af_packet.c 91.0 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 79
#include <net/ip.h>
#include <net/protocol.h>
#include <linux/skbuff.h>
#include <net/sock.h>
#include <linux/errno.h>
#include <linux/timer.h>
#include <asm/system.h>
#include <asm/uaccess.h>
#include <asm/ioctls.h>
#include <asm/page.h>
A
Al Viro 已提交
80
#include <asm/cacheflush.h>
L
Linus Torvalds 已提交
81 82 83 84 85 86
#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>
87
#include <linux/mutex.h>
E
Eric Dumazet 已提交
88
#include <linux/if_vlan.h>
89
#include <linux/virtio_net.h>
90
#include <linux/errqueue.h>
91
#include <linux/net_tstamp.h>
L
Linus Torvalds 已提交
92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111

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

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

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

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

Outgoing, dev->hard_header!=NULL
116 117
   mac_header -> ll header
   data       -> ll header
L
Linus Torvalds 已提交
118 119

Incoming, dev->hard_header==NULL
120 121
   mac_header -> UNKNOWN position. It is very likely, that it points to ll
		 header.  PPP makes it, that is wrong, because introduce
122
		 assymetry between rx and tx paths.
123
   data       -> data
L
Linus Torvalds 已提交
124 125

Outgoing, dev->hard_header==NULL
126 127
   mac_header -> data. ll header is still not built!
   data       -> data
L
Linus Torvalds 已提交
128 129 130 131 132 133 134 135 136

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


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

dev->hard_header != NULL
137 138
   mac_header -> ll header
   data       -> ll header
L
Linus Torvalds 已提交
139 140

dev->hard_header == NULL (ll header is added by device, we cannot control it)
141 142
   mac_header -> data
   data       -> data
L
Linus Torvalds 已提交
143 144 145 146 147 148 149

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

/* Private packet socket structures. */

E
Eric Dumazet 已提交
150
struct packet_mclist {
L
Linus Torvalds 已提交
151 152 153 154 155
	struct packet_mclist	*next;
	int			ifindex;
	int			count;
	unsigned short		type;
	unsigned short		alen;
156 157 158 159 160
	unsigned char		addr[MAX_ADDR_LEN];
};
/* identical to struct packet_mreq except it has
 * a longer address field.
 */
E
Eric Dumazet 已提交
161
struct packet_mreq_max {
162 163 164 165
	int		mr_ifindex;
	unsigned short	mr_type;
	unsigned short	mr_alen;
	unsigned char	mr_address[MAX_ADDR_LEN];
L
Linus Torvalds 已提交
166
};
167

168
static int packet_set_ring(struct sock *sk, union tpacket_req_u *req_u,
J
Johann Baudy 已提交
169 170
		int closing, int tx_ring);

171 172 173

#define V3_ALIGNMENT	(8)

174
#define BLK_HDR_LEN	(ALIGN(sizeof(struct tpacket_block_desc), V3_ALIGNMENT))
175 176 177 178 179

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

/* kbdq - kernel block descriptor queue */
180
struct tpacket_kbdq_core {
181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217
	struct pgv	*pkbdq;
	unsigned int	feature_req_word;
	unsigned int	hdrlen;
	unsigned char	reset_pending_on_curr_blk;
	unsigned char   delete_blk_timer;
	unsigned short	kactive_blk_num;
	unsigned short	blk_sizeof_priv;

	/* last_kactive_blk_num:
	 * trick to see if user-space has caught up
	 * in order to avoid refreshing timer when every single pkt arrives.
	 */
	unsigned short	last_kactive_blk_num;

	char		*pkblk_start;
	char		*pkblk_end;
	int		kblk_size;
	unsigned int	knum_blocks;
	uint64_t	knxt_seq_num;
	char		*prev;
	char		*nxt_offset;
	struct sk_buff	*skb;

	atomic_t	blk_fill_in_prog;

	/* Default is set to 8ms */
#define DEFAULT_PRB_RETIRE_TOV	(8)

	unsigned short  retire_blk_tov;
	unsigned short  version;
	unsigned long	tov_in_jiffies;

	/* timer to retire an outstanding block */
	struct timer_list retire_blk_timer;
};

#define PGV_FROM_VMALLOC 1
218 219 220 221
struct pgv {
	char *buffer;
};

J
Johann Baudy 已提交
222
struct packet_ring_buffer {
223
	struct pgv		*pg_vec;
J
Johann Baudy 已提交
224 225 226 227 228 229 230 231 232
	unsigned int		head;
	unsigned int		frames_per_block;
	unsigned int		frame_size;
	unsigned int		frame_max;

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

233
	struct tpacket_kbdq_core	prb_bdqc;
J
Johann Baudy 已提交
234 235 236
	atomic_t		pending;
};

237 238 239 240 241 242 243 244
#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
245 246
struct packet_sock;
static int tpacket_snd(struct packet_sock *po, struct msghdr *msg);
L
Linus Torvalds 已提交
247

248 249 250 251
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);
252 253 254
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 *,
255
			struct packet_sock *);
256
static void prb_retire_current_block(struct tpacket_kbdq_core *,
257
		struct packet_sock *, unsigned int status);
258 259 260
static int prb_queue_frozen(struct tpacket_kbdq_core *);
static void prb_open_block(struct tpacket_kbdq_core *,
		struct tpacket_block_desc *);
261
static void prb_retire_rx_blk_timer_expired(unsigned long);
262 263 264 265 266 267 268 269 270
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
271 272
static void packet_flush_mclist(struct sock *sk);

D
David S. Miller 已提交
273
struct packet_fanout;
L
Linus Torvalds 已提交
274 275 276
struct packet_sock {
	/* struct sock has to be the first member of packet_sock */
	struct sock		sk;
D
David S. Miller 已提交
277
	struct packet_fanout	*fanout;
L
Linus Torvalds 已提交
278
	struct tpacket_stats	stats;
279
	union  tpacket_stats_u	stats_u;
J
Johann Baudy 已提交
280 281
	struct packet_ring_buffer	rx_ring;
	struct packet_ring_buffer	tx_ring;
L
Linus Torvalds 已提交
282 283
	int			copy_thresh;
	spinlock_t		bind_lock;
284
	struct mutex		pg_vec_lock;
285
	unsigned int		running:1,	/* prot_hook is attached*/
286
				auxdata:1,
287 288
				origdev:1,
				has_vnet_hdr:1;
L
Linus Torvalds 已提交
289
	int			ifindex;	/* bound device		*/
A
Al Viro 已提交
290
	__be16			num;
L
Linus Torvalds 已提交
291 292
	struct packet_mclist	*mclist;
	atomic_t		mapped;
293 294
	enum tpacket_versions	tp_version;
	unsigned int		tp_hdrlen;
295
	unsigned int		tp_reserve;
J
Johann Baudy 已提交
296
	unsigned int		tp_loss:1;
297
	unsigned int		tp_tstamp;
298
	struct packet_type	prot_hook ____cacheline_aligned_in_smp;
L
Linus Torvalds 已提交
299 300
};

D
David S. Miller 已提交
301 302 303 304 305 306 307 308 309
#define PACKET_FANOUT_MAX	256

struct packet_fanout {
#ifdef CONFIG_NET_NS
	struct net		*net;
#endif
	unsigned int		num_members;
	u16			id;
	u8			type;
310
	u8			defrag;
D
David S. Miller 已提交
311 312 313 314 315 316 317 318
	atomic_t		rr_cur;
	struct list_head	list;
	struct sock		*arr[PACKET_FANOUT_MAX];
	spinlock_t		lock;
	atomic_t		sk_ref;
	struct packet_type	prot_hook ____cacheline_aligned_in_smp;
};

319 320 321 322 323 324 325 326 327
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))
328

329
#define GET_PBDQC_FROM_RB(x)	((struct tpacket_kbdq_core *)(&(x)->prb_bdqc))
330
#define GET_PBLOCK_DESC(x, bid)	\
331
	((struct tpacket_block_desc *)((x)->pkbdq[(bid)].buffer))
332
#define GET_CURR_PBLOCK_DESC_FROM_CORE(x)	\
333
	((struct tpacket_block_desc *)((x)->pkbdq[(x)->kactive_blk_num].buffer))
334 335 336 337
#define GET_NEXT_PRB_BLK_NUM(x) \
	(((x)->kactive_blk_num < ((x)->knum_blocks-1)) ? \
	((x)->kactive_blk_num+1) : 0)

338
static struct packet_sock *pkt_sk(struct sock *sk)
339 340 341 342
{
	return (struct packet_sock *)sk;
}

D
David S. Miller 已提交
343 344 345
static void __fanout_unlink(struct sock *sk, struct packet_sock *po);
static void __fanout_link(struct sock *sk, struct packet_sock *po);

346 347 348 349 350 351 352 353
/* 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
354 355 356 357
		if (po->fanout)
			__fanout_link(sk, po);
		else
			dev_add_pack(&po->prot_hook);
358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374
		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
375 376 377 378
	if (po->fanout)
		__fanout_unlink(sk, po);
	else
		__dev_remove_pack(&po->prot_hook);
379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395
	__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);
}

396
static inline __pure struct page *pgv_to_page(void *addr)
397 398 399 400 401 402
{
	if (is_vmalloc_addr(addr))
		return vmalloc_to_page(addr);
	return virt_to_page(addr);
}

J
Johann Baudy 已提交
403
static void __packet_set_status(struct packet_sock *po, void *frame, int status)
L
Linus Torvalds 已提交
404
{
405 406 407 408 409
	union {
		struct tpacket_hdr *h1;
		struct tpacket2_hdr *h2;
		void *raw;
	} h;
L
Linus Torvalds 已提交
410

J
Johann Baudy 已提交
411
	h.raw = frame;
412 413
	switch (po->tp_version) {
	case TPACKET_V1:
J
Johann Baudy 已提交
414
		h.h1->tp_status = status;
415
		flush_dcache_page(pgv_to_page(&h.h1->tp_status));
416 417
		break;
	case TPACKET_V2:
J
Johann Baudy 已提交
418
		h.h2->tp_status = status;
419
		flush_dcache_page(pgv_to_page(&h.h2->tp_status));
420
		break;
421
	case TPACKET_V3:
J
Johann Baudy 已提交
422
	default:
423
		WARN(1, "TPACKET version not supported.\n");
J
Johann Baudy 已提交
424
		BUG();
425
	}
J
Johann Baudy 已提交
426 427

	smp_wmb();
428 429
}

J
Johann Baudy 已提交
430
static int __packet_get_status(struct packet_sock *po, void *frame)
431 432 433 434 435 436 437
{
	union {
		struct tpacket_hdr *h1;
		struct tpacket2_hdr *h2;
		void *raw;
	} h;

J
Johann Baudy 已提交
438 439
	smp_rmb();

440 441 442
	h.raw = frame;
	switch (po->tp_version) {
	case TPACKET_V1:
443
		flush_dcache_page(pgv_to_page(&h.h1->tp_status));
J
Johann Baudy 已提交
444
		return h.h1->tp_status;
445
	case TPACKET_V2:
446
		flush_dcache_page(pgv_to_page(&h.h2->tp_status));
J
Johann Baudy 已提交
447
		return h.h2->tp_status;
448
	case TPACKET_V3:
J
Johann Baudy 已提交
449
	default:
450
		WARN(1, "TPACKET version not supported.\n");
J
Johann Baudy 已提交
451 452
		BUG();
		return 0;
453
	}
L
Linus Torvalds 已提交
454
}
J
Johann Baudy 已提交
455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470

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;

471 472
	h.raw = rb->pg_vec[pg_vec_pos].buffer +
		(frame_offset * rb->frame_size);
J
Johann Baudy 已提交
473 474 475 476 477 478 479

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

	return h.raw;
}

480
static void *packet_current_frame(struct packet_sock *po,
J
Johann Baudy 已提交
481 482 483 484 485 486
		struct packet_ring_buffer *rb,
		int status)
{
	return packet_lookup_frame(po, rb, rb->head, status);
}

487
static void prb_del_retire_blk_timer(struct tpacket_kbdq_core *pkc)
488 489 490 491 492 493 494 495
{
	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)
{
496
	struct tpacket_kbdq_core *pkc;
497 498 499 500 501 502 503 504 505 506 507

	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,
508
		struct tpacket_kbdq_core *pkc,
509 510 511 512 513 514 515 516 517 518
		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)
{
519
	struct tpacket_kbdq_core *pkc;
520 521 522 523 524 525 526 527 528 529 530 531 532

	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;
533 534
	struct ethtool_cmd ecmd;
	int err;
535

536 537 538 539
	rtnl_lock();
	dev = __dev_get_by_index(sock_net(&po->sk), po->ifindex);
	if (unlikely(!dev)) {
		rtnl_unlock();
540
		return DEFAULT_PRB_RETIRE_TOV;
541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561
	}
	err = __ethtool_get_settings(dev, &ecmd);
	rtnl_unlock();
	if (!err) {
		switch (ecmd.speed) {
		case SPEED_10000:
			msec = 1;
			div = 10000/1000;
			break;
		case SPEED_1000:
			msec = 1;
			div = 1000/1000;
			break;
		/*
		 * If the link speed is so slow you don't really
		 * need to worry about perf anyways
		 */
		case SPEED_100:
		case SPEED_10:
		default:
			return DEFAULT_PRB_RETIRE_TOV;
562 563 564 565 566 567 568 569 570 571 572 573 574 575 576
		}
	}

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

	if (div)
		mbits /= div;

	tmo = mbits * msec;

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

577
static void prb_init_ft_ops(struct tpacket_kbdq_core *p1,
578 579 580 581 582 583 584 585 586 587
			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)
{
588 589
	struct tpacket_kbdq_core *p1 = &rb->prb_bdqc;
	struct tpacket_block_desc *pbd;
590 591 592 593 594

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

	p1->knxt_seq_num = 1;
	p1->pkbdq = pg_vec;
595
	pbd = (struct tpacket_block_desc *)pg_vec[0].buffer;
596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618
	p1->pkblk_start	= (char *)pg_vec[0].buffer;
	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.
 */
619
static void _prb_refresh_rx_retire_blk_timer(struct tpacket_kbdq_core *pkc)
620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651
{
	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;
652
	struct tpacket_kbdq_core *pkc = &po->rx_ring.prb_bdqc;
653
	unsigned int frozen;
654
	struct tpacket_block_desc *pbd;
655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717

	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);
}

718
static void prb_flush_block(struct tpacket_kbdq_core *pkc1,
719
		struct tpacket_block_desc *pbd1, __u32 status)
720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760
{
	/* 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.
 */
761 762
static void prb_close_block(struct tpacket_kbdq_core *pkc1,
		struct tpacket_block_desc *pbd1,
763 764 765 766 767
		struct packet_sock *po, unsigned int stat)
{
	__u32 status = TP_STATUS_USER | stat;

	struct tpacket3_hdr *last_pkt;
768
	struct tpacket_hdr_v1 *h1 = &pbd1->hdr.bh1;
769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795

	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);
}

796
static void prb_thaw_queue(struct tpacket_kbdq_core *pkc)
797 798 799 800 801 802 803 804 805 806 807
{
	pkc->reset_pending_on_curr_blk = 0;
}

/*
 * Side effect of opening a block:
 *
 * 1) prb_queue is thawed.
 * 2) retire_blk_timer is refreshed.
 *
 */
808 809
static void prb_open_block(struct tpacket_kbdq_core *pkc1,
	struct tpacket_block_desc *pbd1)
810 811
{
	struct timespec ts;
812
	struct tpacket_hdr_v1 *h1 = &pbd1->hdr.bh1;
813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871

	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;
		pkc1->nxt_offset = (char *)(pkc1->pkblk_start +
		BLK_PLUS_PRIV(pkc1->blk_sizeof_priv));
		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.
 */
872
static void prb_freeze_queue(struct tpacket_kbdq_core *pkc,
873 874 875 876 877 878 879 880 881 882 883 884 885 886
				  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.
 */
887
static void *prb_dispatch_next_block(struct tpacket_kbdq_core *pkc,
888 889
		struct packet_sock *po)
{
890
	struct tpacket_block_desc *pbd;
891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911

	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;
}

912
static void prb_retire_current_block(struct tpacket_kbdq_core *pkc,
913 914
		struct packet_sock *po, unsigned int status)
{
915
	struct tpacket_block_desc *pbd = GET_CURR_PBLOCK_DESC_FROM_CORE(pkc);
916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942

	/* 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();
}

943
static int prb_curr_blk_in_use(struct tpacket_kbdq_core *pkc,
944
				      struct tpacket_block_desc *pbd)
945 946 947 948
{
	return TP_STATUS_USER & BLOCK_STATUS(pbd);
}

949
static int prb_queue_frozen(struct tpacket_kbdq_core *pkc)
950 951 952 953
{
	return pkc->reset_pending_on_curr_blk;
}

954
static void prb_clear_blk_fill_status(struct packet_ring_buffer *rb)
955
{
956
	struct tpacket_kbdq_core *pkc  = GET_PBDQC_FROM_RB(rb);
957 958 959
	atomic_dec(&pkc->blk_fill_in_prog);
}

960
static void prb_fill_rxhash(struct tpacket_kbdq_core *pkc,
961 962 963 964 965
			struct tpacket3_hdr *ppd)
{
	ppd->hv1.tp_rxhash = skb_get_rxhash(pkc->skb);
}

966
static void prb_clear_rxhash(struct tpacket_kbdq_core *pkc,
967 968 969 970 971
			struct tpacket3_hdr *ppd)
{
	ppd->hv1.tp_rxhash = 0;
}

972
static void prb_fill_vlan_info(struct tpacket_kbdq_core *pkc,
973 974 975 976 977 978 979 980 981 982
			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 {
		ppd->hv1.tp_vlan_tci = ppd->tp_status = 0;
	}
}

983
static void prb_run_all_ft_ops(struct tpacket_kbdq_core *pkc,
984 985 986 987 988 989 990 991 992 993
			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);
}

994
static void prb_fill_curr_block(char *curr,
995 996
				struct tpacket_kbdq_core *pkc,
				struct tpacket_block_desc *pbd,
997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017
				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
					    )
{
1018 1019
	struct tpacket_kbdq_core *pkc;
	struct tpacket_block_desc *pbd;
1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073
	char *curr, *end;

	pkc = GET_PBDQC_FROM_RB(((struct packet_ring_buffer *)&po->rx_ring));
	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;
	end = (char *) ((char *)pbd + pkc->kblk_size);

	/* 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;
}

1074
static void *packet_current_rx_frame(struct packet_sock *po,
1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093
					    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();
		return 0;
	}
}

1094
static void *prb_lookup_block(struct packet_sock *po,
1095 1096 1097 1098
				     struct packet_ring_buffer *rb,
				     unsigned int previous,
				     int status)
{
1099 1100
	struct tpacket_kbdq_core *pkc  = GET_PBDQC_FROM_RB(rb);
	struct tpacket_block_desc *pbd = GET_PBLOCK_DESC(pkc, previous);
1101 1102 1103 1104 1105 1106

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

1107
static int prb_previous_blk_num(struct packet_ring_buffer *rb)
1108 1109 1110 1111 1112 1113 1114 1115 1116 1117
{
	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 */
1118
static void *__prb_previous_block(struct packet_sock *po,
1119 1120 1121 1122 1123 1124 1125
					 struct packet_ring_buffer *rb,
					 int status)
{
	unsigned int previous = prb_previous_blk_num(rb);
	return prb_lookup_block(po, rb, previous, status);
}

1126
static void *packet_previous_rx_frame(struct packet_sock *po,
1127 1128 1129 1130 1131 1132 1133 1134 1135
					     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);
}

1136
static void packet_increment_rx_head(struct packet_sock *po,
1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150
					    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;
	}
}

1151
static void *packet_previous_frame(struct packet_sock *po,
J
Johann Baudy 已提交
1152 1153 1154 1155 1156 1157 1158
		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);
}

1159
static void packet_increment_head(struct packet_ring_buffer *buff)
J
Johann Baudy 已提交
1160 1161 1162 1163
{
	buff->head = buff->head != buff->frame_max ? buff->head+1 : 0;
}

L
Linus Torvalds 已提交
1164 1165
static void packet_sock_destruct(struct sock *sk)
{
1166 1167
	skb_queue_purge(&sk->sk_error_queue);

1168 1169
	WARN_ON(atomic_read(&sk->sk_rmem_alloc));
	WARN_ON(atomic_read(&sk->sk_wmem_alloc));
L
Linus Torvalds 已提交
1170 1171

	if (!sock_flag(sk, SOCK_DEAD)) {
E
Eric Dumazet 已提交
1172
		pr_err("Attempt to release alive packet socket: %p\n", sk);
L
Linus Torvalds 已提交
1173 1174 1175
		return;
	}

1176
	sk_refcnt_debug_dec(sk);
L
Linus Torvalds 已提交
1177 1178
}

D
David S. Miller 已提交
1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208
static int fanout_rr_next(struct packet_fanout *f, unsigned int num)
{
	int x = atomic_read(&f->rr_cur) + 1;

	if (x >= num)
		x = 0;

	return x;
}

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

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

	return f->arr[idx];
}

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

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

1209 1210 1211 1212 1213 1214 1215 1216 1217
static struct sock *fanout_demux_cpu(struct packet_fanout *f, struct sk_buff *skb, unsigned int num)
{
	unsigned int cpu = smp_processor_id();

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

static int packet_rcv_fanout(struct sk_buff *skb, struct net_device *dev,
			     struct packet_type *pt, struct net_device *orig_dev)
D
David S. Miller 已提交
1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229
{
	struct packet_fanout *f = pt->af_packet_priv;
	unsigned int num = f->num_members;
	struct packet_sock *po;
	struct sock *sk;

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

1230 1231 1232 1233
	switch (f->type) {
	case PACKET_FANOUT_HASH:
	default:
		if (f->defrag) {
1234
			skb = ip_check_defrag(skb, IP_DEFRAG_AF_PACKET);
1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246
			if (!skb)
				return 0;
		}
		skb_get_rxhash(skb);
		sk = fanout_demux_hash(f, skb, num);
		break;
	case PACKET_FANOUT_LB:
		sk = fanout_demux_lb(f, skb, num);
		break;
	case PACKET_FANOUT_CPU:
		sk = fanout_demux_cpu(f, skb, num);
		break;
D
David S. Miller 已提交
1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283
	}

	po = pkt_sk(sk);

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

static DEFINE_MUTEX(fanout_mutex);
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);
}

1284
static int fanout_add(struct sock *sk, u16 id, u16 type_flags)
D
David S. Miller 已提交
1285 1286 1287
{
	struct packet_sock *po = pkt_sk(sk);
	struct packet_fanout *f, *match;
1288 1289
	u8 type = type_flags & 0xff;
	u8 defrag = (type_flags & PACKET_FANOUT_FLAG_DEFRAG) ? 1 : 0;
D
David S. Miller 已提交
1290 1291 1292 1293 1294
	int err;

	switch (type) {
	case PACKET_FANOUT_HASH:
	case PACKET_FANOUT_LB:
1295
	case PACKET_FANOUT_CPU:
D
David S. Miller 已提交
1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315
		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
1316
	err = -EINVAL;
1317
	if (match && match->defrag != defrag)
E
Eric Dumazet 已提交
1318
		goto out;
D
David S. Miller 已提交
1319
	if (!match) {
E
Eric Dumazet 已提交
1320
		err = -ENOMEM;
D
David S. Miller 已提交
1321
		match = kzalloc(sizeof(*match), GFP_KERNEL);
E
Eric Dumazet 已提交
1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337
		if (!match)
			goto out;
		write_pnet(&match->net, sock_net(sk));
		match->id = id;
		match->type = type;
		match->defrag = defrag;
		atomic_set(&match->rr_cur, 0);
		INIT_LIST_HEAD(&match->list);
		spin_lock_init(&match->lock);
		atomic_set(&match->sk_ref, 0);
		match->prot_hook.type = po->prot_hook.type;
		match->prot_hook.dev = po->prot_hook.dev;
		match->prot_hook.func = packet_rcv_fanout;
		match->prot_hook.af_packet_priv = match;
		dev_add_pack(&match->prot_hook);
		list_add(&match->list, &fanout_list);
D
David S. Miller 已提交
1338
	}
E
Eric Dumazet 已提交
1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349
	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
1350 1351
		}
	}
E
Eric Dumazet 已提交
1352
out:
D
David S. Miller 已提交
1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375
	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;

	po->fanout = NULL;

	mutex_lock(&fanout_mutex);
	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
1376

1377
static const struct proto_ops packet_ops;
L
Linus Torvalds 已提交
1378

1379
static const struct proto_ops packet_ops_spkt;
L
Linus Torvalds 已提交
1380

E
Eric Dumazet 已提交
1381 1382
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
1383 1384 1385 1386 1387 1388 1389 1390 1391 1392
{
	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;
1393

L
Linus Torvalds 已提交
1394 1395 1396 1397 1398 1399 1400
	/*
	 *	Yank back the headers [hope the device set this
	 *	right or kerboom...]
	 *
	 *	Incoming packets have ll header pulled,
	 *	push it back.
	 *
1401
	 *	For outgoing ones skb->data == skb_mac_header(skb)
L
Linus Torvalds 已提交
1402 1403 1404 1405 1406 1407
	 *	so that this procedure is noop.
	 */

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

O
Octavian Purdila 已提交
1408
	if (!net_eq(dev_net(dev), sock_net(sk)))
1409 1410
		goto out;

E
Eric Dumazet 已提交
1411 1412
	skb = skb_share_check(skb, GFP_ATOMIC);
	if (skb == NULL)
L
Linus Torvalds 已提交
1413 1414 1415
		goto oom;

	/* drop any routing info */
E
Eric Dumazet 已提交
1416
	skb_dst_drop(skb);
L
Linus Torvalds 已提交
1417

1418 1419 1420
	/* drop conntrack reference */
	nf_reset(skb);

1421
	spkt = &PACKET_SKB_CB(skb)->sa.pkt;
L
Linus Torvalds 已提交
1422

1423
	skb_push(skb, skb->data - skb_mac_header(skb));
L
Linus Torvalds 已提交
1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437

	/*
	 *	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
1438
	if (sock_queue_rcv_skb(sk, skb) == 0)
L
Linus Torvalds 已提交
1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451
		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
 */
1452

L
Linus Torvalds 已提交
1453 1454 1455 1456
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
1457
	struct sockaddr_pkt *saddr = (struct sockaddr_pkt *)msg->msg_name;
1458
	struct sk_buff *skb = NULL;
L
Linus Torvalds 已提交
1459
	struct net_device *dev;
E
Eric Dumazet 已提交
1460
	__be16 proto = 0;
L
Linus Torvalds 已提交
1461
	int err;
1462

L
Linus Torvalds 已提交
1463
	/*
1464
	 *	Get and verify the address.
L
Linus Torvalds 已提交
1465 1466
	 */

E
Eric Dumazet 已提交
1467
	if (saddr) {
L
Linus Torvalds 已提交
1468
		if (msg->msg_namelen < sizeof(struct sockaddr))
E
Eric Dumazet 已提交
1469 1470 1471 1472 1473
			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
1474 1475

	/*
1476
	 *	Find the device first to size check it
L
Linus Torvalds 已提交
1477 1478 1479
	 */

	saddr->spkt_device[13] = 0;
1480
retry:
E
Eric Dumazet 已提交
1481 1482
	rcu_read_lock();
	dev = dev_get_by_name_rcu(sock_net(sk), saddr->spkt_device);
L
Linus Torvalds 已提交
1483 1484 1485
	err = -ENODEV;
	if (dev == NULL)
		goto out_unlock;
1486

1487 1488 1489 1490
	err = -ENETDOWN;
	if (!(dev->flags & IFF_UP))
		goto out_unlock;

L
Linus Torvalds 已提交
1491
	/*
E
Eric Dumazet 已提交
1492 1493
	 * 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
1494
	 */
1495

L
Linus Torvalds 已提交
1496
	err = -EMSGSIZE;
1497
	if (len > dev->mtu + dev->hard_header_len + VLAN_HLEN)
L
Linus Torvalds 已提交
1498 1499
		goto out_unlock;

1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525
	if (!skb) {
		size_t reserved = LL_RESERVED_SPACE(dev);
		unsigned int hhlen = dev->header_ops ? dev->hard_header_len : 0;

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

		/* Try to align data part correctly */
		if (hhlen) {
			skb->data -= hhlen;
			skb->tail -= hhlen;
			if (len < hhlen)
				skb_reset_network_header(skb);
		}
		err = memcpy_fromiovec(skb_put(skb, len), msg->msg_iov, len);
		if (err)
			goto out_free;
		goto retry;
L
Linus Torvalds 已提交
1526 1527
	}

1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540
	if (len > (dev->mtu + dev->hard_header_len)) {
		/* 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;
		}
	}
1541

L
Linus Torvalds 已提交
1542 1543 1544
	skb->protocol = proto;
	skb->dev = dev;
	skb->priority = sk->sk_priority;
1545
	skb->mark = sk->sk_mark;
1546
	err = sock_tx_timestamp(sk, &skb_shinfo(skb)->tx_flags);
1547 1548
	if (err < 0)
		goto out_unlock;
L
Linus Torvalds 已提交
1549 1550

	dev_queue_xmit(skb);
E
Eric Dumazet 已提交
1551
	rcu_read_unlock();
E
Eric Dumazet 已提交
1552
	return len;
L
Linus Torvalds 已提交
1553 1554

out_unlock:
E
Eric Dumazet 已提交
1555
	rcu_read_unlock();
1556 1557
out_free:
	kfree_skb(skb);
L
Linus Torvalds 已提交
1558 1559 1560
	return err;
}

1561
static unsigned int run_filter(const struct sk_buff *skb,
E
Eric Dumazet 已提交
1562
				      const struct sock *sk,
D
David S. Miller 已提交
1563
				      unsigned int res)
L
Linus Torvalds 已提交
1564 1565
{
	struct sk_filter *filter;
1566

1567 1568
	rcu_read_lock();
	filter = rcu_dereference(sk->sk_filter);
D
David S. Miller 已提交
1569
	if (filter != NULL)
1570
		res = SK_RUN_FILTER(filter, skb);
1571
	rcu_read_unlock();
L
Linus Torvalds 已提交
1572

D
David S. Miller 已提交
1573
	return res;
L
Linus Torvalds 已提交
1574 1575 1576
}

/*
E
Eric Dumazet 已提交
1577 1578 1579 1580 1581 1582 1583 1584 1585
 * 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
1586 1587
 */

E
Eric Dumazet 已提交
1588 1589
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
1590 1591 1592 1593
{
	struct sock *sk;
	struct sockaddr_ll *sll;
	struct packet_sock *po;
E
Eric Dumazet 已提交
1594
	u8 *skb_head = skb->data;
L
Linus Torvalds 已提交
1595
	int skb_len = skb->len;
D
David S. Miller 已提交
1596
	unsigned int snaplen, res;
L
Linus Torvalds 已提交
1597 1598 1599 1600 1601 1602 1603

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

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

O
Octavian Purdila 已提交
1604
	if (!net_eq(dev_net(dev), sock_net(sk)))
1605 1606
		goto drop;

L
Linus Torvalds 已提交
1607 1608
	skb->dev = dev;

1609
	if (dev->header_ops) {
L
Linus Torvalds 已提交
1610
		/* The device has an explicit notion of ll header,
E
Eric Dumazet 已提交
1611 1612 1613 1614 1615
		 * 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
1616 1617
		 */
		if (sk->sk_type != SOCK_DGRAM)
1618
			skb_push(skb, skb->data - skb_mac_header(skb));
L
Linus Torvalds 已提交
1619 1620
		else if (skb->pkt_type == PACKET_OUTGOING) {
			/* Special case: outgoing packets have ll header at head */
1621
			skb_pull(skb, skb_network_offset(skb));
L
Linus Torvalds 已提交
1622 1623 1624 1625 1626
		}
	}

	snaplen = skb->len;

D
David S. Miller 已提交
1627 1628
	res = run_filter(skb, sk, snaplen);
	if (!res)
1629
		goto drop_n_restore;
D
David S. Miller 已提交
1630 1631
	if (snaplen > res)
		snaplen = res;
L
Linus Torvalds 已提交
1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649

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

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

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

1650 1651 1652 1653
	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
1654 1655 1656 1657
	sll->sll_family = AF_PACKET;
	sll->sll_hatype = dev->type;
	sll->sll_protocol = skb->protocol;
	sll->sll_pkttype = skb->pkt_type;
1658
	if (unlikely(po->origdev))
1659 1660 1661
		sll->sll_ifindex = orig_dev->ifindex;
	else
		sll->sll_ifindex = dev->ifindex;
L
Linus Torvalds 已提交
1662

S
Stephen Hemminger 已提交
1663
	sll->sll_halen = dev_parse_header(skb, sll->sll_addr);
L
Linus Torvalds 已提交
1664

1665
	PACKET_SKB_CB(skb)->origlen = skb->len;
1666

L
Linus Torvalds 已提交
1667 1668 1669 1670 1671
	if (pskb_trim(skb, snaplen))
		goto drop_n_acct;

	skb_set_owner_r(skb, sk);
	skb->dev = NULL;
E
Eric Dumazet 已提交
1672
	skb_dst_drop(skb);
L
Linus Torvalds 已提交
1673

1674 1675 1676
	/* drop conntrack reference */
	nf_reset(skb);

L
Linus Torvalds 已提交
1677 1678
	spin_lock(&sk->sk_receive_queue.lock);
	po->stats.tp_packets++;
1679
	skb->dropcount = atomic_read(&sk->sk_drops);
L
Linus Torvalds 已提交
1680 1681 1682 1683 1684 1685
	__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:
1686 1687 1688 1689
	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
1690 1691 1692 1693 1694 1695 1696

drop_n_restore:
	if (skb_head != skb->data && skb_shared(skb)) {
		skb->data = skb_head;
		skb->len = skb_len;
	}
drop:
1697
	consume_skb(skb);
L
Linus Torvalds 已提交
1698 1699 1700
	return 0;
}

E
Eric Dumazet 已提交
1701 1702
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
1703 1704 1705 1706
{
	struct sock *sk;
	struct packet_sock *po;
	struct sockaddr_ll *sll;
1707 1708 1709
	union {
		struct tpacket_hdr *h1;
		struct tpacket2_hdr *h2;
1710
		struct tpacket3_hdr *h3;
1711 1712
		void *raw;
	} h;
E
Eric Dumazet 已提交
1713
	u8 *skb_head = skb->data;
L
Linus Torvalds 已提交
1714
	int skb_len = skb->len;
D
David S. Miller 已提交
1715
	unsigned int snaplen, res;
1716
	unsigned long status = TP_STATUS_USER;
1717
	unsigned short macoff, netoff, hdrlen;
L
Linus Torvalds 已提交
1718
	struct sk_buff *copy_skb = NULL;
1719
	struct timeval tv;
1720
	struct timespec ts;
1721
	struct skb_shared_hwtstamps *shhwtstamps = skb_hwtstamps(skb);
L
Linus Torvalds 已提交
1722 1723 1724 1725 1726 1727 1728

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

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

O
Octavian Purdila 已提交
1729
	if (!net_eq(dev_net(dev), sock_net(sk)))
1730 1731
		goto drop;

1732
	if (dev->header_ops) {
L
Linus Torvalds 已提交
1733
		if (sk->sk_type != SOCK_DGRAM)
1734
			skb_push(skb, skb->data - skb_mac_header(skb));
L
Linus Torvalds 已提交
1735 1736
		else if (skb->pkt_type == PACKET_OUTGOING) {
			/* Special case: outgoing packets have ll header at head */
1737
			skb_pull(skb, skb_network_offset(skb));
L
Linus Torvalds 已提交
1738 1739 1740
		}
	}

1741 1742 1743
	if (skb->ip_summed == CHECKSUM_PARTIAL)
		status |= TP_STATUS_CSUMNOTREADY;

L
Linus Torvalds 已提交
1744 1745
	snaplen = skb->len;

D
David S. Miller 已提交
1746 1747
	res = run_filter(skb, sk, snaplen);
	if (!res)
1748
		goto drop_n_restore;
D
David S. Miller 已提交
1749 1750
	if (snaplen > res)
		snaplen = res;
L
Linus Torvalds 已提交
1751 1752

	if (sk->sk_type == SOCK_DGRAM) {
1753 1754
		macoff = netoff = TPACKET_ALIGN(po->tp_hdrlen) + 16 +
				  po->tp_reserve;
L
Linus Torvalds 已提交
1755
	} else {
1756
		unsigned maclen = skb_network_offset(skb);
1757
		netoff = TPACKET_ALIGN(po->tp_hdrlen +
1758 1759
				       (maclen < 16 ? 16 : maclen)) +
			po->tp_reserve;
L
Linus Torvalds 已提交
1760 1761
		macoff = netoff - maclen;
	}
1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774
	if (po->tp_version <= TPACKET_V2) {
		if (macoff + snaplen > po->rx_ring.frame_size) {
			if (po->copy_thresh &&
				atomic_read(&sk->sk_rmem_alloc) + skb->truesize
				< (unsigned)sk->sk_rcvbuf) {
				if (skb_shared(skb)) {
					copy_skb = skb_clone(skb, GFP_ATOMIC);
				} else {
					copy_skb = skb_get(skb);
					skb_head = skb->data;
				}
				if (copy_skb)
					skb_set_owner_r(copy_skb, sk);
L
Linus Torvalds 已提交
1775
			}
1776 1777 1778
			snaplen = po->rx_ring.frame_size - macoff;
			if ((int)snaplen < 0)
				snaplen = 0;
L
Linus Torvalds 已提交
1779 1780 1781
		}
	}
	spin_lock(&sk->sk_receive_queue.lock);
1782 1783
	h.raw = packet_current_rx_frame(po, skb,
					TP_STATUS_KERNEL, (macoff+snaplen));
1784
	if (!h.raw)
L
Linus Torvalds 已提交
1785
		goto ring_is_full;
1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796
	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
1797 1798 1799 1800 1801 1802 1803
	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);

1804
	skb_copy_bits(skb, 0, h.raw + macoff, snaplen);
L
Linus Torvalds 已提交
1805

1806 1807 1808 1809 1810 1811
	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;
1812 1813 1814 1815 1816 1817 1818
		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)
1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830
			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;
1831 1832 1833 1834 1835 1836 1837
		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)
1838 1839 1840 1841 1842
			ts = ktime_to_timespec(skb->tstamp);
		else
			getnstimeofday(&ts);
		h.h2->tp_sec = ts.tv_sec;
		h.h2->tp_nsec = ts.tv_nsec;
1843 1844 1845 1846 1847 1848
		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;
		}
1849
		h.h2->tp_padding = 0;
1850 1851
		hdrlen = sizeof(*h.h2);
		break;
1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874
	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;
1875 1876 1877
	default:
		BUG();
	}
L
Linus Torvalds 已提交
1878

1879
	sll = h.raw + TPACKET_ALIGN(hdrlen);
S
Stephen Hemminger 已提交
1880
	sll->sll_halen = dev_parse_header(skb, sll->sll_addr);
L
Linus Torvalds 已提交
1881 1882 1883 1884
	sll->sll_family = AF_PACKET;
	sll->sll_hatype = dev->type;
	sll->sll_protocol = skb->protocol;
	sll->sll_pkttype = skb->pkt_type;
1885
	if (unlikely(po->origdev))
1886 1887 1888
		sll->sll_ifindex = orig_dev->ifindex;
	else
		sll->sll_ifindex = dev->ifindex;
L
Linus Torvalds 已提交
1889

R
Ralf Baechle 已提交
1890
	smp_mb();
1891
#if ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE == 1
L
Linus Torvalds 已提交
1892
	{
1893 1894
		u8 *start, *end;

1895 1896 1897 1898 1899 1900
		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));
		}
1901
		smp_wmb();
L
Linus Torvalds 已提交
1902
	}
1903
#endif
1904 1905 1906 1907
	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
1908 1909 1910 1911 1912 1913 1914 1915 1916

	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:
1917
	kfree_skb(skb);
L
Linus Torvalds 已提交
1918 1919 1920 1921 1922 1923 1924
	return 0;

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

	sk->sk_data_ready(sk, 0);
1925
	kfree_skb(copy_skb);
L
Linus Torvalds 已提交
1926 1927 1928
	goto drop_n_restore;
}

J
Johann Baudy 已提交
1929 1930 1931
static void tpacket_destruct_skb(struct sk_buff *skb)
{
	struct packet_sock *po = pkt_sk(skb->sk);
E
Eric Dumazet 已提交
1932
	void *ph;
L
Linus Torvalds 已提交
1933

J
Johann Baudy 已提交
1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944
	if (likely(po->tx_ring.pg_vec)) {
		ph = skb_shinfo(skb)->destructor_arg;
		BUG_ON(__packet_get_status(po, ph) != TP_STATUS_SENDING);
		BUG_ON(atomic_read(&po->tx_ring.pending) == 0);
		atomic_dec(&po->tx_ring.pending);
		__packet_set_status(po, ph, TP_STATUS_AVAILABLE);
	}

	sock_wfree(skb);
}

E
Eric Dumazet 已提交
1945 1946
static int tpacket_fill_skb(struct packet_sock *po, struct sk_buff *skb,
		void *frame, struct net_device *dev, int size_max,
1947
		__be16 proto, unsigned char *addr, int hlen)
J
Johann Baudy 已提交
1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964
{
	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;
1965
	skb->mark = po->sk.sk_mark;
J
Johann Baudy 已提交
1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976
	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
1977
		pr_err("packet size is too long (%d > %d)\n", tp_len, size_max);
J
Johann Baudy 已提交
1978 1979 1980
		return -EMSGSIZE;
	}

1981
	skb_reserve(skb, hlen);
J
Johann Baudy 已提交
1982 1983 1984 1985 1986 1987 1988 1989 1990 1991
	skb_reset_network_header(skb);

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

	if (sock->type == SOCK_DGRAM) {
		err = dev_hard_header(skb, dev, ntohs(proto), addr,
				NULL, tp_len);
		if (unlikely(err < 0))
			return -EINVAL;
E
Eric Dumazet 已提交
1992
	} else if (dev->hard_header_len) {
J
Johann Baudy 已提交
1993 1994
		/* net device doesn't like empty head */
		if (unlikely(tp_len <= dev->hard_header_len)) {
E
Eric Dumazet 已提交
1995 1996
			pr_err("packet size is too short (%d < %d)\n",
			       tp_len, dev->hard_header_len);
J
Johann Baudy 已提交
1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023
			return -EINVAL;
		}

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

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

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

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

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

		if (unlikely(nr_frags >= MAX_SKB_FRAGS)) {
E
Eric Dumazet 已提交
2024 2025
			pr_err("Packet exceed the number of skb frags(%lu)\n",
			       MAX_SKB_FRAGS);
J
Johann Baudy 已提交
2026 2027 2028
			return -EFAULT;
		}

2029 2030
		page = pgv_to_page(data);
		data += len;
J
Johann Baudy 已提交
2031 2032
		flush_dcache_page(page);
		get_page(page);
2033
		skb_fill_page_desc(skb, nr_frags, page, offset, len);
J
Johann Baudy 已提交
2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047
		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;
2048 2049
	bool need_rls_dev = false;
	int err, reserve = 0;
E
Eric Dumazet 已提交
2050 2051
	void *ph;
	struct sockaddr_ll *saddr = (struct sockaddr_ll *)msg->msg_name;
J
Johann Baudy 已提交
2052 2053 2054 2055
	int tp_len, size_max;
	unsigned char *addr;
	int len_sum = 0;
	int status = 0;
2056
	int hlen, tlen;
J
Johann Baudy 已提交
2057 2058 2059 2060 2061

	mutex_lock(&po->pg_vec_lock);

	err = -EBUSY;
	if (saddr == NULL) {
2062
		dev = po->prot_hook.dev;
J
Johann Baudy 已提交
2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074
		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;
2075 2076
		dev = dev_get_by_index(sock_net(&po->sk), saddr->sll_ifindex);
		need_rls_dev = true;
J
Johann Baudy 已提交
2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089
	}

	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
2090
		- (po->tp_hdrlen - sizeof(struct sockaddr_ll));
J
Johann Baudy 已提交
2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104

	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;
2105 2106
		hlen = LL_RESERVED_SPACE(dev);
		tlen = dev->needed_tailroom;
J
Johann Baudy 已提交
2107
		skb = sock_alloc_send_skb(&po->sk,
2108
				hlen + tlen + sizeof(struct sockaddr_ll),
J
Johann Baudy 已提交
2109 2110 2111 2112 2113 2114
				0, &err);

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

		tp_len = tpacket_fill_skb(po, skb, ph, dev, size_max, proto,
2115
				addr, hlen);
J
Johann Baudy 已提交
2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136

		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);
2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150
		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
2151 2152
		packet_increment_head(&po->tx_ring);
		len_sum += tp_len;
2153 2154 2155 2156
	} while (likely((ph != NULL) ||
			((!(msg->msg_flags & MSG_DONTWAIT)) &&
			 (atomic_read(&po->tx_ring.pending))))
		);
J
Johann Baudy 已提交
2157 2158 2159 2160 2161 2162 2163 2164

	err = len_sum;
	goto out_put;

out_status:
	__packet_set_status(po, ph, status);
	kfree_skb(skb);
out_put:
2165 2166
	if (need_rls_dev)
		dev_put(dev);
J
Johann Baudy 已提交
2167 2168 2169 2170 2171
out:
	mutex_unlock(&po->pg_vec_lock);
	return err;
}

2172 2173 2174 2175
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)
2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195
{
	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
2196
static int packet_snd(struct socket *sock,
L
Linus Torvalds 已提交
2197 2198 2199
			  struct msghdr *msg, size_t len)
{
	struct sock *sk = sock->sk;
E
Eric Dumazet 已提交
2200
	struct sockaddr_ll *saddr = (struct sockaddr_ll *)msg->msg_name;
L
Linus Torvalds 已提交
2201 2202
	struct sk_buff *skb;
	struct net_device *dev;
A
Al Viro 已提交
2203
	__be16 proto;
2204
	bool need_rls_dev = false;
L
Linus Torvalds 已提交
2205
	unsigned char *addr;
2206
	int err, reserve = 0;
2207 2208 2209 2210 2211
	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;
2212
	int hlen, tlen;
L
Linus Torvalds 已提交
2213 2214

	/*
2215
	 *	Get and verify the address.
L
Linus Torvalds 已提交
2216
	 */
2217

L
Linus Torvalds 已提交
2218
	if (saddr == NULL) {
2219
		dev = po->prot_hook.dev;
L
Linus Torvalds 已提交
2220 2221 2222 2223 2224 2225
		proto	= po->num;
		addr	= NULL;
	} else {
		err = -EINVAL;
		if (msg->msg_namelen < sizeof(struct sockaddr_ll))
			goto out;
2226 2227
		if (msg->msg_namelen < (saddr->sll_halen + offsetof(struct sockaddr_ll, sll_addr)))
			goto out;
L
Linus Torvalds 已提交
2228 2229
		proto	= saddr->sll_protocol;
		addr	= saddr->sll_addr;
2230 2231
		dev = dev_get_by_index(sock_net(sk), saddr->sll_ifindex);
		need_rls_dev = true;
L
Linus Torvalds 已提交
2232 2233 2234 2235 2236 2237 2238 2239
	}

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

2240 2241 2242 2243
	err = -ENETDOWN;
	if (!(dev->flags & IFF_UP))
		goto out_unlock;

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 2284 2285 2286 2287 2288 2289 2290 2291
	if (po->has_vnet_hdr) {
		vnet_hdr_len = sizeof(vnet_hdr);

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

		len -= vnet_hdr_len;

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

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

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

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

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

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

		}
	}

L
Linus Torvalds 已提交
2292
	err = -EMSGSIZE;
2293
	if (!gso_type && (len > dev->mtu + reserve + VLAN_HLEN))
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 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332
	if (!gso_type && (len > dev->mtu + reserve)) {
		/* 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;
	}

L
Linus Torvalds 已提交
2357 2358 2359 2360 2361 2362 2363 2364
	/*
	 *	Now send it
	 */

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

2365 2366
	if (need_rls_dev)
		dev_put(dev);
L
Linus Torvalds 已提交
2367

E
Eric Dumazet 已提交
2368
	return len;
L
Linus Torvalds 已提交
2369 2370 2371 2372

out_free:
	kfree_skb(skb);
out_unlock:
2373
	if (dev && need_rls_dev)
L
Linus Torvalds 已提交
2374 2375 2376 2377 2378
		dev_put(dev);
out:
	return err;
}

J
Johann Baudy 已提交
2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389
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
2390 2391 2392 2393 2394 2395 2396 2397 2398
/*
 *	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;
2399
	struct net *net;
2400
	union tpacket_req_u req_u;
L
Linus Torvalds 已提交
2401 2402 2403 2404

	if (!sk)
		return 0;

2405
	net = sock_net(sk);
L
Linus Torvalds 已提交
2406 2407
	po = pkt_sk(sk);

2408 2409
	spin_lock_bh(&net->packet.sklist_lock);
	sk_del_node_init_rcu(sk);
2410
	sock_prot_inuse_add(net, sk->sk_prot, -1);
2411
	spin_unlock_bh(&net->packet.sklist_lock);
L
Linus Torvalds 已提交
2412

2413
	spin_lock(&po->bind_lock);
2414
	unregister_prot_hook(sk, false);
2415 2416 2417 2418
	if (po->prot_hook.dev) {
		dev_put(po->prot_hook.dev);
		po->prot_hook.dev = NULL;
	}
2419
	spin_unlock(&po->bind_lock);
L
Linus Torvalds 已提交
2420 2421 2422

	packet_flush_mclist(sk);

2423
	memset(&req_u, 0, sizeof(req_u));
J
Johann Baudy 已提交
2424 2425

	if (po->rx_ring.pg_vec)
2426
		packet_set_ring(sk, &req_u, 1, 0);
J
Johann Baudy 已提交
2427 2428

	if (po->tx_ring.pg_vec)
2429
		packet_set_ring(sk, &req_u, 1, 1);
L
Linus Torvalds 已提交
2430

D
David S. Miller 已提交
2431 2432
	fanout_release(sk);

2433
	synchronize_net();
L
Linus Torvalds 已提交
2434 2435 2436 2437 2438 2439 2440 2441 2442
	/*
	 *	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);
2443
	sk_refcnt_debug_release(sk);
L
Linus Torvalds 已提交
2444 2445 2446 2447 2448 2449 2450 2451 2452

	sock_put(sk);
	return 0;
}

/*
 *	Attach a packet hook.
 */

A
Al Viro 已提交
2453
static int packet_do_bind(struct sock *sk, struct net_device *dev, __be16 protocol)
L
Linus Torvalds 已提交
2454 2455
{
	struct packet_sock *po = pkt_sk(sk);
D
David S. Miller 已提交
2456 2457 2458

	if (po->fanout)
		return -EINVAL;
L
Linus Torvalds 已提交
2459 2460 2461 2462

	lock_sock(sk);

	spin_lock(&po->bind_lock);
2463
	unregister_prot_hook(sk, true);
L
Linus Torvalds 已提交
2464 2465
	po->num = protocol;
	po->prot_hook.type = protocol;
2466 2467
	if (po->prot_hook.dev)
		dev_put(po->prot_hook.dev);
L
Linus Torvalds 已提交
2468 2469 2470 2471 2472 2473 2474
	po->prot_hook.dev = dev;

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

	if (protocol == 0)
		goto out_unlock;

2475
	if (!dev || (dev->flags & IFF_UP)) {
2476
		register_prot_hook(sk);
2477 2478 2479 2480
	} else {
		sk->sk_err = ENETDOWN;
		if (!sock_flag(sk, SOCK_DEAD))
			sk->sk_error_report(sk);
L
Linus Torvalds 已提交
2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492
	}

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
2493 2494
static int packet_bind_spkt(struct socket *sock, struct sockaddr *uaddr,
			    int addr_len)
L
Linus Torvalds 已提交
2495
{
E
Eric Dumazet 已提交
2496
	struct sock *sk = sock->sk;
L
Linus Torvalds 已提交
2497 2498 2499
	char name[15];
	struct net_device *dev;
	int err = -ENODEV;
2500

L
Linus Torvalds 已提交
2501 2502 2503
	/*
	 *	Check legality
	 */
2504

2505
	if (addr_len != sizeof(struct sockaddr))
L
Linus Torvalds 已提交
2506
		return -EINVAL;
E
Eric Dumazet 已提交
2507
	strlcpy(name, uaddr->sa_data, sizeof(name));
L
Linus Torvalds 已提交
2508

2509
	dev = dev_get_by_name(sock_net(sk), name);
2510
	if (dev)
L
Linus Torvalds 已提交
2511 2512 2513 2514 2515 2516
		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
2517 2518
	struct sockaddr_ll *sll = (struct sockaddr_ll *)uaddr;
	struct sock *sk = sock->sk;
L
Linus Torvalds 已提交
2519 2520 2521 2522 2523 2524 2525
	struct net_device *dev = NULL;
	int err;


	/*
	 *	Check legality
	 */
2526

L
Linus Torvalds 已提交
2527 2528 2529 2530 2531 2532 2533
	if (addr_len < sizeof(struct sockaddr_ll))
		return -EINVAL;
	if (sll->sll_family != AF_PACKET)
		return -EINVAL;

	if (sll->sll_ifindex) {
		err = -ENODEV;
2534
		dev = dev_get_by_index(sock_net(sk), sll->sll_ifindex);
L
Linus Torvalds 已提交
2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550
		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),
};

/*
2551
 *	Create a packet of type SOCK_PACKET.
L
Linus Torvalds 已提交
2552 2553
 */

2554 2555
static int packet_create(struct net *net, struct socket *sock, int protocol,
			 int kern)
L
Linus Torvalds 已提交
2556 2557 2558
{
	struct sock *sk;
	struct packet_sock *po;
A
Al Viro 已提交
2559
	__be16 proto = (__force __be16)protocol; /* weird, but documented */
L
Linus Torvalds 已提交
2560 2561 2562 2563
	int err;

	if (!capable(CAP_NET_RAW))
		return -EPERM;
2564 2565
	if (sock->type != SOCK_DGRAM && sock->type != SOCK_RAW &&
	    sock->type != SOCK_PACKET)
L
Linus Torvalds 已提交
2566 2567 2568 2569 2570
		return -ESOCKTNOSUPPORT;

	sock->state = SS_UNCONNECTED;

	err = -ENOBUFS;
2571
	sk = sk_alloc(net, PF_PACKET, GFP_KERNEL, &packet_proto);
L
Linus Torvalds 已提交
2572 2573 2574 2575 2576 2577
	if (sk == NULL)
		goto out;

	sock->ops = &packet_ops;
	if (sock->type == SOCK_PACKET)
		sock->ops = &packet_ops_spkt;
2578

L
Linus Torvalds 已提交
2579 2580 2581 2582
	sock_init_data(sock, sk);

	po = pkt_sk(sk);
	sk->sk_family = PF_PACKET;
A
Al Viro 已提交
2583
	po->num = proto;
L
Linus Torvalds 已提交
2584 2585

	sk->sk_destruct = packet_sock_destruct;
2586
	sk_refcnt_debug_inc(sk);
L
Linus Torvalds 已提交
2587 2588 2589 2590 2591 2592

	/*
	 *	Attach a protocol block
	 */

	spin_lock_init(&po->bind_lock);
2593
	mutex_init(&po->pg_vec_lock);
L
Linus Torvalds 已提交
2594
	po->prot_hook.func = packet_rcv;
2595

L
Linus Torvalds 已提交
2596 2597
	if (sock->type == SOCK_PACKET)
		po->prot_hook.func = packet_rcv_spkt;
2598

L
Linus Torvalds 已提交
2599 2600
	po->prot_hook.af_packet_priv = sk;

A
Al Viro 已提交
2601 2602
	if (proto) {
		po->prot_hook.type = proto;
2603
		register_prot_hook(sk);
L
Linus Torvalds 已提交
2604 2605
	}

2606 2607
	spin_lock_bh(&net->packet.sklist_lock);
	sk_add_node_rcu(sk, &net->packet.sklist);
2608
	sock_prot_inuse_add(net, &packet_proto, 1);
2609 2610
	spin_unlock_bh(&net->packet.sklist_lock);

E
Eric Dumazet 已提交
2611
	return 0;
L
Linus Torvalds 已提交
2612 2613 2614 2615
out:
	return err;
}

2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 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
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
2661 2662 2663 2664 2665 2666 2667 2668 2669 2670 2671
/*
 *	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;
2672
	struct sockaddr_ll *sll;
2673
	int vnet_hdr_len = 0;
L
Linus Torvalds 已提交
2674 2675

	err = -EINVAL;
2676
	if (flags & ~(MSG_PEEK|MSG_DONTWAIT|MSG_TRUNC|MSG_CMSG_COMPAT|MSG_ERRQUEUE))
L
Linus Torvalds 已提交
2677 2678 2679 2680 2681 2682 2683 2684
		goto out;

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

2685 2686 2687 2688 2689
	if (flags & MSG_ERRQUEUE) {
		err = packet_recv_error(sk, msg, len);
		goto out;
	}

L
Linus Torvalds 已提交
2690 2691 2692 2693 2694 2695 2696 2697 2698
	/*
	 *	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
2699
	skb = skb_recv_datagram(sk, flags, flags & MSG_DONTWAIT, &err);
L
Linus Torvalds 已提交
2700 2701

	/*
2702
	 *	An error occurred so return it. Because skb_recv_datagram()
L
Linus Torvalds 已提交
2703 2704 2705 2706
	 *	handles the blocking we don't see and worry about blocking
	 *	retries.
	 */

2707
	if (skb == NULL)
L
Linus Torvalds 已提交
2708 2709
		goto out;

2710 2711 2712 2713 2714
	if (pkt_sk(sk)->has_vnet_hdr) {
		struct virtio_net_hdr vnet_hdr = { 0 };

		err = -EINVAL;
		vnet_hdr_len = sizeof(vnet_hdr);
2715
		if (len < vnet_hdr_len)
2716 2717
			goto out_free;

2718 2719
		len -= vnet_hdr_len;

2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742
		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;
2743
			vnet_hdr.csum_start = skb_checksum_start_offset(skb);
2744
			vnet_hdr.csum_offset = skb->csum_offset;
2745 2746
		} else if (skb->ip_summed == CHECKSUM_UNNECESSARY) {
			vnet_hdr.flags = VIRTIO_NET_HDR_F_DATA_VALID;
2747 2748 2749 2750 2751 2752 2753 2754
		} /* else everything is zero */

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

2755 2756 2757 2758 2759
	/*
	 *	If the address length field is there to be filled in, we fill
	 *	it in now.
	 */

2760
	sll = &PACKET_SKB_CB(skb)->sa.ll;
2761 2762 2763 2764 2765
	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
2766 2767 2768 2769 2770 2771
	/*
	 *	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
2772 2773 2774
	if (copied > len) {
		copied = len;
		msg->msg_flags |= MSG_TRUNC;
L
Linus Torvalds 已提交
2775 2776 2777 2778 2779 2780
	}

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

2781
	sock_recv_ts_and_drops(msg, sk, skb);
L
Linus Torvalds 已提交
2782 2783

	if (msg->msg_name)
2784 2785
		memcpy(msg->msg_name, &PACKET_SKB_CB(skb)->sa,
		       msg->msg_namelen);
L
Linus Torvalds 已提交
2786

2787
	if (pkt_sk(sk)->auxdata) {
2788 2789 2790 2791 2792 2793 2794 2795
		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;
2796
		aux.tp_net = skb_network_offset(skb);
2797 2798 2799 2800 2801 2802
		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;
		}
2803
		aux.tp_padding = 0;
2804
		put_cmsg(msg, SOL_PACKET, PACKET_AUXDATA, sizeof(aux), &aux);
2805 2806
	}

L
Linus Torvalds 已提交
2807 2808 2809 2810
	/*
	 *	Free or return the buffer as appropriate. Again this
	 *	hides all the races and re-entrancy issues from us.
	 */
2811
	err = vnet_hdr_len + ((flags&MSG_TRUNC) ? skb->len : copied);
L
Linus Torvalds 已提交
2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828

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
2829 2830 2831
	rcu_read_lock();
	dev = dev_get_by_index_rcu(sock_net(sk), pkt_sk(sk)->ifindex);
	if (dev)
2832
		strncpy(uaddr->sa_data, dev->name, 14);
E
Eric Dumazet 已提交
2833
	else
L
Linus Torvalds 已提交
2834
		memset(uaddr->sa_data, 0, 14);
E
Eric Dumazet 已提交
2835
	rcu_read_unlock();
L
Linus Torvalds 已提交
2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846
	*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);
2847
	DECLARE_SOCKADDR(struct sockaddr_ll *, sll, uaddr);
L
Linus Torvalds 已提交
2848 2849 2850 2851 2852 2853 2854

	if (peer)
		return -EOPNOTSUPP;

	sll->sll_family = AF_PACKET;
	sll->sll_ifindex = po->ifindex;
	sll->sll_protocol = po->num;
2855
	sll->sll_pkttype = 0;
E
Eric Dumazet 已提交
2856 2857
	rcu_read_lock();
	dev = dev_get_by_index_rcu(sock_net(sk), po->ifindex);
L
Linus Torvalds 已提交
2858 2859 2860 2861 2862 2863 2864 2865
	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
2866
	rcu_read_unlock();
2867
	*uaddr_len = offsetof(struct sockaddr_ll, sll_addr) + sll->sll_halen;
L
Linus Torvalds 已提交
2868 2869 2870 2871

	return 0;
}

2872 2873
static int packet_dev_mc(struct net_device *dev, struct packet_mclist *i,
			 int what)
L
Linus Torvalds 已提交
2874 2875 2876
{
	switch (i->type) {
	case PACKET_MR_MULTICAST:
2877 2878
		if (i->alen != dev->addr_len)
			return -EINVAL;
L
Linus Torvalds 已提交
2879
		if (what > 0)
2880
			return dev_mc_add(dev, i->addr);
L
Linus Torvalds 已提交
2881
		else
2882
			return dev_mc_del(dev, i->addr);
L
Linus Torvalds 已提交
2883 2884
		break;
	case PACKET_MR_PROMISC:
2885
		return dev_set_promiscuity(dev, what);
L
Linus Torvalds 已提交
2886 2887
		break;
	case PACKET_MR_ALLMULTI:
2888
		return dev_set_allmulti(dev, what);
L
Linus Torvalds 已提交
2889
		break;
2890
	case PACKET_MR_UNICAST:
2891 2892
		if (i->alen != dev->addr_len)
			return -EINVAL;
2893
		if (what > 0)
2894
			return dev_uc_add(dev, i->addr);
2895
		else
2896
			return dev_uc_del(dev, i->addr);
2897
		break;
E
Eric Dumazet 已提交
2898 2899
	default:
		break;
L
Linus Torvalds 已提交
2900
	}
2901
	return 0;
L
Linus Torvalds 已提交
2902 2903 2904 2905
}

static void packet_dev_mclist(struct net_device *dev, struct packet_mclist *i, int what)
{
E
Eric Dumazet 已提交
2906
	for ( ; i; i = i->next) {
L
Linus Torvalds 已提交
2907 2908 2909 2910 2911
		if (i->ifindex == dev->ifindex)
			packet_dev_mc(dev, i, what);
	}
}

2912
static int packet_mc_add(struct sock *sk, struct packet_mreq_max *mreq)
L
Linus Torvalds 已提交
2913 2914 2915 2916 2917 2918 2919 2920 2921
{
	struct packet_sock *po = pkt_sk(sk);
	struct packet_mclist *ml, *i;
	struct net_device *dev;
	int err;

	rtnl_lock();

	err = -ENODEV;
2922
	dev = __dev_get_by_index(sock_net(sk), mreq->mr_ifindex);
L
Linus Torvalds 已提交
2923 2924 2925 2926
	if (!dev)
		goto done;

	err = -EINVAL;
2927
	if (mreq->mr_alen > dev->addr_len)
L
Linus Torvalds 已提交
2928 2929 2930
		goto done;

	err = -ENOBUFS;
2931
	i = kmalloc(sizeof(*i), GFP_KERNEL);
L
Linus Torvalds 已提交
2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954
	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;
2955 2956 2957 2958 2959
	err = packet_dev_mc(dev, i, 1);
	if (err) {
		po->mclist = i->next;
		kfree(i);
	}
L
Linus Torvalds 已提交
2960 2961 2962 2963 2964 2965

done:
	rtnl_unlock();
	return err;
}

2966
static int packet_mc_drop(struct sock *sk, struct packet_mreq_max *mreq)
L
Linus Torvalds 已提交
2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978 2979
{
	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;
2980 2981
				dev = __dev_get_by_index(sock_net(sk), ml->ifindex);
				if (dev)
L
Linus Torvalds 已提交
2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005
					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;
3006 3007
		dev = __dev_get_by_index(sock_net(sk), ml->ifindex);
		if (dev != NULL)
L
Linus Torvalds 已提交
3008 3009 3010 3011 3012 3013 3014
			packet_dev_mc(dev, ml, -1);
		kfree(ml);
	}
	rtnl_unlock();
}

static int
3015
packet_setsockopt(struct socket *sock, int level, int optname, char __user *optval, unsigned int optlen)
L
Linus Torvalds 已提交
3016 3017
{
	struct sock *sk = sock->sk;
3018
	struct packet_sock *po = pkt_sk(sk);
L
Linus Torvalds 已提交
3019 3020 3021 3022 3023
	int ret;

	if (level != SOL_PACKET)
		return -ENOPROTOOPT;

J
Johann Baudy 已提交
3024
	switch (optname) {
3025
	case PACKET_ADD_MEMBERSHIP:
L
Linus Torvalds 已提交
3026 3027
	case PACKET_DROP_MEMBERSHIP:
	{
3028 3029 3030 3031
		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
3032
			return -EINVAL;
3033 3034
		if (len > sizeof(mreq))
			len = sizeof(mreq);
E
Eric Dumazet 已提交
3035
		if (copy_from_user(&mreq, optval, len))
L
Linus Torvalds 已提交
3036
			return -EFAULT;
3037 3038
		if (len < (mreq.mr_alen + offsetof(struct packet_mreq, mr_address)))
			return -EINVAL;
L
Linus Torvalds 已提交
3039 3040 3041 3042 3043 3044
		if (optname == PACKET_ADD_MEMBERSHIP)
			ret = packet_mc_add(sk, &mreq);
		else
			ret = packet_mc_drop(sk, &mreq);
		return ret;
	}
3045

L
Linus Torvalds 已提交
3046
	case PACKET_RX_RING:
J
Johann Baudy 已提交
3047
	case PACKET_TX_RING:
L
Linus Torvalds 已提交
3048
	{
3049 3050
		union tpacket_req_u req_u;
		int len;
L
Linus Torvalds 已提交
3051

3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062
		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
3063
			return -EINVAL;
3064 3065
		if (pkt_sk(sk)->has_vnet_hdr)
			return -EINVAL;
3066
		if (copy_from_user(&req_u.req, optval, len))
L
Linus Torvalds 已提交
3067
			return -EFAULT;
3068 3069
		return packet_set_ring(sk, &req_u, 0,
			optname == PACKET_TX_RING);
L
Linus Torvalds 已提交
3070 3071 3072 3073 3074
	}
	case PACKET_COPY_THRESH:
	{
		int val;

E
Eric Dumazet 已提交
3075
		if (optlen != sizeof(val))
L
Linus Torvalds 已提交
3076
			return -EINVAL;
E
Eric Dumazet 已提交
3077
		if (copy_from_user(&val, optval, sizeof(val)))
L
Linus Torvalds 已提交
3078 3079 3080 3081 3082
			return -EFAULT;

		pkt_sk(sk)->copy_thresh = val;
		return 0;
	}
3083 3084 3085 3086 3087 3088
	case PACKET_VERSION:
	{
		int val;

		if (optlen != sizeof(val))
			return -EINVAL;
J
Johann Baudy 已提交
3089
		if (po->rx_ring.pg_vec || po->tx_ring.pg_vec)
3090 3091 3092 3093 3094 3095
			return -EBUSY;
		if (copy_from_user(&val, optval, sizeof(val)))
			return -EFAULT;
		switch (val) {
		case TPACKET_V1:
		case TPACKET_V2:
3096
		case TPACKET_V3:
3097 3098 3099 3100 3101 3102
			po->tp_version = val;
			return 0;
		default:
			return -EINVAL;
		}
	}
3103 3104 3105 3106 3107 3108
	case PACKET_RESERVE:
	{
		unsigned int val;

		if (optlen != sizeof(val))
			return -EINVAL;
J
Johann Baudy 已提交
3109
		if (po->rx_ring.pg_vec || po->tx_ring.pg_vec)
3110 3111 3112 3113 3114 3115
			return -EBUSY;
		if (copy_from_user(&val, optval, sizeof(val)))
			return -EFAULT;
		po->tp_reserve = val;
		return 0;
	}
J
Johann Baudy 已提交
3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128
	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;
	}
3129 3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140
	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;
	}
3141 3142 3143 3144 3145 3146 3147 3148 3149 3150 3151 3152
	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;
	}
3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168
	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;
	}
3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180
	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
3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191
	case PACKET_FANOUT:
	{
		int val;

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

		return fanout_add(sk, val & 0xffff, val >> 16);
	}
L
Linus Torvalds 已提交
3192 3193 3194 3195 3196 3197 3198 3199 3200
	default:
		return -ENOPROTOOPT;
	}
}

static int packet_getsockopt(struct socket *sock, int level, int optname,
			     char __user *optval, int __user *optlen)
{
	int len;
3201
	int val;
L
Linus Torvalds 已提交
3202 3203
	struct sock *sk = sock->sk;
	struct packet_sock *po = pkt_sk(sk);
3204 3205
	void *data;
	struct tpacket_stats st;
3206
	union tpacket_stats_u st_u;
L
Linus Torvalds 已提交
3207 3208 3209 3210

	if (level != SOL_PACKET)
		return -ENOPROTOOPT;

3211 3212
	if (get_user(len, optlen))
		return -EFAULT;
L
Linus Torvalds 已提交
3213 3214 3215

	if (len < 0)
		return -EINVAL;
3216

J
Johann Baudy 已提交
3217
	switch (optname) {
L
Linus Torvalds 已提交
3218
	case PACKET_STATISTICS:
3219 3220 3221 3222 3223 3224
		if (po->tp_version == TPACKET_V3) {
			len = sizeof(struct tpacket_stats_v3);
		} else {
			if (len > sizeof(struct tpacket_stats))
				len = sizeof(struct tpacket_stats);
		}
L
Linus Torvalds 已提交
3225
		spin_lock_bh(&sk->sk_receive_queue.lock);
3226 3227 3228 3229 3230 3231 3232 3233 3234 3235 3236 3237
		if (po->tp_version == TPACKET_V3) {
			memcpy(&st_u.stats3, &po->stats,
			sizeof(struct tpacket_stats));
			st_u.stats3.tp_freeze_q_cnt =
			po->stats_u.stats3.tp_freeze_q_cnt;
			st_u.stats3.tp_packets += po->stats.tp_drops;
			data = &st_u.stats3;
		} else {
			st = po->stats;
			st.tp_packets += st.tp_drops;
			data = &st;
		}
L
Linus Torvalds 已提交
3238 3239
		memset(&po->stats, 0, sizeof(st));
		spin_unlock_bh(&sk->sk_receive_queue.lock);
3240 3241 3242 3243 3244 3245
		break;
	case PACKET_AUXDATA:
		if (len > sizeof(int))
			len = sizeof(int);
		val = po->auxdata;

3246 3247 3248 3249 3250 3251 3252
		data = &val;
		break;
	case PACKET_ORIGDEV:
		if (len > sizeof(int))
			len = sizeof(int);
		val = po->origdev;

3253 3254 3255 3256 3257 3258 3259
		data = &val;
		break;
	case PACKET_VNET_HDR:
		if (len > sizeof(int))
			len = sizeof(int);
		val = po->has_vnet_hdr;

3260
		data = &val;
L
Linus Torvalds 已提交
3261
		break;
3262 3263 3264 3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276 3277 3278 3279
	case PACKET_VERSION:
		if (len > sizeof(int))
			len = sizeof(int);
		val = po->tp_version;
		data = &val;
		break;
	case PACKET_HDRLEN:
		if (len > sizeof(int))
			len = sizeof(int);
		if (copy_from_user(&val, optval, len))
			return -EFAULT;
		switch (val) {
		case TPACKET_V1:
			val = sizeof(struct tpacket_hdr);
			break;
		case TPACKET_V2:
			val = sizeof(struct tpacket2_hdr);
			break;
3280 3281 3282
		case TPACKET_V3:
			val = sizeof(struct tpacket3_hdr);
			break;
3283 3284 3285 3286 3287
		default:
			return -EINVAL;
		}
		data = &val;
		break;
3288 3289 3290 3291 3292 3293
	case PACKET_RESERVE:
		if (len > sizeof(unsigned int))
			len = sizeof(unsigned int);
		val = po->tp_reserve;
		data = &val;
		break;
J
Johann Baudy 已提交
3294 3295 3296 3297 3298 3299
	case PACKET_LOSS:
		if (len > sizeof(unsigned int))
			len = sizeof(unsigned int);
		val = po->tp_loss;
		data = &val;
		break;
3300 3301 3302 3303 3304 3305
	case PACKET_TIMESTAMP:
		if (len > sizeof(int))
			len = sizeof(int);
		val = po->tp_tstamp;
		data = &val;
		break;
D
David S. Miller 已提交
3306 3307 3308 3309 3310 3311 3312 3313 3314
	case PACKET_FANOUT:
		if (len > sizeof(int))
			len = sizeof(int);
		val = (po->fanout ?
		       ((u32)po->fanout->id |
			((u32)po->fanout->type << 16)) :
		       0);
		data = &val;
		break;
L
Linus Torvalds 已提交
3315 3316 3317 3318
	default:
		return -ENOPROTOOPT;
	}

3319 3320
	if (put_user(len, optlen))
		return -EFAULT;
3321 3322
	if (copy_to_user(optval, data, len))
		return -EFAULT;
3323
	return 0;
L
Linus Torvalds 已提交
3324 3325 3326 3327 3328 3329 3330
}


static int packet_notifier(struct notifier_block *this, unsigned long msg, void *data)
{
	struct sock *sk;
	struct hlist_node *node;
3331
	struct net_device *dev = data;
3332
	struct net *net = dev_net(dev);
L
Linus Torvalds 已提交
3333

3334 3335
	rcu_read_lock();
	sk_for_each_rcu(sk, node, &net->packet.sklist) {
L
Linus Torvalds 已提交
3336 3337 3338 3339 3340 3341
		struct packet_sock *po = pkt_sk(sk);

		switch (msg) {
		case NETDEV_UNREGISTER:
			if (po->mclist)
				packet_dev_mclist(dev, po->mclist, -1);
3342 3343
			/* fallthrough */

L
Linus Torvalds 已提交
3344 3345 3346 3347
		case NETDEV_DOWN:
			if (dev->ifindex == po->ifindex) {
				spin_lock(&po->bind_lock);
				if (po->running) {
3348
					__unregister_prot_hook(sk, false);
L
Linus Torvalds 已提交
3349 3350 3351 3352 3353 3354
					sk->sk_err = ENETDOWN;
					if (!sock_flag(sk, SOCK_DEAD))
						sk->sk_error_report(sk);
				}
				if (msg == NETDEV_UNREGISTER) {
					po->ifindex = -1;
3355 3356
					if (po->prot_hook.dev)
						dev_put(po->prot_hook.dev);
L
Linus Torvalds 已提交
3357 3358 3359 3360 3361 3362
					po->prot_hook.dev = NULL;
				}
				spin_unlock(&po->bind_lock);
			}
			break;
		case NETDEV_UP:
3363 3364
			if (dev->ifindex == po->ifindex) {
				spin_lock(&po->bind_lock);
3365 3366
				if (po->num)
					register_prot_hook(sk);
3367
				spin_unlock(&po->bind_lock);
L
Linus Torvalds 已提交
3368 3369 3370 3371
			}
			break;
		}
	}
3372
	rcu_read_unlock();
L
Linus Torvalds 已提交
3373 3374 3375 3376 3377 3378 3379 3380 3381
	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
3382
	switch (cmd) {
E
Eric Dumazet 已提交
3383 3384 3385
	case SIOCOUTQ:
	{
		int amount = sk_wmem_alloc_get(sk);
3386

E
Eric Dumazet 已提交
3387 3388 3389 3390 3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404
		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);
3405

L
Linus Torvalds 已提交
3406
#ifdef CONFIG_INET
E
Eric Dumazet 已提交
3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421
	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
3422 3423
#endif

E
Eric Dumazet 已提交
3424 3425
	default:
		return -ENOIOCTLCMD;
L
Linus Torvalds 已提交
3426 3427 3428 3429
	}
	return 0;
}

E
Eric Dumazet 已提交
3430
static unsigned int packet_poll(struct file *file, struct socket *sock,
L
Linus Torvalds 已提交
3431 3432 3433 3434 3435 3436 3437
				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
3438
	if (po->rx_ring.pg_vec) {
3439 3440
		if (!packet_previous_rx_frame(po, &po->rx_ring,
			TP_STATUS_KERNEL))
L
Linus Torvalds 已提交
3441 3442 3443
			mask |= POLLIN | POLLRDNORM;
	}
	spin_unlock_bh(&sk->sk_receive_queue.lock);
J
Johann Baudy 已提交
3444 3445 3446 3447 3448 3449
	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
3450 3451 3452 3453 3454 3455 3456 3457 3458 3459 3460
	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
3461
	struct socket *sock = file->private_data;
L
Linus Torvalds 已提交
3462
	struct sock *sk = sock->sk;
3463

L
Linus Torvalds 已提交
3464 3465 3466 3467 3468 3469 3470
	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
3471
	struct socket *sock = file->private_data;
L
Linus Torvalds 已提交
3472
	struct sock *sk = sock->sk;
3473

L
Linus Torvalds 已提交
3474 3475 3476 3477
	if (sk)
		atomic_dec(&pkt_sk(sk)->mapped);
}

3478
static const struct vm_operations_struct packet_mmap_ops = {
E
Eric Dumazet 已提交
3479 3480
	.open	=	packet_mm_open,
	.close	=	packet_mm_close,
L
Linus Torvalds 已提交
3481 3482
};

3483 3484
static void free_pg_vec(struct pgv *pg_vec, unsigned int order,
			unsigned int len)
L
Linus Torvalds 已提交
3485 3486 3487
{
	int i;

3488
	for (i = 0; i < len; i++) {
3489
		if (likely(pg_vec[i].buffer)) {
C
Changli Gao 已提交
3490
			if (is_vmalloc_addr(pg_vec[i].buffer))
3491 3492 3493 3494 3495 3496
				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
3497 3498 3499 3500
	}
	kfree(pg_vec);
}

3501
static char *alloc_one_pg_vec_page(unsigned long order)
3502
{
3503 3504 3505 3506 3507 3508 3509 3510 3511 3512 3513 3514
	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
3515
	buffer = vzalloc((1 << order) * PAGE_SIZE);
3516

3517 3518 3519 3520 3521 3522 3523 3524 3525 3526 3527 3528 3529 3530 3531
	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;
3532 3533
}

3534
static struct pgv *alloc_pg_vec(struct tpacket_req *req, int order)
3535 3536
{
	unsigned int block_nr = req->tp_block_nr;
3537
	struct pgv *pg_vec;
3538 3539
	int i;

3540
	pg_vec = kcalloc(block_nr, sizeof(struct pgv), GFP_KERNEL);
3541 3542 3543 3544
	if (unlikely(!pg_vec))
		goto out;

	for (i = 0; i < block_nr; i++) {
C
Changli Gao 已提交
3545
		pg_vec[i].buffer = alloc_one_pg_vec_page(order);
3546
		if (unlikely(!pg_vec[i].buffer))
3547 3548 3549 3550 3551 3552 3553 3554 3555 3556 3557
			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
3558

3559
static int packet_set_ring(struct sock *sk, union tpacket_req_u *req_u,
J
Johann Baudy 已提交
3560
		int closing, int tx_ring)
L
Linus Torvalds 已提交
3561
{
3562
	struct pgv *pg_vec = NULL;
L
Linus Torvalds 已提交
3563
	struct packet_sock *po = pkt_sk(sk);
A
Al Viro 已提交
3564
	int was_running, order = 0;
J
Johann Baudy 已提交
3565 3566
	struct packet_ring_buffer *rb;
	struct sk_buff_head *rb_queue;
A
Al Viro 已提交
3567
	__be16 num;
3568 3569 3570 3571 3572 3573 3574 3575 3576
	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;
	}
3577

J
Johann Baudy 已提交
3578 3579
	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
3580

J
Johann Baudy 已提交
3581 3582 3583 3584 3585 3586 3587
	err = -EBUSY;
	if (!closing) {
		if (atomic_read(&po->mapped))
			goto out;
		if (atomic_read(&rb->pending))
			goto out;
	}
L
Linus Torvalds 已提交
3588

J
Johann Baudy 已提交
3589 3590 3591 3592 3593
	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
3594

3595 3596 3597 3598 3599 3600 3601
		switch (po->tp_version) {
		case TPACKET_V1:
			po->tp_hdrlen = TPACKET_HDRLEN;
			break;
		case TPACKET_V2:
			po->tp_hdrlen = TPACKET2_HDRLEN;
			break;
3602 3603 3604
		case TPACKET_V3:
			po->tp_hdrlen = TPACKET3_HDRLEN;
			break;
3605 3606
		}

J
Johann Baudy 已提交
3607
		err = -EINVAL;
3608
		if (unlikely((int)req->tp_block_size <= 0))
J
Johann Baudy 已提交
3609
			goto out;
3610
		if (unlikely(req->tp_block_size & (PAGE_SIZE - 1)))
J
Johann Baudy 已提交
3611
			goto out;
3612
		if (unlikely(req->tp_frame_size < po->tp_hdrlen +
J
Johann Baudy 已提交
3613 3614
					po->tp_reserve))
			goto out;
3615
		if (unlikely(req->tp_frame_size & (TPACKET_ALIGNMENT - 1)))
J
Johann Baudy 已提交
3616
			goto out;
L
Linus Torvalds 已提交
3617

J
Johann Baudy 已提交
3618 3619 3620 3621 3622 3623
		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
3624 3625

		err = -ENOMEM;
3626 3627 3628
		order = get_order(req->tp_block_size);
		pg_vec = alloc_pg_vec(req, order);
		if (unlikely(!pg_vec))
L
Linus Torvalds 已提交
3629
			goto out;
3630 3631 3632 3633 3634 3635 3636 3637 3638 3639 3640
		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
3641 3642 3643 3644
	}
	/* Done */
	else {
		err = -EINVAL;
3645
		if (unlikely(req->tp_frame_nr))
J
Johann Baudy 已提交
3646
			goto out;
L
Linus Torvalds 已提交
3647 3648 3649 3650 3651 3652 3653 3654 3655 3656
	}

	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;
3657
		__unregister_prot_hook(sk, false);
L
Linus Torvalds 已提交
3658 3659
	}
	spin_unlock(&po->bind_lock);
3660

L
Linus Torvalds 已提交
3661 3662 3663
	synchronize_net();

	err = -EBUSY;
3664
	mutex_lock(&po->pg_vec_lock);
L
Linus Torvalds 已提交
3665 3666
	if (closing || atomic_read(&po->mapped) == 0) {
		err = 0;
J
Johann Baudy 已提交
3667
		spin_lock_bh(&rb_queue->lock);
3668
		swap(rb->pg_vec, pg_vec);
J
Johann Baudy 已提交
3669 3670 3671 3672 3673
		rb->frame_max = (req->tp_frame_nr - 1);
		rb->head = 0;
		rb->frame_size = req->tp_frame_size;
		spin_unlock_bh(&rb_queue->lock);

3674 3675
		swap(rb->pg_vec_order, order);
		swap(rb->pg_vec_len, req->tp_block_nr);
J
Johann Baudy 已提交
3676 3677 3678 3679 3680

		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
3681
		if (atomic_read(&po->mapped))
E
Eric Dumazet 已提交
3682 3683
			pr_err("packet_mmap: vma is busy: %d\n",
			       atomic_read(&po->mapped));
L
Linus Torvalds 已提交
3684
	}
3685
	mutex_unlock(&po->pg_vec_lock);
L
Linus Torvalds 已提交
3686 3687

	spin_lock(&po->bind_lock);
3688
	if (was_running) {
L
Linus Torvalds 已提交
3689
		po->num = num;
3690
		register_prot_hook(sk);
L
Linus Torvalds 已提交
3691 3692
	}
	spin_unlock(&po->bind_lock);
3693 3694 3695 3696 3697
	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
3698 3699 3700 3701 3702 3703 3704 3705
	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
3706 3707
static int packet_mmap(struct file *file, struct socket *sock,
		struct vm_area_struct *vma)
L
Linus Torvalds 已提交
3708 3709 3710
{
	struct sock *sk = sock->sk;
	struct packet_sock *po = pkt_sk(sk);
J
Johann Baudy 已提交
3711 3712
	unsigned long size, expected_size;
	struct packet_ring_buffer *rb;
L
Linus Torvalds 已提交
3713 3714 3715 3716 3717 3718 3719
	unsigned long start;
	int err = -EINVAL;
	int i;

	if (vma->vm_pgoff)
		return -EINVAL;

3720
	mutex_lock(&po->pg_vec_lock);
J
Johann Baudy 已提交
3721 3722 3723 3724 3725 3726 3727 3728 3729 3730 3731

	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
3732
		goto out;
J
Johann Baudy 已提交
3733 3734 3735

	size = vma->vm_end - vma->vm_start;
	if (size != expected_size)
L
Linus Torvalds 已提交
3736 3737 3738
		goto out;

	start = vma->vm_start;
J
Johann Baudy 已提交
3739 3740 3741 3742 3743
	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++) {
3744 3745
			struct page *page;
			void *kaddr = rb->pg_vec[i].buffer;
J
Johann Baudy 已提交
3746 3747
			int pg_num;

C
Changli Gao 已提交
3748 3749
			for (pg_num = 0; pg_num < rb->pg_vec_pages; pg_num++) {
				page = pgv_to_page(kaddr);
J
Johann Baudy 已提交
3750 3751 3752 3753
				err = vm_insert_page(vma, start, page);
				if (unlikely(err))
					goto out;
				start += PAGE_SIZE;
3754
				kaddr += PAGE_SIZE;
J
Johann Baudy 已提交
3755
			}
3756
		}
L
Linus Torvalds 已提交
3757
	}
J
Johann Baudy 已提交
3758

3759
	atomic_inc(&po->mapped);
L
Linus Torvalds 已提交
3760 3761 3762 3763
	vma->vm_ops = &packet_mmap_ops;
	err = 0;

out:
3764
	mutex_unlock(&po->pg_vec_lock);
L
Linus Torvalds 已提交
3765 3766 3767
	return err;
}

3768
static const struct proto_ops packet_ops_spkt = {
L
Linus Torvalds 已提交
3769 3770 3771 3772 3773 3774 3775 3776 3777 3778 3779 3780 3781 3782 3783 3784 3785 3786 3787 3788
	.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,
};

3789
static const struct proto_ops packet_ops = {
L
Linus Torvalds 已提交
3790 3791 3792 3793 3794 3795 3796
	.family =	PF_PACKET,
	.owner =	THIS_MODULE,
	.release =	packet_release,
	.bind =		packet_bind,
	.connect =	sock_no_connect,
	.socketpair =	sock_no_socketpair,
	.accept =	sock_no_accept,
3797
	.getname =	packet_getname,
L
Linus Torvalds 已提交
3798 3799 3800 3801 3802 3803 3804 3805 3806 3807 3808 3809
	.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,
};

3810
static const struct net_proto_family packet_family_ops = {
L
Linus Torvalds 已提交
3811 3812 3813 3814 3815 3816
	.family =	PF_PACKET,
	.create =	packet_create,
	.owner	=	THIS_MODULE,
};

static struct notifier_block packet_netdev_notifier = {
E
Eric Dumazet 已提交
3817
	.notifier_call =	packet_notifier,
L
Linus Torvalds 已提交
3818 3819 3820 3821 3822
};

#ifdef CONFIG_PROC_FS

static void *packet_seq_start(struct seq_file *seq, loff_t *pos)
3823
	__acquires(RCU)
L
Linus Torvalds 已提交
3824
{
3825
	struct net *net = seq_file_net(seq);
3826 3827 3828

	rcu_read_lock();
	return seq_hlist_start_head_rcu(&net->packet.sklist, *pos);
L
Linus Torvalds 已提交
3829 3830 3831 3832
}

static void *packet_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
3833
	struct net *net = seq_file_net(seq);
3834
	return seq_hlist_next_rcu(v, &net->packet.sklist, pos);
L
Linus Torvalds 已提交
3835 3836 3837
}

static void packet_seq_stop(struct seq_file *seq, void *v)
3838
	__releases(RCU)
L
Linus Torvalds 已提交
3839
{
3840
	rcu_read_unlock();
L
Linus Torvalds 已提交
3841 3842
}

3843
static int packet_seq_show(struct seq_file *seq, void *v)
L
Linus Torvalds 已提交
3844 3845 3846 3847
{
	if (v == SEQ_START_TOKEN)
		seq_puts(seq, "sk       RefCnt Type Proto  Iface R Rmem   User   Inode\n");
	else {
3848
		struct sock *s = sk_entry(v);
L
Linus Torvalds 已提交
3849 3850 3851
		const struct packet_sock *po = pkt_sk(s);

		seq_printf(seq,
D
Dan Rosenberg 已提交
3852
			   "%pK %-6d %-4d %04x   %-5d %1d %-6u %-6u %-6lu\n",
L
Linus Torvalds 已提交
3853 3854 3855 3856 3857 3858 3859 3860
			   s,
			   atomic_read(&s->sk_refcnt),
			   s->sk_type,
			   ntohs(po->num),
			   po->ifindex,
			   po->running,
			   atomic_read(&s->sk_rmem_alloc),
			   sock_i_uid(s),
E
Eric Dumazet 已提交
3861
			   sock_i_ino(s));
L
Linus Torvalds 已提交
3862 3863 3864 3865 3866
	}

	return 0;
}

3867
static const struct seq_operations packet_seq_ops = {
L
Linus Torvalds 已提交
3868 3869 3870 3871 3872 3873 3874 3875
	.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)
{
3876 3877
	return seq_open_net(inode, file, &packet_seq_ops,
			    sizeof(struct seq_net_private));
L
Linus Torvalds 已提交
3878 3879
}

3880
static const struct file_operations packet_seq_fops = {
L
Linus Torvalds 已提交
3881 3882 3883 3884
	.owner		= THIS_MODULE,
	.open		= packet_seq_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
3885
	.release	= seq_release_net,
L
Linus Torvalds 已提交
3886 3887 3888 3889
};

#endif

3890
static int __net_init packet_net_init(struct net *net)
3891
{
3892
	spin_lock_init(&net->packet.sklist_lock);
3893
	INIT_HLIST_HEAD(&net->packet.sklist);
3894 3895 3896 3897 3898 3899 3900

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

	return 0;
}

3901
static void __net_exit packet_net_exit(struct net *net)
3902 3903 3904 3905 3906 3907 3908 3909 3910 3911
{
	proc_net_remove(net, "packet");
}

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


L
Linus Torvalds 已提交
3912 3913 3914
static void __exit packet_exit(void)
{
	unregister_netdevice_notifier(&packet_netdev_notifier);
3915
	unregister_pernet_subsys(&packet_net_ops);
L
Linus Torvalds 已提交
3916 3917 3918 3919 3920 3921 3922 3923 3924 3925 3926 3927
	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);
3928
	register_pernet_subsys(&packet_net_ops);
L
Linus Torvalds 已提交
3929 3930 3931 3932 3933 3934 3935 3936 3937
	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);