af_packet.c 90.2 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7
/*
 * INET		An implementation of the TCP/IP protocol suite for the LINUX
 *		operating system.  INET is implemented using the  BSD Socket
 *		interface as the means of communication with the user level.
 *
 *		PACKET - implements raw packet sockets.
 *
8
 * Authors:	Ross Biro
L
Linus Torvalds 已提交
9 10 11
 *		Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
 *		Alan Cox, <gw4pts@gw4pts.ampr.org>
 *
12
 * Fixes:
L
Linus Torvalds 已提交
13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
 *		Alan Cox	:	verify_area() now used correctly
 *		Alan Cox	:	new skbuff lists, look ma no backlogs!
 *		Alan Cox	:	tidied skbuff lists.
 *		Alan Cox	:	Now uses generic datagram routines I
 *					added. Also fixed the peek/read crash
 *					from all old Linux datagram code.
 *		Alan Cox	:	Uses the improved datagram code.
 *		Alan Cox	:	Added NULL's for socket options.
 *		Alan Cox	:	Re-commented the code.
 *		Alan Cox	:	Use new kernel side addressing
 *		Rob Janssen	:	Correct MTU usage.
 *		Dave Platt	:	Counter leaks caused by incorrect
 *					interrupt locking and some slightly
 *					dubious gcc output. Can you read
 *					compiler: it said _VOLATILE_
 *	Richard Kooijman	:	Timestamp fixes.
 *		Alan Cox	:	New buffers. Use sk->mac.raw.
 *		Alan Cox	:	sendmsg/recvmsg support.
 *		Alan Cox	:	Protocol setting support
 *	Alexey Kuznetsov	:	Untied from IPv4 stack.
 *	Cyrus Durgin		:	Fixed kerneld for kmod.
 *	Michal Ostrowski        :       Module initialization cleanup.
35
 *         Ulises Alonso        :       Frame number limit removal and
L
Linus Torvalds 已提交
36
 *                                      packet_set_ring memory leak.
37 38 39
 *		Eric Biederman	:	Allow for > 8 byte hardware addresses.
 *					The convention is that longer addresses
 *					will simply extend the hardware address
40
 *					byte arrays at the end of sockaddr_ll
41
 *					and packet_mreq.
J
Johann Baudy 已提交
42
 *		Johann Baudy	:	Added TX RING.
43 44 45 46
 *		Chetan Loke	:	Implemented TPACKET_V3 block abstraction
 *					layer.
 *					Copyright (C) 2011, <lokec@ccs.neu.edu>
 *
L
Linus Torvalds 已提交
47 48 49 50 51 52 53
 *
 *		This program is free software; you can redistribute it and/or
 *		modify it under the terms of the GNU General Public License
 *		as published by the Free Software Foundation; either version
 *		2 of the License, or (at your option) any later version.
 *
 */
54

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

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

97 98
#include "internal.h"

L
Linus Torvalds 已提交
99 100 101 102 103 104 105 106 107 108 109 110 111 112 113
/*
   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
114 115
   mac_header -> ll header
   data       -> data
L
Linus Torvalds 已提交
116 117

Outgoing, dev->hard_header!=NULL
118 119
   mac_header -> ll header
   data       -> ll header
L
Linus Torvalds 已提交
120 121

Incoming, dev->hard_header==NULL
122 123
   mac_header -> UNKNOWN position. It is very likely, that it points to ll
		 header.  PPP makes it, that is wrong, because introduce
124
		 assymetry between rx and tx paths.
125
   data       -> data
L
Linus Torvalds 已提交
126 127

Outgoing, dev->hard_header==NULL
128 129
   mac_header -> data. ll header is still not built!
   data       -> data
L
Linus Torvalds 已提交
130 131 132 133 134 135 136 137 138

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


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

dev->hard_header != NULL
139 140
   mac_header -> ll header
   data       -> ll header
L
Linus Torvalds 已提交
141 142

dev->hard_header == NULL (ll header is added by device, we cannot control it)
143 144
   mac_header -> data
   data       -> data
L
Linus Torvalds 已提交
145 146 147 148 149 150 151

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

/* Private packet socket structures. */

152 153 154
/* identical to struct packet_mreq except it has
 * a longer address field.
 */
E
Eric Dumazet 已提交
155
struct packet_mreq_max {
156 157 158 159
	int		mr_ifindex;
	unsigned short	mr_type;
	unsigned short	mr_alen;
	unsigned char	mr_address[MAX_ADDR_LEN];
L
Linus Torvalds 已提交
160
};
161

162 163 164 165 166 167 168
union tpacket_uhdr {
	struct tpacket_hdr  *h1;
	struct tpacket2_hdr *h2;
	struct tpacket3_hdr *h3;
	void *raw;
};

169
static int packet_set_ring(struct sock *sk, union tpacket_req_u *req_u,
J
Johann Baudy 已提交
170 171
		int closing, int tx_ring);

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))

#define PGV_FROM_VMALLOC 1
J
Johann Baudy 已提交
180

181 182 183 184 185 186 187 188
#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
189 190
struct packet_sock;
static int tpacket_snd(struct packet_sock *po, struct msghdr *msg);
191 192
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
193

194 195 196 197
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);
198 199 200
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 *,
201
			struct packet_sock *);
202
static void prb_retire_current_block(struct tpacket_kbdq_core *,
203
		struct packet_sock *, unsigned int status);
204 205 206
static int prb_queue_frozen(struct tpacket_kbdq_core *);
static void prb_open_block(struct tpacket_kbdq_core *,
		struct tpacket_block_desc *);
207
static void prb_retire_rx_blk_timer_expired(unsigned long);
208 209 210 211 212 213 214 215 216
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
217 218
static void packet_flush_mclist(struct sock *sk);

219 220 221 222 223 224 225 226 227
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))
228

229
#define GET_PBDQC_FROM_RB(x)	((struct tpacket_kbdq_core *)(&(x)->prb_bdqc))
230
#define GET_PBLOCK_DESC(x, bid)	\
231
	((struct tpacket_block_desc *)((x)->pkbdq[(bid)].buffer))
232
#define GET_CURR_PBLOCK_DESC_FROM_CORE(x)	\
233
	((struct tpacket_block_desc *)((x)->pkbdq[(x)->kactive_blk_num].buffer))
234 235 236 237
#define GET_NEXT_PRB_BLK_NUM(x) \
	(((x)->kactive_blk_num < ((x)->knum_blocks-1)) ? \
	((x)->kactive_blk_num+1) : 0)

D
David S. Miller 已提交
238 239 240
static void __fanout_unlink(struct sock *sk, struct packet_sock *po);
static void __fanout_link(struct sock *sk, struct packet_sock *po);

241 242 243 244 245 246 247 248
/* 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
249 250 251 252
		if (po->fanout)
			__fanout_link(sk, po);
		else
			dev_add_pack(&po->prot_hook);
253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269
		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
270 271 272 273
	if (po->fanout)
		__fanout_unlink(sk, po);
	else
		__dev_remove_pack(&po->prot_hook);
274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290
	__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);
}

291
static inline __pure struct page *pgv_to_page(void *addr)
292 293 294 295 296 297
{
	if (is_vmalloc_addr(addr))
		return vmalloc_to_page(addr);
	return virt_to_page(addr);
}

J
Johann Baudy 已提交
298
static void __packet_set_status(struct packet_sock *po, void *frame, int status)
L
Linus Torvalds 已提交
299
{
300
	union tpacket_uhdr h;
L
Linus Torvalds 已提交
301

J
Johann Baudy 已提交
302
	h.raw = frame;
303 304
	switch (po->tp_version) {
	case TPACKET_V1:
J
Johann Baudy 已提交
305
		h.h1->tp_status = status;
306
		flush_dcache_page(pgv_to_page(&h.h1->tp_status));
307 308
		break;
	case TPACKET_V2:
J
Johann Baudy 已提交
309
		h.h2->tp_status = status;
310
		flush_dcache_page(pgv_to_page(&h.h2->tp_status));
311
		break;
312
	case TPACKET_V3:
J
Johann Baudy 已提交
313
	default:
314
		WARN(1, "TPACKET version not supported.\n");
J
Johann Baudy 已提交
315
		BUG();
316
	}
J
Johann Baudy 已提交
317 318

	smp_wmb();
319 320
}

J
Johann Baudy 已提交
321
static int __packet_get_status(struct packet_sock *po, void *frame)
322
{
323
	union tpacket_uhdr h;
324

J
Johann Baudy 已提交
325 326
	smp_rmb();

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

343 344
static __u32 tpacket_get_timestamp(struct sk_buff *skb, struct timespec *ts,
				   unsigned int flags)
345 346 347 348 349 350
{
	struct skb_shared_hwtstamps *shhwtstamps = skb_hwtstamps(skb);

	if (shhwtstamps) {
		if ((flags & SOF_TIMESTAMPING_SYS_HARDWARE) &&
		    ktime_to_timespec_cond(shhwtstamps->syststamp, ts))
351
			return TP_STATUS_TS_SYS_HARDWARE;
352 353
		if ((flags & SOF_TIMESTAMPING_RAW_HARDWARE) &&
		    ktime_to_timespec_cond(shhwtstamps->hwtstamp, ts))
354
			return TP_STATUS_TS_RAW_HARDWARE;
355 356 357
	}

	if (ktime_to_timespec_cond(skb->tstamp, ts))
358
		return TP_STATUS_TS_SOFTWARE;
359

360
	return 0;
361 362
}

363 364
static __u32 __packet_set_timestamp(struct packet_sock *po, void *frame,
				    struct sk_buff *skb)
365 366 367
{
	union tpacket_uhdr h;
	struct timespec ts;
368
	__u32 ts_status;
369

370 371
	if (!(ts_status = tpacket_get_timestamp(skb, &ts, po->tp_tstamp)))
		return 0;
372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391

	h.raw = frame;
	switch (po->tp_version) {
	case TPACKET_V1:
		h.h1->tp_sec = ts.tv_sec;
		h.h1->tp_usec = ts.tv_nsec / NSEC_PER_USEC;
		break;
	case TPACKET_V2:
		h.h2->tp_sec = ts.tv_sec;
		h.h2->tp_nsec = ts.tv_nsec;
		break;
	case TPACKET_V3:
	default:
		WARN(1, "TPACKET version not supported.\n");
		BUG();
	}

	/* one flush is safe, as both fields always lie on the same cacheline */
	flush_dcache_page(pgv_to_page(&h.h1->tp_sec));
	smp_wmb();
392 393

	return ts_status;
394 395
}

J
Johann Baudy 已提交
396 397 398 399 400 401
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;
402
	union tpacket_uhdr h;
J
Johann Baudy 已提交
403 404 405 406

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

407 408
	h.raw = rb->pg_vec[pg_vec_pos].buffer +
		(frame_offset * rb->frame_size);
J
Johann Baudy 已提交
409 410 411 412 413 414 415

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

	return h.raw;
}

416
static void *packet_current_frame(struct packet_sock *po,
J
Johann Baudy 已提交
417 418 419 420 421 422
		struct packet_ring_buffer *rb,
		int status)
{
	return packet_lookup_frame(po, rb, rb->head, status);
}

423
static void prb_del_retire_blk_timer(struct tpacket_kbdq_core *pkc)
424 425 426 427 428 429 430 431
{
	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)
{
432
	struct tpacket_kbdq_core *pkc;
433 434 435 436 437 438 439 440 441 442 443

	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,
444
		struct tpacket_kbdq_core *pkc,
445 446 447 448 449 450 451 452 453 454
		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)
{
455
	struct tpacket_kbdq_core *pkc;
456 457 458 459 460 461 462 463 464 465 466 467 468

	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;
469 470
	struct ethtool_cmd ecmd;
	int err;
471
	u32 speed;
472

473 474 475 476
	rtnl_lock();
	dev = __dev_get_by_index(sock_net(&po->sk), po->ifindex);
	if (unlikely(!dev)) {
		rtnl_unlock();
477
		return DEFAULT_PRB_RETIRE_TOV;
478 479
	}
	err = __ethtool_get_settings(dev, &ecmd);
480
	speed = ethtool_cmd_speed(&ecmd);
481 482 483 484 485 486
	rtnl_unlock();
	if (!err) {
		/*
		 * If the link speed is so slow you don't really
		 * need to worry about perf anyways
		 */
487
		if (speed < SPEED_1000 || speed == SPEED_UNKNOWN) {
488
			return DEFAULT_PRB_RETIRE_TOV;
489 490 491
		} else {
			msec = 1;
			div = speed / 1000;
492 493 494 495 496 497 498 499 500 501 502 503 504 505 506
		}
	}

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

	if (div)
		mbits /= div;

	tmo = mbits * msec;

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

507
static void prb_init_ft_ops(struct tpacket_kbdq_core *p1,
508 509 510 511 512 513 514 515 516 517
			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)
{
518 519
	struct tpacket_kbdq_core *p1 = &rb->prb_bdqc;
	struct tpacket_block_desc *pbd;
520 521 522 523 524

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

	p1->knxt_seq_num = 1;
	p1->pkbdq = pg_vec;
525
	pbd = (struct tpacket_block_desc *)pg_vec[0].buffer;
J
Joe Perches 已提交
526
	p1->pkblk_start	= pg_vec[0].buffer;
527 528 529 530 531
	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;
532
	po->stats.stats3.tp_freeze_q_cnt = 0;
533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548
	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.
 */
549
static void _prb_refresh_rx_retire_blk_timer(struct tpacket_kbdq_core *pkc)
550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581
{
	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;
582
	struct tpacket_kbdq_core *pkc = &po->rx_ring.prb_bdqc;
583
	unsigned int frozen;
584
	struct tpacket_block_desc *pbd;
585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647

	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);
}

648
static void prb_flush_block(struct tpacket_kbdq_core *pkc1,
649
		struct tpacket_block_desc *pbd1, __u32 status)
650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690
{
	/* 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.
 */
691 692
static void prb_close_block(struct tpacket_kbdq_core *pkc1,
		struct tpacket_block_desc *pbd1,
693 694 695 696 697
		struct packet_sock *po, unsigned int stat)
{
	__u32 status = TP_STATUS_USER | stat;

	struct tpacket3_hdr *last_pkt;
698
	struct tpacket_hdr_v1 *h1 = &pbd1->hdr.bh1;
699

700
	if (po->stats.stats3.tp_drops)
701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725
		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);
}

726
static void prb_thaw_queue(struct tpacket_kbdq_core *pkc)
727 728 729 730 731 732 733 734 735 736 737
{
	pkc->reset_pending_on_curr_blk = 0;
}

/*
 * Side effect of opening a block:
 *
 * 1) prb_queue is thawed.
 * 2) retire_blk_timer is refreshed.
 *
 */
738 739
static void prb_open_block(struct tpacket_kbdq_core *pkc1,
	struct tpacket_block_desc *pbd1)
740 741
{
	struct timespec ts;
742
	struct tpacket_hdr_v1 *h1 = &pbd1->hdr.bh1;
743 744 745

	smp_rmb();

746 747 748
	/* We could have just memset this but we will lose the
	 * flexibility of making the priv area sticky
	 */
749

750 751 752
	BLOCK_SNUM(pbd1) = pkc1->knxt_seq_num++;
	BLOCK_NUM_PKTS(pbd1) = 0;
	BLOCK_LEN(pbd1) = BLK_PLUS_PRIV(pkc1->blk_sizeof_priv);
753

754 755 756 757
	getnstimeofday(&ts);

	h1->ts_first_pkt.ts_sec = ts.tv_sec;
	h1->ts_first_pkt.ts_nsec = ts.tv_nsec;
758

759 760 761 762 763 764 765 766 767 768 769 770 771 772
	pkc1->pkblk_start = (char *)pbd1;
	pkc1->nxt_offset = 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();
773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797
}

/*
 * 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.
 */
798
static void prb_freeze_queue(struct tpacket_kbdq_core *pkc,
799 800 801
				  struct packet_sock *po)
{
	pkc->reset_pending_on_curr_blk = 1;
802
	po->stats.stats3.tp_freeze_q_cnt++;
803 804 805 806 807 808 809 810 811 812
}

#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.
 */
813
static void *prb_dispatch_next_block(struct tpacket_kbdq_core *pkc,
814 815
		struct packet_sock *po)
{
816
	struct tpacket_block_desc *pbd;
817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837

	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;
}

838
static void prb_retire_current_block(struct tpacket_kbdq_core *pkc,
839 840
		struct packet_sock *po, unsigned int status)
{
841
	struct tpacket_block_desc *pbd = GET_CURR_PBLOCK_DESC_FROM_CORE(pkc);
842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864

	/* 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;
	}
}

865
static int prb_curr_blk_in_use(struct tpacket_kbdq_core *pkc,
866
				      struct tpacket_block_desc *pbd)
867 868 869 870
{
	return TP_STATUS_USER & BLOCK_STATUS(pbd);
}

871
static int prb_queue_frozen(struct tpacket_kbdq_core *pkc)
872 873 874 875
{
	return pkc->reset_pending_on_curr_blk;
}

876
static void prb_clear_blk_fill_status(struct packet_ring_buffer *rb)
877
{
878
	struct tpacket_kbdq_core *pkc  = GET_PBDQC_FROM_RB(rb);
879 880 881
	atomic_dec(&pkc->blk_fill_in_prog);
}

882
static void prb_fill_rxhash(struct tpacket_kbdq_core *pkc,
883 884 885 886 887
			struct tpacket3_hdr *ppd)
{
	ppd->hv1.tp_rxhash = skb_get_rxhash(pkc->skb);
}

888
static void prb_clear_rxhash(struct tpacket_kbdq_core *pkc,
889 890 891 892 893
			struct tpacket3_hdr *ppd)
{
	ppd->hv1.tp_rxhash = 0;
}

894
static void prb_fill_vlan_info(struct tpacket_kbdq_core *pkc,
895 896 897 898 899 900
			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 {
901 902
		ppd->hv1.tp_vlan_tci = 0;
		ppd->tp_status = TP_STATUS_AVAILABLE;
903 904 905
	}
}

906
static void prb_run_all_ft_ops(struct tpacket_kbdq_core *pkc,
907 908 909 910 911 912 913 914 915 916
			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);
}

917
static void prb_fill_curr_block(char *curr,
918 919
				struct tpacket_kbdq_core *pkc,
				struct tpacket_block_desc *pbd,
920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940
				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
					    )
{
941 942
	struct tpacket_kbdq_core *pkc;
	struct tpacket_block_desc *pbd;
943 944
	char *curr, *end;

J
Joe Perches 已提交
945
	pkc = GET_PBDQC_FROM_RB(&po->rx_ring);
946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970
	pbd = GET_CURR_PBLOCK_DESC_FROM_CORE(pkc);

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

	smp_mb();
	curr = pkc->nxt_offset;
	pkc->skb = skb;
J
Joe Perches 已提交
971
	end = (char *)pbd + pkc->kblk_size;
972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996

	/* 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;
}

997
static void *packet_current_rx_frame(struct packet_sock *po,
998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012
					    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();
1013
		return NULL;
1014 1015 1016
	}
}

1017
static void *prb_lookup_block(struct packet_sock *po,
1018
				     struct packet_ring_buffer *rb,
1019
				     unsigned int idx,
1020 1021
				     int status)
{
1022
	struct tpacket_kbdq_core *pkc  = GET_PBDQC_FROM_RB(rb);
1023
	struct tpacket_block_desc *pbd = GET_PBLOCK_DESC(pkc, idx);
1024 1025 1026 1027 1028 1029

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

1030
static int prb_previous_blk_num(struct packet_ring_buffer *rb)
1031 1032 1033 1034 1035 1036 1037 1038 1039 1040
{
	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 */
1041
static void *__prb_previous_block(struct packet_sock *po,
1042 1043 1044 1045 1046 1047 1048
					 struct packet_ring_buffer *rb,
					 int status)
{
	unsigned int previous = prb_previous_blk_num(rb);
	return prb_lookup_block(po, rb, previous, status);
}

1049
static void *packet_previous_rx_frame(struct packet_sock *po,
1050 1051 1052 1053 1054 1055 1056 1057 1058
					     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);
}

1059
static void packet_increment_rx_head(struct packet_sock *po,
1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073
					    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;
	}
}

1074
static void *packet_previous_frame(struct packet_sock *po,
J
Johann Baudy 已提交
1075 1076 1077 1078 1079 1080 1081
		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);
}

1082
static void packet_increment_head(struct packet_ring_buffer *buff)
J
Johann Baudy 已提交
1083 1084 1085 1086
{
	buff->head = buff->head != buff->frame_max ? buff->head+1 : 0;
}

1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109
static bool packet_rcv_has_room(struct packet_sock *po, struct sk_buff *skb)
{
	struct sock *sk = &po->sk;
	bool has_room;

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

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

	return has_room;
}

L
Linus Torvalds 已提交
1110 1111
static void packet_sock_destruct(struct sock *sk)
{
1112 1113
	skb_queue_purge(&sk->sk_error_queue);

1114 1115
	WARN_ON(atomic_read(&sk->sk_rmem_alloc));
	WARN_ON(atomic_read(&sk->sk_wmem_alloc));
L
Linus Torvalds 已提交
1116 1117

	if (!sock_flag(sk, SOCK_DEAD)) {
E
Eric Dumazet 已提交
1118
		pr_err("Attempt to release alive packet socket: %p\n", sk);
L
Linus Torvalds 已提交
1119 1120 1121
		return;
	}

1122
	sk_refcnt_debug_dec(sk);
L
Linus Torvalds 已提交
1123 1124
}

D
David S. Miller 已提交
1125 1126 1127 1128 1129 1130 1131 1132 1133 1134
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;
}

1135 1136 1137
static unsigned int fanout_demux_hash(struct packet_fanout *f,
				      struct sk_buff *skb,
				      unsigned int num)
D
David S. Miller 已提交
1138
{
1139
	return (((u64)skb->rxhash) * num) >> 32;
D
David S. Miller 已提交
1140 1141
}

1142 1143 1144
static unsigned int fanout_demux_lb(struct packet_fanout *f,
				    struct sk_buff *skb,
				    unsigned int num)
D
David S. Miller 已提交
1145 1146 1147 1148 1149 1150 1151
{
	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;
1152 1153 1154 1155 1156 1157 1158 1159
	return cur;
}

static unsigned int fanout_demux_cpu(struct packet_fanout *f,
				     struct sk_buff *skb,
				     unsigned int num)
{
	return smp_processor_id() % num;
D
David S. Miller 已提交
1160 1161
}

1162 1163 1164 1165
static unsigned int fanout_demux_rollover(struct packet_fanout *f,
					  struct sk_buff *skb,
					  unsigned int idx, unsigned int skip,
					  unsigned int num)
1166
{
1167
	unsigned int i, j;
1168

1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185
	i = j = min_t(int, f->next[idx], num - 1);
	do {
		if (i != skip && packet_rcv_has_room(pkt_sk(f->arr[i]), skb)) {
			if (i != j)
				f->next[idx] = i;
			return i;
		}
		if (++i == num)
			i = 0;
	} while (i != j);

	return idx;
}

static bool fanout_has_flag(struct packet_fanout *f, u16 flag)
{
	return f->flags & (flag >> 8);
1186 1187 1188 1189
}

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
1190 1191 1192 1193
{
	struct packet_fanout *f = pt->af_packet_priv;
	unsigned int num = f->num_members;
	struct packet_sock *po;
1194
	unsigned int idx;
D
David S. Miller 已提交
1195 1196 1197 1198 1199 1200 1201

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

1202 1203 1204
	switch (f->type) {
	case PACKET_FANOUT_HASH:
	default:
1205
		if (fanout_has_flag(f, PACKET_FANOUT_FLAG_DEFRAG)) {
1206
			skb = ip_check_defrag(skb, IP_DEFRAG_AF_PACKET);
1207 1208 1209 1210
			if (!skb)
				return 0;
		}
		skb_get_rxhash(skb);
1211
		idx = fanout_demux_hash(f, skb, num);
1212 1213
		break;
	case PACKET_FANOUT_LB:
1214
		idx = fanout_demux_lb(f, skb, num);
1215 1216
		break;
	case PACKET_FANOUT_CPU:
1217 1218 1219 1220
		idx = fanout_demux_cpu(f, skb, num);
		break;
	case PACKET_FANOUT_ROLLOVER:
		idx = fanout_demux_rollover(f, skb, 0, (unsigned int) -1, num);
1221
		break;
D
David S. Miller 已提交
1222 1223
	}

1224 1225 1226 1227 1228 1229
	po = pkt_sk(f->arr[idx]);
	if (fanout_has_flag(f, PACKET_FANOUT_FLAG_ROLLOVER) &&
	    unlikely(!packet_rcv_has_room(po, skb))) {
		idx = fanout_demux_rollover(f, skb, idx, idx, num);
		po = pkt_sk(f->arr[idx]);
	}
D
David S. Miller 已提交
1230 1231 1232 1233

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

1234 1235
DEFINE_MUTEX(fanout_mutex);
EXPORT_SYMBOL_GPL(fanout_mutex);
D
David S. Miller 已提交
1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264
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);
}

1265
static bool match_fanout_group(struct packet_type *ptype, struct sock * sk)
1266 1267 1268 1269 1270 1271 1272
{
	if (ptype->af_packet_priv == (void*)((struct packet_sock *)sk)->fanout)
		return true;

	return false;
}

1273
static int fanout_add(struct sock *sk, u16 id, u16 type_flags)
D
David S. Miller 已提交
1274 1275 1276
{
	struct packet_sock *po = pkt_sk(sk);
	struct packet_fanout *f, *match;
1277
	u8 type = type_flags & 0xff;
1278
	u8 flags = type_flags >> 8;
D
David S. Miller 已提交
1279 1280 1281
	int err;

	switch (type) {
1282 1283 1284
	case PACKET_FANOUT_ROLLOVER:
		if (type_flags & PACKET_FANOUT_FLAG_ROLLOVER)
			return -EINVAL;
D
David S. Miller 已提交
1285 1286
	case PACKET_FANOUT_HASH:
	case PACKET_FANOUT_LB:
1287
	case PACKET_FANOUT_CPU:
D
David S. Miller 已提交
1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307
		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
1308
	err = -EINVAL;
1309
	if (match && match->flags != flags)
E
Eric Dumazet 已提交
1310
		goto out;
D
David S. Miller 已提交
1311
	if (!match) {
E
Eric Dumazet 已提交
1312
		err = -ENOMEM;
D
David S. Miller 已提交
1313
		match = kzalloc(sizeof(*match), GFP_KERNEL);
E
Eric Dumazet 已提交
1314 1315 1316 1317 1318
		if (!match)
			goto out;
		write_pnet(&match->net, sock_net(sk));
		match->id = id;
		match->type = type;
1319
		match->flags = flags;
E
Eric Dumazet 已提交
1320 1321 1322 1323 1324 1325 1326 1327
		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;
1328
		match->prot_hook.id_match = match_fanout_group;
E
Eric Dumazet 已提交
1329 1330
		dev_add_pack(&match->prot_hook);
		list_add(&match->list, &fanout_list);
D
David S. Miller 已提交
1331
	}
E
Eric Dumazet 已提交
1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342
	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
1343 1344
		}
	}
E
Eric Dumazet 已提交
1345
out:
D
David S. Miller 已提交
1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358
	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;

1359
	mutex_lock(&fanout_mutex);
D
David S. Miller 已提交
1360 1361 1362 1363 1364 1365 1366 1367 1368
	po->fanout = NULL;

	if (atomic_dec_and_test(&f->sk_ref)) {
		list_del(&f->list);
		dev_remove_pack(&f->prot_hook);
		kfree(f);
	}
	mutex_unlock(&fanout_mutex);
}
L
Linus Torvalds 已提交
1369

1370
static const struct proto_ops packet_ops;
L
Linus Torvalds 已提交
1371

1372
static const struct proto_ops packet_ops_spkt;
L
Linus Torvalds 已提交
1373

E
Eric Dumazet 已提交
1374 1375
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
1376 1377 1378 1379 1380 1381 1382 1383 1384 1385
{
	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;
1386

L
Linus Torvalds 已提交
1387 1388 1389 1390 1391 1392 1393
	/*
	 *	Yank back the headers [hope the device set this
	 *	right or kerboom...]
	 *
	 *	Incoming packets have ll header pulled,
	 *	push it back.
	 *
1394
	 *	For outgoing ones skb->data == skb_mac_header(skb)
L
Linus Torvalds 已提交
1395 1396 1397 1398 1399 1400
	 *	so that this procedure is noop.
	 */

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

O
Octavian Purdila 已提交
1401
	if (!net_eq(dev_net(dev), sock_net(sk)))
1402 1403
		goto out;

E
Eric Dumazet 已提交
1404 1405
	skb = skb_share_check(skb, GFP_ATOMIC);
	if (skb == NULL)
L
Linus Torvalds 已提交
1406 1407 1408
		goto oom;

	/* drop any routing info */
E
Eric Dumazet 已提交
1409
	skb_dst_drop(skb);
L
Linus Torvalds 已提交
1410

1411 1412 1413
	/* drop conntrack reference */
	nf_reset(skb);

1414
	spkt = &PACKET_SKB_CB(skb)->sa.pkt;
L
Linus Torvalds 已提交
1415

1416
	skb_push(skb, skb->data - skb_mac_header(skb));
L
Linus Torvalds 已提交
1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430

	/*
	 *	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
1431
	if (sock_queue_rcv_skb(sk, skb) == 0)
L
Linus Torvalds 已提交
1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444
		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
 */
1445

L
Linus Torvalds 已提交
1446 1447 1448 1449
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
1450
	struct sockaddr_pkt *saddr = (struct sockaddr_pkt *)msg->msg_name;
1451
	struct sk_buff *skb = NULL;
L
Linus Torvalds 已提交
1452
	struct net_device *dev;
E
Eric Dumazet 已提交
1453
	__be16 proto = 0;
L
Linus Torvalds 已提交
1454
	int err;
1455
	int extra_len = 0;
1456

L
Linus Torvalds 已提交
1457
	/*
1458
	 *	Get and verify the address.
L
Linus Torvalds 已提交
1459 1460
	 */

E
Eric Dumazet 已提交
1461
	if (saddr) {
L
Linus Torvalds 已提交
1462
		if (msg->msg_namelen < sizeof(struct sockaddr))
E
Eric Dumazet 已提交
1463 1464 1465 1466 1467
			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
1468 1469

	/*
1470
	 *	Find the device first to size check it
L
Linus Torvalds 已提交
1471 1472
	 */

1473
	saddr->spkt_device[sizeof(saddr->spkt_device) - 1] = 0;
1474
retry:
E
Eric Dumazet 已提交
1475 1476
	rcu_read_lock();
	dev = dev_get_by_name_rcu(sock_net(sk), saddr->spkt_device);
L
Linus Torvalds 已提交
1477 1478 1479
	err = -ENODEV;
	if (dev == NULL)
		goto out_unlock;
1480

1481 1482 1483 1484
	err = -ENETDOWN;
	if (!(dev->flags & IFF_UP))
		goto out_unlock;

L
Linus Torvalds 已提交
1485
	/*
E
Eric Dumazet 已提交
1486 1487
	 * 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
1488
	 */
1489

1490 1491 1492 1493 1494 1495 1496 1497
	if (unlikely(sock_flag(sk, SOCK_NOFCS))) {
		if (!netif_supports_nofcs(dev)) {
			err = -EPROTONOSUPPORT;
			goto out_unlock;
		}
		extra_len = 4; /* We're doing our own CRC */
	}

L
Linus Torvalds 已提交
1498
	err = -EMSGSIZE;
1499
	if (len > dev->mtu + dev->hard_header_len + VLAN_HLEN + extra_len)
L
Linus Torvalds 已提交
1500 1501
		goto out_unlock;

1502 1503
	if (!skb) {
		size_t reserved = LL_RESERVED_SPACE(dev);
1504
		int tlen = dev->needed_tailroom;
1505 1506 1507
		unsigned int hhlen = dev->header_ops ? dev->hard_header_len : 0;

		rcu_read_unlock();
1508
		skb = sock_wmalloc(sk, len + reserved + tlen, 0, GFP_KERNEL);
1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528
		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
1529 1530
	}

1531
	if (len > (dev->mtu + dev->hard_header_len + extra_len)) {
1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543
		/* 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;
		}
	}
1544

L
Linus Torvalds 已提交
1545 1546 1547
	skb->protocol = proto;
	skb->dev = dev;
	skb->priority = sk->sk_priority;
1548
	skb->mark = sk->sk_mark;
1549 1550

	sock_tx_timestamp(sk, &skb_shinfo(skb)->tx_flags);
L
Linus Torvalds 已提交
1551

1552 1553 1554
	if (unlikely(extra_len == 4))
		skb->no_fcs = 1;

1555
	skb_probe_transport_header(skb, 0);
1556

L
Linus Torvalds 已提交
1557
	dev_queue_xmit(skb);
E
Eric Dumazet 已提交
1558
	rcu_read_unlock();
E
Eric Dumazet 已提交
1559
	return len;
L
Linus Torvalds 已提交
1560 1561

out_unlock:
E
Eric Dumazet 已提交
1562
	rcu_read_unlock();
1563 1564
out_free:
	kfree_skb(skb);
L
Linus Torvalds 已提交
1565 1566 1567
	return err;
}

1568
static unsigned int run_filter(const struct sk_buff *skb,
E
Eric Dumazet 已提交
1569
				      const struct sock *sk,
D
David S. Miller 已提交
1570
				      unsigned int res)
L
Linus Torvalds 已提交
1571 1572
{
	struct sk_filter *filter;
1573

1574 1575
	rcu_read_lock();
	filter = rcu_dereference(sk->sk_filter);
D
David S. Miller 已提交
1576
	if (filter != NULL)
1577
		res = SK_RUN_FILTER(filter, skb);
1578
	rcu_read_unlock();
L
Linus Torvalds 已提交
1579

D
David S. Miller 已提交
1580
	return res;
L
Linus Torvalds 已提交
1581 1582 1583
}

/*
E
Eric Dumazet 已提交
1584 1585 1586 1587 1588 1589 1590 1591 1592
 * 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
1593 1594
 */

E
Eric Dumazet 已提交
1595 1596
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
1597 1598 1599 1600
{
	struct sock *sk;
	struct sockaddr_ll *sll;
	struct packet_sock *po;
E
Eric Dumazet 已提交
1601
	u8 *skb_head = skb->data;
L
Linus Torvalds 已提交
1602
	int skb_len = skb->len;
D
David S. Miller 已提交
1603
	unsigned int snaplen, res;
L
Linus Torvalds 已提交
1604 1605 1606 1607 1608 1609 1610

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

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

O
Octavian Purdila 已提交
1611
	if (!net_eq(dev_net(dev), sock_net(sk)))
1612 1613
		goto drop;

L
Linus Torvalds 已提交
1614 1615
	skb->dev = dev;

1616
	if (dev->header_ops) {
L
Linus Torvalds 已提交
1617
		/* The device has an explicit notion of ll header,
E
Eric Dumazet 已提交
1618 1619 1620 1621 1622
		 * 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
1623 1624
		 */
		if (sk->sk_type != SOCK_DGRAM)
1625
			skb_push(skb, skb->data - skb_mac_header(skb));
L
Linus Torvalds 已提交
1626 1627
		else if (skb->pkt_type == PACKET_OUTGOING) {
			/* Special case: outgoing packets have ll header at head */
1628
			skb_pull(skb, skb_network_offset(skb));
L
Linus Torvalds 已提交
1629 1630 1631 1632 1633
		}
	}

	snaplen = skb->len;

D
David S. Miller 已提交
1634 1635
	res = run_filter(skb, sk, snaplen);
	if (!res)
1636
		goto drop_n_restore;
D
David S. Miller 已提交
1637 1638
	if (snaplen > res)
		snaplen = res;
L
Linus Torvalds 已提交
1639

E
Eric Dumazet 已提交
1640
	if (atomic_read(&sk->sk_rmem_alloc) >= sk->sk_rcvbuf)
L
Linus Torvalds 已提交
1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651
		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;
		}
1652
		consume_skb(skb);
L
Linus Torvalds 已提交
1653 1654 1655
		skb = nskb;
	}

1656 1657 1658 1659
	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
1660 1661 1662 1663
	sll->sll_family = AF_PACKET;
	sll->sll_hatype = dev->type;
	sll->sll_protocol = skb->protocol;
	sll->sll_pkttype = skb->pkt_type;
1664
	if (unlikely(po->origdev))
1665 1666 1667
		sll->sll_ifindex = orig_dev->ifindex;
	else
		sll->sll_ifindex = dev->ifindex;
L
Linus Torvalds 已提交
1668

S
Stephen Hemminger 已提交
1669
	sll->sll_halen = dev_parse_header(skb, sll->sll_addr);
L
Linus Torvalds 已提交
1670

1671
	PACKET_SKB_CB(skb)->origlen = skb->len;
1672

L
Linus Torvalds 已提交
1673 1674 1675 1676 1677
	if (pskb_trim(skb, snaplen))
		goto drop_n_acct;

	skb_set_owner_r(skb, sk);
	skb->dev = NULL;
E
Eric Dumazet 已提交
1678
	skb_dst_drop(skb);
L
Linus Torvalds 已提交
1679

1680 1681 1682
	/* drop conntrack reference */
	nf_reset(skb);

L
Linus Torvalds 已提交
1683
	spin_lock(&sk->sk_receive_queue.lock);
1684
	po->stats.stats1.tp_packets++;
1685
	skb->dropcount = atomic_read(&sk->sk_drops);
L
Linus Torvalds 已提交
1686 1687 1688 1689 1690 1691
	__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:
1692
	spin_lock(&sk->sk_receive_queue.lock);
1693
	po->stats.stats1.tp_drops++;
1694 1695
	atomic_inc(&sk->sk_drops);
	spin_unlock(&sk->sk_receive_queue.lock);
L
Linus Torvalds 已提交
1696 1697 1698 1699 1700 1701 1702

drop_n_restore:
	if (skb_head != skb->data && skb_shared(skb)) {
		skb->data = skb_head;
		skb->len = skb_len;
	}
drop:
1703
	consume_skb(skb);
L
Linus Torvalds 已提交
1704 1705 1706
	return 0;
}

E
Eric Dumazet 已提交
1707 1708
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
1709 1710 1711 1712
{
	struct sock *sk;
	struct packet_sock *po;
	struct sockaddr_ll *sll;
1713
	union tpacket_uhdr h;
E
Eric Dumazet 已提交
1714
	u8 *skb_head = skb->data;
L
Linus Torvalds 已提交
1715
	int skb_len = skb->len;
D
David S. Miller 已提交
1716
	unsigned int snaplen, res;
1717
	unsigned long status = TP_STATUS_USER;
1718
	unsigned short macoff, netoff, hdrlen;
L
Linus Torvalds 已提交
1719
	struct sk_buff *copy_skb = NULL;
1720
	struct timespec ts;
1721
	__u32 ts_status;
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 int 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
	if (po->tp_version <= TPACKET_V2) {
		if (macoff + snaplen > po->rx_ring.frame_size) {
			if (po->copy_thresh &&
E
Eric Dumazet 已提交
1765
			    atomic_read(&sk->sk_rmem_alloc) < sk->sk_rcvbuf) {
1766 1767 1768 1769 1770 1771 1772 1773
				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
1774
			}
1775 1776 1777
			snaplen = po->rx_ring.frame_size - macoff;
			if ((int)snaplen < 0)
				snaplen = 0;
L
Linus Torvalds 已提交
1778 1779 1780
		}
	}
	spin_lock(&sk->sk_receive_queue.lock);
1781 1782
	h.raw = packet_current_rx_frame(po, skb,
					TP_STATUS_KERNEL, (macoff+snaplen));
1783
	if (!h.raw)
L
Linus Torvalds 已提交
1784
		goto ring_is_full;
1785 1786 1787 1788 1789 1790 1791 1792
	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.
	 */
1793
		if (po->stats.stats1.tp_drops)
1794 1795
			status |= TP_STATUS_LOSING;
	}
1796
	po->stats.stats1.tp_packets++;
L
Linus Torvalds 已提交
1797 1798 1799 1800 1801 1802
	if (copy_skb) {
		status |= TP_STATUS_COPY;
		__skb_queue_tail(&sk->sk_receive_queue, copy_skb);
	}
	spin_unlock(&sk->sk_receive_queue.lock);

1803
	skb_copy_bits(skb, 0, h.raw + macoff, snaplen);
1804 1805

	if (!(ts_status = tpacket_get_timestamp(skb, &ts, po->tp_tstamp)))
1806
		getnstimeofday(&ts);
L
Linus Torvalds 已提交
1807

1808 1809
	status |= ts_status;

1810 1811 1812 1813 1814 1815
	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;
1816 1817
		h.h1->tp_sec = ts.tv_sec;
		h.h1->tp_usec = ts.tv_nsec / NSEC_PER_USEC;
1818 1819 1820 1821 1822 1823 1824 1825 1826
		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;
		h.h2->tp_sec = ts.tv_sec;
		h.h2->tp_nsec = ts.tv_nsec;
1827 1828 1829 1830 1831 1832
		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;
		}
1833
		h.h2->tp_padding = 0;
1834 1835
		hdrlen = sizeof(*h.h2);
		break;
1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848
	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;
		h.h3->tp_sec  = ts.tv_sec;
		h.h3->tp_nsec = ts.tv_nsec;
		hdrlen = sizeof(*h.h3);
		break;
1849 1850 1851
	default:
		BUG();
	}
L
Linus Torvalds 已提交
1852

1853
	sll = h.raw + TPACKET_ALIGN(hdrlen);
S
Stephen Hemminger 已提交
1854
	sll->sll_halen = dev_parse_header(skb, sll->sll_addr);
L
Linus Torvalds 已提交
1855 1856 1857 1858
	sll->sll_family = AF_PACKET;
	sll->sll_hatype = dev->type;
	sll->sll_protocol = skb->protocol;
	sll->sll_pkttype = skb->pkt_type;
1859
	if (unlikely(po->origdev))
1860 1861 1862
		sll->sll_ifindex = orig_dev->ifindex;
	else
		sll->sll_ifindex = dev->ifindex;
L
Linus Torvalds 已提交
1863

R
Ralf Baechle 已提交
1864
	smp_mb();
1865
#if ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE == 1
L
Linus Torvalds 已提交
1866
	{
1867 1868
		u8 *start, *end;

1869 1870 1871 1872 1873 1874
		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));
		}
1875
		smp_wmb();
L
Linus Torvalds 已提交
1876
	}
1877
#endif
1878 1879 1880 1881
	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
1882 1883 1884 1885 1886 1887 1888 1889 1890

	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:
1891
	kfree_skb(skb);
L
Linus Torvalds 已提交
1892 1893 1894
	return 0;

ring_is_full:
1895
	po->stats.stats1.tp_drops++;
L
Linus Torvalds 已提交
1896 1897 1898
	spin_unlock(&sk->sk_receive_queue.lock);

	sk->sk_data_ready(sk, 0);
1899
	kfree_skb(copy_skb);
L
Linus Torvalds 已提交
1900 1901 1902
	goto drop_n_restore;
}

J
Johann Baudy 已提交
1903 1904 1905
static void tpacket_destruct_skb(struct sk_buff *skb)
{
	struct packet_sock *po = pkt_sk(skb->sk);
E
Eric Dumazet 已提交
1906
	void *ph;
L
Linus Torvalds 已提交
1907

J
Johann Baudy 已提交
1908
	if (likely(po->tx_ring.pg_vec)) {
1909 1910
		__u32 ts;

J
Johann Baudy 已提交
1911 1912 1913
		ph = skb_shinfo(skb)->destructor_arg;
		BUG_ON(atomic_read(&po->tx_ring.pending) == 0);
		atomic_dec(&po->tx_ring.pending);
1914 1915 1916

		ts = __packet_set_timestamp(po, ph, skb);
		__packet_set_status(po, ph, TP_STATUS_AVAILABLE | ts);
J
Johann Baudy 已提交
1917 1918 1919 1920 1921
	}

	sock_wfree(skb);
}

E
Eric Dumazet 已提交
1922 1923
static int tpacket_fill_skb(struct packet_sock *po, struct sk_buff *skb,
		void *frame, struct net_device *dev, int size_max,
1924
		__be16 proto, unsigned char *addr, int hlen)
J
Johann Baudy 已提交
1925
{
1926
	union tpacket_uhdr ph;
1927
	int to_write, offset, len, tp_len, nr_frags, len_max, max_frame_len;
J
Johann Baudy 已提交
1928 1929 1930 1931 1932 1933 1934 1935 1936 1937
	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;
1938
	skb->mark = po->sk.sk_mark;
1939
	sock_tx_timestamp(&po->sk, &skb_shinfo(skb)->tx_flags);
J
Johann Baudy 已提交
1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950
	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;
	}

1951
	skb_reserve(skb, hlen);
J
Johann Baudy 已提交
1952
	skb_reset_network_header(skb);
1953
	skb_probe_transport_header(skb, 0);
1954

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

2005 2006 2007
		if (dev->type == ARPHRD_ETHER)
			skb->protocol = eth_type_trans(skb, dev);

J
Johann Baudy 已提交
2008 2009 2010 2011
		data += dev->hard_header_len;
		to_write -= dev->hard_header_len;
	}

2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023
	max_frame_len = dev->mtu + dev->hard_header_len;
	if (skb->protocol == htons(ETH_P_8021Q))
		max_frame_len += VLAN_HLEN;

	if (size_max > max_frame_len)
		size_max = max_frame_len;

	if (unlikely(tp_len > size_max)) {
		pr_err("packet size is too long (%d > %d)\n", tp_len, size_max);
		return -EMSGSIZE;
	}

J
Johann Baudy 已提交
2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036
	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
2037 2038
			pr_err("Packet exceed the number of skb frags(%lu)\n",
			       MAX_SKB_FRAGS);
J
Johann Baudy 已提交
2039 2040 2041
			return -EFAULT;
		}

2042 2043
		page = pgv_to_page(data);
		data += len;
J
Johann Baudy 已提交
2044 2045
		flush_dcache_page(page);
		get_page(page);
2046
		skb_fill_page_desc(skb, nr_frags, page, offset, len);
J
Johann Baudy 已提交
2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060
		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;
2061
	bool need_rls_dev = false;
2062
	int err;
E
Eric Dumazet 已提交
2063 2064
	void *ph;
	struct sockaddr_ll *saddr = (struct sockaddr_ll *)msg->msg_name;
J
Johann Baudy 已提交
2065 2066 2067
	int tp_len, size_max;
	unsigned char *addr;
	int len_sum = 0;
2068
	int status = TP_STATUS_AVAILABLE;
2069
	int hlen, tlen;
J
Johann Baudy 已提交
2070 2071 2072 2073

	mutex_lock(&po->pg_vec_lock);

	if (saddr == NULL) {
2074
		dev = po->prot_hook.dev;
J
Johann Baudy 已提交
2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086
		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;
2087 2088
		dev = dev_get_by_index(sock_net(&po->sk), saddr->sll_ifindex);
		need_rls_dev = true;
J
Johann Baudy 已提交
2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099
	}

	err = -ENXIO;
	if (unlikely(dev == NULL))
		goto out;

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

	size_max = po->tx_ring.frame_size
2100
		- (po->tp_hdrlen - sizeof(struct sockaddr_ll));
J
Johann Baudy 已提交
2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111

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

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

		status = TP_STATUS_SEND_REQUEST;
2112 2113
		hlen = LL_RESERVED_SPACE(dev);
		tlen = dev->needed_tailroom;
J
Johann Baudy 已提交
2114
		skb = sock_alloc_send_skb(&po->sk,
2115
				hlen + tlen + sizeof(struct sockaddr_ll),
J
Johann Baudy 已提交
2116 2117 2118 2119 2120 2121
				0, &err);

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

		tp_len = tpacket_fill_skb(po, skb, ph, dev, size_max, proto,
2122
				addr, hlen);
J
Johann Baudy 已提交
2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143

		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);
2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157
		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
2158 2159
		packet_increment_head(&po->tx_ring);
		len_sum += tp_len;
2160 2161 2162 2163
	} while (likely((ph != NULL) ||
			((!(msg->msg_flags & MSG_DONTWAIT)) &&
			 (atomic_read(&po->tx_ring.pending))))
		);
J
Johann Baudy 已提交
2164 2165 2166 2167 2168 2169 2170 2171

	err = len_sum;
	goto out_put;

out_status:
	__packet_set_status(po, ph, status);
	kfree_skb(skb);
out_put:
2172 2173
	if (need_rls_dev)
		dev_put(dev);
J
Johann Baudy 已提交
2174 2175 2176 2177 2178
out:
	mutex_unlock(&po->pg_vec_lock);
	return err;
}

2179 2180 2181 2182
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)
2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202
{
	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
2203
static int packet_snd(struct socket *sock,
L
Linus Torvalds 已提交
2204 2205 2206
			  struct msghdr *msg, size_t len)
{
	struct sock *sk = sock->sk;
E
Eric Dumazet 已提交
2207
	struct sockaddr_ll *saddr = (struct sockaddr_ll *)msg->msg_name;
L
Linus Torvalds 已提交
2208 2209
	struct sk_buff *skb;
	struct net_device *dev;
A
Al Viro 已提交
2210
	__be16 proto;
2211
	bool need_rls_dev = false;
L
Linus Torvalds 已提交
2212
	unsigned char *addr;
2213
	int err, reserve = 0;
2214 2215 2216 2217 2218
	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;
2219
	int hlen, tlen;
2220
	int extra_len = 0;
L
Linus Torvalds 已提交
2221 2222

	/*
2223
	 *	Get and verify the address.
L
Linus Torvalds 已提交
2224
	 */
2225

L
Linus Torvalds 已提交
2226
	if (saddr == NULL) {
2227
		dev = po->prot_hook.dev;
L
Linus Torvalds 已提交
2228 2229 2230 2231 2232 2233
		proto	= po->num;
		addr	= NULL;
	} else {
		err = -EINVAL;
		if (msg->msg_namelen < sizeof(struct sockaddr_ll))
			goto out;
2234 2235
		if (msg->msg_namelen < (saddr->sll_halen + offsetof(struct sockaddr_ll, sll_addr)))
			goto out;
L
Linus Torvalds 已提交
2236 2237
		proto	= saddr->sll_protocol;
		addr	= saddr->sll_addr;
2238 2239
		dev = dev_get_by_index(sock_net(sk), saddr->sll_ifindex);
		need_rls_dev = true;
L
Linus Torvalds 已提交
2240 2241 2242 2243 2244 2245 2246 2247
	}

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

2248 2249 2250 2251
	err = -ENETDOWN;
	if (!(dev->flags & IFF_UP))
		goto out_unlock;

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 2292 2293 2294 2295 2296 2297 2298 2299
	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;

		}
	}

2300 2301 2302 2303 2304 2305 2306 2307
	if (unlikely(sock_flag(sk, SOCK_NOFCS))) {
		if (!netif_supports_nofcs(dev)) {
			err = -EPROTONOSUPPORT;
			goto out_unlock;
		}
		extra_len = 4; /* We're doing our own CRC */
	}

L
Linus Torvalds 已提交
2308
	err = -EMSGSIZE;
2309
	if (!gso_type && (len > dev->mtu + reserve + VLAN_HLEN + extra_len))
L
Linus Torvalds 已提交
2310 2311
		goto out_unlock;

2312
	err = -ENOBUFS;
2313 2314 2315
	hlen = LL_RESERVED_SPACE(dev);
	tlen = dev->needed_tailroom;
	skb = packet_alloc_skb(sk, hlen + tlen, hlen, len, vnet_hdr.hdr_len,
2316
			       msg->msg_flags & MSG_DONTWAIT, &err);
E
Eric Dumazet 已提交
2317
	if (skb == NULL)
L
Linus Torvalds 已提交
2318 2319
		goto out_unlock;

2320
	skb_set_network_header(skb, reserve);
L
Linus Torvalds 已提交
2321

2322 2323
	err = -EINVAL;
	if (sock->type == SOCK_DGRAM &&
2324
	    (offset = dev_hard_header(skb, dev, ntohs(proto), addr, NULL, len)) < 0)
2325
		goto out_free;
L
Linus Torvalds 已提交
2326 2327

	/* Returns -EFAULT on error */
2328
	err = skb_copy_datagram_from_iovec(skb, offset, msg->msg_iov, 0, len);
L
Linus Torvalds 已提交
2329 2330
	if (err)
		goto out_free;
2331 2332

	sock_tx_timestamp(sk, &skb_shinfo(skb)->tx_flags);
L
Linus Torvalds 已提交
2333

2334 2335
	if (dev->type == ARPHRD_ETHER) {
		skb->protocol = eth_type_trans(skb, dev);
2336 2337
		if (skb->protocol == htons(ETH_P_8021Q))
			reserve += VLAN_HLEN;
2338 2339 2340 2341 2342
	} else {
		skb->protocol = proto;
		skb->dev = dev;
	}

2343
	if (!gso_type && (len > dev->mtu + reserve + extra_len)) {
2344 2345
		err = -EMSGSIZE;
		goto out_free;
2346 2347
	}

L
Linus Torvalds 已提交
2348
	skb->priority = sk->sk_priority;
2349
	skb->mark = sk->sk_mark;
L
Linus Torvalds 已提交
2350

2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369
	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;
	}

2370
	skb_probe_transport_header(skb, reserve);
2371

2372 2373 2374
	if (unlikely(extra_len == 4))
		skb->no_fcs = 1;

L
Linus Torvalds 已提交
2375 2376 2377 2378 2379 2380 2381 2382
	/*
	 *	Now send it
	 */

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

2383 2384
	if (need_rls_dev)
		dev_put(dev);
L
Linus Torvalds 已提交
2385

E
Eric Dumazet 已提交
2386
	return len;
L
Linus Torvalds 已提交
2387 2388 2389 2390

out_free:
	kfree_skb(skb);
out_unlock:
2391
	if (dev && need_rls_dev)
L
Linus Torvalds 已提交
2392 2393 2394 2395 2396
		dev_put(dev);
out:
	return err;
}

J
Johann Baudy 已提交
2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407
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
2408 2409 2410 2411 2412 2413 2414 2415 2416
/*
 *	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;
2417
	struct net *net;
2418
	union tpacket_req_u req_u;
L
Linus Torvalds 已提交
2419 2420 2421 2422

	if (!sk)
		return 0;

2423
	net = sock_net(sk);
L
Linus Torvalds 已提交
2424 2425
	po = pkt_sk(sk);

2426
	mutex_lock(&net->packet.sklist_lock);
2427
	sk_del_node_init_rcu(sk);
2428 2429 2430
	mutex_unlock(&net->packet.sklist_lock);

	preempt_disable();
2431
	sock_prot_inuse_add(net, sk->sk_prot, -1);
2432
	preempt_enable();
L
Linus Torvalds 已提交
2433

2434
	spin_lock(&po->bind_lock);
2435
	unregister_prot_hook(sk, false);
2436 2437 2438 2439
	if (po->prot_hook.dev) {
		dev_put(po->prot_hook.dev);
		po->prot_hook.dev = NULL;
	}
2440
	spin_unlock(&po->bind_lock);
L
Linus Torvalds 已提交
2441 2442 2443

	packet_flush_mclist(sk);

2444 2445
	if (po->rx_ring.pg_vec) {
		memset(&req_u, 0, sizeof(req_u));
2446
		packet_set_ring(sk, &req_u, 1, 0);
2447
	}
J
Johann Baudy 已提交
2448

2449 2450
	if (po->tx_ring.pg_vec) {
		memset(&req_u, 0, sizeof(req_u));
2451
		packet_set_ring(sk, &req_u, 1, 1);
2452
	}
L
Linus Torvalds 已提交
2453

D
David S. Miller 已提交
2454 2455
	fanout_release(sk);

2456
	synchronize_net();
L
Linus Torvalds 已提交
2457 2458 2459 2460 2461 2462 2463 2464 2465
	/*
	 *	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);
2466
	sk_refcnt_debug_release(sk);
L
Linus Torvalds 已提交
2467 2468 2469 2470 2471 2472 2473 2474 2475

	sock_put(sk);
	return 0;
}

/*
 *	Attach a packet hook.
 */

A
Al Viro 已提交
2476
static int packet_do_bind(struct sock *sk, struct net_device *dev, __be16 protocol)
L
Linus Torvalds 已提交
2477 2478
{
	struct packet_sock *po = pkt_sk(sk);
D
David S. Miller 已提交
2479

2480 2481 2482 2483
	if (po->fanout) {
		if (dev)
			dev_put(dev);

D
David S. Miller 已提交
2484
		return -EINVAL;
2485
	}
L
Linus Torvalds 已提交
2486 2487 2488 2489

	lock_sock(sk);

	spin_lock(&po->bind_lock);
2490
	unregister_prot_hook(sk, true);
L
Linus Torvalds 已提交
2491 2492
	po->num = protocol;
	po->prot_hook.type = protocol;
2493 2494
	if (po->prot_hook.dev)
		dev_put(po->prot_hook.dev);
L
Linus Torvalds 已提交
2495 2496 2497 2498 2499 2500 2501
	po->prot_hook.dev = dev;

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

	if (protocol == 0)
		goto out_unlock;

2502
	if (!dev || (dev->flags & IFF_UP)) {
2503
		register_prot_hook(sk);
2504 2505 2506 2507
	} else {
		sk->sk_err = ENETDOWN;
		if (!sock_flag(sk, SOCK_DEAD))
			sk->sk_error_report(sk);
L
Linus Torvalds 已提交
2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519
	}

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
2520 2521
static int packet_bind_spkt(struct socket *sock, struct sockaddr *uaddr,
			    int addr_len)
L
Linus Torvalds 已提交
2522
{
E
Eric Dumazet 已提交
2523
	struct sock *sk = sock->sk;
L
Linus Torvalds 已提交
2524 2525 2526
	char name[15];
	struct net_device *dev;
	int err = -ENODEV;
2527

L
Linus Torvalds 已提交
2528 2529 2530
	/*
	 *	Check legality
	 */
2531

2532
	if (addr_len != sizeof(struct sockaddr))
L
Linus Torvalds 已提交
2533
		return -EINVAL;
E
Eric Dumazet 已提交
2534
	strlcpy(name, uaddr->sa_data, sizeof(name));
L
Linus Torvalds 已提交
2535

2536
	dev = dev_get_by_name(sock_net(sk), name);
2537
	if (dev)
L
Linus Torvalds 已提交
2538 2539 2540 2541 2542 2543
		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
2544 2545
	struct sockaddr_ll *sll = (struct sockaddr_ll *)uaddr;
	struct sock *sk = sock->sk;
L
Linus Torvalds 已提交
2546 2547 2548 2549 2550 2551 2552
	struct net_device *dev = NULL;
	int err;


	/*
	 *	Check legality
	 */
2553

L
Linus Torvalds 已提交
2554 2555 2556 2557 2558 2559 2560
	if (addr_len < sizeof(struct sockaddr_ll))
		return -EINVAL;
	if (sll->sll_family != AF_PACKET)
		return -EINVAL;

	if (sll->sll_ifindex) {
		err = -ENODEV;
2561
		dev = dev_get_by_index(sock_net(sk), sll->sll_ifindex);
L
Linus Torvalds 已提交
2562 2563 2564 2565 2566 2567 2568 2569 2570 2571 2572 2573 2574 2575 2576 2577
		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),
};

/*
2578
 *	Create a packet of type SOCK_PACKET.
L
Linus Torvalds 已提交
2579 2580
 */

2581 2582
static int packet_create(struct net *net, struct socket *sock, int protocol,
			 int kern)
L
Linus Torvalds 已提交
2583 2584 2585
{
	struct sock *sk;
	struct packet_sock *po;
A
Al Viro 已提交
2586
	__be16 proto = (__force __be16)protocol; /* weird, but documented */
L
Linus Torvalds 已提交
2587 2588
	int err;

2589
	if (!ns_capable(net->user_ns, CAP_NET_RAW))
L
Linus Torvalds 已提交
2590
		return -EPERM;
2591 2592
	if (sock->type != SOCK_DGRAM && sock->type != SOCK_RAW &&
	    sock->type != SOCK_PACKET)
L
Linus Torvalds 已提交
2593 2594 2595 2596 2597
		return -ESOCKTNOSUPPORT;

	sock->state = SS_UNCONNECTED;

	err = -ENOBUFS;
2598
	sk = sk_alloc(net, PF_PACKET, GFP_KERNEL, &packet_proto);
L
Linus Torvalds 已提交
2599 2600 2601 2602 2603 2604
	if (sk == NULL)
		goto out;

	sock->ops = &packet_ops;
	if (sock->type == SOCK_PACKET)
		sock->ops = &packet_ops_spkt;
2605

L
Linus Torvalds 已提交
2606 2607 2608 2609
	sock_init_data(sock, sk);

	po = pkt_sk(sk);
	sk->sk_family = PF_PACKET;
A
Al Viro 已提交
2610
	po->num = proto;
L
Linus Torvalds 已提交
2611 2612

	sk->sk_destruct = packet_sock_destruct;
2613
	sk_refcnt_debug_inc(sk);
L
Linus Torvalds 已提交
2614 2615 2616 2617 2618 2619

	/*
	 *	Attach a protocol block
	 */

	spin_lock_init(&po->bind_lock);
2620
	mutex_init(&po->pg_vec_lock);
L
Linus Torvalds 已提交
2621
	po->prot_hook.func = packet_rcv;
2622

L
Linus Torvalds 已提交
2623 2624
	if (sock->type == SOCK_PACKET)
		po->prot_hook.func = packet_rcv_spkt;
2625

L
Linus Torvalds 已提交
2626 2627
	po->prot_hook.af_packet_priv = sk;

A
Al Viro 已提交
2628 2629
	if (proto) {
		po->prot_hook.type = proto;
2630
		register_prot_hook(sk);
L
Linus Torvalds 已提交
2631 2632
	}

2633
	mutex_lock(&net->packet.sklist_lock);
2634
	sk_add_node_rcu(sk, &net->packet.sklist);
2635 2636 2637
	mutex_unlock(&net->packet.sklist_lock);

	preempt_disable();
2638
	sock_prot_inuse_add(net, &packet_proto, 1);
2639
	preempt_enable();
2640

E
Eric Dumazet 已提交
2641
	return 0;
L
Linus Torvalds 已提交
2642 2643 2644 2645 2646 2647 2648 2649 2650 2651 2652 2653 2654 2655 2656
out:
	return err;
}

/*
 *	Pull a packet from our receive queue and hand it to the user.
 *	If necessary we block.
 */

static int packet_recvmsg(struct kiocb *iocb, struct socket *sock,
			  struct msghdr *msg, size_t len, int flags)
{
	struct sock *sk = sock->sk;
	struct sk_buff *skb;
	int copied, err;
2657
	struct sockaddr_ll *sll;
2658
	int vnet_hdr_len = 0;
L
Linus Torvalds 已提交
2659 2660

	err = -EINVAL;
2661
	if (flags & ~(MSG_PEEK|MSG_DONTWAIT|MSG_TRUNC|MSG_CMSG_COMPAT|MSG_ERRQUEUE))
L
Linus Torvalds 已提交
2662 2663 2664 2665 2666 2667 2668 2669
		goto out;

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

2670
	if (flags & MSG_ERRQUEUE) {
2671 2672
		err = sock_recv_errqueue(sk, msg, len,
					 SOL_PACKET, PACKET_TX_TIMESTAMP);
2673 2674 2675
		goto out;
	}

L
Linus Torvalds 已提交
2676 2677 2678 2679 2680 2681 2682 2683 2684
	/*
	 *	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
2685
	skb = skb_recv_datagram(sk, flags, flags & MSG_DONTWAIT, &err);
L
Linus Torvalds 已提交
2686 2687

	/*
2688
	 *	An error occurred so return it. Because skb_recv_datagram()
L
Linus Torvalds 已提交
2689 2690 2691 2692
	 *	handles the blocking we don't see and worry about blocking
	 *	retries.
	 */

2693
	if (skb == NULL)
L
Linus Torvalds 已提交
2694 2695
		goto out;

2696 2697 2698 2699 2700
	if (pkt_sk(sk)->has_vnet_hdr) {
		struct virtio_net_hdr vnet_hdr = { 0 };

		err = -EINVAL;
		vnet_hdr_len = sizeof(vnet_hdr);
2701
		if (len < vnet_hdr_len)
2702 2703
			goto out_free;

2704 2705
		len -= vnet_hdr_len;

2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728
		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;
2729
			vnet_hdr.csum_start = skb_checksum_start_offset(skb);
2730
			vnet_hdr.csum_offset = skb->csum_offset;
2731 2732
		} else if (skb->ip_summed == CHECKSUM_UNNECESSARY) {
			vnet_hdr.flags = VIRTIO_NET_HDR_F_DATA_VALID;
2733 2734 2735 2736 2737 2738 2739 2740
		} /* else everything is zero */

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

2741 2742 2743 2744 2745
	/*
	 *	If the address length field is there to be filled in, we fill
	 *	it in now.
	 */

2746
	sll = &PACKET_SKB_CB(skb)->sa.ll;
2747 2748 2749 2750 2751
	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
2752 2753 2754 2755 2756 2757
	/*
	 *	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
2758 2759 2760
	if (copied > len) {
		copied = len;
		msg->msg_flags |= MSG_TRUNC;
L
Linus Torvalds 已提交
2761 2762 2763 2764 2765 2766
	}

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

2767
	sock_recv_ts_and_drops(msg, sk, skb);
L
Linus Torvalds 已提交
2768 2769

	if (msg->msg_name)
2770 2771
		memcpy(msg->msg_name, &PACKET_SKB_CB(skb)->sa,
		       msg->msg_namelen);
L
Linus Torvalds 已提交
2772

2773
	if (pkt_sk(sk)->auxdata) {
2774 2775 2776 2777 2778 2779 2780 2781
		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;
2782
		aux.tp_net = skb_network_offset(skb);
2783 2784 2785 2786 2787 2788
		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;
		}
2789
		aux.tp_padding = 0;
2790
		put_cmsg(msg, SOL_PACKET, PACKET_AUXDATA, sizeof(aux), &aux);
2791 2792
	}

L
Linus Torvalds 已提交
2793 2794 2795 2796
	/*
	 *	Free or return the buffer as appropriate. Again this
	 *	hides all the races and re-entrancy issues from us.
	 */
2797
	err = vnet_hdr_len + ((flags&MSG_TRUNC) ? skb->len : copied);
L
Linus Torvalds 已提交
2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814

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;
2815
	memset(uaddr->sa_data, 0, sizeof(uaddr->sa_data));
E
Eric Dumazet 已提交
2816 2817 2818
	rcu_read_lock();
	dev = dev_get_by_index_rcu(sock_net(sk), pkt_sk(sk)->ifindex);
	if (dev)
2819
		strlcpy(uaddr->sa_data, dev->name, sizeof(uaddr->sa_data));
E
Eric Dumazet 已提交
2820
	rcu_read_unlock();
L
Linus Torvalds 已提交
2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831
	*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);
2832
	DECLARE_SOCKADDR(struct sockaddr_ll *, sll, uaddr);
L
Linus Torvalds 已提交
2833 2834 2835 2836 2837 2838 2839

	if (peer)
		return -EOPNOTSUPP;

	sll->sll_family = AF_PACKET;
	sll->sll_ifindex = po->ifindex;
	sll->sll_protocol = po->num;
2840
	sll->sll_pkttype = 0;
E
Eric Dumazet 已提交
2841 2842
	rcu_read_lock();
	dev = dev_get_by_index_rcu(sock_net(sk), po->ifindex);
L
Linus Torvalds 已提交
2843 2844 2845 2846 2847 2848 2849 2850
	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
2851
	rcu_read_unlock();
2852
	*uaddr_len = offsetof(struct sockaddr_ll, sll_addr) + sll->sll_halen;
L
Linus Torvalds 已提交
2853 2854 2855 2856

	return 0;
}

2857 2858
static int packet_dev_mc(struct net_device *dev, struct packet_mclist *i,
			 int what)
L
Linus Torvalds 已提交
2859 2860 2861
{
	switch (i->type) {
	case PACKET_MR_MULTICAST:
2862 2863
		if (i->alen != dev->addr_len)
			return -EINVAL;
L
Linus Torvalds 已提交
2864
		if (what > 0)
2865
			return dev_mc_add(dev, i->addr);
L
Linus Torvalds 已提交
2866
		else
2867
			return dev_mc_del(dev, i->addr);
L
Linus Torvalds 已提交
2868 2869
		break;
	case PACKET_MR_PROMISC:
2870
		return dev_set_promiscuity(dev, what);
L
Linus Torvalds 已提交
2871 2872
		break;
	case PACKET_MR_ALLMULTI:
2873
		return dev_set_allmulti(dev, what);
L
Linus Torvalds 已提交
2874
		break;
2875
	case PACKET_MR_UNICAST:
2876 2877
		if (i->alen != dev->addr_len)
			return -EINVAL;
2878
		if (what > 0)
2879
			return dev_uc_add(dev, i->addr);
2880
		else
2881
			return dev_uc_del(dev, i->addr);
2882
		break;
E
Eric Dumazet 已提交
2883 2884
	default:
		break;
L
Linus Torvalds 已提交
2885
	}
2886
	return 0;
L
Linus Torvalds 已提交
2887 2888 2889 2890
}

static void packet_dev_mclist(struct net_device *dev, struct packet_mclist *i, int what)
{
E
Eric Dumazet 已提交
2891
	for ( ; i; i = i->next) {
L
Linus Torvalds 已提交
2892 2893 2894 2895 2896
		if (i->ifindex == dev->ifindex)
			packet_dev_mc(dev, i, what);
	}
}

2897
static int packet_mc_add(struct sock *sk, struct packet_mreq_max *mreq)
L
Linus Torvalds 已提交
2898 2899 2900 2901 2902 2903 2904 2905 2906
{
	struct packet_sock *po = pkt_sk(sk);
	struct packet_mclist *ml, *i;
	struct net_device *dev;
	int err;

	rtnl_lock();

	err = -ENODEV;
2907
	dev = __dev_get_by_index(sock_net(sk), mreq->mr_ifindex);
L
Linus Torvalds 已提交
2908 2909 2910 2911
	if (!dev)
		goto done;

	err = -EINVAL;
2912
	if (mreq->mr_alen > dev->addr_len)
L
Linus Torvalds 已提交
2913 2914 2915
		goto done;

	err = -ENOBUFS;
2916
	i = kmalloc(sizeof(*i), GFP_KERNEL);
L
Linus Torvalds 已提交
2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937 2938 2939
	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;
2940 2941 2942 2943 2944
	err = packet_dev_mc(dev, i, 1);
	if (err) {
		po->mclist = i->next;
		kfree(i);
	}
L
Linus Torvalds 已提交
2945 2946 2947 2948 2949 2950

done:
	rtnl_unlock();
	return err;
}

2951
static int packet_mc_drop(struct sock *sk, struct packet_mreq_max *mreq)
L
Linus Torvalds 已提交
2952 2953 2954 2955 2956 2957 2958 2959 2960 2961 2962 2963 2964
{
	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;
2965 2966
				dev = __dev_get_by_index(sock_net(sk), ml->ifindex);
				if (dev)
L
Linus Torvalds 已提交
2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978 2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990
					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;
2991 2992
		dev = __dev_get_by_index(sock_net(sk), ml->ifindex);
		if (dev != NULL)
L
Linus Torvalds 已提交
2993 2994 2995 2996 2997 2998 2999
			packet_dev_mc(dev, ml, -1);
		kfree(ml);
	}
	rtnl_unlock();
}

static int
3000
packet_setsockopt(struct socket *sock, int level, int optname, char __user *optval, unsigned int optlen)
L
Linus Torvalds 已提交
3001 3002
{
	struct sock *sk = sock->sk;
3003
	struct packet_sock *po = pkt_sk(sk);
L
Linus Torvalds 已提交
3004 3005 3006 3007 3008
	int ret;

	if (level != SOL_PACKET)
		return -ENOPROTOOPT;

J
Johann Baudy 已提交
3009
	switch (optname) {
3010
	case PACKET_ADD_MEMBERSHIP:
L
Linus Torvalds 已提交
3011 3012
	case PACKET_DROP_MEMBERSHIP:
	{
3013 3014 3015 3016
		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
3017
			return -EINVAL;
3018 3019
		if (len > sizeof(mreq))
			len = sizeof(mreq);
E
Eric Dumazet 已提交
3020
		if (copy_from_user(&mreq, optval, len))
L
Linus Torvalds 已提交
3021
			return -EFAULT;
3022 3023
		if (len < (mreq.mr_alen + offsetof(struct packet_mreq, mr_address)))
			return -EINVAL;
L
Linus Torvalds 已提交
3024 3025 3026 3027 3028 3029
		if (optname == PACKET_ADD_MEMBERSHIP)
			ret = packet_mc_add(sk, &mreq);
		else
			ret = packet_mc_drop(sk, &mreq);
		return ret;
	}
3030

L
Linus Torvalds 已提交
3031
	case PACKET_RX_RING:
J
Johann Baudy 已提交
3032
	case PACKET_TX_RING:
L
Linus Torvalds 已提交
3033
	{
3034 3035
		union tpacket_req_u req_u;
		int len;
L
Linus Torvalds 已提交
3036

3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047
		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
3048
			return -EINVAL;
3049 3050
		if (pkt_sk(sk)->has_vnet_hdr)
			return -EINVAL;
3051
		if (copy_from_user(&req_u.req, optval, len))
L
Linus Torvalds 已提交
3052
			return -EFAULT;
3053 3054
		return packet_set_ring(sk, &req_u, 0,
			optname == PACKET_TX_RING);
L
Linus Torvalds 已提交
3055 3056 3057 3058 3059
	}
	case PACKET_COPY_THRESH:
	{
		int val;

E
Eric Dumazet 已提交
3060
		if (optlen != sizeof(val))
L
Linus Torvalds 已提交
3061
			return -EINVAL;
E
Eric Dumazet 已提交
3062
		if (copy_from_user(&val, optval, sizeof(val)))
L
Linus Torvalds 已提交
3063 3064 3065 3066 3067
			return -EFAULT;

		pkt_sk(sk)->copy_thresh = val;
		return 0;
	}
3068 3069 3070 3071 3072 3073
	case PACKET_VERSION:
	{
		int val;

		if (optlen != sizeof(val))
			return -EINVAL;
J
Johann Baudy 已提交
3074
		if (po->rx_ring.pg_vec || po->tx_ring.pg_vec)
3075 3076 3077 3078 3079 3080
			return -EBUSY;
		if (copy_from_user(&val, optval, sizeof(val)))
			return -EFAULT;
		switch (val) {
		case TPACKET_V1:
		case TPACKET_V2:
3081
		case TPACKET_V3:
3082 3083 3084 3085 3086 3087
			po->tp_version = val;
			return 0;
		default:
			return -EINVAL;
		}
	}
3088 3089 3090 3091 3092 3093
	case PACKET_RESERVE:
	{
		unsigned int val;

		if (optlen != sizeof(val))
			return -EINVAL;
J
Johann Baudy 已提交
3094
		if (po->rx_ring.pg_vec || po->tx_ring.pg_vec)
3095 3096 3097 3098 3099 3100
			return -EBUSY;
		if (copy_from_user(&val, optval, sizeof(val)))
			return -EFAULT;
		po->tp_reserve = val;
		return 0;
	}
J
Johann Baudy 已提交
3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111 3112 3113
	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;
	}
3114 3115 3116 3117 3118 3119 3120 3121 3122 3123 3124 3125
	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;
	}
3126 3127 3128 3129 3130 3131 3132 3133 3134 3135 3136 3137
	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;
	}
3138 3139 3140 3141 3142 3143 3144 3145 3146 3147 3148 3149 3150 3151 3152 3153
	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;
	}
3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165
	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
3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176
	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);
	}
3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189
	case PACKET_TX_HAS_OFF:
	{
		unsigned int val;

		if (optlen != sizeof(val))
			return -EINVAL;
		if (po->rx_ring.pg_vec || po->tx_ring.pg_vec)
			return -EBUSY;
		if (copy_from_user(&val, optval, sizeof(val)))
			return -EFAULT;
		po->tp_tx_has_off = !!val;
		return 0;
	}
L
Linus Torvalds 已提交
3190 3191 3192 3193 3194 3195 3196 3197 3198
	default:
		return -ENOPROTOOPT;
	}
}

static int packet_getsockopt(struct socket *sock, int level, int optname,
			     char __user *optval, int __user *optlen)
{
	int len;
3199
	int val, lv = sizeof(val);
L
Linus Torvalds 已提交
3200 3201
	struct sock *sk = sock->sk;
	struct packet_sock *po = pkt_sk(sk);
3202
	void *data = &val;
3203
	union tpacket_stats_u st;
L
Linus Torvalds 已提交
3204 3205 3206 3207

	if (level != SOL_PACKET)
		return -ENOPROTOOPT;

3208 3209
	if (get_user(len, optlen))
		return -EFAULT;
L
Linus Torvalds 已提交
3210 3211 3212

	if (len < 0)
		return -EINVAL;
3213

J
Johann Baudy 已提交
3214
	switch (optname) {
L
Linus Torvalds 已提交
3215 3216
	case PACKET_STATISTICS:
		spin_lock_bh(&sk->sk_receive_queue.lock);
3217 3218 3219 3220
		memcpy(&st, &po->stats, sizeof(st));
		memset(&po->stats, 0, sizeof(po->stats));
		spin_unlock_bh(&sk->sk_receive_queue.lock);

3221
		if (po->tp_version == TPACKET_V3) {
3222
			lv = sizeof(struct tpacket_stats_v3);
3223
			data = &st.stats3;
3224
		} else {
3225
			lv = sizeof(struct tpacket_stats);
3226
			data = &st.stats1;
3227
		}
3228

3229 3230 3231
		break;
	case PACKET_AUXDATA:
		val = po->auxdata;
3232 3233 3234
		break;
	case PACKET_ORIGDEV:
		val = po->origdev;
3235 3236 3237
		break;
	case PACKET_VNET_HDR:
		val = po->has_vnet_hdr;
L
Linus Torvalds 已提交
3238
		break;
3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251 3252 3253
	case PACKET_VERSION:
		val = po->tp_version;
		break;
	case PACKET_HDRLEN:
		if (len > sizeof(int))
			len = sizeof(int);
		if (copy_from_user(&val, optval, len))
			return -EFAULT;
		switch (val) {
		case TPACKET_V1:
			val = sizeof(struct tpacket_hdr);
			break;
		case TPACKET_V2:
			val = sizeof(struct tpacket2_hdr);
			break;
3254 3255 3256
		case TPACKET_V3:
			val = sizeof(struct tpacket3_hdr);
			break;
3257 3258 3259 3260
		default:
			return -EINVAL;
		}
		break;
3261 3262 3263
	case PACKET_RESERVE:
		val = po->tp_reserve;
		break;
J
Johann Baudy 已提交
3264 3265 3266
	case PACKET_LOSS:
		val = po->tp_loss;
		break;
3267 3268 3269
	case PACKET_TIMESTAMP:
		val = po->tp_tstamp;
		break;
D
David S. Miller 已提交
3270 3271 3272
	case PACKET_FANOUT:
		val = (po->fanout ?
		       ((u32)po->fanout->id |
3273 3274
			((u32)po->fanout->type << 16) |
			((u32)po->fanout->flags << 24)) :
D
David S. Miller 已提交
3275 3276
		       0);
		break;
3277 3278 3279
	case PACKET_TX_HAS_OFF:
		val = po->tp_tx_has_off;
		break;
L
Linus Torvalds 已提交
3280 3281 3282 3283
	default:
		return -ENOPROTOOPT;
	}

3284 3285
	if (len > lv)
		len = lv;
3286 3287
	if (put_user(len, optlen))
		return -EFAULT;
3288 3289
	if (copy_to_user(optval, data, len))
		return -EFAULT;
3290
	return 0;
L
Linus Torvalds 已提交
3291 3292 3293
}


3294 3295
static int packet_notifier(struct notifier_block *this,
			   unsigned long msg, void *ptr)
L
Linus Torvalds 已提交
3296 3297
{
	struct sock *sk;
3298
	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
3299
	struct net *net = dev_net(dev);
L
Linus Torvalds 已提交
3300

3301
	rcu_read_lock();
3302
	sk_for_each_rcu(sk, &net->packet.sklist) {
L
Linus Torvalds 已提交
3303 3304 3305 3306 3307 3308
		struct packet_sock *po = pkt_sk(sk);

		switch (msg) {
		case NETDEV_UNREGISTER:
			if (po->mclist)
				packet_dev_mclist(dev, po->mclist, -1);
3309 3310
			/* fallthrough */

L
Linus Torvalds 已提交
3311 3312 3313 3314
		case NETDEV_DOWN:
			if (dev->ifindex == po->ifindex) {
				spin_lock(&po->bind_lock);
				if (po->running) {
3315
					__unregister_prot_hook(sk, false);
L
Linus Torvalds 已提交
3316 3317 3318 3319 3320 3321
					sk->sk_err = ENETDOWN;
					if (!sock_flag(sk, SOCK_DEAD))
						sk->sk_error_report(sk);
				}
				if (msg == NETDEV_UNREGISTER) {
					po->ifindex = -1;
3322 3323
					if (po->prot_hook.dev)
						dev_put(po->prot_hook.dev);
L
Linus Torvalds 已提交
3324 3325 3326 3327 3328 3329
					po->prot_hook.dev = NULL;
				}
				spin_unlock(&po->bind_lock);
			}
			break;
		case NETDEV_UP:
3330 3331
			if (dev->ifindex == po->ifindex) {
				spin_lock(&po->bind_lock);
3332 3333
				if (po->num)
					register_prot_hook(sk);
3334
				spin_unlock(&po->bind_lock);
L
Linus Torvalds 已提交
3335 3336 3337 3338
			}
			break;
		}
	}
3339
	rcu_read_unlock();
L
Linus Torvalds 已提交
3340 3341 3342 3343 3344 3345 3346 3347 3348
	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
3349
	switch (cmd) {
E
Eric Dumazet 已提交
3350 3351 3352
	case SIOCOUTQ:
	{
		int amount = sk_wmem_alloc_get(sk);
3353

E
Eric Dumazet 已提交
3354 3355 3356 3357 3358 3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371
		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);
3372

L
Linus Torvalds 已提交
3373
#ifdef CONFIG_INET
E
Eric Dumazet 已提交
3374 3375 3376 3377 3378 3379 3380 3381 3382 3383 3384 3385 3386 3387 3388
	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
3389 3390
#endif

E
Eric Dumazet 已提交
3391 3392
	default:
		return -ENOIOCTLCMD;
L
Linus Torvalds 已提交
3393 3394 3395 3396
	}
	return 0;
}

E
Eric Dumazet 已提交
3397
static unsigned int packet_poll(struct file *file, struct socket *sock,
L
Linus Torvalds 已提交
3398 3399 3400 3401 3402 3403 3404
				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
3405
	if (po->rx_ring.pg_vec) {
3406 3407
		if (!packet_previous_rx_frame(po, &po->rx_ring,
			TP_STATUS_KERNEL))
L
Linus Torvalds 已提交
3408 3409 3410
			mask |= POLLIN | POLLRDNORM;
	}
	spin_unlock_bh(&sk->sk_receive_queue.lock);
J
Johann Baudy 已提交
3411 3412 3413 3414 3415 3416
	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
3417 3418 3419 3420 3421 3422 3423 3424 3425 3426 3427
	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
3428
	struct socket *sock = file->private_data;
L
Linus Torvalds 已提交
3429
	struct sock *sk = sock->sk;
3430

L
Linus Torvalds 已提交
3431 3432 3433 3434 3435 3436 3437
	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
3438
	struct socket *sock = file->private_data;
L
Linus Torvalds 已提交
3439
	struct sock *sk = sock->sk;
3440

L
Linus Torvalds 已提交
3441 3442 3443 3444
	if (sk)
		atomic_dec(&pkt_sk(sk)->mapped);
}

3445
static const struct vm_operations_struct packet_mmap_ops = {
E
Eric Dumazet 已提交
3446 3447
	.open	=	packet_mm_open,
	.close	=	packet_mm_close,
L
Linus Torvalds 已提交
3448 3449
};

3450 3451
static void free_pg_vec(struct pgv *pg_vec, unsigned int order,
			unsigned int len)
L
Linus Torvalds 已提交
3452 3453 3454
{
	int i;

3455
	for (i = 0; i < len; i++) {
3456
		if (likely(pg_vec[i].buffer)) {
C
Changli Gao 已提交
3457
			if (is_vmalloc_addr(pg_vec[i].buffer))
3458 3459 3460 3461 3462 3463
				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
3464 3465 3466 3467
	}
	kfree(pg_vec);
}

3468
static char *alloc_one_pg_vec_page(unsigned long order)
3469
{
3470 3471 3472 3473 3474 3475 3476 3477 3478 3479 3480 3481
	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
3482
	buffer = vzalloc((1 << order) * PAGE_SIZE);
3483

3484 3485 3486 3487 3488 3489 3490 3491 3492 3493 3494 3495 3496 3497 3498
	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;
3499 3500
}

3501
static struct pgv *alloc_pg_vec(struct tpacket_req *req, int order)
3502 3503
{
	unsigned int block_nr = req->tp_block_nr;
3504
	struct pgv *pg_vec;
3505 3506
	int i;

3507
	pg_vec = kcalloc(block_nr, sizeof(struct pgv), GFP_KERNEL);
3508 3509 3510 3511
	if (unlikely(!pg_vec))
		goto out;

	for (i = 0; i < block_nr; i++) {
C
Changli Gao 已提交
3512
		pg_vec[i].buffer = alloc_one_pg_vec_page(order);
3513
		if (unlikely(!pg_vec[i].buffer))
3514 3515 3516 3517 3518 3519 3520 3521 3522 3523 3524
			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
3525

3526
static int packet_set_ring(struct sock *sk, union tpacket_req_u *req_u,
J
Johann Baudy 已提交
3527
		int closing, int tx_ring)
L
Linus Torvalds 已提交
3528
{
3529
	struct pgv *pg_vec = NULL;
L
Linus Torvalds 已提交
3530
	struct packet_sock *po = pkt_sk(sk);
A
Al Viro 已提交
3531
	int was_running, order = 0;
J
Johann Baudy 已提交
3532 3533
	struct packet_ring_buffer *rb;
	struct sk_buff_head *rb_queue;
A
Al Viro 已提交
3534
	__be16 num;
3535 3536 3537 3538 3539 3540 3541 3542 3543
	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;
	}
3544

J
Johann Baudy 已提交
3545 3546
	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
3547

J
Johann Baudy 已提交
3548 3549 3550 3551 3552 3553 3554
	err = -EBUSY;
	if (!closing) {
		if (atomic_read(&po->mapped))
			goto out;
		if (atomic_read(&rb->pending))
			goto out;
	}
L
Linus Torvalds 已提交
3555

J
Johann Baudy 已提交
3556 3557 3558 3559 3560
	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
3561

3562 3563 3564 3565 3566 3567 3568
		switch (po->tp_version) {
		case TPACKET_V1:
			po->tp_hdrlen = TPACKET_HDRLEN;
			break;
		case TPACKET_V2:
			po->tp_hdrlen = TPACKET2_HDRLEN;
			break;
3569 3570 3571
		case TPACKET_V3:
			po->tp_hdrlen = TPACKET3_HDRLEN;
			break;
3572 3573
		}

J
Johann Baudy 已提交
3574
		err = -EINVAL;
3575
		if (unlikely((int)req->tp_block_size <= 0))
J
Johann Baudy 已提交
3576
			goto out;
3577
		if (unlikely(req->tp_block_size & (PAGE_SIZE - 1)))
J
Johann Baudy 已提交
3578
			goto out;
3579
		if (unlikely(req->tp_frame_size < po->tp_hdrlen +
J
Johann Baudy 已提交
3580 3581
					po->tp_reserve))
			goto out;
3582
		if (unlikely(req->tp_frame_size & (TPACKET_ALIGNMENT - 1)))
J
Johann Baudy 已提交
3583
			goto out;
L
Linus Torvalds 已提交
3584

J
Johann Baudy 已提交
3585 3586 3587 3588 3589 3590
		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
3591 3592

		err = -ENOMEM;
3593 3594 3595
		order = get_order(req->tp_block_size);
		pg_vec = alloc_pg_vec(req, order);
		if (unlikely(!pg_vec))
L
Linus Torvalds 已提交
3596
			goto out;
3597 3598 3599 3600 3601 3602 3603 3604 3605 3606 3607
		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
3608 3609 3610 3611
	}
	/* Done */
	else {
		err = -EINVAL;
3612
		if (unlikely(req->tp_frame_nr))
J
Johann Baudy 已提交
3613
			goto out;
L
Linus Torvalds 已提交
3614 3615 3616 3617 3618 3619 3620 3621 3622 3623
	}

	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;
3624
		__unregister_prot_hook(sk, false);
L
Linus Torvalds 已提交
3625 3626
	}
	spin_unlock(&po->bind_lock);
3627

L
Linus Torvalds 已提交
3628 3629 3630
	synchronize_net();

	err = -EBUSY;
3631
	mutex_lock(&po->pg_vec_lock);
L
Linus Torvalds 已提交
3632 3633
	if (closing || atomic_read(&po->mapped) == 0) {
		err = 0;
J
Johann Baudy 已提交
3634
		spin_lock_bh(&rb_queue->lock);
3635
		swap(rb->pg_vec, pg_vec);
J
Johann Baudy 已提交
3636 3637 3638 3639 3640
		rb->frame_max = (req->tp_frame_nr - 1);
		rb->head = 0;
		rb->frame_size = req->tp_frame_size;
		spin_unlock_bh(&rb_queue->lock);

3641 3642
		swap(rb->pg_vec_order, order);
		swap(rb->pg_vec_len, req->tp_block_nr);
J
Johann Baudy 已提交
3643 3644 3645 3646 3647

		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
3648
		if (atomic_read(&po->mapped))
E
Eric Dumazet 已提交
3649 3650
			pr_err("packet_mmap: vma is busy: %d\n",
			       atomic_read(&po->mapped));
L
Linus Torvalds 已提交
3651
	}
3652
	mutex_unlock(&po->pg_vec_lock);
L
Linus Torvalds 已提交
3653 3654

	spin_lock(&po->bind_lock);
3655
	if (was_running) {
L
Linus Torvalds 已提交
3656
		po->num = num;
3657
		register_prot_hook(sk);
L
Linus Torvalds 已提交
3658 3659
	}
	spin_unlock(&po->bind_lock);
3660 3661 3662 3663 3664
	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
3665 3666 3667 3668 3669 3670 3671 3672
	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
3673 3674
static int packet_mmap(struct file *file, struct socket *sock,
		struct vm_area_struct *vma)
L
Linus Torvalds 已提交
3675 3676 3677
{
	struct sock *sk = sock->sk;
	struct packet_sock *po = pkt_sk(sk);
J
Johann Baudy 已提交
3678 3679
	unsigned long size, expected_size;
	struct packet_ring_buffer *rb;
L
Linus Torvalds 已提交
3680 3681 3682 3683 3684 3685 3686
	unsigned long start;
	int err = -EINVAL;
	int i;

	if (vma->vm_pgoff)
		return -EINVAL;

3687
	mutex_lock(&po->pg_vec_lock);
J
Johann Baudy 已提交
3688 3689 3690 3691 3692 3693 3694 3695 3696 3697 3698

	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
3699
		goto out;
J
Johann Baudy 已提交
3700 3701 3702

	size = vma->vm_end - vma->vm_start;
	if (size != expected_size)
L
Linus Torvalds 已提交
3703 3704 3705
		goto out;

	start = vma->vm_start;
J
Johann Baudy 已提交
3706 3707 3708 3709 3710
	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++) {
3711 3712
			struct page *page;
			void *kaddr = rb->pg_vec[i].buffer;
J
Johann Baudy 已提交
3713 3714
			int pg_num;

C
Changli Gao 已提交
3715 3716
			for (pg_num = 0; pg_num < rb->pg_vec_pages; pg_num++) {
				page = pgv_to_page(kaddr);
J
Johann Baudy 已提交
3717 3718 3719 3720
				err = vm_insert_page(vma, start, page);
				if (unlikely(err))
					goto out;
				start += PAGE_SIZE;
3721
				kaddr += PAGE_SIZE;
J
Johann Baudy 已提交
3722
			}
3723
		}
L
Linus Torvalds 已提交
3724
	}
J
Johann Baudy 已提交
3725

3726
	atomic_inc(&po->mapped);
L
Linus Torvalds 已提交
3727 3728 3729 3730
	vma->vm_ops = &packet_mmap_ops;
	err = 0;

out:
3731
	mutex_unlock(&po->pg_vec_lock);
L
Linus Torvalds 已提交
3732 3733 3734
	return err;
}

3735
static const struct proto_ops packet_ops_spkt = {
L
Linus Torvalds 已提交
3736 3737 3738 3739 3740 3741 3742 3743 3744 3745 3746 3747 3748 3749 3750 3751 3752 3753 3754 3755
	.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,
};

3756
static const struct proto_ops packet_ops = {
L
Linus Torvalds 已提交
3757 3758 3759 3760 3761 3762 3763
	.family =	PF_PACKET,
	.owner =	THIS_MODULE,
	.release =	packet_release,
	.bind =		packet_bind,
	.connect =	sock_no_connect,
	.socketpair =	sock_no_socketpair,
	.accept =	sock_no_accept,
3764
	.getname =	packet_getname,
L
Linus Torvalds 已提交
3765 3766 3767 3768 3769 3770 3771 3772 3773 3774 3775 3776
	.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,
};

3777
static const struct net_proto_family packet_family_ops = {
L
Linus Torvalds 已提交
3778 3779 3780 3781 3782 3783
	.family =	PF_PACKET,
	.create =	packet_create,
	.owner	=	THIS_MODULE,
};

static struct notifier_block packet_netdev_notifier = {
E
Eric Dumazet 已提交
3784
	.notifier_call =	packet_notifier,
L
Linus Torvalds 已提交
3785 3786 3787 3788 3789
};

#ifdef CONFIG_PROC_FS

static void *packet_seq_start(struct seq_file *seq, loff_t *pos)
3790
	__acquires(RCU)
L
Linus Torvalds 已提交
3791
{
3792
	struct net *net = seq_file_net(seq);
3793 3794 3795

	rcu_read_lock();
	return seq_hlist_start_head_rcu(&net->packet.sklist, *pos);
L
Linus Torvalds 已提交
3796 3797 3798 3799
}

static void *packet_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
3800
	struct net *net = seq_file_net(seq);
3801
	return seq_hlist_next_rcu(v, &net->packet.sklist, pos);
L
Linus Torvalds 已提交
3802 3803 3804
}

static void packet_seq_stop(struct seq_file *seq, void *v)
3805
	__releases(RCU)
L
Linus Torvalds 已提交
3806
{
3807
	rcu_read_unlock();
L
Linus Torvalds 已提交
3808 3809
}

3810
static int packet_seq_show(struct seq_file *seq, void *v)
L
Linus Torvalds 已提交
3811 3812 3813 3814
{
	if (v == SEQ_START_TOKEN)
		seq_puts(seq, "sk       RefCnt Type Proto  Iface R Rmem   User   Inode\n");
	else {
3815
		struct sock *s = sk_entry(v);
L
Linus Torvalds 已提交
3816 3817 3818
		const struct packet_sock *po = pkt_sk(s);

		seq_printf(seq,
D
Dan Rosenberg 已提交
3819
			   "%pK %-6d %-4d %04x   %-5d %1d %-6u %-6u %-6lu\n",
L
Linus Torvalds 已提交
3820 3821 3822 3823 3824 3825 3826
			   s,
			   atomic_read(&s->sk_refcnt),
			   s->sk_type,
			   ntohs(po->num),
			   po->ifindex,
			   po->running,
			   atomic_read(&s->sk_rmem_alloc),
3827
			   from_kuid_munged(seq_user_ns(seq), sock_i_uid(s)),
E
Eric Dumazet 已提交
3828
			   sock_i_ino(s));
L
Linus Torvalds 已提交
3829 3830 3831 3832 3833
	}

	return 0;
}

3834
static const struct seq_operations packet_seq_ops = {
L
Linus Torvalds 已提交
3835 3836 3837 3838 3839 3840 3841 3842
	.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)
{
3843 3844
	return seq_open_net(inode, file, &packet_seq_ops,
			    sizeof(struct seq_net_private));
L
Linus Torvalds 已提交
3845 3846
}

3847
static const struct file_operations packet_seq_fops = {
L
Linus Torvalds 已提交
3848 3849 3850 3851
	.owner		= THIS_MODULE,
	.open		= packet_seq_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
3852
	.release	= seq_release_net,
L
Linus Torvalds 已提交
3853 3854 3855 3856
};

#endif

3857
static int __net_init packet_net_init(struct net *net)
3858
{
3859
	mutex_init(&net->packet.sklist_lock);
3860
	INIT_HLIST_HEAD(&net->packet.sklist);
3861

3862
	if (!proc_create("packet", 0, net->proc_net, &packet_seq_fops))
3863 3864 3865 3866 3867
		return -ENOMEM;

	return 0;
}

3868
static void __net_exit packet_net_exit(struct net *net)
3869
{
3870
	remove_proc_entry("packet", net->proc_net);
3871 3872 3873 3874 3875 3876 3877 3878
}

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


L
Linus Torvalds 已提交
3879 3880 3881
static void __exit packet_exit(void)
{
	unregister_netdevice_notifier(&packet_netdev_notifier);
3882
	unregister_pernet_subsys(&packet_net_ops);
L
Linus Torvalds 已提交
3883 3884 3885 3886 3887 3888 3889 3890 3891 3892 3893 3894
	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);
3895
	register_pernet_subsys(&packet_net_ops);
L
Linus Torvalds 已提交
3896 3897 3898 3899 3900 3901 3902 3903 3904
	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);