icmp.c 28.0 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
 *
 *	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
 *
 */

65 66
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt

L
Linus Torvalds 已提交
67 68 69 70 71 72 73 74
#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>
75
#include <linux/inetdevice.h>
L
Linus Torvalds 已提交
76 77 78
#include <linux/netdevice.h>
#include <linux/string.h>
#include <linux/netfilter_ipv4.h>
79
#include <linux/slab.h>
L
Linus Torvalds 已提交
80 81 82 83 84 85 86 87
#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>
88
#include <net/ping.h>
L
Linus Torvalds 已提交
89 90 91 92 93 94 95
#include <linux/skbuff.h>
#include <net/sock.h>
#include <linux/errno.h>
#include <linux/timer.h>
#include <linux/init.h>
#include <asm/uaccess.h>
#include <net/checksum.h>
96
#include <net/xfrm.h>
97
#include <net/inet_common.h>
98
#include <net/ip_fib.h>
L
Linus Torvalds 已提交
99 100 101 102 103 104 105 106 107 108 109 110

/*
 *	Build xmit assembly blocks
 */

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

	struct {
		struct icmphdr icmph;
111
		__be32	       times[3];
L
Linus Torvalds 已提交
112 113
	} data;
	int head_len;
114
	struct ip_options_data replyopts;
L
Linus Torvalds 已提交
115 116 117 118 119
};

/* 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'. */

120
const struct icmp_err icmp_err_convert[] = {
L
Linus Torvalds 已提交
121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185
	{
		.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
186
EXPORT_SYMBOL(icmp_err_convert);
L
Linus Torvalds 已提交
187 188 189 190 191 192 193 194 195 196

/*
 *	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 */
};

197
static const struct icmp_control icmp_pointers[NR_ICMP_TYPES+1];
L
Linus Torvalds 已提交
198 199 200 201 202 203 204 205

/*
 *	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.
 */
206 207 208 209
static struct sock *icmp_sk(struct net *net)
{
	return net->ipv4.icmp_sk[smp_processor_id()];
}
L
Linus Torvalds 已提交
210

211
static inline struct sock *icmp_xmit_lock(struct net *net)
L
Linus Torvalds 已提交
212
{
213 214
	struct sock *sk;

L
Linus Torvalds 已提交
215 216
	local_bh_disable();

217 218
	sk = icmp_sk(net);

219
	if (unlikely(!spin_trylock(&sk->sk_lock.slock))) {
L
Linus Torvalds 已提交
220 221 222 223
		/* This can happen if the output path signals a
		 * dst_link_failure() for an outgoing ICMP packet.
		 */
		local_bh_enable();
224
		return NULL;
L
Linus Torvalds 已提交
225
	}
226
	return sk;
L
Linus Torvalds 已提交
227 228
}

229
static inline void icmp_xmit_unlock(struct sock *sk)
L
Linus Torvalds 已提交
230
{
231
	spin_unlock_bh(&sk->sk_lock.slock);
L
Linus Torvalds 已提交
232 233 234 235 236 237
}

/*
 *	Send an ICMP frame.
 */

238
static inline bool icmpv4_xrlim_allow(struct net *net, struct rtable *rt,
239
				      struct flowi4 *fl4, int type, int code)
L
Linus Torvalds 已提交
240
{
241
	struct dst_entry *dst = &rt->dst;
242
	bool rc = true;
L
Linus Torvalds 已提交
243 244 245 246 247 248 249 250 251 252

	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))
253
		goto out;
L
Linus Torvalds 已提交
254 255

	/* Limit if icmp type is enabled in ratemask. */
256
	if ((1 << type) & net->ipv4.sysctl_icmp_ratemask) {
257
		struct inet_peer *peer = inet_getpeer_v4(net->ipv4.peers, fl4->daddr, 1);
258
		rc = inet_peer_xrlim_allow(peer,
259
					   net->ipv4.sysctl_icmp_ratelimit);
260
		inet_putpeer(peer);
261
	}
L
Linus Torvalds 已提交
262 263 264 265 266 267 268
out:
	return rc;
}

/*
 *	Maintain the counters used in the SNMP statistics for outgoing ICMP
 */
269
void icmp_out_count(struct net *net, unsigned char type)
L
Linus Torvalds 已提交
270
{
271
	ICMPMSGOUT_INC_STATS(net, type);
272
	ICMP_INC_STATS(net, ICMP_MIB_OUTMSGS);
L
Linus Torvalds 已提交
273 274 275 276 277 278 279 280 281 282
}

/*
 *	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;
283
	__wsum csum;
L
Linus Torvalds 已提交
284 285 286 287 288 289 290 291 292 293 294 295

	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,
296
			    struct flowi4 *fl4,
297
			    struct ipcm_cookie *ipc, struct rtable **rt)
L
Linus Torvalds 已提交
298
{
299
	struct sock *sk;
L
Linus Torvalds 已提交
300 301
	struct sk_buff *skb;

302
	sk = icmp_sk(dev_net((*rt)->dst.dev));
303
	if (ip_append_data(sk, fl4, icmp_glue_bits, icmp_param,
304 305
			   icmp_param->data_len+icmp_param->head_len,
			   icmp_param->head_len,
E
Eric Dumazet 已提交
306 307
			   ipc, rt, MSG_DONTWAIT) < 0) {
		ICMP_INC_STATS_BH(sock_net(sk), ICMP_MIB_OUTERRORS);
308
		ip_flush_pending_frames(sk);
E
Eric Dumazet 已提交
309
	} else if ((skb = skb_peek(&sk->sk_write_queue)) != NULL) {
310
		struct icmphdr *icmph = icmp_hdr(skb);
311
		__wsum csum = 0;
L
Linus Torvalds 已提交
312 313
		struct sk_buff *skb1;

314
		skb_queue_walk(&sk->sk_write_queue, skb1) {
L
Linus Torvalds 已提交
315 316 317 318 319 320 321
			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;
322
		ip_push_pending_frames(sk, fl4);
L
Linus Torvalds 已提交
323 324 325 326 327 328 329 330 331 332
	}
}

/*
 *	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
333
	struct rtable *rt = skb_rtable(skb);
334
	struct net *net = dev_net(rt->dst.dev);
335
	struct flowi4 fl4;
336 337
	struct sock *sk;
	struct inet_sock *inet;
338
	__be32 daddr, saddr;
L
Linus Torvalds 已提交
339

340
	if (ip_options_echo(&icmp_param->replyopts.opt.opt, skb))
341
		return;
L
Linus Torvalds 已提交
342

343 344
	sk = icmp_xmit_lock(net);
	if (sk == NULL)
L
Linus Torvalds 已提交
345
		return;
346
	inet = inet_sk(sk);
L
Linus Torvalds 已提交
347 348 349

	icmp_param->data.icmph.checksum = 0;

350
	inet->tos = ip_hdr(skb)->tos;
351
	daddr = ipc.addr = ip_hdr(skb)->saddr;
352
	saddr = fib_compute_spec_dst(skb);
L
Linus Torvalds 已提交
353
	ipc.opt = NULL;
354
	ipc.tx_flags = 0;
355 356 357 358
	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
359
	}
360 361
	memset(&fl4, 0, sizeof(fl4));
	fl4.daddr = daddr;
362
	fl4.saddr = saddr;
363 364 365 366 367 368
	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;
369
	if (icmpv4_xrlim_allow(net, rt, &fl4, icmp_param->data.icmph.type,
L
Linus Torvalds 已提交
370
			       icmp_param->data.icmph.code))
371
		icmp_push_reply(icmp_param, &fl4, &ipc, &rt);
L
Linus Torvalds 已提交
372 373
	ip_rt_put(rt);
out_unlock:
374
	icmp_xmit_unlock(sk);
L
Linus Torvalds 已提交
375 376
}

377 378 379
static struct rtable *icmp_route_lookup(struct net *net,
					struct flowi4 *fl4,
					struct sk_buff *skb_in,
380
					const struct iphdr *iph,
381 382 383 384 385
					__be32 saddr, u8 tos,
					int type, int code,
					struct icmp_bxm *param)
{
	struct rtable *rt, *rt2;
386
	struct flowi4 fl4_dec;
387 388
	int err;

389 390 391 392 393 394 395 396 397 398
	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);
399 400
	if (IS_ERR(rt))
		return rt;
401 402 403 404

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

405
	rt = (struct rtable *) xfrm_lookup(net, &rt->dst,
406
					   flowi4_to_flowi(fl4), NULL, 0);
407
	if (!IS_ERR(rt)) {
408 409
		if (rt != rt2)
			return rt;
410
	} else if (PTR_ERR(rt) == -EPERM) {
411
		rt = NULL;
412 413
	} else
		return rt;
414

415
	err = xfrm_decode_session_reverse(skb_in, flowi4_to_flowi(&fl4_dec), AF_INET);
416 417 418
	if (err)
		goto relookup_failed;

419 420
	if (inet_addr_type(net, fl4_dec.saddr) == RTN_LOCAL) {
		rt2 = __ip_route_output_key(net, &fl4_dec);
421 422
		if (IS_ERR(rt2))
			err = PTR_ERR(rt2);
423
	} else {
424
		struct flowi4 fl4_2 = {};
425 426
		unsigned long orefdst;

427
		fl4_2.daddr = fl4_dec.saddr;
428
		rt2 = ip_route_output_key(net, &fl4_2);
429 430
		if (IS_ERR(rt2)) {
			err = PTR_ERR(rt2);
431
			goto relookup_failed;
432
		}
433 434
		/* Ugh! */
		orefdst = skb_in->_skb_refdst; /* save old refdst */
435
		err = ip_route_input(skb_in, fl4_dec.daddr, fl4_dec.saddr,
436 437 438 439 440 441 442 443 444 445
				     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;

446
	rt2 = (struct rtable *) xfrm_lookup(net, &rt2->dst,
447
					    flowi4_to_flowi(&fl4_dec), NULL,
448
					    XFRM_LOOKUP_ICMP);
449
	if (!IS_ERR(rt2)) {
450
		dst_release(&rt->dst);
451
		memcpy(fl4, &fl4_dec, sizeof(*fl4));
452
		rt = rt2;
453 454 455 456 457 458 459
	} else if (PTR_ERR(rt2) == -EPERM) {
		if (rt)
			dst_release(&rt->dst);
		return rt2;
	} else {
		err = PTR_ERR(rt2);
		goto relookup_failed;
460 461 462 463 464 465 466 467
	}
	return rt;

relookup_failed:
	if (rt)
		return rt;
	return ERR_PTR(err);
}
L
Linus Torvalds 已提交
468 469 470 471 472 473 474 475 476 477 478 479

/*
 *	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
480
void icmp_send(struct sk_buff *skb_in, int type, int code, __be32 info)
L
Linus Torvalds 已提交
481 482 483 484
{
	struct iphdr *iph;
	int room;
	struct icmp_bxm icmp_param;
E
Eric Dumazet 已提交
485
	struct rtable *rt = skb_rtable(skb_in);
L
Linus Torvalds 已提交
486
	struct ipcm_cookie ipc;
487
	struct flowi4 fl4;
A
Al Viro 已提交
488
	__be32 saddr;
L
Linus Torvalds 已提交
489
	u8  tos;
490
	struct net *net;
491
	struct sock *sk;
L
Linus Torvalds 已提交
492 493 494

	if (!rt)
		goto out;
495
	net = dev_net(rt->dst.dev);
L
Linus Torvalds 已提交
496 497 498 499 500 501

	/*
	 *	Find the original header. It is expected to be valid, of course.
	 *	Check this, icmp_send is called from the most obscure devices
	 *	sometimes.
	 */
502
	iph = ip_hdr(skb_in);
L
Linus Torvalds 已提交
503

504 505
	if ((u8 *)iph < skb_in->head ||
	    (skb_in->network_header + sizeof(*iph)) > skb_in->tail)
L
Linus Torvalds 已提交
506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538
		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,
539
						 skb_network_header(skb_in) +
L
Linus Torvalds 已提交
540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558
						 (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;
		}
	}

559 560
	sk = icmp_xmit_lock(net);
	if (sk == NULL)
L
Linus Torvalds 已提交
561 562 563 564 565 566 567
		return;

	/*
	 *	Construct source address and options.
	 */

	saddr = iph->daddr;
568
	if (!(rt->rt_flags & RTCF_LOCAL)) {
569 570
		struct net_device *dev = NULL;

571
		rcu_read_lock();
572 573
		if (rt_is_input_route(rt) &&
		    net->ipv4.sysctl_icmp_errors_use_inbound_ifaddr)
574
			dev = dev_get_by_index_rcu(net, rt->rt_iif);
575

576
		if (dev)
577
			saddr = inet_select_addr(dev, 0, RT_SCOPE_LINK);
578
		else
579
			saddr = 0;
580
		rcu_read_unlock();
581
	}
L
Linus Torvalds 已提交
582 583 584 585 586

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

587
	if (ip_options_echo(&icmp_param.replyopts.opt.opt, skb_in))
588
		goto out_unlock;
L
Linus Torvalds 已提交
589 590 591 592 593 594 595 596 597 598 599


	/*
	 *	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;
600
	icmp_param.offset = skb_network_offset(skb_in);
601
	inet_sk(sk)->tos = tos;
L
Linus Torvalds 已提交
602
	ipc.addr = iph->saddr;
603
	ipc.opt = &icmp_param.replyopts.opt;
604
	ipc.tx_flags = 0;
L
Linus Torvalds 已提交
605

606
	rt = icmp_route_lookup(net, &fl4, skb_in, iph, saddr, tos,
607 608 609
			       type, code, &icmp_param);
	if (IS_ERR(rt))
		goto out_unlock;
L
Linus Torvalds 已提交
610

611
	if (!icmpv4_xrlim_allow(net, rt, &fl4, type, code))
L
Linus Torvalds 已提交
612 613 614 615
		goto ende;

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

616
	room = dst_mtu(&rt->dst);
L
Linus Torvalds 已提交
617 618
	if (room > 576)
		room = 576;
619
	room -= sizeof(struct iphdr) + icmp_param.replyopts.opt.opt.optlen;
L
Linus Torvalds 已提交
620 621 622 623 624 625 626
	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);

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


637 638 639 640 641 642
static void icmp_socket_deliver(struct sk_buff *skb, u32 info)
{
	const struct iphdr *iph = (const struct iphdr *) skb->data;
	const struct net_protocol *ipprot;
	int protocol = iph->protocol;

643 644 645 646 647 648
	/* 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))
		return;

649 650 651 652 653 654 655 656 657
	raw_icmp_error(skb, protocol, info);

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

L
Linus Torvalds 已提交
658 659 660 661 662 663
/*
 *	Handle ICMP_DEST_UNREACH, ICMP_TIME_EXCEED, and ICMP_QUENCH.
 */

static void icmp_unreach(struct sk_buff *skb)
{
664
	const struct iphdr *iph;
L
Linus Torvalds 已提交
665
	struct icmphdr *icmph;
666
	struct net *net;
667
	u32 info = 0;
668

E
Eric Dumazet 已提交
669
	net = dev_net(skb_dst(skb)->dev);
L
Linus Torvalds 已提交
670 671 672 673 674 675 676 677 678 679

	/*
	 *	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;

680
	icmph = icmp_hdr(skb);
681
	iph   = (const struct iphdr *)skb->data;
L
Linus Torvalds 已提交
682 683 684 685 686 687 688 689 690 691 692 693 694

	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) {
695
				LIMIT_NETDEBUG(KERN_INFO pr_fmt("%pI4: fragmentation needed and DF set\n"),
696
					       &iph->daddr);
L
Linus Torvalds 已提交
697
			} else {
698
				info = ntohs(icmph->un.frag.mtu);
L
Linus Torvalds 已提交
699 700 701 702 703
				if (!info)
					goto out;
			}
			break;
		case ICMP_SR_FAILED:
704
			LIMIT_NETDEBUG(KERN_INFO pr_fmt("%pI4: Source Route Failed\n"),
705
				       &iph->daddr);
L
Linus Torvalds 已提交
706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726
			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
727
	 *	Check the other end isn't violating RFC 1122. Some routers send
L
Linus Torvalds 已提交
728 729 730 731 732
	 *	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.
	 */

733
	if (!net->ipv4.sysctl_icmp_ignore_bogus_error_responses &&
734
	    inet_addr_type(net, iph->daddr) == RTN_BROADCAST) {
735 736 737 738
		net_warn_ratelimited("%pI4 sent an invalid ICMP type %u, code %u error to a broadcast: %pI4 on %s\n",
				     &ip_hdr(skb)->saddr,
				     icmph->type, icmph->code,
				     &iph->daddr, skb->dev->name);
L
Linus Torvalds 已提交
739 740 741
		goto out;
	}

742
	icmp_socket_deliver(skb, info);
L
Linus Torvalds 已提交
743 744 745 746

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)
{
758 759 760 761
	if (skb->len < sizeof(struct iphdr)) {
		ICMP_INC_STATS_BH(dev_net(skb->dev), ICMP_MIB_INERRORS);
		return;
	}
L
Linus Torvalds 已提交
762 763

	if (!pskb_may_pull(skb, sizeof(struct iphdr)))
764
		return;
765

766
	icmp_socket_deliver(skb, icmp_hdr(skb)->un.gateway);
L
Linus Torvalds 已提交
767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782
}

/*
 *	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)
{
783 784
	struct net *net;

E
Eric Dumazet 已提交
785
	net = dev_net(skb_dst(skb)->dev);
786
	if (!net->ipv4.sysctl_icmp_echo_ignore_all) {
L
Linus Torvalds 已提交
787 788
		struct icmp_bxm icmp_param;

789
		icmp_param.data.icmph	   = *icmp_hdr(skb);
L
Linus Torvalds 已提交
790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807
		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)
{
808
	struct timespec tv;
L
Linus Torvalds 已提交
809 810 811 812 813 814 815 816 817 818
	struct icmp_bxm icmp_param;
	/*
	 *	Too short.
	 */
	if (skb->len < 4)
		goto out_err;

	/*
	 *	Fill in the current time as ms since midnight UT:
	 */
819 820 821
	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
822 823 824
	icmp_param.data.times[2] = icmp_param.data.times[1];
	if (skb_copy_bits(skb, 0, &icmp_param.data.times[0], 4))
		BUG();
825
	icmp_param.data.icmph	   = *icmp_hdr(skb);
L
Linus Torvalds 已提交
826 827 828 829 830 831 832 833 834 835
	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
836
	ICMP_INC_STATS_BH(dev_net(skb_dst(skb)->dev), ICMP_MIB_INERRORS);
L
Linus Torvalds 已提交
837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876
	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
877
	net_dbg_ratelimited("a guy asks for address mask. Who is it?\n");
L
Linus Torvalds 已提交
878 879 880 881 882 883
#endif
}

/*
 * RFC1812 (4.3.3.9).	A router SHOULD listen all replies, and complain
 *			loudly if an inconsistency is found.
884
 * called with rcu_read_lock()
L
Linus Torvalds 已提交
885 886 887 888
 */

static void icmp_address_reply(struct sk_buff *skb)
{
E
Eric Dumazet 已提交
889
	struct rtable *rt = skb_rtable(skb);
L
Linus Torvalds 已提交
890 891 892 893 894
	struct net_device *dev = skb->dev;
	struct in_device *in_dev;
	struct in_ifaddr *ifa;

	if (skb->len < 4 || !(rt->rt_flags&RTCF_DIRECTSRC))
895
		return;
L
Linus Torvalds 已提交
896

897
	in_dev = __in_dev_get_rcu(dev);
L
Linus Torvalds 已提交
898
	if (!in_dev)
899 900
		return;

L
Linus Torvalds 已提交
901 902 903
	if (in_dev->ifa_list &&
	    IN_DEV_LOG_MARTIANS(in_dev) &&
	    IN_DEV_FORWARD(in_dev)) {
A
Al Viro 已提交
904
		__be32 _mask, *mp;
L
Linus Torvalds 已提交
905 906

		mp = skb_header_pointer(skb, 0, sizeof(_mask), &_mask);
907
		BUG_ON(mp == NULL);
L
Linus Torvalds 已提交
908 909
		for (ifa = in_dev->ifa_list; ifa; ifa = ifa->ifa_next) {
			if (*mp == ifa->ifa_mask &&
910
			    inet_ifa_match(ip_hdr(skb)->saddr, ifa))
L
Linus Torvalds 已提交
911 912
				break;
		}
913 914 915 916
		if (!ifa)
			net_info_ratelimited("Wrong address mask %pI4 from %s/%pI4\n",
					     mp,
					     dev->name, &ip_hdr(skb)->saddr);
L
Linus Torvalds 已提交
917 918 919 920 921 922 923 924 925 926 927 928 929
	}
}

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
930
	struct rtable *rt = skb_rtable(skb);
931
	struct net *net = dev_net(rt->dst.dev);
L
Linus Torvalds 已提交
932

933
	if (!xfrm4_policy_check(NULL, XFRM_POLICY_IN, skb)) {
934
		struct sec_path *sp = skb_sec_path(skb);
935 936
		int nh;

937
		if (!(sp && sp->xvec[sp->len - 1]->props.flags &
938 939 940
				 XFRM_STATE_ICMP))
			goto drop;

941 942 943 944 945 946 947 948 949 950 951 952
		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);
	}

953
	ICMP_INC_STATS_BH(net, ICMP_MIB_INMSGS);
L
Linus Torvalds 已提交
954 955

	switch (skb->ip_summed) {
956
	case CHECKSUM_COMPLETE:
957
		if (!csum_fold(skb->csum))
L
Linus Torvalds 已提交
958
			break;
959
		/* fall through */
L
Linus Torvalds 已提交
960
	case CHECKSUM_NONE:
961 962
		skb->csum = 0;
		if (__skb_checksum_complete(skb))
L
Linus Torvalds 已提交
963 964 965
			goto error;
	}

966 967
	if (!pskb_pull(skb, sizeof(*icmph)))
		goto error;
L
Linus Torvalds 已提交
968

969
	icmph = icmp_hdr(skb);
L
Linus Torvalds 已提交
970

971
	ICMPMSGIN_INC_STATS_BH(net, icmph->type);
L
Linus Torvalds 已提交
972 973 974 975 976 977 978 979 980 981 982 983 984 985
	/*
	 *	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
	 */

986
	if (rt->rt_flags & (RTCF_BROADCAST | RTCF_MULTICAST)) {
L
Linus Torvalds 已提交
987 988 989 990 991 992
		/*
		 *	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.
		 */
993 994
		if ((icmph->type == ICMP_ECHO ||
		     icmph->type == ICMP_TIMESTAMP) &&
995
		    net->ipv4.sysctl_icmp_echo_ignore_broadcasts) {
L
Linus Torvalds 已提交
996 997 998 999 1000 1001 1002
			goto error;
		}
		if (icmph->type != ICMP_ECHO &&
		    icmph->type != ICMP_TIMESTAMP &&
		    icmph->type != ICMP_ADDRESS &&
		    icmph->type != ICMP_ADDRESSREPLY) {
			goto error;
1003
		}
L
Linus Torvalds 已提交
1004 1005 1006 1007 1008 1009 1010 1011
	}

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

drop:
	kfree_skb(skb);
	return 0;
error:
1012
	ICMP_INC_STATS_BH(net, ICMP_MIB_INERRORS);
L
Linus Torvalds 已提交
1013 1014 1015 1016 1017 1018
	goto drop;
}

/*
 *	This table is the definition of how we handle ICMP.
 */
1019
static const struct icmp_control icmp_pointers[NR_ICMP_TYPES + 1] = {
L
Linus Torvalds 已提交
1020
	[ICMP_ECHOREPLY] = {
1021
		.handler = ping_rcv,
L
Linus Torvalds 已提交
1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078
	},
	[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,
	},
1079
	[ICMP_INFO_REPLY] = {
L
Linus Torvalds 已提交
1080 1081 1082 1083 1084 1085 1086 1087 1088 1089
		.handler = icmp_discard,
	},
	[ICMP_ADDRESS] = {
		.handler = icmp_address,
	},
	[ICMP_ADDRESSREPLY] = {
		.handler = icmp_address_reply,
	},
};

1090
static void __net_exit icmp_sk_exit(struct net *net)
L
Linus Torvalds 已提交
1091 1092 1093
{
	int i;

1094
	for_each_possible_cpu(i)
1095
		inet_ctl_sock_destroy(net->ipv4.icmp_sk[i]);
1096 1097
	kfree(net->ipv4.icmp_sk);
	net->ipv4.icmp_sk = NULL;
1098
}
L
Linus Torvalds 已提交
1099

1100
static int __net_init icmp_sk_init(struct net *net)
1101 1102 1103
{
	int i, err;

1104 1105 1106
	net->ipv4.icmp_sk =
		kzalloc(nr_cpu_ids * sizeof(struct sock *), GFP_KERNEL);
	if (net->ipv4.icmp_sk == NULL)
1107 1108
		return -ENOMEM;

1109
	for_each_possible_cpu(i) {
1110
		struct sock *sk;
L
Linus Torvalds 已提交
1111

1112 1113
		err = inet_ctl_sock_create(&sk, PF_INET,
					   SOCK_RAW, IPPROTO_ICMP, net);
L
Linus Torvalds 已提交
1114
		if (err < 0)
1115
			goto fail;
L
Linus Torvalds 已提交
1116

1117
		net->ipv4.icmp_sk[i] = sk;
L
Linus Torvalds 已提交
1118 1119

		/* Enough space for 2 64K ICMP packets, including
E
Eric Dumazet 已提交
1120
		 * sk_buff/skb_shared_info struct overhead.
L
Linus Torvalds 已提交
1121
		 */
E
Eric Dumazet 已提交
1122
		sk->sk_sndbuf =	2 * SKB_TRUESIZE(64 * 1024);
L
Linus Torvalds 已提交
1123

1124 1125 1126 1127
		/*
		 * Speedup sock_wfree()
		 */
		sock_set_flag(sk, SOCK_USE_WRITE_QUEUE);
1128
		inet_sk(sk)->pmtudisc = IP_PMTUDISC_DONT;
L
Linus Torvalds 已提交
1129
	}
1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153

	/* 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;

1154 1155 1156
	return 0;

fail:
1157
	for_each_possible_cpu(i)
1158
		inet_ctl_sock_destroy(net->ipv4.icmp_sk[i]);
1159
	kfree(net->ipv4.icmp_sk);
1160
	return err;
L
Linus Torvalds 已提交
1161 1162
}

1163 1164 1165 1166 1167 1168 1169
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
1170
	return register_pernet_subsys(&icmp_sk_ops);
1171
}