icmp.c 28.3 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3
/*
 *	NET3:	Implementation of the ICMP protocol layer.
 *
4
 *		Alan Cox, <alan@lxorguk.ukuu.org.uk>
L
Linus Torvalds 已提交
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72
 *
 *	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.
 *
 *	Some of the function names and the icmp unreach table for this
 *	module were derived from [icmp.c 1.0.11 06/02/93] by
 *	Ross Biro, Fred N. van Kempen, Mark Evans, Alan Cox, Gerhard Koerting.
 *	Other than that this module is a complete rewrite.
 *
 *	Fixes:
 *	Clemens Fruhwirth	:	introduce global icmp rate limiting
 *					with icmp type masking ability instead
 *					of broken per type icmp timeouts.
 *		Mike Shaver	:	RFC1122 checks.
 *		Alan Cox	:	Multicast ping reply as self.
 *		Alan Cox	:	Fix atomicity lockup in ip_build_xmit
 *					call.
 *		Alan Cox	:	Added 216,128 byte paths to the MTU
 *					code.
 *		Martin Mares	:	RFC1812 checks.
 *		Martin Mares	:	Can be configured to follow redirects
 *					if acting as a router _without_ a
 *					routing protocol (RFC 1812).
 *		Martin Mares	:	Echo requests may be configured to
 *					be ignored (RFC 1812).
 *		Martin Mares	:	Limitation of ICMP error message
 *					transmit rate (RFC 1812).
 *		Martin Mares	:	TOS and Precedence set correctly
 *					(RFC 1812).
 *		Martin Mares	:	Now copying as much data from the
 *					original packet as we can without
 *					exceeding 576 bytes (RFC 1812).
 *	Willy Konynenberg	:	Transparent proxying support.
 *		Keith Owens	:	RFC1191 correction for 4.2BSD based
 *					path MTU bug.
 *		Thomas Quinot	:	ICMP Dest Unreach codes up to 15 are
 *					valid (RFC 1812).
 *		Andi Kleen	:	Check all packet lengths properly
 *					and moved all kfree_skb() up to
 *					icmp_rcv.
 *		Andi Kleen	:	Move the rate limit bookkeeping
 *					into the dest entry and use a token
 *					bucket filter (thanks to ANK). Make
 *					the rates sysctl configurable.
 *		Yu Tianli	:	Fixed two ugly bugs in icmp_send
 *					- IP option length was accounted wrongly
 *					- ICMP header length was not accounted
 *					  at all.
 *              Tristan Greaves :       Added sysctl option to ignore bogus
 *              			broadcast responses from broken routers.
 *
 * To Fix:
 *
 *	- Should use skb_pull() instead of all the manual checking.
 *	  This would also greatly simply some upper layer error handlers. --AK
 *
 */

#include <linux/module.h>
#include <linux/types.h>
#include <linux/jiffies.h>
#include <linux/kernel.h>
#include <linux/fcntl.h>
#include <linux/socket.h>
#include <linux/in.h>
#include <linux/inet.h>
73
#include <linux/inetdevice.h>
L
Linus Torvalds 已提交
74 75 76
#include <linux/netdevice.h>
#include <linux/string.h>
#include <linux/netfilter_ipv4.h>
77
#include <linux/slab.h>
L
Linus Torvalds 已提交
78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93
#include <net/snmp.h>
#include <net/ip.h>
#include <net/route.h>
#include <net/protocol.h>
#include <net/icmp.h>
#include <net/tcp.h>
#include <net/udp.h>
#include <net/raw.h>
#include <linux/skbuff.h>
#include <net/sock.h>
#include <linux/errno.h>
#include <linux/timer.h>
#include <linux/init.h>
#include <asm/system.h>
#include <asm/uaccess.h>
#include <net/checksum.h>
94
#include <net/xfrm.h>
95
#include <net/inet_common.h>
L
Linus Torvalds 已提交
96 97 98 99 100 101 102 103 104 105 106 107

/*
 *	Build xmit assembly blocks
 */

struct icmp_bxm {
	struct sk_buff *skb;
	int offset;
	int data_len;

	struct {
		struct icmphdr icmph;
108
		__be32	       times[3];
L
Linus Torvalds 已提交
109 110
	} data;
	int head_len;
111
	struct ip_options_data replyopts;
L
Linus Torvalds 已提交
112 113 114 115 116
};

/* An array of errno for error messages from dest unreach. */
/* RFC 1122: 3.2.2.1 States that NET_UNREACH, HOST_UNREACH and SR_FAILED MUST be considered 'transient errs'. */

117
const struct icmp_err icmp_err_convert[] = {
L
Linus Torvalds 已提交
118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182
	{
		.errno = ENETUNREACH,	/* ICMP_NET_UNREACH */
		.fatal = 0,
	},
	{
		.errno = EHOSTUNREACH,	/* ICMP_HOST_UNREACH */
		.fatal = 0,
	},
	{
		.errno = ENOPROTOOPT	/* ICMP_PROT_UNREACH */,
		.fatal = 1,
	},
	{
		.errno = ECONNREFUSED,	/* ICMP_PORT_UNREACH */
		.fatal = 1,
	},
	{
		.errno = EMSGSIZE,	/* ICMP_FRAG_NEEDED */
		.fatal = 0,
	},
	{
		.errno = EOPNOTSUPP,	/* ICMP_SR_FAILED */
		.fatal = 0,
	},
	{
		.errno = ENETUNREACH,	/* ICMP_NET_UNKNOWN */
		.fatal = 1,
	},
	{
		.errno = EHOSTDOWN,	/* ICMP_HOST_UNKNOWN */
		.fatal = 1,
	},
	{
		.errno = ENONET,	/* ICMP_HOST_ISOLATED */
		.fatal = 1,
	},
	{
		.errno = ENETUNREACH,	/* ICMP_NET_ANO	*/
		.fatal = 1,
	},
	{
		.errno = EHOSTUNREACH,	/* ICMP_HOST_ANO */
		.fatal = 1,
	},
	{
		.errno = ENETUNREACH,	/* ICMP_NET_UNR_TOS */
		.fatal = 0,
	},
	{
		.errno = EHOSTUNREACH,	/* ICMP_HOST_UNR_TOS */
		.fatal = 0,
	},
	{
		.errno = EHOSTUNREACH,	/* ICMP_PKT_FILTERED */
		.fatal = 1,
	},
	{
		.errno = EHOSTUNREACH,	/* ICMP_PREC_VIOLATION */
		.fatal = 1,
	},
	{
		.errno = EHOSTUNREACH,	/* ICMP_PREC_CUTOFF */
		.fatal = 1,
	},
};
E
Eric Dumazet 已提交
183
EXPORT_SYMBOL(icmp_err_convert);
L
Linus Torvalds 已提交
184 185 186 187 188 189 190 191 192 193

/*
 *	ICMP control array. This specifies what to do with each ICMP.
 */

struct icmp_control {
	void (*handler)(struct sk_buff *skb);
	short   error;		/* This ICMP is classed as an error message */
};

194
static const struct icmp_control icmp_pointers[NR_ICMP_TYPES+1];
L
Linus Torvalds 已提交
195 196 197 198 199 200 201 202

/*
 *	The ICMP socket(s). This is the most convenient way to flow control
 *	our ICMP output as well as maintain a clean interface throughout
 *	all layers. All Socketless IP sends will soon be gone.
 *
 *	On SMP we have one ICMP socket per-cpu.
 */
203 204 205 206
static struct sock *icmp_sk(struct net *net)
{
	return net->ipv4.icmp_sk[smp_processor_id()];
}
L
Linus Torvalds 已提交
207

208
static inline struct sock *icmp_xmit_lock(struct net *net)
L
Linus Torvalds 已提交
209
{
210 211
	struct sock *sk;

L
Linus Torvalds 已提交
212 213
	local_bh_disable();

214 215
	sk = icmp_sk(net);

216
	if (unlikely(!spin_trylock(&sk->sk_lock.slock))) {
L
Linus Torvalds 已提交
217 218 219 220
		/* This can happen if the output path signals a
		 * dst_link_failure() for an outgoing ICMP packet.
		 */
		local_bh_enable();
221
		return NULL;
L
Linus Torvalds 已提交
222
	}
223
	return sk;
L
Linus Torvalds 已提交
224 225
}

226
static inline void icmp_xmit_unlock(struct sock *sk)
L
Linus Torvalds 已提交
227
{
228
	spin_unlock_bh(&sk->sk_lock.slock);
L
Linus Torvalds 已提交
229 230 231 232 233 234
}

/*
 *	Send an ICMP frame.
 */

235
static inline bool icmpv4_xrlim_allow(struct net *net, struct rtable *rt,
236
		int type, int code)
L
Linus Torvalds 已提交
237
{
238
	struct dst_entry *dst = &rt->dst;
239
	bool rc = true;
L
Linus Torvalds 已提交
240 241 242 243 244 245 246 247 248 249

	if (type > NR_ICMP_TYPES)
		goto out;

	/* Don't limit PMTU discovery. */
	if (type == ICMP_DEST_UNREACH && code == ICMP_FRAG_NEEDED)
		goto out;

	/* No rate limit on loopback */
	if (dst->dev && (dst->dev->flags&IFF_LOOPBACK))
250
		goto out;
L
Linus Torvalds 已提交
251 252

	/* Limit if icmp type is enabled in ratemask. */
253 254 255 256 257 258
	if ((1 << type) & net->ipv4.sysctl_icmp_ratemask) {
		if (!rt->peer)
			rt_bind_peer(rt, 1);
		rc = inet_peer_xrlim_allow(rt->peer,
					   net->ipv4.sysctl_icmp_ratelimit);
	}
L
Linus Torvalds 已提交
259 260 261 262 263 264 265
out:
	return rc;
}

/*
 *	Maintain the counters used in the SNMP statistics for outgoing ICMP
 */
266
void icmp_out_count(struct net *net, unsigned char type)
L
Linus Torvalds 已提交
267
{
268
	ICMPMSGOUT_INC_STATS(net, type);
269
	ICMP_INC_STATS(net, ICMP_MIB_OUTMSGS);
L
Linus Torvalds 已提交
270 271 272 273 274 275 276 277 278 279
}

/*
 *	Checksum each fragment, and on the first include the headers and final
 *	checksum.
 */
static int icmp_glue_bits(void *from, char *to, int offset, int len, int odd,
			  struct sk_buff *skb)
{
	struct icmp_bxm *icmp_param = (struct icmp_bxm *)from;
280
	__wsum csum;
L
Linus Torvalds 已提交
281 282 283 284 285 286 287 288 289 290 291 292

	csum = skb_copy_and_csum_bits(icmp_param->skb,
				      icmp_param->offset + offset,
				      to, len, 0);

	skb->csum = csum_block_add(skb->csum, csum, odd);
	if (icmp_pointers[icmp_param->data.icmph.type].error)
		nf_ct_attach(skb, icmp_param->skb);
	return 0;
}

static void icmp_push_reply(struct icmp_bxm *icmp_param,
293
			    struct flowi4 *fl4,
294
			    struct ipcm_cookie *ipc, struct rtable **rt)
L
Linus Torvalds 已提交
295
{
296
	struct sock *sk;
L
Linus Torvalds 已提交
297 298
	struct sk_buff *skb;

299
	sk = icmp_sk(dev_net((*rt)->dst.dev));
300
	if (ip_append_data(sk, fl4, icmp_glue_bits, icmp_param,
301 302
			   icmp_param->data_len+icmp_param->head_len,
			   icmp_param->head_len,
E
Eric Dumazet 已提交
303 304
			   ipc, rt, MSG_DONTWAIT) < 0) {
		ICMP_INC_STATS_BH(sock_net(sk), ICMP_MIB_OUTERRORS);
305
		ip_flush_pending_frames(sk);
E
Eric Dumazet 已提交
306
	} else if ((skb = skb_peek(&sk->sk_write_queue)) != NULL) {
307
		struct icmphdr *icmph = icmp_hdr(skb);
308
		__wsum csum = 0;
L
Linus Torvalds 已提交
309 310
		struct sk_buff *skb1;

311
		skb_queue_walk(&sk->sk_write_queue, skb1) {
L
Linus Torvalds 已提交
312 313 314 315 316 317 318
			csum = csum_add(csum, skb1->csum);
		}
		csum = csum_partial_copy_nocheck((void *)&icmp_param->data,
						 (char *)icmph,
						 icmp_param->head_len, csum);
		icmph->checksum = csum_fold(csum);
		skb->ip_summed = CHECKSUM_NONE;
319
		ip_push_pending_frames(sk, fl4);
L
Linus Torvalds 已提交
320 321 322 323 324 325 326 327 328 329
	}
}

/*
 *	Driving logic for building and sending ICMP messages.
 */

static void icmp_reply(struct icmp_bxm *icmp_param, struct sk_buff *skb)
{
	struct ipcm_cookie ipc;
E
Eric Dumazet 已提交
330
	struct rtable *rt = skb_rtable(skb);
331
	struct net *net = dev_net(rt->dst.dev);
332
	struct flowi4 fl4;
333 334
	struct sock *sk;
	struct inet_sock *inet;
A
Al Viro 已提交
335
	__be32 daddr;
L
Linus Torvalds 已提交
336

337
	if (ip_options_echo(&icmp_param->replyopts.opt.opt, skb))
338
		return;
L
Linus Torvalds 已提交
339

340 341
	sk = icmp_xmit_lock(net);
	if (sk == NULL)
L
Linus Torvalds 已提交
342
		return;
343
	inet = inet_sk(sk);
L
Linus Torvalds 已提交
344 345 346

	icmp_param->data.icmph.checksum = 0;

347
	inet->tos = ip_hdr(skb)->tos;
348
	daddr = ipc.addr = ip_hdr(skb)->saddr;
L
Linus Torvalds 已提交
349
	ipc.opt = NULL;
350
	ipc.tx_flags = 0;
351 352 353 354
	if (icmp_param->replyopts.opt.opt.optlen) {
		ipc.opt = &icmp_param->replyopts.opt;
		if (ipc.opt->opt.srr)
			daddr = icmp_param->replyopts.opt.opt.faddr;
L
Linus Torvalds 已提交
355
	}
356 357 358 359 360 361 362 363 364
	memset(&fl4, 0, sizeof(fl4));
	fl4.daddr = daddr;
	fl4.saddr = rt->rt_spec_dst;
	fl4.flowi4_tos = RT_TOS(ip_hdr(skb)->tos);
	fl4.flowi4_proto = IPPROTO_ICMP;
	security_skb_classify_flow(skb, flowi4_to_flowi(&fl4));
	rt = ip_route_output_key(net, &fl4);
	if (IS_ERR(rt))
		goto out_unlock;
365
	if (icmpv4_xrlim_allow(net, rt, icmp_param->data.icmph.type,
L
Linus Torvalds 已提交
366
			       icmp_param->data.icmph.code))
367
		icmp_push_reply(icmp_param, &fl4, &ipc, &rt);
L
Linus Torvalds 已提交
368 369
	ip_rt_put(rt);
out_unlock:
370
	icmp_xmit_unlock(sk);
L
Linus Torvalds 已提交
371 372
}

373 374 375
static struct rtable *icmp_route_lookup(struct net *net,
					struct flowi4 *fl4,
					struct sk_buff *skb_in,
376
					const struct iphdr *iph,
377 378 379 380 381 382 383
					__be32 saddr, u8 tos,
					int type, int code,
					struct icmp_bxm *param)
{
	struct rtable *rt, *rt2;
	int err;

384 385 386 387 388 389 390 391 392 393
	memset(fl4, 0, sizeof(*fl4));
	fl4->daddr = (param->replyopts.opt.opt.srr ?
		      param->replyopts.opt.opt.faddr : iph->saddr);
	fl4->saddr = saddr;
	fl4->flowi4_tos = RT_TOS(tos);
	fl4->flowi4_proto = IPPROTO_ICMP;
	fl4->fl4_icmp_type = type;
	fl4->fl4_icmp_code = code;
	security_skb_classify_flow(skb_in, flowi4_to_flowi(fl4));
	rt = __ip_route_output_key(net, fl4);
394 395
	if (IS_ERR(rt))
		return rt;
396 397 398 399

	/* No need to clone since we're just using its address. */
	rt2 = rt;

400
	rt = (struct rtable *) xfrm_lookup(net, &rt->dst,
401
					   flowi4_to_flowi(fl4), NULL, 0);
402
	if (!IS_ERR(rt)) {
403 404
		if (rt != rt2)
			return rt;
405
	} else if (PTR_ERR(rt) == -EPERM) {
406
		rt = NULL;
407 408
	} else
		return rt;
409

410
	err = xfrm_decode_session_reverse(skb_in, flowi4_to_flowi(fl4), AF_INET);
411 412 413
	if (err)
		goto relookup_failed;

414 415
	if (inet_addr_type(net, fl4->saddr) == RTN_LOCAL) {
		rt2 = __ip_route_output_key(net, fl4);
416 417
		if (IS_ERR(rt2))
			err = PTR_ERR(rt2);
418
	} else {
419
		struct flowi4 fl4_2 = {};
420 421
		unsigned long orefdst;

422
		fl4_2.daddr = fl4->saddr;
423
		rt2 = ip_route_output_key(net, &fl4_2);
424 425
		if (IS_ERR(rt2)) {
			err = PTR_ERR(rt2);
426
			goto relookup_failed;
427
		}
428 429
		/* Ugh! */
		orefdst = skb_in->_skb_refdst; /* save old refdst */
430
		err = ip_route_input(skb_in, fl4->daddr, fl4->saddr,
431 432 433 434 435 436 437 438 439 440
				     RT_TOS(tos), rt2->dst.dev);

		dst_release(&rt2->dst);
		rt2 = skb_rtable(skb_in);
		skb_in->_skb_refdst = orefdst; /* restore old refdst */
	}

	if (err)
		goto relookup_failed;

441
	rt2 = (struct rtable *) xfrm_lookup(net, &rt2->dst,
442
					    flowi4_to_flowi(fl4), NULL,
443
					    XFRM_LOOKUP_ICMP);
444
	if (!IS_ERR(rt2)) {
445 446
		dst_release(&rt->dst);
		rt = rt2;
447 448 449 450 451 452 453
	} else if (PTR_ERR(rt2) == -EPERM) {
		if (rt)
			dst_release(&rt->dst);
		return rt2;
	} else {
		err = PTR_ERR(rt2);
		goto relookup_failed;
454 455 456 457 458 459 460 461
	}
	return rt;

relookup_failed:
	if (rt)
		return rt;
	return ERR_PTR(err);
}
L
Linus Torvalds 已提交
462 463 464 465 466 467 468 469 470 471 472 473

/*
 *	Send an ICMP message in response to a situation
 *
 *	RFC 1122: 3.2.2	MUST send at least the IP header and 8 bytes of header.
 *		  MAY send more (we do).
 *			MUST NOT change this header information.
 *			MUST NOT reply to a multicast/broadcast IP address.
 *			MUST NOT reply to a multicast/broadcast MAC address.
 *			MUST reply to only the first fragment.
 */

A
Al Viro 已提交
474
void icmp_send(struct sk_buff *skb_in, int type, int code, __be32 info)
L
Linus Torvalds 已提交
475 476 477 478
{
	struct iphdr *iph;
	int room;
	struct icmp_bxm icmp_param;
E
Eric Dumazet 已提交
479
	struct rtable *rt = skb_rtable(skb_in);
L
Linus Torvalds 已提交
480
	struct ipcm_cookie ipc;
481
	struct flowi4 fl4;
A
Al Viro 已提交
482
	__be32 saddr;
L
Linus Torvalds 已提交
483
	u8  tos;
484
	struct net *net;
485
	struct sock *sk;
L
Linus Torvalds 已提交
486 487 488

	if (!rt)
		goto out;
489
	net = dev_net(rt->dst.dev);
L
Linus Torvalds 已提交
490 491 492 493 494 495

	/*
	 *	Find the original header. It is expected to be valid, of course.
	 *	Check this, icmp_send is called from the most obscure devices
	 *	sometimes.
	 */
496
	iph = ip_hdr(skb_in);
L
Linus Torvalds 已提交
497

498 499
	if ((u8 *)iph < skb_in->head ||
	    (skb_in->network_header + sizeof(*iph)) > skb_in->tail)
L
Linus Torvalds 已提交
500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532
		goto out;

	/*
	 *	No replies to physical multicast/broadcast
	 */
	if (skb_in->pkt_type != PACKET_HOST)
		goto out;

	/*
	 *	Now check at the protocol level
	 */
	if (rt->rt_flags & (RTCF_BROADCAST | RTCF_MULTICAST))
		goto out;

	/*
	 *	Only reply to fragment 0. We byte re-order the constant
	 *	mask for efficiency.
	 */
	if (iph->frag_off & htons(IP_OFFSET))
		goto out;

	/*
	 *	If we send an ICMP error to an ICMP error a mess would result..
	 */
	if (icmp_pointers[type].error) {
		/*
		 *	We are an error, check if we are replying to an
		 *	ICMP error
		 */
		if (iph->protocol == IPPROTO_ICMP) {
			u8 _inner_type, *itp;

			itp = skb_header_pointer(skb_in,
533
						 skb_network_header(skb_in) +
L
Linus Torvalds 已提交
534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552
						 (iph->ihl << 2) +
						 offsetof(struct icmphdr,
							  type) -
						 skb_in->data,
						 sizeof(_inner_type),
						 &_inner_type);
			if (itp == NULL)
				goto out;

			/*
			 *	Assume any unknown ICMP type is an error. This
			 *	isn't specified by the RFC, but think about it..
			 */
			if (*itp > NR_ICMP_TYPES ||
			    icmp_pointers[*itp].error)
				goto out;
		}
	}

553 554
	sk = icmp_xmit_lock(net);
	if (sk == NULL)
L
Linus Torvalds 已提交
555 556 557 558 559 560 561
		return;

	/*
	 *	Construct source address and options.
	 */

	saddr = iph->daddr;
562
	if (!(rt->rt_flags & RTCF_LOCAL)) {
563 564
		struct net_device *dev = NULL;

565
		rcu_read_lock();
566 567
		if (rt_is_input_route(rt) &&
		    net->ipv4.sysctl_icmp_errors_use_inbound_ifaddr)
568
			dev = dev_get_by_index_rcu(net, rt->rt_iif);
569

570
		if (dev)
571
			saddr = inet_select_addr(dev, 0, RT_SCOPE_LINK);
572
		else
573
			saddr = 0;
574
		rcu_read_unlock();
575
	}
L
Linus Torvalds 已提交
576 577 578 579 580

	tos = icmp_pointers[type].error ? ((iph->tos & IPTOS_TOS_MASK) |
					   IPTOS_PREC_INTERNETCONTROL) :
					  iph->tos;

581
	if (ip_options_echo(&icmp_param.replyopts.opt.opt, skb_in))
582
		goto out_unlock;
L
Linus Torvalds 已提交
583 584 585 586 587 588 589 590 591 592 593


	/*
	 *	Prepare data for ICMP header.
	 */

	icmp_param.data.icmph.type	 = type;
	icmp_param.data.icmph.code	 = code;
	icmp_param.data.icmph.un.gateway = info;
	icmp_param.data.icmph.checksum	 = 0;
	icmp_param.skb	  = skb_in;
594
	icmp_param.offset = skb_network_offset(skb_in);
595
	inet_sk(sk)->tos = tos;
L
Linus Torvalds 已提交
596
	ipc.addr = iph->saddr;
597
	ipc.opt = &icmp_param.replyopts.opt;
598
	ipc.tx_flags = 0;
L
Linus Torvalds 已提交
599

600
	rt = icmp_route_lookup(net, &fl4, skb_in, iph, saddr, tos,
601 602 603
			       type, code, &icmp_param);
	if (IS_ERR(rt))
		goto out_unlock;
L
Linus Torvalds 已提交
604

605
	if (!icmpv4_xrlim_allow(net, rt, type, code))
L
Linus Torvalds 已提交
606 607 608 609
		goto ende;

	/* RFC says return as much as we can without exceeding 576 bytes. */

610
	room = dst_mtu(&rt->dst);
L
Linus Torvalds 已提交
611 612
	if (room > 576)
		room = 576;
613
	room -= sizeof(struct iphdr) + icmp_param.replyopts.opt.opt.optlen;
L
Linus Torvalds 已提交
614 615 616 617 618 619 620
	room -= sizeof(struct icmphdr);

	icmp_param.data_len = skb_in->len - icmp_param.offset;
	if (icmp_param.data_len > room)
		icmp_param.data_len = room;
	icmp_param.head_len = sizeof(struct icmphdr);

621
	icmp_push_reply(&icmp_param, &fl4, &ipc, &rt);
L
Linus Torvalds 已提交
622 623 624
ende:
	ip_rt_put(rt);
out_unlock:
625
	icmp_xmit_unlock(sk);
L
Linus Torvalds 已提交
626 627
out:;
}
E
Eric Dumazet 已提交
628
EXPORT_SYMBOL(icmp_send);
L
Linus Torvalds 已提交
629 630 631 632 633 634 635 636


/*
 *	Handle ICMP_DEST_UNREACH, ICMP_TIME_EXCEED, and ICMP_QUENCH.
 */

static void icmp_unreach(struct sk_buff *skb)
{
637
	const struct iphdr *iph;
L
Linus Torvalds 已提交
638 639
	struct icmphdr *icmph;
	int hash, protocol;
640
	const struct net_protocol *ipprot;
L
Linus Torvalds 已提交
641
	u32 info = 0;
642 643
	struct net *net;

E
Eric Dumazet 已提交
644
	net = dev_net(skb_dst(skb)->dev);
L
Linus Torvalds 已提交
645 646 647 648 649 650 651 652 653 654

	/*
	 *	Incomplete header ?
	 * 	Only checks for the IP header, there should be an
	 *	additional check for longer headers in upper levels.
	 */

	if (!pskb_may_pull(skb, sizeof(struct iphdr)))
		goto out_err;

655
	icmph = icmp_hdr(skb);
656
	iph   = (const struct iphdr *)skb->data;
L
Linus Torvalds 已提交
657 658 659 660 661 662 663 664 665 666 667 668 669

	if (iph->ihl < 5) /* Mangled header, drop. */
		goto out_err;

	if (icmph->type == ICMP_DEST_UNREACH) {
		switch (icmph->code & 15) {
		case ICMP_NET_UNREACH:
		case ICMP_HOST_UNREACH:
		case ICMP_PROT_UNREACH:
		case ICMP_PORT_UNREACH:
			break;
		case ICMP_FRAG_NEEDED:
			if (ipv4_config.no_pmtu_disc) {
670 671
				LIMIT_NETDEBUG(KERN_INFO "ICMP: %pI4: fragmentation needed and DF set.\n",
					       &iph->daddr);
L
Linus Torvalds 已提交
672
			} else {
673
				info = ip_rt_frag_needed(net, iph,
674 675
							 ntohs(icmph->un.frag.mtu),
							 skb->dev);
L
Linus Torvalds 已提交
676 677 678 679 680
				if (!info)
					goto out;
			}
			break;
		case ICMP_SR_FAILED:
681 682
			LIMIT_NETDEBUG(KERN_INFO "ICMP: %pI4: Source Route Failed.\n",
				       &iph->daddr);
L
Linus Torvalds 已提交
683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703
			break;
		default:
			break;
		}
		if (icmph->code > NR_ICMP_UNREACH)
			goto out;
	} else if (icmph->type == ICMP_PARAMETERPROB)
		info = ntohl(icmph->un.gateway) >> 24;

	/*
	 *	Throw it at our lower layers
	 *
	 *	RFC 1122: 3.2.2 MUST extract the protocol ID from the passed
	 *		  header.
	 *	RFC 1122: 3.2.2.1 MUST pass ICMP unreach messages to the
	 *		  transport layer.
	 *	RFC 1122: 3.2.2.2 MUST pass ICMP time expired messages to
	 *		  transport layer.
	 */

	/*
L
Lucas De Marchi 已提交
704
	 *	Check the other end isn't violating RFC 1122. Some routers send
L
Linus Torvalds 已提交
705 706 707 708 709
	 *	bogus responses to broadcast frames. If you see this message
	 *	first check your netmask matches at both ends, if it does then
	 *	get the other vendor to fix their kit.
	 */

710
	if (!net->ipv4.sysctl_icmp_ignore_bogus_error_responses &&
711
	    inet_addr_type(net, iph->daddr) == RTN_BROADCAST) {
L
Linus Torvalds 已提交
712
		if (net_ratelimit())
713
			printk(KERN_WARNING "%pI4 sent an invalid ICMP "
L
Linus Torvalds 已提交
714
					    "type %u, code %u "
715 716
					    "error to a broadcast: %pI4 on %s\n",
			       &ip_hdr(skb)->saddr,
L
Linus Torvalds 已提交
717
			       icmph->type, icmph->code,
718
			       &iph->daddr,
L
Linus Torvalds 已提交
719 720 721 722 723 724 725 726 727 728
			       skb->dev->name);
		goto out;
	}

	/* Checkin full IP header plus 8 bytes of protocol to
	 * avoid additional coding at protocol handlers.
	 */
	if (!pskb_may_pull(skb, iph->ihl * 4 + 8))
		goto out;

729
	iph = (const struct iphdr *)skb->data;
L
Linus Torvalds 已提交
730 731 732 733 734
	protocol = iph->protocol;

	/*
	 *	Deliver ICMP message to raw sockets. Pretty useless feature?
	 */
735
	raw_icmp_error(skb, protocol, info);
L
Linus Torvalds 已提交
736 737 738 739 740 741 742 743 744 745 746

	hash = protocol & (MAX_INET_PROTOS - 1);
	rcu_read_lock();
	ipprot = rcu_dereference(inet_protos[hash]);
	if (ipprot && ipprot->err_handler)
		ipprot->err_handler(skb, info);
	rcu_read_unlock();

out:
	return;
out_err:
747
	ICMP_INC_STATS_BH(net, ICMP_MIB_INERRORS);
L
Linus Torvalds 已提交
748 749 750 751 752 753 754 755 756 757
	goto out;
}


/*
 *	Handle ICMP_REDIRECT.
 */

static void icmp_redirect(struct sk_buff *skb)
{
758
	const struct iphdr *iph;
L
Linus Torvalds 已提交
759 760 761 762 763 764 765 766 767 768

	if (skb->len < sizeof(struct iphdr))
		goto out_err;

	/*
	 *	Get the copied header of the packet that caused the redirect
	 */
	if (!pskb_may_pull(skb, sizeof(struct iphdr)))
		goto out;

769
	iph = (const struct iphdr *)skb->data;
L
Linus Torvalds 已提交
770

771
	switch (icmp_hdr(skb)->code & 7) {
L
Linus Torvalds 已提交
772 773 774 775 776 777 778
	case ICMP_REDIR_NET:
	case ICMP_REDIR_NETTOS:
		/*
		 * As per RFC recommendations now handle it as a host redirect.
		 */
	case ICMP_REDIR_HOST:
	case ICMP_REDIR_HOSTTOS:
779
		ip_rt_redirect(ip_hdr(skb)->saddr, iph->daddr,
780
			       icmp_hdr(skb)->un.gateway,
781
			       iph->saddr, skb->dev);
L
Linus Torvalds 已提交
782
		break;
783
	}
L
Linus Torvalds 已提交
784 785 786
out:
	return;
out_err:
787
	ICMP_INC_STATS_BH(dev_net(skb->dev), ICMP_MIB_INERRORS);
L
Linus Torvalds 已提交
788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804
	goto out;
}

/*
 *	Handle ICMP_ECHO ("ping") requests.
 *
 *	RFC 1122: 3.2.2.6 MUST have an echo server that answers ICMP echo
 *		  requests.
 *	RFC 1122: 3.2.2.6 Data received in the ICMP_ECHO request MUST be
 *		  included in the reply.
 *	RFC 1812: 4.3.3.6 SHOULD have a config option for silently ignoring
 *		  echo requests, MUST have default=NOT.
 *	See also WRT handling of options once they are done and working.
 */

static void icmp_echo(struct sk_buff *skb)
{
805 806
	struct net *net;

E
Eric Dumazet 已提交
807
	net = dev_net(skb_dst(skb)->dev);
808
	if (!net->ipv4.sysctl_icmp_echo_ignore_all) {
L
Linus Torvalds 已提交
809 810
		struct icmp_bxm icmp_param;

811
		icmp_param.data.icmph	   = *icmp_hdr(skb);
L
Linus Torvalds 已提交
812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829
		icmp_param.data.icmph.type = ICMP_ECHOREPLY;
		icmp_param.skb		   = skb;
		icmp_param.offset	   = 0;
		icmp_param.data_len	   = skb->len;
		icmp_param.head_len	   = sizeof(struct icmphdr);
		icmp_reply(&icmp_param, skb);
	}
}

/*
 *	Handle ICMP Timestamp requests.
 *	RFC 1122: 3.2.2.8 MAY implement ICMP timestamp requests.
 *		  SHOULD be in the kernel for minimum random latency.
 *		  MUST be accurate to a few minutes.
 *		  MUST be updated at least at 15Hz.
 */
static void icmp_timestamp(struct sk_buff *skb)
{
830
	struct timespec tv;
L
Linus Torvalds 已提交
831 832 833 834 835 836 837 838 839 840
	struct icmp_bxm icmp_param;
	/*
	 *	Too short.
	 */
	if (skb->len < 4)
		goto out_err;

	/*
	 *	Fill in the current time as ms since midnight UT:
	 */
841 842 843
	getnstimeofday(&tv);
	icmp_param.data.times[1] = htonl((tv.tv_sec % 86400) * MSEC_PER_SEC +
					 tv.tv_nsec / NSEC_PER_MSEC);
L
Linus Torvalds 已提交
844 845 846
	icmp_param.data.times[2] = icmp_param.data.times[1];
	if (skb_copy_bits(skb, 0, &icmp_param.data.times[0], 4))
		BUG();
847
	icmp_param.data.icmph	   = *icmp_hdr(skb);
L
Linus Torvalds 已提交
848 849 850 851 852 853 854 855 856 857
	icmp_param.data.icmph.type = ICMP_TIMESTAMPREPLY;
	icmp_param.data.icmph.code = 0;
	icmp_param.skb		   = skb;
	icmp_param.offset	   = 0;
	icmp_param.data_len	   = 0;
	icmp_param.head_len	   = sizeof(struct icmphdr) + 12;
	icmp_reply(&icmp_param, skb);
out:
	return;
out_err:
E
Eric Dumazet 已提交
858
	ICMP_INC_STATS_BH(dev_net(skb_dst(skb)->dev), ICMP_MIB_INERRORS);
L
Linus Torvalds 已提交
859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906
	goto out;
}


/*
 *	Handle ICMP_ADDRESS_MASK requests.  (RFC950)
 *
 * RFC1122 (3.2.2.9).  A host MUST only send replies to
 * ADDRESS_MASK requests if it's been configured as an address mask
 * agent.  Receiving a request doesn't constitute implicit permission to
 * act as one. Of course, implementing this correctly requires (SHOULD)
 * a way to turn the functionality on and off.  Another one for sysctl(),
 * I guess. -- MS
 *
 * RFC1812 (4.3.3.9).	A router MUST implement it.
 *			A router SHOULD have switch turning it on/off.
 *		      	This switch MUST be ON by default.
 *
 * Gratuitous replies, zero-source replies are not implemented,
 * that complies with RFC. DO NOT implement them!!! All the idea
 * of broadcast addrmask replies as specified in RFC950 is broken.
 * The problem is that it is not uncommon to have several prefixes
 * on one physical interface. Moreover, addrmask agent can even be
 * not aware of existing another prefixes.
 * If source is zero, addrmask agent cannot choose correct prefix.
 * Gratuitous mask announcements suffer from the same problem.
 * RFC1812 explains it, but still allows to use ADDRMASK,
 * that is pretty silly. --ANK
 *
 * All these rules are so bizarre, that I removed kernel addrmask
 * support at all. It is wrong, it is obsolete, nobody uses it in
 * any case. --ANK
 *
 * Furthermore you can do it with a usermode address agent program
 * anyway...
 */

static void icmp_address(struct sk_buff *skb)
{
#if 0
	if (net_ratelimit())
		printk(KERN_DEBUG "a guy asks for address mask. Who is it?\n");
#endif
}

/*
 * RFC1812 (4.3.3.9).	A router SHOULD listen all replies, and complain
 *			loudly if an inconsistency is found.
907
 * called with rcu_read_lock()
L
Linus Torvalds 已提交
908 909 910 911
 */

static void icmp_address_reply(struct sk_buff *skb)
{
E
Eric Dumazet 已提交
912
	struct rtable *rt = skb_rtable(skb);
L
Linus Torvalds 已提交
913 914 915 916 917
	struct net_device *dev = skb->dev;
	struct in_device *in_dev;
	struct in_ifaddr *ifa;

	if (skb->len < 4 || !(rt->rt_flags&RTCF_DIRECTSRC))
918
		return;
L
Linus Torvalds 已提交
919

920
	in_dev = __in_dev_get_rcu(dev);
L
Linus Torvalds 已提交
921
	if (!in_dev)
922 923
		return;

L
Linus Torvalds 已提交
924 925 926
	if (in_dev->ifa_list &&
	    IN_DEV_LOG_MARTIANS(in_dev) &&
	    IN_DEV_FORWARD(in_dev)) {
A
Al Viro 已提交
927
		__be32 _mask, *mp;
L
Linus Torvalds 已提交
928 929

		mp = skb_header_pointer(skb, 0, sizeof(_mask), &_mask);
930
		BUG_ON(mp == NULL);
L
Linus Torvalds 已提交
931 932
		for (ifa = in_dev->ifa_list; ifa; ifa = ifa->ifa_next) {
			if (*mp == ifa->ifa_mask &&
933
			    inet_ifa_match(ip_hdr(skb)->saddr, ifa))
L
Linus Torvalds 已提交
934 935 936
				break;
		}
		if (!ifa && net_ratelimit()) {
937
			printk(KERN_INFO "Wrong address mask %pI4 from %s/%pI4\n",
938
			       mp, dev->name, &ip_hdr(skb)->saddr);
L
Linus Torvalds 已提交
939 940 941 942 943 944 945 946 947 948 949 950 951 952
		}
	}
}

static void icmp_discard(struct sk_buff *skb)
{
}

/*
 *	Deal with incoming ICMP packets.
 */
int icmp_rcv(struct sk_buff *skb)
{
	struct icmphdr *icmph;
E
Eric Dumazet 已提交
953
	struct rtable *rt = skb_rtable(skb);
954
	struct net *net = dev_net(rt->dst.dev);
L
Linus Torvalds 已提交
955

956
	if (!xfrm4_policy_check(NULL, XFRM_POLICY_IN, skb)) {
957
		struct sec_path *sp = skb_sec_path(skb);
958 959
		int nh;

960
		if (!(sp && sp->xvec[sp->len - 1]->props.flags &
961 962 963
				 XFRM_STATE_ICMP))
			goto drop;

964 965 966 967 968 969 970 971 972 973 974 975
		if (!pskb_may_pull(skb, sizeof(*icmph) + sizeof(struct iphdr)))
			goto drop;

		nh = skb_network_offset(skb);
		skb_set_network_header(skb, sizeof(*icmph));

		if (!xfrm4_policy_check_reverse(NULL, XFRM_POLICY_IN, skb))
			goto drop;

		skb_set_network_header(skb, nh);
	}

976
	ICMP_INC_STATS_BH(net, ICMP_MIB_INMSGS);
L
Linus Torvalds 已提交
977 978

	switch (skb->ip_summed) {
979
	case CHECKSUM_COMPLETE:
980
		if (!csum_fold(skb->csum))
L
Linus Torvalds 已提交
981
			break;
982
		/* fall through */
L
Linus Torvalds 已提交
983
	case CHECKSUM_NONE:
984 985
		skb->csum = 0;
		if (__skb_checksum_complete(skb))
L
Linus Torvalds 已提交
986 987 988
			goto error;
	}

989 990
	if (!pskb_pull(skb, sizeof(*icmph)))
		goto error;
L
Linus Torvalds 已提交
991

992
	icmph = icmp_hdr(skb);
L
Linus Torvalds 已提交
993

994
	ICMPMSGIN_INC_STATS_BH(net, icmph->type);
L
Linus Torvalds 已提交
995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008
	/*
	 *	18 is the highest 'known' ICMP type. Anything else is a mystery
	 *
	 *	RFC 1122: 3.2.2  Unknown ICMP messages types MUST be silently
	 *		  discarded.
	 */
	if (icmph->type > NR_ICMP_TYPES)
		goto error;


	/*
	 *	Parse the ICMP message
	 */

1009
	if (rt->rt_flags & (RTCF_BROADCAST | RTCF_MULTICAST)) {
L
Linus Torvalds 已提交
1010 1011 1012 1013 1014 1015
		/*
		 *	RFC 1122: 3.2.2.6 An ICMP_ECHO to broadcast MAY be
		 *	  silently ignored (we let user decide with a sysctl).
		 *	RFC 1122: 3.2.2.8 An ICMP_TIMESTAMP MAY be silently
		 *	  discarded if to broadcast/multicast.
		 */
1016 1017
		if ((icmph->type == ICMP_ECHO ||
		     icmph->type == ICMP_TIMESTAMP) &&
1018
		    net->ipv4.sysctl_icmp_echo_ignore_broadcasts) {
L
Linus Torvalds 已提交
1019 1020 1021 1022 1023 1024 1025
			goto error;
		}
		if (icmph->type != ICMP_ECHO &&
		    icmph->type != ICMP_TIMESTAMP &&
		    icmph->type != ICMP_ADDRESS &&
		    icmph->type != ICMP_ADDRESSREPLY) {
			goto error;
1026
		}
L
Linus Torvalds 已提交
1027 1028 1029 1030 1031 1032 1033 1034
	}

	icmp_pointers[icmph->type].handler(skb);

drop:
	kfree_skb(skb);
	return 0;
error:
1035
	ICMP_INC_STATS_BH(net, ICMP_MIB_INERRORS);
L
Linus Torvalds 已提交
1036 1037 1038 1039 1040 1041
	goto drop;
}

/*
 *	This table is the definition of how we handle ICMP.
 */
1042
static const struct icmp_control icmp_pointers[NR_ICMP_TYPES + 1] = {
L
Linus Torvalds 已提交
1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101
	[ICMP_ECHOREPLY] = {
		.handler = icmp_discard,
	},
	[1] = {
		.handler = icmp_discard,
		.error = 1,
	},
	[2] = {
		.handler = icmp_discard,
		.error = 1,
	},
	[ICMP_DEST_UNREACH] = {
		.handler = icmp_unreach,
		.error = 1,
	},
	[ICMP_SOURCE_QUENCH] = {
		.handler = icmp_unreach,
		.error = 1,
	},
	[ICMP_REDIRECT] = {
		.handler = icmp_redirect,
		.error = 1,
	},
	[6] = {
		.handler = icmp_discard,
		.error = 1,
	},
	[7] = {
		.handler = icmp_discard,
		.error = 1,
	},
	[ICMP_ECHO] = {
		.handler = icmp_echo,
	},
	[9] = {
		.handler = icmp_discard,
		.error = 1,
	},
	[10] = {
		.handler = icmp_discard,
		.error = 1,
	},
	[ICMP_TIME_EXCEEDED] = {
		.handler = icmp_unreach,
		.error = 1,
	},
	[ICMP_PARAMETERPROB] = {
		.handler = icmp_unreach,
		.error = 1,
	},
	[ICMP_TIMESTAMP] = {
		.handler = icmp_timestamp,
	},
	[ICMP_TIMESTAMPREPLY] = {
		.handler = icmp_discard,
	},
	[ICMP_INFO_REQUEST] = {
		.handler = icmp_discard,
	},
1102
	[ICMP_INFO_REPLY] = {
L
Linus Torvalds 已提交
1103 1104 1105 1106 1107 1108 1109 1110 1111 1112
		.handler = icmp_discard,
	},
	[ICMP_ADDRESS] = {
		.handler = icmp_address,
	},
	[ICMP_ADDRESSREPLY] = {
		.handler = icmp_address_reply,
	},
};

1113
static void __net_exit icmp_sk_exit(struct net *net)
L
Linus Torvalds 已提交
1114 1115 1116
{
	int i;

1117
	for_each_possible_cpu(i)
1118
		inet_ctl_sock_destroy(net->ipv4.icmp_sk[i]);
1119 1120
	kfree(net->ipv4.icmp_sk);
	net->ipv4.icmp_sk = NULL;
1121
}
L
Linus Torvalds 已提交
1122

1123
static int __net_init icmp_sk_init(struct net *net)
1124 1125 1126
{
	int i, err;

1127 1128 1129
	net->ipv4.icmp_sk =
		kzalloc(nr_cpu_ids * sizeof(struct sock *), GFP_KERNEL);
	if (net->ipv4.icmp_sk == NULL)
1130 1131
		return -ENOMEM;

1132
	for_each_possible_cpu(i) {
1133
		struct sock *sk;
L
Linus Torvalds 已提交
1134

1135 1136
		err = inet_ctl_sock_create(&sk, PF_INET,
					   SOCK_RAW, IPPROTO_ICMP, net);
L
Linus Torvalds 已提交
1137
		if (err < 0)
1138
			goto fail;
L
Linus Torvalds 已提交
1139

1140
		net->ipv4.icmp_sk[i] = sk;
L
Linus Torvalds 已提交
1141 1142 1143 1144

		/* Enough space for 2 64K ICMP packets, including
		 * sk_buff struct overhead.
		 */
1145
		sk->sk_sndbuf =
L
Linus Torvalds 已提交
1146 1147
			(2 * ((64 * 1024) + sizeof(struct sk_buff)));

1148 1149 1150 1151
		/*
		 * Speedup sock_wfree()
		 */
		sock_set_flag(sk, SOCK_USE_WRITE_QUEUE);
1152
		inet_sk(sk)->pmtudisc = IP_PMTUDISC_DONT;
L
Linus Torvalds 已提交
1153
	}
1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177

	/* Control parameters for ECHO replies. */
	net->ipv4.sysctl_icmp_echo_ignore_all = 0;
	net->ipv4.sysctl_icmp_echo_ignore_broadcasts = 1;

	/* Control parameter - ignore bogus broadcast responses? */
	net->ipv4.sysctl_icmp_ignore_bogus_error_responses = 1;

	/*
	 * 	Configurable global rate limit.
	 *
	 *	ratelimit defines tokens/packet consumed for dst->rate_token
	 *	bucket ratemask defines which icmp types are ratelimited by
	 *	setting	it's bit position.
	 *
	 *	default:
	 *	dest unreachable (3), source quench (4),
	 *	time exceeded (11), parameter problem (12)
	 */

	net->ipv4.sysctl_icmp_ratelimit = 1 * HZ;
	net->ipv4.sysctl_icmp_ratemask = 0x1818;
	net->ipv4.sysctl_icmp_errors_use_inbound_ifaddr = 0;

1178 1179 1180
	return 0;

fail:
1181
	for_each_possible_cpu(i)
1182
		inet_ctl_sock_destroy(net->ipv4.icmp_sk[i]);
1183
	kfree(net->ipv4.icmp_sk);
1184
	return err;
L
Linus Torvalds 已提交
1185 1186
}

1187 1188 1189 1190 1191 1192 1193
static struct pernet_operations __net_initdata icmp_sk_ops = {
       .init = icmp_sk_init,
       .exit = icmp_sk_exit,
};

int __init icmp_init(void)
{
E
Eric W. Biederman 已提交
1194
	return register_pernet_subsys(&icmp_sk_ops);
1195
}