af_packet.c 88.7 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.
43 44 45 46
 *		Chetan Loke	:	Implemented TPACKET_V3 block abstraction
 *					layer.
 *					Copyright (C) 2011, <lokec@ccs.neu.edu>
 *
L
Linus Torvalds 已提交
47 48 49 50 51 52 53
 *
 *		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.
 *
 */
54

L
Linus Torvalds 已提交
55 56
#include <linux/types.h>
#include <linux/mm.h>
57
#include <linux/capability.h>
L
Linus Torvalds 已提交
58 59 60 61 62 63 64
#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>
65
#include <linux/kernel.h>
L
Linus Torvalds 已提交
66
#include <linux/kmod.h>
67
#include <linux/slab.h>
68
#include <linux/vmalloc.h>
69
#include <net/net_namespace.h>
L
Linus Torvalds 已提交
70 71 72 73 74 75 76 77 78
#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/uaccess.h>
#include <asm/ioctls.h>
#include <asm/page.h>
A
Al Viro 已提交
79
#include <asm/cacheflush.h>
L
Linus Torvalds 已提交
80 81 82 83 84 85
#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>
86
#include <linux/mutex.h>
E
Eric Dumazet 已提交
87
#include <linux/if_vlan.h>
88
#include <linux/virtio_net.h>
89
#include <linux/errqueue.h>
90
#include <linux/net_tstamp.h>
L
Linus Torvalds 已提交
91 92 93 94 95

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

96 97
#include "internal.h"

L
Linus Torvalds 已提交
98 99 100 101 102 103 104 105 106 107 108 109 110 111 112
/*
   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
113 114
   mac_header -> ll header
   data       -> data
L
Linus Torvalds 已提交
115 116

Outgoing, dev->hard_header!=NULL
117 118
   mac_header -> ll header
   data       -> ll header
L
Linus Torvalds 已提交
119 120

Incoming, dev->hard_header==NULL
121 122
   mac_header -> UNKNOWN position. It is very likely, that it points to ll
		 header.  PPP makes it, that is wrong, because introduce
123
		 assymetry between rx and tx paths.
124
   data       -> data
L
Linus Torvalds 已提交
125 126

Outgoing, dev->hard_header==NULL
127 128
   mac_header -> data. ll header is still not built!
   data       -> data
L
Linus Torvalds 已提交
129 130 131 132 133 134 135 136 137

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


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

dev->hard_header != NULL
138 139
   mac_header -> ll header
   data       -> ll header
L
Linus Torvalds 已提交
140 141

dev->hard_header == NULL (ll header is added by device, we cannot control it)
142 143
   mac_header -> data
   data       -> data
L
Linus Torvalds 已提交
144 145 146 147 148 149 150

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

/* Private packet socket structures. */

151 152 153
/* identical to struct packet_mreq except it has
 * a longer address field.
 */
E
Eric Dumazet 已提交
154
struct packet_mreq_max {
155 156 157 158
	int		mr_ifindex;
	unsigned short	mr_type;
	unsigned short	mr_alen;
	unsigned char	mr_address[MAX_ADDR_LEN];
L
Linus Torvalds 已提交
159
};
160

161
static int packet_set_ring(struct sock *sk, union tpacket_req_u *req_u,
J
Johann Baudy 已提交
162 163
		int closing, int tx_ring);

164 165 166

#define V3_ALIGNMENT	(8)

167
#define BLK_HDR_LEN	(ALIGN(sizeof(struct tpacket_block_desc), V3_ALIGNMENT))
168 169 170 171 172

#define BLK_PLUS_PRIV(sz_of_priv) \
	(BLK_HDR_LEN + ALIGN((sz_of_priv), V3_ALIGNMENT))

#define PGV_FROM_VMALLOC 1
J
Johann Baudy 已提交
173

174 175 176 177 178 179 180 181
#define BLOCK_STATUS(x)	((x)->hdr.bh1.block_status)
#define BLOCK_NUM_PKTS(x)	((x)->hdr.bh1.num_pkts)
#define BLOCK_O2FP(x)		((x)->hdr.bh1.offset_to_first_pkt)
#define BLOCK_LEN(x)		((x)->hdr.bh1.blk_len)
#define BLOCK_SNUM(x)		((x)->hdr.bh1.seq_num)
#define BLOCK_O2PRIV(x)	((x)->offset_to_priv)
#define BLOCK_PRIV(x)		((void *)((char *)(x) + BLOCK_O2PRIV(x)))

J
Johann Baudy 已提交
182 183
struct packet_sock;
static int tpacket_snd(struct packet_sock *po, struct msghdr *msg);
L
Linus Torvalds 已提交
184

185 186 187 188
static void *packet_previous_frame(struct packet_sock *po,
		struct packet_ring_buffer *rb,
		int status);
static void packet_increment_head(struct packet_ring_buffer *buff);
189 190 191
static int prb_curr_blk_in_use(struct tpacket_kbdq_core *,
			struct tpacket_block_desc *);
static void *prb_dispatch_next_block(struct tpacket_kbdq_core *,
192
			struct packet_sock *);
193
static void prb_retire_current_block(struct tpacket_kbdq_core *,
194
		struct packet_sock *, unsigned int status);
195 196 197
static int prb_queue_frozen(struct tpacket_kbdq_core *);
static void prb_open_block(struct tpacket_kbdq_core *,
		struct tpacket_block_desc *);
198
static void prb_retire_rx_blk_timer_expired(unsigned long);
199 200 201 202 203 204 205 206 207
static void _prb_refresh_rx_retire_blk_timer(struct tpacket_kbdq_core *);
static void prb_init_blk_timer(struct packet_sock *,
		struct tpacket_kbdq_core *,
		void (*func) (unsigned long));
static void prb_fill_rxhash(struct tpacket_kbdq_core *, struct tpacket3_hdr *);
static void prb_clear_rxhash(struct tpacket_kbdq_core *,
		struct tpacket3_hdr *);
static void prb_fill_vlan_info(struct tpacket_kbdq_core *,
		struct tpacket3_hdr *);
L
Linus Torvalds 已提交
208 209
static void packet_flush_mclist(struct sock *sk);

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

220
#define GET_PBDQC_FROM_RB(x)	((struct tpacket_kbdq_core *)(&(x)->prb_bdqc))
221
#define GET_PBLOCK_DESC(x, bid)	\
222
	((struct tpacket_block_desc *)((x)->pkbdq[(bid)].buffer))
223
#define GET_CURR_PBLOCK_DESC_FROM_CORE(x)	\
224
	((struct tpacket_block_desc *)((x)->pkbdq[(x)->kactive_blk_num].buffer))
225 226 227 228
#define GET_NEXT_PRB_BLK_NUM(x) \
	(((x)->kactive_blk_num < ((x)->knum_blocks-1)) ? \
	((x)->kactive_blk_num+1) : 0)

D
David S. Miller 已提交
229 230 231
static void __fanout_unlink(struct sock *sk, struct packet_sock *po);
static void __fanout_link(struct sock *sk, struct packet_sock *po);

232 233 234 235 236 237 238 239
/* register_prot_hook must be invoked with the po->bind_lock held,
 * or from a context in which asynchronous accesses to the packet
 * socket is not possible (packet_create()).
 */
static void register_prot_hook(struct sock *sk)
{
	struct packet_sock *po = pkt_sk(sk);
	if (!po->running) {
D
David S. Miller 已提交
240 241 242 243
		if (po->fanout)
			__fanout_link(sk, po);
		else
			dev_add_pack(&po->prot_hook);
244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260
		sock_hold(sk);
		po->running = 1;
	}
}

/* {,__}unregister_prot_hook() must be invoked with the po->bind_lock
 * held.   If the sync parameter is true, we will temporarily drop
 * the po->bind_lock and do a synchronize_net to make sure no
 * asynchronous packet processing paths still refer to the elements
 * of po->prot_hook.  If the sync parameter is false, it is the
 * callers responsibility to take care of this.
 */
static void __unregister_prot_hook(struct sock *sk, bool sync)
{
	struct packet_sock *po = pkt_sk(sk);

	po->running = 0;
D
David S. Miller 已提交
261 262 263 264
	if (po->fanout)
		__fanout_unlink(sk, po);
	else
		__dev_remove_pack(&po->prot_hook);
265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281
	__sock_put(sk);

	if (sync) {
		spin_unlock(&po->bind_lock);
		synchronize_net();
		spin_lock(&po->bind_lock);
	}
}

static void unregister_prot_hook(struct sock *sk, bool sync)
{
	struct packet_sock *po = pkt_sk(sk);

	if (po->running)
		__unregister_prot_hook(sk, sync);
}

282
static inline __pure struct page *pgv_to_page(void *addr)
283 284 285 286 287 288
{
	if (is_vmalloc_addr(addr))
		return vmalloc_to_page(addr);
	return virt_to_page(addr);
}

J
Johann Baudy 已提交
289
static void __packet_set_status(struct packet_sock *po, void *frame, int status)
L
Linus Torvalds 已提交
290
{
291 292 293 294 295
	union {
		struct tpacket_hdr *h1;
		struct tpacket2_hdr *h2;
		void *raw;
	} h;
L
Linus Torvalds 已提交
296

J
Johann Baudy 已提交
297
	h.raw = frame;
298 299
	switch (po->tp_version) {
	case TPACKET_V1:
J
Johann Baudy 已提交
300
		h.h1->tp_status = status;
301
		flush_dcache_page(pgv_to_page(&h.h1->tp_status));
302 303
		break;
	case TPACKET_V2:
J
Johann Baudy 已提交
304
		h.h2->tp_status = status;
305
		flush_dcache_page(pgv_to_page(&h.h2->tp_status));
306
		break;
307
	case TPACKET_V3:
J
Johann Baudy 已提交
308
	default:
309
		WARN(1, "TPACKET version not supported.\n");
J
Johann Baudy 已提交
310
		BUG();
311
	}
J
Johann Baudy 已提交
312 313

	smp_wmb();
314 315
}

J
Johann Baudy 已提交
316
static int __packet_get_status(struct packet_sock *po, void *frame)
317 318 319 320 321 322 323
{
	union {
		struct tpacket_hdr *h1;
		struct tpacket2_hdr *h2;
		void *raw;
	} h;

J
Johann Baudy 已提交
324 325
	smp_rmb();

326 327 328
	h.raw = frame;
	switch (po->tp_version) {
	case TPACKET_V1:
329
		flush_dcache_page(pgv_to_page(&h.h1->tp_status));
J
Johann Baudy 已提交
330
		return h.h1->tp_status;
331
	case TPACKET_V2:
332
		flush_dcache_page(pgv_to_page(&h.h2->tp_status));
J
Johann Baudy 已提交
333
		return h.h2->tp_status;
334
	case TPACKET_V3:
J
Johann Baudy 已提交
335
	default:
336
		WARN(1, "TPACKET version not supported.\n");
J
Johann Baudy 已提交
337 338
		BUG();
		return 0;
339
	}
L
Linus Torvalds 已提交
340
}
J
Johann Baudy 已提交
341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356

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;

357 358
	h.raw = rb->pg_vec[pg_vec_pos].buffer +
		(frame_offset * rb->frame_size);
J
Johann Baudy 已提交
359 360 361 362 363 364 365

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

	return h.raw;
}

366
static void *packet_current_frame(struct packet_sock *po,
J
Johann Baudy 已提交
367 368 369 370 371 372
		struct packet_ring_buffer *rb,
		int status)
{
	return packet_lookup_frame(po, rb, rb->head, status);
}

373
static void prb_del_retire_blk_timer(struct tpacket_kbdq_core *pkc)
374 375 376 377 378 379 380 381
{
	del_timer_sync(&pkc->retire_blk_timer);
}

static void prb_shutdown_retire_blk_timer(struct packet_sock *po,
		int tx_ring,
		struct sk_buff_head *rb_queue)
{
382
	struct tpacket_kbdq_core *pkc;
383 384 385 386 387 388 389 390 391 392 393

	pkc = tx_ring ? &po->tx_ring.prb_bdqc : &po->rx_ring.prb_bdqc;

	spin_lock(&rb_queue->lock);
	pkc->delete_blk_timer = 1;
	spin_unlock(&rb_queue->lock);

	prb_del_retire_blk_timer(pkc);
}

static void prb_init_blk_timer(struct packet_sock *po,
394
		struct tpacket_kbdq_core *pkc,
395 396 397 398 399 400 401 402 403 404
		void (*func) (unsigned long))
{
	init_timer(&pkc->retire_blk_timer);
	pkc->retire_blk_timer.data = (long)po;
	pkc->retire_blk_timer.function = func;
	pkc->retire_blk_timer.expires = jiffies;
}

static void prb_setup_retire_blk_timer(struct packet_sock *po, int tx_ring)
{
405
	struct tpacket_kbdq_core *pkc;
406 407 408 409 410 411 412 413 414 415 416 417 418

	if (tx_ring)
		BUG();

	pkc = tx_ring ? &po->tx_ring.prb_bdqc : &po->rx_ring.prb_bdqc;
	prb_init_blk_timer(po, pkc, prb_retire_rx_blk_timer_expired);
}

static int prb_calc_retire_blk_tmo(struct packet_sock *po,
				int blk_size_in_bytes)
{
	struct net_device *dev;
	unsigned int mbits = 0, msec = 0, div = 0, tmo = 0;
419 420
	struct ethtool_cmd ecmd;
	int err;
421
	u32 speed;
422

423 424 425 426
	rtnl_lock();
	dev = __dev_get_by_index(sock_net(&po->sk), po->ifindex);
	if (unlikely(!dev)) {
		rtnl_unlock();
427
		return DEFAULT_PRB_RETIRE_TOV;
428 429
	}
	err = __ethtool_get_settings(dev, &ecmd);
430
	speed = ethtool_cmd_speed(&ecmd);
431 432 433 434 435 436
	rtnl_unlock();
	if (!err) {
		/*
		 * If the link speed is so slow you don't really
		 * need to worry about perf anyways
		 */
437
		if (speed < SPEED_1000 || speed == SPEED_UNKNOWN) {
438
			return DEFAULT_PRB_RETIRE_TOV;
439 440 441
		} else {
			msec = 1;
			div = speed / 1000;
442 443 444 445 446 447 448 449 450 451 452 453 454 455 456
		}
	}

	mbits = (blk_size_in_bytes * 8) / (1024 * 1024);

	if (div)
		mbits /= div;

	tmo = mbits * msec;

	if (div)
		return tmo+1;
	return tmo;
}

457
static void prb_init_ft_ops(struct tpacket_kbdq_core *p1,
458 459 460 461 462 463 464 465 466 467
			union tpacket_req_u *req_u)
{
	p1->feature_req_word = req_u->req3.tp_feature_req_word;
}

static void init_prb_bdqc(struct packet_sock *po,
			struct packet_ring_buffer *rb,
			struct pgv *pg_vec,
			union tpacket_req_u *req_u, int tx_ring)
{
468 469
	struct tpacket_kbdq_core *p1 = &rb->prb_bdqc;
	struct tpacket_block_desc *pbd;
470 471 472 473 474

	memset(p1, 0x0, sizeof(*p1));

	p1->knxt_seq_num = 1;
	p1->pkbdq = pg_vec;
475
	pbd = (struct tpacket_block_desc *)pg_vec[0].buffer;
J
Joe Perches 已提交
476
	p1->pkblk_start	= pg_vec[0].buffer;
477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498
	p1->kblk_size = req_u->req3.tp_block_size;
	p1->knum_blocks	= req_u->req3.tp_block_nr;
	p1->hdrlen = po->tp_hdrlen;
	p1->version = po->tp_version;
	p1->last_kactive_blk_num = 0;
	po->stats_u.stats3.tp_freeze_q_cnt = 0;
	if (req_u->req3.tp_retire_blk_tov)
		p1->retire_blk_tov = req_u->req3.tp_retire_blk_tov;
	else
		p1->retire_blk_tov = prb_calc_retire_blk_tmo(po,
						req_u->req3.tp_block_size);
	p1->tov_in_jiffies = msecs_to_jiffies(p1->retire_blk_tov);
	p1->blk_sizeof_priv = req_u->req3.tp_sizeof_priv;

	prb_init_ft_ops(p1, req_u);
	prb_setup_retire_blk_timer(po, tx_ring);
	prb_open_block(p1, pbd);
}

/*  Do NOT update the last_blk_num first.
 *  Assumes sk_buff_head lock is held.
 */
499
static void _prb_refresh_rx_retire_blk_timer(struct tpacket_kbdq_core *pkc)
500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531
{
	mod_timer(&pkc->retire_blk_timer,
			jiffies + pkc->tov_in_jiffies);
	pkc->last_kactive_blk_num = pkc->kactive_blk_num;
}

/*
 * Timer logic:
 * 1) We refresh the timer only when we open a block.
 *    By doing this we don't waste cycles refreshing the timer
 *	  on packet-by-packet basis.
 *
 * With a 1MB block-size, on a 1Gbps line, it will take
 * i) ~8 ms to fill a block + ii) memcpy etc.
 * In this cut we are not accounting for the memcpy time.
 *
 * So, if the user sets the 'tmo' to 10ms then the timer
 * will never fire while the block is still getting filled
 * (which is what we want). However, the user could choose
 * to close a block early and that's fine.
 *
 * But when the timer does fire, we check whether or not to refresh it.
 * Since the tmo granularity is in msecs, it is not too expensive
 * to refresh the timer, lets say every '8' msecs.
 * Either the user can set the 'tmo' or we can derive it based on
 * a) line-speed and b) block-size.
 * prb_calc_retire_blk_tmo() calculates the tmo.
 *
 */
static void prb_retire_rx_blk_timer_expired(unsigned long data)
{
	struct packet_sock *po = (struct packet_sock *)data;
532
	struct tpacket_kbdq_core *pkc = &po->rx_ring.prb_bdqc;
533
	unsigned int frozen;
534
	struct tpacket_block_desc *pbd;
535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597

	spin_lock(&po->sk.sk_receive_queue.lock);

	frozen = prb_queue_frozen(pkc);
	pbd = GET_CURR_PBLOCK_DESC_FROM_CORE(pkc);

	if (unlikely(pkc->delete_blk_timer))
		goto out;

	/* We only need to plug the race when the block is partially filled.
	 * tpacket_rcv:
	 *		lock(); increment BLOCK_NUM_PKTS; unlock()
	 *		copy_bits() is in progress ...
	 *		timer fires on other cpu:
	 *		we can't retire the current block because copy_bits
	 *		is in progress.
	 *
	 */
	if (BLOCK_NUM_PKTS(pbd)) {
		while (atomic_read(&pkc->blk_fill_in_prog)) {
			/* Waiting for skb_copy_bits to finish... */
			cpu_relax();
		}
	}

	if (pkc->last_kactive_blk_num == pkc->kactive_blk_num) {
		if (!frozen) {
			prb_retire_current_block(pkc, po, TP_STATUS_BLK_TMO);
			if (!prb_dispatch_next_block(pkc, po))
				goto refresh_timer;
			else
				goto out;
		} else {
			/* Case 1. Queue was frozen because user-space was
			 *	   lagging behind.
			 */
			if (prb_curr_blk_in_use(pkc, pbd)) {
				/*
				 * Ok, user-space is still behind.
				 * So just refresh the timer.
				 */
				goto refresh_timer;
			} else {
			       /* Case 2. queue was frozen,user-space caught up,
				* now the link went idle && the timer fired.
				* We don't have a block to close.So we open this
				* block and restart the timer.
				* opening a block thaws the queue,restarts timer
				* Thawing/timer-refresh is a side effect.
				*/
				prb_open_block(pkc, pbd);
				goto out;
			}
		}
	}

refresh_timer:
	_prb_refresh_rx_retire_blk_timer(pkc);

out:
	spin_unlock(&po->sk.sk_receive_queue.lock);
}

598
static void prb_flush_block(struct tpacket_kbdq_core *pkc1,
599
		struct tpacket_block_desc *pbd1, __u32 status)
600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640
{
	/* Flush everything minus the block header */

#if ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE == 1
	u8 *start, *end;

	start = (u8 *)pbd1;

	/* Skip the block header(we know header WILL fit in 4K) */
	start += PAGE_SIZE;

	end = (u8 *)PAGE_ALIGN((unsigned long)pkc1->pkblk_end);
	for (; start < end; start += PAGE_SIZE)
		flush_dcache_page(pgv_to_page(start));

	smp_wmb();
#endif

	/* Now update the block status. */

	BLOCK_STATUS(pbd1) = status;

	/* Flush the block header */

#if ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE == 1
	start = (u8 *)pbd1;
	flush_dcache_page(pgv_to_page(start));

	smp_wmb();
#endif
}

/*
 * Side effect:
 *
 * 1) flush the block
 * 2) Increment active_blk_num
 *
 * Note:We DONT refresh the timer on purpose.
 *	Because almost always the next block will be opened.
 */
641 642
static void prb_close_block(struct tpacket_kbdq_core *pkc1,
		struct tpacket_block_desc *pbd1,
643 644 645 646 647
		struct packet_sock *po, unsigned int stat)
{
	__u32 status = TP_STATUS_USER | stat;

	struct tpacket3_hdr *last_pkt;
648
	struct tpacket_hdr_v1 *h1 = &pbd1->hdr.bh1;
649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675

	if (po->stats.tp_drops)
		status |= TP_STATUS_LOSING;

	last_pkt = (struct tpacket3_hdr *)pkc1->prev;
	last_pkt->tp_next_offset = 0;

	/* Get the ts of the last pkt */
	if (BLOCK_NUM_PKTS(pbd1)) {
		h1->ts_last_pkt.ts_sec = last_pkt->tp_sec;
		h1->ts_last_pkt.ts_nsec	= last_pkt->tp_nsec;
	} else {
		/* Ok, we tmo'd - so get the current time */
		struct timespec ts;
		getnstimeofday(&ts);
		h1->ts_last_pkt.ts_sec = ts.tv_sec;
		h1->ts_last_pkt.ts_nsec	= ts.tv_nsec;
	}

	smp_wmb();

	/* Flush the block */
	prb_flush_block(pkc1, pbd1, status);

	pkc1->kactive_blk_num = GET_NEXT_PRB_BLK_NUM(pkc1);
}

676
static void prb_thaw_queue(struct tpacket_kbdq_core *pkc)
677 678 679 680 681 682 683 684 685 686 687
{
	pkc->reset_pending_on_curr_blk = 0;
}

/*
 * Side effect of opening a block:
 *
 * 1) prb_queue is thawed.
 * 2) retire_blk_timer is refreshed.
 *
 */
688 689
static void prb_open_block(struct tpacket_kbdq_core *pkc1,
	struct tpacket_block_desc *pbd1)
690 691
{
	struct timespec ts;
692
	struct tpacket_hdr_v1 *h1 = &pbd1->hdr.bh1;
693 694 695 696 697 698 699 700 701 702 703 704 705 706 707

	smp_rmb();

	if (likely(TP_STATUS_KERNEL == BLOCK_STATUS(pbd1))) {

		/* We could have just memset this but we will lose the
		 * flexibility of making the priv area sticky
		 */
		BLOCK_SNUM(pbd1) = pkc1->knxt_seq_num++;
		BLOCK_NUM_PKTS(pbd1) = 0;
		BLOCK_LEN(pbd1) = BLK_PLUS_PRIV(pkc1->blk_sizeof_priv);
		getnstimeofday(&ts);
		h1->ts_first_pkt.ts_sec = ts.tv_sec;
		h1->ts_first_pkt.ts_nsec = ts.tv_nsec;
		pkc1->pkblk_start = (char *)pbd1;
J
Joe Perches 已提交
708
		pkc1->nxt_offset = pkc1->pkblk_start + BLK_PLUS_PRIV(pkc1->blk_sizeof_priv);
709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750
		BLOCK_O2FP(pbd1) = (__u32)BLK_PLUS_PRIV(pkc1->blk_sizeof_priv);
		BLOCK_O2PRIV(pbd1) = BLK_HDR_LEN;
		pbd1->version = pkc1->version;
		pkc1->prev = pkc1->nxt_offset;
		pkc1->pkblk_end = pkc1->pkblk_start + pkc1->kblk_size;
		prb_thaw_queue(pkc1);
		_prb_refresh_rx_retire_blk_timer(pkc1);

		smp_wmb();

		return;
	}

	WARN(1, "ERROR block:%p is NOT FREE status:%d kactive_blk_num:%d\n",
		pbd1, BLOCK_STATUS(pbd1), pkc1->kactive_blk_num);
	dump_stack();
	BUG();
}

/*
 * Queue freeze logic:
 * 1) Assume tp_block_nr = 8 blocks.
 * 2) At time 't0', user opens Rx ring.
 * 3) Some time past 't0', kernel starts filling blocks starting from 0 .. 7
 * 4) user-space is either sleeping or processing block '0'.
 * 5) tpacket_rcv is currently filling block '7', since there is no space left,
 *    it will close block-7,loop around and try to fill block '0'.
 *    call-flow:
 *    __packet_lookup_frame_in_block
 *      prb_retire_current_block()
 *      prb_dispatch_next_block()
 *        |->(BLOCK_STATUS == USER) evaluates to true
 *    5.1) Since block-0 is currently in-use, we just freeze the queue.
 * 6) Now there are two cases:
 *    6.1) Link goes idle right after the queue is frozen.
 *         But remember, the last open_block() refreshed the timer.
 *         When this timer expires,it will refresh itself so that we can
 *         re-open block-0 in near future.
 *    6.2) Link is busy and keeps on receiving packets. This is a simple
 *         case and __packet_lookup_frame_in_block will check if block-0
 *         is free and can now be re-used.
 */
751
static void prb_freeze_queue(struct tpacket_kbdq_core *pkc,
752 753 754 755 756 757 758 759 760 761 762 763 764 765
				  struct packet_sock *po)
{
	pkc->reset_pending_on_curr_blk = 1;
	po->stats_u.stats3.tp_freeze_q_cnt++;
}

#define TOTAL_PKT_LEN_INCL_ALIGN(length) (ALIGN((length), V3_ALIGNMENT))

/*
 * If the next block is free then we will dispatch it
 * and return a good offset.
 * Else, we will freeze the queue.
 * So, caller must check the return value.
 */
766
static void *prb_dispatch_next_block(struct tpacket_kbdq_core *pkc,
767 768
		struct packet_sock *po)
{
769
	struct tpacket_block_desc *pbd;
770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790

	smp_rmb();

	/* 1. Get current block num */
	pbd = GET_CURR_PBLOCK_DESC_FROM_CORE(pkc);

	/* 2. If this block is currently in_use then freeze the queue */
	if (TP_STATUS_USER & BLOCK_STATUS(pbd)) {
		prb_freeze_queue(pkc, po);
		return NULL;
	}

	/*
	 * 3.
	 * open this block and return the offset where the first packet
	 * needs to get stored.
	 */
	prb_open_block(pkc, pbd);
	return (void *)pkc->nxt_offset;
}

791
static void prb_retire_current_block(struct tpacket_kbdq_core *pkc,
792 793
		struct packet_sock *po, unsigned int status)
{
794
	struct tpacket_block_desc *pbd = GET_CURR_PBLOCK_DESC_FROM_CORE(pkc);
795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821

	/* retire/close the current block */
	if (likely(TP_STATUS_KERNEL == BLOCK_STATUS(pbd))) {
		/*
		 * Plug the case where copy_bits() is in progress on
		 * cpu-0 and tpacket_rcv() got invoked on cpu-1, didn't
		 * have space to copy the pkt in the current block and
		 * called prb_retire_current_block()
		 *
		 * We don't need to worry about the TMO case because
		 * the timer-handler already handled this case.
		 */
		if (!(status & TP_STATUS_BLK_TMO)) {
			while (atomic_read(&pkc->blk_fill_in_prog)) {
				/* Waiting for skb_copy_bits to finish... */
				cpu_relax();
			}
		}
		prb_close_block(pkc, pbd, po, status);
		return;
	}

	WARN(1, "ERROR-pbd[%d]:%p\n", pkc->kactive_blk_num, pbd);
	dump_stack();
	BUG();
}

822
static int prb_curr_blk_in_use(struct tpacket_kbdq_core *pkc,
823
				      struct tpacket_block_desc *pbd)
824 825 826 827
{
	return TP_STATUS_USER & BLOCK_STATUS(pbd);
}

828
static int prb_queue_frozen(struct tpacket_kbdq_core *pkc)
829 830 831 832
{
	return pkc->reset_pending_on_curr_blk;
}

833
static void prb_clear_blk_fill_status(struct packet_ring_buffer *rb)
834
{
835
	struct tpacket_kbdq_core *pkc  = GET_PBDQC_FROM_RB(rb);
836 837 838
	atomic_dec(&pkc->blk_fill_in_prog);
}

839
static void prb_fill_rxhash(struct tpacket_kbdq_core *pkc,
840 841 842 843 844
			struct tpacket3_hdr *ppd)
{
	ppd->hv1.tp_rxhash = skb_get_rxhash(pkc->skb);
}

845
static void prb_clear_rxhash(struct tpacket_kbdq_core *pkc,
846 847 848 849 850
			struct tpacket3_hdr *ppd)
{
	ppd->hv1.tp_rxhash = 0;
}

851
static void prb_fill_vlan_info(struct tpacket_kbdq_core *pkc,
852 853 854 855 856 857
			struct tpacket3_hdr *ppd)
{
	if (vlan_tx_tag_present(pkc->skb)) {
		ppd->hv1.tp_vlan_tci = vlan_tx_tag_get(pkc->skb);
		ppd->tp_status = TP_STATUS_VLAN_VALID;
	} else {
858 859
		ppd->hv1.tp_vlan_tci = 0;
		ppd->tp_status = TP_STATUS_AVAILABLE;
860 861 862
	}
}

863
static void prb_run_all_ft_ops(struct tpacket_kbdq_core *pkc,
864 865 866 867 868 869 870 871 872 873
			struct tpacket3_hdr *ppd)
{
	prb_fill_vlan_info(pkc, ppd);

	if (pkc->feature_req_word & TP_FT_REQ_FILL_RXHASH)
		prb_fill_rxhash(pkc, ppd);
	else
		prb_clear_rxhash(pkc, ppd);
}

874
static void prb_fill_curr_block(char *curr,
875 876
				struct tpacket_kbdq_core *pkc,
				struct tpacket_block_desc *pbd,
877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897
				unsigned int len)
{
	struct tpacket3_hdr *ppd;

	ppd  = (struct tpacket3_hdr *)curr;
	ppd->tp_next_offset = TOTAL_PKT_LEN_INCL_ALIGN(len);
	pkc->prev = curr;
	pkc->nxt_offset += TOTAL_PKT_LEN_INCL_ALIGN(len);
	BLOCK_LEN(pbd) += TOTAL_PKT_LEN_INCL_ALIGN(len);
	BLOCK_NUM_PKTS(pbd) += 1;
	atomic_inc(&pkc->blk_fill_in_prog);
	prb_run_all_ft_ops(pkc, ppd);
}

/* Assumes caller has the sk->rx_queue.lock */
static void *__packet_lookup_frame_in_block(struct packet_sock *po,
					    struct sk_buff *skb,
						int status,
					    unsigned int len
					    )
{
898 899
	struct tpacket_kbdq_core *pkc;
	struct tpacket_block_desc *pbd;
900 901
	char *curr, *end;

J
Joe Perches 已提交
902
	pkc = GET_PBDQC_FROM_RB(&po->rx_ring);
903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927
	pbd = GET_CURR_PBLOCK_DESC_FROM_CORE(pkc);

	/* Queue is frozen when user space is lagging behind */
	if (prb_queue_frozen(pkc)) {
		/*
		 * Check if that last block which caused the queue to freeze,
		 * is still in_use by user-space.
		 */
		if (prb_curr_blk_in_use(pkc, pbd)) {
			/* Can't record this packet */
			return NULL;
		} else {
			/*
			 * Ok, the block was released by user-space.
			 * Now let's open that block.
			 * opening a block also thaws the queue.
			 * Thawing is a side effect.
			 */
			prb_open_block(pkc, pbd);
		}
	}

	smp_mb();
	curr = pkc->nxt_offset;
	pkc->skb = skb;
J
Joe Perches 已提交
928
	end = (char *)pbd + pkc->kblk_size;
929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953

	/* first try the current block */
	if (curr+TOTAL_PKT_LEN_INCL_ALIGN(len) < end) {
		prb_fill_curr_block(curr, pkc, pbd, len);
		return (void *)curr;
	}

	/* Ok, close the current block */
	prb_retire_current_block(pkc, po, 0);

	/* Now, try to dispatch the next block */
	curr = (char *)prb_dispatch_next_block(pkc, po);
	if (curr) {
		pbd = GET_CURR_PBLOCK_DESC_FROM_CORE(pkc);
		prb_fill_curr_block(curr, pkc, pbd, len);
		return (void *)curr;
	}

	/*
	 * No free blocks are available.user_space hasn't caught up yet.
	 * Queue was just frozen and now this packet will get dropped.
	 */
	return NULL;
}

954
static void *packet_current_rx_frame(struct packet_sock *po,
955 956 957 958 959 960 961 962 963 964 965 966 967 968 969
					    struct sk_buff *skb,
					    int status, unsigned int len)
{
	char *curr = NULL;
	switch (po->tp_version) {
	case TPACKET_V1:
	case TPACKET_V2:
		curr = packet_lookup_frame(po, &po->rx_ring,
					po->rx_ring.head, status);
		return curr;
	case TPACKET_V3:
		return __packet_lookup_frame_in_block(po, skb, status, len);
	default:
		WARN(1, "TPACKET version not supported\n");
		BUG();
970
		return NULL;
971 972 973
	}
}

974
static void *prb_lookup_block(struct packet_sock *po,
975 976 977 978
				     struct packet_ring_buffer *rb,
				     unsigned int previous,
				     int status)
{
979 980
	struct tpacket_kbdq_core *pkc  = GET_PBDQC_FROM_RB(rb);
	struct tpacket_block_desc *pbd = GET_PBLOCK_DESC(pkc, previous);
981 982 983 984 985 986

	if (status != BLOCK_STATUS(pbd))
		return NULL;
	return pbd;
}

987
static int prb_previous_blk_num(struct packet_ring_buffer *rb)
988 989 990 991 992 993 994 995 996 997
{
	unsigned int prev;
	if (rb->prb_bdqc.kactive_blk_num)
		prev = rb->prb_bdqc.kactive_blk_num-1;
	else
		prev = rb->prb_bdqc.knum_blocks-1;
	return prev;
}

/* Assumes caller has held the rx_queue.lock */
998
static void *__prb_previous_block(struct packet_sock *po,
999 1000 1001 1002 1003 1004 1005
					 struct packet_ring_buffer *rb,
					 int status)
{
	unsigned int previous = prb_previous_blk_num(rb);
	return prb_lookup_block(po, rb, previous, status);
}

1006
static void *packet_previous_rx_frame(struct packet_sock *po,
1007 1008 1009 1010 1011 1012 1013 1014 1015
					     struct packet_ring_buffer *rb,
					     int status)
{
	if (po->tp_version <= TPACKET_V2)
		return packet_previous_frame(po, rb, status);

	return __prb_previous_block(po, rb, status);
}

1016
static void packet_increment_rx_head(struct packet_sock *po,
1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030
					    struct packet_ring_buffer *rb)
{
	switch (po->tp_version) {
	case TPACKET_V1:
	case TPACKET_V2:
		return packet_increment_head(rb);
	case TPACKET_V3:
	default:
		WARN(1, "TPACKET version not supported.\n");
		BUG();
		return;
	}
}

1031
static void *packet_previous_frame(struct packet_sock *po,
J
Johann Baudy 已提交
1032 1033 1034 1035 1036 1037 1038
		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);
}

1039
static void packet_increment_head(struct packet_ring_buffer *buff)
J
Johann Baudy 已提交
1040 1041 1042 1043
{
	buff->head = buff->head != buff->frame_max ? buff->head+1 : 0;
}

L
Linus Torvalds 已提交
1044 1045
static void packet_sock_destruct(struct sock *sk)
{
1046 1047
	skb_queue_purge(&sk->sk_error_queue);

1048 1049
	WARN_ON(atomic_read(&sk->sk_rmem_alloc));
	WARN_ON(atomic_read(&sk->sk_wmem_alloc));
L
Linus Torvalds 已提交
1050 1051

	if (!sock_flag(sk, SOCK_DEAD)) {
E
Eric Dumazet 已提交
1052
		pr_err("Attempt to release alive packet socket: %p\n", sk);
L
Linus Torvalds 已提交
1053 1054 1055
		return;
	}

1056
	sk_refcnt_debug_dec(sk);
L
Linus Torvalds 已提交
1057 1058
}

D
David S. Miller 已提交
1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088
static int fanout_rr_next(struct packet_fanout *f, unsigned int num)
{
	int x = atomic_read(&f->rr_cur) + 1;

	if (x >= num)
		x = 0;

	return x;
}

static struct sock *fanout_demux_hash(struct packet_fanout *f, struct sk_buff *skb, unsigned int num)
{
	u32 idx, hash = skb->rxhash;

	idx = ((u64)hash * num) >> 32;

	return f->arr[idx];
}

static struct sock *fanout_demux_lb(struct packet_fanout *f, struct sk_buff *skb, unsigned int num)
{
	int cur, old;

	cur = atomic_read(&f->rr_cur);
	while ((old = atomic_cmpxchg(&f->rr_cur, cur,
				     fanout_rr_next(f, num))) != cur)
		cur = old;
	return f->arr[cur];
}

1089 1090 1091 1092 1093 1094 1095 1096 1097
static struct sock *fanout_demux_cpu(struct packet_fanout *f, struct sk_buff *skb, unsigned int num)
{
	unsigned int cpu = smp_processor_id();

	return f->arr[cpu % num];
}

static int packet_rcv_fanout(struct sk_buff *skb, struct net_device *dev,
			     struct packet_type *pt, struct net_device *orig_dev)
D
David S. Miller 已提交
1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109
{
	struct packet_fanout *f = pt->af_packet_priv;
	unsigned int num = f->num_members;
	struct packet_sock *po;
	struct sock *sk;

	if (!net_eq(dev_net(dev), read_pnet(&f->net)) ||
	    !num) {
		kfree_skb(skb);
		return 0;
	}

1110 1111 1112 1113
	switch (f->type) {
	case PACKET_FANOUT_HASH:
	default:
		if (f->defrag) {
1114
			skb = ip_check_defrag(skb, IP_DEFRAG_AF_PACKET);
1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126
			if (!skb)
				return 0;
		}
		skb_get_rxhash(skb);
		sk = fanout_demux_hash(f, skb, num);
		break;
	case PACKET_FANOUT_LB:
		sk = fanout_demux_lb(f, skb, num);
		break;
	case PACKET_FANOUT_CPU:
		sk = fanout_demux_cpu(f, skb, num);
		break;
D
David S. Miller 已提交
1127 1128 1129 1130 1131 1132 1133
	}

	po = pkt_sk(sk);

	return po->prot_hook.func(skb, dev, &po->prot_hook, orig_dev);
}

1134 1135
DEFINE_MUTEX(fanout_mutex);
EXPORT_SYMBOL_GPL(fanout_mutex);
D
David S. Miller 已提交
1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164
static LIST_HEAD(fanout_list);

static void __fanout_link(struct sock *sk, struct packet_sock *po)
{
	struct packet_fanout *f = po->fanout;

	spin_lock(&f->lock);
	f->arr[f->num_members] = sk;
	smp_wmb();
	f->num_members++;
	spin_unlock(&f->lock);
}

static void __fanout_unlink(struct sock *sk, struct packet_sock *po)
{
	struct packet_fanout *f = po->fanout;
	int i;

	spin_lock(&f->lock);
	for (i = 0; i < f->num_members; i++) {
		if (f->arr[i] == sk)
			break;
	}
	BUG_ON(i >= f->num_members);
	f->arr[i] = f->arr[f->num_members - 1];
	f->num_members--;
	spin_unlock(&f->lock);
}

1165
static bool match_fanout_group(struct packet_type *ptype, struct sock * sk)
1166 1167 1168 1169 1170 1171 1172
{
	if (ptype->af_packet_priv == (void*)((struct packet_sock *)sk)->fanout)
		return true;

	return false;
}

1173
static int fanout_add(struct sock *sk, u16 id, u16 type_flags)
D
David S. Miller 已提交
1174 1175 1176
{
	struct packet_sock *po = pkt_sk(sk);
	struct packet_fanout *f, *match;
1177 1178
	u8 type = type_flags & 0xff;
	u8 defrag = (type_flags & PACKET_FANOUT_FLAG_DEFRAG) ? 1 : 0;
D
David S. Miller 已提交
1179 1180 1181 1182 1183
	int err;

	switch (type) {
	case PACKET_FANOUT_HASH:
	case PACKET_FANOUT_LB:
1184
	case PACKET_FANOUT_CPU:
D
David S. Miller 已提交
1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204
		break;
	default:
		return -EINVAL;
	}

	if (!po->running)
		return -EINVAL;

	if (po->fanout)
		return -EALREADY;

	mutex_lock(&fanout_mutex);
	match = NULL;
	list_for_each_entry(f, &fanout_list, list) {
		if (f->id == id &&
		    read_pnet(&f->net) == sock_net(sk)) {
			match = f;
			break;
		}
	}
E
Eric Dumazet 已提交
1205
	err = -EINVAL;
1206
	if (match && match->defrag != defrag)
E
Eric Dumazet 已提交
1207
		goto out;
D
David S. Miller 已提交
1208
	if (!match) {
E
Eric Dumazet 已提交
1209
		err = -ENOMEM;
D
David S. Miller 已提交
1210
		match = kzalloc(sizeof(*match), GFP_KERNEL);
E
Eric Dumazet 已提交
1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224
		if (!match)
			goto out;
		write_pnet(&match->net, sock_net(sk));
		match->id = id;
		match->type = type;
		match->defrag = defrag;
		atomic_set(&match->rr_cur, 0);
		INIT_LIST_HEAD(&match->list);
		spin_lock_init(&match->lock);
		atomic_set(&match->sk_ref, 0);
		match->prot_hook.type = po->prot_hook.type;
		match->prot_hook.dev = po->prot_hook.dev;
		match->prot_hook.func = packet_rcv_fanout;
		match->prot_hook.af_packet_priv = match;
1225
		match->prot_hook.id_match = match_fanout_group;
E
Eric Dumazet 已提交
1226 1227
		dev_add_pack(&match->prot_hook);
		list_add(&match->list, &fanout_list);
D
David S. Miller 已提交
1228
	}
E
Eric Dumazet 已提交
1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239
	err = -EINVAL;
	if (match->type == type &&
	    match->prot_hook.type == po->prot_hook.type &&
	    match->prot_hook.dev == po->prot_hook.dev) {
		err = -ENOSPC;
		if (atomic_read(&match->sk_ref) < PACKET_FANOUT_MAX) {
			__dev_remove_pack(&po->prot_hook);
			po->fanout = match;
			atomic_inc(&match->sk_ref);
			__fanout_link(sk, po);
			err = 0;
D
David S. Miller 已提交
1240 1241
		}
	}
E
Eric Dumazet 已提交
1242
out:
D
David S. Miller 已提交
1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255
	mutex_unlock(&fanout_mutex);
	return err;
}

static void fanout_release(struct sock *sk)
{
	struct packet_sock *po = pkt_sk(sk);
	struct packet_fanout *f;

	f = po->fanout;
	if (!f)
		return;

1256
	mutex_lock(&fanout_mutex);
D
David S. Miller 已提交
1257 1258 1259 1260 1261 1262 1263 1264 1265
	po->fanout = NULL;

	if (atomic_dec_and_test(&f->sk_ref)) {
		list_del(&f->list);
		dev_remove_pack(&f->prot_hook);
		kfree(f);
	}
	mutex_unlock(&fanout_mutex);
}
L
Linus Torvalds 已提交
1266

1267
static const struct proto_ops packet_ops;
L
Linus Torvalds 已提交
1268

1269
static const struct proto_ops packet_ops_spkt;
L
Linus Torvalds 已提交
1270

E
Eric Dumazet 已提交
1271 1272
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
1273 1274 1275 1276 1277 1278 1279 1280 1281 1282
{
	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;
1283

L
Linus Torvalds 已提交
1284 1285 1286 1287 1288 1289 1290
	/*
	 *	Yank back the headers [hope the device set this
	 *	right or kerboom...]
	 *
	 *	Incoming packets have ll header pulled,
	 *	push it back.
	 *
1291
	 *	For outgoing ones skb->data == skb_mac_header(skb)
L
Linus Torvalds 已提交
1292 1293 1294 1295 1296 1297
	 *	so that this procedure is noop.
	 */

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

O
Octavian Purdila 已提交
1298
	if (!net_eq(dev_net(dev), sock_net(sk)))
1299 1300
		goto out;

E
Eric Dumazet 已提交
1301 1302
	skb = skb_share_check(skb, GFP_ATOMIC);
	if (skb == NULL)
L
Linus Torvalds 已提交
1303 1304 1305
		goto oom;

	/* drop any routing info */
E
Eric Dumazet 已提交
1306
	skb_dst_drop(skb);
L
Linus Torvalds 已提交
1307

1308 1309 1310
	/* drop conntrack reference */
	nf_reset(skb);

1311
	spkt = &PACKET_SKB_CB(skb)->sa.pkt;
L
Linus Torvalds 已提交
1312

1313
	skb_push(skb, skb->data - skb_mac_header(skb));
L
Linus Torvalds 已提交
1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327

	/*
	 *	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
1328
	if (sock_queue_rcv_skb(sk, skb) == 0)
L
Linus Torvalds 已提交
1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341
		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
 */
1342

L
Linus Torvalds 已提交
1343 1344 1345 1346
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
1347
	struct sockaddr_pkt *saddr = (struct sockaddr_pkt *)msg->msg_name;
1348
	struct sk_buff *skb = NULL;
L
Linus Torvalds 已提交
1349
	struct net_device *dev;
E
Eric Dumazet 已提交
1350
	__be16 proto = 0;
L
Linus Torvalds 已提交
1351
	int err;
1352
	int extra_len = 0;
1353

L
Linus Torvalds 已提交
1354
	/*
1355
	 *	Get and verify the address.
L
Linus Torvalds 已提交
1356 1357
	 */

E
Eric Dumazet 已提交
1358
	if (saddr) {
L
Linus Torvalds 已提交
1359
		if (msg->msg_namelen < sizeof(struct sockaddr))
E
Eric Dumazet 已提交
1360 1361 1362 1363 1364
			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
1365 1366

	/*
1367
	 *	Find the device first to size check it
L
Linus Torvalds 已提交
1368 1369
	 */

1370
	saddr->spkt_device[sizeof(saddr->spkt_device) - 1] = 0;
1371
retry:
E
Eric Dumazet 已提交
1372 1373
	rcu_read_lock();
	dev = dev_get_by_name_rcu(sock_net(sk), saddr->spkt_device);
L
Linus Torvalds 已提交
1374 1375 1376
	err = -ENODEV;
	if (dev == NULL)
		goto out_unlock;
1377

1378 1379 1380 1381
	err = -ENETDOWN;
	if (!(dev->flags & IFF_UP))
		goto out_unlock;

L
Linus Torvalds 已提交
1382
	/*
E
Eric Dumazet 已提交
1383 1384
	 * 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
1385
	 */
1386

1387 1388 1389 1390 1391 1392 1393 1394
	if (unlikely(sock_flag(sk, SOCK_NOFCS))) {
		if (!netif_supports_nofcs(dev)) {
			err = -EPROTONOSUPPORT;
			goto out_unlock;
		}
		extra_len = 4; /* We're doing our own CRC */
	}

L
Linus Torvalds 已提交
1395
	err = -EMSGSIZE;
1396
	if (len > dev->mtu + dev->hard_header_len + VLAN_HLEN + extra_len)
L
Linus Torvalds 已提交
1397 1398
		goto out_unlock;

1399 1400
	if (!skb) {
		size_t reserved = LL_RESERVED_SPACE(dev);
1401
		int tlen = dev->needed_tailroom;
1402 1403 1404
		unsigned int hhlen = dev->header_ops ? dev->hard_header_len : 0;

		rcu_read_unlock();
1405
		skb = sock_wmalloc(sk, len + reserved + tlen, 0, GFP_KERNEL);
1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425
		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
1426 1427
	}

1428
	if (len > (dev->mtu + dev->hard_header_len + extra_len)) {
1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440
		/* Earlier code assumed this would be a VLAN pkt,
		 * double-check this now that we have the actual
		 * packet in hand.
		 */
		struct ethhdr *ehdr;
		skb_reset_mac_header(skb);
		ehdr = eth_hdr(skb);
		if (ehdr->h_proto != htons(ETH_P_8021Q)) {
			err = -EMSGSIZE;
			goto out_unlock;
		}
	}
1441

L
Linus Torvalds 已提交
1442 1443 1444
	skb->protocol = proto;
	skb->dev = dev;
	skb->priority = sk->sk_priority;
1445
	skb->mark = sk->sk_mark;
1446
	err = sock_tx_timestamp(sk, &skb_shinfo(skb)->tx_flags);
1447 1448
	if (err < 0)
		goto out_unlock;
L
Linus Torvalds 已提交
1449

1450 1451 1452
	if (unlikely(extra_len == 4))
		skb->no_fcs = 1;

L
Linus Torvalds 已提交
1453
	dev_queue_xmit(skb);
E
Eric Dumazet 已提交
1454
	rcu_read_unlock();
E
Eric Dumazet 已提交
1455
	return len;
L
Linus Torvalds 已提交
1456 1457

out_unlock:
E
Eric Dumazet 已提交
1458
	rcu_read_unlock();
1459 1460
out_free:
	kfree_skb(skb);
L
Linus Torvalds 已提交
1461 1462 1463
	return err;
}

1464
static unsigned int run_filter(const struct sk_buff *skb,
E
Eric Dumazet 已提交
1465
				      const struct sock *sk,
D
David S. Miller 已提交
1466
				      unsigned int res)
L
Linus Torvalds 已提交
1467 1468
{
	struct sk_filter *filter;
1469

1470 1471
	rcu_read_lock();
	filter = rcu_dereference(sk->sk_filter);
D
David S. Miller 已提交
1472
	if (filter != NULL)
1473
		res = SK_RUN_FILTER(filter, skb);
1474
	rcu_read_unlock();
L
Linus Torvalds 已提交
1475

D
David S. Miller 已提交
1476
	return res;
L
Linus Torvalds 已提交
1477 1478 1479
}

/*
E
Eric Dumazet 已提交
1480 1481 1482 1483 1484 1485 1486 1487 1488
 * 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
1489 1490
 */

E
Eric Dumazet 已提交
1491 1492
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
1493 1494 1495 1496
{
	struct sock *sk;
	struct sockaddr_ll *sll;
	struct packet_sock *po;
E
Eric Dumazet 已提交
1497
	u8 *skb_head = skb->data;
L
Linus Torvalds 已提交
1498
	int skb_len = skb->len;
D
David S. Miller 已提交
1499
	unsigned int snaplen, res;
L
Linus Torvalds 已提交
1500 1501 1502 1503 1504 1505 1506

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

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

O
Octavian Purdila 已提交
1507
	if (!net_eq(dev_net(dev), sock_net(sk)))
1508 1509
		goto drop;

L
Linus Torvalds 已提交
1510 1511
	skb->dev = dev;

1512
	if (dev->header_ops) {
L
Linus Torvalds 已提交
1513
		/* The device has an explicit notion of ll header,
E
Eric Dumazet 已提交
1514 1515 1516 1517 1518
		 * 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
1519 1520
		 */
		if (sk->sk_type != SOCK_DGRAM)
1521
			skb_push(skb, skb->data - skb_mac_header(skb));
L
Linus Torvalds 已提交
1522 1523
		else if (skb->pkt_type == PACKET_OUTGOING) {
			/* Special case: outgoing packets have ll header at head */
1524
			skb_pull(skb, skb_network_offset(skb));
L
Linus Torvalds 已提交
1525 1526 1527 1528 1529
		}
	}

	snaplen = skb->len;

D
David S. Miller 已提交
1530 1531
	res = run_filter(skb, sk, snaplen);
	if (!res)
1532
		goto drop_n_restore;
D
David S. Miller 已提交
1533 1534
	if (snaplen > res)
		snaplen = res;
L
Linus Torvalds 已提交
1535

E
Eric Dumazet 已提交
1536
	if (atomic_read(&sk->sk_rmem_alloc) >= sk->sk_rcvbuf)
L
Linus Torvalds 已提交
1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547
		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;
		}
1548
		consume_skb(skb);
L
Linus Torvalds 已提交
1549 1550 1551
		skb = nskb;
	}

1552 1553 1554 1555
	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
1556 1557 1558 1559
	sll->sll_family = AF_PACKET;
	sll->sll_hatype = dev->type;
	sll->sll_protocol = skb->protocol;
	sll->sll_pkttype = skb->pkt_type;
1560
	if (unlikely(po->origdev))
1561 1562 1563
		sll->sll_ifindex = orig_dev->ifindex;
	else
		sll->sll_ifindex = dev->ifindex;
L
Linus Torvalds 已提交
1564

S
Stephen Hemminger 已提交
1565
	sll->sll_halen = dev_parse_header(skb, sll->sll_addr);
L
Linus Torvalds 已提交
1566

1567
	PACKET_SKB_CB(skb)->origlen = skb->len;
1568

L
Linus Torvalds 已提交
1569 1570 1571 1572 1573
	if (pskb_trim(skb, snaplen))
		goto drop_n_acct;

	skb_set_owner_r(skb, sk);
	skb->dev = NULL;
E
Eric Dumazet 已提交
1574
	skb_dst_drop(skb);
L
Linus Torvalds 已提交
1575

1576 1577 1578
	/* drop conntrack reference */
	nf_reset(skb);

L
Linus Torvalds 已提交
1579 1580
	spin_lock(&sk->sk_receive_queue.lock);
	po->stats.tp_packets++;
1581
	skb->dropcount = atomic_read(&sk->sk_drops);
L
Linus Torvalds 已提交
1582 1583 1584 1585 1586 1587
	__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:
1588 1589 1590 1591
	spin_lock(&sk->sk_receive_queue.lock);
	po->stats.tp_drops++;
	atomic_inc(&sk->sk_drops);
	spin_unlock(&sk->sk_receive_queue.lock);
L
Linus Torvalds 已提交
1592 1593 1594 1595 1596 1597 1598

drop_n_restore:
	if (skb_head != skb->data && skb_shared(skb)) {
		skb->data = skb_head;
		skb->len = skb_len;
	}
drop:
1599
	consume_skb(skb);
L
Linus Torvalds 已提交
1600 1601 1602
	return 0;
}

E
Eric Dumazet 已提交
1603 1604
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
1605 1606 1607 1608
{
	struct sock *sk;
	struct packet_sock *po;
	struct sockaddr_ll *sll;
1609 1610 1611
	union {
		struct tpacket_hdr *h1;
		struct tpacket2_hdr *h2;
1612
		struct tpacket3_hdr *h3;
1613 1614
		void *raw;
	} h;
E
Eric Dumazet 已提交
1615
	u8 *skb_head = skb->data;
L
Linus Torvalds 已提交
1616
	int skb_len = skb->len;
D
David S. Miller 已提交
1617
	unsigned int snaplen, res;
1618
	unsigned long status = TP_STATUS_USER;
1619
	unsigned short macoff, netoff, hdrlen;
L
Linus Torvalds 已提交
1620
	struct sk_buff *copy_skb = NULL;
1621
	struct timeval tv;
1622
	struct timespec ts;
1623
	struct skb_shared_hwtstamps *shhwtstamps = skb_hwtstamps(skb);
L
Linus Torvalds 已提交
1624 1625 1626 1627 1628 1629 1630

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

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

O
Octavian Purdila 已提交
1631
	if (!net_eq(dev_net(dev), sock_net(sk)))
1632 1633
		goto drop;

1634
	if (dev->header_ops) {
L
Linus Torvalds 已提交
1635
		if (sk->sk_type != SOCK_DGRAM)
1636
			skb_push(skb, skb->data - skb_mac_header(skb));
L
Linus Torvalds 已提交
1637 1638
		else if (skb->pkt_type == PACKET_OUTGOING) {
			/* Special case: outgoing packets have ll header at head */
1639
			skb_pull(skb, skb_network_offset(skb));
L
Linus Torvalds 已提交
1640 1641 1642
		}
	}

1643 1644 1645
	if (skb->ip_summed == CHECKSUM_PARTIAL)
		status |= TP_STATUS_CSUMNOTREADY;

L
Linus Torvalds 已提交
1646 1647
	snaplen = skb->len;

D
David S. Miller 已提交
1648 1649
	res = run_filter(skb, sk, snaplen);
	if (!res)
1650
		goto drop_n_restore;
D
David S. Miller 已提交
1651 1652
	if (snaplen > res)
		snaplen = res;
L
Linus Torvalds 已提交
1653 1654

	if (sk->sk_type == SOCK_DGRAM) {
1655 1656
		macoff = netoff = TPACKET_ALIGN(po->tp_hdrlen) + 16 +
				  po->tp_reserve;
L
Linus Torvalds 已提交
1657
	} else {
1658
		unsigned int maclen = skb_network_offset(skb);
1659
		netoff = TPACKET_ALIGN(po->tp_hdrlen +
1660 1661
				       (maclen < 16 ? 16 : maclen)) +
			po->tp_reserve;
L
Linus Torvalds 已提交
1662 1663
		macoff = netoff - maclen;
	}
1664 1665 1666
	if (po->tp_version <= TPACKET_V2) {
		if (macoff + snaplen > po->rx_ring.frame_size) {
			if (po->copy_thresh &&
E
Eric Dumazet 已提交
1667
			    atomic_read(&sk->sk_rmem_alloc) < sk->sk_rcvbuf) {
1668 1669 1670 1671 1672 1673 1674 1675
				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);
L
Linus Torvalds 已提交
1676
			}
1677 1678 1679
			snaplen = po->rx_ring.frame_size - macoff;
			if ((int)snaplen < 0)
				snaplen = 0;
L
Linus Torvalds 已提交
1680 1681 1682
		}
	}
	spin_lock(&sk->sk_receive_queue.lock);
1683 1684
	h.raw = packet_current_rx_frame(po, skb,
					TP_STATUS_KERNEL, (macoff+snaplen));
1685
	if (!h.raw)
L
Linus Torvalds 已提交
1686
		goto ring_is_full;
1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697
	if (po->tp_version <= TPACKET_V2) {
		packet_increment_rx_head(po, &po->rx_ring);
	/*
	 * LOSING will be reported till you read the stats,
	 * because it's COR - Clear On Read.
	 * Anyways, moving it for V1/V2 only as V3 doesn't need this
	 * at packet level.
	 */
		if (po->stats.tp_drops)
			status |= TP_STATUS_LOSING;
	}
L
Linus Torvalds 已提交
1698 1699 1700 1701 1702 1703 1704
	po->stats.tp_packets++;
	if (copy_skb) {
		status |= TP_STATUS_COPY;
		__skb_queue_tail(&sk->sk_receive_queue, copy_skb);
	}
	spin_unlock(&sk->sk_receive_queue.lock);

1705
	skb_copy_bits(skb, 0, h.raw + macoff, snaplen);
L
Linus Torvalds 已提交
1706

1707 1708 1709 1710 1711 1712
	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;
1713 1714 1715 1716 1717 1718 1719
		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)
1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731
			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;
1732 1733 1734 1735 1736 1737 1738
		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)
1739 1740 1741 1742 1743
			ts = ktime_to_timespec(skb->tstamp);
		else
			getnstimeofday(&ts);
		h.h2->tp_sec = ts.tv_sec;
		h.h2->tp_nsec = ts.tv_nsec;
1744 1745 1746 1747 1748 1749
		if (vlan_tx_tag_present(skb)) {
			h.h2->tp_vlan_tci = vlan_tx_tag_get(skb);
			status |= TP_STATUS_VLAN_VALID;
		} else {
			h.h2->tp_vlan_tci = 0;
		}
1750
		h.h2->tp_padding = 0;
1751 1752
		hdrlen = sizeof(*h.h2);
		break;
1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775
	case TPACKET_V3:
		/* tp_nxt_offset,vlan are already populated above.
		 * So DONT clear those fields here
		 */
		h.h3->tp_status |= status;
		h.h3->tp_len = skb->len;
		h.h3->tp_snaplen = snaplen;
		h.h3->tp_mac = macoff;
		h.h3->tp_net = netoff;
		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)
			ts = ktime_to_timespec(skb->tstamp);
		else
			getnstimeofday(&ts);
		h.h3->tp_sec  = ts.tv_sec;
		h.h3->tp_nsec = ts.tv_nsec;
		hdrlen = sizeof(*h.h3);
		break;
1776 1777 1778
	default:
		BUG();
	}
L
Linus Torvalds 已提交
1779

1780
	sll = h.raw + TPACKET_ALIGN(hdrlen);
S
Stephen Hemminger 已提交
1781
	sll->sll_halen = dev_parse_header(skb, sll->sll_addr);
L
Linus Torvalds 已提交
1782 1783 1784 1785
	sll->sll_family = AF_PACKET;
	sll->sll_hatype = dev->type;
	sll->sll_protocol = skb->protocol;
	sll->sll_pkttype = skb->pkt_type;
1786
	if (unlikely(po->origdev))
1787 1788 1789
		sll->sll_ifindex = orig_dev->ifindex;
	else
		sll->sll_ifindex = dev->ifindex;
L
Linus Torvalds 已提交
1790

R
Ralf Baechle 已提交
1791
	smp_mb();
1792
#if ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE == 1
L
Linus Torvalds 已提交
1793
	{
1794 1795
		u8 *start, *end;

1796 1797 1798 1799 1800 1801
		if (po->tp_version <= TPACKET_V2) {
			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));
		}
1802
		smp_wmb();
L
Linus Torvalds 已提交
1803
	}
1804
#endif
1805 1806 1807 1808
	if (po->tp_version <= TPACKET_V2)
		__packet_set_status(po, h.raw, status);
	else
		prb_clear_blk_fill_status(&po->rx_ring);
L
Linus Torvalds 已提交
1809 1810 1811 1812 1813 1814 1815 1816 1817

	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:
1818
	kfree_skb(skb);
L
Linus Torvalds 已提交
1819 1820 1821 1822 1823 1824 1825
	return 0;

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

	sk->sk_data_ready(sk, 0);
1826
	kfree_skb(copy_skb);
L
Linus Torvalds 已提交
1827 1828 1829
	goto drop_n_restore;
}

J
Johann Baudy 已提交
1830 1831 1832
static void tpacket_destruct_skb(struct sk_buff *skb)
{
	struct packet_sock *po = pkt_sk(skb->sk);
E
Eric Dumazet 已提交
1833
	void *ph;
L
Linus Torvalds 已提交
1834

J
Johann Baudy 已提交
1835 1836 1837 1838 1839 1840 1841 1842 1843 1844
	if (likely(po->tx_ring.pg_vec)) {
		ph = skb_shinfo(skb)->destructor_arg;
		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
1845 1846
static int tpacket_fill_skb(struct packet_sock *po, struct sk_buff *skb,
		void *frame, struct net_device *dev, int size_max,
1847
		__be16 proto, unsigned char *addr, int hlen)
J
Johann Baudy 已提交
1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864
{
	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;
1865
	skb->mark = po->sk.sk_mark;
J
Johann Baudy 已提交
1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876
	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
1877
		pr_err("packet size is too long (%d > %d)\n", tp_len, size_max);
J
Johann Baudy 已提交
1878 1879 1880
		return -EMSGSIZE;
	}

1881
	skb_reserve(skb, hlen);
J
Johann Baudy 已提交
1882 1883 1884 1885 1886 1887 1888 1889 1890 1891
	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
1892
	} else if (dev->hard_header_len) {
J
Johann Baudy 已提交
1893 1894
		/* net device doesn't like empty head */
		if (unlikely(tp_len <= dev->hard_header_len)) {
E
Eric Dumazet 已提交
1895 1896
			pr_err("packet size is too short (%d < %d)\n",
			       tp_len, dev->hard_header_len);
J
Johann Baudy 已提交
1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922
			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;
	}

	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
1923 1924
			pr_err("Packet exceed the number of skb frags(%lu)\n",
			       MAX_SKB_FRAGS);
J
Johann Baudy 已提交
1925 1926 1927
			return -EFAULT;
		}

1928 1929
		page = pgv_to_page(data);
		data += len;
J
Johann Baudy 已提交
1930 1931
		flush_dcache_page(page);
		get_page(page);
1932
		skb_fill_page_desc(skb, nr_frags, page, offset, len);
J
Johann Baudy 已提交
1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946
		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 sk_buff *skb;
	struct net_device *dev;
	__be16 proto;
1947 1948
	bool need_rls_dev = false;
	int err, reserve = 0;
E
Eric Dumazet 已提交
1949 1950
	void *ph;
	struct sockaddr_ll *saddr = (struct sockaddr_ll *)msg->msg_name;
J
Johann Baudy 已提交
1951 1952 1953
	int tp_len, size_max;
	unsigned char *addr;
	int len_sum = 0;
1954
	int status = TP_STATUS_AVAILABLE;
1955
	int hlen, tlen;
J
Johann Baudy 已提交
1956 1957 1958 1959

	mutex_lock(&po->pg_vec_lock);

	if (saddr == NULL) {
1960
		dev = po->prot_hook.dev;
J
Johann Baudy 已提交
1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972
		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;
		proto	= saddr->sll_protocol;
		addr	= saddr->sll_addr;
1973 1974
		dev = dev_get_by_index(sock_net(&po->sk), saddr->sll_ifindex);
		need_rls_dev = true;
J
Johann Baudy 已提交
1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987
	}

	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
1988
		- (po->tp_hdrlen - sizeof(struct sockaddr_ll));
J
Johann Baudy 已提交
1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002

	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;
2003 2004
		hlen = LL_RESERVED_SPACE(dev);
		tlen = dev->needed_tailroom;
J
Johann Baudy 已提交
2005
		skb = sock_alloc_send_skb(&po->sk,
2006
				hlen + tlen + sizeof(struct sockaddr_ll),
J
Johann Baudy 已提交
2007 2008 2009 2010 2011 2012
				0, &err);

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

		tp_len = tpacket_fill_skb(po, skb, ph, dev, size_max, proto,
2013
				addr, hlen);
J
Johann Baudy 已提交
2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034

		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);
2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048
		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
2049 2050
		packet_increment_head(&po->tx_ring);
		len_sum += tp_len;
2051 2052 2053 2054
	} while (likely((ph != NULL) ||
			((!(msg->msg_flags & MSG_DONTWAIT)) &&
			 (atomic_read(&po->tx_ring.pending))))
		);
J
Johann Baudy 已提交
2055 2056 2057 2058 2059 2060 2061 2062

	err = len_sum;
	goto out_put;

out_status:
	__packet_set_status(po, ph, status);
	kfree_skb(skb);
out_put:
2063 2064
	if (need_rls_dev)
		dev_put(dev);
J
Johann Baudy 已提交
2065 2066 2067 2068 2069
out:
	mutex_unlock(&po->pg_vec_lock);
	return err;
}

2070 2071 2072 2073
static 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)
2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093
{
	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
2094
static int packet_snd(struct socket *sock,
L
Linus Torvalds 已提交
2095 2096 2097
			  struct msghdr *msg, size_t len)
{
	struct sock *sk = sock->sk;
E
Eric Dumazet 已提交
2098
	struct sockaddr_ll *saddr = (struct sockaddr_ll *)msg->msg_name;
L
Linus Torvalds 已提交
2099 2100
	struct sk_buff *skb;
	struct net_device *dev;
A
Al Viro 已提交
2101
	__be16 proto;
2102
	bool need_rls_dev = false;
L
Linus Torvalds 已提交
2103
	unsigned char *addr;
2104
	int err, reserve = 0;
2105 2106 2107 2108 2109
	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;
2110
	int hlen, tlen;
2111
	int extra_len = 0;
L
Linus Torvalds 已提交
2112 2113

	/*
2114
	 *	Get and verify the address.
L
Linus Torvalds 已提交
2115
	 */
2116

L
Linus Torvalds 已提交
2117
	if (saddr == NULL) {
2118
		dev = po->prot_hook.dev;
L
Linus Torvalds 已提交
2119 2120 2121 2122 2123 2124
		proto	= po->num;
		addr	= NULL;
	} else {
		err = -EINVAL;
		if (msg->msg_namelen < sizeof(struct sockaddr_ll))
			goto out;
2125 2126
		if (msg->msg_namelen < (saddr->sll_halen + offsetof(struct sockaddr_ll, sll_addr)))
			goto out;
L
Linus Torvalds 已提交
2127 2128
		proto	= saddr->sll_protocol;
		addr	= saddr->sll_addr;
2129 2130
		dev = dev_get_by_index(sock_net(sk), saddr->sll_ifindex);
		need_rls_dev = true;
L
Linus Torvalds 已提交
2131 2132 2133 2134 2135 2136 2137 2138
	}

	err = -ENXIO;
	if (dev == NULL)
		goto out_unlock;
	if (sock->type == SOCK_RAW)
		reserve = dev->hard_header_len;

2139 2140 2141 2142
	err = -ENETDOWN;
	if (!(dev->flags & IFF_UP))
		goto out_unlock;

2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190
	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;

		}
	}

2191 2192 2193 2194 2195 2196 2197 2198
	if (unlikely(sock_flag(sk, SOCK_NOFCS))) {
		if (!netif_supports_nofcs(dev)) {
			err = -EPROTONOSUPPORT;
			goto out_unlock;
		}
		extra_len = 4; /* We're doing our own CRC */
	}

L
Linus Torvalds 已提交
2199
	err = -EMSGSIZE;
2200
	if (!gso_type && (len > dev->mtu + reserve + VLAN_HLEN + extra_len))
L
Linus Torvalds 已提交
2201 2202
		goto out_unlock;

2203
	err = -ENOBUFS;
2204 2205 2206
	hlen = LL_RESERVED_SPACE(dev);
	tlen = dev->needed_tailroom;
	skb = packet_alloc_skb(sk, hlen + tlen, hlen, len, vnet_hdr.hdr_len,
2207
			       msg->msg_flags & MSG_DONTWAIT, &err);
E
Eric Dumazet 已提交
2208
	if (skb == NULL)
L
Linus Torvalds 已提交
2209 2210
		goto out_unlock;

2211
	skb_set_network_header(skb, reserve);
L
Linus Torvalds 已提交
2212

2213 2214
	err = -EINVAL;
	if (sock->type == SOCK_DGRAM &&
2215
	    (offset = dev_hard_header(skb, dev, ntohs(proto), addr, NULL, len)) < 0)
2216
		goto out_free;
L
Linus Torvalds 已提交
2217 2218

	/* Returns -EFAULT on error */
2219
	err = skb_copy_datagram_from_iovec(skb, offset, msg->msg_iov, 0, len);
L
Linus Torvalds 已提交
2220 2221
	if (err)
		goto out_free;
2222
	err = sock_tx_timestamp(sk, &skb_shinfo(skb)->tx_flags);
2223 2224
	if (err < 0)
		goto out_free;
L
Linus Torvalds 已提交
2225

2226
	if (!gso_type && (len > dev->mtu + reserve + extra_len)) {
2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239
		/* Earlier code assumed this would be a VLAN pkt,
		 * double-check this now that we have the actual
		 * packet in hand.
		 */
		struct ethhdr *ehdr;
		skb_reset_mac_header(skb);
		ehdr = eth_hdr(skb);
		if (ehdr->h_proto != htons(ETH_P_8021Q)) {
			err = -EMSGSIZE;
			goto out_free;
		}
	}

L
Linus Torvalds 已提交
2240 2241 2242
	skb->protocol = proto;
	skb->dev = dev;
	skb->priority = sk->sk_priority;
2243
	skb->mark = sk->sk_mark;
L
Linus Torvalds 已提交
2244

2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263
	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;
	}

2264 2265 2266
	if (unlikely(extra_len == 4))
		skb->no_fcs = 1;

L
Linus Torvalds 已提交
2267 2268 2269 2270 2271 2272 2273 2274
	/*
	 *	Now send it
	 */

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

2275 2276
	if (need_rls_dev)
		dev_put(dev);
L
Linus Torvalds 已提交
2277

E
Eric Dumazet 已提交
2278
	return len;
L
Linus Torvalds 已提交
2279 2280 2281 2282

out_free:
	kfree_skb(skb);
out_unlock:
2283
	if (dev && need_rls_dev)
L
Linus Torvalds 已提交
2284 2285 2286 2287 2288
		dev_put(dev);
out:
	return err;
}

J
Johann Baudy 已提交
2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299
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
2300 2301 2302 2303 2304 2305 2306 2307 2308
/*
 *	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;
2309
	struct net *net;
2310
	union tpacket_req_u req_u;
L
Linus Torvalds 已提交
2311 2312 2313 2314

	if (!sk)
		return 0;

2315
	net = sock_net(sk);
L
Linus Torvalds 已提交
2316 2317
	po = pkt_sk(sk);

2318
	mutex_lock(&net->packet.sklist_lock);
2319
	sk_del_node_init_rcu(sk);
2320 2321 2322
	mutex_unlock(&net->packet.sklist_lock);

	preempt_disable();
2323
	sock_prot_inuse_add(net, sk->sk_prot, -1);
2324
	preempt_enable();
L
Linus Torvalds 已提交
2325

2326
	spin_lock(&po->bind_lock);
2327
	unregister_prot_hook(sk, false);
2328 2329 2330 2331
	if (po->prot_hook.dev) {
		dev_put(po->prot_hook.dev);
		po->prot_hook.dev = NULL;
	}
2332
	spin_unlock(&po->bind_lock);
L
Linus Torvalds 已提交
2333 2334 2335

	packet_flush_mclist(sk);

2336
	memset(&req_u, 0, sizeof(req_u));
J
Johann Baudy 已提交
2337 2338

	if (po->rx_ring.pg_vec)
2339
		packet_set_ring(sk, &req_u, 1, 0);
J
Johann Baudy 已提交
2340 2341

	if (po->tx_ring.pg_vec)
2342
		packet_set_ring(sk, &req_u, 1, 1);
L
Linus Torvalds 已提交
2343

D
David S. Miller 已提交
2344 2345
	fanout_release(sk);

2346
	synchronize_net();
L
Linus Torvalds 已提交
2347 2348 2349 2350 2351 2352 2353 2354 2355
	/*
	 *	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);
2356
	sk_refcnt_debug_release(sk);
L
Linus Torvalds 已提交
2357 2358 2359 2360 2361 2362 2363 2364 2365

	sock_put(sk);
	return 0;
}

/*
 *	Attach a packet hook.
 */

A
Al Viro 已提交
2366
static int packet_do_bind(struct sock *sk, struct net_device *dev, __be16 protocol)
L
Linus Torvalds 已提交
2367 2368
{
	struct packet_sock *po = pkt_sk(sk);
D
David S. Miller 已提交
2369

2370 2371 2372 2373
	if (po->fanout) {
		if (dev)
			dev_put(dev);

D
David S. Miller 已提交
2374
		return -EINVAL;
2375
	}
L
Linus Torvalds 已提交
2376 2377 2378 2379

	lock_sock(sk);

	spin_lock(&po->bind_lock);
2380
	unregister_prot_hook(sk, true);
L
Linus Torvalds 已提交
2381 2382
	po->num = protocol;
	po->prot_hook.type = protocol;
2383 2384
	if (po->prot_hook.dev)
		dev_put(po->prot_hook.dev);
L
Linus Torvalds 已提交
2385 2386 2387 2388 2389 2390 2391
	po->prot_hook.dev = dev;

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

	if (protocol == 0)
		goto out_unlock;

2392
	if (!dev || (dev->flags & IFF_UP)) {
2393
		register_prot_hook(sk);
2394 2395 2396 2397
	} else {
		sk->sk_err = ENETDOWN;
		if (!sock_flag(sk, SOCK_DEAD))
			sk->sk_error_report(sk);
L
Linus Torvalds 已提交
2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409
	}

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
2410 2411
static int packet_bind_spkt(struct socket *sock, struct sockaddr *uaddr,
			    int addr_len)
L
Linus Torvalds 已提交
2412
{
E
Eric Dumazet 已提交
2413
	struct sock *sk = sock->sk;
L
Linus Torvalds 已提交
2414 2415 2416
	char name[15];
	struct net_device *dev;
	int err = -ENODEV;
2417

L
Linus Torvalds 已提交
2418 2419 2420
	/*
	 *	Check legality
	 */
2421

2422
	if (addr_len != sizeof(struct sockaddr))
L
Linus Torvalds 已提交
2423
		return -EINVAL;
E
Eric Dumazet 已提交
2424
	strlcpy(name, uaddr->sa_data, sizeof(name));
L
Linus Torvalds 已提交
2425

2426
	dev = dev_get_by_name(sock_net(sk), name);
2427
	if (dev)
L
Linus Torvalds 已提交
2428 2429 2430 2431 2432 2433
		err = packet_do_bind(sk, dev, pkt_sk(sk)->num);
	return err;
}

static int packet_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len)
{
E
Eric Dumazet 已提交
2434 2435
	struct sockaddr_ll *sll = (struct sockaddr_ll *)uaddr;
	struct sock *sk = sock->sk;
L
Linus Torvalds 已提交
2436 2437 2438 2439 2440 2441 2442
	struct net_device *dev = NULL;
	int err;


	/*
	 *	Check legality
	 */
2443

L
Linus Torvalds 已提交
2444 2445 2446 2447 2448 2449 2450
	if (addr_len < sizeof(struct sockaddr_ll))
		return -EINVAL;
	if (sll->sll_family != AF_PACKET)
		return -EINVAL;

	if (sll->sll_ifindex) {
		err = -ENODEV;
2451
		dev = dev_get_by_index(sock_net(sk), sll->sll_ifindex);
L
Linus Torvalds 已提交
2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467
		if (dev == NULL)
			goto out;
	}
	err = packet_do_bind(sk, dev, sll->sll_protocol ? : pkt_sk(sk)->num);

out:
	return err;
}

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

/*
2468
 *	Create a packet of type SOCK_PACKET.
L
Linus Torvalds 已提交
2469 2470
 */

2471 2472
static int packet_create(struct net *net, struct socket *sock, int protocol,
			 int kern)
L
Linus Torvalds 已提交
2473 2474 2475
{
	struct sock *sk;
	struct packet_sock *po;
A
Al Viro 已提交
2476
	__be16 proto = (__force __be16)protocol; /* weird, but documented */
L
Linus Torvalds 已提交
2477 2478 2479 2480
	int err;

	if (!capable(CAP_NET_RAW))
		return -EPERM;
2481 2482
	if (sock->type != SOCK_DGRAM && sock->type != SOCK_RAW &&
	    sock->type != SOCK_PACKET)
L
Linus Torvalds 已提交
2483 2484 2485 2486 2487
		return -ESOCKTNOSUPPORT;

	sock->state = SS_UNCONNECTED;

	err = -ENOBUFS;
2488
	sk = sk_alloc(net, PF_PACKET, GFP_KERNEL, &packet_proto);
L
Linus Torvalds 已提交
2489 2490 2491 2492 2493 2494
	if (sk == NULL)
		goto out;

	sock->ops = &packet_ops;
	if (sock->type == SOCK_PACKET)
		sock->ops = &packet_ops_spkt;
2495

L
Linus Torvalds 已提交
2496 2497 2498 2499
	sock_init_data(sock, sk);

	po = pkt_sk(sk);
	sk->sk_family = PF_PACKET;
A
Al Viro 已提交
2500
	po->num = proto;
L
Linus Torvalds 已提交
2501 2502

	sk->sk_destruct = packet_sock_destruct;
2503
	sk_refcnt_debug_inc(sk);
L
Linus Torvalds 已提交
2504 2505 2506 2507 2508 2509

	/*
	 *	Attach a protocol block
	 */

	spin_lock_init(&po->bind_lock);
2510
	mutex_init(&po->pg_vec_lock);
L
Linus Torvalds 已提交
2511
	po->prot_hook.func = packet_rcv;
2512

L
Linus Torvalds 已提交
2513 2514
	if (sock->type == SOCK_PACKET)
		po->prot_hook.func = packet_rcv_spkt;
2515

L
Linus Torvalds 已提交
2516 2517
	po->prot_hook.af_packet_priv = sk;

A
Al Viro 已提交
2518 2519
	if (proto) {
		po->prot_hook.type = proto;
2520
		register_prot_hook(sk);
L
Linus Torvalds 已提交
2521 2522
	}

2523
	mutex_lock(&net->packet.sklist_lock);
2524
	sk_add_node_rcu(sk, &net->packet.sklist);
2525 2526 2527
	mutex_unlock(&net->packet.sklist_lock);

	preempt_disable();
2528
	sock_prot_inuse_add(net, &packet_proto, 1);
2529
	preempt_enable();
2530

E
Eric Dumazet 已提交
2531
	return 0;
L
Linus Torvalds 已提交
2532 2533 2534 2535
out:
	return err;
}

2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568 2569 2570 2571 2572 2573 2574 2575 2576 2577 2578 2579 2580
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
2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591
/*
 *	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;
2592
	struct sockaddr_ll *sll;
2593
	int vnet_hdr_len = 0;
L
Linus Torvalds 已提交
2594 2595

	err = -EINVAL;
2596
	if (flags & ~(MSG_PEEK|MSG_DONTWAIT|MSG_TRUNC|MSG_CMSG_COMPAT|MSG_ERRQUEUE))
L
Linus Torvalds 已提交
2597 2598 2599 2600 2601 2602 2603 2604
		goto out;

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

2605 2606 2607 2608 2609
	if (flags & MSG_ERRQUEUE) {
		err = packet_recv_error(sk, msg, len);
		goto out;
	}

L
Linus Torvalds 已提交
2610 2611 2612 2613 2614 2615 2616 2617 2618
	/*
	 *	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
2619
	skb = skb_recv_datagram(sk, flags, flags & MSG_DONTWAIT, &err);
L
Linus Torvalds 已提交
2620 2621

	/*
2622
	 *	An error occurred so return it. Because skb_recv_datagram()
L
Linus Torvalds 已提交
2623 2624 2625 2626
	 *	handles the blocking we don't see and worry about blocking
	 *	retries.
	 */

2627
	if (skb == NULL)
L
Linus Torvalds 已提交
2628 2629
		goto out;

2630 2631 2632 2633 2634
	if (pkt_sk(sk)->has_vnet_hdr) {
		struct virtio_net_hdr vnet_hdr = { 0 };

		err = -EINVAL;
		vnet_hdr_len = sizeof(vnet_hdr);
2635
		if (len < vnet_hdr_len)
2636 2637
			goto out_free;

2638 2639
		len -= vnet_hdr_len;

2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 2662
		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;
2663
			vnet_hdr.csum_start = skb_checksum_start_offset(skb);
2664
			vnet_hdr.csum_offset = skb->csum_offset;
2665 2666
		} else if (skb->ip_summed == CHECKSUM_UNNECESSARY) {
			vnet_hdr.flags = VIRTIO_NET_HDR_F_DATA_VALID;
2667 2668 2669 2670 2671 2672 2673 2674
		} /* else everything is zero */

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

2675 2676 2677 2678 2679
	/*
	 *	If the address length field is there to be filled in, we fill
	 *	it in now.
	 */

2680
	sll = &PACKET_SKB_CB(skb)->sa.ll;
2681 2682 2683 2684 2685
	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
2686 2687 2688 2689 2690 2691
	/*
	 *	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
2692 2693 2694
	if (copied > len) {
		copied = len;
		msg->msg_flags |= MSG_TRUNC;
L
Linus Torvalds 已提交
2695 2696 2697 2698 2699 2700
	}

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

2701
	sock_recv_ts_and_drops(msg, sk, skb);
L
Linus Torvalds 已提交
2702 2703

	if (msg->msg_name)
2704 2705
		memcpy(msg->msg_name, &PACKET_SKB_CB(skb)->sa,
		       msg->msg_namelen);
L
Linus Torvalds 已提交
2706

2707
	if (pkt_sk(sk)->auxdata) {
2708 2709 2710 2711 2712 2713 2714 2715
		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;
2716
		aux.tp_net = skb_network_offset(skb);
2717 2718 2719 2720 2721 2722
		if (vlan_tx_tag_present(skb)) {
			aux.tp_vlan_tci = vlan_tx_tag_get(skb);
			aux.tp_status |= TP_STATUS_VLAN_VALID;
		} else {
			aux.tp_vlan_tci = 0;
		}
2723
		aux.tp_padding = 0;
2724
		put_cmsg(msg, SOL_PACKET, PACKET_AUXDATA, sizeof(aux), &aux);
2725 2726
	}

L
Linus Torvalds 已提交
2727 2728 2729 2730
	/*
	 *	Free or return the buffer as appropriate. Again this
	 *	hides all the races and re-entrancy issues from us.
	 */
2731
	err = vnet_hdr_len + ((flags&MSG_TRUNC) ? skb->len : copied);
L
Linus Torvalds 已提交
2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748

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
2749 2750 2751
	rcu_read_lock();
	dev = dev_get_by_index_rcu(sock_net(sk), pkt_sk(sk)->ifindex);
	if (dev)
2752
		strncpy(uaddr->sa_data, dev->name, 14);
E
Eric Dumazet 已提交
2753
	else
L
Linus Torvalds 已提交
2754
		memset(uaddr->sa_data, 0, 14);
E
Eric Dumazet 已提交
2755
	rcu_read_unlock();
L
Linus Torvalds 已提交
2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766
	*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);
2767
	DECLARE_SOCKADDR(struct sockaddr_ll *, sll, uaddr);
L
Linus Torvalds 已提交
2768 2769 2770 2771 2772 2773 2774

	if (peer)
		return -EOPNOTSUPP;

	sll->sll_family = AF_PACKET;
	sll->sll_ifindex = po->ifindex;
	sll->sll_protocol = po->num;
2775
	sll->sll_pkttype = 0;
E
Eric Dumazet 已提交
2776 2777
	rcu_read_lock();
	dev = dev_get_by_index_rcu(sock_net(sk), po->ifindex);
L
Linus Torvalds 已提交
2778 2779 2780 2781 2782 2783 2784 2785
	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
2786
	rcu_read_unlock();
2787
	*uaddr_len = offsetof(struct sockaddr_ll, sll_addr) + sll->sll_halen;
L
Linus Torvalds 已提交
2788 2789 2790 2791

	return 0;
}

2792 2793
static int packet_dev_mc(struct net_device *dev, struct packet_mclist *i,
			 int what)
L
Linus Torvalds 已提交
2794 2795 2796
{
	switch (i->type) {
	case PACKET_MR_MULTICAST:
2797 2798
		if (i->alen != dev->addr_len)
			return -EINVAL;
L
Linus Torvalds 已提交
2799
		if (what > 0)
2800
			return dev_mc_add(dev, i->addr);
L
Linus Torvalds 已提交
2801
		else
2802
			return dev_mc_del(dev, i->addr);
L
Linus Torvalds 已提交
2803 2804
		break;
	case PACKET_MR_PROMISC:
2805
		return dev_set_promiscuity(dev, what);
L
Linus Torvalds 已提交
2806 2807
		break;
	case PACKET_MR_ALLMULTI:
2808
		return dev_set_allmulti(dev, what);
L
Linus Torvalds 已提交
2809
		break;
2810
	case PACKET_MR_UNICAST:
2811 2812
		if (i->alen != dev->addr_len)
			return -EINVAL;
2813
		if (what > 0)
2814
			return dev_uc_add(dev, i->addr);
2815
		else
2816
			return dev_uc_del(dev, i->addr);
2817
		break;
E
Eric Dumazet 已提交
2818 2819
	default:
		break;
L
Linus Torvalds 已提交
2820
	}
2821
	return 0;
L
Linus Torvalds 已提交
2822 2823 2824 2825
}

static void packet_dev_mclist(struct net_device *dev, struct packet_mclist *i, int what)
{
E
Eric Dumazet 已提交
2826
	for ( ; i; i = i->next) {
L
Linus Torvalds 已提交
2827 2828 2829 2830 2831
		if (i->ifindex == dev->ifindex)
			packet_dev_mc(dev, i, what);
	}
}

2832
static int packet_mc_add(struct sock *sk, struct packet_mreq_max *mreq)
L
Linus Torvalds 已提交
2833 2834 2835 2836 2837 2838 2839 2840 2841
{
	struct packet_sock *po = pkt_sk(sk);
	struct packet_mclist *ml, *i;
	struct net_device *dev;
	int err;

	rtnl_lock();

	err = -ENODEV;
2842
	dev = __dev_get_by_index(sock_net(sk), mreq->mr_ifindex);
L
Linus Torvalds 已提交
2843 2844 2845 2846
	if (!dev)
		goto done;

	err = -EINVAL;
2847
	if (mreq->mr_alen > dev->addr_len)
L
Linus Torvalds 已提交
2848 2849 2850
		goto done;

	err = -ENOBUFS;
2851
	i = kmalloc(sizeof(*i), GFP_KERNEL);
L
Linus Torvalds 已提交
2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 2874
	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;
2875 2876 2877 2878 2879
	err = packet_dev_mc(dev, i, 1);
	if (err) {
		po->mclist = i->next;
		kfree(i);
	}
L
Linus Torvalds 已提交
2880 2881 2882 2883 2884 2885

done:
	rtnl_unlock();
	return err;
}

2886
static int packet_mc_drop(struct sock *sk, struct packet_mreq_max *mreq)
L
Linus Torvalds 已提交
2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899
{
	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;
2900 2901
				dev = __dev_get_by_index(sock_net(sk), ml->ifindex);
				if (dev)
L
Linus Torvalds 已提交
2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925
					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;
2926 2927
		dev = __dev_get_by_index(sock_net(sk), ml->ifindex);
		if (dev != NULL)
L
Linus Torvalds 已提交
2928 2929 2930 2931 2932 2933 2934
			packet_dev_mc(dev, ml, -1);
		kfree(ml);
	}
	rtnl_unlock();
}

static int
2935
packet_setsockopt(struct socket *sock, int level, int optname, char __user *optval, unsigned int optlen)
L
Linus Torvalds 已提交
2936 2937
{
	struct sock *sk = sock->sk;
2938
	struct packet_sock *po = pkt_sk(sk);
L
Linus Torvalds 已提交
2939 2940 2941 2942 2943
	int ret;

	if (level != SOL_PACKET)
		return -ENOPROTOOPT;

J
Johann Baudy 已提交
2944
	switch (optname) {
2945
	case PACKET_ADD_MEMBERSHIP:
L
Linus Torvalds 已提交
2946 2947
	case PACKET_DROP_MEMBERSHIP:
	{
2948 2949 2950 2951
		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
2952
			return -EINVAL;
2953 2954
		if (len > sizeof(mreq))
			len = sizeof(mreq);
E
Eric Dumazet 已提交
2955
		if (copy_from_user(&mreq, optval, len))
L
Linus Torvalds 已提交
2956
			return -EFAULT;
2957 2958
		if (len < (mreq.mr_alen + offsetof(struct packet_mreq, mr_address)))
			return -EINVAL;
L
Linus Torvalds 已提交
2959 2960 2961 2962 2963 2964
		if (optname == PACKET_ADD_MEMBERSHIP)
			ret = packet_mc_add(sk, &mreq);
		else
			ret = packet_mc_drop(sk, &mreq);
		return ret;
	}
2965

L
Linus Torvalds 已提交
2966
	case PACKET_RX_RING:
J
Johann Baudy 已提交
2967
	case PACKET_TX_RING:
L
Linus Torvalds 已提交
2968
	{
2969 2970
		union tpacket_req_u req_u;
		int len;
L
Linus Torvalds 已提交
2971

2972 2973 2974 2975 2976 2977 2978 2979 2980 2981 2982
		switch (po->tp_version) {
		case TPACKET_V1:
		case TPACKET_V2:
			len = sizeof(req_u.req);
			break;
		case TPACKET_V3:
		default:
			len = sizeof(req_u.req3);
			break;
		}
		if (optlen < len)
L
Linus Torvalds 已提交
2983
			return -EINVAL;
2984 2985
		if (pkt_sk(sk)->has_vnet_hdr)
			return -EINVAL;
2986
		if (copy_from_user(&req_u.req, optval, len))
L
Linus Torvalds 已提交
2987
			return -EFAULT;
2988 2989
		return packet_set_ring(sk, &req_u, 0,
			optname == PACKET_TX_RING);
L
Linus Torvalds 已提交
2990 2991 2992 2993 2994
	}
	case PACKET_COPY_THRESH:
	{
		int val;

E
Eric Dumazet 已提交
2995
		if (optlen != sizeof(val))
L
Linus Torvalds 已提交
2996
			return -EINVAL;
E
Eric Dumazet 已提交
2997
		if (copy_from_user(&val, optval, sizeof(val)))
L
Linus Torvalds 已提交
2998 2999 3000 3001 3002
			return -EFAULT;

		pkt_sk(sk)->copy_thresh = val;
		return 0;
	}
3003 3004 3005 3006 3007 3008
	case PACKET_VERSION:
	{
		int val;

		if (optlen != sizeof(val))
			return -EINVAL;
J
Johann Baudy 已提交
3009
		if (po->rx_ring.pg_vec || po->tx_ring.pg_vec)
3010 3011 3012 3013 3014 3015
			return -EBUSY;
		if (copy_from_user(&val, optval, sizeof(val)))
			return -EFAULT;
		switch (val) {
		case TPACKET_V1:
		case TPACKET_V2:
3016
		case TPACKET_V3:
3017 3018 3019 3020 3021 3022
			po->tp_version = val;
			return 0;
		default:
			return -EINVAL;
		}
	}
3023 3024 3025 3026 3027 3028
	case PACKET_RESERVE:
	{
		unsigned int val;

		if (optlen != sizeof(val))
			return -EINVAL;
J
Johann Baudy 已提交
3029
		if (po->rx_ring.pg_vec || po->tx_ring.pg_vec)
3030 3031 3032 3033 3034 3035
			return -EBUSY;
		if (copy_from_user(&val, optval, sizeof(val)))
			return -EFAULT;
		po->tp_reserve = val;
		return 0;
	}
J
Johann Baudy 已提交
3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048
	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;
	}
3049 3050 3051 3052 3053 3054 3055 3056 3057 3058 3059 3060
	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;
	}
3061 3062 3063 3064 3065 3066 3067 3068 3069 3070 3071 3072
	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;
	}
3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 3087 3088
	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;
	}
3089 3090 3091 3092 3093 3094 3095 3096 3097 3098 3099 3100
	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;
	}
D
David S. Miller 已提交
3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111
	case PACKET_FANOUT:
	{
		int val;

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

		return fanout_add(sk, val & 0xffff, val >> 16);
	}
L
Linus Torvalds 已提交
3112 3113 3114 3115 3116 3117 3118 3119 3120
	default:
		return -ENOPROTOOPT;
	}
}

static int packet_getsockopt(struct socket *sock, int level, int optname,
			     char __user *optval, int __user *optlen)
{
	int len;
3121
	int val, lv = sizeof(val);
L
Linus Torvalds 已提交
3122 3123
	struct sock *sk = sock->sk;
	struct packet_sock *po = pkt_sk(sk);
3124
	void *data = &val;
3125
	struct tpacket_stats st;
3126
	union tpacket_stats_u st_u;
L
Linus Torvalds 已提交
3127 3128 3129 3130

	if (level != SOL_PACKET)
		return -ENOPROTOOPT;

3131 3132
	if (get_user(len, optlen))
		return -EFAULT;
L
Linus Torvalds 已提交
3133 3134 3135

	if (len < 0)
		return -EINVAL;
3136

J
Johann Baudy 已提交
3137
	switch (optname) {
L
Linus Torvalds 已提交
3138 3139
	case PACKET_STATISTICS:
		spin_lock_bh(&sk->sk_receive_queue.lock);
3140
		if (po->tp_version == TPACKET_V3) {
3141
			lv = sizeof(struct tpacket_stats_v3);
3142
			memcpy(&st_u.stats3, &po->stats,
3143
			       sizeof(struct tpacket_stats));
3144
			st_u.stats3.tp_freeze_q_cnt =
3145
					po->stats_u.stats3.tp_freeze_q_cnt;
3146 3147 3148
			st_u.stats3.tp_packets += po->stats.tp_drops;
			data = &st_u.stats3;
		} else {
3149
			lv = sizeof(struct tpacket_stats);
3150 3151 3152 3153
			st = po->stats;
			st.tp_packets += st.tp_drops;
			data = &st;
		}
L
Linus Torvalds 已提交
3154 3155
		memset(&po->stats, 0, sizeof(st));
		spin_unlock_bh(&sk->sk_receive_queue.lock);
3156 3157 3158
		break;
	case PACKET_AUXDATA:
		val = po->auxdata;
3159 3160 3161
		break;
	case PACKET_ORIGDEV:
		val = po->origdev;
3162 3163 3164
		break;
	case PACKET_VNET_HDR:
		val = po->has_vnet_hdr;
L
Linus Torvalds 已提交
3165
		break;
3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180
	case PACKET_VERSION:
		val = po->tp_version;
		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;
3181 3182 3183
		case TPACKET_V3:
			val = sizeof(struct tpacket3_hdr);
			break;
3184 3185 3186 3187
		default:
			return -EINVAL;
		}
		break;
3188 3189 3190
	case PACKET_RESERVE:
		val = po->tp_reserve;
		break;
J
Johann Baudy 已提交
3191 3192 3193
	case PACKET_LOSS:
		val = po->tp_loss;
		break;
3194 3195 3196
	case PACKET_TIMESTAMP:
		val = po->tp_tstamp;
		break;
D
David S. Miller 已提交
3197 3198 3199 3200 3201 3202
	case PACKET_FANOUT:
		val = (po->fanout ?
		       ((u32)po->fanout->id |
			((u32)po->fanout->type << 16)) :
		       0);
		break;
L
Linus Torvalds 已提交
3203 3204 3205 3206
	default:
		return -ENOPROTOOPT;
	}

3207 3208
	if (len > lv)
		len = lv;
3209 3210
	if (put_user(len, optlen))
		return -EFAULT;
3211 3212
	if (copy_to_user(optval, data, len))
		return -EFAULT;
3213
	return 0;
L
Linus Torvalds 已提交
3214 3215 3216 3217 3218 3219 3220
}


static int packet_notifier(struct notifier_block *this, unsigned long msg, void *data)
{
	struct sock *sk;
	struct hlist_node *node;
3221
	struct net_device *dev = data;
3222
	struct net *net = dev_net(dev);
L
Linus Torvalds 已提交
3223

3224 3225
	rcu_read_lock();
	sk_for_each_rcu(sk, node, &net->packet.sklist) {
L
Linus Torvalds 已提交
3226 3227 3228 3229 3230 3231
		struct packet_sock *po = pkt_sk(sk);

		switch (msg) {
		case NETDEV_UNREGISTER:
			if (po->mclist)
				packet_dev_mclist(dev, po->mclist, -1);
3232 3233
			/* fallthrough */

L
Linus Torvalds 已提交
3234 3235 3236 3237
		case NETDEV_DOWN:
			if (dev->ifindex == po->ifindex) {
				spin_lock(&po->bind_lock);
				if (po->running) {
3238
					__unregister_prot_hook(sk, false);
L
Linus Torvalds 已提交
3239 3240 3241 3242 3243 3244
					sk->sk_err = ENETDOWN;
					if (!sock_flag(sk, SOCK_DEAD))
						sk->sk_error_report(sk);
				}
				if (msg == NETDEV_UNREGISTER) {
					po->ifindex = -1;
3245 3246
					if (po->prot_hook.dev)
						dev_put(po->prot_hook.dev);
L
Linus Torvalds 已提交
3247 3248 3249 3250 3251 3252
					po->prot_hook.dev = NULL;
				}
				spin_unlock(&po->bind_lock);
			}
			break;
		case NETDEV_UP:
3253 3254
			if (dev->ifindex == po->ifindex) {
				spin_lock(&po->bind_lock);
3255 3256
				if (po->num)
					register_prot_hook(sk);
3257
				spin_unlock(&po->bind_lock);
L
Linus Torvalds 已提交
3258 3259 3260 3261
			}
			break;
		}
	}
3262
	rcu_read_unlock();
L
Linus Torvalds 已提交
3263 3264 3265 3266 3267 3268 3269 3270 3271
	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
3272
	switch (cmd) {
E
Eric Dumazet 已提交
3273 3274 3275
	case SIOCOUTQ:
	{
		int amount = sk_wmem_alloc_get(sk);
3276

E
Eric Dumazet 已提交
3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288 3289 3290 3291 3292 3293 3294
		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);
3295

L
Linus Torvalds 已提交
3296
#ifdef CONFIG_INET
E
Eric Dumazet 已提交
3297 3298 3299 3300 3301 3302 3303 3304 3305 3306 3307 3308 3309 3310 3311
	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
3312 3313
#endif

E
Eric Dumazet 已提交
3314 3315
	default:
		return -ENOIOCTLCMD;
L
Linus Torvalds 已提交
3316 3317 3318 3319
	}
	return 0;
}

E
Eric Dumazet 已提交
3320
static unsigned int packet_poll(struct file *file, struct socket *sock,
L
Linus Torvalds 已提交
3321 3322 3323 3324 3325 3326 3327
				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
3328
	if (po->rx_ring.pg_vec) {
3329 3330
		if (!packet_previous_rx_frame(po, &po->rx_ring,
			TP_STATUS_KERNEL))
L
Linus Torvalds 已提交
3331 3332 3333
			mask |= POLLIN | POLLRDNORM;
	}
	spin_unlock_bh(&sk->sk_receive_queue.lock);
J
Johann Baudy 已提交
3334 3335 3336 3337 3338 3339
	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
3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350
	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
3351
	struct socket *sock = file->private_data;
L
Linus Torvalds 已提交
3352
	struct sock *sk = sock->sk;
3353

L
Linus Torvalds 已提交
3354 3355 3356 3357 3358 3359 3360
	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
3361
	struct socket *sock = file->private_data;
L
Linus Torvalds 已提交
3362
	struct sock *sk = sock->sk;
3363

L
Linus Torvalds 已提交
3364 3365 3366 3367
	if (sk)
		atomic_dec(&pkt_sk(sk)->mapped);
}

3368
static const struct vm_operations_struct packet_mmap_ops = {
E
Eric Dumazet 已提交
3369 3370
	.open	=	packet_mm_open,
	.close	=	packet_mm_close,
L
Linus Torvalds 已提交
3371 3372
};

3373 3374
static void free_pg_vec(struct pgv *pg_vec, unsigned int order,
			unsigned int len)
L
Linus Torvalds 已提交
3375 3376 3377
{
	int i;

3378
	for (i = 0; i < len; i++) {
3379
		if (likely(pg_vec[i].buffer)) {
C
Changli Gao 已提交
3380
			if (is_vmalloc_addr(pg_vec[i].buffer))
3381 3382 3383 3384 3385 3386
				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
3387 3388 3389 3390
	}
	kfree(pg_vec);
}

3391
static char *alloc_one_pg_vec_page(unsigned long order)
3392
{
3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404
	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
3405
	buffer = vzalloc((1 << order) * PAGE_SIZE);
3406

3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421
	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;
3422 3423
}

3424
static struct pgv *alloc_pg_vec(struct tpacket_req *req, int order)
3425 3426
{
	unsigned int block_nr = req->tp_block_nr;
3427
	struct pgv *pg_vec;
3428 3429
	int i;

3430
	pg_vec = kcalloc(block_nr, sizeof(struct pgv), GFP_KERNEL);
3431 3432 3433 3434
	if (unlikely(!pg_vec))
		goto out;

	for (i = 0; i < block_nr; i++) {
C
Changli Gao 已提交
3435
		pg_vec[i].buffer = alloc_one_pg_vec_page(order);
3436
		if (unlikely(!pg_vec[i].buffer))
3437 3438 3439 3440 3441 3442 3443 3444 3445 3446 3447
			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
3448

3449
static int packet_set_ring(struct sock *sk, union tpacket_req_u *req_u,
J
Johann Baudy 已提交
3450
		int closing, int tx_ring)
L
Linus Torvalds 已提交
3451
{
3452
	struct pgv *pg_vec = NULL;
L
Linus Torvalds 已提交
3453
	struct packet_sock *po = pkt_sk(sk);
A
Al Viro 已提交
3454
	int was_running, order = 0;
J
Johann Baudy 已提交
3455 3456
	struct packet_ring_buffer *rb;
	struct sk_buff_head *rb_queue;
A
Al Viro 已提交
3457
	__be16 num;
3458 3459 3460 3461 3462 3463 3464 3465 3466
	int err = -EINVAL;
	/* Added to avoid minimal code churn */
	struct tpacket_req *req = &req_u->req;

	/* Opening a Tx-ring is NOT supported in TPACKET_V3 */
	if (!closing && tx_ring && (po->tp_version > TPACKET_V2)) {
		WARN(1, "Tx-ring is not supported.\n");
		goto out;
	}
3467

J
Johann Baudy 已提交
3468 3469
	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
3470

J
Johann Baudy 已提交
3471 3472 3473 3474 3475 3476 3477
	err = -EBUSY;
	if (!closing) {
		if (atomic_read(&po->mapped))
			goto out;
		if (atomic_read(&rb->pending))
			goto out;
	}
L
Linus Torvalds 已提交
3478

J
Johann Baudy 已提交
3479 3480 3481 3482 3483
	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
3484

3485 3486 3487 3488 3489 3490 3491
		switch (po->tp_version) {
		case TPACKET_V1:
			po->tp_hdrlen = TPACKET_HDRLEN;
			break;
		case TPACKET_V2:
			po->tp_hdrlen = TPACKET2_HDRLEN;
			break;
3492 3493 3494
		case TPACKET_V3:
			po->tp_hdrlen = TPACKET3_HDRLEN;
			break;
3495 3496
		}

J
Johann Baudy 已提交
3497
		err = -EINVAL;
3498
		if (unlikely((int)req->tp_block_size <= 0))
J
Johann Baudy 已提交
3499
			goto out;
3500
		if (unlikely(req->tp_block_size & (PAGE_SIZE - 1)))
J
Johann Baudy 已提交
3501
			goto out;
3502
		if (unlikely(req->tp_frame_size < po->tp_hdrlen +
J
Johann Baudy 已提交
3503 3504
					po->tp_reserve))
			goto out;
3505
		if (unlikely(req->tp_frame_size & (TPACKET_ALIGNMENT - 1)))
J
Johann Baudy 已提交
3506
			goto out;
L
Linus Torvalds 已提交
3507

J
Johann Baudy 已提交
3508 3509 3510 3511 3512 3513
		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
3514 3515

		err = -ENOMEM;
3516 3517 3518
		order = get_order(req->tp_block_size);
		pg_vec = alloc_pg_vec(req, order);
		if (unlikely(!pg_vec))
L
Linus Torvalds 已提交
3519
			goto out;
3520 3521 3522 3523 3524 3525 3526 3527 3528 3529 3530
		switch (po->tp_version) {
		case TPACKET_V3:
		/* Transmit path is not supported. We checked
		 * it above but just being paranoid
		 */
			if (!tx_ring)
				init_prb_bdqc(po, rb, pg_vec, req_u, tx_ring);
				break;
		default:
			break;
		}
J
Johann Baudy 已提交
3531 3532 3533 3534
	}
	/* Done */
	else {
		err = -EINVAL;
3535
		if (unlikely(req->tp_frame_nr))
J
Johann Baudy 已提交
3536
			goto out;
L
Linus Torvalds 已提交
3537 3538 3539 3540 3541 3542 3543 3544 3545 3546
	}

	lock_sock(sk);

	/* Detach socket from network */
	spin_lock(&po->bind_lock);
	was_running = po->running;
	num = po->num;
	if (was_running) {
		po->num = 0;
3547
		__unregister_prot_hook(sk, false);
L
Linus Torvalds 已提交
3548 3549
	}
	spin_unlock(&po->bind_lock);
3550

L
Linus Torvalds 已提交
3551 3552 3553
	synchronize_net();

	err = -EBUSY;
3554
	mutex_lock(&po->pg_vec_lock);
L
Linus Torvalds 已提交
3555 3556
	if (closing || atomic_read(&po->mapped) == 0) {
		err = 0;
J
Johann Baudy 已提交
3557
		spin_lock_bh(&rb_queue->lock);
3558
		swap(rb->pg_vec, pg_vec);
J
Johann Baudy 已提交
3559 3560 3561 3562 3563
		rb->frame_max = (req->tp_frame_nr - 1);
		rb->head = 0;
		rb->frame_size = req->tp_frame_size;
		spin_unlock_bh(&rb_queue->lock);

3564 3565
		swap(rb->pg_vec_order, order);
		swap(rb->pg_vec_len, req->tp_block_nr);
J
Johann Baudy 已提交
3566 3567 3568 3569 3570

		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
3571
		if (atomic_read(&po->mapped))
E
Eric Dumazet 已提交
3572 3573
			pr_err("packet_mmap: vma is busy: %d\n",
			       atomic_read(&po->mapped));
L
Linus Torvalds 已提交
3574
	}
3575
	mutex_unlock(&po->pg_vec_lock);
L
Linus Torvalds 已提交
3576 3577

	spin_lock(&po->bind_lock);
3578
	if (was_running) {
L
Linus Torvalds 已提交
3579
		po->num = num;
3580
		register_prot_hook(sk);
L
Linus Torvalds 已提交
3581 3582
	}
	spin_unlock(&po->bind_lock);
3583 3584 3585 3586 3587
	if (closing && (po->tp_version > TPACKET_V2)) {
		/* Because we don't support block-based V3 on tx-ring */
		if (!tx_ring)
			prb_shutdown_retire_blk_timer(po, tx_ring, rb_queue);
	}
L
Linus Torvalds 已提交
3588 3589 3590 3591 3592 3593 3594 3595
	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
3596 3597
static int packet_mmap(struct file *file, struct socket *sock,
		struct vm_area_struct *vma)
L
Linus Torvalds 已提交
3598 3599 3600
{
	struct sock *sk = sock->sk;
	struct packet_sock *po = pkt_sk(sk);
J
Johann Baudy 已提交
3601 3602
	unsigned long size, expected_size;
	struct packet_ring_buffer *rb;
L
Linus Torvalds 已提交
3603 3604 3605 3606 3607 3608 3609
	unsigned long start;
	int err = -EINVAL;
	int i;

	if (vma->vm_pgoff)
		return -EINVAL;

3610
	mutex_lock(&po->pg_vec_lock);
J
Johann Baudy 已提交
3611 3612 3613 3614 3615 3616 3617 3618 3619 3620 3621

	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
3622
		goto out;
J
Johann Baudy 已提交
3623 3624 3625

	size = vma->vm_end - vma->vm_start;
	if (size != expected_size)
L
Linus Torvalds 已提交
3626 3627 3628
		goto out;

	start = vma->vm_start;
J
Johann Baudy 已提交
3629 3630 3631 3632 3633
	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++) {
3634 3635
			struct page *page;
			void *kaddr = rb->pg_vec[i].buffer;
J
Johann Baudy 已提交
3636 3637
			int pg_num;

C
Changli Gao 已提交
3638 3639
			for (pg_num = 0; pg_num < rb->pg_vec_pages; pg_num++) {
				page = pgv_to_page(kaddr);
J
Johann Baudy 已提交
3640 3641 3642 3643
				err = vm_insert_page(vma, start, page);
				if (unlikely(err))
					goto out;
				start += PAGE_SIZE;
3644
				kaddr += PAGE_SIZE;
J
Johann Baudy 已提交
3645
			}
3646
		}
L
Linus Torvalds 已提交
3647
	}
J
Johann Baudy 已提交
3648

3649
	atomic_inc(&po->mapped);
L
Linus Torvalds 已提交
3650 3651 3652 3653
	vma->vm_ops = &packet_mmap_ops;
	err = 0;

out:
3654
	mutex_unlock(&po->pg_vec_lock);
L
Linus Torvalds 已提交
3655 3656 3657
	return err;
}

3658
static const struct proto_ops packet_ops_spkt = {
L
Linus Torvalds 已提交
3659 3660 3661 3662 3663 3664 3665 3666 3667 3668 3669 3670 3671 3672 3673 3674 3675 3676 3677 3678
	.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,
};

3679
static const struct proto_ops packet_ops = {
L
Linus Torvalds 已提交
3680 3681 3682 3683 3684 3685 3686
	.family =	PF_PACKET,
	.owner =	THIS_MODULE,
	.release =	packet_release,
	.bind =		packet_bind,
	.connect =	sock_no_connect,
	.socketpair =	sock_no_socketpair,
	.accept =	sock_no_accept,
3687
	.getname =	packet_getname,
L
Linus Torvalds 已提交
3688 3689 3690 3691 3692 3693 3694 3695 3696 3697 3698 3699
	.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,
};

3700
static const struct net_proto_family packet_family_ops = {
L
Linus Torvalds 已提交
3701 3702 3703 3704 3705 3706
	.family =	PF_PACKET,
	.create =	packet_create,
	.owner	=	THIS_MODULE,
};

static struct notifier_block packet_netdev_notifier = {
E
Eric Dumazet 已提交
3707
	.notifier_call =	packet_notifier,
L
Linus Torvalds 已提交
3708 3709 3710 3711 3712
};

#ifdef CONFIG_PROC_FS

static void *packet_seq_start(struct seq_file *seq, loff_t *pos)
3713
	__acquires(RCU)
L
Linus Torvalds 已提交
3714
{
3715
	struct net *net = seq_file_net(seq);
3716 3717 3718

	rcu_read_lock();
	return seq_hlist_start_head_rcu(&net->packet.sklist, *pos);
L
Linus Torvalds 已提交
3719 3720 3721 3722
}

static void *packet_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
3723
	struct net *net = seq_file_net(seq);
3724
	return seq_hlist_next_rcu(v, &net->packet.sklist, pos);
L
Linus Torvalds 已提交
3725 3726 3727
}

static void packet_seq_stop(struct seq_file *seq, void *v)
3728
	__releases(RCU)
L
Linus Torvalds 已提交
3729
{
3730
	rcu_read_unlock();
L
Linus Torvalds 已提交
3731 3732
}

3733
static int packet_seq_show(struct seq_file *seq, void *v)
L
Linus Torvalds 已提交
3734 3735 3736 3737
{
	if (v == SEQ_START_TOKEN)
		seq_puts(seq, "sk       RefCnt Type Proto  Iface R Rmem   User   Inode\n");
	else {
3738
		struct sock *s = sk_entry(v);
L
Linus Torvalds 已提交
3739 3740 3741
		const struct packet_sock *po = pkt_sk(s);

		seq_printf(seq,
D
Dan Rosenberg 已提交
3742
			   "%pK %-6d %-4d %04x   %-5d %1d %-6u %-6u %-6lu\n",
L
Linus Torvalds 已提交
3743 3744 3745 3746 3747 3748 3749
			   s,
			   atomic_read(&s->sk_refcnt),
			   s->sk_type,
			   ntohs(po->num),
			   po->ifindex,
			   po->running,
			   atomic_read(&s->sk_rmem_alloc),
3750
			   from_kuid_munged(seq_user_ns(seq), sock_i_uid(s)),
E
Eric Dumazet 已提交
3751
			   sock_i_ino(s));
L
Linus Torvalds 已提交
3752 3753 3754 3755 3756
	}

	return 0;
}

3757
static const struct seq_operations packet_seq_ops = {
L
Linus Torvalds 已提交
3758 3759 3760 3761 3762 3763 3764 3765
	.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)
{
3766 3767
	return seq_open_net(inode, file, &packet_seq_ops,
			    sizeof(struct seq_net_private));
L
Linus Torvalds 已提交
3768 3769
}

3770
static const struct file_operations packet_seq_fops = {
L
Linus Torvalds 已提交
3771 3772 3773 3774
	.owner		= THIS_MODULE,
	.open		= packet_seq_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
3775
	.release	= seq_release_net,
L
Linus Torvalds 已提交
3776 3777 3778 3779
};

#endif

3780
static int __net_init packet_net_init(struct net *net)
3781
{
3782
	mutex_init(&net->packet.sklist_lock);
3783
	INIT_HLIST_HEAD(&net->packet.sklist);
3784 3785 3786 3787 3788 3789 3790

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

	return 0;
}

3791
static void __net_exit packet_net_exit(struct net *net)
3792 3793 3794 3795 3796 3797 3798 3799 3800 3801
{
	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
3802 3803 3804
static void __exit packet_exit(void)
{
	unregister_netdevice_notifier(&packet_netdev_notifier);
3805
	unregister_pernet_subsys(&packet_net_ops);
L
Linus Torvalds 已提交
3806 3807 3808 3809 3810 3811 3812 3813 3814 3815 3816 3817
	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);
3818
	register_pernet_subsys(&packet_net_ops);
L
Linus Torvalds 已提交
3819 3820 3821 3822 3823 3824 3825 3826 3827
	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);