filter->insns);
530
	rcu_read_unlock_bh();
L
Linus Torvalds 已提交
531

D
David S. Miller 已提交
532
	return res;
L
Linus Torvalds 已提交
533 534 535
}

/*
E
Eric Dumazet 已提交
536 537 538 539 540 541 542 543 544
 * This function makes lazy skb cloning in hope that most of packets
 * are discarded by BPF.
 *
 * Note tricky part: we DO mangle shared skb! skb->data, skb->len
 * and skb->cb are mangled. It works because (and until) packets
 * falling here are owned by current CPU. Output packets are cloned
 * by dev_queue_xmit_nit(), input packets are processed by net_bh
 * sequencially, so that if we return skb to original state on exit,
 * we will not harm anyone.
L
Linus Torvalds 已提交
545 546
 */

E
Eric Dumazet 已提交
547 548
static int packet_rcv(struct sk_buff *skb, struct net_device *dev,
		      struct packet_type *pt, struct net_device *orig_dev)
L
Linus Torvalds 已提交
549 550 551 552
{
	struct sock *sk;
	struct sockaddr_ll *sll;
	struct packet_sock *po;
E
Eric Dumazet 已提交
553
	u8 *skb_head = skb->data;
L
Linus Torvalds 已提交
554
	int skb_len = skb->len;
D
David S. Miller 已提交
555
	unsigned int snaplen, res;
L
Linus Torvalds 已提交
556 557 558 559 560 561 562

	if (skb->pkt_type == PACKET_LOOPBACK)
		goto drop;

	sk = pt->af_packet_priv;
	po = pkt_sk(sk);

O
Octavian Purdila 已提交
563
	if (!net_eq(dev_net(dev), sock_net(sk)))
564 565
		goto drop;

L
Linus Torvalds 已提交
566 567
	skb->dev = dev;

568
	if (dev->header_ops) {
L
Linus Torvalds 已提交
569
		/* The device has an explicit notion of ll header,
E
Eric Dumazet 已提交
570 571 572 573 574
		 * exported to higher levels.
		 *
		 * Otherwise, the device hides details of its frame
		 * structure, so that corresponding packet head is
		 * never delivered to user.
L
Linus Torvalds 已提交
575 576
		 */
		if (sk->sk_type != SOCK_DGRAM)
577
			skb_push(skb, skb->data - skb_mac_header(skb));
L
Linus Torvalds 已提交
578 579
		else if (skb->pkt_type == PACKET_OUTGOING) {
			/* Special case: outgoing packets have ll header at head */
580
			skb_pull(skb, skb_network_offset(skb));
L
Linus Torvalds 已提交
581 582 583 584 585
		}
	}

	snaplen = skb->len;

D
David S. Miller 已提交
586 587
	res = run_filter(skb, sk, snaplen);
	if (!res)
588
		goto drop_n_restore;
D
David S. Miller 已提交
589 590
	if (snaplen > res)
		snaplen = res;
L
Linus Torvalds 已提交
591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608

	if (atomic_read(&sk->sk_rmem_alloc) + skb->truesize >=
	    (unsigned)sk->sk_rcvbuf)
		goto drop_n_acct;

	if (skb_shared(skb)) {
		struct sk_buff *nskb = skb_clone(skb, GFP_ATOMIC);
		if (nskb == NULL)
			goto drop_n_acct;

		if (skb_head != skb->data) {
			skb->data = skb_head;
			skb->len = skb_len;
		}
		kfree_skb(skb);
		skb = nskb;
	}

609 610 611 612
	BUILD_BUG_ON(sizeof(*PACKET_SKB_CB(skb)) + MAX_ADDR_LEN - 8 >
		     sizeof(skb->cb));

	sll = &PACKET_SKB_CB(skb)->sa.ll;
L
Linus Torvalds 已提交
613 614 615 616
	sll->sll_family = AF_PACKET;
	sll->sll_hatype = dev->type;
	sll->sll_protocol = skb->protocol;
	sll->sll_pkttype = skb->pkt_type;
617
	if (unlikely(po->origdev))
618 619 620
		sll->sll_ifindex = orig_dev->ifindex;
	else
		sll->sll_ifindex = dev->ifindex;
L
Linus Torvalds 已提交
621

S
Stephen Hemminger 已提交
622
	sll->sll_halen = dev_parse_header(skb, sll->sll_addr);
L
Linus Torvalds 已提交
623

624
	PACKET_SKB_CB(skb)->origlen = skb->len;
625

L
Linus Torvalds 已提交
626 627 628 629 630
	if (pskb_trim(skb, snaplen))
		goto drop_n_acct;

	skb_set_owner_r(skb, sk);
	skb->dev = NULL;
E
Eric Dumazet 已提交
631
	skb_dst_drop(skb);
L
Linus Torvalds 已提交
632

633 634 635
	/* drop conntrack reference */
	nf_reset(skb);

L
Linus Torvalds 已提交
636 637
	spin_lock(&sk->sk_receive_queue.lock);
	po->stats.tp_packets++;
638
	skb->dropcount = atomic_read(&sk->sk_drops);
L
Linus Torvalds 已提交
639 640 641 642 643 644
	__skb_queue_tail(&sk->sk_receive_queue, skb);
	spin_unlock(&sk->sk_receive_queue.lock);
	sk->sk_data_ready(sk, skb->len);
	return 0;

drop_n_acct:
645
	po->stats.tp_drops = atomic_inc_return(&sk->sk_drops);
L
Linus Torvalds 已提交
646 647 648 649 650 651 652

drop_n_restore:
	if (skb_head != skb->data && skb_shared(skb)) {
		skb->data = skb_head;
		skb->len = skb_len;
	}
drop:
653
	consume_skb(skb);
L
Linus Torvalds 已提交
654 655 656
	return 0;
}

E
Eric Dumazet 已提交
657 658
static int tpacket_rcv(struct sk_buff *skb, struct net_device *dev,
		       struct packet_type *pt, struct net_device *orig_dev)
L
Linus Torvalds 已提交
659 660 661 662
{
	struct sock *sk;
	struct packet_sock *po;
	struct sockaddr_ll *sll;
663 664 665 666 667
	union {
		struct tpacket_hdr *h1;
		struct tpacket2_hdr *h2;
		void *raw;
	} h;
E
Eric Dumazet 已提交
668
	u8 *skb_head = skb->data;
L
Linus Torvalds 已提交
669
	int skb_len = skb->len;
D
David S. Miller 已提交
670
	unsigned int snaplen, res;
L
Linus Torvalds 已提交
671
	unsigned long status = TP_STATUS_LOSING|TP_STATUS_USER;
672
	unsigned short macoff, netoff, hdrlen;
L
Linus Torvalds 已提交
673
	struct sk_buff *copy_skb = NULL;
674
	struct timeval tv;
675
	struct timespec ts;
676
	struct skb_shared_hwtstamps *shhwtstamps = skb_hwtstamps(skb);
L
Linus Torvalds 已提交
677 678 679 680 681 682 683

	if (skb->pkt_type == PACKET_LOOPBACK)
		goto drop;

	sk = pt->af_packet_priv;
	po = pkt_sk(sk);

O
Octavian Purdila 已提交
684
	if (!net_eq(dev_net(dev), sock_net(sk)))
685 686
		goto drop;

687
	if (dev->header_ops) {
L
Linus Torvalds 已提交
688
		if (sk->sk_type != SOCK_DGRAM)
689
			skb_push(skb, skb->data - skb_mac_header(skb));
L
Linus Torvalds 已提交
690 691
		else if (skb->pkt_type == PACKET_OUTGOING) {
			/* Special case: outgoing packets have ll header at head */
692
			skb_pull(skb, skb_network_offset(skb));
L
Linus Torvalds 已提交
693 694 695
		}
	}

696 697 698
	if (skb->ip_summed == CHECKSUM_PARTIAL)
		status |= TP_STATUS_CSUMNOTREADY;

L
Linus Torvalds 已提交
699 700
	snaplen = skb->len;

D
David S. Miller 已提交
701 702
	res = run_filter(skb, sk, snaplen);
	if (!res)
703
		goto drop_n_restore;
D
David S. Miller 已提交
704 705
	if (snaplen > res)
		snaplen = res;
L
Linus Torvalds 已提交
706 707

	if (sk->sk_type == SOCK_DGRAM) {
708 709
		macoff = netoff = TPACKET_ALIGN(po->tp_hdrlen) + 16 +
				  po->tp_reserve;
L
Linus Torvalds 已提交
710
	} else {
711
		unsigned maclen = skb_network_offset(skb);
712
		netoff = TPACKET_ALIGN(po->tp_hdrlen +
713 714
				       (maclen < 16 ? 16 : maclen)) +
			po->tp_reserve;
L
Linus Torvalds 已提交
715 716 717
		macoff = netoff - maclen;
	}

J
Johann Baudy 已提交
718
	if (macoff + snaplen > po->rx_ring.frame_size) {
L
Linus Torvalds 已提交
719 720 721 722 723 724 725 726 727 728 729 730
		if (po->copy_thresh &&
		    atomic_read(&sk->sk_rmem_alloc) + skb->truesize <
		    (unsigned)sk->sk_rcvbuf) {
			if (skb_shared(skb)) {
				copy_skb = skb_clone(skb, GFP_ATOMIC);
			} else {
				copy_skb = skb_get(skb);
				skb_head = skb->data;
			}
			if (copy_skb)
				skb_set_owner_r(copy_skb, sk);
		}
J
Johann Baudy 已提交
731
		snaplen = po->rx_ring.frame_size - macoff;
L
Linus Torvalds 已提交
732 733 734 735 736
		if ((int)snaplen < 0)
			snaplen = 0;
	}

	spin_lock(&sk->sk_receive_queue.lock);
J
Johann Baudy 已提交
737
	h.raw = packet_current_frame(po, &po->rx_ring, TP_STATUS_KERNEL);
738
	if (!h.raw)
L
Linus Torvalds 已提交
739
		goto ring_is_full;
J
Johann Baudy 已提交
740
	packet_increment_head(&po->rx_ring);
L
Linus Torvalds 已提交
741 742 743 744 745 746 747 748 749
	po->stats.tp_packets++;
	if (copy_skb) {
		status |= TP_STATUS_COPY;
		__skb_queue_tail(&sk->sk_receive_queue, copy_skb);
	}
	if (!po->stats.tp_drops)
		status &= ~TP_STATUS_LOSING;
	spin_unlock(&sk->sk_receive_queue.lock);

750
	skb_copy_bits(skb, 0, h.raw + macoff, snaplen);
L
Linus Torvalds 已提交
751

752 753 754 755 756 757
	switch (po->tp_version) {
	case TPACKET_V1:
		h.h1->tp_len = skb->len;
		h.h1->tp_snaplen = snaplen;
		h.h1->tp_mac = macoff;
		h.h1->tp_net = netoff;
758 759 760 761 762 763 764
		if ((po->tp_tstamp & SOF_TIMESTAMPING_SYS_HARDWARE)
				&& shhwtstamps->syststamp.tv64)
			tv = ktime_to_timeval(shhwtstamps->syststamp);
		else if ((po->tp_tstamp & SOF_TIMESTAMPING_RAW_HARDWARE)
				&& shhwtstamps->hwtstamp.tv64)
			tv = ktime_to_timeval(shhwtstamps->hwtstamp);
		else if (skb->tstamp.tv64)
765 766 767 768 769 770 771 772 773 774 775 776
			tv = ktime_to_timeval(skb->tstamp);
		else
			do_gettimeofday(&tv);
		h.h1->tp_sec = tv.tv_sec;
		h.h1->tp_usec = tv.tv_usec;
		hdrlen = sizeof(*h.h1);
		break;
	case TPACKET_V2:
		h.h2->tp_len = skb->len;
		h.h2->tp_snaplen = snaplen;
		h.h2->tp_mac = macoff;
		h.h2->tp_net = netoff;
777 778 779 780 781 782 783
		if ((po->tp_tstamp & SOF_TIMESTAMPING_SYS_HARDWARE)
				&& shhwtstamps->syststamp.tv64)
			ts = ktime_to_timespec(shhwtstamps->syststamp);
		else if ((po->tp_tstamp & SOF_TIMESTAMPING_RAW_HARDWARE)
				&& shhwtstamps->hwtstamp.tv64)
			ts = ktime_to_timespec(shhwtstamps->hwtstamp);
		else if (skb->tstamp.tv64)
784 785 786 787 788
			ts = ktime_to_timespec(skb->tstamp);
		else
			getnstimeofday(&ts);
		h.h2->tp_sec = ts.tv_sec;
		h.h2->tp_nsec = ts.tv_nsec;
E
Eric Dumazet 已提交
789
		h.h2->tp_vlan_tci = vlan_tx_tag_get(skb);
790 791 792 793 794
		hdrlen = sizeof(*h.h2);
		break;
	default:
		BUG();
	}
L
Linus Torvalds 已提交
795

796
	sll = h.raw + TPACKET_ALIGN(hdrlen);
S
Stephen Hemminger 已提交
797
	sll->sll_halen = dev_parse_header(skb, sll->sll_addr);
L
Linus Torvalds 已提交
798 799 800 801
	sll->sll_family = AF_PACKET;
	sll->sll_hatype = dev->type;
	sll->sll_protocol = skb->protocol;
	sll->sll_pkttype = skb->pkt_type;
802
	if (unlikely(po->origdev))
803 804 805
		sll->sll_ifindex = orig_dev->ifindex;
	else
		sll->sll_ifindex = dev->ifindex;
L
Linus Torvalds 已提交
806

807
	__packet_set_status(po, h.raw, status);
R
Ralf Baechle 已提交
808
	smp_mb();
L
Linus Torvalds 已提交
809
	{
810 811 812 813 814
		u8 *start, *end;

		end = (u8 *)PAGE_ALIGN((unsigned long)h.raw + macoff + snaplen);
		for (start = h.raw; start < end; start += PAGE_SIZE)
			flush_dcache_page(pgv_to_page(start));
L
Linus Torvalds 已提交
815 816 817 818 819 820 821 822 823 824
	}

	sk->sk_data_ready(sk, 0);

drop_n_restore:
	if (skb_head != skb->data && skb_shared(skb)) {
		skb->data = skb_head;
		skb->len = skb_len;
	}
drop:
825
	kfree_skb(skb);
L
Linus Torvalds 已提交
826 827 828 829 830 831 832
	return 0;

ring_is_full:
	po->stats.tp_drops++;
	spin_unlock(&sk->sk_receive_queue.lock);

	sk->sk_data_ready(sk, 0);
833
	kfree_skb(copy_skb);
L
Linus Torvalds 已提交
834 835 836
	goto drop_n_restore;
}

J
Johann Baudy 已提交
837 838 839
static void tpacket_destruct_skb(struct sk_buff *skb)
{
	struct packet_sock *po = pkt_sk(skb->sk);
E
Eric Dumazet 已提交
840
	void *ph;
L
Linus Torvalds 已提交
841

J
Johann Baudy 已提交
842
	BUG_ON(skb == NULL);
L
Linus Torvalds 已提交
843

J
Johann Baudy 已提交
844 845 846 847 848 849 850 851 852 853 854
	if (likely(po->tx_ring.pg_vec)) {
		ph = skb_shinfo(skb)->destructor_arg;
		BUG_ON(__packet_get_status(po, ph) != TP_STATUS_SENDING);
		BUG_ON(atomic_read(&po->tx_ring.pending) == 0);
		atomic_dec(&po->tx_ring.pending);
		__packet_set_status(po, ph, TP_STATUS_AVAILABLE);
	}

	sock_wfree(skb);
}

E
Eric Dumazet 已提交
855 856 857
static int tpacket_fill_skb(struct packet_sock *po, struct sk_buff *skb,
		void *frame, struct net_device *dev, int size_max,
		__be16 proto, unsigned char *addr)
J
Johann Baudy 已提交
858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874
{
	union {
		struct tpacket_hdr *h1;
		struct tpacket2_hdr *h2;
		void *raw;
	} ph;
	int to_write, offset, len, tp_len, nr_frags, len_max;
	struct socket *sock = po->sk.sk_socket;
	struct page *page;
	void *data;
	int err;

	ph.raw = frame;

	skb->protocol = proto;
	skb->dev = dev;
	skb->priority = po->sk.sk_priority;
875
	skb->mark = po->sk.sk_mark;
J
Johann Baudy 已提交
876 877 878 879 880 881 882 883 884 885 886
	skb_shinfo(skb)->destructor_arg = ph.raw;

	switch (po->tp_version) {
	case TPACKET_V2:
		tp_len = ph.h2->tp_len;
		break;
	default:
		tp_len = ph.h1->tp_len;
		break;
	}
	if (unlikely(tp_len > size_max)) {
E
Eric Dumazet 已提交
887
		pr_err("packet size is too long (%d > %d)\n", tp_len, size_max);
J
Johann Baudy 已提交
888 889 890 891 892 893 894 895 896 897 898 899 900 901
		return -EMSGSIZE;
	}

	skb_reserve(skb, LL_RESERVED_SPACE(dev));
	skb_reset_network_header(skb);

	data = ph.raw + po->tp_hdrlen - sizeof(struct sockaddr_ll);
	to_write = tp_len;

	if (sock->type == SOCK_DGRAM) {
		err = dev_hard_header(skb, dev, ntohs(proto), addr,
				NULL, tp_len);
		if (unlikely(err < 0))
			return -EINVAL;
E
Eric Dumazet 已提交
902
	} else if (dev->hard_header_len) {
J
Johann Baudy 已提交
903 904
		/* net device doesn't like empty head */
		if (unlikely(tp_len <= dev->hard_header_len)) {
E
Eric Dumazet 已提交
905 906
			pr_err("packet size is too short (%d < %d)\n",
			       tp_len, dev->hard_header_len);
J
Johann Baudy 已提交
907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933
			return -EINVAL;
		}

		skb_push(skb, dev->hard_header_len);
		err = skb_store_bits(skb, 0, data,
				dev->hard_header_len);
		if (unlikely(err))
			return err;

		data += dev->hard_header_len;
		to_write -= dev->hard_header_len;
	}

	err = -EFAULT;
	offset = offset_in_page(data);
	len_max = PAGE_SIZE - offset;
	len = ((to_write > len_max) ? len_max : to_write);

	skb->data_len = to_write;
	skb->len += to_write;
	skb->truesize += to_write;
	atomic_add(to_write, &po->sk.sk_wmem_alloc);

	while (likely(to_write)) {
		nr_frags = skb_shinfo(skb)->nr_frags;

		if (unlikely(nr_frags >= MAX_SKB_FRAGS)) {
E
Eric Dumazet 已提交
934 935
			pr_err("Packet exceed the number of skb frags(%lu)\n",
			       MAX_SKB_FRAGS);
J
Johann Baudy 已提交
936 937 938
			return -EFAULT;
		}

939 940
		page = pgv_to_page(data);
		data += len;
J
Johann Baudy 已提交
941 942
		flush_dcache_page(page);
		get_page(page);
943
		skb_fill_page_desc(skb, nr_frags, page, offset, len);
J
Johann Baudy 已提交
944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959
		to_write -= len;
		offset = 0;
		len_max = PAGE_SIZE;
		len = ((to_write > len_max) ? len_max : to_write);
	}

	return tp_len;
}

static int tpacket_snd(struct packet_sock *po, struct msghdr *msg)
{
	struct socket *sock;
	struct sk_buff *skb;
	struct net_device *dev;
	__be16 proto;
	int ifindex, err, reserve = 0;
E
Eric Dumazet 已提交
960 961
	void *ph;
	struct sockaddr_ll *saddr = (struct sockaddr_ll *)msg->msg_name;
J
Johann Baudy 已提交
962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000
	int tp_len, size_max;
	unsigned char *addr;
	int len_sum = 0;
	int status = 0;

	sock = po->sk.sk_socket;

	mutex_lock(&po->pg_vec_lock);

	err = -EBUSY;
	if (saddr == NULL) {
		ifindex	= po->ifindex;
		proto	= po->num;
		addr	= NULL;
	} else {
		err = -EINVAL;
		if (msg->msg_namelen < sizeof(struct sockaddr_ll))
			goto out;
		if (msg->msg_namelen < (saddr->sll_halen
					+ offsetof(struct sockaddr_ll,
						sll_addr)))
			goto out;
		ifindex	= saddr->sll_ifindex;
		proto	= saddr->sll_protocol;
		addr	= saddr->sll_addr;
	}

	dev = dev_get_by_index(sock_net(&po->sk), ifindex);
	err = -ENXIO;
	if (unlikely(dev == NULL))
		goto out;

	reserve = dev->hard_header_len;

	err = -ENETDOWN;
	if (unlikely(!(dev->flags & IFF_UP)))
		goto out_put;

	size_max = po->tx_ring.frame_size
1001
		- (po->tp_hdrlen - sizeof(struct sockaddr_ll));
J
Johann Baudy 已提交
1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046

	if (size_max > dev->mtu + reserve)
		size_max = dev->mtu + reserve;

	do {
		ph = packet_current_frame(po, &po->tx_ring,
				TP_STATUS_SEND_REQUEST);

		if (unlikely(ph == NULL)) {
			schedule();
			continue;
		}

		status = TP_STATUS_SEND_REQUEST;
		skb = sock_alloc_send_skb(&po->sk,
				LL_ALLOCATED_SPACE(dev)
				+ sizeof(struct sockaddr_ll),
				0, &err);

		if (unlikely(skb == NULL))
			goto out_status;

		tp_len = tpacket_fill_skb(po, skb, ph, dev, size_max, proto,
				addr);

		if (unlikely(tp_len < 0)) {
			if (po->tp_loss) {
				__packet_set_status(po, ph,
						TP_STATUS_AVAILABLE);
				packet_increment_head(&po->tx_ring);
				kfree_skb(skb);
				continue;
			} else {
				status = TP_STATUS_WRONG_FORMAT;
				err = tp_len;
				goto out_status;
			}
		}

		skb->destructor = tpacket_destruct_skb;
		__packet_set_status(po, ph, TP_STATUS_SENDING);
		atomic_inc(&po->tx_ring.pending);

		status = TP_STATUS_SEND_REQUEST;
		err = dev_queue_xmit(skb);
1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060
		if (unlikely(err > 0)) {
			err = net_xmit_errno(err);
			if (err && __packet_get_status(po, ph) ==
				   TP_STATUS_AVAILABLE) {
				/* skb was destructed already */
				skb = NULL;
				goto out_status;
			}
			/*
			 * skb was dropped but not destructed yet;
			 * let's treat it like congestion or err < 0
			 */
			err = 0;
		}
J
Johann Baudy 已提交
1061 1062
		packet_increment_head(&po->tx_ring);
		len_sum += tp_len;
1063 1064 1065 1066
	} while (likely((ph != NULL) ||
			((!(msg->msg_flags & MSG_DONTWAIT)) &&
			 (atomic_read(&po->tx_ring.pending))))
		);
J
Johann Baudy 已提交
1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080

	err = len_sum;
	goto out_put;

out_status:
	__packet_set_status(po, ph, status);
	kfree_skb(skb);
out_put:
	dev_put(dev);
out:
	mutex_unlock(&po->pg_vec_lock);
	return err;
}

1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104
static inline struct sk_buff *packet_alloc_skb(struct sock *sk, size_t prepad,
					       size_t reserve, size_t len,
					       size_t linear, int noblock,
					       int *err)
{
	struct sk_buff *skb;

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

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

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

	return skb;
}

J
Johann Baudy 已提交
1105
static int packet_snd(struct socket *sock,
L
Linus Torvalds 已提交
1106 1107 1108
			  struct msghdr *msg, size_t len)
{
	struct sock *sk = sock->sk;
E
Eric Dumazet 已提交
1109
	struct sockaddr_ll *saddr = (struct sockaddr_ll *)msg->msg_name;
L
Linus Torvalds 已提交
1110 1111
	struct sk_buff *skb;
	struct net_device *dev;
A
Al Viro 已提交
1112
	__be16 proto;
L
Linus Torvalds 已提交
1113 1114
	unsigned char *addr;
	int ifindex, err, reserve = 0;
1115 1116 1117 1118 1119
	struct virtio_net_hdr vnet_hdr = { 0 };
	int offset = 0;
	int vnet_hdr_len;
	struct packet_sock *po = pkt_sk(sk);
	unsigned short gso_type = 0;
L
Linus Torvalds 已提交
1120 1121

	/*
1122
	 *	Get and verify the address.
L
Linus Torvalds 已提交
1123
	 */
1124

L
Linus Torvalds 已提交
1125 1126 1127 1128 1129 1130 1131 1132
	if (saddr == NULL) {
		ifindex	= po->ifindex;
		proto	= po->num;
		addr	= NULL;
	} else {
		err = -EINVAL;
		if (msg->msg_namelen < sizeof(struct sockaddr_ll))
			goto out;
1133 1134
		if (msg->msg_namelen < (saddr->sll_halen + offsetof(struct sockaddr_ll, sll_addr)))
			goto out;
L
Linus Torvalds 已提交
1135 1136 1137 1138 1139 1140
		ifindex	= saddr->sll_ifindex;
		proto	= saddr->sll_protocol;
		addr	= saddr->sll_addr;
	}


1141
	dev = dev_get_by_index(sock_net(sk), ifindex);
L
Linus Torvalds 已提交
1142 1143 1144 1145 1146 1147
	err = -ENXIO;
	if (dev == NULL)
		goto out_unlock;
	if (sock->type == SOCK_RAW)
		reserve = dev->hard_header_len;

1148 1149 1150 1151
	err = -ENETDOWN;
	if (!(dev->flags & IFF_UP))
		goto out_unlock;

1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199
	if (po->has_vnet_hdr) {
		vnet_hdr_len = sizeof(vnet_hdr);

		err = -EINVAL;
		if (len < vnet_hdr_len)
			goto out_unlock;

		len -= vnet_hdr_len;

		err = memcpy_fromiovec((void *)&vnet_hdr, msg->msg_iov,
				       vnet_hdr_len);
		if (err < 0)
			goto out_unlock;

		if ((vnet_hdr.flags & VIRTIO_NET_HDR_F_NEEDS_CSUM) &&
		    (vnet_hdr.csum_start + vnet_hdr.csum_offset + 2 >
		      vnet_hdr.hdr_len))
			vnet_hdr.hdr_len = vnet_hdr.csum_start +
						 vnet_hdr.csum_offset + 2;

		err = -EINVAL;
		if (vnet_hdr.hdr_len > len)
			goto out_unlock;

		if (vnet_hdr.gso_type != VIRTIO_NET_HDR_GSO_NONE) {
			switch (vnet_hdr.gso_type & ~VIRTIO_NET_HDR_GSO_ECN) {
			case VIRTIO_NET_HDR_GSO_TCPV4:
				gso_type = SKB_GSO_TCPV4;
				break;
			case VIRTIO_NET_HDR_GSO_TCPV6:
				gso_type = SKB_GSO_TCPV6;
				break;
			case VIRTIO_NET_HDR_GSO_UDP:
				gso_type = SKB_GSO_UDP;
				break;
			default:
				goto out_unlock;
			}

			if (vnet_hdr.gso_type & VIRTIO_NET_HDR_GSO_ECN)
				gso_type |= SKB_GSO_TCP_ECN;

			if (vnet_hdr.gso_size == 0)
				goto out_unlock;

		}
	}

L
Linus Torvalds 已提交
1200
	err = -EMSGSIZE;
1201
	if (!gso_type && (len > dev->mtu+reserve))
L
Linus Torvalds 已提交
1202 1203
		goto out_unlock;

1204 1205 1206 1207
	err = -ENOBUFS;
	skb = packet_alloc_skb(sk, LL_ALLOCATED_SPACE(dev),
			       LL_RESERVED_SPACE(dev), len, vnet_hdr.hdr_len,
			       msg->msg_flags & MSG_DONTWAIT, &err);
E
Eric Dumazet 已提交
1208
	if (skb == NULL)
L
Linus Torvalds 已提交
1209 1210
		goto out_unlock;

1211
	skb_set_network_header(skb, reserve);
L
Linus Torvalds 已提交
1212

1213 1214
	err = -EINVAL;
	if (sock->type == SOCK_DGRAM &&
1215
	    (offset = dev_hard_header(skb, dev, ntohs(proto), addr, NULL, len)) < 0)
1216
		goto out_free;
L
Linus Torvalds 已提交
1217 1218

	/* Returns -EFAULT on error */
1219
	err = skb_copy_datagram_from_iovec(skb, offset, msg->msg_iov, 0, len);
L
Linus Torvalds 已提交
1220 1221
	if (err)
		goto out_free;
1222
	err = sock_tx_timestamp(sk, &skb_shinfo(skb)->tx_flags);
1223 1224
	if (err < 0)
		goto out_free;
L
Linus Torvalds 已提交
1225 1226 1227 1228

	skb->protocol = proto;
	skb->dev = dev;
	skb->priority = sk->sk_priority;
1229
	skb->mark = sk->sk_mark;
L
Linus Torvalds 已提交
1230

1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249
	if (po->has_vnet_hdr) {
		if (vnet_hdr.flags & VIRTIO_NET_HDR_F_NEEDS_CSUM) {
			if (!skb_partial_csum_set(skb, vnet_hdr.csum_start,
						  vnet_hdr.csum_offset)) {
				err = -EINVAL;
				goto out_free;
			}
		}

		skb_shinfo(skb)->gso_size = vnet_hdr.gso_size;
		skb_shinfo(skb)->gso_type = gso_type;

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

		len += vnet_hdr_len;
	}

L
Linus Torvalds 已提交
1250 1251 1252 1253 1254 1255 1256 1257 1258 1259
	/*
	 *	Now send it
	 */

	err = dev_queue_xmit(skb);
	if (err > 0 && (err = net_xmit_errno(err)) != 0)
		goto out_unlock;

	dev_put(dev);

E
Eric Dumazet 已提交
1260
	return len;
L
Linus Torvalds 已提交
1261 1262 1263 1264 1265 1266 1267 1268 1269 1270

out_free:
	kfree_skb(skb);
out_unlock:
	if (dev)
		dev_put(dev);
out:
	return err;
}

J
Johann Baudy 已提交
1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281
static int packet_sendmsg(struct kiocb *iocb, struct socket *sock,
		struct msghdr *msg, size_t len)
{
	struct sock *sk = sock->sk;
	struct packet_sock *po = pkt_sk(sk);
	if (po->tx_ring.pg_vec)
		return tpacket_snd(po, msg);
	else
		return packet_snd(sock, msg, len);
}

L
Linus Torvalds 已提交
1282 1283 1284 1285 1286 1287 1288 1289 1290
/*
 *	Close a PACKET socket. This is fairly simple. We immediately go
 *	to 'closed' state and remove our protocol entry in the device list.
 */

static int packet_release(struct socket *sock)
{
	struct sock *sk = sock->sk;
	struct packet_sock *po;
1291
	struct net *net;
J
Johann Baudy 已提交
1292
	struct tpacket_req req;
L
Linus Torvalds 已提交
1293 1294 1295 1296

	if (!sk)
		return 0;

1297
	net = sock_net(sk);
L
Linus Torvalds 已提交
1298 1299
	po = pkt_sk(sk);

1300 1301
	spin_lock_bh(&net->packet.sklist_lock);
	sk_del_node_init_rcu(sk);
1302
	sock_prot_inuse_add(net, sk->sk_prot, -1);
1303
	spin_unlock_bh(&net->packet.sklist_lock);
L
Linus Torvalds 已提交
1304

1305
	spin_lock(&po->bind_lock);
L
Linus Torvalds 已提交
1306 1307
	if (po->running) {
		/*
1308
		 * Remove from protocol table
L
Linus Torvalds 已提交
1309 1310 1311
		 */
		po->running = 0;
		po->num = 0;
1312
		__dev_remove_pack(&po->prot_hook);
L
Linus Torvalds 已提交
1313 1314
		__sock_put(sk);
	}
1315
	spin_unlock(&po->bind_lock);
L
Linus Torvalds 已提交
1316 1317 1318

	packet_flush_mclist(sk);

J
Johann Baudy 已提交
1319 1320 1321 1322 1323 1324 1325
	memset(&req, 0, sizeof(req));

	if (po->rx_ring.pg_vec)
		packet_set_ring(sk, &req, 1, 0);

	if (po->tx_ring.pg_vec)
		packet_set_ring(sk, &req, 1, 1);
L
Linus Torvalds 已提交
1326

1327
	synchronize_net();
L
Linus Torvalds 已提交
1328 1329 1330 1331 1332 1333 1334 1335 1336
	/*
	 *	Now the socket is dead. No more input will appear.
	 */
	sock_orphan(sk);
	sock->sk = NULL;

	/* Purge queues */

	skb_queue_purge(&sk->sk_receive_queue);
1337
	sk_refcnt_debug_release(sk);
L
Linus Torvalds 已提交
1338 1339 1340 1341 1342 1343 1344 1345 1346

	sock_put(sk);
	return 0;
}

/*
 *	Attach a packet hook.
 */

A
Al Viro 已提交
1347
static int packet_do_bind(struct sock *sk, struct net_device *dev, __be16 protocol)
L
Linus Torvalds 已提交
1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374
{
	struct packet_sock *po = pkt_sk(sk);
	/*
	 *	Detach an existing hook if present.
	 */

	lock_sock(sk);

	spin_lock(&po->bind_lock);
	if (po->running) {
		__sock_put(sk);
		po->running = 0;
		po->num = 0;
		spin_unlock(&po->bind_lock);
		dev_remove_pack(&po->prot_hook);
		spin_lock(&po->bind_lock);
	}

	po->num = protocol;
	po->prot_hook.type = protocol;
	po->prot_hook.dev = dev;

	po->ifindex = dev ? dev->ifindex : 0;

	if (protocol == 0)
		goto out_unlock;

1375
	if (!dev || (dev->flags & IFF_UP)) {
L
Linus Torvalds 已提交
1376 1377 1378
		dev_add_pack(&po->prot_hook);
		sock_hold(sk);
		po->running = 1;
1379 1380 1381 1382
	} else {
		sk->sk_err = ENETDOWN;
		if (!sock_flag(sk, SOCK_DEAD))
			sk->sk_error_report(sk);
L
Linus Torvalds 已提交
1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394
	}

out_unlock:
	spin_unlock(&po->bind_lock);
	release_sock(sk);
	return 0;
}

/*
 *	Bind a packet socket to a device
 */

E
Eric Dumazet 已提交
1395 1396
static int packet_bind_spkt(struct socket *sock, struct sockaddr *uaddr,
			    int addr_len)
L
Linus Torvalds 已提交
1397
{
E
Eric Dumazet 已提交
1398
	struct sock *sk = sock->sk;
L
Linus Torvalds 已提交
1399 1400 1401
	char name[15];
	struct net_device *dev;
	int err = -ENODEV;
1402

L
Linus Torvalds 已提交
1403 1404 1405
	/*
	 *	Check legality
	 */
1406

1407
	if (addr_len != sizeof(struct sockaddr))
L
Linus Torvalds 已提交
1408
		return -EINVAL;
E
Eric Dumazet 已提交
1409
	strlcpy(name, uaddr->sa_data, sizeof(name));
L
Linus Torvalds 已提交
1410

1411
	dev = dev_get_by_name(sock_net(sk), name);
L
Linus Torvalds 已提交
1412 1413 1414 1415 1416 1417 1418 1419 1420
	if (dev) {
		err = packet_do_bind(sk, dev, pkt_sk(sk)->num);
		dev_put(dev);
	}
	return err;
}

static int packet_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len)
{
E
Eric Dumazet 已提交
1421 1422
	struct sockaddr_ll *sll = (struct sockaddr_ll *)uaddr;
	struct sock *sk = sock->sk;
L
Linus Torvalds 已提交
1423 1424 1425 1426 1427 1428 1429
	struct net_device *dev = NULL;
	int err;


	/*
	 *	Check legality
	 */
1430

L
Linus Torvalds 已提交
1431 1432 1433 1434 1435 1436 1437
	if (addr_len < sizeof(struct sockaddr_ll))
		return -EINVAL;
	if (sll->sll_family != AF_PACKET)
		return -EINVAL;

	if (sll->sll_ifindex) {
		err = -ENODEV;
1438
		dev = dev_get_by_index(sock_net(sk), sll->sll_ifindex);
L
Linus Torvalds 已提交
1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456
		if (dev == NULL)
			goto out;
	}
	err = packet_do_bind(sk, dev, sll->sll_protocol ? : pkt_sk(sk)->num);
	if (dev)
		dev_put(dev);

out:
	return err;
}

static struct proto packet_proto = {
	.name	  = "PACKET",
	.owner	  = THIS_MODULE,
	.obj_size = sizeof(struct packet_sock),
};

/*
1457
 *	Create a packet of type SOCK_PACKET.
L
Linus Torvalds 已提交
1458 1459
 */

1460 1461
static int packet_create(struct net *net, struct socket *sock, int protocol,
			 int kern)
L
Linus Torvalds 已提交
1462 1463 1464
{
	struct sock *sk;
	struct packet_sock *po;
A
Al Viro 已提交
1465
	__be16 proto = (__force __be16)protocol; /* weird, but documented */
L
Linus Torvalds 已提交
1466 1467 1468 1469
	int err;

	if (!capable(CAP_NET_RAW))
		return -EPERM;
1470 1471
	if (sock->type != SOCK_DGRAM && sock->type != SOCK_RAW &&
	    sock->type != SOCK_PACKET)
L
Linus Torvalds 已提交
1472 1473 1474 1475 1476
		return -ESOCKTNOSUPPORT;

	sock->state = SS_UNCONNECTED;

	err = -ENOBUFS;
1477
	sk = sk_alloc(net, PF_PACKET, GFP_KERNEL, &packet_proto);
L
Linus Torvalds 已提交
1478 1479 1480 1481 1482 1483
	if (sk == NULL)
		goto out;

	sock->ops = &packet_ops;
	if (sock->type == SOCK_PACKET)
		sock->ops = &packet_ops_spkt;
1484

L
Linus Torvalds 已提交
1485 1486 1487 1488
	sock_init_data(sock, sk);

	po = pkt_sk(sk);
	sk->sk_family = PF_PACKET;
A
Al Viro 已提交
1489
	po->num = proto;
L
Linus Torvalds 已提交
1490 1491

	sk->sk_destruct = packet_sock_destruct;
1492
	sk_refcnt_debug_inc(sk);
L
Linus Torvalds 已提交
1493 1494 1495 1496 1497 1498

	/*
	 *	Attach a protocol block
	 */

	spin_lock_init(&po->bind_lock);
1499
	mutex_init(&po->pg_vec_lock);
L
Linus Torvalds 已提交
1500
	po->prot_hook.func = packet_rcv;
1501

L
Linus Torvalds 已提交
1502 1503
	if (sock->type == SOCK_PACKET)
		po->prot_hook.func = packet_rcv_spkt;
1504

L
Linus Torvalds 已提交
1505 1506
	po->prot_hook.af_packet_priv = sk;

A
Al Viro 已提交
1507 1508
	if (proto) {
		po->prot_hook.type = proto;
L
Linus Torvalds 已提交
1509 1510 1511 1512 1513
		dev_add_pack(&po->prot_hook);
		sock_hold(sk);
		po->running = 1;
	}

1514 1515
	spin_lock_bh(&net->packet.sklist_lock);
	sk_add_node_rcu(sk, &net->packet.sklist);
1516
	sock_prot_inuse_add(net, &packet_proto, 1);
1517 1518
	spin_unlock_bh(&net->packet.sklist_lock);

E
Eric Dumazet 已提交
1519
	return 0;
L
Linus Torvalds 已提交
1520 1521 1522 1523
out:
	return err;
}

1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568
static int packet_recv_error(struct sock *sk, struct msghdr *msg, int len)
{
	struct sock_exterr_skb *serr;
	struct sk_buff *skb, *skb2;
	int copied, err;

	err = -EAGAIN;
	skb = skb_dequeue(&sk->sk_error_queue);
	if (skb == NULL)
		goto out;

	copied = skb->len;
	if (copied > len) {
		msg->msg_flags |= MSG_TRUNC;
		copied = len;
	}
	err = skb_copy_datagram_iovec(skb, 0, msg->msg_iov, copied);
	if (err)
		goto out_free_skb;

	sock_recv_timestamp(msg, sk, skb);

	serr = SKB_EXT_ERR(skb);
	put_cmsg(msg, SOL_PACKET, PACKET_TX_TIMESTAMP,
		 sizeof(serr->ee), &serr->ee);

	msg->msg_flags |= MSG_ERRQUEUE;
	err = copied;

	/* Reset and regenerate socket error */
	spin_lock_bh(&sk->sk_error_queue.lock);
	sk->sk_err = 0;
	if ((skb2 = skb_peek(&sk->sk_error_queue)) != NULL) {
		sk->sk_err = SKB_EXT_ERR(skb2)->ee.ee_errno;
		spin_unlock_bh(&sk->sk_error_queue.lock);
		sk->sk_error_report(sk);
	} else
		spin_unlock_bh(&sk->sk_error_queue.lock);

out_free_skb:
	kfree_skb(skb);
out:
	return err;
}

L
Linus Torvalds 已提交
1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579
/*
 *	Pull a packet from our receive queue and hand it to the user.
 *	If necessary we block.
 */

static int packet_recvmsg(struct kiocb *iocb, struct socket *sock,
			  struct msghdr *msg, size_t len, int flags)
{
	struct sock *sk = sock->sk;
	struct sk_buff *skb;
	int copied, err;
1580
	struct sockaddr_ll *sll;
1581
	int vnet_hdr_len = 0;
L
Linus Torvalds 已提交
1582 1583

	err = -EINVAL;
1584
	if (flags & ~(MSG_PEEK|MSG_DONTWAIT|MSG_TRUNC|MSG_CMSG_COMPAT|MSG_ERRQUEUE))
L
Linus Torvalds 已提交
1585 1586 1587 1588 1589 1590 1591 1592
		goto out;

#if 0
	/* What error should we return now? EUNATTACH? */
	if (pkt_sk(sk)->ifindex < 0)
		return -ENODEV;
#endif

1593 1594 1595 1596 1597
	if (flags & MSG_ERRQUEUE) {
		err = packet_recv_error(sk, msg, len);
		goto out;
	}

L
Linus Torvalds 已提交
1598 1599 1600 1601 1602 1603 1604 1605 1606
	/*
	 *	Call the generic datagram receiver. This handles all sorts
	 *	of horrible races and re-entrancy so we can forget about it
	 *	in the protocol layers.
	 *
	 *	Now it will return ENETDOWN, if device have just gone down,
	 *	but then it will block.
	 */

E
Eric Dumazet 已提交
1607
	skb = skb_recv_datagram(sk, flags, flags & MSG_DONTWAIT, &err);
L
Linus Torvalds 已提交
1608 1609

	/*
1610
	 *	An error occurred so return it. Because skb_recv_datagram()
L
Linus Torvalds 已提交
1611 1612 1613 1614
	 *	handles the blocking we don't see and worry about blocking
	 *	retries.
	 */

1615
	if (skb == NULL)
L
Linus Torvalds 已提交
1616 1617
		goto out;

1618 1619 1620 1621 1622
	if (pkt_sk(sk)->has_vnet_hdr) {
		struct virtio_net_hdr vnet_hdr = { 0 };

		err = -EINVAL;
		vnet_hdr_len = sizeof(vnet_hdr);
1623
		if (len < vnet_hdr_len)
1624 1625
			goto out_free;

1626 1627
		len -= vnet_hdr_len;

1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661
		if (skb_is_gso(skb)) {
			struct skb_shared_info *sinfo = skb_shinfo(skb);

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

		if (skb->ip_summed == CHECKSUM_PARTIAL) {
			vnet_hdr.flags = VIRTIO_NET_HDR_F_NEEDS_CSUM;
			vnet_hdr.csum_start = skb->csum_start -
							skb_headroom(skb);
			vnet_hdr.csum_offset = skb->csum_offset;
		} /* else everything is zero */

		err = memcpy_toiovec(msg->msg_iov, (void *)&vnet_hdr,
				     vnet_hdr_len);
		if (err < 0)
			goto out_free;
	}

1662 1663 1664 1665 1666
	/*
	 *	If the address length field is there to be filled in, we fill
	 *	it in now.
	 */

1667
	sll = &PACKET_SKB_CB(skb)->sa.ll;
1668 1669 1670 1671 1672
	if (sock->type == SOCK_PACKET)
		msg->msg_namelen = sizeof(struct sockaddr_pkt);
	else
		msg->msg_namelen = sll->sll_halen + offsetof(struct sockaddr_ll, sll_addr);

L
Linus Torvalds 已提交
1673 1674 1675 1676 1677 1678
	/*
	 *	You lose any data beyond the buffer you gave. If it worries a
	 *	user program they can ask the device for its MTU anyway.
	 */

	copied = skb->len;
E
Eric Dumazet 已提交
1679 1680 1681
	if (copied > len) {
		copied = len;
		msg->msg_flags |= MSG_TRUNC;
L
Linus Torvalds 已提交
1682 1683 1684 1685 1686 1687
	}

	err = skb_copy_datagram_iovec(skb, 0, msg->msg_iov, copied);
	if (err)
		goto out_free;

1688
	sock_recv_ts_and_drops(msg, sk, skb);
L
Linus Torvalds 已提交
1689 1690

	if (msg->msg_name)
1691 1692
		memcpy(msg->msg_name, &PACKET_SKB_CB(skb)->sa,
		       msg->msg_namelen);
L
Linus Torvalds 已提交
1693

1694
	if (pkt_sk(sk)->auxdata) {
1695 1696 1697 1698 1699 1700 1701 1702
		struct tpacket_auxdata aux;

		aux.tp_status = TP_STATUS_USER;
		if (skb->ip_summed == CHECKSUM_PARTIAL)
			aux.tp_status |= TP_STATUS_CSUMNOTREADY;
		aux.tp_len = PACKET_SKB_CB(skb)->origlen;
		aux.tp_snaplen = skb->len;
		aux.tp_mac = 0;
1703
		aux.tp_net = skb_network_offset(skb);
E
Eric Dumazet 已提交
1704
		aux.tp_vlan_tci = vlan_tx_tag_get(skb);
1705 1706

		put_cmsg(msg, SOL_PACKET, PACKET_AUXDATA, sizeof(aux), &aux);
1707 1708
	}

L
Linus Torvalds 已提交
1709 1710 1711 1712
	/*
	 *	Free or return the buffer as appropriate. Again this
	 *	hides all the races and re-entrancy issues from us.
	 */
1713
	err = vnet_hdr_len + ((flags&MSG_TRUNC) ? skb->len : copied);
L
Linus Torvalds 已提交
1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730

out_free:
	skb_free_datagram(sk, skb);
out:
	return err;
}

static int packet_getname_spkt(struct socket *sock, struct sockaddr *uaddr,
			       int *uaddr_len, int peer)
{
	struct net_device *dev;
	struct sock *sk	= sock->sk;

	if (peer)
		return -EOPNOTSUPP;

	uaddr->sa_family = AF_PACKET;
E
Eric Dumazet 已提交
1731 1732 1733
	rcu_read_lock();
	dev = dev_get_by_index_rcu(sock_net(sk), pkt_sk(sk)->ifindex);
	if (dev)
1734
		strncpy(uaddr->sa_data, dev->name, 14);
E
Eric Dumazet 已提交
1735
	else
L
Linus Torvalds 已提交
1736
		memset(uaddr->sa_data, 0, 14);
E
Eric Dumazet 已提交
1737
	rcu_read_unlock();
L
Linus Torvalds 已提交
1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748
	*uaddr_len = sizeof(*uaddr);

	return 0;
}

static int packet_getname(struct socket *sock, struct sockaddr *uaddr,
			  int *uaddr_len, int peer)
{
	struct net_device *dev;
	struct sock *sk = sock->sk;
	struct packet_sock *po = pkt_sk(sk);
1749
	DECLARE_SOCKADDR(struct sockaddr_ll *, sll, uaddr);
L
Linus Torvalds 已提交
1750 1751 1752 1753 1754 1755 1756

	if (peer)
		return -EOPNOTSUPP;

	sll->sll_family = AF_PACKET;
	sll->sll_ifindex = po->ifindex;
	sll->sll_protocol = po->num;
1757
	sll->sll_pkttype = 0;
E
Eric Dumazet 已提交
1758 1759
	rcu_read_lock();
	dev = dev_get_by_index_rcu(sock_net(sk), po->ifindex);
L
Linus Torvalds 已提交
1760 1761 1762 1763 1764 1765 1766 1767
	if (dev) {
		sll->sll_hatype = dev->type;
		sll->sll_halen = dev->addr_len;
		memcpy(sll->sll_addr, dev->dev_addr, dev->addr_len);
	} else {
		sll->sll_hatype = 0;	/* Bad: we have no ARPHRD_UNSPEC */
		sll->sll_halen = 0;
	}
E
Eric Dumazet 已提交
1768
	rcu_read_unlock();
1769
	*uaddr_len = offsetof(struct sockaddr_ll, sll_addr) + sll->sll_halen;
L
Linus Torvalds 已提交
1770 1771 1772 1773

	return 0;
}

1774 1775
static int packet_dev_mc(struct net_device *dev, struct packet_mclist *i,
			 int what)
L
Linus Torvalds 已提交
1776 1777 1778
{
	switch (i->type) {
	case PACKET_MR_MULTICAST:
1779 1780
		if (i->alen != dev->addr_len)
			return -EINVAL;
L
Linus Torvalds 已提交
1781
		if (what > 0)
1782
			return dev_mc_add(dev, i->addr);
L
Linus Torvalds 已提交
1783
		else
1784
			return dev_mc_del(dev, i->addr);
L
Linus Torvalds 已提交
1785 1786
		break;
	case PACKET_MR_PROMISC:
1787
		return dev_set_promiscuity(dev, what);
L
Linus Torvalds 已提交
1788 1789
		break;
	case PACKET_MR_ALLMULTI:
1790
		return dev_set_allmulti(dev, what);
L
Linus Torvalds 已提交
1791
		break;
1792
	case PACKET_MR_UNICAST:
1793 1794
		if (i->alen != dev->addr_len)
			return -EINVAL;
1795
		if (what > 0)
1796
			return dev_uc_add(dev, i->addr);
1797
		else
1798
			return dev_uc_del(dev, i->addr);
1799
		break;
E
Eric Dumazet 已提交
1800 1801
	default:
		break;
L
Linus Torvalds 已提交
1802
	}
1803
	return 0;
L
Linus Torvalds 已提交
1804 1805 1806 1807
}

static void packet_dev_mclist(struct net_device *dev, struct packet_mclist *i, int what)
{
E
Eric Dumazet 已提交
1808
	for ( ; i; i = i->next) {
L
Linus Torvalds 已提交
1809 1810 1811 1812 1813
		if (i->ifindex == dev->ifindex)
			packet_dev_mc(dev, i, what);
	}
}

1814
static int packet_mc_add(struct sock *sk, struct packet_mreq_max *mreq)
L
Linus Torvalds 已提交
1815 1816 1817 1818 1819 1820 1821 1822 1823
{
	struct packet_sock *po = pkt_sk(sk);
	struct packet_mclist *ml, *i;
	struct net_device *dev;
	int err;

	rtnl_lock();

	err = -ENODEV;
1824
	dev = __dev_get_by_index(sock_net(sk), mreq->mr_ifindex);
L
Linus Torvalds 已提交
1825 1826 1827 1828
	if (!dev)
		goto done;

	err = -EINVAL;
1829
	if (mreq->mr_alen > dev->addr_len)
L
Linus Torvalds 已提交
1830 1831 1832
		goto done;

	err = -ENOBUFS;
1833
	i = kmalloc(sizeof(*i), GFP_KERNEL);
L
Linus Torvalds 已提交
1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856
	if (i == NULL)
		goto done;

	err = 0;
	for (ml = po->mclist; ml; ml = ml->next) {
		if (ml->ifindex == mreq->mr_ifindex &&
		    ml->type == mreq->mr_type &&
		    ml->alen == mreq->mr_alen &&
		    memcmp(ml->addr, mreq->mr_address, ml->alen) == 0) {
			ml->count++;
			/* Free the new element ... */
			kfree(i);
			goto done;
		}
	}

	i->type = mreq->mr_type;
	i->ifindex = mreq->mr_ifindex;
	i->alen = mreq->mr_alen;
	memcpy(i->addr, mreq->mr_address, i->alen);
	i->count = 1;
	i->next = po->mclist;
	po->mclist = i;
1857 1858 1859 1860 1861
	err = packet_dev_mc(dev, i, 1);
	if (err) {
		po->mclist = i->next;
		kfree(i);
	}
L
Linus Torvalds 已提交
1862 1863 1864 1865 1866 1867

done:
	rtnl_unlock();
	return err;
}

1868
static int packet_mc_drop(struct sock *sk, struct packet_mreq_max *mreq)
L
Linus Torvalds 已提交
1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881
{
	struct packet_mclist *ml, **mlp;

	rtnl_lock();

	for (mlp = &pkt_sk(sk)->mclist; (ml = *mlp) != NULL; mlp = &ml->next) {
		if (ml->ifindex == mreq->mr_ifindex &&
		    ml->type == mreq->mr_type &&
		    ml->alen == mreq->mr_alen &&
		    memcmp(ml->addr, mreq->mr_address, ml->alen) == 0) {
			if (--ml->count == 0) {
				struct net_device *dev;
				*mlp = ml->next;
1882 1883
				dev = __dev_get_by_index(sock_net(sk), ml->ifindex);
				if (dev)
L
Linus Torvalds 已提交
1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907
					packet_dev_mc(dev, ml, -1);
				kfree(ml);
			}
			rtnl_unlock();
			return 0;
		}
	}
	rtnl_unlock();
	return -EADDRNOTAVAIL;
}

static void packet_flush_mclist(struct sock *sk)
{
	struct packet_sock *po = pkt_sk(sk);
	struct packet_mclist *ml;

	if (!po->mclist)
		return;

	rtnl_lock();
	while ((ml = po->mclist) != NULL) {
		struct net_device *dev;

		po->mclist = ml->next;
1908 1909
		dev = __dev_get_by_index(sock_net(sk), ml->ifindex);
		if (dev != NULL)
L
Linus Torvalds 已提交
1910 1911 1912 1913 1914 1915 1916
			packet_dev_mc(dev, ml, -1);
		kfree(ml);
	}
	rtnl_unlock();
}

static int
1917
packet_setsockopt(struct socket *sock, int level, int optname, char __user *optval, unsigned int optlen)
L
Linus Torvalds 已提交
1918 1919
{
	struct sock *sk = sock->sk;
1920
	struct packet_sock *po = pkt_sk(sk);
L
Linus Torvalds 已提交
1921 1922 1923 1924 1925
	int ret;

	if (level != SOL_PACKET)
		return -ENOPROTOOPT;

J
Johann Baudy 已提交
1926
	switch (optname) {
1927
	case PACKET_ADD_MEMBERSHIP:
L
Linus Torvalds 已提交
1928 1929
	case PACKET_DROP_MEMBERSHIP:
	{
1930 1931 1932 1933
		struct packet_mreq_max mreq;
		int len = optlen;
		memset(&mreq, 0, sizeof(mreq));
		if (len < sizeof(struct packet_mreq))
L
Linus Torvalds 已提交
1934
			return -EINVAL;
1935 1936
		if (len > sizeof(mreq))
			len = sizeof(mreq);
E
Eric Dumazet 已提交
1937
		if (copy_from_user(&mreq, optval, len))
L
Linus Torvalds 已提交
1938
			return -EFAULT;
1939 1940
		if (len < (mreq.mr_alen + offsetof(struct packet_mreq, mr_address)))
			return -EINVAL;
L
Linus Torvalds 已提交
1941 1942 1943 1944 1945 1946
		if (optname == PACKET_ADD_MEMBERSHIP)
			ret = packet_mc_add(sk, &mreq);
		else
			ret = packet_mc_drop(sk, &mreq);
		return ret;
	}
1947

L
Linus Torvalds 已提交
1948
	case PACKET_RX_RING:
J
Johann Baudy 已提交
1949
	case PACKET_TX_RING:
L
Linus Torvalds 已提交
1950 1951 1952
	{
		struct tpacket_req req;

E
Eric Dumazet 已提交
1953
		if (optlen < sizeof(req))
L
Linus Torvalds 已提交
1954
			return -EINVAL;
1955 1956
		if (pkt_sk(sk)->has_vnet_hdr)
			return -EINVAL;
E
Eric Dumazet 已提交
1957
		if (copy_from_user(&req, optval, sizeof(req)))
L
Linus Torvalds 已提交
1958
			return -EFAULT;
J
Johann Baudy 已提交
1959
		return packet_set_ring(sk, &req, 0, optname == PACKET_TX_RING);
L
Linus Torvalds 已提交
1960 1961 1962 1963 1964
	}
	case PACKET_COPY_THRESH:
	{
		int val;

E
Eric Dumazet 已提交
1965
		if (optlen != sizeof(val))
L
Linus Torvalds 已提交
1966
			return -EINVAL;
E
Eric Dumazet 已提交
1967
		if (copy_from_user(&val, optval, sizeof(val)))
L
Linus Torvalds 已提交
1968 1969 1970 1971 1972
			return -EFAULT;

		pkt_sk(sk)->copy_thresh = val;
		return 0;
	}
1973 1974 1975 1976 1977 1978
	case PACKET_VERSION:
	{
		int val;

		if (optlen != sizeof(val))
			return -EINVAL;
J
Johann Baudy 已提交
1979
		if (po->rx_ring.pg_vec || po->tx_ring.pg_vec)
1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991
			return -EBUSY;
		if (copy_from_user(&val, optval, sizeof(val)))
			return -EFAULT;
		switch (val) {
		case TPACKET_V1:
		case TPACKET_V2:
			po->tp_version = val;
			return 0;
		default:
			return -EINVAL;
		}
	}
1992 1993 1994 1995 1996 1997
	case PACKET_RESERVE:
	{
		unsigned int val;

		if (optlen != sizeof(val))
			return -EINVAL;
J
Johann Baudy 已提交
1998
		if (po->rx_ring.pg_vec || po->tx_ring.pg_vec)
1999 2000 2001 2002 2003 2004
			return -EBUSY;
		if (copy_from_user(&val, optval, sizeof(val)))
			return -EFAULT;
		po->tp_reserve = val;
		return 0;
	}
J
Johann Baudy 已提交
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017
	case PACKET_LOSS:
	{
		unsigned int val;

		if (optlen != sizeof(val))
			return -EINVAL;
		if (po->rx_ring.pg_vec || po->tx_ring.pg_vec)
			return -EBUSY;
		if (copy_from_user(&val, optval, sizeof(val)))
			return -EFAULT;
		po->tp_loss = !!val;
		return 0;
	}
2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029
	case PACKET_AUXDATA:
	{
		int val;

		if (optlen < sizeof(val))
			return -EINVAL;
		if (copy_from_user(&val, optval, sizeof(val)))
			return -EFAULT;

		po->auxdata = !!val;
		return 0;
	}
2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041
	case PACKET_ORIGDEV:
	{
		int val;

		if (optlen < sizeof(val))
			return -EINVAL;
		if (copy_from_user(&val, optval, sizeof(val)))
			return -EFAULT;

		po->origdev = !!val;
		return 0;
	}
2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057
	case PACKET_VNET_HDR:
	{
		int val;

		if (sock->type != SOCK_RAW)
			return -EINVAL;
		if (po->rx_ring.pg_vec || po->tx_ring.pg_vec)
			return -EBUSY;
		if (optlen < sizeof(val))
			return -EINVAL;
		if (copy_from_user(&val, optval, sizeof(val)))
			return -EFAULT;

		po->has_vnet_hdr = !!val;
		return 0;
	}
2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069
	case PACKET_TIMESTAMP:
	{
		int val;

		if (optlen != sizeof(val))
			return -EINVAL;
		if (copy_from_user(&val, optval, sizeof(val)))
			return -EFAULT;

		po->tp_tstamp = val;
		return 0;
	}
L
Linus Torvalds 已提交
2070 2071 2072 2073 2074 2075 2076 2077 2078
	default:
		return -ENOPROTOOPT;
	}
}

static int packet_getsockopt(struct socket *sock, int level, int optname,
			     char __user *optval, int __user *optlen)
{
	int len;
2079
	int val;
L
Linus Torvalds 已提交
2080 2081
	struct sock *sk = sock->sk;
	struct packet_sock *po = pkt_sk(sk);
2082 2083
	void *data;
	struct tpacket_stats st;
L
Linus Torvalds 已提交
2084 2085 2086 2087

	if (level != SOL_PACKET)
		return -ENOPROTOOPT;

2088 2089
	if (get_user(len, optlen))
		return -EFAULT;
L
Linus Torvalds 已提交
2090 2091 2092

	if (len < 0)
		return -EINVAL;
2093

J
Johann Baudy 已提交
2094
	switch (optname) {
L
Linus Torvalds 已提交
2095 2096 2097 2098 2099 2100 2101 2102 2103
	case PACKET_STATISTICS:
		if (len > sizeof(struct tpacket_stats))
			len = sizeof(struct tpacket_stats);
		spin_lock_bh(&sk->sk_receive_queue.lock);
		st = po->stats;
		memset(&po->stats, 0, sizeof(st));
		spin_unlock_bh(&sk->sk_receive_queue.lock);
		st.tp_packets += st.tp_drops;

2104 2105 2106 2107 2108 2109 2110
		data = &st;
		break;
	case PACKET_AUXDATA:
		if (len > sizeof(int))
			len = sizeof(int);
		val = po->auxdata;

2111 2112 2113 2114 2115 2116 2117
		data = &val;
		break;
	case PACKET_ORIGDEV:
		if (len > sizeof(int))
			len = sizeof(int);
		val = po->origdev;

2118 2119 2120 2121 2122 2123 2124
		data = &val;
		break;
	case PACKET_VNET_HDR:
		if (len > sizeof(int))
			len = sizeof(int);
		val = po->has_vnet_hdr;

2125
		data = &val;
L
Linus Torvalds 已提交
2126
		break;
2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149
	case PACKET_VERSION:
		if (len > sizeof(int))
			len = sizeof(int);
		val = po->tp_version;
		data = &val;
		break;
	case PACKET_HDRLEN:
		if (len > sizeof(int))
			len = sizeof(int);
		if (copy_from_user(&val, optval, len))
			return -EFAULT;
		switch (val) {
		case TPACKET_V1:
			val = sizeof(struct tpacket_hdr);
			break;
		case TPACKET_V2:
			val = sizeof(struct tpacket2_hdr);
			break;
		default:
			return -EINVAL;
		}
		data = &val;
		break;
2150 2151 2152 2153 2154 2155
	case PACKET_RESERVE:
		if (len > sizeof(unsigned int))
			len = sizeof(unsigned int);
		val = po->tp_reserve;
		data = &val;
		break;
J
Johann Baudy 已提交
2156 2157 2158 2159 2160 2161
	case PACKET_LOSS:
		if (len > sizeof(unsigned int))
			len = sizeof(unsigned int);
		val = po->tp_loss;
		data = &val;
		break;
2162 2163 2164 2165 2166 2167
	case PACKET_TIMESTAMP:
		if (len > sizeof(int))
			len = sizeof(int);
		val = po->tp_tstamp;
		data = &val;
		break;
L
Linus Torvalds 已提交
2168 2169 2170 2171
	default:
		return -ENOPROTOOPT;
	}

2172 2173
	if (put_user(len, optlen))
		return -EFAULT;
2174 2175
	if (copy_to_user(optval, data, len))
		return -EFAULT;
2176
	return 0;
L
Linus Torvalds 已提交
2177 2178 2179 2180 2181 2182 2183
}


static int packet_notifier(struct notifier_block *this, unsigned long msg, void *data)
{
	struct sock *sk;
	struct hlist_node *node;
2184
	struct net_device *dev = data;
2185
	struct net *net = dev_net(dev);
L
Linus Torvalds 已提交
2186

2187 2188
	rcu_read_lock();
	sk_for_each_rcu(sk, node, &net->packet.sklist) {
L
Linus Torvalds 已提交
2189 2190 2191 2192 2193 2194
		struct packet_sock *po = pkt_sk(sk);

		switch (msg) {
		case NETDEV_UNREGISTER:
			if (po->mclist)
				packet_dev_mclist(dev, po->mclist, -1);
2195 2196
			/* fallthrough */

L
Linus Torvalds 已提交
2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215
		case NETDEV_DOWN:
			if (dev->ifindex == po->ifindex) {
				spin_lock(&po->bind_lock);
				if (po->running) {
					__dev_remove_pack(&po->prot_hook);
					__sock_put(sk);
					po->running = 0;
					sk->sk_err = ENETDOWN;
					if (!sock_flag(sk, SOCK_DEAD))
						sk->sk_error_report(sk);
				}
				if (msg == NETDEV_UNREGISTER) {
					po->ifindex = -1;
					po->prot_hook.dev = NULL;
				}
				spin_unlock(&po->bind_lock);
			}
			break;
		case NETDEV_UP:
2216 2217 2218 2219 2220 2221 2222 2223
			if (dev->ifindex == po->ifindex) {
				spin_lock(&po->bind_lock);
				if (po->num && !po->running) {
					dev_add_pack(&po->prot_hook);
					sock_hold(sk);
					po->running = 1;
				}
				spin_unlock(&po->bind_lock);
L
Linus Torvalds 已提交
2224 2225 2226 2227
			}
			break;
		}
	}
2228
	rcu_read_unlock();
L
Linus Torvalds 已提交
2229 2230 2231 2232 2233 2234 2235 2236 2237
	return NOTIFY_DONE;
}


static int packet_ioctl(struct socket *sock, unsigned int cmd,
			unsigned long arg)
{
	struct sock *sk = sock->sk;

J
Johann Baudy 已提交
2238
	switch (cmd) {
E
Eric Dumazet 已提交
2239 2240 2241
	case SIOCOUTQ:
	{
		int amount = sk_wmem_alloc_get(sk);
2242

E
Eric Dumazet 已提交
2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260
		return put_user(amount, (int __user *)arg);
	}
	case SIOCINQ:
	{
		struct sk_buff *skb;
		int amount = 0;

		spin_lock_bh(&sk->sk_receive_queue.lock);
		skb = skb_peek(&sk->sk_receive_queue);
		if (skb)
			amount = skb->len;
		spin_unlock_bh(&sk->sk_receive_queue.lock);
		return put_user(amount, (int __user *)arg);
	}
	case SIOCGSTAMP:
		return sock_get_timestamp(sk, (struct timeval __user *)arg);
	case SIOCGSTAMPNS:
		return sock_get_timestampns(sk, (struct timespec __user *)arg);
2261

L
Linus Torvalds 已提交
2262
#ifdef CONFIG_INET
E
Eric Dumazet 已提交
2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277
	case SIOCADDRT:
	case SIOCDELRT:
	case SIOCDARP:
	case SIOCGARP:
	case SIOCSARP:
	case SIOCGIFADDR:
	case SIOCSIFADDR:
	case SIOCGIFBRDADDR:
	case SIOCSIFBRDADDR:
	case SIOCGIFNETMASK:
	case SIOCSIFNETMASK:
	case SIOCGIFDSTADDR:
	case SIOCSIFDSTADDR:
	case SIOCSIFFLAGS:
		return inet_dgram_ops.ioctl(sock, cmd, arg);
L
Linus Torvalds 已提交
2278 2279
#endif

E
Eric Dumazet 已提交
2280 2281
	default:
		return -ENOIOCTLCMD;
L
Linus Torvalds 已提交
2282 2283 2284 2285
	}
	return 0;
}

E
Eric Dumazet 已提交
2286
static unsigned int packet_poll(struct file *file, struct socket *sock,
L
Linus Torvalds 已提交
2287 2288 2289 2290 2291 2292 2293
				poll_table *wait)
{
	struct sock *sk = sock->sk;
	struct packet_sock *po = pkt_sk(sk);
	unsigned int mask = datagram_poll(file, sock, wait);

	spin_lock_bh(&sk->sk_receive_queue.lock);
J
Johann Baudy 已提交
2294 2295
	if (po->rx_ring.pg_vec) {
		if (!packet_previous_frame(po, &po->rx_ring, TP_STATUS_KERNEL))
L
Linus Torvalds 已提交
2296 2297 2298
			mask |= POLLIN | POLLRDNORM;
	}
	spin_unlock_bh(&sk->sk_receive_queue.lock);
J
Johann Baudy 已提交
2299 2300 2301 2302 2303 2304
	spin_lock_bh(&sk->sk_write_queue.lock);
	if (po->tx_ring.pg_vec) {
		if (packet_current_frame(po, &po->tx_ring, TP_STATUS_AVAILABLE))
			mask |= POLLOUT | POLLWRNORM;
	}
	spin_unlock_bh(&sk->sk_write_queue.lock);
L
Linus Torvalds 已提交
2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315
	return mask;
}


/* Dirty? Well, I still did not learn better way to account
 * for user mmaps.
 */

static void packet_mm_open(struct vm_area_struct *vma)
{
	struct file *file = vma->vm_file;
E
Eric Dumazet 已提交
2316
	struct socket *sock = file->private_data;
L
Linus Torvalds 已提交
2317
	struct sock *sk = sock->sk;
2318

L
Linus Torvalds 已提交
2319 2320 2321 2322 2323 2324 2325
	if (sk)
		atomic_inc(&pkt_sk(sk)->mapped);
}

static void packet_mm_close(struct vm_area_struct *vma)
{
	struct file *file = vma->vm_file;
E
Eric Dumazet 已提交
2326
	struct socket *sock = file->private_data;
L
Linus Torvalds 已提交
2327
	struct sock *sk = sock->sk;
2328

L
Linus Torvalds 已提交
2329 2330 2331 2332
	if (sk)
		atomic_dec(&pkt_sk(sk)->mapped);
}

2333
static const struct vm_operations_struct packet_mmap_ops = {
E
Eric Dumazet 已提交
2334 2335
	.open	=	packet_mm_open,
	.close	=	packet_mm_close,
L
Linus Torvalds 已提交
2336 2337
};

2338 2339
static void free_pg_vec(struct pgv *pg_vec, unsigned int order,
			unsigned int len)
L
Linus Torvalds 已提交
2340 2341 2342
{
	int i;

2343
	for (i = 0; i < len; i++) {
2344
		if (likely(pg_vec[i].buffer)) {
C
Changli Gao 已提交
2345
			if (is_vmalloc_addr(pg_vec[i].buffer))
2346 2347 2348 2349 2350 2351
				vfree(pg_vec[i].buffer);
			else
				free_pages((unsigned long)pg_vec[i].buffer,
					   order);
			pg_vec[i].buffer = NULL;
		}
L
Linus Torvalds 已提交
2352 2353 2354 2355
	}
	kfree(pg_vec);
}

C
Changli Gao 已提交
2356
static inline char *alloc_one_pg_vec_page(unsigned long order)
2357
{
2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369
	char *buffer = NULL;
	gfp_t gfp_flags = GFP_KERNEL | __GFP_COMP |
			  __GFP_ZERO | __GFP_NOWARN | __GFP_NORETRY;

	buffer = (char *) __get_free_pages(gfp_flags, order);

	if (buffer)
		return buffer;

	/*
	 * __get_free_pages failed, fall back to vmalloc
	 */
E
Eric Dumazet 已提交
2370
	buffer = vzalloc((1 << order) * PAGE_SIZE);
2371

2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386
	if (buffer)
		return buffer;

	/*
	 * vmalloc failed, lets dig into swap here
	 */
	gfp_flags &= ~__GFP_NORETRY;
	buffer = (char *)__get_free_pages(gfp_flags, order);
	if (buffer)
		return buffer;

	/*
	 * complete and utter failure
	 */
	return NULL;
2387 2388
}

2389
static struct pgv *alloc_pg_vec(struct tpacket_req *req, int order)
2390 2391
{
	unsigned int block_nr = req->tp_block_nr;
2392
	struct pgv *pg_vec;
2393 2394
	int i;

2395
	pg_vec = kcalloc(block_nr, sizeof(struct pgv), GFP_KERNEL);
2396 2397 2398 2399
	if (unlikely(!pg_vec))
		goto out;

	for (i = 0; i < block_nr; i++) {
C
Changli Gao 已提交
2400
		pg_vec[i].buffer = alloc_one_pg_vec_page(order);
2401
		if (unlikely(!pg_vec[i].buffer))
2402 2403 2404 2405 2406 2407 2408 2409
			goto out_free_pgvec;
	}

out:
	return pg_vec;

out_free_pgvec:
	free_pg_vec(pg_vec, order, block_nr);
2410
	kfree(pg_vec);
2411 2412 2413
	pg_vec = NULL;
	goto out;
}
L
Linus Torvalds 已提交
2414

J
Johann Baudy 已提交
2415 2416
static int packet_set_ring(struct sock *sk, struct tpacket_req *req,
		int closing, int tx_ring)
L
Linus Torvalds 已提交
2417
{
2418
	struct pgv *pg_vec = NULL;
L
Linus Torvalds 已提交
2419
	struct packet_sock *po = pkt_sk(sk);
A
Al Viro 已提交
2420
	int was_running, order = 0;
J
Johann Baudy 已提交
2421 2422
	struct packet_ring_buffer *rb;
	struct sk_buff_head *rb_queue;
A
Al Viro 已提交
2423
	__be16 num;
J
Johann Baudy 已提交
2424
	int err;
2425

J
Johann Baudy 已提交
2426 2427
	rb = tx_ring ? &po->tx_ring : &po->rx_ring;
	rb_queue = tx_ring ? &sk->sk_write_queue : &sk->sk_receive_queue;
L
Linus Torvalds 已提交
2428

J
Johann Baudy 已提交
2429 2430 2431 2432 2433 2434 2435
	err = -EBUSY;
	if (!closing) {
		if (atomic_read(&po->mapped))
			goto out;
		if (atomic_read(&rb->pending))
			goto out;
	}
L
Linus Torvalds 已提交
2436

J
Johann Baudy 已提交
2437 2438 2439 2440 2441
	if (req->tp_block_nr) {
		/* Sanity tests and some calculations */
		err = -EBUSY;
		if (unlikely(rb->pg_vec))
			goto out;
L
Linus Torvalds 已提交
2442

2443 2444 2445 2446 2447 2448 2449 2450 2451
		switch (po->tp_version) {
		case TPACKET_V1:
			po->tp_hdrlen = TPACKET_HDRLEN;
			break;
		case TPACKET_V2:
			po->tp_hdrlen = TPACKET2_HDRLEN;
			break;
		}

J
Johann Baudy 已提交
2452
		err = -EINVAL;
2453
		if (unlikely((int)req->tp_block_size <= 0))
J
Johann Baudy 已提交
2454
			goto out;
2455
		if (unlikely(req->tp_block_size & (PAGE_SIZE - 1)))
J
Johann Baudy 已提交
2456
			goto out;
2457
		if (unlikely(req->tp_frame_size < po->tp_hdrlen +
J
Johann Baudy 已提交
2458 2459
					po->tp_reserve))
			goto out;
2460
		if (unlikely(req->tp_frame_size & (TPACKET_ALIGNMENT - 1)))
J
Johann Baudy 已提交
2461
			goto out;
L
Linus Torvalds 已提交
2462

J
Johann Baudy 已提交
2463 2464 2465 2466 2467 2468
		rb->frames_per_block = req->tp_block_size/req->tp_frame_size;
		if (unlikely(rb->frames_per_block <= 0))
			goto out;
		if (unlikely((rb->frames_per_block * req->tp_block_nr) !=
					req->tp_frame_nr))
			goto out;
L
Linus Torvalds 已提交
2469 2470

		err = -ENOMEM;
2471 2472 2473
		order = get_order(req->tp_block_size);
		pg_vec = alloc_pg_vec(req, order);
		if (unlikely(!pg_vec))
L
Linus Torvalds 已提交
2474
			goto out;
J
Johann Baudy 已提交
2475 2476 2477 2478
	}
	/* Done */
	else {
		err = -EINVAL;
2479
		if (unlikely(req->tp_frame_nr))
J
Johann Baudy 已提交
2480
			goto out;
L
Linus Torvalds 已提交
2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495
	}

	lock_sock(sk);

	/* Detach socket from network */
	spin_lock(&po->bind_lock);
	was_running = po->running;
	num = po->num;
	if (was_running) {
		__dev_remove_pack(&po->prot_hook);
		po->num = 0;
		po->running = 0;
		__sock_put(sk);
	}
	spin_unlock(&po->bind_lock);
2496

L
Linus Torvalds 已提交
2497 2498 2499
	synchronize_net();

	err = -EBUSY;
2500
	mutex_lock(&po->pg_vec_lock);
L
Linus Torvalds 已提交
2501 2502 2503
	if (closing || atomic_read(&po->mapped) == 0) {
		err = 0;
#define XC(a, b) ({ __typeof__ ((a)) __t; __t = (a); (a) = (b); __t; })
J
Johann Baudy 已提交
2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517
		spin_lock_bh(&rb_queue->lock);
		pg_vec = XC(rb->pg_vec, pg_vec);
		rb->frame_max = (req->tp_frame_nr - 1);
		rb->head = 0;
		rb->frame_size = req->tp_frame_size;
		spin_unlock_bh(&rb_queue->lock);

		order = XC(rb->pg_vec_order, order);
		req->tp_block_nr = XC(rb->pg_vec_len, req->tp_block_nr);

		rb->pg_vec_pages = req->tp_block_size/PAGE_SIZE;
		po->prot_hook.func = (po->rx_ring.pg_vec) ?
						tpacket_rcv : packet_rcv;
		skb_queue_purge(rb_queue);
L
Linus Torvalds 已提交
2518 2519
#undef XC
		if (atomic_read(&po->mapped))
E
Eric Dumazet 已提交
2520 2521
			pr_err("packet_mmap: vma is busy: %d\n",
			       atomic_read(&po->mapped));
L
Linus Torvalds 已提交
2522
	}
2523
	mutex_unlock(&po->pg_vec_lock);
L
Linus Torvalds 已提交
2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541

	spin_lock(&po->bind_lock);
	if (was_running && !po->running) {
		sock_hold(sk);
		po->running = 1;
		po->num = num;
		dev_add_pack(&po->prot_hook);
	}
	spin_unlock(&po->bind_lock);

	release_sock(sk);

	if (pg_vec)
		free_pg_vec(pg_vec, order, req->tp_block_nr);
out:
	return err;
}

J
Johann Baudy 已提交
2542 2543
static int packet_mmap(struct file *file, struct socket *sock,
		struct vm_area_struct *vma)
L
Linus Torvalds 已提交
2544 2545 2546
{
	struct sock *sk = sock->sk;
	struct packet_sock *po = pkt_sk(sk);
J
Johann Baudy 已提交
2547 2548
	unsigned long size, expected_size;
	struct packet_ring_buffer *rb;
L
Linus Torvalds 已提交
2549 2550 2551 2552 2553 2554 2555
	unsigned long start;
	int err = -EINVAL;
	int i;

	if (vma->vm_pgoff)
		return -EINVAL;

2556
	mutex_lock(&po->pg_vec_lock);
J
Johann Baudy 已提交
2557 2558 2559 2560 2561 2562 2563 2564 2565 2566 2567

	expected_size = 0;
	for (rb = &po->rx_ring; rb <= &po->tx_ring; rb++) {
		if (rb->pg_vec) {
			expected_size += rb->pg_vec_len
						* rb->pg_vec_pages
						* PAGE_SIZE;
		}
	}

	if (expected_size == 0)
L
Linus Torvalds 已提交
2568
		goto out;
J
Johann Baudy 已提交
2569 2570 2571

	size = vma->vm_end - vma->vm_start;
	if (size != expected_size)
L
Linus Torvalds 已提交
2572 2573 2574
		goto out;

	start = vma->vm_start;
J
Johann Baudy 已提交
2575 2576 2577 2578 2579
	for (rb = &po->rx_ring; rb <= &po->tx_ring; rb++) {
		if (rb->pg_vec == NULL)
			continue;

		for (i = 0; i < rb->pg_vec_len; i++) {
2580 2581
			struct page *page;
			void *kaddr = rb->pg_vec[i].buffer;
J
Johann Baudy 已提交
2582 2583
			int pg_num;

C
Changli Gao 已提交
2584 2585
			for (pg_num = 0; pg_num < rb->pg_vec_pages; pg_num++) {
				page = pgv_to_page(kaddr);
J
Johann Baudy 已提交
2586 2587 2588 2589
				err = vm_insert_page(vma, start, page);
				if (unlikely(err))
					goto out;
				start += PAGE_SIZE;
2590
				kaddr += PAGE_SIZE;
J
Johann Baudy 已提交
2591
			}
2592
		}
L
Linus Torvalds 已提交
2593
	}
J
Johann Baudy 已提交
2594

2595
	atomic_inc(&po->mapped);
L
Linus Torvalds 已提交
2596 2597 2598 2599
	vma->vm_ops = &packet_mmap_ops;
	err = 0;

out:
2600
	mutex_unlock(&po->pg_vec_lock);
L
Linus Torvalds 已提交
2601 2602 2603
	return err;
}

2604
static const struct proto_ops packet_ops_spkt = {
L
Linus Torvalds 已提交
2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624
	.family =	PF_PACKET,
	.owner =	THIS_MODULE,
	.release =	packet_release,
	.bind =		packet_bind_spkt,
	.connect =	sock_no_connect,
	.socketpair =	sock_no_socketpair,
	.accept =	sock_no_accept,
	.getname =	packet_getname_spkt,
	.poll =		datagram_poll,
	.ioctl =	packet_ioctl,
	.listen =	sock_no_listen,
	.shutdown =	sock_no_shutdown,
	.setsockopt =	sock_no_setsockopt,
	.getsockopt =	sock_no_getsockopt,
	.sendmsg =	packet_sendmsg_spkt,
	.recvmsg =	packet_recvmsg,
	.mmap =		sock_no_mmap,
	.sendpage =	sock_no_sendpage,
};

2625
static const struct proto_ops packet_ops = {
L
Linus Torvalds 已提交
2626 2627 2628 2629 2630 2631 2632
	.family =	PF_PACKET,
	.owner =	THIS_MODULE,
	.release =	packet_release,
	.bind =		packet_bind,
	.connect =	sock_no_connect,
	.socketpair =	sock_no_socketpair,
	.accept =	sock_no_accept,
2633
	.getname =	packet_getname,
L
Linus Torvalds 已提交
2634 2635 2636 2637 2638 2639 2640 2641 2642 2643 2644 2645
	.poll =		packet_poll,
	.ioctl =	packet_ioctl,
	.listen =	sock_no_listen,
	.shutdown =	sock_no_shutdown,
	.setsockopt =	packet_setsockopt,
	.getsockopt =	packet_getsockopt,
	.sendmsg =	packet_sendmsg,
	.recvmsg =	packet_recvmsg,
	.mmap =		packet_mmap,
	.sendpage =	sock_no_sendpage,
};

2646
static const struct net_proto_family packet_family_ops = {
L
Linus Torvalds 已提交
2647 2648 2649 2650 2651 2652
	.family =	PF_PACKET,
	.create =	packet_create,
	.owner	=	THIS_MODULE,
};

static struct notifier_block packet_netdev_notifier = {
E
Eric Dumazet 已提交
2653
	.notifier_call =	packet_notifier,
L
Linus Torvalds 已提交
2654 2655 2656 2657 2658
};

#ifdef CONFIG_PROC_FS

static void *packet_seq_start(struct seq_file *seq, loff_t *pos)
2659
	__acquires(RCU)
L
Linus Torvalds 已提交
2660
{
2661
	struct net *net = seq_file_net(seq);
2662 2663 2664

	rcu_read_lock();
	return seq_hlist_start_head_rcu(&net->packet.sklist, *pos);
L
Linus Torvalds 已提交
2665 2666 2667 2668
}

static void *packet_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
2669
	struct net *net = seq_file_net(seq);
2670
	return seq_hlist_next_rcu(v, &net->packet.sklist, pos);
L
Linus Torvalds 已提交
2671 2672 2673
}

static void packet_seq_stop(struct seq_file *seq, void *v)
2674
	__releases(RCU)
L
Linus Torvalds 已提交
2675
{
2676
	rcu_read_unlock();
L
Linus Torvalds 已提交
2677 2678
}

2679
static int packet_seq_show(struct seq_file *seq, void *v)
L
Linus Torvalds 已提交
2680 2681 2682 2683
{
	if (v == SEQ_START_TOKEN)
		seq_puts(seq, "sk       RefCnt Type Proto  Iface R Rmem   User   Inode\n");
	else {
2684
		struct sock *s = sk_entry(v);
L
Linus Torvalds 已提交
2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696
		const struct packet_sock *po = pkt_sk(s);

		seq_printf(seq,
			   "%p %-6d %-4d %04x   %-5d %1d %-6u %-6u %-6lu\n",
			   s,
			   atomic_read(&s->sk_refcnt),
			   s->sk_type,
			   ntohs(po->num),
			   po->ifindex,
			   po->running,
			   atomic_read(&s->sk_rmem_alloc),
			   sock_i_uid(s),
E
Eric Dumazet 已提交
2697
			   sock_i_ino(s));
L
Linus Torvalds 已提交
2698 2699 2700 2701 2702
	}

	return 0;
}

2703
static const struct seq_operations packet_seq_ops = {
L
Linus Torvalds 已提交
2704 2705 2706 2707 2708 2709 2710 2711
	.start	= packet_seq_start,
	.next	= packet_seq_next,
	.stop	= packet_seq_stop,
	.show	= packet_seq_show,
};

static int packet_seq_open(struct inode *inode, struct file *file)
{
2712 2713
	return seq_open_net(inode, file, &packet_seq_ops,
			    sizeof(struct seq_net_private));
L
Linus Torvalds 已提交
2714 2715
}

2716
static const struct file_operations packet_seq_fops = {
L
Linus Torvalds 已提交
2717 2718 2719 2720
	.owner		= THIS_MODULE,
	.open		= packet_seq_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
2721
	.release	= seq_release_net,
L
Linus Torvalds 已提交
2722 2723 2724 2725
};

#endif

2726
static int __net_init packet_net_init(struct net *net)
2727
{
2728
	spin_lock_init(&net->packet.sklist_lock);
2729
	INIT_HLIST_HEAD(&net->packet.sklist);
2730 2731 2732 2733 2734 2735 2736

	if (!proc_net_fops_create(net, "packet", 0, &packet_seq_fops))
		return -ENOMEM;

	return 0;
}

2737
static void __net_exit packet_net_exit(struct net *net)
2738 2739 2740 2741 2742 2743 2744 2745 2746 2747
{
	proc_net_remove(net, "packet");
}

static struct pernet_operations packet_net_ops = {
	.init = packet_net_init,
	.exit = packet_net_exit,
};


L
Linus Torvalds 已提交
2748 2749 2750
static void __exit packet_exit(void)
{
	unregister_netdevice_notifier(&packet_netdev_notifier);
2751
	unregister_pernet_subsys(&packet_net_ops);
L
Linus Torvalds 已提交
2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763
	sock_unregister(PF_PACKET);
	proto_unregister(&packet_proto);
}

static int __init packet_init(void)
{
	int rc = proto_register(&packet_proto, 0);

	if (rc != 0)
		goto out;

	sock_register(&packet_family_ops);
2764
	register_pernet_subsys(&packet_net_ops);
L
Linus Torvalds 已提交
2765 2766 2767 2768 2769 2770 2771 2772 2773
	register_netdevice_notifier(&packet_netdev_notifier);
out:
	return rc;
}

module_init(packet_init);
module_exit(packet_exit);
MODULE_LICENSE("GPL");
MODULE_ALIAS_NETPROTO(PF_PACKET);