af_packet.c 91.4 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);
184 185
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
186

187 188 189 190
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);
191 192 193
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 *,
194
			struct packet_sock *);
195
static void prb_retire_current_block(struct tpacket_kbdq_core *,
196
		struct packet_sock *, unsigned int status);
197 198 199
static int prb_queue_frozen(struct tpacket_kbdq_core *);
static void prb_open_block(struct tpacket_kbdq_core *,
		struct tpacket_block_desc *);
200
static void prb_retire_rx_blk_timer_expired(unsigned long);
201 202 203 204 205 206 207 208 209
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
210 211
static void packet_flush_mclist(struct sock *sk);

212 213 214 215 216 217 218 219 220
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))
221

222
#define GET_PBDQC_FROM_RB(x)	((struct tpacket_kbdq_core *)(&(x)->prb_bdqc))
223
#define GET_PBLOCK_DESC(x, bid)	\
224
	((struct tpacket_block_desc *)((x)->pkbdq[(bid)].buffer))
225
#define GET_CURR_PBLOCK_DESC_FROM_CORE(x)	\
226
	((struct tpacket_block_desc *)((x)->pkbdq[(x)->kactive_blk_num].buffer))
227 228 229 230
#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
231 232 233
static void __fanout_unlink(struct sock *sk, struct packet_sock *po);
static void __fanout_link(struct sock *sk, struct packet_sock *po);

234 235 236 237 238 239 240 241
/* 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
242 243 244 245
		if (po->fanout)
			__fanout_link(sk, po);
		else
			dev_add_pack(&po->prot_hook);
246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262
		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
263 264 265 266
	if (po->fanout)
		__fanout_unlink(sk, po);
	else
		__dev_remove_pack(&po->prot_hook);
267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283
	__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);
}

284
static inline __pure struct page *pgv_to_page(void *addr)
285 286 287 288 289 290
{
	if (is_vmalloc_addr(addr))
		return vmalloc_to_page(addr);
	return virt_to_page(addr);
}

J
Johann Baudy 已提交
291
static void __packet_set_status(struct packet_sock *po, void *frame, int status)
L
Linus Torvalds 已提交
292
{
293 294 295 296 297
	union {
		struct tpacket_hdr *h1;
		struct tpacket2_hdr *h2;
		void *raw;
	} h;
L
Linus Torvalds 已提交
298

J
Johann Baudy 已提交
299
	h.raw = frame;
300 301
	switch (po->tp_version) {
	case TPACKET_V1:
J
Johann Baudy 已提交
302
		h.h1->tp_status = status;
303
		flush_dcache_page(pgv_to_page(&h.h1->tp_status));
304 305
		break;
	case TPACKET_V2:
J
Johann Baudy 已提交
306
		h.h2->tp_status = status;
307
		flush_dcache_page(pgv_to_page(&h.h2->tp_status));
308
		break;
309
	case TPACKET_V3:
J
Johann Baudy 已提交
310
	default:
311
		WARN(1, "TPACKET version not supported.\n");
J
Johann Baudy 已提交
312
		BUG();
313
	}
J
Johann Baudy 已提交
314 315

	smp_wmb();
316 317
}

J
Johann Baudy 已提交
318
static int __packet_get_status(struct packet_sock *po, void *frame)
319 320 321 322 323 324 325
{
	union {
		struct tpacket_hdr *h1;
		struct tpacket2_hdr *h2;
		void *raw;
	} h;

J
Johann Baudy 已提交
326 327
	smp_rmb();

328 329 330
	h.raw = frame;
	switch (po->tp_version) {
	case TPACKET_V1:
331
		flush_dcache_page(pgv_to_page(&h.h1->tp_status));
J
Johann Baudy 已提交
332
		return h.h1->tp_status;
333
	case TPACKET_V2:
334
		flush_dcache_page(pgv_to_page(&h.h2->tp_status));
J
Johann Baudy 已提交
335
		return h.h2->tp_status;
336
	case TPACKET_V3:
J
Johann Baudy 已提交
337
	default:
338
		WARN(1, "TPACKET version not supported.\n");
J
Johann Baudy 已提交
339 340
		BUG();
		return 0;
341
	}
L
Linus Torvalds 已提交
342
}
J
Johann Baudy 已提交
343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358

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;

359 360
	h.raw = rb->pg_vec[pg_vec_pos].buffer +
		(frame_offset * rb->frame_size);
J
Johann Baudy 已提交
361 362 363 364 365 366 367

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

	return h.raw;
}

368
static void *packet_current_frame(struct packet_sock *po,
J
Johann Baudy 已提交
369 370 371 372 373 374
		struct packet_ring_buffer *rb,
		int status)
{
	return packet_lookup_frame(po, rb, rb->head, status);
}

375
static void prb_del_retire_blk_timer(struct tpacket_kbdq_core *pkc)
376 377 378 379 380 381 382 383
{
	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)
{
384
	struct tpacket_kbdq_core *pkc;
385 386 387 388 389 390 391 392 393 394 395

	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,
396
		struct tpacket_kbdq_core *pkc,
397 398 399 400 401 402 403 404 405 406
		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)
{
407
	struct tpacket_kbdq_core *pkc;
408 409 410 411 412 413 414 415 416 417 418 419 420

	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;
421 422
	struct ethtool_cmd ecmd;
	int err;
423
	u32 speed;
424

425 426 427 428
	rtnl_lock();
	dev = __dev_get_by_index(sock_net(&po->sk), po->ifindex);
	if (unlikely(!dev)) {
		rtnl_unlock();
429
		return DEFAULT_PRB_RETIRE_TOV;
430 431
	}
	err = __ethtool_get_settings(dev, &ecmd);
432
	speed = ethtool_cmd_speed(&ecmd);
433 434 435 436 437 438
	rtnl_unlock();
	if (!err) {
		/*
		 * If the link speed is so slow you don't really
		 * need to worry about perf anyways
		 */
439
		if (speed < SPEED_1000 || speed == SPEED_UNKNOWN) {
440
			return DEFAULT_PRB_RETIRE_TOV;
441 442 443
		} else {
			msec = 1;
			div = speed / 1000;
444 445 446 447 448 449 450 451 452 453 454 455 456 457 458
		}
	}

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

	if (div)
		mbits /= div;

	tmo = mbits * msec;

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

459
static void prb_init_ft_ops(struct tpacket_kbdq_core *p1,
460 461 462 463 464 465 466 467 468 469
			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)
{
470 471
	struct tpacket_kbdq_core *p1 = &rb->prb_bdqc;
	struct tpacket_block_desc *pbd;
472 473 474 475 476

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

	p1->knxt_seq_num = 1;
	p1->pkbdq = pg_vec;
477
	pbd = (struct tpacket_block_desc *)pg_vec[0].buffer;
J
Joe Perches 已提交
478
	p1->pkblk_start	= pg_vec[0].buffer;
479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500
	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.
 */
501
static void _prb_refresh_rx_retire_blk_timer(struct tpacket_kbdq_core *pkc)
502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533
{
	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;
534
	struct tpacket_kbdq_core *pkc = &po->rx_ring.prb_bdqc;
535
	unsigned int frozen;
536
	struct tpacket_block_desc *pbd;
537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599

	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);
}

600
static void prb_flush_block(struct tpacket_kbdq_core *pkc1,
601
		struct tpacket_block_desc *pbd1, __u32 status)
602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642
{
	/* 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.
 */
643 644
static void prb_close_block(struct tpacket_kbdq_core *pkc1,
		struct tpacket_block_desc *pbd1,
645 646 647 648 649
		struct packet_sock *po, unsigned int stat)
{
	__u32 status = TP_STATUS_USER | stat;

	struct tpacket3_hdr *last_pkt;
650
	struct tpacket_hdr_v1 *h1 = &pbd1->hdr.bh1;
651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677

	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);
}

678
static void prb_thaw_queue(struct tpacket_kbdq_core *pkc)
679 680 681 682 683 684 685 686 687 688 689
{
	pkc->reset_pending_on_curr_blk = 0;
}

/*
 * Side effect of opening a block:
 *
 * 1) prb_queue is thawed.
 * 2) retire_blk_timer is refreshed.
 *
 */
690 691
static void prb_open_block(struct tpacket_kbdq_core *pkc1,
	struct tpacket_block_desc *pbd1)
692 693
{
	struct timespec ts;
694
	struct tpacket_hdr_v1 *h1 = &pbd1->hdr.bh1;
695 696 697 698 699 700 701 702 703 704 705 706 707 708 709

	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
710
		pkc1->nxt_offset = pkc1->pkblk_start + BLK_PLUS_PRIV(pkc1->blk_sizeof_priv);
711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752
		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.
 */
753
static void prb_freeze_queue(struct tpacket_kbdq_core *pkc,
754 755 756 757 758 759 760 761 762 763 764 765 766 767
				  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.
 */
768
static void *prb_dispatch_next_block(struct tpacket_kbdq_core *pkc,
769 770
		struct packet_sock *po)
{
771
	struct tpacket_block_desc *pbd;
772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792

	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;
}

793
static void prb_retire_current_block(struct tpacket_kbdq_core *pkc,
794 795
		struct packet_sock *po, unsigned int status)
{
796
	struct tpacket_block_desc *pbd = GET_CURR_PBLOCK_DESC_FROM_CORE(pkc);
797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823

	/* 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();
}

824
static int prb_curr_blk_in_use(struct tpacket_kbdq_core *pkc,
825
				      struct tpacket_block_desc *pbd)
826 827 828 829
{
	return TP_STATUS_USER & BLOCK_STATUS(pbd);
}

830
static int prb_queue_frozen(struct tpacket_kbdq_core *pkc)
831 832 833 834
{
	return pkc->reset_pending_on_curr_blk;
}

835
static void prb_clear_blk_fill_status(struct packet_ring_buffer *rb)
836
{
837
	struct tpacket_kbdq_core *pkc  = GET_PBDQC_FROM_RB(rb);
838 839 840
	atomic_dec(&pkc->blk_fill_in_prog);
}

841
static void prb_fill_rxhash(struct tpacket_kbdq_core *pkc,
842 843 844 845 846
			struct tpacket3_hdr *ppd)
{
	ppd->hv1.tp_rxhash = skb_get_rxhash(pkc->skb);
}

847
static void prb_clear_rxhash(struct tpacket_kbdq_core *pkc,
848 849 850 851 852
			struct tpacket3_hdr *ppd)
{
	ppd->hv1.tp_rxhash = 0;
}

853
static void prb_fill_vlan_info(struct tpacket_kbdq_core *pkc,
854 855 856 857 858 859
			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 {
860 861
		ppd->hv1.tp_vlan_tci = 0;
		ppd->tp_status = TP_STATUS_AVAILABLE;
862 863 864
	}
}

865
static void prb_run_all_ft_ops(struct tpacket_kbdq_core *pkc,
866 867 868 869 870 871 872 873 874 875
			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);
}

876
static void prb_fill_curr_block(char *curr,
877 878
				struct tpacket_kbdq_core *pkc,
				struct tpacket_block_desc *pbd,
879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899
				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
					    )
{
900 901
	struct tpacket_kbdq_core *pkc;
	struct tpacket_block_desc *pbd;
902 903
	char *curr, *end;

J
Joe Perches 已提交
904
	pkc = GET_PBDQC_FROM_RB(&po->rx_ring);
905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929
	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
930
	end = (char *)pbd + pkc->kblk_size;
931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955

	/* 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;
}

956
static void *packet_current_rx_frame(struct packet_sock *po,
957 958 959 960 961 962 963 964 965 966 967 968 969 970 971
					    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();
972
		return NULL;
973 974 975
	}
}

976
static void *prb_lookup_block(struct packet_sock *po,
977
				     struct packet_ring_buffer *rb,
978
				     unsigned int idx,
979 980
				     int status)
{
981
	struct tpacket_kbdq_core *pkc  = GET_PBDQC_FROM_RB(rb);
982
	struct tpacket_block_desc *pbd = GET_PBLOCK_DESC(pkc, idx);
983 984 985 986 987 988

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

989
static int prb_previous_blk_num(struct packet_ring_buffer *rb)
990 991 992 993 994 995 996 997 998 999
{
	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 */
1000
static void *__prb_previous_block(struct packet_sock *po,
1001 1002 1003 1004 1005 1006 1007
					 struct packet_ring_buffer *rb,
					 int status)
{
	unsigned int previous = prb_previous_blk_num(rb);
	return prb_lookup_block(po, rb, previous, status);
}

1008
static void *packet_previous_rx_frame(struct packet_sock *po,
1009 1010 1011 1012 1013 1014 1015 1016 1017
					     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);
}

1018
static void packet_increment_rx_head(struct packet_sock *po,
1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032
					    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;
	}
}

1033
static void *packet_previous_frame(struct packet_sock *po,
J
Johann Baudy 已提交
1034 1035 1036 1037 1038 1039 1040
		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);
}

1041
static void packet_increment_head(struct packet_ring_buffer *buff)
J
Johann Baudy 已提交
1042 1043 1044 1045
{
	buff->head = buff->head != buff->frame_max ? buff->head+1 : 0;
}

1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068
static bool packet_rcv_has_room(struct packet_sock *po, struct sk_buff *skb)
{
	struct sock *sk = &po->sk;
	bool has_room;

	if (po->prot_hook.func != tpacket_rcv)
		return (atomic_read(&sk->sk_rmem_alloc) + skb->truesize)
			<= sk->sk_rcvbuf;

	spin_lock(&sk->sk_receive_queue.lock);
	if (po->tp_version == TPACKET_V3)
		has_room = prb_lookup_block(po, &po->rx_ring,
					    po->rx_ring.prb_bdqc.kactive_blk_num,
					    TP_STATUS_KERNEL);
	else
		has_room = packet_lookup_frame(po, &po->rx_ring,
					       po->rx_ring.head,
					       TP_STATUS_KERNEL);
	spin_unlock(&sk->sk_receive_queue.lock);

	return has_room;
}

L
Linus Torvalds 已提交
1069 1070
static void packet_sock_destruct(struct sock *sk)
{
1071 1072
	skb_queue_purge(&sk->sk_error_queue);

1073 1074
	WARN_ON(atomic_read(&sk->sk_rmem_alloc));
	WARN_ON(atomic_read(&sk->sk_wmem_alloc));
L
Linus Torvalds 已提交
1075 1076

	if (!sock_flag(sk, SOCK_DEAD)) {
E
Eric Dumazet 已提交
1077
		pr_err("Attempt to release alive packet socket: %p\n", sk);
L
Linus Torvalds 已提交
1078 1079 1080
		return;
	}

1081
	sk_refcnt_debug_dec(sk);
L
Linus Torvalds 已提交
1082 1083
}

D
David S. Miller 已提交
1084 1085 1086 1087 1088 1089 1090 1091 1092 1093
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;
}

1094 1095 1096
static unsigned int fanout_demux_hash(struct packet_fanout *f,
				      struct sk_buff *skb,
				      unsigned int num)
D
David S. Miller 已提交
1097
{
1098
	return (((u64)skb->rxhash) * num) >> 32;
D
David S. Miller 已提交
1099 1100
}

1101 1102 1103
static unsigned int fanout_demux_lb(struct packet_fanout *f,
				    struct sk_buff *skb,
				    unsigned int num)
D
David S. Miller 已提交
1104 1105 1106 1107 1108 1109 1110
{
	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;
1111 1112 1113 1114 1115 1116 1117 1118
	return cur;
}

static unsigned int fanout_demux_cpu(struct packet_fanout *f,
				     struct sk_buff *skb,
				     unsigned int num)
{
	return smp_processor_id() % num;
D
David S. Miller 已提交
1119 1120
}

1121 1122 1123 1124
static unsigned int fanout_demux_rollover(struct packet_fanout *f,
					  struct sk_buff *skb,
					  unsigned int idx, unsigned int skip,
					  unsigned int num)
1125
{
1126
	unsigned int i, j;
1127

1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144
	i = j = min_t(int, f->next[idx], num - 1);
	do {
		if (i != skip && packet_rcv_has_room(pkt_sk(f->arr[i]), skb)) {
			if (i != j)
				f->next[idx] = i;
			return i;
		}
		if (++i == num)
			i = 0;
	} while (i != j);

	return idx;
}

static bool fanout_has_flag(struct packet_fanout *f, u16 flag)
{
	return f->flags & (flag >> 8);
1145 1146 1147 1148
}

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
1149 1150 1151 1152
{
	struct packet_fanout *f = pt->af_packet_priv;
	unsigned int num = f->num_members;
	struct packet_sock *po;
1153
	unsigned int idx;
D
David S. Miller 已提交
1154 1155 1156 1157 1158 1159 1160

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

1161 1162 1163
	switch (f->type) {
	case PACKET_FANOUT_HASH:
	default:
1164
		if (fanout_has_flag(f, PACKET_FANOUT_FLAG_DEFRAG)) {
1165
			skb = ip_check_defrag(skb, IP_DEFRAG_AF_PACKET);
1166 1167 1168 1169
			if (!skb)
				return 0;
		}
		skb_get_rxhash(skb);
1170
		idx = fanout_demux_hash(f, skb, num);
1171 1172
		break;
	case PACKET_FANOUT_LB:
1173
		idx = fanout_demux_lb(f, skb, num);
1174 1175
		break;
	case PACKET_FANOUT_CPU:
1176 1177 1178 1179
		idx = fanout_demux_cpu(f, skb, num);
		break;
	case PACKET_FANOUT_ROLLOVER:
		idx = fanout_demux_rollover(f, skb, 0, (unsigned int) -1, num);
1180
		break;
D
David S. Miller 已提交
1181 1182
	}

1183 1184 1185 1186 1187 1188
	po = pkt_sk(f->arr[idx]);
	if (fanout_has_flag(f, PACKET_FANOUT_FLAG_ROLLOVER) &&
	    unlikely(!packet_rcv_has_room(po, skb))) {
		idx = fanout_demux_rollover(f, skb, idx, idx, num);
		po = pkt_sk(f->arr[idx]);
	}
D
David S. Miller 已提交
1189 1190 1191 1192

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

1193 1194
DEFINE_MUTEX(fanout_mutex);
EXPORT_SYMBOL_GPL(fanout_mutex);
D
David S. Miller 已提交
1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223
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);
}

1224
static bool match_fanout_group(struct packet_type *ptype, struct sock * sk)
1225 1226 1227 1228 1229 1230 1231
{
	if (ptype->af_packet_priv == (void*)((struct packet_sock *)sk)->fanout)
		return true;

	return false;
}

1232
static int fanout_add(struct sock *sk, u16 id, u16 type_flags)
D
David S. Miller 已提交
1233 1234 1235
{
	struct packet_sock *po = pkt_sk(sk);
	struct packet_fanout *f, *match;
1236
	u8 type = type_flags & 0xff;
1237
	u8 flags = type_flags >> 8;
D
David S. Miller 已提交
1238 1239 1240
	int err;

	switch (type) {
1241 1242 1243
	case PACKET_FANOUT_ROLLOVER:
		if (type_flags & PACKET_FANOUT_FLAG_ROLLOVER)
			return -EINVAL;
D
David S. Miller 已提交
1244 1245
	case PACKET_FANOUT_HASH:
	case PACKET_FANOUT_LB:
1246
	case PACKET_FANOUT_CPU:
D
David S. Miller 已提交
1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266
		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
1267
	err = -EINVAL;
1268
	if (match && match->flags != flags)
E
Eric Dumazet 已提交
1269
		goto out;
D
David S. Miller 已提交
1270
	if (!match) {
E
Eric Dumazet 已提交
1271
		err = -ENOMEM;
D
David S. Miller 已提交
1272
		match = kzalloc(sizeof(*match), GFP_KERNEL);
E
Eric Dumazet 已提交
1273 1274 1275 1276 1277
		if (!match)
			goto out;
		write_pnet(&match->net, sock_net(sk));
		match->id = id;
		match->type = type;
1278
		match->flags = flags;
E
Eric Dumazet 已提交
1279 1280 1281 1282 1283 1284 1285 1286
		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;
1287
		match->prot_hook.id_match = match_fanout_group;
E
Eric Dumazet 已提交
1288 1289
		dev_add_pack(&match->prot_hook);
		list_add(&match->list, &fanout_list);
D
David S. Miller 已提交
1290
	}
E
Eric Dumazet 已提交
1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301
	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
1302 1303
		}
	}
E
Eric Dumazet 已提交
1304
out:
D
David S. Miller 已提交
1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317
	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;

1318
	mutex_lock(&fanout_mutex);
D
David S. Miller 已提交
1319 1320 1321 1322 1323 1324 1325 1326 1327
	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
1328

1329
static const struct proto_ops packet_ops;
L
Linus Torvalds 已提交
1330

1331
static const struct proto_ops packet_ops_spkt;
L
Linus Torvalds 已提交
1332

E
Eric Dumazet 已提交
1333 1334
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
1335 1336 1337 1338 1339 1340 1341 1342 1343 1344
{
	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;
1345

L
Linus Torvalds 已提交
1346 1347 1348 1349 1350 1351 1352
	/*
	 *	Yank back the headers [hope the device set this
	 *	right or kerboom...]
	 *
	 *	Incoming packets have ll header pulled,
	 *	push it back.
	 *
1353
	 *	For outgoing ones skb->data == skb_mac_header(skb)
L
Linus Torvalds 已提交
1354 1355 1356 1357 1358 1359
	 *	so that this procedure is noop.
	 */

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

O
Octavian Purdila 已提交
1360
	if (!net_eq(dev_net(dev), sock_net(sk)))
1361 1362
		goto out;

E
Eric Dumazet 已提交
1363 1364
	skb = skb_share_check(skb, GFP_ATOMIC);
	if (skb == NULL)
L
Linus Torvalds 已提交
1365 1366 1367
		goto oom;

	/* drop any routing info */
E
Eric Dumazet 已提交
1368
	skb_dst_drop(skb);
L
Linus Torvalds 已提交
1369

1370 1371 1372
	/* drop conntrack reference */
	nf_reset(skb);

1373
	spkt = &PACKET_SKB_CB(skb)->sa.pkt;
L
Linus Torvalds 已提交
1374

1375
	skb_push(skb, skb->data - skb_mac_header(skb));
L
Linus Torvalds 已提交
1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389

	/*
	 *	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
1390
	if (sock_queue_rcv_skb(sk, skb) == 0)
L
Linus Torvalds 已提交
1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403
		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
 */
1404

L
Linus Torvalds 已提交
1405 1406 1407 1408
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
1409
	struct sockaddr_pkt *saddr = (struct sockaddr_pkt *)msg->msg_name;
1410
	struct sk_buff *skb = NULL;
L
Linus Torvalds 已提交
1411
	struct net_device *dev;
E
Eric Dumazet 已提交
1412
	__be16 proto = 0;
L
Linus Torvalds 已提交
1413
	int err;
1414
	int extra_len = 0;
1415

L
Linus Torvalds 已提交
1416
	/*
1417
	 *	Get and verify the address.
L
Linus Torvalds 已提交
1418 1419
	 */

E
Eric Dumazet 已提交
1420
	if (saddr) {
L
Linus Torvalds 已提交
1421
		if (msg->msg_namelen < sizeof(struct sockaddr))
E
Eric Dumazet 已提交
1422 1423 1424 1425 1426
			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
1427 1428

	/*
1429
	 *	Find the device first to size check it
L
Linus Torvalds 已提交
1430 1431
	 */

1432
	saddr->spkt_device[sizeof(saddr->spkt_device) - 1] = 0;
1433
retry:
E
Eric Dumazet 已提交
1434 1435
	rcu_read_lock();
	dev = dev_get_by_name_rcu(sock_net(sk), saddr->spkt_device);
L
Linus Torvalds 已提交
1436 1437 1438
	err = -ENODEV;
	if (dev == NULL)
		goto out_unlock;
1439

1440 1441 1442 1443
	err = -ENETDOWN;
	if (!(dev->flags & IFF_UP))
		goto out_unlock;

L
Linus Torvalds 已提交
1444
	/*
E
Eric Dumazet 已提交
1445 1446
	 * 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
1447
	 */
1448

1449 1450 1451 1452 1453 1454 1455 1456
	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
1457
	err = -EMSGSIZE;
1458
	if (len > dev->mtu + dev->hard_header_len + VLAN_HLEN + extra_len)
L
Linus Torvalds 已提交
1459 1460
		goto out_unlock;

1461 1462
	if (!skb) {
		size_t reserved = LL_RESERVED_SPACE(dev);
1463
		int tlen = dev->needed_tailroom;
1464 1465 1466
		unsigned int hhlen = dev->header_ops ? dev->hard_header_len : 0;

		rcu_read_unlock();
1467
		skb = sock_wmalloc(sk, len + reserved + tlen, 0, GFP_KERNEL);
1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487
		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
1488 1489
	}

1490
	if (len > (dev->mtu + dev->hard_header_len + extra_len)) {
1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502
		/* 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;
		}
	}
1503

L
Linus Torvalds 已提交
1504 1505 1506
	skb->protocol = proto;
	skb->dev = dev;
	skb->priority = sk->sk_priority;
1507
	skb->mark = sk->sk_mark;
1508
	err = sock_tx_timestamp(sk, &skb_shinfo(skb)->tx_flags);
1509 1510
	if (err < 0)
		goto out_unlock;
L
Linus Torvalds 已提交
1511

1512 1513 1514
	if (unlikely(extra_len == 4))
		skb->no_fcs = 1;

1515
	skb_probe_transport_header(skb, 0);
1516

L
Linus Torvalds 已提交
1517
	dev_queue_xmit(skb);
E
Eric Dumazet 已提交
1518
	rcu_read_unlock();
E
Eric Dumazet 已提交
1519
	return len;
L
Linus Torvalds 已提交
1520 1521

out_unlock:
E
Eric Dumazet 已提交
1522
	rcu_read_unlock();
1523 1524
out_free:
	kfree_skb(skb);
L
Linus Torvalds 已提交
1525 1526 1527
	return err;
}

1528
static unsigned int run_filter(const struct sk_buff *skb,
E
Eric Dumazet 已提交
1529
				      const struct sock *sk,
D
David S. Miller 已提交
1530
				      unsigned int res)
L
Linus Torvalds 已提交
1531 1532
{
	struct sk_filter *filter;
1533

1534 1535
	rcu_read_lock();
	filter = rcu_dereference(sk->sk_filter);
D
David S. Miller 已提交
1536
	if (filter != NULL)
1537
		res = SK_RUN_FILTER(filter, skb);
1538
	rcu_read_unlock();
L
Linus Torvalds 已提交
1539

D
David S. Miller 已提交
1540
	return res;
L
Linus Torvalds 已提交
1541 1542 1543
}

/*
E
Eric Dumazet 已提交
1544 1545 1546 1547 1548 1549 1550 1551 1552
 * 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
1553 1554
 */

E
Eric Dumazet 已提交
1555 1556
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
1557 1558 1559 1560
{
	struct sock *sk;
	struct sockaddr_ll *sll;
	struct packet_sock *po;
E
Eric Dumazet 已提交
1561
	u8 *skb_head = skb->data;
L
Linus Torvalds 已提交
1562
	int skb_len = skb->len;
D
David S. Miller 已提交
1563
	unsigned int snaplen, res;
L
Linus Torvalds 已提交
1564 1565 1566 1567 1568 1569 1570

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

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

O
Octavian Purdila 已提交
1571
	if (!net_eq(dev_net(dev), sock_net(sk)))
1572 1573
		goto drop;

L
Linus Torvalds 已提交
1574 1575
	skb->dev = dev;

1576
	if (dev->header_ops) {
L
Linus Torvalds 已提交
1577
		/* The device has an explicit notion of ll header,
E
Eric Dumazet 已提交
1578 1579 1580 1581 1582
		 * 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
1583 1584
		 */
		if (sk->sk_type != SOCK_DGRAM)
1585
			skb_push(skb, skb->data - skb_mac_header(skb));
L
Linus Torvalds 已提交
1586 1587
		else if (skb->pkt_type == PACKET_OUTGOING) {
			/* Special case: outgoing packets have ll header at head */
1588
			skb_pull(skb, skb_network_offset(skb));
L
Linus Torvalds 已提交
1589 1590 1591 1592 1593
		}
	}

	snaplen = skb->len;

D
David S. Miller 已提交
1594 1595
	res = run_filter(skb, sk, snaplen);
	if (!res)
1596
		goto drop_n_restore;
D
David S. Miller 已提交
1597 1598
	if (snaplen > res)
		snaplen = res;
L
Linus Torvalds 已提交
1599

E
Eric Dumazet 已提交
1600
	if (atomic_read(&sk->sk_rmem_alloc) >= sk->sk_rcvbuf)
L
Linus Torvalds 已提交
1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611
		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;
		}
1612
		consume_skb(skb);
L
Linus Torvalds 已提交
1613 1614 1615
		skb = nskb;
	}

1616 1617 1618 1619
	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
1620 1621 1622 1623
	sll->sll_family = AF_PACKET;
	sll->sll_hatype = dev->type;
	sll->sll_protocol = skb->protocol;
	sll->sll_pkttype = skb->pkt_type;
1624
	if (unlikely(po->origdev))
1625 1626 1627
		sll->sll_ifindex = orig_dev->ifindex;
	else
		sll->sll_ifindex = dev->ifindex;
L
Linus Torvalds 已提交
1628

S
Stephen Hemminger 已提交
1629
	sll->sll_halen = dev_parse_header(skb, sll->sll_addr);
L
Linus Torvalds 已提交
1630

1631
	PACKET_SKB_CB(skb)->origlen = skb->len;
1632

L
Linus Torvalds 已提交
1633 1634 1635 1636 1637
	if (pskb_trim(skb, snaplen))
		goto drop_n_acct;

	skb_set_owner_r(skb, sk);
	skb->dev = NULL;
E
Eric Dumazet 已提交
1638
	skb_dst_drop(skb);
L
Linus Torvalds 已提交
1639

1640 1641 1642
	/* drop conntrack reference */
	nf_reset(skb);

L
Linus Torvalds 已提交
1643 1644
	spin_lock(&sk->sk_receive_queue.lock);
	po->stats.tp_packets++;
1645
	skb->dropcount = atomic_read(&sk->sk_drops);
L
Linus Torvalds 已提交
1646 1647 1648 1649 1650 1651
	__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:
1652 1653 1654 1655
	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
1656 1657 1658 1659 1660 1661 1662

drop_n_restore:
	if (skb_head != skb->data && skb_shared(skb)) {
		skb->data = skb_head;
		skb->len = skb_len;
	}
drop:
1663
	consume_skb(skb);
L
Linus Torvalds 已提交
1664 1665 1666
	return 0;
}

E
Eric Dumazet 已提交
1667 1668
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
1669 1670 1671 1672
{
	struct sock *sk;
	struct packet_sock *po;
	struct sockaddr_ll *sll;
1673 1674 1675
	union {
		struct tpacket_hdr *h1;
		struct tpacket2_hdr *h2;
1676
		struct tpacket3_hdr *h3;
1677 1678
		void *raw;
	} h;
E
Eric Dumazet 已提交
1679
	u8 *skb_head = skb->data;
L
Linus Torvalds 已提交
1680
	int skb_len = skb->len;
D
David S. Miller 已提交
1681
	unsigned int snaplen, res;
1682
	unsigned long status = TP_STATUS_USER;
1683
	unsigned short macoff, netoff, hdrlen;
L
Linus Torvalds 已提交
1684
	struct sk_buff *copy_skb = NULL;
1685
	struct timeval tv;
1686
	struct timespec ts;
1687
	struct skb_shared_hwtstamps *shhwtstamps = skb_hwtstamps(skb);
L
Linus Torvalds 已提交
1688 1689 1690 1691 1692 1693 1694

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

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

O
Octavian Purdila 已提交
1695
	if (!net_eq(dev_net(dev), sock_net(sk)))
1696 1697
		goto drop;

1698
	if (dev->header_ops) {
L
Linus Torvalds 已提交
1699
		if (sk->sk_type != SOCK_DGRAM)
1700
			skb_push(skb, skb->data - skb_mac_header(skb));
L
Linus Torvalds 已提交
1701 1702
		else if (skb->pkt_type == PACKET_OUTGOING) {
			/* Special case: outgoing packets have ll header at head */
1703
			skb_pull(skb, skb_network_offset(skb));
L
Linus Torvalds 已提交
1704 1705 1706
		}
	}

1707 1708 1709
	if (skb->ip_summed == CHECKSUM_PARTIAL)
		status |= TP_STATUS_CSUMNOTREADY;

L
Linus Torvalds 已提交
1710 1711
	snaplen = skb->len;

D
David S. Miller 已提交
1712 1713
	res = run_filter(skb, sk, snaplen);
	if (!res)
1714
		goto drop_n_restore;
D
David S. Miller 已提交
1715 1716
	if (snaplen > res)
		snaplen = res;
L
Linus Torvalds 已提交
1717 1718

	if (sk->sk_type == SOCK_DGRAM) {
1719 1720
		macoff = netoff = TPACKET_ALIGN(po->tp_hdrlen) + 16 +
				  po->tp_reserve;
L
Linus Torvalds 已提交
1721
	} else {
1722
		unsigned int maclen = skb_network_offset(skb);
1723
		netoff = TPACKET_ALIGN(po->tp_hdrlen +
1724 1725
				       (maclen < 16 ? 16 : maclen)) +
			po->tp_reserve;
L
Linus Torvalds 已提交
1726 1727
		macoff = netoff - maclen;
	}
1728 1729 1730
	if (po->tp_version <= TPACKET_V2) {
		if (macoff + snaplen > po->rx_ring.frame_size) {
			if (po->copy_thresh &&
E
Eric Dumazet 已提交
1731
			    atomic_read(&sk->sk_rmem_alloc) < sk->sk_rcvbuf) {
1732 1733 1734 1735 1736 1737 1738 1739
				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
1740
			}
1741 1742 1743
			snaplen = po->rx_ring.frame_size - macoff;
			if ((int)snaplen < 0)
				snaplen = 0;
L
Linus Torvalds 已提交
1744 1745 1746
		}
	}
	spin_lock(&sk->sk_receive_queue.lock);
1747 1748
	h.raw = packet_current_rx_frame(po, skb,
					TP_STATUS_KERNEL, (macoff+snaplen));
1749
	if (!h.raw)
L
Linus Torvalds 已提交
1750
		goto ring_is_full;
1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761
	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
1762 1763 1764 1765 1766 1767 1768
	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);

1769
	skb_copy_bits(skb, 0, h.raw + macoff, snaplen);
L
Linus Torvalds 已提交
1770

1771 1772 1773 1774 1775 1776
	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;
1777 1778 1779 1780 1781 1782 1783
		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)
1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795
			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;
1796 1797 1798 1799 1800 1801 1802
		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)
1803 1804 1805 1806 1807
			ts = ktime_to_timespec(skb->tstamp);
		else
			getnstimeofday(&ts);
		h.h2->tp_sec = ts.tv_sec;
		h.h2->tp_nsec = ts.tv_nsec;
1808 1809 1810 1811 1812 1813
		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;
		}
1814
		h.h2->tp_padding = 0;
1815 1816
		hdrlen = sizeof(*h.h2);
		break;
1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839
	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;
1840 1841 1842
	default:
		BUG();
	}
L
Linus Torvalds 已提交
1843

1844
	sll = h.raw + TPACKET_ALIGN(hdrlen);
S
Stephen Hemminger 已提交
1845
	sll->sll_halen = dev_parse_header(skb, sll->sll_addr);
L
Linus Torvalds 已提交
1846 1847 1848 1849
	sll->sll_family = AF_PACKET;
	sll->sll_hatype = dev->type;
	sll->sll_protocol = skb->protocol;
	sll->sll_pkttype = skb->pkt_type;
1850
	if (unlikely(po->origdev))
1851 1852 1853
		sll->sll_ifindex = orig_dev->ifindex;
	else
		sll->sll_ifindex = dev->ifindex;
L
Linus Torvalds 已提交
1854

R
Ralf Baechle 已提交
1855
	smp_mb();
1856
#if ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE == 1
L
Linus Torvalds 已提交
1857
	{
1858 1859
		u8 *start, *end;

1860 1861 1862 1863 1864 1865
		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));
		}
1866
		smp_wmb();
L
Linus Torvalds 已提交
1867
	}
1868
#endif
1869 1870 1871 1872
	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
1873 1874 1875 1876 1877 1878 1879 1880 1881

	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:
1882
	kfree_skb(skb);
L
Linus Torvalds 已提交
1883 1884 1885 1886 1887 1888 1889
	return 0;

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

	sk->sk_data_ready(sk, 0);
1890
	kfree_skb(copy_skb);
L
Linus Torvalds 已提交
1891 1892 1893
	goto drop_n_restore;
}

J
Johann Baudy 已提交
1894 1895 1896
static void tpacket_destruct_skb(struct sk_buff *skb)
{
	struct packet_sock *po = pkt_sk(skb->sk);
E
Eric Dumazet 已提交
1897
	void *ph;
L
Linus Torvalds 已提交
1898

J
Johann Baudy 已提交
1899 1900 1901 1902 1903 1904 1905 1906 1907 1908
	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
1909 1910
static int tpacket_fill_skb(struct packet_sock *po, struct sk_buff *skb,
		void *frame, struct net_device *dev, int size_max,
1911
		__be16 proto, unsigned char *addr, int hlen)
J
Johann Baudy 已提交
1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928
{
	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;
1929
	skb->mark = po->sk.sk_mark;
J
Johann Baudy 已提交
1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940
	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
1941
		pr_err("packet size is too long (%d > %d)\n", tp_len, size_max);
J
Johann Baudy 已提交
1942 1943 1944
		return -EMSGSIZE;
	}

1945
	skb_reserve(skb, hlen);
J
Johann Baudy 已提交
1946
	skb_reset_network_header(skb);
1947
	skb_probe_transport_header(skb, 0);
1948

1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977
	if (po->tp_tx_has_off) {
		int off_min, off_max, off;
		off_min = po->tp_hdrlen - sizeof(struct sockaddr_ll);
		off_max = po->tx_ring.frame_size - tp_len;
		if (sock->type == SOCK_DGRAM) {
			switch (po->tp_version) {
			case TPACKET_V2:
				off = ph.h2->tp_net;
				break;
			default:
				off = ph.h1->tp_net;
				break;
			}
		} else {
			switch (po->tp_version) {
			case TPACKET_V2:
				off = ph.h2->tp_mac;
				break;
			default:
				off = ph.h1->tp_mac;
				break;
			}
		}
		if (unlikely((off < off_min) || (off_max < off)))
			return -EINVAL;
		data = ph.raw + off;
	} else {
		data = ph.raw + po->tp_hdrlen - sizeof(struct sockaddr_ll);
	}
J
Johann Baudy 已提交
1978 1979 1980 1981 1982 1983 1984
	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
1985
	} else if (dev->hard_header_len) {
J
Johann Baudy 已提交
1986 1987
		/* net device doesn't like empty head */
		if (unlikely(tp_len <= dev->hard_header_len)) {
E
Eric Dumazet 已提交
1988 1989
			pr_err("packet size is too short (%d < %d)\n",
			       tp_len, dev->hard_header_len);
J
Johann Baudy 已提交
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
			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
2016 2017
			pr_err("Packet exceed the number of skb frags(%lu)\n",
			       MAX_SKB_FRAGS);
J
Johann Baudy 已提交
2018 2019 2020
			return -EFAULT;
		}

2021 2022
		page = pgv_to_page(data);
		data += len;
J
Johann Baudy 已提交
2023 2024
		flush_dcache_page(page);
		get_page(page);
2025
		skb_fill_page_desc(skb, nr_frags, page, offset, len);
J
Johann Baudy 已提交
2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039
		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;
2040 2041
	bool need_rls_dev = false;
	int err, reserve = 0;
E
Eric Dumazet 已提交
2042 2043
	void *ph;
	struct sockaddr_ll *saddr = (struct sockaddr_ll *)msg->msg_name;
J
Johann Baudy 已提交
2044 2045 2046
	int tp_len, size_max;
	unsigned char *addr;
	int len_sum = 0;
2047
	int status = TP_STATUS_AVAILABLE;
2048
	int hlen, tlen;
J
Johann Baudy 已提交
2049 2050 2051 2052

	mutex_lock(&po->pg_vec_lock);

	if (saddr == NULL) {
2053
		dev = po->prot_hook.dev;
J
Johann Baudy 已提交
2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065
		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;
2066 2067
		dev = dev_get_by_index(sock_net(&po->sk), saddr->sll_ifindex);
		need_rls_dev = true;
J
Johann Baudy 已提交
2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080
	}

	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
2081
		- (po->tp_hdrlen - sizeof(struct sockaddr_ll));
J
Johann Baudy 已提交
2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095

	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;
2096 2097
		hlen = LL_RESERVED_SPACE(dev);
		tlen = dev->needed_tailroom;
J
Johann Baudy 已提交
2098
		skb = sock_alloc_send_skb(&po->sk,
2099
				hlen + tlen + sizeof(struct sockaddr_ll),
J
Johann Baudy 已提交
2100 2101 2102 2103 2104 2105
				0, &err);

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

		tp_len = tpacket_fill_skb(po, skb, ph, dev, size_max, proto,
2106
				addr, hlen);
J
Johann Baudy 已提交
2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127

		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);
2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141
		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
2142 2143
		packet_increment_head(&po->tx_ring);
		len_sum += tp_len;
2144 2145 2146 2147
	} while (likely((ph != NULL) ||
			((!(msg->msg_flags & MSG_DONTWAIT)) &&
			 (atomic_read(&po->tx_ring.pending))))
		);
J
Johann Baudy 已提交
2148 2149 2150 2151 2152 2153 2154 2155

	err = len_sum;
	goto out_put;

out_status:
	__packet_set_status(po, ph, status);
	kfree_skb(skb);
out_put:
2156 2157
	if (need_rls_dev)
		dev_put(dev);
J
Johann Baudy 已提交
2158 2159 2160 2161 2162
out:
	mutex_unlock(&po->pg_vec_lock);
	return err;
}

2163 2164 2165 2166
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)
2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186
{
	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
2187
static int packet_snd(struct socket *sock,
L
Linus Torvalds 已提交
2188 2189 2190
			  struct msghdr *msg, size_t len)
{
	struct sock *sk = sock->sk;
E
Eric Dumazet 已提交
2191
	struct sockaddr_ll *saddr = (struct sockaddr_ll *)msg->msg_name;
L
Linus Torvalds 已提交
2192 2193
	struct sk_buff *skb;
	struct net_device *dev;
A
Al Viro 已提交
2194
	__be16 proto;
2195
	bool need_rls_dev = false;
L
Linus Torvalds 已提交
2196
	unsigned char *addr;
2197
	int err, reserve = 0;
2198 2199 2200 2201 2202
	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;
2203
	int hlen, tlen;
2204
	int extra_len = 0;
L
Linus Torvalds 已提交
2205 2206

	/*
2207
	 *	Get and verify the address.
L
Linus Torvalds 已提交
2208
	 */
2209

L
Linus Torvalds 已提交
2210
	if (saddr == NULL) {
2211
		dev = po->prot_hook.dev;
L
Linus Torvalds 已提交
2212 2213 2214 2215 2216 2217
		proto	= po->num;
		addr	= NULL;
	} else {
		err = -EINVAL;
		if (msg->msg_namelen < sizeof(struct sockaddr_ll))
			goto out;
2218 2219
		if (msg->msg_namelen < (saddr->sll_halen + offsetof(struct sockaddr_ll, sll_addr)))
			goto out;
L
Linus Torvalds 已提交
2220 2221
		proto	= saddr->sll_protocol;
		addr	= saddr->sll_addr;
2222 2223
		dev = dev_get_by_index(sock_net(sk), saddr->sll_ifindex);
		need_rls_dev = true;
L
Linus Torvalds 已提交
2224 2225 2226 2227 2228 2229 2230 2231
	}

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

2232 2233 2234 2235
	err = -ENETDOWN;
	if (!(dev->flags & IFF_UP))
		goto out_unlock;

2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283
	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;

		}
	}

2284 2285 2286 2287 2288 2289 2290 2291
	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
2292
	err = -EMSGSIZE;
2293
	if (!gso_type && (len > dev->mtu + reserve + VLAN_HLEN + extra_len))
L
Linus Torvalds 已提交
2294 2295
		goto out_unlock;

2296
	err = -ENOBUFS;
2297 2298 2299
	hlen = LL_RESERVED_SPACE(dev);
	tlen = dev->needed_tailroom;
	skb = packet_alloc_skb(sk, hlen + tlen, hlen, len, vnet_hdr.hdr_len,
2300
			       msg->msg_flags & MSG_DONTWAIT, &err);
E
Eric Dumazet 已提交
2301
	if (skb == NULL)
L
Linus Torvalds 已提交
2302 2303
		goto out_unlock;

2304
	skb_set_network_header(skb, reserve);
L
Linus Torvalds 已提交
2305

2306 2307
	err = -EINVAL;
	if (sock->type == SOCK_DGRAM &&
2308
	    (offset = dev_hard_header(skb, dev, ntohs(proto), addr, NULL, len)) < 0)
2309
		goto out_free;
L
Linus Torvalds 已提交
2310 2311

	/* Returns -EFAULT on error */
2312
	err = skb_copy_datagram_from_iovec(skb, offset, msg->msg_iov, 0, len);
L
Linus Torvalds 已提交
2313 2314
	if (err)
		goto out_free;
2315
	err = sock_tx_timestamp(sk, &skb_shinfo(skb)->tx_flags);
2316 2317
	if (err < 0)
		goto out_free;
L
Linus Torvalds 已提交
2318

2319
	if (!gso_type && (len > dev->mtu + reserve + extra_len)) {
2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332
		/* 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
2333 2334 2335
	skb->protocol = proto;
	skb->dev = dev;
	skb->priority = sk->sk_priority;
2336
	skb->mark = sk->sk_mark;
L
Linus Torvalds 已提交
2337

2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356
	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;
	}

2357
	skb_probe_transport_header(skb, reserve);
2358

2359 2360 2361
	if (unlikely(extra_len == 4))
		skb->no_fcs = 1;

L
Linus Torvalds 已提交
2362 2363 2364 2365 2366 2367 2368 2369
	/*
	 *	Now send it
	 */

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

2370 2371
	if (need_rls_dev)
		dev_put(dev);
L
Linus Torvalds 已提交
2372

E
Eric Dumazet 已提交
2373
	return len;
L
Linus Torvalds 已提交
2374 2375 2376 2377

out_free:
	kfree_skb(skb);
out_unlock:
2378
	if (dev && need_rls_dev)
L
Linus Torvalds 已提交
2379 2380 2381 2382 2383
		dev_put(dev);
out:
	return err;
}

J
Johann Baudy 已提交
2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394
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
2395 2396 2397 2398 2399 2400 2401 2402 2403
/*
 *	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;
2404
	struct net *net;
2405
	union tpacket_req_u req_u;
L
Linus Torvalds 已提交
2406 2407 2408 2409

	if (!sk)
		return 0;

2410
	net = sock_net(sk);
L
Linus Torvalds 已提交
2411 2412
	po = pkt_sk(sk);

2413
	mutex_lock(&net->packet.sklist_lock);
2414
	sk_del_node_init_rcu(sk);
2415 2416 2417
	mutex_unlock(&net->packet.sklist_lock);

	preempt_disable();
2418
	sock_prot_inuse_add(net, sk->sk_prot, -1);
2419
	preempt_enable();
L
Linus Torvalds 已提交
2420

2421
	spin_lock(&po->bind_lock);
2422
	unregister_prot_hook(sk, false);
2423 2424 2425 2426
	if (po->prot_hook.dev) {
		dev_put(po->prot_hook.dev);
		po->prot_hook.dev = NULL;
	}
2427
	spin_unlock(&po->bind_lock);
L
Linus Torvalds 已提交
2428 2429 2430

	packet_flush_mclist(sk);

2431 2432
	if (po->rx_ring.pg_vec) {
		memset(&req_u, 0, sizeof(req_u));
2433
		packet_set_ring(sk, &req_u, 1, 0);
2434
	}
J
Johann Baudy 已提交
2435

2436 2437
	if (po->tx_ring.pg_vec) {
		memset(&req_u, 0, sizeof(req_u));
2438
		packet_set_ring(sk, &req_u, 1, 1);
2439
	}
L
Linus Torvalds 已提交
2440

D
David S. Miller 已提交
2441 2442
	fanout_release(sk);

2443
	synchronize_net();
L
Linus Torvalds 已提交
2444 2445 2446 2447 2448 2449 2450 2451 2452
	/*
	 *	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);
2453
	sk_refcnt_debug_release(sk);
L
Linus Torvalds 已提交
2454 2455 2456 2457 2458 2459 2460 2461 2462

	sock_put(sk);
	return 0;
}

/*
 *	Attach a packet hook.
 */

A
Al Viro 已提交
2463
static int packet_do_bind(struct sock *sk, struct net_device *dev, __be16 protocol)
L
Linus Torvalds 已提交
2464 2465
{
	struct packet_sock *po = pkt_sk(sk);
D
David S. Miller 已提交
2466

2467 2468 2469 2470
	if (po->fanout) {
		if (dev)
			dev_put(dev);

D
David S. Miller 已提交
2471
		return -EINVAL;
2472
	}
L
Linus Torvalds 已提交
2473 2474 2475 2476

	lock_sock(sk);

	spin_lock(&po->bind_lock);
2477
	unregister_prot_hook(sk, true);
L
Linus Torvalds 已提交
2478 2479
	po->num = protocol;
	po->prot_hook.type = protocol;
2480 2481
	if (po->prot_hook.dev)
		dev_put(po->prot_hook.dev);
L
Linus Torvalds 已提交
2482 2483 2484 2485 2486 2487 2488
	po->prot_hook.dev = dev;

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

	if (protocol == 0)
		goto out_unlock;

2489
	if (!dev || (dev->flags & IFF_UP)) {
2490
		register_prot_hook(sk);
2491 2492 2493 2494
	} else {
		sk->sk_err = ENETDOWN;
		if (!sock_flag(sk, SOCK_DEAD))
			sk->sk_error_report(sk);
L
Linus Torvalds 已提交
2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506
	}

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
2507 2508
static int packet_bind_spkt(struct socket *sock, struct sockaddr *uaddr,
			    int addr_len)
L
Linus Torvalds 已提交
2509
{
E
Eric Dumazet 已提交
2510
	struct sock *sk = sock->sk;
L
Linus Torvalds 已提交
2511 2512 2513
	char name[15];
	struct net_device *dev;
	int err = -ENODEV;
2514

L
Linus Torvalds 已提交
2515 2516 2517
	/*
	 *	Check legality
	 */
2518

2519
	if (addr_len != sizeof(struct sockaddr))
L
Linus Torvalds 已提交
2520
		return -EINVAL;
E
Eric Dumazet 已提交
2521
	strlcpy(name, uaddr->sa_data, sizeof(name));
L
Linus Torvalds 已提交
2522

2523
	dev = dev_get_by_name(sock_net(sk), name);
2524
	if (dev)
L
Linus Torvalds 已提交
2525 2526 2527 2528 2529 2530
		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
2531 2532
	struct sockaddr_ll *sll = (struct sockaddr_ll *)uaddr;
	struct sock *sk = sock->sk;
L
Linus Torvalds 已提交
2533 2534 2535 2536 2537 2538 2539
	struct net_device *dev = NULL;
	int err;


	/*
	 *	Check legality
	 */
2540

L
Linus Torvalds 已提交
2541 2542 2543 2544 2545 2546 2547
	if (addr_len < sizeof(struct sockaddr_ll))
		return -EINVAL;
	if (sll->sll_family != AF_PACKET)
		return -EINVAL;

	if (sll->sll_ifindex) {
		err = -ENODEV;
2548
		dev = dev_get_by_index(sock_net(sk), sll->sll_ifindex);
L
Linus Torvalds 已提交
2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564
		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),
};

/*
2565
 *	Create a packet of type SOCK_PACKET.
L
Linus Torvalds 已提交
2566 2567
 */

2568 2569
static int packet_create(struct net *net, struct socket *sock, int protocol,
			 int kern)
L
Linus Torvalds 已提交
2570 2571 2572
{
	struct sock *sk;
	struct packet_sock *po;
A
Al Viro 已提交
2573
	__be16 proto = (__force __be16)protocol; /* weird, but documented */
L
Linus Torvalds 已提交
2574 2575
	int err;

2576
	if (!ns_capable(net->user_ns, CAP_NET_RAW))
L
Linus Torvalds 已提交
2577
		return -EPERM;
2578 2579
	if (sock->type != SOCK_DGRAM && sock->type != SOCK_RAW &&
	    sock->type != SOCK_PACKET)
L
Linus Torvalds 已提交
2580 2581 2582 2583 2584
		return -ESOCKTNOSUPPORT;

	sock->state = SS_UNCONNECTED;

	err = -ENOBUFS;
2585
	sk = sk_alloc(net, PF_PACKET, GFP_KERNEL, &packet_proto);
L
Linus Torvalds 已提交
2586 2587 2588 2589 2590 2591
	if (sk == NULL)
		goto out;

	sock->ops = &packet_ops;
	if (sock->type == SOCK_PACKET)
		sock->ops = &packet_ops_spkt;
2592

L
Linus Torvalds 已提交
2593 2594 2595 2596
	sock_init_data(sock, sk);

	po = pkt_sk(sk);
	sk->sk_family = PF_PACKET;
A
Al Viro 已提交
2597
	po->num = proto;
L
Linus Torvalds 已提交
2598 2599

	sk->sk_destruct = packet_sock_destruct;
2600
	sk_refcnt_debug_inc(sk);
L
Linus Torvalds 已提交
2601 2602 2603 2604 2605 2606

	/*
	 *	Attach a protocol block
	 */

	spin_lock_init(&po->bind_lock);
2607
	mutex_init(&po->pg_vec_lock);
L
Linus Torvalds 已提交
2608
	po->prot_hook.func = packet_rcv;
2609

L
Linus Torvalds 已提交
2610 2611
	if (sock->type == SOCK_PACKET)
		po->prot_hook.func = packet_rcv_spkt;
2612

L
Linus Torvalds 已提交
2613 2614
	po->prot_hook.af_packet_priv = sk;

A
Al Viro 已提交
2615 2616
	if (proto) {
		po->prot_hook.type = proto;
2617
		register_prot_hook(sk);
L
Linus Torvalds 已提交
2618 2619
	}

2620
	mutex_lock(&net->packet.sklist_lock);
2621
	sk_add_node_rcu(sk, &net->packet.sklist);
2622 2623 2624
	mutex_unlock(&net->packet.sklist_lock);

	preempt_disable();
2625
	sock_prot_inuse_add(net, &packet_proto, 1);
2626
	preempt_enable();
2627

E
Eric Dumazet 已提交
2628
	return 0;
L
Linus Torvalds 已提交
2629 2630 2631 2632
out:
	return err;
}

2633 2634 2635 2636 2637 2638 2639 2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 2662 2663 2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 2677
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
2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688
/*
 *	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;
2689
	struct sockaddr_ll *sll;
2690
	int vnet_hdr_len = 0;
L
Linus Torvalds 已提交
2691 2692

	err = -EINVAL;
2693
	if (flags & ~(MSG_PEEK|MSG_DONTWAIT|MSG_TRUNC|MSG_CMSG_COMPAT|MSG_ERRQUEUE))
L
Linus Torvalds 已提交
2694 2695 2696 2697 2698 2699 2700 2701
		goto out;

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

2702 2703 2704 2705 2706
	if (flags & MSG_ERRQUEUE) {
		err = packet_recv_error(sk, msg, len);
		goto out;
	}

L
Linus Torvalds 已提交
2707 2708 2709 2710 2711 2712 2713 2714 2715
	/*
	 *	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
2716
	skb = skb_recv_datagram(sk, flags, flags & MSG_DONTWAIT, &err);
L
Linus Torvalds 已提交
2717 2718

	/*
2719
	 *	An error occurred so return it. Because skb_recv_datagram()
L
Linus Torvalds 已提交
2720 2721 2722 2723
	 *	handles the blocking we don't see and worry about blocking
	 *	retries.
	 */

2724
	if (skb == NULL)
L
Linus Torvalds 已提交
2725 2726
		goto out;

2727 2728 2729 2730 2731
	if (pkt_sk(sk)->has_vnet_hdr) {
		struct virtio_net_hdr vnet_hdr = { 0 };

		err = -EINVAL;
		vnet_hdr_len = sizeof(vnet_hdr);
2732
		if (len < vnet_hdr_len)
2733 2734
			goto out_free;

2735 2736
		len -= vnet_hdr_len;

2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759
		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;
2760
			vnet_hdr.csum_start = skb_checksum_start_offset(skb);
2761
			vnet_hdr.csum_offset = skb->csum_offset;
2762 2763
		} else if (skb->ip_summed == CHECKSUM_UNNECESSARY) {
			vnet_hdr.flags = VIRTIO_NET_HDR_F_DATA_VALID;
2764 2765 2766 2767 2768 2769 2770 2771
		} /* else everything is zero */

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

2772 2773 2774 2775 2776
	/*
	 *	If the address length field is there to be filled in, we fill
	 *	it in now.
	 */

2777
	sll = &PACKET_SKB_CB(skb)->sa.ll;
2778 2779 2780 2781 2782
	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
2783 2784 2785 2786 2787 2788
	/*
	 *	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
2789 2790 2791
	if (copied > len) {
		copied = len;
		msg->msg_flags |= MSG_TRUNC;
L
Linus Torvalds 已提交
2792 2793 2794 2795 2796 2797
	}

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

2798
	sock_recv_ts_and_drops(msg, sk, skb);
L
Linus Torvalds 已提交
2799 2800

	if (msg->msg_name)
2801 2802
		memcpy(msg->msg_name, &PACKET_SKB_CB(skb)->sa,
		       msg->msg_namelen);
L
Linus Torvalds 已提交
2803

2804
	if (pkt_sk(sk)->auxdata) {
2805 2806 2807 2808 2809 2810 2811 2812
		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;
2813
		aux.tp_net = skb_network_offset(skb);
2814 2815 2816 2817 2818 2819
		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;
		}
2820
		aux.tp_padding = 0;
2821
		put_cmsg(msg, SOL_PACKET, PACKET_AUXDATA, sizeof(aux), &aux);
2822 2823
	}

L
Linus Torvalds 已提交
2824 2825 2826 2827
	/*
	 *	Free or return the buffer as appropriate. Again this
	 *	hides all the races and re-entrancy issues from us.
	 */
2828
	err = vnet_hdr_len + ((flags&MSG_TRUNC) ? skb->len : copied);
L
Linus Torvalds 已提交
2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845

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
2846 2847 2848
	rcu_read_lock();
	dev = dev_get_by_index_rcu(sock_net(sk), pkt_sk(sk)->ifindex);
	if (dev)
2849
		strncpy(uaddr->sa_data, dev->name, 14);
E
Eric Dumazet 已提交
2850
	else
L
Linus Torvalds 已提交
2851
		memset(uaddr->sa_data, 0, 14);
E
Eric Dumazet 已提交
2852
	rcu_read_unlock();
L
Linus Torvalds 已提交
2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863
	*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);
2864
	DECLARE_SOCKADDR(struct sockaddr_ll *, sll, uaddr);
L
Linus Torvalds 已提交
2865 2866 2867 2868 2869 2870 2871

	if (peer)
		return -EOPNOTSUPP;

	sll->sll_family = AF_PACKET;
	sll->sll_ifindex = po->ifindex;
	sll->sll_protocol = po->num;
2872
	sll->sll_pkttype = 0;
E
Eric Dumazet 已提交
2873 2874
	rcu_read_lock();
	dev = dev_get_by_index_rcu(sock_net(sk), po->ifindex);
L
Linus Torvalds 已提交
2875 2876 2877 2878 2879 2880 2881 2882
	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
2883
	rcu_read_unlock();
2884
	*uaddr_len = offsetof(struct sockaddr_ll, sll_addr) + sll->sll_halen;
L
Linus Torvalds 已提交
2885 2886 2887 2888

	return 0;
}

2889 2890
static int packet_dev_mc(struct net_device *dev, struct packet_mclist *i,
			 int what)
L
Linus Torvalds 已提交
2891 2892 2893
{
	switch (i->type) {
	case PACKET_MR_MULTICAST:
2894 2895
		if (i->alen != dev->addr_len)
			return -EINVAL;
L
Linus Torvalds 已提交
2896
		if (what > 0)
2897
			return dev_mc_add(dev, i->addr);
L
Linus Torvalds 已提交
2898
		else
2899
			return dev_mc_del(dev, i->addr);
L
Linus Torvalds 已提交
2900 2901
		break;
	case PACKET_MR_PROMISC:
2902
		return dev_set_promiscuity(dev, what);
L
Linus Torvalds 已提交
2903 2904
		break;
	case PACKET_MR_ALLMULTI:
2905
		return dev_set_allmulti(dev, what);
L
Linus Torvalds 已提交
2906
		break;
2907
	case PACKET_MR_UNICAST:
2908 2909
		if (i->alen != dev->addr_len)
			return -EINVAL;
2910
		if (what > 0)
2911
			return dev_uc_add(dev, i->addr);
2912
		else
2913
			return dev_uc_del(dev, i->addr);
2914
		break;
E
Eric Dumazet 已提交
2915 2916
	default:
		break;
L
Linus Torvalds 已提交
2917
	}
2918
	return 0;
L
Linus Torvalds 已提交
2919 2920 2921 2922
}

static void packet_dev_mclist(struct net_device *dev, struct packet_mclist *i, int what)
{
E
Eric Dumazet 已提交
2923
	for ( ; i; i = i->next) {
L
Linus Torvalds 已提交
2924 2925 2926 2927 2928
		if (i->ifindex == dev->ifindex)
			packet_dev_mc(dev, i, what);
	}
}

2929
static int packet_mc_add(struct sock *sk, struct packet_mreq_max *mreq)
L
Linus Torvalds 已提交
2930 2931 2932 2933 2934 2935 2936 2937 2938
{
	struct packet_sock *po = pkt_sk(sk);
	struct packet_mclist *ml, *i;
	struct net_device *dev;
	int err;

	rtnl_lock();

	err = -ENODEV;
2939
	dev = __dev_get_by_index(sock_net(sk), mreq->mr_ifindex);
L
Linus Torvalds 已提交
2940 2941 2942 2943
	if (!dev)
		goto done;

	err = -EINVAL;
2944
	if (mreq->mr_alen > dev->addr_len)
L
Linus Torvalds 已提交
2945 2946 2947
		goto done;

	err = -ENOBUFS;
2948
	i = kmalloc(sizeof(*i), GFP_KERNEL);
L
Linus Torvalds 已提交
2949 2950 2951 2952 2953 2954 2955 2956 2957 2958 2959 2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971
	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;
2972 2973 2974 2975 2976
	err = packet_dev_mc(dev, i, 1);
	if (err) {
		po->mclist = i->next;
		kfree(i);
	}
L
Linus Torvalds 已提交
2977 2978 2979 2980 2981 2982

done:
	rtnl_unlock();
	return err;
}

2983
static int packet_mc_drop(struct sock *sk, struct packet_mreq_max *mreq)
L
Linus Torvalds 已提交
2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996
{
	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;
2997 2998
				dev = __dev_get_by_index(sock_net(sk), ml->ifindex);
				if (dev)
L
Linus Torvalds 已提交
2999 3000 3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022
					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;
3023 3024
		dev = __dev_get_by_index(sock_net(sk), ml->ifindex);
		if (dev != NULL)
L
Linus Torvalds 已提交
3025 3026 3027 3028 3029 3030 3031
			packet_dev_mc(dev, ml, -1);
		kfree(ml);
	}
	rtnl_unlock();
}

static int
3032
packet_setsockopt(struct socket *sock, int level, int optname, char __user *optval, unsigned int optlen)
L
Linus Torvalds 已提交
3033 3034
{
	struct sock *sk = sock->sk;
3035
	struct packet_sock *po = pkt_sk(sk);
L
Linus Torvalds 已提交
3036 3037 3038 3039 3040
	int ret;

	if (level != SOL_PACKET)
		return -ENOPROTOOPT;

J
Johann Baudy 已提交
3041
	switch (optname) {
3042
	case PACKET_ADD_MEMBERSHIP:
L
Linus Torvalds 已提交
3043 3044
	case PACKET_DROP_MEMBERSHIP:
	{
3045 3046 3047 3048
		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
3049
			return -EINVAL;
3050 3051
		if (len > sizeof(mreq))
			len = sizeof(mreq);
E
Eric Dumazet 已提交
3052
		if (copy_from_user(&mreq, optval, len))
L
Linus Torvalds 已提交
3053
			return -EFAULT;
3054 3055
		if (len < (mreq.mr_alen + offsetof(struct packet_mreq, mr_address)))
			return -EINVAL;
L
Linus Torvalds 已提交
3056 3057 3058 3059 3060 3061
		if (optname == PACKET_ADD_MEMBERSHIP)
			ret = packet_mc_add(sk, &mreq);
		else
			ret = packet_mc_drop(sk, &mreq);
		return ret;
	}
3062

L
Linus Torvalds 已提交
3063
	case PACKET_RX_RING:
J
Johann Baudy 已提交
3064
	case PACKET_TX_RING:
L
Linus Torvalds 已提交
3065
	{
3066 3067
		union tpacket_req_u req_u;
		int len;
L
Linus Torvalds 已提交
3068

3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079
		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
3080
			return -EINVAL;
3081 3082
		if (pkt_sk(sk)->has_vnet_hdr)
			return -EINVAL;
3083
		if (copy_from_user(&req_u.req, optval, len))
L
Linus Torvalds 已提交
3084
			return -EFAULT;
3085 3086
		return packet_set_ring(sk, &req_u, 0,
			optname == PACKET_TX_RING);
L
Linus Torvalds 已提交
3087 3088 3089 3090 3091
	}
	case PACKET_COPY_THRESH:
	{
		int val;

E
Eric Dumazet 已提交
3092
		if (optlen != sizeof(val))
L
Linus Torvalds 已提交
3093
			return -EINVAL;
E
Eric Dumazet 已提交
3094
		if (copy_from_user(&val, optval, sizeof(val)))
L
Linus Torvalds 已提交
3095 3096 3097 3098 3099
			return -EFAULT;

		pkt_sk(sk)->copy_thresh = val;
		return 0;
	}
3100 3101 3102 3103 3104 3105
	case PACKET_VERSION:
	{
		int val;

		if (optlen != sizeof(val))
			return -EINVAL;
J
Johann Baudy 已提交
3106
		if (po->rx_ring.pg_vec || po->tx_ring.pg_vec)
3107 3108 3109 3110 3111 3112
			return -EBUSY;
		if (copy_from_user(&val, optval, sizeof(val)))
			return -EFAULT;
		switch (val) {
		case TPACKET_V1:
		case TPACKET_V2:
3113
		case TPACKET_V3:
3114 3115 3116 3117 3118 3119
			po->tp_version = val;
			return 0;
		default:
			return -EINVAL;
		}
	}
3120 3121 3122 3123 3124 3125
	case PACKET_RESERVE:
	{
		unsigned int val;

		if (optlen != sizeof(val))
			return -EINVAL;
J
Johann Baudy 已提交
3126
		if (po->rx_ring.pg_vec || po->tx_ring.pg_vec)
3127 3128 3129 3130 3131 3132
			return -EBUSY;
		if (copy_from_user(&val, optval, sizeof(val)))
			return -EFAULT;
		po->tp_reserve = val;
		return 0;
	}
J
Johann Baudy 已提交
3133 3134 3135 3136 3137 3138 3139 3140 3141 3142 3143 3144 3145
	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;
	}
3146 3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157
	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;
	}
3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169
	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;
	}
3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185
	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;
	}
3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 3196 3197
	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
3198 3199 3200 3201 3202 3203 3204 3205 3206 3207 3208
	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);
	}
3209 3210 3211 3212 3213 3214 3215 3216 3217 3218 3219 3220 3221
	case PACKET_TX_HAS_OFF:
	{
		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_tx_has_off = !!val;
		return 0;
	}
L
Linus Torvalds 已提交
3222 3223 3224 3225 3226 3227 3228 3229 3230
	default:
		return -ENOPROTOOPT;
	}
}

static int packet_getsockopt(struct socket *sock, int level, int optname,
			     char __user *optval, int __user *optlen)
{
	int len;
3231
	int val, lv = sizeof(val);
L
Linus Torvalds 已提交
3232 3233
	struct sock *sk = sock->sk;
	struct packet_sock *po = pkt_sk(sk);
3234
	void *data = &val;
3235
	struct tpacket_stats st;
3236
	union tpacket_stats_u st_u;
L
Linus Torvalds 已提交
3237 3238 3239 3240

	if (level != SOL_PACKET)
		return -ENOPROTOOPT;

3241 3242
	if (get_user(len, optlen))
		return -EFAULT;
L
Linus Torvalds 已提交
3243 3244 3245

	if (len < 0)
		return -EINVAL;
3246

J
Johann Baudy 已提交
3247
	switch (optname) {
L
Linus Torvalds 已提交
3248 3249
	case PACKET_STATISTICS:
		spin_lock_bh(&sk->sk_receive_queue.lock);
3250
		if (po->tp_version == TPACKET_V3) {
3251
			lv = sizeof(struct tpacket_stats_v3);
3252
			memcpy(&st_u.stats3, &po->stats,
3253
			       sizeof(struct tpacket_stats));
3254
			st_u.stats3.tp_freeze_q_cnt =
3255
					po->stats_u.stats3.tp_freeze_q_cnt;
3256 3257 3258
			st_u.stats3.tp_packets += po->stats.tp_drops;
			data = &st_u.stats3;
		} else {
3259
			lv = sizeof(struct tpacket_stats);
3260 3261 3262 3263
			st = po->stats;
			st.tp_packets += st.tp_drops;
			data = &st;
		}
L
Linus Torvalds 已提交
3264 3265
		memset(&po->stats, 0, sizeof(st));
		spin_unlock_bh(&sk->sk_receive_queue.lock);
3266 3267 3268
		break;
	case PACKET_AUXDATA:
		val = po->auxdata;
3269 3270 3271
		break;
	case PACKET_ORIGDEV:
		val = po->origdev;
3272 3273 3274
		break;
	case PACKET_VNET_HDR:
		val = po->has_vnet_hdr;
L
Linus Torvalds 已提交
3275
		break;
3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288 3289 3290
	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;
3291 3292 3293
		case TPACKET_V3:
			val = sizeof(struct tpacket3_hdr);
			break;
3294 3295 3296 3297
		default:
			return -EINVAL;
		}
		break;
3298 3299 3300
	case PACKET_RESERVE:
		val = po->tp_reserve;
		break;
J
Johann Baudy 已提交
3301 3302 3303
	case PACKET_LOSS:
		val = po->tp_loss;
		break;
3304 3305 3306
	case PACKET_TIMESTAMP:
		val = po->tp_tstamp;
		break;
D
David S. Miller 已提交
3307 3308 3309
	case PACKET_FANOUT:
		val = (po->fanout ?
		       ((u32)po->fanout->id |
3310 3311
			((u32)po->fanout->type << 16) |
			((u32)po->fanout->flags << 24)) :
D
David S. Miller 已提交
3312 3313
		       0);
		break;
3314 3315 3316
	case PACKET_TX_HAS_OFF:
		val = po->tp_tx_has_off;
		break;
L
Linus Torvalds 已提交
3317 3318 3319 3320
	default:
		return -ENOPROTOOPT;
	}

3321 3322
	if (len > lv)
		len = lv;
3323 3324
	if (put_user(len, optlen))
		return -EFAULT;
3325 3326
	if (copy_to_user(optval, data, len))
		return -EFAULT;
3327
	return 0;
L
Linus Torvalds 已提交
3328 3329 3330 3331 3332 3333
}


static int packet_notifier(struct notifier_block *this, unsigned long msg, void *data)
{
	struct sock *sk;
3334
	struct net_device *dev = data;
3335
	struct net *net = dev_net(dev);
L
Linus Torvalds 已提交
3336

3337
	rcu_read_lock();
3338
	sk_for_each_rcu(sk, &net->packet.sklist) {
L
Linus Torvalds 已提交
3339 3340 3341 3342 3343 3344
		struct packet_sock *po = pkt_sk(sk);

		switch (msg) {
		case NETDEV_UNREGISTER:
			if (po->mclist)
				packet_dev_mclist(dev, po->mclist, -1);
3345 3346
			/* fallthrough */

L
Linus Torvalds 已提交
3347 3348 3349 3350
		case NETDEV_DOWN:
			if (dev->ifindex == po->ifindex) {
				spin_lock(&po->bind_lock);
				if (po->running) {
3351
					__unregister_prot_hook(sk, false);
L
Linus Torvalds 已提交
3352 3353 3354 3355 3356 3357
					sk->sk_err = ENETDOWN;
					if (!sock_flag(sk, SOCK_DEAD))
						sk->sk_error_report(sk);
				}
				if (msg == NETDEV_UNREGISTER) {
					po->ifindex = -1;
3358 3359
					if (po->prot_hook.dev)
						dev_put(po->prot_hook.dev);
L
Linus Torvalds 已提交
3360 3361 3362 3363 3364 3365
					po->prot_hook.dev = NULL;
				}
				spin_unlock(&po->bind_lock);
			}
			break;
		case NETDEV_UP:
3366 3367
			if (dev->ifindex == po->ifindex) {
				spin_lock(&po->bind_lock);
3368 3369
				if (po->num)
					register_prot_hook(sk);
3370
				spin_unlock(&po->bind_lock);
L
Linus Torvalds 已提交
3371 3372 3373 3374
			}
			break;
		}
	}
3375
	rcu_read_unlock();
L
Linus Torvalds 已提交
3376 3377 3378 3379 3380 3381 3382 3383 3384
	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
3385
	switch (cmd) {
E
Eric Dumazet 已提交
3386 3387 3388
	case SIOCOUTQ:
	{
		int amount = sk_wmem_alloc_get(sk);
3389

E
Eric Dumazet 已提交
3390 3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404 3405 3406 3407
		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);
3408

L
Linus Torvalds 已提交
3409
#ifdef CONFIG_INET
E
Eric Dumazet 已提交
3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423 3424
	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
3425 3426
#endif

E
Eric Dumazet 已提交
3427 3428
	default:
		return -ENOIOCTLCMD;
L
Linus Torvalds 已提交
3429 3430 3431 3432
	}
	return 0;
}

E
Eric Dumazet 已提交
3433
static unsigned int packet_poll(struct file *file, struct socket *sock,
L
Linus Torvalds 已提交
3434 3435 3436 3437 3438 3439 3440
				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
3441
	if (po->rx_ring.pg_vec) {
3442 3443
		if (!packet_previous_rx_frame(po, &po->rx_ring,
			TP_STATUS_KERNEL))
L
Linus Torvalds 已提交
3444 3445 3446
			mask |= POLLIN | POLLRDNORM;
	}
	spin_unlock_bh(&sk->sk_receive_queue.lock);
J
Johann Baudy 已提交
3447 3448 3449 3450 3451 3452
	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
3453 3454 3455 3456 3457 3458 3459 3460 3461 3462 3463
	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
3464
	struct socket *sock = file->private_data;
L
Linus Torvalds 已提交
3465
	struct sock *sk = sock->sk;
3466

L
Linus Torvalds 已提交
3467 3468 3469 3470 3471 3472 3473
	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
3474
	struct socket *sock = file->private_data;
L
Linus Torvalds 已提交
3475
	struct sock *sk = sock->sk;
3476

L
Linus Torvalds 已提交
3477 3478 3479 3480
	if (sk)
		atomic_dec(&pkt_sk(sk)->mapped);
}

3481
static const struct vm_operations_struct packet_mmap_ops = {
E
Eric Dumazet 已提交
3482 3483
	.open	=	packet_mm_open,
	.close	=	packet_mm_close,
L
Linus Torvalds 已提交
3484 3485
};

3486 3487
static void free_pg_vec(struct pgv *pg_vec, unsigned int order,
			unsigned int len)
L
Linus Torvalds 已提交
3488 3489 3490
{
	int i;

3491
	for (i = 0; i < len; i++) {
3492
		if (likely(pg_vec[i].buffer)) {
C
Changli Gao 已提交
3493
			if (is_vmalloc_addr(pg_vec[i].buffer))
3494 3495 3496 3497 3498 3499
				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
3500 3501 3502 3503
	}
	kfree(pg_vec);
}

3504
static char *alloc_one_pg_vec_page(unsigned long order)
3505
{
3506 3507 3508 3509 3510 3511 3512 3513 3514 3515 3516 3517
	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
3518
	buffer = vzalloc((1 << order) * PAGE_SIZE);
3519

3520 3521 3522 3523 3524 3525 3526 3527 3528 3529 3530 3531 3532 3533 3534
	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;
3535 3536
}

3537
static struct pgv *alloc_pg_vec(struct tpacket_req *req, int order)
3538 3539
{
	unsigned int block_nr = req->tp_block_nr;
3540
	struct pgv *pg_vec;
3541 3542
	int i;

3543
	pg_vec = kcalloc(block_nr, sizeof(struct pgv), GFP_KERNEL);
3544 3545 3546 3547
	if (unlikely(!pg_vec))
		goto out;

	for (i = 0; i < block_nr; i++) {
C
Changli Gao 已提交
3548
		pg_vec[i].buffer = alloc_one_pg_vec_page(order);
3549
		if (unlikely(!pg_vec[i].buffer))
3550 3551 3552 3553 3554 3555 3556 3557 3558 3559 3560
			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
3561

3562
static int packet_set_ring(struct sock *sk, union tpacket_req_u *req_u,
J
Johann Baudy 已提交
3563
		int closing, int tx_ring)
L
Linus Torvalds 已提交
3564
{
3565
	struct pgv *pg_vec = NULL;
L
Linus Torvalds 已提交
3566
	struct packet_sock *po = pkt_sk(sk);
A
Al Viro 已提交
3567
	int was_running, order = 0;
J
Johann Baudy 已提交
3568 3569
	struct packet_ring_buffer *rb;
	struct sk_buff_head *rb_queue;
A
Al Viro 已提交
3570
	__be16 num;
3571 3572 3573 3574 3575 3576 3577 3578 3579
	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;
	}
3580

J
Johann Baudy 已提交
3581 3582
	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
3583

J
Johann Baudy 已提交
3584 3585 3586 3587 3588 3589 3590
	err = -EBUSY;
	if (!closing) {
		if (atomic_read(&po->mapped))
			goto out;
		if (atomic_read(&rb->pending))
			goto out;
	}
L
Linus Torvalds 已提交
3591

J
Johann Baudy 已提交
3592 3593 3594 3595 3596
	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
3597

3598 3599 3600 3601 3602 3603 3604
		switch (po->tp_version) {
		case TPACKET_V1:
			po->tp_hdrlen = TPACKET_HDRLEN;
			break;
		case TPACKET_V2:
			po->tp_hdrlen = TPACKET2_HDRLEN;
			break;
3605 3606 3607
		case TPACKET_V3:
			po->tp_hdrlen = TPACKET3_HDRLEN;
			break;
3608 3609
		}

J
Johann Baudy 已提交
3610
		err = -EINVAL;
3611
		if (unlikely((int)req->tp_block_size <= 0))
J
Johann Baudy 已提交
3612
			goto out;
3613
		if (unlikely(req->tp_block_size & (PAGE_SIZE - 1)))
J
Johann Baudy 已提交
3614
			goto out;
3615
		if (unlikely(req->tp_frame_size < po->tp_hdrlen +
J
Johann Baudy 已提交
3616 3617
					po->tp_reserve))
			goto out;
3618
		if (unlikely(req->tp_frame_size & (TPACKET_ALIGNMENT - 1)))
J
Johann Baudy 已提交
3619
			goto out;
L
Linus Torvalds 已提交
3620

J
Johann Baudy 已提交
3621 3622 3623 3624 3625 3626
		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
3627 3628

		err = -ENOMEM;
3629 3630 3631
		order = get_order(req->tp_block_size);
		pg_vec = alloc_pg_vec(req, order);
		if (unlikely(!pg_vec))
L
Linus Torvalds 已提交
3632
			goto out;
3633 3634 3635 3636 3637 3638 3639 3640 3641 3642 3643
		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
3644 3645 3646 3647
	}
	/* Done */
	else {
		err = -EINVAL;
3648
		if (unlikely(req->tp_frame_nr))
J
Johann Baudy 已提交
3649
			goto out;
L
Linus Torvalds 已提交
3650 3651 3652 3653 3654 3655 3656 3657 3658 3659
	}

	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;
3660
		__unregister_prot_hook(sk, false);
L
Linus Torvalds 已提交
3661 3662
	}
	spin_unlock(&po->bind_lock);
3663

L
Linus Torvalds 已提交
3664 3665 3666
	synchronize_net();

	err = -EBUSY;
3667
	mutex_lock(&po->pg_vec_lock);
L
Linus Torvalds 已提交
3668 3669
	if (closing || atomic_read(&po->mapped) == 0) {
		err = 0;
J
Johann Baudy 已提交
3670
		spin_lock_bh(&rb_queue->lock);
3671
		swap(rb->pg_vec, pg_vec);
J
Johann Baudy 已提交
3672 3673 3674 3675 3676
		rb->frame_max = (req->tp_frame_nr - 1);
		rb->head = 0;
		rb->frame_size = req->tp_frame_size;
		spin_unlock_bh(&rb_queue->lock);

3677 3678
		swap(rb->pg_vec_order, order);
		swap(rb->pg_vec_len, req->tp_block_nr);
J
Johann Baudy 已提交
3679 3680 3681 3682 3683

		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
3684
		if (atomic_read(&po->mapped))
E
Eric Dumazet 已提交
3685 3686
			pr_err("packet_mmap: vma is busy: %d\n",
			       atomic_read(&po->mapped));
L
Linus Torvalds 已提交
3687
	}
3688
	mutex_unlock(&po->pg_vec_lock);
L
Linus Torvalds 已提交
3689 3690

	spin_lock(&po->bind_lock);
3691
	if (was_running) {
L
Linus Torvalds 已提交
3692
		po->num = num;
3693
		register_prot_hook(sk);
L
Linus Torvalds 已提交
3694 3695
	}
	spin_unlock(&po->bind_lock);
3696 3697 3698 3699 3700
	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
3701 3702 3703 3704 3705 3706 3707 3708
	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
3709 3710
static int packet_mmap(struct file *file, struct socket *sock,
		struct vm_area_struct *vma)
L
Linus Torvalds 已提交
3711 3712 3713
{
	struct sock *sk = sock->sk;
	struct packet_sock *po = pkt_sk(sk);
J
Johann Baudy 已提交
3714 3715
	unsigned long size, expected_size;
	struct packet_ring_buffer *rb;
L
Linus Torvalds 已提交
3716 3717 3718 3719 3720 3721 3722
	unsigned long start;
	int err = -EINVAL;
	int i;

	if (vma->vm_pgoff)
		return -EINVAL;

3723
	mutex_lock(&po->pg_vec_lock);
J
Johann Baudy 已提交
3724 3725 3726 3727 3728 3729 3730 3731 3732 3733 3734

	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
3735
		goto out;
J
Johann Baudy 已提交
3736 3737 3738

	size = vma->vm_end - vma->vm_start;
	if (size != expected_size)
L
Linus Torvalds 已提交
3739 3740 3741
		goto out;

	start = vma->vm_start;
J
Johann Baudy 已提交
3742 3743 3744 3745 3746
	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++) {
3747 3748
			struct page *page;
			void *kaddr = rb->pg_vec[i].buffer;
J
Johann Baudy 已提交
3749 3750
			int pg_num;

C
Changli Gao 已提交
3751 3752
			for (pg_num = 0; pg_num < rb->pg_vec_pages; pg_num++) {
				page = pgv_to_page(kaddr);
J
Johann Baudy 已提交
3753 3754 3755 3756
				err = vm_insert_page(vma, start, page);
				if (unlikely(err))
					goto out;
				start += PAGE_SIZE;
3757
				kaddr += PAGE_SIZE;
J
Johann Baudy 已提交
3758
			}
3759
		}
L
Linus Torvalds 已提交
3760
	}
J
Johann Baudy 已提交
3761

3762
	atomic_inc(&po->mapped);
L
Linus Torvalds 已提交
3763 3764 3765 3766
	vma->vm_ops = &packet_mmap_ops;
	err = 0;

out:
3767
	mutex_unlock(&po->pg_vec_lock);
L
Linus Torvalds 已提交
3768 3769 3770
	return err;
}

3771
static const struct proto_ops packet_ops_spkt = {
L
Linus Torvalds 已提交
3772 3773 3774 3775 3776 3777 3778 3779 3780 3781 3782 3783 3784 3785 3786 3787 3788 3789 3790 3791
	.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,
};

3792
static const struct proto_ops packet_ops = {
L
Linus Torvalds 已提交
3793 3794 3795 3796 3797 3798 3799
	.family =	PF_PACKET,
	.owner =	THIS_MODULE,
	.release =	packet_release,
	.bind =		packet_bind,
	.connect =	sock_no_connect,
	.socketpair =	sock_no_socketpair,
	.accept =	sock_no_accept,
3800
	.getname =	packet_getname,
L
Linus Torvalds 已提交
3801 3802 3803 3804 3805 3806 3807 3808 3809 3810 3811 3812
	.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,
};

3813
static const struct net_proto_family packet_family_ops = {
L
Linus Torvalds 已提交
3814 3815 3816 3817 3818 3819
	.family =	PF_PACKET,
	.create =	packet_create,
	.owner	=	THIS_MODULE,
};

static struct notifier_block packet_netdev_notifier = {
E
Eric Dumazet 已提交
3820
	.notifier_call =	packet_notifier,
L
Linus Torvalds 已提交
3821 3822 3823 3824 3825
};

#ifdef CONFIG_PROC_FS

static void *packet_seq_start(struct seq_file *seq, loff_t *pos)
3826
	__acquires(RCU)
L
Linus Torvalds 已提交
3827
{
3828
	struct net *net = seq_file_net(seq);
3829 3830 3831

	rcu_read_lock();
	return seq_hlist_start_head_rcu(&net->packet.sklist, *pos);
L
Linus Torvalds 已提交
3832 3833 3834 3835
}

static void *packet_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
3836
	struct net *net = seq_file_net(seq);
3837
	return seq_hlist_next_rcu(v, &net->packet.sklist, pos);
L
Linus Torvalds 已提交
3838 3839 3840
}

static void packet_seq_stop(struct seq_file *seq, void *v)
3841
	__releases(RCU)
L
Linus Torvalds 已提交
3842
{
3843
	rcu_read_unlock();
L
Linus Torvalds 已提交
3844 3845
}

3846
static int packet_seq_show(struct seq_file *seq, void *v)
L
Linus Torvalds 已提交
3847 3848 3849 3850
{
	if (v == SEQ_START_TOKEN)
		seq_puts(seq, "sk       RefCnt Type Proto  Iface R Rmem   User   Inode\n");
	else {
3851
		struct sock *s = sk_entry(v);
L
Linus Torvalds 已提交
3852 3853 3854
		const struct packet_sock *po = pkt_sk(s);

		seq_printf(seq,
D
Dan Rosenberg 已提交
3855
			   "%pK %-6d %-4d %04x   %-5d %1d %-6u %-6u %-6lu\n",
L
Linus Torvalds 已提交
3856 3857 3858 3859 3860 3861 3862
			   s,
			   atomic_read(&s->sk_refcnt),
			   s->sk_type,
			   ntohs(po->num),
			   po->ifindex,
			   po->running,
			   atomic_read(&s->sk_rmem_alloc),
3863
			   from_kuid_munged(seq_user_ns(seq), sock_i_uid(s)),
E
Eric Dumazet 已提交
3864
			   sock_i_ino(s));
L
Linus Torvalds 已提交
3865 3866 3867 3868 3869
	}

	return 0;
}

3870
static const struct seq_operations packet_seq_ops = {
L
Linus Torvalds 已提交
3871 3872 3873 3874 3875 3876 3877 3878
	.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)
{
3879 3880
	return seq_open_net(inode, file, &packet_seq_ops,
			    sizeof(struct seq_net_private));
L
Linus Torvalds 已提交
3881 3882
}

3883
static const struct file_operations packet_seq_fops = {
L
Linus Torvalds 已提交
3884 3885 3886 3887
	.owner		= THIS_MODULE,
	.open		= packet_seq_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
3888
	.release	= seq_release_net,
L
Linus Torvalds 已提交
3889 3890 3891 3892
};

#endif

3893
static int __net_init packet_net_init(struct net *net)
3894
{
3895
	mutex_init(&net->packet.sklist_lock);
3896
	INIT_HLIST_HEAD(&net->packet.sklist);
3897

3898
	if (!proc_create("packet", 0, net->proc_net, &packet_seq_fops))
3899 3900 3901 3902 3903
		return -ENOMEM;

	return 0;
}

3904
static void __net_exit packet_net_exit(struct net *net)
3905
{
3906
	remove_proc_entry("packet", net->proc_net);
3907 3908 3909 3910 3911 3912 3913 3914
}

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


L
Linus Torvalds 已提交
3915 3916 3917
static void __exit packet_exit(void)
{
	unregister_netdevice_notifier(&packet_netdev_notifier);
3918
	unregister_pernet_subsys(&packet_net_ops);
L
Linus Torvalds 已提交
3919 3920 3921 3922 3923 3924 3925 3926 3927 3928 3929 3930
	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);
3931
	register_pernet_subsys(&packet_net_ops);
L
Linus Torvalds 已提交
3932 3933 3934 3935 3936 3937 3938 3939 3940
	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);