af_packet.c 54.8 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 <net/net_namespace.h>
L
Linus Torvalds 已提交
64 65 66 67 68 69 70 71 72 73
#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
74
#include <asm/cacheflush.h>
L
Linus Torvalds 已提交
75 76 77 78 79 80
#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>
81
#include <linux/mutex.h>
E
Eric Dumazet 已提交
82
#include <linux/if_vlan.h>
L
Linus Torvalds 已提交
83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102

#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
103 104
   mac_header -> ll header
   data       -> data
L
Linus Torvalds 已提交
105 106

Outgoing, dev->hard_header!=NULL
107 108
   mac_header -> ll header
   data       -> ll header
L
Linus Torvalds 已提交
109 110

Incoming, dev->hard_header==NULL
111 112
   mac_header -> UNKNOWN position. It is very likely, that it points to ll
		 header.  PPP makes it, that is wrong, because introduce
113
		 assymetry between rx and tx paths.
114
   data       -> data
L
Linus Torvalds 已提交
115 116

Outgoing, dev->hard_header==NULL
117 118
   mac_header -> data. ll header is still not built!
   data       -> data
L
Linus Torvalds 已提交
119 120 121 122 123 124 125 126 127

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


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

dev->hard_header != NULL
128 129
   mac_header -> ll header
   data       -> ll header
L
Linus Torvalds 已提交
130 131

dev->hard_header == NULL (ll header is added by device, we cannot control it)
132 133
   mac_header -> data
   data       -> data
L
Linus Torvalds 已提交
134 135 136 137 138 139 140

   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
141
struct packet_mclist {
L
Linus Torvalds 已提交
142 143 144 145 146
	struct packet_mclist	*next;
	int			ifindex;
	int			count;
	unsigned short		type;
	unsigned short		alen;
147 148 149 150 151
	unsigned char		addr[MAX_ADDR_LEN];
};
/* identical to struct packet_mreq except it has
 * a longer address field.
 */
E
Eric Dumazet 已提交
152
struct packet_mreq_max {
153 154 155 156
	int		mr_ifindex;
	unsigned short	mr_type;
	unsigned short	mr_alen;
	unsigned char	mr_address[MAX_ADDR_LEN];
L
Linus Torvalds 已提交
157
};
158

L
Linus Torvalds 已提交
159
#ifdef CONFIG_PACKET_MMAP
J
Johann Baudy 已提交
160 161 162 163
static int packet_set_ring(struct sock *sk, struct tpacket_req *req,
		int closing, int tx_ring);

struct packet_ring_buffer {
E
Eric Dumazet 已提交
164
	char			**pg_vec;
J
Johann Baudy 已提交
165 166 167 168 169 170 171 172 173 174 175 176 177 178
	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
179 180 181 182 183 184 185 186 187
#endif

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;
#ifdef CONFIG_PACKET_MMAP
J
Johann Baudy 已提交
188 189
	struct packet_ring_buffer	rx_ring;
	struct packet_ring_buffer	tx_ring;
L
Linus Torvalds 已提交
190 191 192
	int			copy_thresh;
#endif
	spinlock_t		bind_lock;
193
	struct mutex		pg_vec_lock;
194
	unsigned int		running:1,	/* prot_hook is attached*/
195 196
				auxdata:1,
				origdev:1;
L
Linus Torvalds 已提交
197
	int			ifindex;	/* bound device		*/
A
Al Viro 已提交
198
	__be16			num;
L
Linus Torvalds 已提交
199 200 201
	struct packet_mclist	*mclist;
#ifdef CONFIG_PACKET_MMAP
	atomic_t		mapped;
202 203
	enum tpacket_versions	tp_version;
	unsigned int		tp_hdrlen;
204
	unsigned int		tp_reserve;
J
Johann Baudy 已提交
205
	unsigned int		tp_loss:1;
L
Linus Torvalds 已提交
206
#endif
207
	struct packet_type	prot_hook ____cacheline_aligned_in_smp;
L
Linus Torvalds 已提交
208 209
};

210 211 212 213 214 215 216 217 218
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))
219

L
Linus Torvalds 已提交
220 221
#ifdef CONFIG_PACKET_MMAP

J
Johann Baudy 已提交
222
static void __packet_set_status(struct packet_sock *po, void *frame, int status)
L
Linus Torvalds 已提交
223
{
224 225 226 227 228
	union {
		struct tpacket_hdr *h1;
		struct tpacket2_hdr *h2;
		void *raw;
	} h;
L
Linus Torvalds 已提交
229

J
Johann Baudy 已提交
230
	h.raw = frame;
231 232
	switch (po->tp_version) {
	case TPACKET_V1:
J
Johann Baudy 已提交
233 234
		h.h1->tp_status = status;
		flush_dcache_page(virt_to_page(&h.h1->tp_status));
235 236
		break;
	case TPACKET_V2:
J
Johann Baudy 已提交
237 238
		h.h2->tp_status = status;
		flush_dcache_page(virt_to_page(&h.h2->tp_status));
239
		break;
J
Johann Baudy 已提交
240
	default:
E
Eric Dumazet 已提交
241
		pr_err("TPACKET version not supported\n");
J
Johann Baudy 已提交
242
		BUG();
243
	}
J
Johann Baudy 已提交
244 245

	smp_wmb();
246 247
}

J
Johann Baudy 已提交
248
static int __packet_get_status(struct packet_sock *po, void *frame)
249 250 251 252 253 254 255
{
	union {
		struct tpacket_hdr *h1;
		struct tpacket2_hdr *h2;
		void *raw;
	} h;

J
Johann Baudy 已提交
256 257
	smp_rmb();

258 259 260
	h.raw = frame;
	switch (po->tp_version) {
	case TPACKET_V1:
J
Johann Baudy 已提交
261 262
		flush_dcache_page(virt_to_page(&h.h1->tp_status));
		return h.h1->tp_status;
263
	case TPACKET_V2:
J
Johann Baudy 已提交
264 265 266
		flush_dcache_page(virt_to_page(&h.h2->tp_status));
		return h.h2->tp_status;
	default:
E
Eric Dumazet 已提交
267
		pr_err("TPACKET version not supported\n");
J
Johann Baudy 已提交
268 269
		BUG();
		return 0;
270
	}
L
Linus Torvalds 已提交
271
}
J
Johann Baudy 已提交
272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315

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;

	h.raw = rb->pg_vec[pg_vec_pos] + (frame_offset * rb->frame_size);

	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
316 317 318 319 320 321 322 323 324
#endif

static inline struct packet_sock *pkt_sk(struct sock *sk)
{
	return (struct packet_sock *)sk;
}

static void packet_sock_destruct(struct sock *sk)
{
325 326
	WARN_ON(atomic_read(&sk->sk_rmem_alloc));
	WARN_ON(atomic_read(&sk->sk_wmem_alloc));
L
Linus Torvalds 已提交
327 328

	if (!sock_flag(sk, SOCK_DEAD)) {
E
Eric Dumazet 已提交
329
		pr_err("Attempt to release alive packet socket: %p\n", sk);
L
Linus Torvalds 已提交
330 331 332
		return;
	}

333
	sk_refcnt_debug_dec(sk);
L
Linus Torvalds 已提交
334 335 336
}


337
static const struct proto_ops packet_ops;
L
Linus Torvalds 已提交
338

339
static const struct proto_ops packet_ops_spkt;
L
Linus Torvalds 已提交
340

E
Eric Dumazet 已提交
341 342
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
343 344 345 346 347 348 349 350 351 352
{
	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;
353

L
Linus Torvalds 已提交
354 355 356 357 358 359 360
	/*
	 *	Yank back the headers [hope the device set this
	 *	right or kerboom...]
	 *
	 *	Incoming packets have ll header pulled,
	 *	push it back.
	 *
361
	 *	For outgoing ones skb->data == skb_mac_header(skb)
L
Linus Torvalds 已提交
362 363 364 365 366 367
	 *	so that this procedure is noop.
	 */

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

O
Octavian Purdila 已提交
368
	if (!net_eq(dev_net(dev), sock_net(sk)))
369 370
		goto out;

E
Eric Dumazet 已提交
371 372
	skb = skb_share_check(skb, GFP_ATOMIC);
	if (skb == NULL)
L
Linus Torvalds 已提交
373 374 375
		goto oom;

	/* drop any routing info */
E
Eric Dumazet 已提交
376
	skb_dst_drop(skb);
L
Linus Torvalds 已提交
377

378 379 380
	/* drop conntrack reference */
	nf_reset(skb);

381
	spkt = &PACKET_SKB_CB(skb)->sa.pkt;
L
Linus Torvalds 已提交
382

383
	skb_push(skb, skb->data - skb_mac_header(skb));
L
Linus Torvalds 已提交
384 385 386 387 388 389 390 391 392 393 394 395 396 397

	/*
	 *	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
398
	if (sock_queue_rcv_skb(sk, skb) == 0)
L
Linus Torvalds 已提交
399 400 401 402 403 404 405 406 407 408 409 410 411
		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
 */
412

L
Linus Torvalds 已提交
413 414 415 416
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
417
	struct sockaddr_pkt *saddr = (struct sockaddr_pkt *)msg->msg_name;
L
Linus Torvalds 已提交
418 419
	struct sk_buff *skb;
	struct net_device *dev;
E
Eric Dumazet 已提交
420
	__be16 proto = 0;
L
Linus Torvalds 已提交
421
	int err;
422

L
Linus Torvalds 已提交
423
	/*
424
	 *	Get and verify the address.
L
Linus Torvalds 已提交
425 426
	 */

E
Eric Dumazet 已提交
427
	if (saddr) {
L
Linus Torvalds 已提交
428
		if (msg->msg_namelen < sizeof(struct sockaddr))
E
Eric Dumazet 已提交
429 430 431 432 433
			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
434 435

	/*
436
	 *	Find the device first to size check it
L
Linus Torvalds 已提交
437 438 439
	 */

	saddr->spkt_device[13] = 0;
E
Eric Dumazet 已提交
440 441
	rcu_read_lock();
	dev = dev_get_by_name_rcu(sock_net(sk), saddr->spkt_device);
L
Linus Torvalds 已提交
442 443 444
	err = -ENODEV;
	if (dev == NULL)
		goto out_unlock;
445

446 447 448 449
	err = -ENETDOWN;
	if (!(dev->flags & IFF_UP))
		goto out_unlock;

L
Linus Torvalds 已提交
450
	/*
E
Eric Dumazet 已提交
451 452
	 * 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
453
	 */
454

L
Linus Torvalds 已提交
455
	err = -EMSGSIZE;
456
	if (len > dev->mtu + dev->hard_header_len)
L
Linus Torvalds 已提交
457 458 459 460 461 462
		goto out_unlock;

	err = -ENOBUFS;
	skb = sock_wmalloc(sk, len + LL_RESERVED_SPACE(dev), 0, GFP_KERNEL);

	/*
E
Eric Dumazet 已提交
463 464 465
	 * If the write buffer is full, then tough. At this level the user
	 * gets to deal with the problem - do your own algorithmic backoffs.
	 * That's far more flexible.
L
Linus Torvalds 已提交
466
	 */
467 468

	if (skb == NULL)
L
Linus Torvalds 已提交
469 470 471
		goto out_unlock;

	/*
472
	 *	Fill it in
L
Linus Torvalds 已提交
473
	 */
474

L
Linus Torvalds 已提交
475 476 477 478 479
	/* 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, LL_RESERVED_SPACE(dev));
480
	skb_reset_network_header(skb);
L
Linus Torvalds 已提交
481 482

	/* Try to align data part correctly */
483
	if (dev->header_ops) {
L
Linus Torvalds 已提交
484 485 486
		skb->data -= dev->hard_header_len;
		skb->tail -= dev->hard_header_len;
		if (len < dev->hard_header_len)
487
			skb_reset_network_header(skb);
L
Linus Torvalds 已提交
488 489 490
	}

	/* Returns -EFAULT on error */
E
Eric Dumazet 已提交
491
	err = memcpy_fromiovec(skb_put(skb, len), msg->msg_iov, len);
L
Linus Torvalds 已提交
492 493 494
	skb->protocol = proto;
	skb->dev = dev;
	skb->priority = sk->sk_priority;
495
	skb->mark = sk->sk_mark;
L
Linus Torvalds 已提交
496 497 498 499 500 501 502 503
	if (err)
		goto out_free;

	/*
	 *	Now send it
	 */

	dev_queue_xmit(skb);
E
Eric Dumazet 已提交
504
	rcu_read_unlock();
E
Eric Dumazet 已提交
505
	return len;
L
Linus Torvalds 已提交
506 507 508 509

out_free:
	kfree_skb(skb);
out_unlock:
E
Eric Dumazet 已提交
510
	rcu_read_unlock();
L
Linus Torvalds 已提交
511 512 513
	return err;
}

D
David S. Miller 已提交
514 515
static inline unsigned int run_filter(struct sk_buff *skb, struct sock *sk,
				      unsigned int res)
L
Linus Torvalds 已提交
516 517
{
	struct sk_filter *filter;
518 519 520

	rcu_read_lock_bh();
	filter = rcu_dereference(sk->sk_filter);
D
David S. Miller 已提交
521 522
	if (filter != NULL)
		res = sk_run_filter(skb, filter->insns, filter->len);
523
	rcu_read_unlock_bh();
L
Linus Torvalds 已提交
524

D
David S. Miller 已提交
525
	return res;
L
Linus Torvalds 已提交
526 527 528 529 530 531 532 533 534 535 536 537 538 539
}

/*
   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.
 */

E
Eric Dumazet 已提交
540 541
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
542 543 544 545
{
	struct sock *sk;
	struct sockaddr_ll *sll;
	struct packet_sock *po;
E
Eric Dumazet 已提交
546
	u8 *skb_head = skb->data;
L
Linus Torvalds 已提交
547
	int skb_len = skb->len;
D
David S. Miller 已提交
548
	unsigned int snaplen, res;
L
Linus Torvalds 已提交
549 550 551 552 553 554 555

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

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

O
Octavian Purdila 已提交
556
	if (!net_eq(dev_net(dev), sock_net(sk)))
557 558
		goto drop;

L
Linus Torvalds 已提交
559 560
	skb->dev = dev;

561
	if (dev->header_ops) {
L
Linus Torvalds 已提交
562 563 564 565 566 567 568 569
		/* The device has an explicit notion of ll header,
		   exported to higher levels.

		   Otherwise, the device hides datails of it frame
		   structure, so that corresponding packet head
		   never delivered to user.
		 */
		if (sk->sk_type != SOCK_DGRAM)
570
			skb_push(skb, skb->data - skb_mac_header(skb));
L
Linus Torvalds 已提交
571 572
		else if (skb->pkt_type == PACKET_OUTGOING) {
			/* Special case: outgoing packets have ll header at head */
573
			skb_pull(skb, skb_network_offset(skb));
L
Linus Torvalds 已提交
574 575 576 577 578
		}
	}

	snaplen = skb->len;

D
David S. Miller 已提交
579 580
	res = run_filter(skb, sk, snaplen);
	if (!res)
581
		goto drop_n_restore;
D
David S. Miller 已提交
582 583
	if (snaplen > res)
		snaplen = res;
L
Linus Torvalds 已提交
584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601

	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;
	}

602 603 604 605
	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
606 607 608 609
	sll->sll_family = AF_PACKET;
	sll->sll_hatype = dev->type;
	sll->sll_protocol = skb->protocol;
	sll->sll_pkttype = skb->pkt_type;
610
	if (unlikely(po->origdev))
611 612 613
		sll->sll_ifindex = orig_dev->ifindex;
	else
		sll->sll_ifindex = dev->ifindex;
L
Linus Torvalds 已提交
614

S
Stephen Hemminger 已提交
615
	sll->sll_halen = dev_parse_header(skb, sll->sll_addr);
L
Linus Torvalds 已提交
616

617
	PACKET_SKB_CB(skb)->origlen = skb->len;
618

L
Linus Torvalds 已提交
619 620 621 622 623
	if (pskb_trim(skb, snaplen))
		goto drop_n_acct;

	skb_set_owner_r(skb, sk);
	skb->dev = NULL;
E
Eric Dumazet 已提交
624
	skb_dst_drop(skb);
L
Linus Torvalds 已提交
625

626 627 628
	/* drop conntrack reference */
	nf_reset(skb);

L
Linus Torvalds 已提交
629 630
	spin_lock(&sk->sk_receive_queue.lock);
	po->stats.tp_packets++;
631
	skb->dropcount = atomic_read(&sk->sk_drops);
L
Linus Torvalds 已提交
632 633 634 635 636 637
	__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:
638
	po->stats.tp_drops = atomic_inc_return(&sk->sk_drops);
L
Linus Torvalds 已提交
639 640 641 642 643 644 645

drop_n_restore:
	if (skb_head != skb->data && skb_shared(skb)) {
		skb->data = skb_head;
		skb->len = skb_len;
	}
drop:
646
	consume_skb(skb);
L
Linus Torvalds 已提交
647 648 649 650
	return 0;
}

#ifdef CONFIG_PACKET_MMAP
E
Eric Dumazet 已提交
651 652
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
653 654 655 656
{
	struct sock *sk;
	struct packet_sock *po;
	struct sockaddr_ll *sll;
657 658 659 660 661
	union {
		struct tpacket_hdr *h1;
		struct tpacket2_hdr *h2;
		void *raw;
	} h;
E
Eric Dumazet 已提交
662
	u8 *skb_head = skb->data;
L
Linus Torvalds 已提交
663
	int skb_len = skb->len;
D
David S. Miller 已提交
664
	unsigned int snaplen, res;
L
Linus Torvalds 已提交
665
	unsigned long status = TP_STATUS_LOSING|TP_STATUS_USER;
666
	unsigned short macoff, netoff, hdrlen;
L
Linus Torvalds 已提交
667
	struct sk_buff *copy_skb = NULL;
668
	struct timeval tv;
669
	struct timespec ts;
L
Linus Torvalds 已提交
670 671 672 673 674 675 676

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

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

O
Octavian Purdila 已提交
677
	if (!net_eq(dev_net(dev), sock_net(sk)))
678 679
		goto drop;

680
	if (dev->header_ops) {
L
Linus Torvalds 已提交
681
		if (sk->sk_type != SOCK_DGRAM)
682
			skb_push(skb, skb->data - skb_mac_header(skb));
L
Linus Torvalds 已提交
683 684
		else if (skb->pkt_type == PACKET_OUTGOING) {
			/* Special case: outgoing packets have ll header at head */
685
			skb_pull(skb, skb_network_offset(skb));
L
Linus Torvalds 已提交
686 687 688
		}
	}

689 690 691
	if (skb->ip_summed == CHECKSUM_PARTIAL)
		status |= TP_STATUS_CSUMNOTREADY;

L
Linus Torvalds 已提交
692 693
	snaplen = skb->len;

D
David S. Miller 已提交
694 695
	res = run_filter(skb, sk, snaplen);
	if (!res)
696
		goto drop_n_restore;
D
David S. Miller 已提交
697 698
	if (snaplen > res)
		snaplen = res;
L
Linus Torvalds 已提交
699 700

	if (sk->sk_type == SOCK_DGRAM) {
701 702
		macoff = netoff = TPACKET_ALIGN(po->tp_hdrlen) + 16 +
				  po->tp_reserve;
L
Linus Torvalds 已提交
703
	} else {
704
		unsigned maclen = skb_network_offset(skb);
705
		netoff = TPACKET_ALIGN(po->tp_hdrlen +
706 707
				       (maclen < 16 ? 16 : maclen)) +
			po->tp_reserve;
L
Linus Torvalds 已提交
708 709 710
		macoff = netoff - maclen;
	}

J
Johann Baudy 已提交
711
	if (macoff + snaplen > po->rx_ring.frame_size) {
L
Linus Torvalds 已提交
712 713 714 715 716 717 718 719 720 721 722 723
		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
724
		snaplen = po->rx_ring.frame_size - macoff;
L
Linus Torvalds 已提交
725 726 727 728 729
		if ((int)snaplen < 0)
			snaplen = 0;
	}

	spin_lock(&sk->sk_receive_queue.lock);
J
Johann Baudy 已提交
730
	h.raw = packet_current_frame(po, &po->rx_ring, TP_STATUS_KERNEL);
731
	if (!h.raw)
L
Linus Torvalds 已提交
732
		goto ring_is_full;
J
Johann Baudy 已提交
733
	packet_increment_head(&po->rx_ring);
L
Linus Torvalds 已提交
734 735 736 737 738 739 740 741 742
	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);

743
	skb_copy_bits(skb, 0, h.raw + macoff, snaplen);
L
Linus Torvalds 已提交
744

745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769
	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;
		if (skb->tstamp.tv64)
			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;
		if (skb->tstamp.tv64)
			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
770
		h.h2->tp_vlan_tci = vlan_tx_tag_get(skb);
771 772 773 774 775
		hdrlen = sizeof(*h.h2);
		break;
	default:
		BUG();
	}
L
Linus Torvalds 已提交
776

777
	sll = h.raw + TPACKET_ALIGN(hdrlen);
S
Stephen Hemminger 已提交
778
	sll->sll_halen = dev_parse_header(skb, sll->sll_addr);
L
Linus Torvalds 已提交
779 780 781 782
	sll->sll_family = AF_PACKET;
	sll->sll_hatype = dev->type;
	sll->sll_protocol = skb->protocol;
	sll->sll_pkttype = skb->pkt_type;
783
	if (unlikely(po->origdev))
784 785 786
		sll->sll_ifindex = orig_dev->ifindex;
	else
		sll->sll_ifindex = dev->ifindex;
L
Linus Torvalds 已提交
787

788
	__packet_set_status(po, h.raw, status);
R
Ralf Baechle 已提交
789
	smp_mb();
L
Linus Torvalds 已提交
790 791
	{
		struct page *p_start, *p_end;
792
		u8 *h_end = h.raw + macoff + snaplen - 1;
L
Linus Torvalds 已提交
793

794
		p_start = virt_to_page(h.raw);
L
Linus Torvalds 已提交
795 796 797 798 799 800 801 802 803 804 805 806 807 808 809
		p_end = virt_to_page(h_end);
		while (p_start <= p_end) {
			flush_dcache_page(p_start);
			p_start++;
		}
	}

	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:
810
	kfree_skb(skb);
L
Linus Torvalds 已提交
811 812 813 814 815 816 817
	return 0;

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

	sk->sk_data_ready(sk, 0);
818
	kfree_skb(copy_skb);
L
Linus Torvalds 已提交
819 820 821
	goto drop_n_restore;
}

J
Johann Baudy 已提交
822 823 824
static void tpacket_destruct_skb(struct sk_buff *skb)
{
	struct packet_sock *po = pkt_sk(skb->sk);
E
Eric Dumazet 已提交
825
	void *ph;
L
Linus Torvalds 已提交
826

J
Johann Baudy 已提交
827
	BUG_ON(skb == NULL);
L
Linus Torvalds 已提交
828

J
Johann Baudy 已提交
829 830 831 832 833 834 835 836 837 838 839
	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
840 841 842
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
843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859
{
	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;
860
	skb->mark = po->sk.sk_mark;
J
Johann Baudy 已提交
861 862 863 864 865 866 867 868 869 870 871
	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
872
		pr_err("packet size is too long (%d > %d)\n", tp_len, size_max);
J
Johann Baudy 已提交
873 874 875 876 877 878 879 880 881 882 883 884 885 886
		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
887
	} else if (dev->hard_header_len) {
J
Johann Baudy 已提交
888 889
		/* net device doesn't like empty head */
		if (unlikely(tp_len <= dev->hard_header_len)) {
E
Eric Dumazet 已提交
890 891
			pr_err("packet size is too short (%d < %d)\n",
			       tp_len, dev->hard_header_len);
J
Johann Baudy 已提交
892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919
			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;
	page = virt_to_page(data);
	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
920 921
			pr_err("Packet exceed the number of skb frags(%lu)\n",
			       MAX_SKB_FRAGS);
J
Johann Baudy 已提交
922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945
			return -EFAULT;
		}

		flush_dcache_page(page);
		get_page(page);
		skb_fill_page_desc(skb,
				nr_frags,
				page++, offset, len);
		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
946 947
	void *ph;
	struct sockaddr_ll *saddr = (struct sockaddr_ll *)msg->msg_name;
J
Johann Baudy 已提交
948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986
	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
987
		- (po->tp_hdrlen - sizeof(struct sockaddr_ll));
J
Johann Baudy 已提交
988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036

	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);
		if (unlikely(err > 0 && (err = net_xmit_errno(err)) != 0))
			goto out_xmit;
		packet_increment_head(&po->tx_ring);
		len_sum += tp_len;
1037 1038 1039 1040
	} while (likely((ph != NULL) ||
			((!(msg->msg_flags & MSG_DONTWAIT)) &&
			 (atomic_read(&po->tx_ring.pending))))
		);
J
Johann Baudy 已提交
1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059

	err = len_sum;
	goto out_put;

out_xmit:
	skb->destructor = sock_wfree;
	atomic_dec(&po->tx_ring.pending);
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;
}
#endif

static int packet_snd(struct socket *sock,
L
Linus Torvalds 已提交
1060 1061 1062
			  struct msghdr *msg, size_t len)
{
	struct sock *sk = sock->sk;
E
Eric Dumazet 已提交
1063
	struct sockaddr_ll *saddr = (struct sockaddr_ll *)msg->msg_name;
L
Linus Torvalds 已提交
1064 1065
	struct sk_buff *skb;
	struct net_device *dev;
A
Al Viro 已提交
1066
	__be16 proto;
L
Linus Torvalds 已提交
1067 1068 1069 1070
	unsigned char *addr;
	int ifindex, err, reserve = 0;

	/*
1071
	 *	Get and verify the address.
L
Linus Torvalds 已提交
1072
	 */
1073

L
Linus Torvalds 已提交
1074 1075 1076 1077 1078 1079 1080 1081 1082 1083
	if (saddr == NULL) {
		struct packet_sock *po = pkt_sk(sk);

		ifindex	= po->ifindex;
		proto	= po->num;
		addr	= NULL;
	} else {
		err = -EINVAL;
		if (msg->msg_namelen < sizeof(struct sockaddr_ll))
			goto out;
1084 1085
		if (msg->msg_namelen < (saddr->sll_halen + offsetof(struct sockaddr_ll, sll_addr)))
			goto out;
L
Linus Torvalds 已提交
1086 1087 1088 1089 1090 1091
		ifindex	= saddr->sll_ifindex;
		proto	= saddr->sll_protocol;
		addr	= saddr->sll_addr;
	}


1092
	dev = dev_get_by_index(sock_net(sk), ifindex);
L
Linus Torvalds 已提交
1093 1094 1095 1096 1097 1098
	err = -ENXIO;
	if (dev == NULL)
		goto out_unlock;
	if (sock->type == SOCK_RAW)
		reserve = dev->hard_header_len;

1099 1100 1101 1102
	err = -ENETDOWN;
	if (!(dev->flags & IFF_UP))
		goto out_unlock;

L
Linus Torvalds 已提交
1103 1104 1105 1106
	err = -EMSGSIZE;
	if (len > dev->mtu+reserve)
		goto out_unlock;

1107
	skb = sock_alloc_send_skb(sk, len + LL_ALLOCATED_SPACE(dev),
L
Linus Torvalds 已提交
1108
				msg->msg_flags & MSG_DONTWAIT, &err);
E
Eric Dumazet 已提交
1109
	if (skb == NULL)
L
Linus Torvalds 已提交
1110 1111 1112
		goto out_unlock;

	skb_reserve(skb, LL_RESERVED_SPACE(dev));
1113
	skb_reset_network_header(skb);
L
Linus Torvalds 已提交
1114

1115 1116 1117 1118
	err = -EINVAL;
	if (sock->type == SOCK_DGRAM &&
	    dev_hard_header(skb, dev, ntohs(proto), addr, NULL, len) < 0)
		goto out_free;
L
Linus Torvalds 已提交
1119 1120

	/* Returns -EFAULT on error */
E
Eric Dumazet 已提交
1121
	err = memcpy_fromiovec(skb_put(skb, len), msg->msg_iov, len);
L
Linus Torvalds 已提交
1122 1123 1124 1125 1126 1127
	if (err)
		goto out_free;

	skb->protocol = proto;
	skb->dev = dev;
	skb->priority = sk->sk_priority;
1128
	skb->mark = sk->sk_mark;
L
Linus Torvalds 已提交
1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139

	/*
	 *	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
1140
	return len;
L
Linus Torvalds 已提交
1141 1142 1143 1144 1145 1146 1147 1148 1149 1150

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

J
Johann Baudy 已提交
1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163
static int packet_sendmsg(struct kiocb *iocb, struct socket *sock,
		struct msghdr *msg, size_t len)
{
#ifdef CONFIG_PACKET_MMAP
	struct sock *sk = sock->sk;
	struct packet_sock *po = pkt_sk(sk);
	if (po->tx_ring.pg_vec)
		return tpacket_snd(po, msg);
	else
#endif
		return packet_snd(sock, msg, len);
}

L
Linus Torvalds 已提交
1164 1165 1166 1167 1168 1169 1170 1171 1172
/*
 *	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;
1173
	struct net *net;
J
Johann Baudy 已提交
1174 1175 1176
#ifdef CONFIG_PACKET_MMAP
	struct tpacket_req req;
#endif
L
Linus Torvalds 已提交
1177 1178 1179 1180

	if (!sk)
		return 0;

1181
	net = sock_net(sk);
L
Linus Torvalds 已提交
1182 1183
	po = pkt_sk(sk);

1184
	write_lock_bh(&net->packet.sklist_lock);
L
Linus Torvalds 已提交
1185
	sk_del_node_init(sk);
1186
	sock_prot_inuse_add(net, sk->sk_prot, -1);
1187
	write_unlock_bh(&net->packet.sklist_lock);
L
Linus Torvalds 已提交
1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205

	/*
	 *	Unhook packet receive handler.
	 */

	if (po->running) {
		/*
		 *	Remove the protocol hook
		 */
		dev_remove_pack(&po->prot_hook);
		po->running = 0;
		po->num = 0;
		__sock_put(sk);
	}

	packet_flush_mclist(sk);

#ifdef CONFIG_PACKET_MMAP
J
Johann Baudy 已提交
1206 1207 1208 1209 1210 1211 1212
	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
1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224
#endif

	/*
	 *	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);
1225
	sk_refcnt_debug_release(sk);
L
Linus Torvalds 已提交
1226 1227 1228 1229 1230 1231 1232 1233 1234

	sock_put(sk);
	return 0;
}

/*
 *	Attach a packet hook.
 */

A
Al Viro 已提交
1235
static int packet_do_bind(struct sock *sk, struct net_device *dev, __be16 protocol)
L
Linus Torvalds 已提交
1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262
{
	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;

1263
	if (!dev || (dev->flags & IFF_UP)) {
L
Linus Torvalds 已提交
1264 1265 1266
		dev_add_pack(&po->prot_hook);
		sock_hold(sk);
		po->running = 1;
1267 1268 1269 1270
	} else {
		sk->sk_err = ENETDOWN;
		if (!sock_flag(sk, SOCK_DEAD))
			sk->sk_error_report(sk);
L
Linus Torvalds 已提交
1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282
	}

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
1283 1284
static int packet_bind_spkt(struct socket *sock, struct sockaddr *uaddr,
			    int addr_len)
L
Linus Torvalds 已提交
1285
{
E
Eric Dumazet 已提交
1286
	struct sock *sk = sock->sk;
L
Linus Torvalds 已提交
1287 1288 1289
	char name[15];
	struct net_device *dev;
	int err = -ENODEV;
1290

L
Linus Torvalds 已提交
1291 1292 1293
	/*
	 *	Check legality
	 */
1294

1295
	if (addr_len != sizeof(struct sockaddr))
L
Linus Torvalds 已提交
1296
		return -EINVAL;
E
Eric Dumazet 已提交
1297
	strlcpy(name, uaddr->sa_data, sizeof(name));
L
Linus Torvalds 已提交
1298

1299
	dev = dev_get_by_name(sock_net(sk), name);
L
Linus Torvalds 已提交
1300 1301 1302 1303 1304 1305 1306 1307 1308
	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
1309 1310
	struct sockaddr_ll *sll = (struct sockaddr_ll *)uaddr;
	struct sock *sk = sock->sk;
L
Linus Torvalds 已提交
1311 1312 1313 1314 1315 1316 1317
	struct net_device *dev = NULL;
	int err;


	/*
	 *	Check legality
	 */
1318

L
Linus Torvalds 已提交
1319 1320 1321 1322 1323 1324 1325
	if (addr_len < sizeof(struct sockaddr_ll))
		return -EINVAL;
	if (sll->sll_family != AF_PACKET)
		return -EINVAL;

	if (sll->sll_ifindex) {
		err = -ENODEV;
1326
		dev = dev_get_by_index(sock_net(sk), sll->sll_ifindex);
L
Linus Torvalds 已提交
1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344
		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),
};

/*
1345
 *	Create a packet of type SOCK_PACKET.
L
Linus Torvalds 已提交
1346 1347
 */

1348 1349
static int packet_create(struct net *net, struct socket *sock, int protocol,
			 int kern)
L
Linus Torvalds 已提交
1350 1351 1352
{
	struct sock *sk;
	struct packet_sock *po;
A
Al Viro 已提交
1353
	__be16 proto = (__force __be16)protocol; /* weird, but documented */
L
Linus Torvalds 已提交
1354 1355 1356 1357
	int err;

	if (!capable(CAP_NET_RAW))
		return -EPERM;
1358 1359
	if (sock->type != SOCK_DGRAM && sock->type != SOCK_RAW &&
	    sock->type != SOCK_PACKET)
L
Linus Torvalds 已提交
1360 1361 1362 1363 1364
		return -ESOCKTNOSUPPORT;

	sock->state = SS_UNCONNECTED;

	err = -ENOBUFS;
1365
	sk = sk_alloc(net, PF_PACKET, GFP_KERNEL, &packet_proto);
L
Linus Torvalds 已提交
1366 1367 1368 1369 1370 1371
	if (sk == NULL)
		goto out;

	sock->ops = &packet_ops;
	if (sock->type == SOCK_PACKET)
		sock->ops = &packet_ops_spkt;
1372

L
Linus Torvalds 已提交
1373 1374 1375 1376
	sock_init_data(sock, sk);

	po = pkt_sk(sk);
	sk->sk_family = PF_PACKET;
A
Al Viro 已提交
1377
	po->num = proto;
L
Linus Torvalds 已提交
1378 1379

	sk->sk_destruct = packet_sock_destruct;
1380
	sk_refcnt_debug_inc(sk);
L
Linus Torvalds 已提交
1381 1382 1383 1384 1385 1386

	/*
	 *	Attach a protocol block
	 */

	spin_lock_init(&po->bind_lock);
1387
	mutex_init(&po->pg_vec_lock);
L
Linus Torvalds 已提交
1388
	po->prot_hook.func = packet_rcv;
1389

L
Linus Torvalds 已提交
1390 1391
	if (sock->type == SOCK_PACKET)
		po->prot_hook.func = packet_rcv_spkt;
1392

L
Linus Torvalds 已提交
1393 1394
	po->prot_hook.af_packet_priv = sk;

A
Al Viro 已提交
1395 1396
	if (proto) {
		po->prot_hook.type = proto;
L
Linus Torvalds 已提交
1397 1398 1399 1400 1401
		dev_add_pack(&po->prot_hook);
		sock_hold(sk);
		po->running = 1;
	}

1402 1403
	write_lock_bh(&net->packet.sklist_lock);
	sk_add_node(sk, &net->packet.sklist);
1404
	sock_prot_inuse_add(net, &packet_proto, 1);
1405
	write_unlock_bh(&net->packet.sklist_lock);
E
Eric Dumazet 已提交
1406
	return 0;
L
Linus Torvalds 已提交
1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421
out:
	return err;
}

/*
 *	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;
1422
	struct sockaddr_ll *sll;
L
Linus Torvalds 已提交
1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442

	err = -EINVAL;
	if (flags & ~(MSG_PEEK|MSG_DONTWAIT|MSG_TRUNC|MSG_CMSG_COMPAT))
		goto out;

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

	/*
	 *	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
1443
	skb = skb_recv_datagram(sk, flags, flags & MSG_DONTWAIT, &err);
L
Linus Torvalds 已提交
1444 1445

	/*
1446
	 *	An error occurred so return it. Because skb_recv_datagram()
L
Linus Torvalds 已提交
1447 1448 1449 1450
	 *	handles the blocking we don't see and worry about blocking
	 *	retries.
	 */

1451
	if (skb == NULL)
L
Linus Torvalds 已提交
1452 1453
		goto out;

1454 1455 1456 1457 1458
	/*
	 *	If the address length field is there to be filled in, we fill
	 *	it in now.
	 */

1459
	sll = &PACKET_SKB_CB(skb)->sa.ll;
1460 1461 1462 1463 1464
	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
1465 1466 1467 1468 1469 1470
	/*
	 *	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
1471 1472 1473
	if (copied > len) {
		copied = len;
		msg->msg_flags |= MSG_TRUNC;
L
Linus Torvalds 已提交
1474 1475 1476 1477 1478 1479
	}

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

1480
	sock_recv_ts_and_drops(msg, sk, skb);
L
Linus Torvalds 已提交
1481 1482

	if (msg->msg_name)
1483 1484
		memcpy(msg->msg_name, &PACKET_SKB_CB(skb)->sa,
		       msg->msg_namelen);
L
Linus Torvalds 已提交
1485

1486
	if (pkt_sk(sk)->auxdata) {
1487 1488 1489 1490 1491 1492 1493 1494
		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;
1495
		aux.tp_net = skb_network_offset(skb);
E
Eric Dumazet 已提交
1496
		aux.tp_vlan_tci = vlan_tx_tag_get(skb);
1497 1498

		put_cmsg(msg, SOL_PACKET, PACKET_AUXDATA, sizeof(aux), &aux);
1499 1500
	}

L
Linus Torvalds 已提交
1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522
	/*
	 *	Free or return the buffer as appropriate. Again this
	 *	hides all the races and re-entrancy issues from us.
	 */
	err = (flags&MSG_TRUNC) ? skb->len : copied;

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
1523 1524 1525
	rcu_read_lock();
	dev = dev_get_by_index_rcu(sock_net(sk), pkt_sk(sk)->ifindex);
	if (dev)
L
Linus Torvalds 已提交
1526
		strlcpy(uaddr->sa_data, dev->name, 15);
E
Eric Dumazet 已提交
1527
	else
L
Linus Torvalds 已提交
1528
		memset(uaddr->sa_data, 0, 14);
E
Eric Dumazet 已提交
1529
	rcu_read_unlock();
L
Linus Torvalds 已提交
1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540
	*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);
1541
	DECLARE_SOCKADDR(struct sockaddr_ll *, sll, uaddr);
L
Linus Torvalds 已提交
1542 1543 1544 1545 1546 1547 1548

	if (peer)
		return -EOPNOTSUPP;

	sll->sll_family = AF_PACKET;
	sll->sll_ifindex = po->ifindex;
	sll->sll_protocol = po->num;
E
Eric Dumazet 已提交
1549 1550
	rcu_read_lock();
	dev = dev_get_by_index_rcu(sock_net(sk), po->ifindex);
L
Linus Torvalds 已提交
1551 1552 1553 1554 1555 1556 1557 1558
	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
1559
	rcu_read_unlock();
1560
	*uaddr_len = offsetof(struct sockaddr_ll, sll_addr) + sll->sll_halen;
L
Linus Torvalds 已提交
1561 1562 1563 1564

	return 0;
}

1565 1566
static int packet_dev_mc(struct net_device *dev, struct packet_mclist *i,
			 int what)
L
Linus Torvalds 已提交
1567 1568 1569 1570
{
	switch (i->type) {
	case PACKET_MR_MULTICAST:
		if (what > 0)
1571
			return dev_mc_add(dev, i->addr, i->alen, 0);
L
Linus Torvalds 已提交
1572
		else
1573
			return dev_mc_delete(dev, i->addr, i->alen, 0);
L
Linus Torvalds 已提交
1574 1575
		break;
	case PACKET_MR_PROMISC:
1576
		return dev_set_promiscuity(dev, what);
L
Linus Torvalds 已提交
1577 1578
		break;
	case PACKET_MR_ALLMULTI:
1579
		return dev_set_allmulti(dev, what);
L
Linus Torvalds 已提交
1580
		break;
1581 1582
	case PACKET_MR_UNICAST:
		if (what > 0)
J
Jiri Pirko 已提交
1583
			return dev_unicast_add(dev, i->addr);
1584
		else
J
Jiri Pirko 已提交
1585
			return dev_unicast_delete(dev, i->addr);
1586
		break;
E
Eric Dumazet 已提交
1587 1588
	default:
		break;
L
Linus Torvalds 已提交
1589
	}
1590
	return 0;
L
Linus Torvalds 已提交
1591 1592 1593 1594
}

static void packet_dev_mclist(struct net_device *dev, struct packet_mclist *i, int what)
{
E
Eric Dumazet 已提交
1595
	for ( ; i; i = i->next) {
L
Linus Torvalds 已提交
1596 1597 1598 1599 1600
		if (i->ifindex == dev->ifindex)
			packet_dev_mc(dev, i, what);
	}
}

1601
static int packet_mc_add(struct sock *sk, struct packet_mreq_max *mreq)
L
Linus Torvalds 已提交
1602 1603 1604 1605 1606 1607 1608 1609 1610
{
	struct packet_sock *po = pkt_sk(sk);
	struct packet_mclist *ml, *i;
	struct net_device *dev;
	int err;

	rtnl_lock();

	err = -ENODEV;
1611
	dev = __dev_get_by_index(sock_net(sk), mreq->mr_ifindex);
L
Linus Torvalds 已提交
1612 1613 1614 1615 1616 1617 1618 1619
	if (!dev)
		goto done;

	err = -EINVAL;
	if (mreq->mr_alen > dev->addr_len)
		goto done;

	err = -ENOBUFS;
1620
	i = kmalloc(sizeof(*i), GFP_KERNEL);
L
Linus Torvalds 已提交
1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643
	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;
1644 1645 1646 1647 1648
	err = packet_dev_mc(dev, i, 1);
	if (err) {
		po->mclist = i->next;
		kfree(i);
	}
L
Linus Torvalds 已提交
1649 1650 1651 1652 1653 1654

done:
	rtnl_unlock();
	return err;
}

1655
static int packet_mc_drop(struct sock *sk, struct packet_mreq_max *mreq)
L
Linus Torvalds 已提交
1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668
{
	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;
1669 1670
				dev = __dev_get_by_index(sock_net(sk), ml->ifindex);
				if (dev)
L
Linus Torvalds 已提交
1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694
					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;
1695 1696
		dev = __dev_get_by_index(sock_net(sk), ml->ifindex);
		if (dev != NULL)
L
Linus Torvalds 已提交
1697 1698 1699 1700 1701 1702 1703
			packet_dev_mc(dev, ml, -1);
		kfree(ml);
	}
	rtnl_unlock();
}

static int
1704
packet_setsockopt(struct socket *sock, int level, int optname, char __user *optval, unsigned int optlen)
L
Linus Torvalds 已提交
1705 1706
{
	struct sock *sk = sock->sk;
1707
	struct packet_sock *po = pkt_sk(sk);
L
Linus Torvalds 已提交
1708 1709 1710 1711 1712
	int ret;

	if (level != SOL_PACKET)
		return -ENOPROTOOPT;

J
Johann Baudy 已提交
1713
	switch (optname) {
1714
	case PACKET_ADD_MEMBERSHIP:
L
Linus Torvalds 已提交
1715 1716
	case PACKET_DROP_MEMBERSHIP:
	{
1717 1718 1719 1720
		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
1721
			return -EINVAL;
1722 1723
		if (len > sizeof(mreq))
			len = sizeof(mreq);
E
Eric Dumazet 已提交
1724
		if (copy_from_user(&mreq, optval, len))
L
Linus Torvalds 已提交
1725
			return -EFAULT;
1726 1727
		if (len < (mreq.mr_alen + offsetof(struct packet_mreq, mr_address)))
			return -EINVAL;
L
Linus Torvalds 已提交
1728 1729 1730 1731 1732 1733
		if (optname == PACKET_ADD_MEMBERSHIP)
			ret = packet_mc_add(sk, &mreq);
		else
			ret = packet_mc_drop(sk, &mreq);
		return ret;
	}
1734

L
Linus Torvalds 已提交
1735 1736
#ifdef CONFIG_PACKET_MMAP
	case PACKET_RX_RING:
J
Johann Baudy 已提交
1737
	case PACKET_TX_RING:
L
Linus Torvalds 已提交
1738 1739 1740
	{
		struct tpacket_req req;

E
Eric Dumazet 已提交
1741
		if (optlen < sizeof(req))
L
Linus Torvalds 已提交
1742
			return -EINVAL;
E
Eric Dumazet 已提交
1743
		if (copy_from_user(&req, optval, sizeof(req)))
L
Linus Torvalds 已提交
1744
			return -EFAULT;
J
Johann Baudy 已提交
1745
		return packet_set_ring(sk, &req, 0, optname == PACKET_TX_RING);
L
Linus Torvalds 已提交
1746 1747 1748 1749 1750
	}
	case PACKET_COPY_THRESH:
	{
		int val;

E
Eric Dumazet 已提交
1751
		if (optlen != sizeof(val))
L
Linus Torvalds 已提交
1752
			return -EINVAL;
E
Eric Dumazet 已提交
1753
		if (copy_from_user(&val, optval, sizeof(val)))
L
Linus Torvalds 已提交
1754 1755 1756 1757 1758
			return -EFAULT;

		pkt_sk(sk)->copy_thresh = val;
		return 0;
	}
1759 1760 1761 1762 1763 1764
	case PACKET_VERSION:
	{
		int val;

		if (optlen != sizeof(val))
			return -EINVAL;
J
Johann Baudy 已提交
1765
		if (po->rx_ring.pg_vec || po->tx_ring.pg_vec)
1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777
			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;
		}
	}
1778 1779 1780 1781 1782 1783
	case PACKET_RESERVE:
	{
		unsigned int val;

		if (optlen != sizeof(val))
			return -EINVAL;
J
Johann Baudy 已提交
1784
		if (po->rx_ring.pg_vec || po->tx_ring.pg_vec)
1785 1786 1787 1788 1789 1790
			return -EBUSY;
		if (copy_from_user(&val, optval, sizeof(val)))
			return -EFAULT;
		po->tp_reserve = val;
		return 0;
	}
J
Johann Baudy 已提交
1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803
	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;
	}
L
Linus Torvalds 已提交
1804
#endif
1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816
	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;
	}
1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828
	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;
	}
L
Linus Torvalds 已提交
1829 1830 1831 1832 1833 1834 1835 1836 1837
	default:
		return -ENOPROTOOPT;
	}
}

static int packet_getsockopt(struct socket *sock, int level, int optname,
			     char __user *optval, int __user *optlen)
{
	int len;
1838
	int val;
L
Linus Torvalds 已提交
1839 1840
	struct sock *sk = sock->sk;
	struct packet_sock *po = pkt_sk(sk);
1841 1842
	void *data;
	struct tpacket_stats st;
L
Linus Torvalds 已提交
1843 1844 1845 1846

	if (level != SOL_PACKET)
		return -ENOPROTOOPT;

1847 1848
	if (get_user(len, optlen))
		return -EFAULT;
L
Linus Torvalds 已提交
1849 1850 1851

	if (len < 0)
		return -EINVAL;
1852

J
Johann Baudy 已提交
1853
	switch (optname) {
L
Linus Torvalds 已提交
1854 1855 1856 1857 1858 1859 1860 1861 1862
	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;

1863 1864 1865 1866 1867 1868 1869
		data = &st;
		break;
	case PACKET_AUXDATA:
		if (len > sizeof(int))
			len = sizeof(int);
		val = po->auxdata;

1870 1871 1872 1873 1874 1875 1876
		data = &val;
		break;
	case PACKET_ORIGDEV:
		if (len > sizeof(int))
			len = sizeof(int);
		val = po->origdev;

1877
		data = &val;
L
Linus Torvalds 已提交
1878
		break;
1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902
#ifdef CONFIG_PACKET_MMAP
	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;
1903 1904 1905 1906 1907 1908
	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
1909 1910 1911 1912 1913 1914
	case PACKET_LOSS:
		if (len > sizeof(unsigned int))
			len = sizeof(unsigned int);
		val = po->tp_loss;
		data = &val;
		break;
1915
#endif
L
Linus Torvalds 已提交
1916 1917 1918 1919
	default:
		return -ENOPROTOOPT;
	}

1920 1921
	if (put_user(len, optlen))
		return -EFAULT;
1922 1923
	if (copy_to_user(optval, data, len))
		return -EFAULT;
1924
	return 0;
L
Linus Torvalds 已提交
1925 1926 1927 1928 1929 1930 1931
}


static int packet_notifier(struct notifier_block *this, unsigned long msg, void *data)
{
	struct sock *sk;
	struct hlist_node *node;
1932
	struct net_device *dev = data;
1933
	struct net *net = dev_net(dev);
L
Linus Torvalds 已提交
1934

1935 1936
	read_lock(&net->packet.sklist_lock);
	sk_for_each(sk, node, &net->packet.sklist) {
L
Linus Torvalds 已提交
1937 1938 1939 1940 1941 1942
		struct packet_sock *po = pkt_sk(sk);

		switch (msg) {
		case NETDEV_UNREGISTER:
			if (po->mclist)
				packet_dev_mclist(dev, po->mclist, -1);
1943 1944
			/* fallthrough */

L
Linus Torvalds 已提交
1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974
		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:
			spin_lock(&po->bind_lock);
			if (dev->ifindex == po->ifindex && po->num &&
			    !po->running) {
				dev_add_pack(&po->prot_hook);
				sock_hold(sk);
				po->running = 1;
			}
			spin_unlock(&po->bind_lock);
			break;
		}
	}
1975
	read_unlock(&net->packet.sklist_lock);
L
Linus Torvalds 已提交
1976 1977 1978 1979 1980 1981 1982 1983 1984
	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
1985
	switch (cmd) {
E
Eric Dumazet 已提交
1986 1987 1988
	case SIOCOUTQ:
	{
		int amount = sk_wmem_alloc_get(sk);
1989

E
Eric Dumazet 已提交
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007
		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);
2008

L
Linus Torvalds 已提交
2009
#ifdef CONFIG_INET
E
Eric Dumazet 已提交
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026
	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:
		if (!net_eq(sock_net(sk), &init_net))
			return -ENOIOCTLCMD;
		return inet_dgram_ops.ioctl(sock, cmd, arg);
L
Linus Torvalds 已提交
2027 2028
#endif

E
Eric Dumazet 已提交
2029 2030
	default:
		return -ENOIOCTLCMD;
L
Linus Torvalds 已提交
2031 2032 2033 2034 2035 2036 2037 2038 2039
	}
	return 0;
}

#ifndef CONFIG_PACKET_MMAP
#define packet_mmap sock_no_mmap
#define packet_poll datagram_poll
#else

E
Eric Dumazet 已提交
2040
static unsigned int packet_poll(struct file *file, struct socket *sock,
L
Linus Torvalds 已提交
2041 2042 2043 2044 2045 2046 2047
				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
2048 2049
	if (po->rx_ring.pg_vec) {
		if (!packet_previous_frame(po, &po->rx_ring, TP_STATUS_KERNEL))
L
Linus Torvalds 已提交
2050 2051 2052
			mask |= POLLIN | POLLRDNORM;
	}
	spin_unlock_bh(&sk->sk_receive_queue.lock);
J
Johann Baudy 已提交
2053 2054 2055 2056 2057 2058
	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
2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069
	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
2070
	struct socket *sock = file->private_data;
L
Linus Torvalds 已提交
2071
	struct sock *sk = sock->sk;
2072

L
Linus Torvalds 已提交
2073 2074 2075 2076 2077 2078 2079
	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
2080
	struct socket *sock = file->private_data;
L
Linus Torvalds 已提交
2081
	struct sock *sk = sock->sk;
2082

L
Linus Torvalds 已提交
2083 2084 2085 2086
	if (sk)
		atomic_dec(&pkt_sk(sk)->mapped);
}

2087
static const struct vm_operations_struct packet_mmap_ops = {
E
Eric Dumazet 已提交
2088 2089
	.open	=	packet_mm_open,
	.close	=	packet_mm_close,
L
Linus Torvalds 已提交
2090 2091
};

2092
static void free_pg_vec(char **pg_vec, unsigned int order, unsigned int len)
L
Linus Torvalds 已提交
2093 2094 2095
{
	int i;

2096 2097 2098
	for (i = 0; i < len; i++) {
		if (likely(pg_vec[i]))
			free_pages((unsigned long) pg_vec[i], order);
L
Linus Torvalds 已提交
2099 2100 2101 2102
	}
	kfree(pg_vec);
}

2103 2104
static inline char *alloc_one_pg_vec_page(unsigned long order)
{
2105 2106 2107
	gfp_t gfp_flags = GFP_KERNEL | __GFP_COMP | __GFP_ZERO | __GFP_NOWARN;

	return (char *) __get_free_pages(gfp_flags, order);
2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133
}

static char **alloc_pg_vec(struct tpacket_req *req, int order)
{
	unsigned int block_nr = req->tp_block_nr;
	char **pg_vec;
	int i;

	pg_vec = kzalloc(block_nr * sizeof(char *), GFP_KERNEL);
	if (unlikely(!pg_vec))
		goto out;

	for (i = 0; i < block_nr; i++) {
		pg_vec[i] = alloc_one_pg_vec_page(order);
		if (unlikely(!pg_vec[i]))
			goto out_free_pgvec;
	}

out:
	return pg_vec;

out_free_pgvec:
	free_pg_vec(pg_vec, order, block_nr);
	pg_vec = NULL;
	goto out;
}
L
Linus Torvalds 已提交
2134

J
Johann Baudy 已提交
2135 2136
static int packet_set_ring(struct sock *sk, struct tpacket_req *req,
		int closing, int tx_ring)
L
Linus Torvalds 已提交
2137 2138 2139
{
	char **pg_vec = NULL;
	struct packet_sock *po = pkt_sk(sk);
A
Al Viro 已提交
2140
	int was_running, order = 0;
J
Johann Baudy 已提交
2141 2142
	struct packet_ring_buffer *rb;
	struct sk_buff_head *rb_queue;
A
Al Viro 已提交
2143
	__be16 num;
J
Johann Baudy 已提交
2144
	int err;
2145

J
Johann Baudy 已提交
2146 2147
	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
2148

J
Johann Baudy 已提交
2149 2150 2151 2152 2153 2154 2155
	err = -EBUSY;
	if (!closing) {
		if (atomic_read(&po->mapped))
			goto out;
		if (atomic_read(&rb->pending))
			goto out;
	}
L
Linus Torvalds 已提交
2156

J
Johann Baudy 已提交
2157 2158 2159 2160 2161
	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
2162

2163 2164 2165 2166 2167 2168 2169 2170 2171
		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
2172
		err = -EINVAL;
2173
		if (unlikely((int)req->tp_block_size <= 0))
J
Johann Baudy 已提交
2174
			goto out;
2175
		if (unlikely(req->tp_block_size & (PAGE_SIZE - 1)))
J
Johann Baudy 已提交
2176
			goto out;
2177
		if (unlikely(req->tp_frame_size < po->tp_hdrlen +
J
Johann Baudy 已提交
2178 2179
					po->tp_reserve))
			goto out;
2180
		if (unlikely(req->tp_frame_size & (TPACKET_ALIGNMENT - 1)))
J
Johann Baudy 已提交
2181
			goto out;
L
Linus Torvalds 已提交
2182

J
Johann Baudy 已提交
2183 2184 2185 2186 2187 2188
		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
2189 2190

		err = -ENOMEM;
2191 2192 2193
		order = get_order(req->tp_block_size);
		pg_vec = alloc_pg_vec(req, order);
		if (unlikely(!pg_vec))
L
Linus Torvalds 已提交
2194
			goto out;
J
Johann Baudy 已提交
2195 2196 2197 2198
	}
	/* Done */
	else {
		err = -EINVAL;
2199
		if (unlikely(req->tp_frame_nr))
J
Johann Baudy 已提交
2200
			goto out;
L
Linus Torvalds 已提交
2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215
	}

	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);
2216

L
Linus Torvalds 已提交
2217 2218 2219
	synchronize_net();

	err = -EBUSY;
2220
	mutex_lock(&po->pg_vec_lock);
L
Linus Torvalds 已提交
2221 2222 2223
	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
2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237
		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
2238 2239
#undef XC
		if (atomic_read(&po->mapped))
E
Eric Dumazet 已提交
2240 2241
			pr_err("packet_mmap: vma is busy: %d\n",
			       atomic_read(&po->mapped));
L
Linus Torvalds 已提交
2242
	}
2243
	mutex_unlock(&po->pg_vec_lock);
L
Linus Torvalds 已提交
2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261

	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
2262 2263
static int packet_mmap(struct file *file, struct socket *sock,
		struct vm_area_struct *vma)
L
Linus Torvalds 已提交
2264 2265 2266
{
	struct sock *sk = sock->sk;
	struct packet_sock *po = pkt_sk(sk);
J
Johann Baudy 已提交
2267 2268
	unsigned long size, expected_size;
	struct packet_ring_buffer *rb;
L
Linus Torvalds 已提交
2269 2270 2271 2272 2273 2274 2275
	unsigned long start;
	int err = -EINVAL;
	int i;

	if (vma->vm_pgoff)
		return -EINVAL;

2276
	mutex_lock(&po->pg_vec_lock);
J
Johann Baudy 已提交
2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287

	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
2288
		goto out;
J
Johann Baudy 已提交
2289 2290 2291

	size = vma->vm_end - vma->vm_start;
	if (size != expected_size)
L
Linus Torvalds 已提交
2292 2293 2294
		goto out;

	start = vma->vm_start;
J
Johann Baudy 已提交
2295 2296 2297 2298 2299 2300 2301 2302 2303
	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++) {
			struct page *page = virt_to_page(rb->pg_vec[i]);
			int pg_num;

			for (pg_num = 0; pg_num < rb->pg_vec_pages;
E
Eric Dumazet 已提交
2304
					pg_num++, page++) {
J
Johann Baudy 已提交
2305 2306 2307 2308 2309
				err = vm_insert_page(vma, start, page);
				if (unlikely(err))
					goto out;
				start += PAGE_SIZE;
			}
2310
		}
L
Linus Torvalds 已提交
2311
	}
J
Johann Baudy 已提交
2312

2313
	atomic_inc(&po->mapped);
L
Linus Torvalds 已提交
2314 2315 2316 2317
	vma->vm_ops = &packet_mmap_ops;
	err = 0;

out:
2318
	mutex_unlock(&po->pg_vec_lock);
L
Linus Torvalds 已提交
2319 2320 2321 2322 2323
	return err;
}
#endif


2324
static const struct proto_ops packet_ops_spkt = {
L
Linus Torvalds 已提交
2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344
	.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,
};

2345
static const struct proto_ops packet_ops = {
L
Linus Torvalds 已提交
2346 2347 2348 2349 2350 2351 2352
	.family =	PF_PACKET,
	.owner =	THIS_MODULE,
	.release =	packet_release,
	.bind =		packet_bind,
	.connect =	sock_no_connect,
	.socketpair =	sock_no_socketpair,
	.accept =	sock_no_accept,
2353
	.getname =	packet_getname,
L
Linus Torvalds 已提交
2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365
	.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,
};

2366
static const struct net_proto_family packet_family_ops = {
L
Linus Torvalds 已提交
2367 2368 2369 2370 2371 2372
	.family =	PF_PACKET,
	.create =	packet_create,
	.owner	=	THIS_MODULE,
};

static struct notifier_block packet_netdev_notifier = {
E
Eric Dumazet 已提交
2373
	.notifier_call =	packet_notifier,
L
Linus Torvalds 已提交
2374 2375 2376
};

#ifdef CONFIG_PROC_FS
2377
static inline struct sock *packet_seq_idx(struct net *net, loff_t off)
L
Linus Torvalds 已提交
2378 2379 2380 2381
{
	struct sock *s;
	struct hlist_node *node;

2382
	sk_for_each(s, node, &net->packet.sklist) {
L
Linus Torvalds 已提交
2383 2384 2385 2386 2387 2388 2389
		if (!off--)
			return s;
	}
	return NULL;
}

static void *packet_seq_start(struct seq_file *seq, loff_t *pos)
2390
	__acquires(seq_file_net(seq)->packet.sklist_lock)
L
Linus Torvalds 已提交
2391
{
2392
	struct net *net = seq_file_net(seq);
2393
	read_lock(&net->packet.sklist_lock);
2394
	return *pos ? packet_seq_idx(net, *pos - 1) : SEQ_START_TOKEN;
L
Linus Torvalds 已提交
2395 2396 2397 2398
}

static void *packet_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
2399
	struct net *net = seq_file_net(seq);
L
Linus Torvalds 已提交
2400
	++*pos;
2401
	return  (v == SEQ_START_TOKEN)
2402
		? sk_head(&net->packet.sklist)
E
Eric Dumazet 已提交
2403
		: sk_next((struct sock *)v) ;
L
Linus Torvalds 已提交
2404 2405 2406
}

static void packet_seq_stop(struct seq_file *seq, void *v)
2407
	__releases(seq_file_net(seq)->packet.sklist_lock)
L
Linus Torvalds 已提交
2408
{
2409
	struct net *net = seq_file_net(seq);
2410
	read_unlock(&net->packet.sklist_lock);
L
Linus Torvalds 已提交
2411 2412
}

2413
static int packet_seq_show(struct seq_file *seq, void *v)
L
Linus Torvalds 已提交
2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430
{
	if (v == SEQ_START_TOKEN)
		seq_puts(seq, "sk       RefCnt Type Proto  Iface R Rmem   User   Inode\n");
	else {
		struct sock *s = v;
		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
2431
			   sock_i_ino(s));
L
Linus Torvalds 已提交
2432 2433 2434 2435 2436
	}

	return 0;
}

2437
static const struct seq_operations packet_seq_ops = {
L
Linus Torvalds 已提交
2438 2439 2440 2441 2442 2443 2444 2445
	.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)
{
2446 2447
	return seq_open_net(inode, file, &packet_seq_ops,
			    sizeof(struct seq_net_private));
L
Linus Torvalds 已提交
2448 2449
}

2450
static const struct file_operations packet_seq_fops = {
L
Linus Torvalds 已提交
2451 2452 2453 2454
	.owner		= THIS_MODULE,
	.open		= packet_seq_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
2455
	.release	= seq_release_net,
L
Linus Torvalds 已提交
2456 2457 2458 2459
};

#endif

2460 2461
static int packet_net_init(struct net *net)
{
2462 2463
	rwlock_init(&net->packet.sklist_lock);
	INIT_HLIST_HEAD(&net->packet.sklist);
2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481

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

	return 0;
}

static void packet_net_exit(struct net *net)
{
	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
2482 2483 2484
static void __exit packet_exit(void)
{
	unregister_netdevice_notifier(&packet_netdev_notifier);
2485
	unregister_pernet_subsys(&packet_net_ops);
L
Linus Torvalds 已提交
2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497
	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);
2498
	register_pernet_subsys(&packet_net_ops);
L
Linus Torvalds 已提交
2499 2500 2501 2502 2503 2504 2505 2506 2507
	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);