af_packet.c 95.7 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7
/*
 * INET		An implementation of the TCP/IP protocol suite for the LINUX
 *		operating system.  INET is implemented using the  BSD Socket
 *		interface as the means of communication with the user level.
 *
 *		PACKET - implements raw packet sockets.
 *
8
 * Authors:	Ross Biro
L
Linus Torvalds 已提交
9 10 11
 *		Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
 *		Alan Cox, <gw4pts@gw4pts.ampr.org>
 *
12
 * Fixes:
L
Linus Torvalds 已提交
13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
 *		Alan Cox	:	verify_area() now used correctly
 *		Alan Cox	:	new skbuff lists, look ma no backlogs!
 *		Alan Cox	:	tidied skbuff lists.
 *		Alan Cox	:	Now uses generic datagram routines I
 *					added. Also fixed the peek/read crash
 *					from all old Linux datagram code.
 *		Alan Cox	:	Uses the improved datagram code.
 *		Alan Cox	:	Added NULL's for socket options.
 *		Alan Cox	:	Re-commented the code.
 *		Alan Cox	:	Use new kernel side addressing
 *		Rob Janssen	:	Correct MTU usage.
 *		Dave Platt	:	Counter leaks caused by incorrect
 *					interrupt locking and some slightly
 *					dubious gcc output. Can you read
 *					compiler: it said _VOLATILE_
 *	Richard Kooijman	:	Timestamp fixes.
 *		Alan Cox	:	New buffers. Use sk->mac.raw.
 *		Alan Cox	:	sendmsg/recvmsg support.
 *		Alan Cox	:	Protocol setting support
 *	Alexey Kuznetsov	:	Untied from IPv4 stack.
 *	Cyrus Durgin		:	Fixed kerneld for kmod.
 *	Michal Ostrowski        :       Module initialization cleanup.
35
 *         Ulises Alonso        :       Frame number limit removal and
L
Linus Torvalds 已提交
36
 *                                      packet_set_ring memory leak.
37 38 39
 *		Eric Biederman	:	Allow for > 8 byte hardware addresses.
 *					The convention is that longer addresses
 *					will simply extend the hardware address
40
 *					byte arrays at the end of sockaddr_ll
41
 *					and packet_mreq.
J
Johann Baudy 已提交
42
 *		Johann Baudy	:	Added TX RING.
43 44 45 46
 *		Chetan Loke	:	Implemented TPACKET_V3 block abstraction
 *					layer.
 *					Copyright (C) 2011, <lokec@ccs.neu.edu>
 *
L
Linus Torvalds 已提交
47 48 49 50 51 52 53
 *
 *		This program is free software; you can redistribute it and/or
 *		modify it under the terms of the GNU General Public License
 *		as published by the Free Software Foundation; either version
 *		2 of the License, or (at your option) any later version.
 *
 */
54

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

96 97
#include "internal.h"

L
Linus Torvalds 已提交
98 99 100 101 102 103 104 105 106 107 108 109 110 111 112
/*
   Assumptions:
   - if device has no dev->hard_header routine, it adds and removes ll header
     inside itself. In this case ll header is invisible outside of device,
     but higher levels still should reserve dev->hard_header_len.
     Some devices are enough clever to reallocate skb, when header
     will not fit to reserved space (tunnel), another ones are silly
     (PPP).
   - packet socket receives packets with pulled ll header,
     so that SOCK_RAW should push it back.

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

Incoming, dev->hard_header!=NULL
113 114
   mac_header -> ll header
   data       -> data
L
Linus Torvalds 已提交
115 116

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

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

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

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


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

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

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

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

/* Private packet socket structures. */

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

161 162 163 164 165 166 167
union tpacket_uhdr {
	struct tpacket_hdr  *h1;
	struct tpacket2_hdr *h2;
	struct tpacket3_hdr *h3;
	void *raw;
};

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

171 172
#define V3_ALIGNMENT	(8)

173
#define BLK_HDR_LEN	(ALIGN(sizeof(struct tpacket_block_desc), V3_ALIGNMENT))
174 175 176 177 178

#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
179

180 181 182 183 184 185 186 187
#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
188 189
struct packet_sock;
static int tpacket_snd(struct packet_sock *po, struct msghdr *msg);
190 191
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
192

193 194 195 196
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);
197 198 199
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 *,
200
			struct packet_sock *);
201
static void prb_retire_current_block(struct tpacket_kbdq_core *,
202
		struct packet_sock *, unsigned int status);
203 204 205
static int prb_queue_frozen(struct tpacket_kbdq_core *);
static void prb_open_block(struct tpacket_kbdq_core *,
		struct tpacket_block_desc *);
206
static void prb_retire_rx_blk_timer_expired(unsigned long);
207 208 209 210 211 212 213 214 215
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
216 217
static void packet_flush_mclist(struct sock *sk);

218 219 220 221 222 223 224 225 226
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))
227

228
#define GET_PBDQC_FROM_RB(x)	((struct tpacket_kbdq_core *)(&(x)->prb_bdqc))
229
#define GET_PBLOCK_DESC(x, bid)	\
230
	((struct tpacket_block_desc *)((x)->pkbdq[(bid)].buffer))
231
#define GET_CURR_PBLOCK_DESC_FROM_CORE(x)	\
232
	((struct tpacket_block_desc *)((x)->pkbdq[(x)->kactive_blk_num].buffer))
233 234 235 236
#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
237 238 239
static void __fanout_unlink(struct sock *sk, struct packet_sock *po);
static void __fanout_link(struct sock *sk, struct packet_sock *po);

240 241 242 243 244 245
static int packet_direct_xmit(struct sk_buff *skb)
{
	struct net_device *dev = skb->dev;
	const struct net_device_ops *ops = dev->netdev_ops;
	netdev_features_t features;
	struct netdev_queue *txq;
246
	int ret = NETDEV_TX_BUSY;
247 248 249
	u16 queue_map;

	if (unlikely(!netif_running(dev) ||
250 251
		     !netif_carrier_ok(dev)))
		goto drop;
252 253 254

	features = netif_skb_features(skb);
	if (skb_needs_linearize(skb, features) &&
255 256
	    __skb_linearize(skb))
		goto drop;
257 258 259 260

	queue_map = skb_get_queue_mapping(skb);
	txq = netdev_get_tx_queue(dev, queue_map);

261 262 263
	local_bh_disable();

	HARD_TX_LOCK(dev, txq, smp_processor_id());
264
	if (!netif_xmit_frozen_or_drv_stopped(txq)) {
265 266 267
		ret = ops->ndo_start_xmit(skb, dev);
		if (ret == NETDEV_TX_OK)
			txq_trans_update(txq);
268
	}
269
	HARD_TX_UNLOCK(dev, txq);
270

271 272 273
	local_bh_enable();

	if (!dev_xmit_complete(ret))
274
		kfree_skb(skb);
275

276
	return ret;
277
drop:
278
	atomic_long_inc(&dev->tx_dropped);
279 280
	kfree_skb(skb);
	return NET_XMIT_DROP;
281 282
}

283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306
static struct net_device *packet_cached_dev_get(struct packet_sock *po)
{
	struct net_device *dev;

	rcu_read_lock();
	dev = rcu_dereference(po->cached_dev);
	if (likely(dev))
		dev_hold(dev);
	rcu_read_unlock();

	return dev;
}

static void packet_cached_dev_assign(struct packet_sock *po,
				     struct net_device *dev)
{
	rcu_assign_pointer(po->cached_dev, dev);
}

static void packet_cached_dev_reset(struct packet_sock *po)
{
	RCU_INIT_POINTER(po->cached_dev, NULL);
}

307 308 309 310 311
static bool packet_use_direct_xmit(const struct packet_sock *po)
{
	return po->xmit == packet_direct_xmit;
}

312
static u16 __packet_pick_tx_queue(struct net_device *dev, struct sk_buff *skb)
313
{
314
	return (u16) raw_smp_processor_id() % dev->real_num_tx_queues;
315 316
}

317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332
static void packet_pick_tx_queue(struct net_device *dev, struct sk_buff *skb)
{
	const struct net_device_ops *ops = dev->netdev_ops;
	u16 queue_index;

	if (ops->ndo_select_queue) {
		queue_index = ops->ndo_select_queue(dev, skb, NULL,
						    __packet_pick_tx_queue);
		queue_index = netdev_cap_txqueue(dev, queue_index);
	} else {
		queue_index = __packet_pick_tx_queue(dev, skb);
	}

	skb_set_queue_mapping(skb, queue_index);
}

333 334 335 336 337 338 339
/* 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);
340

341
	if (!po->running) {
342
		if (po->fanout)
D
David S. Miller 已提交
343
			__fanout_link(sk, po);
344
		else
D
David S. Miller 已提交
345
			dev_add_pack(&po->prot_hook);
346

347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363
		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;
364 365

	if (po->fanout)
D
David S. Miller 已提交
366
		__fanout_unlink(sk, po);
367
	else
D
David S. Miller 已提交
368
		__dev_remove_pack(&po->prot_hook);
369

370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386
	__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);
}

387
static inline __pure struct page *pgv_to_page(void *addr)
388 389 390 391 392 393
{
	if (is_vmalloc_addr(addr))
		return vmalloc_to_page(addr);
	return virt_to_page(addr);
}

J
Johann Baudy 已提交
394
static void __packet_set_status(struct packet_sock *po, void *frame, int status)
L
Linus Torvalds 已提交
395
{
396
	union tpacket_uhdr h;
L
Linus Torvalds 已提交
397

J
Johann Baudy 已提交
398
	h.raw = frame;
399 400
	switch (po->tp_version) {
	case TPACKET_V1:
J
Johann Baudy 已提交
401
		h.h1->tp_status = status;
402
		flush_dcache_page(pgv_to_page(&h.h1->tp_status));
403 404
		break;
	case TPACKET_V2:
J
Johann Baudy 已提交
405
		h.h2->tp_status = status;
406
		flush_dcache_page(pgv_to_page(&h.h2->tp_status));
407
		break;
408
	case TPACKET_V3:
J
Johann Baudy 已提交
409
	default:
410
		WARN(1, "TPACKET version not supported.\n");
J
Johann Baudy 已提交
411
		BUG();
412
	}
J
Johann Baudy 已提交
413 414

	smp_wmb();
415 416
}

J
Johann Baudy 已提交
417
static int __packet_get_status(struct packet_sock *po, void *frame)
418
{
419
	union tpacket_uhdr h;
420

J
Johann Baudy 已提交
421 422
	smp_rmb();

423 424 425
	h.raw = frame;
	switch (po->tp_version) {
	case TPACKET_V1:
426
		flush_dcache_page(pgv_to_page(&h.h1->tp_status));
J
Johann Baudy 已提交
427
		return h.h1->tp_status;
428
	case TPACKET_V2:
429
		flush_dcache_page(pgv_to_page(&h.h2->tp_status));
J
Johann Baudy 已提交
430
		return h.h2->tp_status;
431
	case TPACKET_V3:
J
Johann Baudy 已提交
432
	default:
433
		WARN(1, "TPACKET version not supported.\n");
J
Johann Baudy 已提交
434 435
		BUG();
		return 0;
436
	}
L
Linus Torvalds 已提交
437
}
J
Johann Baudy 已提交
438

439 440
static __u32 tpacket_get_timestamp(struct sk_buff *skb, struct timespec *ts,
				   unsigned int flags)
441 442 443 444 445 446
{
	struct skb_shared_hwtstamps *shhwtstamps = skb_hwtstamps(skb);

	if (shhwtstamps) {
		if ((flags & SOF_TIMESTAMPING_SYS_HARDWARE) &&
		    ktime_to_timespec_cond(shhwtstamps->syststamp, ts))
447
			return TP_STATUS_TS_SYS_HARDWARE;
448 449
		if ((flags & SOF_TIMESTAMPING_RAW_HARDWARE) &&
		    ktime_to_timespec_cond(shhwtstamps->hwtstamp, ts))
450
			return TP_STATUS_TS_RAW_HARDWARE;
451 452 453
	}

	if (ktime_to_timespec_cond(skb->tstamp, ts))
454
		return TP_STATUS_TS_SOFTWARE;
455

456
	return 0;
457 458
}

459 460
static __u32 __packet_set_timestamp(struct packet_sock *po, void *frame,
				    struct sk_buff *skb)
461 462 463
{
	union tpacket_uhdr h;
	struct timespec ts;
464
	__u32 ts_status;
465

466 467
	if (!(ts_status = tpacket_get_timestamp(skb, &ts, po->tp_tstamp)))
		return 0;
468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487

	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();
488 489

	return ts_status;
490 491
}

J
Johann Baudy 已提交
492 493 494 495 496 497
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;
498
	union tpacket_uhdr h;
J
Johann Baudy 已提交
499 500 501 502

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

503 504
	h.raw = rb->pg_vec[pg_vec_pos].buffer +
		(frame_offset * rb->frame_size);
J
Johann Baudy 已提交
505 506 507 508 509 510 511

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

	return h.raw;
}

512
static void *packet_current_frame(struct packet_sock *po,
J
Johann Baudy 已提交
513 514 515 516 517 518
		struct packet_ring_buffer *rb,
		int status)
{
	return packet_lookup_frame(po, rb, rb->head, status);
}

519
static void prb_del_retire_blk_timer(struct tpacket_kbdq_core *pkc)
520 521 522 523 524 525 526 527
{
	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)
{
528
	struct tpacket_kbdq_core *pkc;
529

530 531
	pkc = tx_ring ? GET_PBDQC_FROM_RB(&po->tx_ring) :
			GET_PBDQC_FROM_RB(&po->rx_ring);
532

533
	spin_lock_bh(&rb_queue->lock);
534
	pkc->delete_blk_timer = 1;
535
	spin_unlock_bh(&rb_queue->lock);
536 537 538 539 540

	prb_del_retire_blk_timer(pkc);
}

static void prb_init_blk_timer(struct packet_sock *po,
541
		struct tpacket_kbdq_core *pkc,
542 543 544 545 546 547 548 549 550 551
		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)
{
552
	struct tpacket_kbdq_core *pkc;
553 554 555 556

	if (tx_ring)
		BUG();

557 558
	pkc = tx_ring ? GET_PBDQC_FROM_RB(&po->tx_ring) :
			GET_PBDQC_FROM_RB(&po->rx_ring);
559 560 561 562 563 564 565 566
	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;
567 568
	struct ethtool_cmd ecmd;
	int err;
569
	u32 speed;
570

571 572 573 574
	rtnl_lock();
	dev = __dev_get_by_index(sock_net(&po->sk), po->ifindex);
	if (unlikely(!dev)) {
		rtnl_unlock();
575
		return DEFAULT_PRB_RETIRE_TOV;
576 577
	}
	err = __ethtool_get_settings(dev, &ecmd);
578
	speed = ethtool_cmd_speed(&ecmd);
579 580 581 582 583 584
	rtnl_unlock();
	if (!err) {
		/*
		 * If the link speed is so slow you don't really
		 * need to worry about perf anyways
		 */
585
		if (speed < SPEED_1000 || speed == SPEED_UNKNOWN) {
586
			return DEFAULT_PRB_RETIRE_TOV;
587 588 589
		} else {
			msec = 1;
			div = speed / 1000;
590 591 592 593 594 595 596 597 598 599 600 601 602 603 604
		}
	}

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

	if (div)
		mbits /= div;

	tmo = mbits * msec;

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

605
static void prb_init_ft_ops(struct tpacket_kbdq_core *p1,
606 607 608 609 610 611 612 613 614 615
			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)
{
616
	struct tpacket_kbdq_core *p1 = GET_PBDQC_FROM_RB(rb);
617
	struct tpacket_block_desc *pbd;
618 619 620 621 622

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

	p1->knxt_seq_num = 1;
	p1->pkbdq = pg_vec;
623
	pbd = (struct tpacket_block_desc *)pg_vec[0].buffer;
J
Joe Perches 已提交
624
	p1->pkblk_start	= pg_vec[0].buffer;
625 626 627 628 629
	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;
630
	po->stats.stats3.tp_freeze_q_cnt = 0;
631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646
	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.
 */
647
static void _prb_refresh_rx_retire_blk_timer(struct tpacket_kbdq_core *pkc)
648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679
{
	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;
680
	struct tpacket_kbdq_core *pkc = GET_PBDQC_FROM_RB(&po->rx_ring);
681
	unsigned int frozen;
682
	struct tpacket_block_desc *pbd;
683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745

	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);
}

746
static void prb_flush_block(struct tpacket_kbdq_core *pkc1,
747
		struct tpacket_block_desc *pbd1, __u32 status)
748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788
{
	/* 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.
 */
789 790
static void prb_close_block(struct tpacket_kbdq_core *pkc1,
		struct tpacket_block_desc *pbd1,
791 792 793 794 795
		struct packet_sock *po, unsigned int stat)
{
	__u32 status = TP_STATUS_USER | stat;

	struct tpacket3_hdr *last_pkt;
796
	struct tpacket_hdr_v1 *h1 = &pbd1->hdr.bh1;
797

798
	if (po->stats.stats3.tp_drops)
799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823
		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);
}

824
static void prb_thaw_queue(struct tpacket_kbdq_core *pkc)
825 826 827 828 829 830 831 832 833 834 835
{
	pkc->reset_pending_on_curr_blk = 0;
}

/*
 * Side effect of opening a block:
 *
 * 1) prb_queue is thawed.
 * 2) retire_blk_timer is refreshed.
 *
 */
836 837
static void prb_open_block(struct tpacket_kbdq_core *pkc1,
	struct tpacket_block_desc *pbd1)
838 839
{
	struct timespec ts;
840
	struct tpacket_hdr_v1 *h1 = &pbd1->hdr.bh1;
841 842 843

	smp_rmb();

844 845 846
	/* We could have just memset this but we will lose the
	 * flexibility of making the priv area sticky
	 */
847

848 849 850
	BLOCK_SNUM(pbd1) = pkc1->knxt_seq_num++;
	BLOCK_NUM_PKTS(pbd1) = 0;
	BLOCK_LEN(pbd1) = BLK_PLUS_PRIV(pkc1->blk_sizeof_priv);
851

852 853 854 855
	getnstimeofday(&ts);

	h1->ts_first_pkt.ts_sec = ts.tv_sec;
	h1->ts_first_pkt.ts_nsec = ts.tv_nsec;
856

857 858 859 860 861 862 863 864 865 866 867 868 869 870
	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();
871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895
}

/*
 * 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.
 */
896
static void prb_freeze_queue(struct tpacket_kbdq_core *pkc,
897 898 899
				  struct packet_sock *po)
{
	pkc->reset_pending_on_curr_blk = 1;
900
	po->stats.stats3.tp_freeze_q_cnt++;
901 902 903 904 905 906 907 908 909 910
}

#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.
 */
911
static void *prb_dispatch_next_block(struct tpacket_kbdq_core *pkc,
912 913
		struct packet_sock *po)
{
914
	struct tpacket_block_desc *pbd;
915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935

	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;
}

936
static void prb_retire_current_block(struct tpacket_kbdq_core *pkc,
937 938
		struct packet_sock *po, unsigned int status)
{
939
	struct tpacket_block_desc *pbd = GET_CURR_PBLOCK_DESC_FROM_CORE(pkc);
940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962

	/* 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;
	}
}

963
static int prb_curr_blk_in_use(struct tpacket_kbdq_core *pkc,
964
				      struct tpacket_block_desc *pbd)
965 966 967 968
{
	return TP_STATUS_USER & BLOCK_STATUS(pbd);
}

969
static int prb_queue_frozen(struct tpacket_kbdq_core *pkc)
970 971 972 973
{
	return pkc->reset_pending_on_curr_blk;
}

974
static void prb_clear_blk_fill_status(struct packet_ring_buffer *rb)
975
{
976
	struct tpacket_kbdq_core *pkc  = GET_PBDQC_FROM_RB(rb);
977 978 979
	atomic_dec(&pkc->blk_fill_in_prog);
}

980
static void prb_fill_rxhash(struct tpacket_kbdq_core *pkc,
981 982
			struct tpacket3_hdr *ppd)
{
983
	ppd->hv1.tp_rxhash = skb_get_hash(pkc->skb);
984 985
}

986
static void prb_clear_rxhash(struct tpacket_kbdq_core *pkc,
987 988 989 990 991
			struct tpacket3_hdr *ppd)
{
	ppd->hv1.tp_rxhash = 0;
}

992
static void prb_fill_vlan_info(struct tpacket_kbdq_core *pkc,
993 994 995 996
			struct tpacket3_hdr *ppd)
{
	if (vlan_tx_tag_present(pkc->skb)) {
		ppd->hv1.tp_vlan_tci = vlan_tx_tag_get(pkc->skb);
997 998
		ppd->hv1.tp_vlan_tpid = ntohs(pkc->skb->vlan_proto);
		ppd->tp_status = TP_STATUS_VLAN_VALID | TP_STATUS_VLAN_TPID_VALID;
999
	} else {
1000
		ppd->hv1.tp_vlan_tci = 0;
1001
		ppd->hv1.tp_vlan_tpid = 0;
1002
		ppd->tp_status = TP_STATUS_AVAILABLE;
1003 1004 1005
	}
}

1006
static void prb_run_all_ft_ops(struct tpacket_kbdq_core *pkc,
1007 1008
			struct tpacket3_hdr *ppd)
{
1009
	ppd->hv1.tp_padding = 0;
1010 1011 1012 1013 1014 1015 1016 1017
	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);
}

1018
static void prb_fill_curr_block(char *curr,
1019 1020
				struct tpacket_kbdq_core *pkc,
				struct tpacket_block_desc *pbd,
1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041
				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
					    )
{
1042 1043
	struct tpacket_kbdq_core *pkc;
	struct tpacket_block_desc *pbd;
1044 1045
	char *curr, *end;

J
Joe Perches 已提交
1046
	pkc = GET_PBDQC_FROM_RB(&po->rx_ring);
1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071
	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
1072
	end = (char *)pbd + pkc->kblk_size;
1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097

	/* 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;
}

1098
static void *packet_current_rx_frame(struct packet_sock *po,
1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113
					    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();
1114
		return NULL;
1115 1116 1117
	}
}

1118
static void *prb_lookup_block(struct packet_sock *po,
1119
				     struct packet_ring_buffer *rb,
1120
				     unsigned int idx,
1121 1122
				     int status)
{
1123
	struct tpacket_kbdq_core *pkc  = GET_PBDQC_FROM_RB(rb);
1124
	struct tpacket_block_desc *pbd = GET_PBLOCK_DESC(pkc, idx);
1125 1126 1127 1128 1129 1130

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

1131
static int prb_previous_blk_num(struct packet_ring_buffer *rb)
1132 1133 1134 1135 1136 1137 1138 1139 1140 1141
{
	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 */
1142
static void *__prb_previous_block(struct packet_sock *po,
1143 1144 1145 1146 1147 1148 1149
					 struct packet_ring_buffer *rb,
					 int status)
{
	unsigned int previous = prb_previous_blk_num(rb);
	return prb_lookup_block(po, rb, previous, status);
}

1150
static void *packet_previous_rx_frame(struct packet_sock *po,
1151 1152 1153 1154 1155 1156 1157 1158 1159
					     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);
}

1160
static void packet_increment_rx_head(struct packet_sock *po,
1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174
					    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;
	}
}

1175
static void *packet_previous_frame(struct packet_sock *po,
J
Johann Baudy 已提交
1176 1177 1178 1179 1180 1181 1182
		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);
}

1183
static void packet_increment_head(struct packet_ring_buffer *buff)
J
Johann Baudy 已提交
1184 1185 1186 1187
{
	buff->head = buff->head != buff->frame_max ? buff->head+1 : 0;
}

1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228
static void packet_inc_pending(struct packet_ring_buffer *rb)
{
	this_cpu_inc(*rb->pending_refcnt);
}

static void packet_dec_pending(struct packet_ring_buffer *rb)
{
	this_cpu_dec(*rb->pending_refcnt);
}

static unsigned int packet_read_pending(const struct packet_ring_buffer *rb)
{
	unsigned int refcnt = 0;
	int cpu;

	/* We don't use pending refcount in rx_ring. */
	if (rb->pending_refcnt == NULL)
		return 0;

	for_each_possible_cpu(cpu)
		refcnt += *per_cpu_ptr(rb->pending_refcnt, cpu);

	return refcnt;
}

static int packet_alloc_pending(struct packet_sock *po)
{
	po->rx_ring.pending_refcnt = NULL;

	po->tx_ring.pending_refcnt = alloc_percpu(unsigned int);
	if (unlikely(po->tx_ring.pending_refcnt == NULL))
		return -ENOBUFS;

	return 0;
}

static void packet_free_pending(struct packet_sock *po)
{
	free_percpu(po->tx_ring.pending_refcnt);
}

1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251
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
1252 1253
static void packet_sock_destruct(struct sock *sk)
{
1254 1255
	skb_queue_purge(&sk->sk_error_queue);

1256 1257
	WARN_ON(atomic_read(&sk->sk_rmem_alloc));
	WARN_ON(atomic_read(&sk->sk_wmem_alloc));
L
Linus Torvalds 已提交
1258 1259

	if (!sock_flag(sk, SOCK_DEAD)) {
E
Eric Dumazet 已提交
1260
		pr_err("Attempt to release alive packet socket: %p\n", sk);
L
Linus Torvalds 已提交
1261 1262 1263
		return;
	}

1264
	sk_refcnt_debug_dec(sk);
L
Linus Torvalds 已提交
1265 1266
}

D
David S. Miller 已提交
1267 1268 1269 1270 1271 1272 1273 1274 1275 1276
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;
}

1277 1278 1279
static unsigned int fanout_demux_hash(struct packet_fanout *f,
				      struct sk_buff *skb,
				      unsigned int num)
D
David S. Miller 已提交
1280
{
1281
	return reciprocal_scale(skb_get_hash(skb), num);
D
David S. Miller 已提交
1282 1283
}

1284 1285 1286
static unsigned int fanout_demux_lb(struct packet_fanout *f,
				    struct sk_buff *skb,
				    unsigned int num)
D
David S. Miller 已提交
1287 1288 1289 1290 1291 1292 1293
{
	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;
1294 1295 1296 1297 1298 1299 1300 1301
	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
1302 1303
}

1304 1305 1306 1307
static unsigned int fanout_demux_rnd(struct packet_fanout *f,
				     struct sk_buff *skb,
				     unsigned int num)
{
1308
	return prandom_u32_max(num);
1309 1310
}

1311 1312 1313 1314
static unsigned int fanout_demux_rollover(struct packet_fanout *f,
					  struct sk_buff *skb,
					  unsigned int idx, unsigned int skip,
					  unsigned int num)
1315
{
1316
	unsigned int i, j;
1317

1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331
	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;
}

1332 1333 1334 1335 1336 1337 1338
static unsigned int fanout_demux_qm(struct packet_fanout *f,
				    struct sk_buff *skb,
				    unsigned int num)
{
	return skb_get_queue_mapping(skb) % num;
}

1339 1340 1341
static bool fanout_has_flag(struct packet_fanout *f, u16 flag)
{
	return f->flags & (flag >> 8);
1342 1343 1344 1345
}

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
1346 1347 1348 1349
{
	struct packet_fanout *f = pt->af_packet_priv;
	unsigned int num = f->num_members;
	struct packet_sock *po;
1350
	unsigned int idx;
D
David S. Miller 已提交
1351 1352 1353 1354 1355 1356 1357

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

1358 1359 1360
	switch (f->type) {
	case PACKET_FANOUT_HASH:
	default:
1361
		if (fanout_has_flag(f, PACKET_FANOUT_FLAG_DEFRAG)) {
1362
			skb = ip_check_defrag(skb, IP_DEFRAG_AF_PACKET);
1363 1364 1365
			if (!skb)
				return 0;
		}
1366
		idx = fanout_demux_hash(f, skb, num);
1367 1368
		break;
	case PACKET_FANOUT_LB:
1369
		idx = fanout_demux_lb(f, skb, num);
1370 1371
		break;
	case PACKET_FANOUT_CPU:
1372 1373
		idx = fanout_demux_cpu(f, skb, num);
		break;
1374 1375 1376
	case PACKET_FANOUT_RND:
		idx = fanout_demux_rnd(f, skb, num);
		break;
1377 1378 1379
	case PACKET_FANOUT_QM:
		idx = fanout_demux_qm(f, skb, num);
		break;
1380 1381
	case PACKET_FANOUT_ROLLOVER:
		idx = fanout_demux_rollover(f, skb, 0, (unsigned int) -1, num);
1382
		break;
D
David S. Miller 已提交
1383 1384
	}

1385 1386 1387 1388 1389 1390
	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
1391 1392 1393 1394

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

1395 1396
DEFINE_MUTEX(fanout_mutex);
EXPORT_SYMBOL_GPL(fanout_mutex);
D
David S. Miller 已提交
1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425
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);
}

1426
static bool match_fanout_group(struct packet_type *ptype, struct sock *sk)
1427
{
1428
	if (ptype->af_packet_priv == (void *)((struct packet_sock *)sk)->fanout)
1429 1430 1431 1432 1433
		return true;

	return false;
}

1434
static int fanout_add(struct sock *sk, u16 id, u16 type_flags)
D
David S. Miller 已提交
1435 1436 1437
{
	struct packet_sock *po = pkt_sk(sk);
	struct packet_fanout *f, *match;
1438
	u8 type = type_flags & 0xff;
1439
	u8 flags = type_flags >> 8;
D
David S. Miller 已提交
1440 1441 1442
	int err;

	switch (type) {
1443 1444 1445
	case PACKET_FANOUT_ROLLOVER:
		if (type_flags & PACKET_FANOUT_FLAG_ROLLOVER)
			return -EINVAL;
D
David S. Miller 已提交
1446 1447
	case PACKET_FANOUT_HASH:
	case PACKET_FANOUT_LB:
1448
	case PACKET_FANOUT_CPU:
1449
	case PACKET_FANOUT_RND:
1450
	case PACKET_FANOUT_QM:
D
David S. Miller 已提交
1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470
		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
1471
	err = -EINVAL;
1472
	if (match && match->flags != flags)
E
Eric Dumazet 已提交
1473
		goto out;
D
David S. Miller 已提交
1474
	if (!match) {
E
Eric Dumazet 已提交
1475
		err = -ENOMEM;
D
David S. Miller 已提交
1476
		match = kzalloc(sizeof(*match), GFP_KERNEL);
E
Eric Dumazet 已提交
1477 1478 1479 1480 1481
		if (!match)
			goto out;
		write_pnet(&match->net, sock_net(sk));
		match->id = id;
		match->type = type;
1482
		match->flags = flags;
E
Eric Dumazet 已提交
1483 1484 1485 1486 1487 1488 1489 1490
		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;
1491
		match->prot_hook.id_match = match_fanout_group;
E
Eric Dumazet 已提交
1492 1493
		dev_add_pack(&match->prot_hook);
		list_add(&match->list, &fanout_list);
D
David S. Miller 已提交
1494
	}
E
Eric Dumazet 已提交
1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505
	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
1506 1507
		}
	}
E
Eric Dumazet 已提交
1508
out:
D
David S. Miller 已提交
1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521
	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;

1522
	mutex_lock(&fanout_mutex);
D
David S. Miller 已提交
1523 1524 1525 1526 1527 1528 1529 1530 1531
	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
1532

1533
static const struct proto_ops packet_ops;
L
Linus Torvalds 已提交
1534

1535
static const struct proto_ops packet_ops_spkt;
L
Linus Torvalds 已提交
1536

E
Eric Dumazet 已提交
1537 1538
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
1539 1540 1541 1542 1543 1544 1545 1546 1547 1548
{
	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;
1549

L
Linus Torvalds 已提交
1550 1551 1552 1553 1554 1555 1556
	/*
	 *	Yank back the headers [hope the device set this
	 *	right or kerboom...]
	 *
	 *	Incoming packets have ll header pulled,
	 *	push it back.
	 *
1557
	 *	For outgoing ones skb->data == skb_mac_header(skb)
L
Linus Torvalds 已提交
1558 1559 1560 1561 1562 1563
	 *	so that this procedure is noop.
	 */

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

O
Octavian Purdila 已提交
1564
	if (!net_eq(dev_net(dev), sock_net(sk)))
1565 1566
		goto out;

E
Eric Dumazet 已提交
1567 1568
	skb = skb_share_check(skb, GFP_ATOMIC);
	if (skb == NULL)
L
Linus Torvalds 已提交
1569 1570 1571
		goto oom;

	/* drop any routing info */
E
Eric Dumazet 已提交
1572
	skb_dst_drop(skb);
L
Linus Torvalds 已提交
1573

1574 1575 1576
	/* drop conntrack reference */
	nf_reset(skb);

1577
	spkt = &PACKET_SKB_CB(skb)->sa.pkt;
L
Linus Torvalds 已提交
1578

1579
	skb_push(skb, skb->data - skb_mac_header(skb));
L
Linus Torvalds 已提交
1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593

	/*
	 *	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
1594
	if (sock_queue_rcv_skb(sk, skb) == 0)
L
Linus Torvalds 已提交
1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607
		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
 */
1608

L
Linus Torvalds 已提交
1609 1610 1611 1612
static int packet_sendmsg_spkt(struct kiocb *iocb, struct socket *sock,
			       struct msghdr *msg, size_t len)
{
	struct sock *sk = sock->sk;
1613
	DECLARE_SOCKADDR(struct sockaddr_pkt *, saddr, msg->msg_name);
1614
	struct sk_buff *skb = NULL;
L
Linus Torvalds 已提交
1615
	struct net_device *dev;
E
Eric Dumazet 已提交
1616
	__be16 proto = 0;
L
Linus Torvalds 已提交
1617
	int err;
1618
	int extra_len = 0;
1619

L
Linus Torvalds 已提交
1620
	/*
1621
	 *	Get and verify the address.
L
Linus Torvalds 已提交
1622 1623
	 */

E
Eric Dumazet 已提交
1624
	if (saddr) {
L
Linus Torvalds 已提交
1625
		if (msg->msg_namelen < sizeof(struct sockaddr))
E
Eric Dumazet 已提交
1626 1627 1628 1629 1630
			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
1631 1632

	/*
1633
	 *	Find the device first to size check it
L
Linus Torvalds 已提交
1634 1635
	 */

1636
	saddr->spkt_device[sizeof(saddr->spkt_device) - 1] = 0;
1637
retry:
E
Eric Dumazet 已提交
1638 1639
	rcu_read_lock();
	dev = dev_get_by_name_rcu(sock_net(sk), saddr->spkt_device);
L
Linus Torvalds 已提交
1640 1641 1642
	err = -ENODEV;
	if (dev == NULL)
		goto out_unlock;
1643

1644 1645 1646 1647
	err = -ENETDOWN;
	if (!(dev->flags & IFF_UP))
		goto out_unlock;

L
Linus Torvalds 已提交
1648
	/*
E
Eric Dumazet 已提交
1649 1650
	 * 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
1651
	 */
1652

1653 1654 1655 1656 1657 1658 1659 1660
	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
1661
	err = -EMSGSIZE;
1662
	if (len > dev->mtu + dev->hard_header_len + VLAN_HLEN + extra_len)
L
Linus Torvalds 已提交
1663 1664
		goto out_unlock;

1665 1666
	if (!skb) {
		size_t reserved = LL_RESERVED_SPACE(dev);
1667
		int tlen = dev->needed_tailroom;
1668 1669 1670
		unsigned int hhlen = dev->header_ops ? dev->hard_header_len : 0;

		rcu_read_unlock();
1671
		skb = sock_wmalloc(sk, len + reserved + tlen, 0, GFP_KERNEL);
1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691
		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
1692 1693
	}

1694
	if (len > (dev->mtu + dev->hard_header_len + extra_len)) {
1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706
		/* 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;
		}
	}
1707

L
Linus Torvalds 已提交
1708 1709 1710
	skb->protocol = proto;
	skb->dev = dev;
	skb->priority = sk->sk_priority;
1711
	skb->mark = sk->sk_mark;
1712 1713

	sock_tx_timestamp(sk, &skb_shinfo(skb)->tx_flags);
L
Linus Torvalds 已提交
1714

1715 1716 1717
	if (unlikely(extra_len == 4))
		skb->no_fcs = 1;

1718
	skb_probe_transport_header(skb, 0);
1719

L
Linus Torvalds 已提交
1720
	dev_queue_xmit(skb);
E
Eric Dumazet 已提交
1721
	rcu_read_unlock();
E
Eric Dumazet 已提交
1722
	return len;
L
Linus Torvalds 已提交
1723 1724

out_unlock:
E
Eric Dumazet 已提交
1725
	rcu_read_unlock();
1726 1727
out_free:
	kfree_skb(skb);
L
Linus Torvalds 已提交
1728 1729 1730
	return err;
}

1731
static unsigned int run_filter(const struct sk_buff *skb,
E
Eric Dumazet 已提交
1732
				      const struct sock *sk,
D
David S. Miller 已提交
1733
				      unsigned int res)
L
Linus Torvalds 已提交
1734 1735
{
	struct sk_filter *filter;
1736

1737 1738
	rcu_read_lock();
	filter = rcu_dereference(sk->sk_filter);
D
David S. Miller 已提交
1739
	if (filter != NULL)
1740
		res = SK_RUN_FILTER(filter, skb);
1741
	rcu_read_unlock();
L
Linus Torvalds 已提交
1742

D
David S. Miller 已提交
1743
	return res;
L
Linus Torvalds 已提交
1744 1745 1746
}

/*
E
Eric Dumazet 已提交
1747 1748 1749 1750 1751 1752 1753 1754 1755
 * 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
1756 1757
 */

E
Eric Dumazet 已提交
1758 1759
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
1760 1761 1762 1763
{
	struct sock *sk;
	struct sockaddr_ll *sll;
	struct packet_sock *po;
E
Eric Dumazet 已提交
1764
	u8 *skb_head = skb->data;
L
Linus Torvalds 已提交
1765
	int skb_len = skb->len;
D
David S. Miller 已提交
1766
	unsigned int snaplen, res;
L
Linus Torvalds 已提交
1767 1768 1769 1770 1771 1772 1773

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

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

O
Octavian Purdila 已提交
1774
	if (!net_eq(dev_net(dev), sock_net(sk)))
1775 1776
		goto drop;

L
Linus Torvalds 已提交
1777 1778
	skb->dev = dev;

1779
	if (dev->header_ops) {
L
Linus Torvalds 已提交
1780
		/* The device has an explicit notion of ll header,
E
Eric Dumazet 已提交
1781 1782 1783 1784 1785
		 * 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
1786 1787
		 */
		if (sk->sk_type != SOCK_DGRAM)
1788
			skb_push(skb, skb->data - skb_mac_header(skb));
L
Linus Torvalds 已提交
1789 1790
		else if (skb->pkt_type == PACKET_OUTGOING) {
			/* Special case: outgoing packets have ll header at head */
1791
			skb_pull(skb, skb_network_offset(skb));
L
Linus Torvalds 已提交
1792 1793 1794 1795 1796
		}
	}

	snaplen = skb->len;

D
David S. Miller 已提交
1797 1798
	res = run_filter(skb, sk, snaplen);
	if (!res)
1799
		goto drop_n_restore;
D
David S. Miller 已提交
1800 1801
	if (snaplen > res)
		snaplen = res;
L
Linus Torvalds 已提交
1802

E
Eric Dumazet 已提交
1803
	if (atomic_read(&sk->sk_rmem_alloc) >= sk->sk_rcvbuf)
L
Linus Torvalds 已提交
1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814
		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;
		}
1815
		consume_skb(skb);
L
Linus Torvalds 已提交
1816 1817 1818
		skb = nskb;
	}

1819 1820 1821 1822
	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
1823 1824 1825 1826
	sll->sll_family = AF_PACKET;
	sll->sll_hatype = dev->type;
	sll->sll_protocol = skb->protocol;
	sll->sll_pkttype = skb->pkt_type;
1827
	if (unlikely(po->origdev))
1828 1829 1830
		sll->sll_ifindex = orig_dev->ifindex;
	else
		sll->sll_ifindex = dev->ifindex;
L
Linus Torvalds 已提交
1831

S
Stephen Hemminger 已提交
1832
	sll->sll_halen = dev_parse_header(skb, sll->sll_addr);
L
Linus Torvalds 已提交
1833

1834
	PACKET_SKB_CB(skb)->origlen = skb->len;
1835

L
Linus Torvalds 已提交
1836 1837 1838 1839 1840
	if (pskb_trim(skb, snaplen))
		goto drop_n_acct;

	skb_set_owner_r(skb, sk);
	skb->dev = NULL;
E
Eric Dumazet 已提交
1841
	skb_dst_drop(skb);
L
Linus Torvalds 已提交
1842

1843 1844 1845
	/* drop conntrack reference */
	nf_reset(skb);

L
Linus Torvalds 已提交
1846
	spin_lock(&sk->sk_receive_queue.lock);
1847
	po->stats.stats1.tp_packets++;
1848
	skb->dropcount = atomic_read(&sk->sk_drops);
L
Linus Torvalds 已提交
1849 1850
	__skb_queue_tail(&sk->sk_receive_queue, skb);
	spin_unlock(&sk->sk_receive_queue.lock);
1851
	sk->sk_data_ready(sk);
L
Linus Torvalds 已提交
1852 1853 1854
	return 0;

drop_n_acct:
1855
	spin_lock(&sk->sk_receive_queue.lock);
1856
	po->stats.stats1.tp_drops++;
1857 1858
	atomic_inc(&sk->sk_drops);
	spin_unlock(&sk->sk_receive_queue.lock);
L
Linus Torvalds 已提交
1859 1860 1861 1862 1863 1864 1865

drop_n_restore:
	if (skb_head != skb->data && skb_shared(skb)) {
		skb->data = skb_head;
		skb->len = skb_len;
	}
drop:
1866
	consume_skb(skb);
L
Linus Torvalds 已提交
1867 1868 1869
	return 0;
}

E
Eric Dumazet 已提交
1870 1871
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
1872 1873 1874 1875
{
	struct sock *sk;
	struct packet_sock *po;
	struct sockaddr_ll *sll;
1876
	union tpacket_uhdr h;
E
Eric Dumazet 已提交
1877
	u8 *skb_head = skb->data;
L
Linus Torvalds 已提交
1878
	int skb_len = skb->len;
D
David S. Miller 已提交
1879
	unsigned int snaplen, res;
1880
	unsigned long status = TP_STATUS_USER;
1881
	unsigned short macoff, netoff, hdrlen;
L
Linus Torvalds 已提交
1882
	struct sk_buff *copy_skb = NULL;
1883
	struct timespec ts;
1884
	__u32 ts_status;
L
Linus Torvalds 已提交
1885

1886 1887 1888 1889 1890 1891 1892
	/* struct tpacket{2,3}_hdr is aligned to a multiple of TPACKET_ALIGNMENT.
	 * We may add members to them until current aligned size without forcing
	 * userspace to call getsockopt(..., PACKET_HDRLEN, ...).
	 */
	BUILD_BUG_ON(TPACKET_ALIGN(sizeof(*h.h2)) != 32);
	BUILD_BUG_ON(TPACKET_ALIGN(sizeof(*h.h3)) != 48);

L
Linus Torvalds 已提交
1893 1894 1895 1896 1897 1898
	if (skb->pkt_type == PACKET_LOOPBACK)
		goto drop;

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

O
Octavian Purdila 已提交
1899
	if (!net_eq(dev_net(dev), sock_net(sk)))
1900 1901
		goto drop;

1902
	if (dev->header_ops) {
L
Linus Torvalds 已提交
1903
		if (sk->sk_type != SOCK_DGRAM)
1904
			skb_push(skb, skb->data - skb_mac_header(skb));
L
Linus Torvalds 已提交
1905 1906
		else if (skb->pkt_type == PACKET_OUTGOING) {
			/* Special case: outgoing packets have ll header at head */
1907
			skb_pull(skb, skb_network_offset(skb));
L
Linus Torvalds 已提交
1908 1909 1910
		}
	}

1911 1912 1913
	if (skb->ip_summed == CHECKSUM_PARTIAL)
		status |= TP_STATUS_CSUMNOTREADY;

L
Linus Torvalds 已提交
1914 1915
	snaplen = skb->len;

D
David S. Miller 已提交
1916 1917
	res = run_filter(skb, sk, snaplen);
	if (!res)
1918
		goto drop_n_restore;
D
David S. Miller 已提交
1919 1920
	if (snaplen > res)
		snaplen = res;
L
Linus Torvalds 已提交
1921 1922

	if (sk->sk_type == SOCK_DGRAM) {
1923 1924
		macoff = netoff = TPACKET_ALIGN(po->tp_hdrlen) + 16 +
				  po->tp_reserve;
L
Linus Torvalds 已提交
1925
	} else {
1926
		unsigned int maclen = skb_network_offset(skb);
1927
		netoff = TPACKET_ALIGN(po->tp_hdrlen +
1928 1929
				       (maclen < 16 ? 16 : maclen)) +
			po->tp_reserve;
L
Linus Torvalds 已提交
1930 1931
		macoff = netoff - maclen;
	}
1932 1933 1934
	if (po->tp_version <= TPACKET_V2) {
		if (macoff + snaplen > po->rx_ring.frame_size) {
			if (po->copy_thresh &&
E
Eric Dumazet 已提交
1935
			    atomic_read(&sk->sk_rmem_alloc) < sk->sk_rcvbuf) {
1936 1937 1938 1939 1940 1941 1942 1943
				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
1944
			}
1945 1946 1947
			snaplen = po->rx_ring.frame_size - macoff;
			if ((int)snaplen < 0)
				snaplen = 0;
L
Linus Torvalds 已提交
1948 1949 1950
		}
	}
	spin_lock(&sk->sk_receive_queue.lock);
1951 1952
	h.raw = packet_current_rx_frame(po, skb,
					TP_STATUS_KERNEL, (macoff+snaplen));
1953
	if (!h.raw)
L
Linus Torvalds 已提交
1954
		goto ring_is_full;
1955 1956 1957 1958 1959 1960 1961 1962
	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.
	 */
1963
		if (po->stats.stats1.tp_drops)
1964 1965
			status |= TP_STATUS_LOSING;
	}
1966
	po->stats.stats1.tp_packets++;
L
Linus Torvalds 已提交
1967 1968 1969 1970 1971 1972
	if (copy_skb) {
		status |= TP_STATUS_COPY;
		__skb_queue_tail(&sk->sk_receive_queue, copy_skb);
	}
	spin_unlock(&sk->sk_receive_queue.lock);

1973
	skb_copy_bits(skb, 0, h.raw + macoff, snaplen);
1974 1975

	if (!(ts_status = tpacket_get_timestamp(skb, &ts, po->tp_tstamp)))
1976
		getnstimeofday(&ts);
L
Linus Torvalds 已提交
1977

1978 1979
	status |= ts_status;

1980 1981 1982 1983 1984 1985
	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;
1986 1987
		h.h1->tp_sec = ts.tv_sec;
		h.h1->tp_usec = ts.tv_nsec / NSEC_PER_USEC;
1988 1989 1990 1991 1992 1993 1994 1995 1996
		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;
1997 1998
		if (vlan_tx_tag_present(skb)) {
			h.h2->tp_vlan_tci = vlan_tx_tag_get(skb);
1999 2000
			h.h2->tp_vlan_tpid = ntohs(skb->vlan_proto);
			status |= TP_STATUS_VLAN_VALID | TP_STATUS_VLAN_TPID_VALID;
2001 2002
		} else {
			h.h2->tp_vlan_tci = 0;
2003
			h.h2->tp_vlan_tpid = 0;
2004
		}
2005
		memset(h.h2->tp_padding, 0, sizeof(h.h2->tp_padding));
2006 2007
		hdrlen = sizeof(*h.h2);
		break;
2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
	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;
2019
		memset(h.h3->tp_padding, 0, sizeof(h.h3->tp_padding));
2020 2021
		hdrlen = sizeof(*h.h3);
		break;
2022 2023 2024
	default:
		BUG();
	}
L
Linus Torvalds 已提交
2025

2026
	sll = h.raw + TPACKET_ALIGN(hdrlen);
S
Stephen Hemminger 已提交
2027
	sll->sll_halen = dev_parse_header(skb, sll->sll_addr);
L
Linus Torvalds 已提交
2028 2029 2030 2031
	sll->sll_family = AF_PACKET;
	sll->sll_hatype = dev->type;
	sll->sll_protocol = skb->protocol;
	sll->sll_pkttype = skb->pkt_type;
2032
	if (unlikely(po->origdev))
2033 2034 2035
		sll->sll_ifindex = orig_dev->ifindex;
	else
		sll->sll_ifindex = dev->ifindex;
L
Linus Torvalds 已提交
2036

R
Ralf Baechle 已提交
2037
	smp_mb();
2038

2039
#if ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE == 1
2040
	if (po->tp_version <= TPACKET_V2) {
2041 2042
		u8 *start, *end;

2043 2044 2045 2046 2047
		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));
L
Linus Torvalds 已提交
2048
	}
2049
	smp_wmb();
2050
#endif
2051

2052 2053 2054 2055
	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
2056

2057
	sk->sk_data_ready(sk);
L
Linus Torvalds 已提交
2058 2059 2060 2061 2062 2063 2064

drop_n_restore:
	if (skb_head != skb->data && skb_shared(skb)) {
		skb->data = skb_head;
		skb->len = skb_len;
	}
drop:
2065
	kfree_skb(skb);
L
Linus Torvalds 已提交
2066 2067 2068
	return 0;

ring_is_full:
2069
	po->stats.stats1.tp_drops++;
L
Linus Torvalds 已提交
2070 2071
	spin_unlock(&sk->sk_receive_queue.lock);

2072
	sk->sk_data_ready(sk);
2073
	kfree_skb(copy_skb);
L
Linus Torvalds 已提交
2074 2075 2076
	goto drop_n_restore;
}

J
Johann Baudy 已提交
2077 2078 2079
static void tpacket_destruct_skb(struct sk_buff *skb)
{
	struct packet_sock *po = pkt_sk(skb->sk);
L
Linus Torvalds 已提交
2080

J
Johann Baudy 已提交
2081
	if (likely(po->tx_ring.pg_vec)) {
2082
		void *ph;
2083 2084
		__u32 ts;

J
Johann Baudy 已提交
2085
		ph = skb_shinfo(skb)->destructor_arg;
2086
		packet_dec_pending(&po->tx_ring);
2087 2088 2089

		ts = __packet_set_timestamp(po, ph, skb);
		__packet_set_status(po, ph, TP_STATUS_AVAILABLE | ts);
J
Johann Baudy 已提交
2090 2091 2092 2093 2094
	}

	sock_wfree(skb);
}

E
Eric Dumazet 已提交
2095 2096
static int tpacket_fill_skb(struct packet_sock *po, struct sk_buff *skb,
		void *frame, struct net_device *dev, int size_max,
2097
		__be16 proto, unsigned char *addr, int hlen)
J
Johann Baudy 已提交
2098
{
2099
	union tpacket_uhdr ph;
2100
	int to_write, offset, len, tp_len, nr_frags, len_max;
J
Johann Baudy 已提交
2101 2102 2103 2104 2105 2106 2107 2108 2109 2110
	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;
2111
	skb->mark = po->sk.sk_mark;
2112
	sock_tx_timestamp(&po->sk, &skb_shinfo(skb)->tx_flags);
J
Johann Baudy 已提交
2113 2114 2115 2116 2117 2118 2119 2120 2121 2122
	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;
	}
2123 2124 2125 2126
	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
2127

2128
	skb_reserve(skb, hlen);
J
Johann Baudy 已提交
2129
	skb_reset_network_header(skb);
2130

2131 2132 2133
	if (!packet_use_direct_xmit(po))
		skb_probe_transport_header(skb, 0);
	if (unlikely(po->tp_tx_has_off)) {
2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161
		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
2162 2163 2164 2165 2166 2167 2168
	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
2169
	} else if (dev->hard_header_len) {
J
Johann Baudy 已提交
2170 2171
		/* net device doesn't like empty head */
		if (unlikely(tp_len <= dev->hard_header_len)) {
E
Eric Dumazet 已提交
2172 2173
			pr_err("packet size is too short (%d < %d)\n",
			       tp_len, dev->hard_header_len);
J
Johann Baudy 已提交
2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199
			return -EINVAL;
		}

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

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

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

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

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

		if (unlikely(nr_frags >= MAX_SKB_FRAGS)) {
E
Eric Dumazet 已提交
2200 2201
			pr_err("Packet exceed the number of skb frags(%lu)\n",
			       MAX_SKB_FRAGS);
J
Johann Baudy 已提交
2202 2203 2204
			return -EFAULT;
		}

2205 2206
		page = pgv_to_page(data);
		data += len;
J
Johann Baudy 已提交
2207 2208
		flush_dcache_page(page);
		get_page(page);
2209
		skb_fill_page_desc(skb, nr_frags, page, offset, len);
J
Johann Baudy 已提交
2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223
		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;
2224
	int err, reserve = 0;
E
Eric Dumazet 已提交
2225
	void *ph;
2226
	DECLARE_SOCKADDR(struct sockaddr_ll *, saddr, msg->msg_name);
2227
	bool need_wait = !(msg->msg_flags & MSG_DONTWAIT);
J
Johann Baudy 已提交
2228 2229 2230
	int tp_len, size_max;
	unsigned char *addr;
	int len_sum = 0;
2231
	int status = TP_STATUS_AVAILABLE;
2232
	int hlen, tlen;
J
Johann Baudy 已提交
2233 2234 2235

	mutex_lock(&po->pg_vec_lock);

2236
	if (likely(saddr == NULL)) {
2237
		dev	= packet_cached_dev_get(po);
J
Johann Baudy 已提交
2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249
		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;
2250
		dev = dev_get_by_index(sock_net(&po->sk), saddr->sll_ifindex);
J
Johann Baudy 已提交
2251 2252 2253 2254 2255 2256 2257 2258 2259
	}

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

2260
	reserve = dev->hard_header_len + VLAN_HLEN;
J
Johann Baudy 已提交
2261
	size_max = po->tx_ring.frame_size
2262
		- (po->tp_hdrlen - sizeof(struct sockaddr_ll));
J
Johann Baudy 已提交
2263

2264 2265 2266
	if (size_max > dev->mtu + reserve)
		size_max = dev->mtu + reserve;

J
Johann Baudy 已提交
2267 2268
	do {
		ph = packet_current_frame(po, &po->tx_ring,
2269
					  TP_STATUS_SEND_REQUEST);
J
Johann Baudy 已提交
2270
		if (unlikely(ph == NULL)) {
2271 2272
			if (need_wait && need_resched())
				schedule();
J
Johann Baudy 已提交
2273 2274 2275 2276
			continue;
		}

		status = TP_STATUS_SEND_REQUEST;
2277 2278
		hlen = LL_RESERVED_SPACE(dev);
		tlen = dev->needed_tailroom;
J
Johann Baudy 已提交
2279
		skb = sock_alloc_send_skb(&po->sk,
2280
				hlen + tlen + sizeof(struct sockaddr_ll),
J
Johann Baudy 已提交
2281 2282 2283 2284 2285 2286
				0, &err);

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

		tp_len = tpacket_fill_skb(po, skb, ph, dev, size_max, proto,
2287 2288 2289 2290 2291 2292 2293
					  addr, hlen);
		if (tp_len > dev->mtu + dev->hard_header_len) {
			struct ethhdr *ehdr;
			/* Earlier code assumed this would be a VLAN pkt,
			 * double-check this now that we have the actual
			 * packet in hand.
			 */
J
Johann Baudy 已提交
2294

2295 2296 2297 2298 2299
			skb_reset_mac_header(skb);
			ehdr = eth_hdr(skb);
			if (ehdr->h_proto != htons(ETH_P_8021Q))
				tp_len = -EMSGSIZE;
		}
J
Johann Baudy 已提交
2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313
		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;
			}
		}

2314 2315
		packet_pick_tx_queue(dev, skb);

J
Johann Baudy 已提交
2316 2317
		skb->destructor = tpacket_destruct_skb;
		__packet_set_status(po, ph, TP_STATUS_SENDING);
2318
		packet_inc_pending(&po->tx_ring);
J
Johann Baudy 已提交
2319 2320

		status = TP_STATUS_SEND_REQUEST;
2321
		err = po->xmit(skb);
2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335
		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
2336 2337
		packet_increment_head(&po->tx_ring);
		len_sum += tp_len;
2338 2339 2340 2341 2342 2343 2344 2345
	} while (likely((ph != NULL) ||
		/* Note: packet_read_pending() might be slow if we have
		 * to call it as it's per_cpu variable, but in fast-path
		 * we already short-circuit the loop with the first
		 * condition, and luckily don't have to go that path
		 * anyway.
		 */
		 (need_wait && packet_read_pending(&po->tx_ring))));
J
Johann Baudy 已提交
2346 2347 2348 2349 2350 2351 2352 2353

	err = len_sum;
	goto out_put;

out_status:
	__packet_set_status(po, ph, status);
	kfree_skb(skb);
out_put:
2354
	dev_put(dev);
J
Johann Baudy 已提交
2355 2356 2357 2358 2359
out:
	mutex_unlock(&po->pg_vec_lock);
	return err;
}

2360 2361 2362 2363
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)
2364 2365 2366 2367 2368 2369 2370 2371
{
	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,
2372
				   err, 0);
2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383
	if (!skb)
		return NULL;

	skb_reserve(skb, reserve);
	skb_put(skb, linear);
	skb->data_len = len - linear;
	skb->len += len - linear;

	return skb;
}

2384
static int packet_snd(struct socket *sock, struct msghdr *msg, size_t len)
L
Linus Torvalds 已提交
2385 2386
{
	struct sock *sk = sock->sk;
2387
	DECLARE_SOCKADDR(struct sockaddr_ll *, saddr, msg->msg_name);
L
Linus Torvalds 已提交
2388 2389
	struct sk_buff *skb;
	struct net_device *dev;
A
Al Viro 已提交
2390
	__be16 proto;
L
Linus Torvalds 已提交
2391
	unsigned char *addr;
2392
	int err, reserve = 0;
2393 2394 2395 2396 2397
	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;
2398
	int hlen, tlen;
2399
	int extra_len = 0;
L
Linus Torvalds 已提交
2400 2401

	/*
2402
	 *	Get and verify the address.
L
Linus Torvalds 已提交
2403
	 */
2404

2405
	if (likely(saddr == NULL)) {
2406
		dev	= packet_cached_dev_get(po);
L
Linus Torvalds 已提交
2407 2408 2409 2410 2411 2412
		proto	= po->num;
		addr	= NULL;
	} else {
		err = -EINVAL;
		if (msg->msg_namelen < sizeof(struct sockaddr_ll))
			goto out;
2413 2414
		if (msg->msg_namelen < (saddr->sll_halen + offsetof(struct sockaddr_ll, sll_addr)))
			goto out;
L
Linus Torvalds 已提交
2415 2416
		proto	= saddr->sll_protocol;
		addr	= saddr->sll_addr;
2417
		dev = dev_get_by_index(sock_net(sk), saddr->sll_ifindex);
L
Linus Torvalds 已提交
2418 2419 2420
	}

	err = -ENXIO;
2421
	if (unlikely(dev == NULL))
L
Linus Torvalds 已提交
2422
		goto out_unlock;
2423
	err = -ENETDOWN;
2424
	if (unlikely(!(dev->flags & IFF_UP)))
2425 2426
		goto out_unlock;

2427 2428
	if (sock->type == SOCK_RAW)
		reserve = dev->hard_header_len;
2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476
	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;

		}
	}

2477 2478 2479 2480 2481 2482 2483 2484
	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
2485
	err = -EMSGSIZE;
2486
	if (!gso_type && (len > dev->mtu + reserve + VLAN_HLEN + extra_len))
L
Linus Torvalds 已提交
2487 2488
		goto out_unlock;

2489
	err = -ENOBUFS;
2490 2491 2492
	hlen = LL_RESERVED_SPACE(dev);
	tlen = dev->needed_tailroom;
	skb = packet_alloc_skb(sk, hlen + tlen, hlen, len, vnet_hdr.hdr_len,
2493
			       msg->msg_flags & MSG_DONTWAIT, &err);
E
Eric Dumazet 已提交
2494
	if (skb == NULL)
L
Linus Torvalds 已提交
2495 2496
		goto out_unlock;

2497
	skb_set_network_header(skb, reserve);
L
Linus Torvalds 已提交
2498

2499 2500
	err = -EINVAL;
	if (sock->type == SOCK_DGRAM &&
2501
	    (offset = dev_hard_header(skb, dev, ntohs(proto), addr, NULL, len)) < 0)
2502
		goto out_free;
L
Linus Torvalds 已提交
2503 2504

	/* Returns -EFAULT on error */
2505
	err = skb_copy_datagram_from_iovec(skb, offset, msg->msg_iov, 0, len);
L
Linus Torvalds 已提交
2506 2507
	if (err)
		goto out_free;
2508 2509

	sock_tx_timestamp(sk, &skb_shinfo(skb)->tx_flags);
L
Linus Torvalds 已提交
2510

2511
	if (!gso_type && (len > dev->mtu + reserve + extra_len)) {
2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522
		/* Earlier code assumed this would be a VLAN pkt,
		 * double-check this now that we have the actual
		 * packet in hand.
		 */
		struct ethhdr *ehdr;
		skb_reset_mac_header(skb);
		ehdr = eth_hdr(skb);
		if (ehdr->h_proto != htons(ETH_P_8021Q)) {
			err = -EMSGSIZE;
			goto out_free;
		}
2523 2524
	}

2525 2526
	skb->protocol = proto;
	skb->dev = dev;
L
Linus Torvalds 已提交
2527
	skb->priority = sk->sk_priority;
2528
	skb->mark = sk->sk_mark;
2529 2530

	packet_pick_tx_queue(dev, skb);
L
Linus Torvalds 已提交
2531

2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550
	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;
	}

2551 2552
	if (!packet_use_direct_xmit(po))
		skb_probe_transport_header(skb, reserve);
2553 2554 2555
	if (unlikely(extra_len == 4))
		skb->no_fcs = 1;

2556
	err = po->xmit(skb);
L
Linus Torvalds 已提交
2557 2558 2559
	if (err > 0 && (err = net_xmit_errno(err)) != 0)
		goto out_unlock;

2560
	dev_put(dev);
L
Linus Torvalds 已提交
2561

E
Eric Dumazet 已提交
2562
	return len;
L
Linus Torvalds 已提交
2563 2564 2565 2566

out_free:
	kfree_skb(skb);
out_unlock:
2567
	if (dev)
L
Linus Torvalds 已提交
2568 2569 2570 2571 2572
		dev_put(dev);
out:
	return err;
}

J
Johann Baudy 已提交
2573 2574 2575 2576 2577
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);
2578

J
Johann Baudy 已提交
2579 2580 2581 2582 2583 2584
	if (po->tx_ring.pg_vec)
		return tpacket_snd(po, msg);
	else
		return packet_snd(sock, msg, len);
}

L
Linus Torvalds 已提交
2585 2586 2587 2588 2589 2590 2591 2592 2593
/*
 *	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;
2594
	struct net *net;
2595
	union tpacket_req_u req_u;
L
Linus Torvalds 已提交
2596 2597 2598 2599

	if (!sk)
		return 0;

2600
	net = sock_net(sk);
L
Linus Torvalds 已提交
2601 2602
	po = pkt_sk(sk);

2603
	mutex_lock(&net->packet.sklist_lock);
2604
	sk_del_node_init_rcu(sk);
2605 2606 2607
	mutex_unlock(&net->packet.sklist_lock);

	preempt_disable();
2608
	sock_prot_inuse_add(net, sk->sk_prot, -1);
2609
	preempt_enable();
L
Linus Torvalds 已提交
2610

2611
	spin_lock(&po->bind_lock);
2612
	unregister_prot_hook(sk, false);
2613 2614
	packet_cached_dev_reset(po);

2615 2616 2617 2618
	if (po->prot_hook.dev) {
		dev_put(po->prot_hook.dev);
		po->prot_hook.dev = NULL;
	}
2619
	spin_unlock(&po->bind_lock);
L
Linus Torvalds 已提交
2620 2621 2622

	packet_flush_mclist(sk);

2623 2624
	if (po->rx_ring.pg_vec) {
		memset(&req_u, 0, sizeof(req_u));
2625
		packet_set_ring(sk, &req_u, 1, 0);
2626
	}
J
Johann Baudy 已提交
2627

2628 2629
	if (po->tx_ring.pg_vec) {
		memset(&req_u, 0, sizeof(req_u));
2630
		packet_set_ring(sk, &req_u, 1, 1);
2631
	}
L
Linus Torvalds 已提交
2632

D
David S. Miller 已提交
2633 2634
	fanout_release(sk);

2635
	synchronize_net();
L
Linus Torvalds 已提交
2636 2637 2638 2639 2640 2641 2642 2643 2644
	/*
	 *	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);
2645
	packet_free_pending(po);
2646
	sk_refcnt_debug_release(sk);
L
Linus Torvalds 已提交
2647 2648 2649 2650 2651 2652 2653 2654 2655

	sock_put(sk);
	return 0;
}

/*
 *	Attach a packet hook.
 */

2656
static int packet_do_bind(struct sock *sk, struct net_device *dev, __be16 proto)
L
Linus Torvalds 已提交
2657 2658
{
	struct packet_sock *po = pkt_sk(sk);
2659 2660 2661
	const struct net_device *dev_curr;
	__be16 proto_curr;
	bool need_rehook;
D
David S. Miller 已提交
2662

2663 2664 2665 2666
	if (po->fanout) {
		if (dev)
			dev_put(dev);

D
David S. Miller 已提交
2667
		return -EINVAL;
2668
	}
L
Linus Torvalds 已提交
2669 2670 2671

	lock_sock(sk);
	spin_lock(&po->bind_lock);
2672

2673 2674 2675 2676 2677 2678 2679
	proto_curr = po->prot_hook.type;
	dev_curr = po->prot_hook.dev;

	need_rehook = proto_curr != proto || dev_curr != dev;

	if (need_rehook) {
		unregister_prot_hook(sk, true);
L
Linus Torvalds 已提交
2680

2681 2682
		po->num = proto;
		po->prot_hook.type = proto;
L
Linus Torvalds 已提交
2683

2684 2685 2686 2687 2688 2689 2690 2691
		if (po->prot_hook.dev)
			dev_put(po->prot_hook.dev);

		po->prot_hook.dev = dev;

		po->ifindex = dev ? dev->ifindex : 0;
		packet_cached_dev_assign(po, dev);
	}
2692

2693
	if (proto == 0 || !need_rehook)
L
Linus Torvalds 已提交
2694 2695
		goto out_unlock;

2696
	if (!dev || (dev->flags & IFF_UP)) {
2697
		register_prot_hook(sk);
2698 2699 2700 2701
	} else {
		sk->sk_err = ENETDOWN;
		if (!sock_flag(sk, SOCK_DEAD))
			sk->sk_error_report(sk);
L
Linus Torvalds 已提交
2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713
	}

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
2714 2715
static int packet_bind_spkt(struct socket *sock, struct sockaddr *uaddr,
			    int addr_len)
L
Linus Torvalds 已提交
2716
{
E
Eric Dumazet 已提交
2717
	struct sock *sk = sock->sk;
L
Linus Torvalds 已提交
2718 2719 2720
	char name[15];
	struct net_device *dev;
	int err = -ENODEV;
2721

L
Linus Torvalds 已提交
2722 2723 2724
	/*
	 *	Check legality
	 */
2725

2726
	if (addr_len != sizeof(struct sockaddr))
L
Linus Torvalds 已提交
2727
		return -EINVAL;
E
Eric Dumazet 已提交
2728
	strlcpy(name, uaddr->sa_data, sizeof(name));
L
Linus Torvalds 已提交
2729

2730
	dev = dev_get_by_name(sock_net(sk), name);
2731
	if (dev)
L
Linus Torvalds 已提交
2732 2733 2734 2735 2736 2737
		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
2738 2739
	struct sockaddr_ll *sll = (struct sockaddr_ll *)uaddr;
	struct sock *sk = sock->sk;
L
Linus Torvalds 已提交
2740 2741 2742 2743 2744 2745 2746
	struct net_device *dev = NULL;
	int err;


	/*
	 *	Check legality
	 */
2747

L
Linus Torvalds 已提交
2748 2749 2750 2751 2752 2753 2754
	if (addr_len < sizeof(struct sockaddr_ll))
		return -EINVAL;
	if (sll->sll_family != AF_PACKET)
		return -EINVAL;

	if (sll->sll_ifindex) {
		err = -ENODEV;
2755
		dev = dev_get_by_index(sock_net(sk), sll->sll_ifindex);
L
Linus Torvalds 已提交
2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771
		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),
};

/*
2772
 *	Create a packet of type SOCK_PACKET.
L
Linus Torvalds 已提交
2773 2774
 */

2775 2776
static int packet_create(struct net *net, struct socket *sock, int protocol,
			 int kern)
L
Linus Torvalds 已提交
2777 2778 2779
{
	struct sock *sk;
	struct packet_sock *po;
A
Al Viro 已提交
2780
	__be16 proto = (__force __be16)protocol; /* weird, but documented */
L
Linus Torvalds 已提交
2781 2782
	int err;

2783
	if (!ns_capable(net->user_ns, CAP_NET_RAW))
L
Linus Torvalds 已提交
2784
		return -EPERM;
2785 2786
	if (sock->type != SOCK_DGRAM && sock->type != SOCK_RAW &&
	    sock->type != SOCK_PACKET)
L
Linus Torvalds 已提交
2787 2788 2789 2790 2791
		return -ESOCKTNOSUPPORT;

	sock->state = SS_UNCONNECTED;

	err = -ENOBUFS;
2792
	sk = sk_alloc(net, PF_PACKET, GFP_KERNEL, &packet_proto);
L
Linus Torvalds 已提交
2793 2794 2795 2796 2797 2798
	if (sk == NULL)
		goto out;

	sock->ops = &packet_ops;
	if (sock->type == SOCK_PACKET)
		sock->ops = &packet_ops_spkt;
2799

L
Linus Torvalds 已提交
2800 2801 2802 2803
	sock_init_data(sock, sk);

	po = pkt_sk(sk);
	sk->sk_family = PF_PACKET;
A
Al Viro 已提交
2804
	po->num = proto;
2805
	po->xmit = dev_queue_xmit;
2806

2807 2808 2809 2810
	err = packet_alloc_pending(po);
	if (err)
		goto out2;

2811
	packet_cached_dev_reset(po);
L
Linus Torvalds 已提交
2812 2813

	sk->sk_destruct = packet_sock_destruct;
2814
	sk_refcnt_debug_inc(sk);
L
Linus Torvalds 已提交
2815 2816 2817 2818 2819 2820

	/*
	 *	Attach a protocol block
	 */

	spin_lock_init(&po->bind_lock);
2821
	mutex_init(&po->pg_vec_lock);
L
Linus Torvalds 已提交
2822
	po->prot_hook.func = packet_rcv;
2823

L
Linus Torvalds 已提交
2824 2825
	if (sock->type == SOCK_PACKET)
		po->prot_hook.func = packet_rcv_spkt;
2826

L
Linus Torvalds 已提交
2827 2828
	po->prot_hook.af_packet_priv = sk;

A
Al Viro 已提交
2829 2830
	if (proto) {
		po->prot_hook.type = proto;
2831
		register_prot_hook(sk);
L
Linus Torvalds 已提交
2832 2833
	}

2834
	mutex_lock(&net->packet.sklist_lock);
2835
	sk_add_node_rcu(sk, &net->packet.sklist);
2836 2837 2838
	mutex_unlock(&net->packet.sklist_lock);

	preempt_disable();
2839
	sock_prot_inuse_add(net, &packet_proto, 1);
2840
	preempt_enable();
2841

E
Eric Dumazet 已提交
2842
	return 0;
2843 2844
out2:
	sk_free(sk);
L
Linus Torvalds 已提交
2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859
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;
2860
	int vnet_hdr_len = 0;
L
Linus Torvalds 已提交
2861 2862

	err = -EINVAL;
2863
	if (flags & ~(MSG_PEEK|MSG_DONTWAIT|MSG_TRUNC|MSG_CMSG_COMPAT|MSG_ERRQUEUE))
L
Linus Torvalds 已提交
2864 2865 2866 2867 2868 2869 2870 2871
		goto out;

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

2872
	if (flags & MSG_ERRQUEUE) {
2873 2874
		err = sock_recv_errqueue(sk, msg, len,
					 SOL_PACKET, PACKET_TX_TIMESTAMP);
2875 2876 2877
		goto out;
	}

L
Linus Torvalds 已提交
2878 2879 2880 2881 2882 2883 2884 2885 2886
	/*
	 *	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
2887
	skb = skb_recv_datagram(sk, flags, flags & MSG_DONTWAIT, &err);
L
Linus Torvalds 已提交
2888 2889

	/*
2890
	 *	An error occurred so return it. Because skb_recv_datagram()
L
Linus Torvalds 已提交
2891 2892 2893 2894
	 *	handles the blocking we don't see and worry about blocking
	 *	retries.
	 */

2895
	if (skb == NULL)
L
Linus Torvalds 已提交
2896 2897
		goto out;

2898 2899 2900 2901 2902
	if (pkt_sk(sk)->has_vnet_hdr) {
		struct virtio_net_hdr vnet_hdr = { 0 };

		err = -EINVAL;
		vnet_hdr_len = sizeof(vnet_hdr);
2903
		if (len < vnet_hdr_len)
2904 2905
			goto out_free;

2906 2907
		len -= vnet_hdr_len;

2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930
		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;
2931
			vnet_hdr.csum_start = skb_checksum_start_offset(skb);
2932
			vnet_hdr.csum_offset = skb->csum_offset;
2933 2934
		} else if (skb->ip_summed == CHECKSUM_UNNECESSARY) {
			vnet_hdr.flags = VIRTIO_NET_HDR_F_DATA_VALID;
2935 2936 2937 2938 2939 2940 2941 2942
		} /* else everything is zero */

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

2943 2944 2945
	/* You lose any data beyond the buffer you gave. If it worries
	 * a user program they can ask the device for its MTU
	 * anyway.
L
Linus Torvalds 已提交
2946 2947
	 */
	copied = skb->len;
E
Eric Dumazet 已提交
2948 2949 2950
	if (copied > len) {
		copied = len;
		msg->msg_flags |= MSG_TRUNC;
L
Linus Torvalds 已提交
2951 2952 2953 2954 2955 2956
	}

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

2957
	sock_recv_ts_and_drops(msg, sk, skb);
L
Linus Torvalds 已提交
2958

2959 2960 2961 2962 2963
	if (msg->msg_name) {
		/* If the address length field is there to be filled
		 * in, we fill it in now.
		 */
		if (sock->type == SOCK_PACKET) {
2964
			__sockaddr_check_size(sizeof(struct sockaddr_pkt));
2965 2966 2967 2968 2969 2970
			msg->msg_namelen = sizeof(struct sockaddr_pkt);
		} else {
			struct sockaddr_ll *sll = &PACKET_SKB_CB(skb)->sa.ll;
			msg->msg_namelen = sll->sll_halen +
				offsetof(struct sockaddr_ll, sll_addr);
		}
2971 2972
		memcpy(msg->msg_name, &PACKET_SKB_CB(skb)->sa,
		       msg->msg_namelen);
2973
	}
L
Linus Torvalds 已提交
2974

2975
	if (pkt_sk(sk)->auxdata) {
2976 2977 2978 2979 2980 2981 2982 2983
		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;
2984
		aux.tp_net = skb_network_offset(skb);
2985 2986
		if (vlan_tx_tag_present(skb)) {
			aux.tp_vlan_tci = vlan_tx_tag_get(skb);
2987 2988
			aux.tp_vlan_tpid = ntohs(skb->vlan_proto);
			aux.tp_status |= TP_STATUS_VLAN_VALID | TP_STATUS_VLAN_TPID_VALID;
2989 2990
		} else {
			aux.tp_vlan_tci = 0;
2991
			aux.tp_vlan_tpid = 0;
2992
		}
2993
		put_cmsg(msg, SOL_PACKET, PACKET_AUXDATA, sizeof(aux), &aux);
2994 2995
	}

L
Linus Torvalds 已提交
2996 2997 2998 2999
	/*
	 *	Free or return the buffer as appropriate. Again this
	 *	hides all the races and re-entrancy issues from us.
	 */
3000
	err = vnet_hdr_len + ((flags&MSG_TRUNC) ? skb->len : copied);
L
Linus Torvalds 已提交
3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017

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;
3018
	memset(uaddr->sa_data, 0, sizeof(uaddr->sa_data));
E
Eric Dumazet 已提交
3019 3020 3021
	rcu_read_lock();
	dev = dev_get_by_index_rcu(sock_net(sk), pkt_sk(sk)->ifindex);
	if (dev)
3022
		strlcpy(uaddr->sa_data, dev->name, sizeof(uaddr->sa_data));
E
Eric Dumazet 已提交
3023
	rcu_read_unlock();
L
Linus Torvalds 已提交
3024 3025 3026 3027 3028 3029 3030 3031 3032 3033 3034
	*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);
3035
	DECLARE_SOCKADDR(struct sockaddr_ll *, sll, uaddr);
L
Linus Torvalds 已提交
3036 3037 3038 3039 3040 3041 3042

	if (peer)
		return -EOPNOTSUPP;

	sll->sll_family = AF_PACKET;
	sll->sll_ifindex = po->ifindex;
	sll->sll_protocol = po->num;
3043
	sll->sll_pkttype = 0;
E
Eric Dumazet 已提交
3044 3045
	rcu_read_lock();
	dev = dev_get_by_index_rcu(sock_net(sk), po->ifindex);
L
Linus Torvalds 已提交
3046 3047 3048 3049 3050 3051 3052 3053
	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
3054
	rcu_read_unlock();
3055
	*uaddr_len = offsetof(struct sockaddr_ll, sll_addr) + sll->sll_halen;
L
Linus Torvalds 已提交
3056 3057 3058 3059

	return 0;
}

3060 3061
static int packet_dev_mc(struct net_device *dev, struct packet_mclist *i,
			 int what)
L
Linus Torvalds 已提交
3062 3063 3064
{
	switch (i->type) {
	case PACKET_MR_MULTICAST:
3065 3066
		if (i->alen != dev->addr_len)
			return -EINVAL;
L
Linus Torvalds 已提交
3067
		if (what > 0)
3068
			return dev_mc_add(dev, i->addr);
L
Linus Torvalds 已提交
3069
		else
3070
			return dev_mc_del(dev, i->addr);
L
Linus Torvalds 已提交
3071 3072
		break;
	case PACKET_MR_PROMISC:
3073
		return dev_set_promiscuity(dev, what);
L
Linus Torvalds 已提交
3074
	case PACKET_MR_ALLMULTI:
3075
		return dev_set_allmulti(dev, what);
3076
	case PACKET_MR_UNICAST:
3077 3078
		if (i->alen != dev->addr_len)
			return -EINVAL;
3079
		if (what > 0)
3080
			return dev_uc_add(dev, i->addr);
3081
		else
3082
			return dev_uc_del(dev, i->addr);
3083
		break;
E
Eric Dumazet 已提交
3084 3085
	default:
		break;
L
Linus Torvalds 已提交
3086
	}
3087
	return 0;
L
Linus Torvalds 已提交
3088 3089 3090 3091
}

static void packet_dev_mclist(struct net_device *dev, struct packet_mclist *i, int what)
{
E
Eric Dumazet 已提交
3092
	for ( ; i; i = i->next) {
L
Linus Torvalds 已提交
3093 3094 3095 3096 3097
		if (i->ifindex == dev->ifindex)
			packet_dev_mc(dev, i, what);
	}
}

3098
static int packet_mc_add(struct sock *sk, struct packet_mreq_max *mreq)
L
Linus Torvalds 已提交
3099 3100 3101 3102 3103 3104 3105 3106 3107
{
	struct packet_sock *po = pkt_sk(sk);
	struct packet_mclist *ml, *i;
	struct net_device *dev;
	int err;

	rtnl_lock();

	err = -ENODEV;
3108
	dev = __dev_get_by_index(sock_net(sk), mreq->mr_ifindex);
L
Linus Torvalds 已提交
3109 3110 3111 3112
	if (!dev)
		goto done;

	err = -EINVAL;
3113
	if (mreq->mr_alen > dev->addr_len)
L
Linus Torvalds 已提交
3114 3115 3116
		goto done;

	err = -ENOBUFS;
3117
	i = kmalloc(sizeof(*i), GFP_KERNEL);
L
Linus Torvalds 已提交
3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140
	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;
3141 3142 3143 3144 3145
	err = packet_dev_mc(dev, i, 1);
	if (err) {
		po->mclist = i->next;
		kfree(i);
	}
L
Linus Torvalds 已提交
3146 3147 3148 3149 3150 3151

done:
	rtnl_unlock();
	return err;
}

3152
static int packet_mc_drop(struct sock *sk, struct packet_mreq_max *mreq)
L
Linus Torvalds 已提交
3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165
{
	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;
3166 3167
				dev = __dev_get_by_index(sock_net(sk), ml->ifindex);
				if (dev)
L
Linus Torvalds 已提交
3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191
					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;
3192 3193
		dev = __dev_get_by_index(sock_net(sk), ml->ifindex);
		if (dev != NULL)
L
Linus Torvalds 已提交
3194 3195 3196 3197 3198 3199 3200
			packet_dev_mc(dev, ml, -1);
		kfree(ml);
	}
	rtnl_unlock();
}

static int
3201
packet_setsockopt(struct socket *sock, int level, int optname, char __user *optval, unsigned int optlen)
L
Linus Torvalds 已提交
3202 3203
{
	struct sock *sk = sock->sk;
3204
	struct packet_sock *po = pkt_sk(sk);
L
Linus Torvalds 已提交
3205 3206 3207 3208 3209
	int ret;

	if (level != SOL_PACKET)
		return -ENOPROTOOPT;

J
Johann Baudy 已提交
3210
	switch (optname) {
3211
	case PACKET_ADD_MEMBERSHIP:
L
Linus Torvalds 已提交
3212 3213
	case PACKET_DROP_MEMBERSHIP:
	{
3214 3215 3216 3217
		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
3218
			return -EINVAL;
3219 3220
		if (len > sizeof(mreq))
			len = sizeof(mreq);
E
Eric Dumazet 已提交
3221
		if (copy_from_user(&mreq, optval, len))
L
Linus Torvalds 已提交
3222
			return -EFAULT;
3223 3224
		if (len < (mreq.mr_alen + offsetof(struct packet_mreq, mr_address)))
			return -EINVAL;
L
Linus Torvalds 已提交
3225 3226 3227 3228 3229 3230
		if (optname == PACKET_ADD_MEMBERSHIP)
			ret = packet_mc_add(sk, &mreq);
		else
			ret = packet_mc_drop(sk, &mreq);
		return ret;
	}
3231

L
Linus Torvalds 已提交
3232
	case PACKET_RX_RING:
J
Johann Baudy 已提交
3233
	case PACKET_TX_RING:
L
Linus Torvalds 已提交
3234
	{
3235 3236
		union tpacket_req_u req_u;
		int len;
L
Linus Torvalds 已提交
3237

3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248
		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
3249
			return -EINVAL;
3250 3251
		if (pkt_sk(sk)->has_vnet_hdr)
			return -EINVAL;
3252
		if (copy_from_user(&req_u.req, optval, len))
L
Linus Torvalds 已提交
3253
			return -EFAULT;
3254 3255
		return packet_set_ring(sk, &req_u, 0,
			optname == PACKET_TX_RING);
L
Linus Torvalds 已提交
3256 3257 3258 3259 3260
	}
	case PACKET_COPY_THRESH:
	{
		int val;

E
Eric Dumazet 已提交
3261
		if (optlen != sizeof(val))
L
Linus Torvalds 已提交
3262
			return -EINVAL;
E
Eric Dumazet 已提交
3263
		if (copy_from_user(&val, optval, sizeof(val)))
L
Linus Torvalds 已提交
3264 3265 3266 3267 3268
			return -EFAULT;

		pkt_sk(sk)->copy_thresh = val;
		return 0;
	}
3269 3270 3271 3272 3273 3274
	case PACKET_VERSION:
	{
		int val;

		if (optlen != sizeof(val))
			return -EINVAL;
J
Johann Baudy 已提交
3275
		if (po->rx_ring.pg_vec || po->tx_ring.pg_vec)
3276 3277 3278 3279 3280 3281
			return -EBUSY;
		if (copy_from_user(&val, optval, sizeof(val)))
			return -EFAULT;
		switch (val) {
		case TPACKET_V1:
		case TPACKET_V2:
3282
		case TPACKET_V3:
3283 3284 3285 3286 3287 3288
			po->tp_version = val;
			return 0;
		default:
			return -EINVAL;
		}
	}
3289 3290 3291 3292 3293 3294
	case PACKET_RESERVE:
	{
		unsigned int val;

		if (optlen != sizeof(val))
			return -EINVAL;
J
Johann Baudy 已提交
3295
		if (po->rx_ring.pg_vec || po->tx_ring.pg_vec)
3296 3297 3298 3299 3300 3301
			return -EBUSY;
		if (copy_from_user(&val, optval, sizeof(val)))
			return -EFAULT;
		po->tp_reserve = val;
		return 0;
	}
J
Johann Baudy 已提交
3302 3303 3304 3305 3306 3307 3308 3309 3310 3311 3312 3313 3314
	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;
	}
3315 3316 3317 3318 3319 3320 3321 3322 3323 3324 3325 3326
	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;
	}
3327 3328 3329 3330 3331 3332 3333 3334 3335 3336 3337 3338
	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;
	}
3339 3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 3353 3354
	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;
	}
3355 3356 3357 3358 3359 3360 3361 3362 3363 3364 3365 3366
	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
3367 3368 3369 3370 3371 3372 3373 3374 3375 3376 3377
	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);
	}
3378 3379 3380 3381 3382 3383 3384 3385 3386 3387 3388 3389 3390
	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;
	}
3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402
	case PACKET_QDISC_BYPASS:
	{
		int val;

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

		po->xmit = val ? packet_direct_xmit : dev_queue_xmit;
		return 0;
	}
L
Linus Torvalds 已提交
3403 3404 3405 3406 3407 3408 3409 3410 3411
	default:
		return -ENOPROTOOPT;
	}
}

static int packet_getsockopt(struct socket *sock, int level, int optname,
			     char __user *optval, int __user *optlen)
{
	int len;
3412
	int val, lv = sizeof(val);
L
Linus Torvalds 已提交
3413 3414
	struct sock *sk = sock->sk;
	struct packet_sock *po = pkt_sk(sk);
3415
	void *data = &val;
3416
	union tpacket_stats_u st;
L
Linus Torvalds 已提交
3417 3418 3419 3420

	if (level != SOL_PACKET)
		return -ENOPROTOOPT;

3421 3422
	if (get_user(len, optlen))
		return -EFAULT;
L
Linus Torvalds 已提交
3423 3424 3425

	if (len < 0)
		return -EINVAL;
3426

J
Johann Baudy 已提交
3427
	switch (optname) {
L
Linus Torvalds 已提交
3428 3429
	case PACKET_STATISTICS:
		spin_lock_bh(&sk->sk_receive_queue.lock);
3430 3431 3432 3433
		memcpy(&st, &po->stats, sizeof(st));
		memset(&po->stats, 0, sizeof(po->stats));
		spin_unlock_bh(&sk->sk_receive_queue.lock);

3434
		if (po->tp_version == TPACKET_V3) {
3435
			lv = sizeof(struct tpacket_stats_v3);
3436
			st.stats3.tp_packets += st.stats3.tp_drops;
3437
			data = &st.stats3;
3438
		} else {
3439
			lv = sizeof(struct tpacket_stats);
3440
			st.stats1.tp_packets += st.stats1.tp_drops;
3441
			data = &st.stats1;
3442
		}
3443

3444 3445 3446
		break;
	case PACKET_AUXDATA:
		val = po->auxdata;
3447 3448 3449
		break;
	case PACKET_ORIGDEV:
		val = po->origdev;
3450 3451 3452
		break;
	case PACKET_VNET_HDR:
		val = po->has_vnet_hdr;
L
Linus Torvalds 已提交
3453
		break;
3454 3455 3456 3457 3458 3459 3460 3461 3462 3463 3464 3465 3466 3467 3468
	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;
3469 3470 3471
		case TPACKET_V3:
			val = sizeof(struct tpacket3_hdr);
			break;
3472 3473 3474 3475
		default:
			return -EINVAL;
		}
		break;
3476 3477 3478
	case PACKET_RESERVE:
		val = po->tp_reserve;
		break;
J
Johann Baudy 已提交
3479 3480 3481
	case PACKET_LOSS:
		val = po->tp_loss;
		break;
3482 3483 3484
	case PACKET_TIMESTAMP:
		val = po->tp_tstamp;
		break;
D
David S. Miller 已提交
3485 3486 3487
	case PACKET_FANOUT:
		val = (po->fanout ?
		       ((u32)po->fanout->id |
3488 3489
			((u32)po->fanout->type << 16) |
			((u32)po->fanout->flags << 24)) :
D
David S. Miller 已提交
3490 3491
		       0);
		break;
3492 3493 3494
	case PACKET_TX_HAS_OFF:
		val = po->tp_tx_has_off;
		break;
3495 3496 3497
	case PACKET_QDISC_BYPASS:
		val = packet_use_direct_xmit(po);
		break;
L
Linus Torvalds 已提交
3498 3499 3500 3501
	default:
		return -ENOPROTOOPT;
	}

3502 3503
	if (len > lv)
		len = lv;
3504 3505
	if (put_user(len, optlen))
		return -EFAULT;
3506 3507
	if (copy_to_user(optval, data, len))
		return -EFAULT;
3508
	return 0;
L
Linus Torvalds 已提交
3509 3510 3511
}


3512 3513
static int packet_notifier(struct notifier_block *this,
			   unsigned long msg, void *ptr)
L
Linus Torvalds 已提交
3514 3515
{
	struct sock *sk;
3516
	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
3517
	struct net *net = dev_net(dev);
L
Linus Torvalds 已提交
3518

3519
	rcu_read_lock();
3520
	sk_for_each_rcu(sk, &net->packet.sklist) {
L
Linus Torvalds 已提交
3521 3522 3523 3524 3525 3526
		struct packet_sock *po = pkt_sk(sk);

		switch (msg) {
		case NETDEV_UNREGISTER:
			if (po->mclist)
				packet_dev_mclist(dev, po->mclist, -1);
3527 3528
			/* fallthrough */

L
Linus Torvalds 已提交
3529 3530 3531 3532
		case NETDEV_DOWN:
			if (dev->ifindex == po->ifindex) {
				spin_lock(&po->bind_lock);
				if (po->running) {
3533
					__unregister_prot_hook(sk, false);
L
Linus Torvalds 已提交
3534 3535 3536 3537 3538
					sk->sk_err = ENETDOWN;
					if (!sock_flag(sk, SOCK_DEAD))
						sk->sk_error_report(sk);
				}
				if (msg == NETDEV_UNREGISTER) {
3539
					packet_cached_dev_reset(po);
L
Linus Torvalds 已提交
3540
					po->ifindex = -1;
3541 3542
					if (po->prot_hook.dev)
						dev_put(po->prot_hook.dev);
L
Linus Torvalds 已提交
3543 3544 3545 3546 3547 3548
					po->prot_hook.dev = NULL;
				}
				spin_unlock(&po->bind_lock);
			}
			break;
		case NETDEV_UP:
3549 3550
			if (dev->ifindex == po->ifindex) {
				spin_lock(&po->bind_lock);
3551 3552
				if (po->num)
					register_prot_hook(sk);
3553
				spin_unlock(&po->bind_lock);
L
Linus Torvalds 已提交
3554 3555 3556 3557
			}
			break;
		}
	}
3558
	rcu_read_unlock();
L
Linus Torvalds 已提交
3559 3560 3561 3562 3563 3564 3565 3566 3567
	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
3568
	switch (cmd) {
E
Eric Dumazet 已提交
3569 3570 3571
	case SIOCOUTQ:
	{
		int amount = sk_wmem_alloc_get(sk);
3572

E
Eric Dumazet 已提交
3573 3574 3575 3576 3577 3578 3579 3580 3581 3582 3583 3584 3585 3586 3587 3588 3589 3590
		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);
3591

L
Linus Torvalds 已提交
3592
#ifdef CONFIG_INET
E
Eric Dumazet 已提交
3593 3594 3595 3596 3597 3598 3599 3600 3601 3602 3603 3604 3605 3606 3607
	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
3608 3609
#endif

E
Eric Dumazet 已提交
3610 3611
	default:
		return -ENOIOCTLCMD;
L
Linus Torvalds 已提交
3612 3613 3614 3615
	}
	return 0;
}

E
Eric Dumazet 已提交
3616
static unsigned int packet_poll(struct file *file, struct socket *sock,
L
Linus Torvalds 已提交
3617 3618 3619 3620 3621 3622 3623
				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
3624
	if (po->rx_ring.pg_vec) {
3625 3626
		if (!packet_previous_rx_frame(po, &po->rx_ring,
			TP_STATUS_KERNEL))
L
Linus Torvalds 已提交
3627 3628 3629
			mask |= POLLIN | POLLRDNORM;
	}
	spin_unlock_bh(&sk->sk_receive_queue.lock);
J
Johann Baudy 已提交
3630 3631 3632 3633 3634 3635
	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
3636 3637 3638 3639 3640 3641 3642 3643 3644 3645 3646
	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
3647
	struct socket *sock = file->private_data;
L
Linus Torvalds 已提交
3648
	struct sock *sk = sock->sk;
3649

L
Linus Torvalds 已提交
3650 3651 3652 3653 3654 3655 3656
	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
3657
	struct socket *sock = file->private_data;
L
Linus Torvalds 已提交
3658
	struct sock *sk = sock->sk;
3659

L
Linus Torvalds 已提交
3660 3661 3662 3663
	if (sk)
		atomic_dec(&pkt_sk(sk)->mapped);
}

3664
static const struct vm_operations_struct packet_mmap_ops = {
E
Eric Dumazet 已提交
3665 3666
	.open	=	packet_mm_open,
	.close	=	packet_mm_close,
L
Linus Torvalds 已提交
3667 3668
};

3669 3670
static void free_pg_vec(struct pgv *pg_vec, unsigned int order,
			unsigned int len)
L
Linus Torvalds 已提交
3671 3672 3673
{
	int i;

3674
	for (i = 0; i < len; i++) {
3675
		if (likely(pg_vec[i].buffer)) {
C
Changli Gao 已提交
3676
			if (is_vmalloc_addr(pg_vec[i].buffer))
3677 3678 3679 3680 3681 3682
				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
3683 3684 3685 3686
	}
	kfree(pg_vec);
}

3687
static char *alloc_one_pg_vec_page(unsigned long order)
3688
{
3689
	char *buffer;
3690 3691 3692 3693 3694 3695 3696
	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;

3697
	/* __get_free_pages failed, fall back to vmalloc */
E
Eric Dumazet 已提交
3698
	buffer = vzalloc((1 << order) * PAGE_SIZE);
3699 3700 3701
	if (buffer)
		return buffer;

3702
	/* vmalloc failed, lets dig into swap here */
3703
	gfp_flags &= ~__GFP_NORETRY;
3704
	buffer = (char *) __get_free_pages(gfp_flags, order);
3705 3706 3707
	if (buffer)
		return buffer;

3708
	/* complete and utter failure */
3709
	return NULL;
3710 3711
}

3712
static struct pgv *alloc_pg_vec(struct tpacket_req *req, int order)
3713 3714
{
	unsigned int block_nr = req->tp_block_nr;
3715
	struct pgv *pg_vec;
3716 3717
	int i;

3718
	pg_vec = kcalloc(block_nr, sizeof(struct pgv), GFP_KERNEL);
3719 3720 3721 3722
	if (unlikely(!pg_vec))
		goto out;

	for (i = 0; i < block_nr; i++) {
C
Changli Gao 已提交
3723
		pg_vec[i].buffer = alloc_one_pg_vec_page(order);
3724
		if (unlikely(!pg_vec[i].buffer))
3725 3726 3727 3728 3729 3730 3731 3732 3733 3734 3735
			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
3736

3737
static int packet_set_ring(struct sock *sk, union tpacket_req_u *req_u,
J
Johann Baudy 已提交
3738
		int closing, int tx_ring)
L
Linus Torvalds 已提交
3739
{
3740
	struct pgv *pg_vec = NULL;
L
Linus Torvalds 已提交
3741
	struct packet_sock *po = pkt_sk(sk);
A
Al Viro 已提交
3742
	int was_running, order = 0;
J
Johann Baudy 已提交
3743 3744
	struct packet_ring_buffer *rb;
	struct sk_buff_head *rb_queue;
A
Al Viro 已提交
3745
	__be16 num;
3746 3747 3748 3749 3750 3751 3752 3753 3754
	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;
	}
3755

J
Johann Baudy 已提交
3756 3757
	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
3758

J
Johann Baudy 已提交
3759 3760 3761 3762
	err = -EBUSY;
	if (!closing) {
		if (atomic_read(&po->mapped))
			goto out;
3763
		if (packet_read_pending(rb))
J
Johann Baudy 已提交
3764 3765
			goto out;
	}
L
Linus Torvalds 已提交
3766

J
Johann Baudy 已提交
3767 3768 3769 3770 3771
	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
3772

3773 3774 3775 3776 3777 3778 3779
		switch (po->tp_version) {
		case TPACKET_V1:
			po->tp_hdrlen = TPACKET_HDRLEN;
			break;
		case TPACKET_V2:
			po->tp_hdrlen = TPACKET2_HDRLEN;
			break;
3780 3781 3782
		case TPACKET_V3:
			po->tp_hdrlen = TPACKET3_HDRLEN;
			break;
3783 3784
		}

J
Johann Baudy 已提交
3785
		err = -EINVAL;
3786
		if (unlikely((int)req->tp_block_size <= 0))
J
Johann Baudy 已提交
3787
			goto out;
3788
		if (unlikely(req->tp_block_size & (PAGE_SIZE - 1)))
J
Johann Baudy 已提交
3789
			goto out;
3790
		if (unlikely(req->tp_frame_size < po->tp_hdrlen +
J
Johann Baudy 已提交
3791 3792
					po->tp_reserve))
			goto out;
3793
		if (unlikely(req->tp_frame_size & (TPACKET_ALIGNMENT - 1)))
J
Johann Baudy 已提交
3794
			goto out;
L
Linus Torvalds 已提交
3795

J
Johann Baudy 已提交
3796 3797 3798 3799 3800 3801
		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
3802 3803

		err = -ENOMEM;
3804 3805 3806
		order = get_order(req->tp_block_size);
		pg_vec = alloc_pg_vec(req, order);
		if (unlikely(!pg_vec))
L
Linus Torvalds 已提交
3807
			goto out;
3808 3809 3810 3811 3812 3813 3814
		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);
3815
			break;
3816 3817 3818
		default:
			break;
		}
J
Johann Baudy 已提交
3819 3820 3821 3822
	}
	/* Done */
	else {
		err = -EINVAL;
3823
		if (unlikely(req->tp_frame_nr))
J
Johann Baudy 已提交
3824
			goto out;
L
Linus Torvalds 已提交
3825 3826 3827 3828 3829 3830 3831 3832 3833 3834
	}

	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;
3835
		__unregister_prot_hook(sk, false);
L
Linus Torvalds 已提交
3836 3837
	}
	spin_unlock(&po->bind_lock);
3838

L
Linus Torvalds 已提交
3839 3840 3841
	synchronize_net();

	err = -EBUSY;
3842
	mutex_lock(&po->pg_vec_lock);
L
Linus Torvalds 已提交
3843 3844
	if (closing || atomic_read(&po->mapped) == 0) {
		err = 0;
J
Johann Baudy 已提交
3845
		spin_lock_bh(&rb_queue->lock);
3846
		swap(rb->pg_vec, pg_vec);
J
Johann Baudy 已提交
3847 3848 3849 3850 3851
		rb->frame_max = (req->tp_frame_nr - 1);
		rb->head = 0;
		rb->frame_size = req->tp_frame_size;
		spin_unlock_bh(&rb_queue->lock);

3852 3853
		swap(rb->pg_vec_order, order);
		swap(rb->pg_vec_len, req->tp_block_nr);
J
Johann Baudy 已提交
3854 3855 3856 3857 3858

		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
3859
		if (atomic_read(&po->mapped))
E
Eric Dumazet 已提交
3860 3861
			pr_err("packet_mmap: vma is busy: %d\n",
			       atomic_read(&po->mapped));
L
Linus Torvalds 已提交
3862
	}
3863
	mutex_unlock(&po->pg_vec_lock);
L
Linus Torvalds 已提交
3864 3865

	spin_lock(&po->bind_lock);
3866
	if (was_running) {
L
Linus Torvalds 已提交
3867
		po->num = num;
3868
		register_prot_hook(sk);
L
Linus Torvalds 已提交
3869 3870
	}
	spin_unlock(&po->bind_lock);
3871 3872 3873 3874 3875
	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
3876 3877 3878 3879 3880 3881 3882 3883
	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
3884 3885
static int packet_mmap(struct file *file, struct socket *sock,
		struct vm_area_struct *vma)
L
Linus Torvalds 已提交
3886 3887 3888
{
	struct sock *sk = sock->sk;
	struct packet_sock *po = pkt_sk(sk);
J
Johann Baudy 已提交
3889 3890
	unsigned long size, expected_size;
	struct packet_ring_buffer *rb;
L
Linus Torvalds 已提交
3891 3892 3893 3894 3895 3896 3897
	unsigned long start;
	int err = -EINVAL;
	int i;

	if (vma->vm_pgoff)
		return -EINVAL;

3898
	mutex_lock(&po->pg_vec_lock);
J
Johann Baudy 已提交
3899 3900 3901 3902 3903 3904 3905 3906 3907 3908 3909

	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
3910
		goto out;
J
Johann Baudy 已提交
3911 3912 3913

	size = vma->vm_end - vma->vm_start;
	if (size != expected_size)
L
Linus Torvalds 已提交
3914 3915 3916
		goto out;

	start = vma->vm_start;
J
Johann Baudy 已提交
3917 3918 3919 3920 3921
	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++) {
3922 3923
			struct page *page;
			void *kaddr = rb->pg_vec[i].buffer;
J
Johann Baudy 已提交
3924 3925
			int pg_num;

C
Changli Gao 已提交
3926 3927
			for (pg_num = 0; pg_num < rb->pg_vec_pages; pg_num++) {
				page = pgv_to_page(kaddr);
J
Johann Baudy 已提交
3928 3929 3930 3931
				err = vm_insert_page(vma, start, page);
				if (unlikely(err))
					goto out;
				start += PAGE_SIZE;
3932
				kaddr += PAGE_SIZE;
J
Johann Baudy 已提交
3933
			}
3934
		}
L
Linus Torvalds 已提交
3935
	}
J
Johann Baudy 已提交
3936

3937
	atomic_inc(&po->mapped);
L
Linus Torvalds 已提交
3938 3939 3940 3941
	vma->vm_ops = &packet_mmap_ops;
	err = 0;

out:
3942
	mutex_unlock(&po->pg_vec_lock);
L
Linus Torvalds 已提交
3943 3944 3945
	return err;
}

3946
static const struct proto_ops packet_ops_spkt = {
L
Linus Torvalds 已提交
3947 3948 3949 3950 3951 3952 3953 3954 3955 3956 3957 3958 3959 3960 3961 3962 3963 3964 3965 3966
	.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,
};

3967
static const struct proto_ops packet_ops = {
L
Linus Torvalds 已提交
3968 3969 3970 3971 3972 3973 3974
	.family =	PF_PACKET,
	.owner =	THIS_MODULE,
	.release =	packet_release,
	.bind =		packet_bind,
	.connect =	sock_no_connect,
	.socketpair =	sock_no_socketpair,
	.accept =	sock_no_accept,
3975
	.getname =	packet_getname,
L
Linus Torvalds 已提交
3976 3977 3978 3979 3980 3981 3982 3983 3984 3985 3986 3987
	.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,
};

3988
static const struct net_proto_family packet_family_ops = {
L
Linus Torvalds 已提交
3989 3990 3991 3992 3993 3994
	.family =	PF_PACKET,
	.create =	packet_create,
	.owner	=	THIS_MODULE,
};

static struct notifier_block packet_netdev_notifier = {
E
Eric Dumazet 已提交
3995
	.notifier_call =	packet_notifier,
L
Linus Torvalds 已提交
3996 3997 3998 3999 4000
};

#ifdef CONFIG_PROC_FS

static void *packet_seq_start(struct seq_file *seq, loff_t *pos)
4001
	__acquires(RCU)
L
Linus Torvalds 已提交
4002
{
4003
	struct net *net = seq_file_net(seq);
4004 4005 4006

	rcu_read_lock();
	return seq_hlist_start_head_rcu(&net->packet.sklist, *pos);
L
Linus Torvalds 已提交
4007 4008 4009 4010
}

static void *packet_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
4011
	struct net *net = seq_file_net(seq);
4012
	return seq_hlist_next_rcu(v, &net->packet.sklist, pos);
L
Linus Torvalds 已提交
4013 4014 4015
}

static void packet_seq_stop(struct seq_file *seq, void *v)
4016
	__releases(RCU)
L
Linus Torvalds 已提交
4017
{
4018
	rcu_read_unlock();
L
Linus Torvalds 已提交
4019 4020
}

4021
static int packet_seq_show(struct seq_file *seq, void *v)
L
Linus Torvalds 已提交
4022 4023 4024 4025
{
	if (v == SEQ_START_TOKEN)
		seq_puts(seq, "sk       RefCnt Type Proto  Iface R Rmem   User   Inode\n");
	else {
4026
		struct sock *s = sk_entry(v);
L
Linus Torvalds 已提交
4027 4028 4029
		const struct packet_sock *po = pkt_sk(s);

		seq_printf(seq,
D
Dan Rosenberg 已提交
4030
			   "%pK %-6d %-4d %04x   %-5d %1d %-6u %-6u %-6lu\n",
L
Linus Torvalds 已提交
4031 4032 4033 4034 4035 4036 4037
			   s,
			   atomic_read(&s->sk_refcnt),
			   s->sk_type,
			   ntohs(po->num),
			   po->ifindex,
			   po->running,
			   atomic_read(&s->sk_rmem_alloc),
4038
			   from_kuid_munged(seq_user_ns(seq), sock_i_uid(s)),
E
Eric Dumazet 已提交
4039
			   sock_i_ino(s));
L
Linus Torvalds 已提交
4040 4041 4042 4043 4044
	}

	return 0;
}

4045
static const struct seq_operations packet_seq_ops = {
L
Linus Torvalds 已提交
4046 4047 4048 4049 4050 4051 4052 4053
	.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)
{
4054 4055
	return seq_open_net(inode, file, &packet_seq_ops,
			    sizeof(struct seq_net_private));
L
Linus Torvalds 已提交
4056 4057
}

4058
static const struct file_operations packet_seq_fops = {
L
Linus Torvalds 已提交
4059 4060 4061 4062
	.owner		= THIS_MODULE,
	.open		= packet_seq_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
4063
	.release	= seq_release_net,
L
Linus Torvalds 已提交
4064 4065 4066 4067
};

#endif

4068
static int __net_init packet_net_init(struct net *net)
4069
{
4070
	mutex_init(&net->packet.sklist_lock);
4071
	INIT_HLIST_HEAD(&net->packet.sklist);
4072

4073
	if (!proc_create("packet", 0, net->proc_net, &packet_seq_fops))
4074 4075 4076 4077 4078
		return -ENOMEM;

	return 0;
}

4079
static void __net_exit packet_net_exit(struct net *net)
4080
{
4081
	remove_proc_entry("packet", net->proc_net);
4082 4083 4084 4085 4086 4087 4088 4089
}

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


L
Linus Torvalds 已提交
4090 4091 4092
static void __exit packet_exit(void)
{
	unregister_netdevice_notifier(&packet_netdev_notifier);
4093
	unregister_pernet_subsys(&packet_net_ops);
L
Linus Torvalds 已提交
4094 4095 4096 4097 4098 4099 4100 4101 4102 4103 4104 4105
	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);
4106
	register_pernet_subsys(&packet_net_ops);
L
Linus Torvalds 已提交
4107 4108 4109 4110 4111 4112 4113 4114 4115
	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);