1 // SPDX-License-Identifier: GPL-2.0-or-later
2 /*
3  *	NET3:	Implementation of the ICMP protocol layer.
4  *
5  *		Alan Cox, <alan@lxorguk.ukuu.org.uk>
6  *
7  *	Some of the function names and the icmp unreach table for this
8  *	module were derived from [icmp.c 1.0.11 06/02/93] by
9  *	Ross Biro, Fred N. van Kempen, Mark Evans, Alan Cox, Gerhard Koerting.
10  *	Other than that this module is a complete rewrite.
11  *
12  *	Fixes:
13  *	Clemens Fruhwirth	:	introduce global icmp rate limiting
14  *					with icmp type masking ability instead
15  *					of broken per type icmp timeouts.
16  *		Mike Shaver	:	RFC1122 checks.
17  *		Alan Cox	:	Multicast ping reply as self.
18  *		Alan Cox	:	Fix atomicity lockup in ip_build_xmit
19  *					call.
20  *		Alan Cox	:	Added 216,128 byte paths to the MTU
21  *					code.
22  *		Martin Mares	:	RFC1812 checks.
23  *		Martin Mares	:	Can be configured to follow redirects
24  *					if acting as a router _without_ a
25  *					routing protocol (RFC 1812).
26  *		Martin Mares	:	Echo requests may be configured to
27  *					be ignored (RFC 1812).
28  *		Martin Mares	:	Limitation of ICMP error message
29  *					transmit rate (RFC 1812).
30  *		Martin Mares	:	TOS and Precedence set correctly
31  *					(RFC 1812).
32  *		Martin Mares	:	Now copying as much data from the
33  *					original packet as we can without
34  *					exceeding 576 bytes (RFC 1812).
35  *	Willy Konynenberg	:	Transparent proxying support.
36  *		Keith Owens	:	RFC1191 correction for 4.2BSD based
37  *					path MTU bug.
38  *		Thomas Quinot	:	ICMP Dest Unreach codes up to 15 are
39  *					valid (RFC 1812).
40  *		Andi Kleen	:	Check all packet lengths properly
41  *					and moved all kfree_skb() up to
42  *					icmp_rcv.
43  *		Andi Kleen	:	Move the rate limit bookkeeping
44  *					into the dest entry and use a token
45  *					bucket filter (thanks to ANK). Make
46  *					the rates sysctl configurable.
47  *		Yu Tianli	:	Fixed two ugly bugs in icmp_send
48  *					- IP option length was accounted wrongly
49  *					- ICMP header length was not accounted
50  *					  at all.
51  *              Tristan Greaves :       Added sysctl option to ignore bogus
52  *              			broadcast responses from broken routers.
53  *
54  * To Fix:
55  *
56  *	- Should use skb_pull() instead of all the manual checking.
57  *	  This would also greatly simply some upper layer error handlers. --AK
58  */
59 
60 #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
61 
62 #include <linux/module.h>
63 #include <linux/types.h>
64 #include <linux/jiffies.h>
65 #include <linux/kernel.h>
66 #include <linux/fcntl.h>
67 #include <linux/socket.h>
68 #include <linux/in.h>
69 #include <linux/inet.h>
70 #include <linux/inetdevice.h>
71 #include <linux/netdevice.h>
72 #include <linux/string.h>
73 #include <linux/netfilter_ipv4.h>
74 #include <linux/slab.h>
75 #include <net/snmp.h>
76 #include <net/ip.h>
77 #include <net/route.h>
78 #include <net/protocol.h>
79 #include <net/icmp.h>
80 #include <net/tcp.h>
81 #include <net/udp.h>
82 #include <net/raw.h>
83 #include <net/ping.h>
84 #include <linux/skbuff.h>
85 #include <net/sock.h>
86 #include <linux/errno.h>
87 #include <linux/timer.h>
88 #include <linux/init.h>
89 #include <linux/uaccess.h>
90 #include <net/checksum.h>
91 #include <net/xfrm.h>
92 #include <net/inet_common.h>
93 #include <net/ip_fib.h>
94 #include <net/l3mdev.h>
95 
96 /*
97  *	Build xmit assembly blocks
98  */
99 
100 struct icmp_bxm {
101 	struct sk_buff *skb;
102 	int offset;
103 	int data_len;
104 
105 	struct {
106 		struct icmphdr icmph;
107 		__be32	       times[3];
108 	} data;
109 	int head_len;
110 	struct ip_options_data replyopts;
111 };
112 
113 /* An array of errno for error messages from dest unreach. */
114 /* RFC 1122: 3.2.2.1 States that NET_UNREACH, HOST_UNREACH and SR_FAILED MUST be considered 'transient errs'. */
115 
116 const struct icmp_err icmp_err_convert[] = {
117 	{
118 		.errno = ENETUNREACH,	/* ICMP_NET_UNREACH */
119 		.fatal = 0,
120 	},
121 	{
122 		.errno = EHOSTUNREACH,	/* ICMP_HOST_UNREACH */
123 		.fatal = 0,
124 	},
125 	{
126 		.errno = ENOPROTOOPT	/* ICMP_PROT_UNREACH */,
127 		.fatal = 1,
128 	},
129 	{
130 		.errno = ECONNREFUSED,	/* ICMP_PORT_UNREACH */
131 		.fatal = 1,
132 	},
133 	{
134 		.errno = EMSGSIZE,	/* ICMP_FRAG_NEEDED */
135 		.fatal = 0,
136 	},
137 	{
138 		.errno = EOPNOTSUPP,	/* ICMP_SR_FAILED */
139 		.fatal = 0,
140 	},
141 	{
142 		.errno = ENETUNREACH,	/* ICMP_NET_UNKNOWN */
143 		.fatal = 1,
144 	},
145 	{
146 		.errno = EHOSTDOWN,	/* ICMP_HOST_UNKNOWN */
147 		.fatal = 1,
148 	},
149 	{
150 		.errno = ENONET,	/* ICMP_HOST_ISOLATED */
151 		.fatal = 1,
152 	},
153 	{
154 		.errno = ENETUNREACH,	/* ICMP_NET_ANO	*/
155 		.fatal = 1,
156 	},
157 	{
158 		.errno = EHOSTUNREACH,	/* ICMP_HOST_ANO */
159 		.fatal = 1,
160 	},
161 	{
162 		.errno = ENETUNREACH,	/* ICMP_NET_UNR_TOS */
163 		.fatal = 0,
164 	},
165 	{
166 		.errno = EHOSTUNREACH,	/* ICMP_HOST_UNR_TOS */
167 		.fatal = 0,
168 	},
169 	{
170 		.errno = EHOSTUNREACH,	/* ICMP_PKT_FILTERED */
171 		.fatal = 1,
172 	},
173 	{
174 		.errno = EHOSTUNREACH,	/* ICMP_PREC_VIOLATION */
175 		.fatal = 1,
176 	},
177 	{
178 		.errno = EHOSTUNREACH,	/* ICMP_PREC_CUTOFF */
179 		.fatal = 1,
180 	},
181 };
182 EXPORT_SYMBOL(icmp_err_convert);
183 
184 /*
185  *	ICMP control array. This specifies what to do with each ICMP.
186  */
187 
188 struct icmp_control {
189 	enum skb_drop_reason (*handler)(struct sk_buff *skb);
190 	short   error;		/* This ICMP is classed as an error message */
191 };
192 
193 static const struct icmp_control icmp_pointers[NR_ICMP_TYPES+1];
194 
195 static DEFINE_PER_CPU(struct sock *, ipv4_icmp_sk);
196 
197 /* Called with BH disabled */
icmp_xmit_lock(struct net * net)198 static inline struct sock *icmp_xmit_lock(struct net *net)
199 {
200 	struct sock *sk;
201 
202 	sk = this_cpu_read(ipv4_icmp_sk);
203 
204 	if (unlikely(!spin_trylock(&sk->sk_lock.slock))) {
205 		/* This can happen if the output path signals a
206 		 * dst_link_failure() for an outgoing ICMP packet.
207 		 */
208 		return NULL;
209 	}
210 	sock_net_set(sk, net);
211 	return sk;
212 }
213 
icmp_xmit_unlock(struct sock * sk)214 static inline void icmp_xmit_unlock(struct sock *sk)
215 {
216 	sock_net_set(sk, &init_net);
217 	spin_unlock(&sk->sk_lock.slock);
218 }
219 
220 int sysctl_icmp_msgs_per_sec __read_mostly = 1000;
221 int sysctl_icmp_msgs_burst __read_mostly = 50;
222 
223 static struct {
224 	spinlock_t	lock;
225 	u32		credit;
226 	u32		stamp;
227 } icmp_global = {
228 	.lock		= __SPIN_LOCK_UNLOCKED(icmp_global.lock),
229 };
230 
231 /**
232  * icmp_global_allow - Are we allowed to send one more ICMP message ?
233  *
234  * Uses a token bucket to limit our ICMP messages to ~sysctl_icmp_msgs_per_sec.
235  * Returns false if we reached the limit and can not send another packet.
236  * Note: called with BH disabled
237  */
icmp_global_allow(void)238 bool icmp_global_allow(void)
239 {
240 	u32 credit, delta, incr = 0, now = (u32)jiffies;
241 	bool rc = false;
242 
243 	/* Check if token bucket is empty and cannot be refilled
244 	 * without taking the spinlock. The READ_ONCE() are paired
245 	 * with the following WRITE_ONCE() in this same function.
246 	 */
247 	if (!READ_ONCE(icmp_global.credit)) {
248 		delta = min_t(u32, now - READ_ONCE(icmp_global.stamp), HZ);
249 		if (delta < HZ / 50)
250 			return false;
251 	}
252 
253 	spin_lock(&icmp_global.lock);
254 	delta = min_t(u32, now - icmp_global.stamp, HZ);
255 	if (delta >= HZ / 50) {
256 		incr = READ_ONCE(sysctl_icmp_msgs_per_sec) * delta / HZ;
257 		if (incr)
258 			WRITE_ONCE(icmp_global.stamp, now);
259 	}
260 	credit = min_t(u32, icmp_global.credit + incr,
261 		       READ_ONCE(sysctl_icmp_msgs_burst));
262 	if (credit) {
263 		/* We want to use a credit of one in average, but need to randomize
264 		 * it for security reasons.
265 		 */
266 		credit = max_t(int, credit - get_random_u32_below(3), 0);
267 		rc = true;
268 	}
269 	WRITE_ONCE(icmp_global.credit, credit);
270 	spin_unlock(&icmp_global.lock);
271 	return rc;
272 }
273 EXPORT_SYMBOL(icmp_global_allow);
274 
icmpv4_mask_allow(struct net * net,int type,int code)275 static bool icmpv4_mask_allow(struct net *net, int type, int code)
276 {
277 	if (type > NR_ICMP_TYPES)
278 		return true;
279 
280 	/* Don't limit PMTU discovery. */
281 	if (type == ICMP_DEST_UNREACH && code == ICMP_FRAG_NEEDED)
282 		return true;
283 
284 	/* Limit if icmp type is enabled in ratemask. */
285 	if (!((1 << type) & READ_ONCE(net->ipv4.sysctl_icmp_ratemask)))
286 		return true;
287 
288 	return false;
289 }
290 
icmpv4_global_allow(struct net * net,int type,int code)291 static bool icmpv4_global_allow(struct net *net, int type, int code)
292 {
293 	if (icmpv4_mask_allow(net, type, code))
294 		return true;
295 
296 	if (icmp_global_allow())
297 		return true;
298 
299 	__ICMP_INC_STATS(net, ICMP_MIB_RATELIMITGLOBAL);
300 	return false;
301 }
302 
303 /*
304  *	Send an ICMP frame.
305  */
306 
icmpv4_xrlim_allow(struct net * net,struct rtable * rt,struct flowi4 * fl4,int type,int code)307 static bool icmpv4_xrlim_allow(struct net *net, struct rtable *rt,
308 			       struct flowi4 *fl4, int type, int code)
309 {
310 	struct dst_entry *dst = &rt->dst;
311 	struct inet_peer *peer;
312 	bool rc = true;
313 	int vif;
314 
315 	if (icmpv4_mask_allow(net, type, code))
316 		goto out;
317 
318 	/* No rate limit on loopback */
319 	if (dst->dev && (dst->dev->flags&IFF_LOOPBACK))
320 		goto out;
321 
322 	vif = l3mdev_master_ifindex(dst->dev);
323 	peer = inet_getpeer_v4(net->ipv4.peers, fl4->daddr, vif, 1);
324 	rc = inet_peer_xrlim_allow(peer,
325 				   READ_ONCE(net->ipv4.sysctl_icmp_ratelimit));
326 	if (peer)
327 		inet_putpeer(peer);
328 out:
329 	if (!rc)
330 		__ICMP_INC_STATS(net, ICMP_MIB_RATELIMITHOST);
331 	return rc;
332 }
333 
334 /*
335  *	Maintain the counters used in the SNMP statistics for outgoing ICMP
336  */
icmp_out_count(struct net * net,unsigned char type)337 void icmp_out_count(struct net *net, unsigned char type)
338 {
339 	ICMPMSGOUT_INC_STATS(net, type);
340 	ICMP_INC_STATS(net, ICMP_MIB_OUTMSGS);
341 }
342 
343 /*
344  *	Checksum each fragment, and on the first include the headers and final
345  *	checksum.
346  */
icmp_glue_bits(void * from,char * to,int offset,int len,int odd,struct sk_buff * skb)347 static int icmp_glue_bits(void *from, char *to, int offset, int len, int odd,
348 			  struct sk_buff *skb)
349 {
350 	struct icmp_bxm *icmp_param = from;
351 	__wsum csum;
352 
353 	csum = skb_copy_and_csum_bits(icmp_param->skb,
354 				      icmp_param->offset + offset,
355 				      to, len);
356 
357 	skb->csum = csum_block_add(skb->csum, csum, odd);
358 	if (icmp_pointers[icmp_param->data.icmph.type].error)
359 		nf_ct_attach(skb, icmp_param->skb);
360 	return 0;
361 }
362 
icmp_push_reply(struct sock * sk,struct icmp_bxm * icmp_param,struct flowi4 * fl4,struct ipcm_cookie * ipc,struct rtable ** rt)363 static void icmp_push_reply(struct sock *sk,
364 			    struct icmp_bxm *icmp_param,
365 			    struct flowi4 *fl4,
366 			    struct ipcm_cookie *ipc, struct rtable **rt)
367 {
368 	struct sk_buff *skb;
369 
370 	if (ip_append_data(sk, fl4, icmp_glue_bits, icmp_param,
371 			   icmp_param->data_len+icmp_param->head_len,
372 			   icmp_param->head_len,
373 			   ipc, rt, MSG_DONTWAIT) < 0) {
374 		__ICMP_INC_STATS(sock_net(sk), ICMP_MIB_OUTERRORS);
375 		ip_flush_pending_frames(sk);
376 	} else if ((skb = skb_peek(&sk->sk_write_queue)) != NULL) {
377 		struct icmphdr *icmph = icmp_hdr(skb);
378 		__wsum csum;
379 		struct sk_buff *skb1;
380 
381 		csum = csum_partial_copy_nocheck((void *)&icmp_param->data,
382 						 (char *)icmph,
383 						 icmp_param->head_len);
384 		skb_queue_walk(&sk->sk_write_queue, skb1) {
385 			csum = csum_add(csum, skb1->csum);
386 		}
387 		icmph->checksum = csum_fold(csum);
388 		skb->ip_summed = CHECKSUM_NONE;
389 		ip_push_pending_frames(sk, fl4);
390 	}
391 }
392 
393 /*
394  *	Driving logic for building and sending ICMP messages.
395  */
396 
icmp_reply(struct icmp_bxm * icmp_param,struct sk_buff * skb)397 static void icmp_reply(struct icmp_bxm *icmp_param, struct sk_buff *skb)
398 {
399 	struct ipcm_cookie ipc;
400 	struct rtable *rt = skb_rtable(skb);
401 	struct net *net = dev_net(rt->dst.dev);
402 	struct flowi4 fl4;
403 	struct sock *sk;
404 	struct inet_sock *inet;
405 	__be32 daddr, saddr;
406 	u32 mark = IP4_REPLY_MARK(net, skb->mark);
407 	int type = icmp_param->data.icmph.type;
408 	int code = icmp_param->data.icmph.code;
409 
410 	if (ip_options_echo(net, &icmp_param->replyopts.opt.opt, skb))
411 		return;
412 
413 	/* Needed by both icmp_global_allow and icmp_xmit_lock */
414 	local_bh_disable();
415 
416 	/* global icmp_msgs_per_sec */
417 	if (!icmpv4_global_allow(net, type, code))
418 		goto out_bh_enable;
419 
420 	sk = icmp_xmit_lock(net);
421 	if (!sk)
422 		goto out_bh_enable;
423 	inet = inet_sk(sk);
424 
425 	icmp_param->data.icmph.checksum = 0;
426 
427 	ipcm_init(&ipc);
428 	inet->tos = ip_hdr(skb)->tos;
429 	ipc.sockc.mark = mark;
430 	daddr = ipc.addr = ip_hdr(skb)->saddr;
431 	saddr = fib_compute_spec_dst(skb);
432 
433 	if (icmp_param->replyopts.opt.opt.optlen) {
434 		ipc.opt = &icmp_param->replyopts.opt;
435 		if (ipc.opt->opt.srr)
436 			daddr = icmp_param->replyopts.opt.opt.faddr;
437 	}
438 	memset(&fl4, 0, sizeof(fl4));
439 	fl4.daddr = daddr;
440 	fl4.saddr = saddr;
441 	fl4.flowi4_mark = mark;
442 	fl4.flowi4_uid = sock_net_uid(net, NULL);
443 	fl4.flowi4_tos = RT_TOS(ip_hdr(skb)->tos);
444 	fl4.flowi4_proto = IPPROTO_ICMP;
445 	fl4.flowi4_oif = l3mdev_master_ifindex(skb->dev);
446 	security_skb_classify_flow(skb, flowi4_to_flowi_common(&fl4));
447 	rt = ip_route_output_key(net, &fl4);
448 	if (IS_ERR(rt))
449 		goto out_unlock;
450 	if (icmpv4_xrlim_allow(net, rt, &fl4, type, code))
451 		icmp_push_reply(sk, icmp_param, &fl4, &ipc, &rt);
452 	ip_rt_put(rt);
453 out_unlock:
454 	icmp_xmit_unlock(sk);
455 out_bh_enable:
456 	local_bh_enable();
457 }
458 
459 /*
460  * The device used for looking up which routing table to use for sending an ICMP
461  * error is preferably the source whenever it is set, which should ensure the
462  * icmp error can be sent to the source host, else lookup using the routing
463  * table of the destination device, else use the main routing table (index 0).
464  */
icmp_get_route_lookup_dev(struct sk_buff * skb)465 static struct net_device *icmp_get_route_lookup_dev(struct sk_buff *skb)
466 {
467 	struct net_device *route_lookup_dev = NULL;
468 
469 	if (skb->dev)
470 		route_lookup_dev = skb->dev;
471 	else if (skb_dst(skb))
472 		route_lookup_dev = skb_dst(skb)->dev;
473 	return route_lookup_dev;
474 }
475 
icmp_route_lookup(struct net * net,struct flowi4 * fl4,struct sk_buff * skb_in,const struct iphdr * iph,__be32 saddr,u8 tos,u32 mark,int type,int code,struct icmp_bxm * param)476 static struct rtable *icmp_route_lookup(struct net *net,
477 					struct flowi4 *fl4,
478 					struct sk_buff *skb_in,
479 					const struct iphdr *iph,
480 					__be32 saddr, u8 tos, u32 mark,
481 					int type, int code,
482 					struct icmp_bxm *param)
483 {
484 	struct net_device *route_lookup_dev;
485 	struct rtable *rt, *rt2;
486 	struct flowi4 fl4_dec;
487 	int err;
488 
489 	memset(fl4, 0, sizeof(*fl4));
490 	fl4->daddr = (param->replyopts.opt.opt.srr ?
491 		      param->replyopts.opt.opt.faddr : iph->saddr);
492 	fl4->saddr = saddr;
493 	fl4->flowi4_mark = mark;
494 	fl4->flowi4_uid = sock_net_uid(net, NULL);
495 	fl4->flowi4_tos = RT_TOS(tos);
496 	fl4->flowi4_proto = IPPROTO_ICMP;
497 	fl4->fl4_icmp_type = type;
498 	fl4->fl4_icmp_code = code;
499 	route_lookup_dev = icmp_get_route_lookup_dev(skb_in);
500 	fl4->flowi4_oif = l3mdev_master_ifindex(route_lookup_dev);
501 
502 	security_skb_classify_flow(skb_in, flowi4_to_flowi_common(fl4));
503 	rt = ip_route_output_key_hash(net, fl4, skb_in);
504 	if (IS_ERR(rt))
505 		return rt;
506 
507 	/* No need to clone since we're just using its address. */
508 	rt2 = rt;
509 
510 	rt = (struct rtable *) xfrm_lookup(net, &rt->dst,
511 					   flowi4_to_flowi(fl4), NULL, 0);
512 	if (!IS_ERR(rt)) {
513 		if (rt != rt2)
514 			return rt;
515 	} else if (PTR_ERR(rt) == -EPERM) {
516 		rt = NULL;
517 	} else
518 		return rt;
519 
520 	err = xfrm_decode_session_reverse(skb_in, flowi4_to_flowi(&fl4_dec), AF_INET);
521 	if (err)
522 		goto relookup_failed;
523 
524 	if (inet_addr_type_dev_table(net, route_lookup_dev,
525 				     fl4_dec.saddr) == RTN_LOCAL) {
526 		rt2 = __ip_route_output_key(net, &fl4_dec);
527 		if (IS_ERR(rt2))
528 			err = PTR_ERR(rt2);
529 	} else {
530 		struct flowi4 fl4_2 = {};
531 		unsigned long orefdst;
532 
533 		fl4_2.daddr = fl4_dec.saddr;
534 		rt2 = ip_route_output_key(net, &fl4_2);
535 		if (IS_ERR(rt2)) {
536 			err = PTR_ERR(rt2);
537 			goto relookup_failed;
538 		}
539 		/* Ugh! */
540 		orefdst = skb_in->_skb_refdst; /* save old refdst */
541 		skb_dst_set(skb_in, NULL);
542 		err = ip_route_input(skb_in, fl4_dec.daddr, fl4_dec.saddr,
543 				     RT_TOS(tos), rt2->dst.dev);
544 
545 		dst_release(&rt2->dst);
546 		rt2 = skb_rtable(skb_in);
547 		skb_in->_skb_refdst = orefdst; /* restore old refdst */
548 	}
549 
550 	if (err)
551 		goto relookup_failed;
552 
553 	rt2 = (struct rtable *) xfrm_lookup(net, &rt2->dst,
554 					    flowi4_to_flowi(&fl4_dec), NULL,
555 					    XFRM_LOOKUP_ICMP);
556 	if (!IS_ERR(rt2)) {
557 		dst_release(&rt->dst);
558 		memcpy(fl4, &fl4_dec, sizeof(*fl4));
559 		rt = rt2;
560 	} else if (PTR_ERR(rt2) == -EPERM) {
561 		if (rt)
562 			dst_release(&rt->dst);
563 		return rt2;
564 	} else {
565 		err = PTR_ERR(rt2);
566 		goto relookup_failed;
567 	}
568 	return rt;
569 
570 relookup_failed:
571 	if (rt)
572 		return rt;
573 	return ERR_PTR(err);
574 }
575 
576 /*
577  *	Send an ICMP message in response to a situation
578  *
579  *	RFC 1122: 3.2.2	MUST send at least the IP header and 8 bytes of header.
580  *		  MAY send more (we do).
581  *			MUST NOT change this header information.
582  *			MUST NOT reply to a multicast/broadcast IP address.
583  *			MUST NOT reply to a multicast/broadcast MAC address.
584  *			MUST reply to only the first fragment.
585  */
586 
__icmp_send(struct sk_buff * skb_in,int type,int code,__be32 info,const struct ip_options * opt)587 void __icmp_send(struct sk_buff *skb_in, int type, int code, __be32 info,
588 		 const struct ip_options *opt)
589 {
590 	struct iphdr *iph;
591 	int room;
592 	struct icmp_bxm icmp_param;
593 	struct rtable *rt = skb_rtable(skb_in);
594 	struct ipcm_cookie ipc;
595 	struct flowi4 fl4;
596 	__be32 saddr;
597 	u8  tos;
598 	u32 mark;
599 	struct net *net;
600 	struct sock *sk;
601 
602 	if (!rt)
603 		goto out;
604 
605 	if (rt->dst.dev)
606 		net = dev_net(rt->dst.dev);
607 	else if (skb_in->dev)
608 		net = dev_net(skb_in->dev);
609 	else
610 		goto out;
611 
612 	/*
613 	 *	Find the original header. It is expected to be valid, of course.
614 	 *	Check this, icmp_send is called from the most obscure devices
615 	 *	sometimes.
616 	 */
617 	iph = ip_hdr(skb_in);
618 
619 	if ((u8 *)iph < skb_in->head ||
620 	    (skb_network_header(skb_in) + sizeof(*iph)) >
621 	    skb_tail_pointer(skb_in))
622 		goto out;
623 
624 	/*
625 	 *	No replies to physical multicast/broadcast
626 	 */
627 	if (skb_in->pkt_type != PACKET_HOST)
628 		goto out;
629 
630 	/*
631 	 *	Now check at the protocol level
632 	 */
633 	if (rt->rt_flags & (RTCF_BROADCAST | RTCF_MULTICAST))
634 		goto out;
635 
636 	/*
637 	 *	Only reply to fragment 0. We byte re-order the constant
638 	 *	mask for efficiency.
639 	 */
640 	if (iph->frag_off & htons(IP_OFFSET))
641 		goto out;
642 
643 	/*
644 	 *	If we send an ICMP error to an ICMP error a mess would result..
645 	 */
646 	if (icmp_pointers[type].error) {
647 		/*
648 		 *	We are an error, check if we are replying to an
649 		 *	ICMP error
650 		 */
651 		if (iph->protocol == IPPROTO_ICMP) {
652 			u8 _inner_type, *itp;
653 
654 			itp = skb_header_pointer(skb_in,
655 						 skb_network_header(skb_in) +
656 						 (iph->ihl << 2) +
657 						 offsetof(struct icmphdr,
658 							  type) -
659 						 skb_in->data,
660 						 sizeof(_inner_type),
661 						 &_inner_type);
662 			if (!itp)
663 				goto out;
664 
665 			/*
666 			 *	Assume any unknown ICMP type is an error. This
667 			 *	isn't specified by the RFC, but think about it..
668 			 */
669 			if (*itp > NR_ICMP_TYPES ||
670 			    icmp_pointers[*itp].error)
671 				goto out;
672 		}
673 	}
674 
675 	/* Needed by both icmp_global_allow and icmp_xmit_lock */
676 	local_bh_disable();
677 
678 	/* Check global sysctl_icmp_msgs_per_sec ratelimit, unless
679 	 * incoming dev is loopback.  If outgoing dev change to not be
680 	 * loopback, then peer ratelimit still work (in icmpv4_xrlim_allow)
681 	 */
682 	if (!(skb_in->dev && (skb_in->dev->flags&IFF_LOOPBACK)) &&
683 	      !icmpv4_global_allow(net, type, code))
684 		goto out_bh_enable;
685 
686 	sk = icmp_xmit_lock(net);
687 	if (!sk)
688 		goto out_bh_enable;
689 
690 	/*
691 	 *	Construct source address and options.
692 	 */
693 
694 	saddr = iph->daddr;
695 	if (!(rt->rt_flags & RTCF_LOCAL)) {
696 		struct net_device *dev = NULL;
697 
698 		rcu_read_lock();
699 		if (rt_is_input_route(rt) &&
700 		    READ_ONCE(net->ipv4.sysctl_icmp_errors_use_inbound_ifaddr))
701 			dev = dev_get_by_index_rcu(net, inet_iif(skb_in));
702 
703 		if (dev)
704 			saddr = inet_select_addr(dev, iph->saddr,
705 						 RT_SCOPE_LINK);
706 		else
707 			saddr = 0;
708 		rcu_read_unlock();
709 	}
710 
711 	tos = icmp_pointers[type].error ? (RT_TOS(iph->tos) |
712 					   IPTOS_PREC_INTERNETCONTROL) :
713 					   iph->tos;
714 	mark = IP4_REPLY_MARK(net, skb_in->mark);
715 
716 	if (__ip_options_echo(net, &icmp_param.replyopts.opt.opt, skb_in, opt))
717 		goto out_unlock;
718 
719 
720 	/*
721 	 *	Prepare data for ICMP header.
722 	 */
723 
724 	icmp_param.data.icmph.type	 = type;
725 	icmp_param.data.icmph.code	 = code;
726 	icmp_param.data.icmph.un.gateway = info;
727 	icmp_param.data.icmph.checksum	 = 0;
728 	icmp_param.skb	  = skb_in;
729 	icmp_param.offset = skb_network_offset(skb_in);
730 	inet_sk(sk)->tos = tos;
731 	ipcm_init(&ipc);
732 	ipc.addr = iph->saddr;
733 	ipc.opt = &icmp_param.replyopts.opt;
734 	ipc.sockc.mark = mark;
735 
736 	rt = icmp_route_lookup(net, &fl4, skb_in, iph, saddr, tos, mark,
737 			       type, code, &icmp_param);
738 	if (IS_ERR(rt))
739 		goto out_unlock;
740 
741 	/* peer icmp_ratelimit */
742 	if (!icmpv4_xrlim_allow(net, rt, &fl4, type, code))
743 		goto ende;
744 
745 	/* RFC says return as much as we can without exceeding 576 bytes. */
746 
747 	room = dst_mtu(&rt->dst);
748 	if (room > 576)
749 		room = 576;
750 	room -= sizeof(struct iphdr) + icmp_param.replyopts.opt.opt.optlen;
751 	room -= sizeof(struct icmphdr);
752 
753 	icmp_param.data_len = skb_in->len - icmp_param.offset;
754 	if (icmp_param.data_len > room)
755 		icmp_param.data_len = room;
756 	icmp_param.head_len = sizeof(struct icmphdr);
757 
758 	/* if we don't have a source address at this point, fall back to the
759 	 * dummy address instead of sending out a packet with a source address
760 	 * of 0.0.0.0
761 	 */
762 	if (!fl4.saddr)
763 		fl4.saddr = htonl(INADDR_DUMMY);
764 
765 	icmp_push_reply(sk, &icmp_param, &fl4, &ipc, &rt);
766 ende:
767 	ip_rt_put(rt);
768 out_unlock:
769 	icmp_xmit_unlock(sk);
770 out_bh_enable:
771 	local_bh_enable();
772 out:;
773 }
774 EXPORT_SYMBOL(__icmp_send);
775 
776 #if IS_ENABLED(CONFIG_NF_NAT)
777 #include <net/netfilter/nf_conntrack.h>
icmp_ndo_send(struct sk_buff * skb_in,int type,int code,__be32 info)778 void icmp_ndo_send(struct sk_buff *skb_in, int type, int code, __be32 info)
779 {
780 	struct sk_buff *cloned_skb = NULL;
781 	struct ip_options opts = { 0 };
782 	enum ip_conntrack_info ctinfo;
783 	struct nf_conn *ct;
784 	__be32 orig_ip;
785 
786 	ct = nf_ct_get(skb_in, &ctinfo);
787 	if (!ct || !(ct->status & IPS_SRC_NAT)) {
788 		__icmp_send(skb_in, type, code, info, &opts);
789 		return;
790 	}
791 
792 	if (skb_shared(skb_in))
793 		skb_in = cloned_skb = skb_clone(skb_in, GFP_ATOMIC);
794 
795 	if (unlikely(!skb_in || skb_network_header(skb_in) < skb_in->head ||
796 	    (skb_network_header(skb_in) + sizeof(struct iphdr)) >
797 	    skb_tail_pointer(skb_in) || skb_ensure_writable(skb_in,
798 	    skb_network_offset(skb_in) + sizeof(struct iphdr))))
799 		goto out;
800 
801 	orig_ip = ip_hdr(skb_in)->saddr;
802 	ip_hdr(skb_in)->saddr = ct->tuplehash[0].tuple.src.u3.ip;
803 	__icmp_send(skb_in, type, code, info, &opts);
804 	ip_hdr(skb_in)->saddr = orig_ip;
805 out:
806 	consume_skb(cloned_skb);
807 }
808 EXPORT_SYMBOL(icmp_ndo_send);
809 #endif
810 
icmp_socket_deliver(struct sk_buff * skb,u32 info)811 static void icmp_socket_deliver(struct sk_buff *skb, u32 info)
812 {
813 	const struct iphdr *iph = (const struct iphdr *)skb->data;
814 	const struct net_protocol *ipprot;
815 	int protocol = iph->protocol;
816 
817 	/* Checkin full IP header plus 8 bytes of protocol to
818 	 * avoid additional coding at protocol handlers.
819 	 */
820 	if (!pskb_may_pull(skb, iph->ihl * 4 + 8)) {
821 		__ICMP_INC_STATS(dev_net(skb->dev), ICMP_MIB_INERRORS);
822 		return;
823 	}
824 
825 	raw_icmp_error(skb, protocol, info);
826 
827 	ipprot = rcu_dereference(inet_protos[protocol]);
828 	if (ipprot && ipprot->err_handler)
829 		ipprot->err_handler(skb, info);
830 }
831 
icmp_tag_validation(int proto)832 static bool icmp_tag_validation(int proto)
833 {
834 	bool ok;
835 
836 	rcu_read_lock();
837 	ok = rcu_dereference(inet_protos[proto])->icmp_strict_tag_validation;
838 	rcu_read_unlock();
839 	return ok;
840 }
841 
842 /*
843  *	Handle ICMP_DEST_UNREACH, ICMP_TIME_EXCEEDED, ICMP_QUENCH, and
844  *	ICMP_PARAMETERPROB.
845  */
846 
icmp_unreach(struct sk_buff * skb)847 static enum skb_drop_reason icmp_unreach(struct sk_buff *skb)
848 {
849 	enum skb_drop_reason reason = SKB_NOT_DROPPED_YET;
850 	const struct iphdr *iph;
851 	struct icmphdr *icmph;
852 	struct net *net;
853 	u32 info = 0;
854 
855 	net = dev_net(skb_dst(skb)->dev);
856 
857 	/*
858 	 *	Incomplete header ?
859 	 * 	Only checks for the IP header, there should be an
860 	 *	additional check for longer headers in upper levels.
861 	 */
862 
863 	if (!pskb_may_pull(skb, sizeof(struct iphdr)))
864 		goto out_err;
865 
866 	icmph = icmp_hdr(skb);
867 	iph   = (const struct iphdr *)skb->data;
868 
869 	if (iph->ihl < 5)  { /* Mangled header, drop. */
870 		reason = SKB_DROP_REASON_IP_INHDR;
871 		goto out_err;
872 	}
873 
874 	switch (icmph->type) {
875 	case ICMP_DEST_UNREACH:
876 		switch (icmph->code & 15) {
877 		case ICMP_NET_UNREACH:
878 		case ICMP_HOST_UNREACH:
879 		case ICMP_PROT_UNREACH:
880 		case ICMP_PORT_UNREACH:
881 			break;
882 		case ICMP_FRAG_NEEDED:
883 			/* for documentation of the ip_no_pmtu_disc
884 			 * values please see
885 			 * Documentation/networking/ip-sysctl.rst
886 			 */
887 			switch (READ_ONCE(net->ipv4.sysctl_ip_no_pmtu_disc)) {
888 			default:
889 				net_dbg_ratelimited("%pI4: fragmentation needed and DF set\n",
890 						    &iph->daddr);
891 				break;
892 			case 2:
893 				goto out;
894 			case 3:
895 				if (!icmp_tag_validation(iph->protocol))
896 					goto out;
897 				fallthrough;
898 			case 0:
899 				info = ntohs(icmph->un.frag.mtu);
900 			}
901 			break;
902 		case ICMP_SR_FAILED:
903 			net_dbg_ratelimited("%pI4: Source Route Failed\n",
904 					    &iph->daddr);
905 			break;
906 		default:
907 			break;
908 		}
909 		if (icmph->code > NR_ICMP_UNREACH)
910 			goto out;
911 		break;
912 	case ICMP_PARAMETERPROB:
913 		info = ntohl(icmph->un.gateway) >> 24;
914 		break;
915 	case ICMP_TIME_EXCEEDED:
916 		__ICMP_INC_STATS(net, ICMP_MIB_INTIMEEXCDS);
917 		if (icmph->code == ICMP_EXC_FRAGTIME)
918 			goto out;
919 		break;
920 	}
921 
922 	/*
923 	 *	Throw it at our lower layers
924 	 *
925 	 *	RFC 1122: 3.2.2 MUST extract the protocol ID from the passed
926 	 *		  header.
927 	 *	RFC 1122: 3.2.2.1 MUST pass ICMP unreach messages to the
928 	 *		  transport layer.
929 	 *	RFC 1122: 3.2.2.2 MUST pass ICMP time expired messages to
930 	 *		  transport layer.
931 	 */
932 
933 	/*
934 	 *	Check the other end isn't violating RFC 1122. Some routers send
935 	 *	bogus responses to broadcast frames. If you see this message
936 	 *	first check your netmask matches at both ends, if it does then
937 	 *	get the other vendor to fix their kit.
938 	 */
939 
940 	if (!READ_ONCE(net->ipv4.sysctl_icmp_ignore_bogus_error_responses) &&
941 	    inet_addr_type_dev_table(net, skb->dev, iph->daddr) == RTN_BROADCAST) {
942 		net_warn_ratelimited("%pI4 sent an invalid ICMP type %u, code %u error to a broadcast: %pI4 on %s\n",
943 				     &ip_hdr(skb)->saddr,
944 				     icmph->type, icmph->code,
945 				     &iph->daddr, skb->dev->name);
946 		goto out;
947 	}
948 
949 	icmp_socket_deliver(skb, info);
950 
951 out:
952 	return reason;
953 out_err:
954 	__ICMP_INC_STATS(net, ICMP_MIB_INERRORS);
955 	return reason ?: SKB_DROP_REASON_NOT_SPECIFIED;
956 }
957 
958 
959 /*
960  *	Handle ICMP_REDIRECT.
961  */
962 
icmp_redirect(struct sk_buff * skb)963 static enum skb_drop_reason icmp_redirect(struct sk_buff *skb)
964 {
965 	if (skb->len < sizeof(struct iphdr)) {
966 		__ICMP_INC_STATS(dev_net(skb->dev), ICMP_MIB_INERRORS);
967 		return SKB_DROP_REASON_PKT_TOO_SMALL;
968 	}
969 
970 	if (!pskb_may_pull(skb, sizeof(struct iphdr))) {
971 		/* there aught to be a stat */
972 		return SKB_DROP_REASON_NOMEM;
973 	}
974 
975 	icmp_socket_deliver(skb, ntohl(icmp_hdr(skb)->un.gateway));
976 	return SKB_NOT_DROPPED_YET;
977 }
978 
979 /*
980  *	Handle ICMP_ECHO ("ping") and ICMP_EXT_ECHO ("PROBE") requests.
981  *
982  *	RFC 1122: 3.2.2.6 MUST have an echo server that answers ICMP echo
983  *		  requests.
984  *	RFC 1122: 3.2.2.6 Data received in the ICMP_ECHO request MUST be
985  *		  included in the reply.
986  *	RFC 1812: 4.3.3.6 SHOULD have a config option for silently ignoring
987  *		  echo requests, MUST have default=NOT.
988  *	RFC 8335: 8 MUST have a config option to enable/disable ICMP
989  *		  Extended Echo Functionality, MUST be disabled by default
990  *	See also WRT handling of options once they are done and working.
991  */
992 
icmp_echo(struct sk_buff * skb)993 static enum skb_drop_reason icmp_echo(struct sk_buff *skb)
994 {
995 	struct icmp_bxm icmp_param;
996 	struct net *net;
997 
998 	net = dev_net(skb_dst(skb)->dev);
999 	/* should there be an ICMP stat for ignored echos? */
1000 	if (READ_ONCE(net->ipv4.sysctl_icmp_echo_ignore_all))
1001 		return SKB_NOT_DROPPED_YET;
1002 
1003 	icmp_param.data.icmph	   = *icmp_hdr(skb);
1004 	icmp_param.skb		   = skb;
1005 	icmp_param.offset	   = 0;
1006 	icmp_param.data_len	   = skb->len;
1007 	icmp_param.head_len	   = sizeof(struct icmphdr);
1008 
1009 	if (icmp_param.data.icmph.type == ICMP_ECHO)
1010 		icmp_param.data.icmph.type = ICMP_ECHOREPLY;
1011 	else if (!icmp_build_probe(skb, &icmp_param.data.icmph))
1012 		return SKB_NOT_DROPPED_YET;
1013 
1014 	icmp_reply(&icmp_param, skb);
1015 	return SKB_NOT_DROPPED_YET;
1016 }
1017 
1018 /*	Helper for icmp_echo and icmpv6_echo_reply.
1019  *	Searches for net_device that matches PROBE interface identifier
1020  *		and builds PROBE reply message in icmphdr.
1021  *
1022  *	Returns false if PROBE responses are disabled via sysctl
1023  */
1024 
icmp_build_probe(struct sk_buff * skb,struct icmphdr * icmphdr)1025 bool icmp_build_probe(struct sk_buff *skb, struct icmphdr *icmphdr)
1026 {
1027 	struct icmp_ext_hdr *ext_hdr, _ext_hdr;
1028 	struct icmp_ext_echo_iio *iio, _iio;
1029 	struct net *net = dev_net(skb->dev);
1030 	struct net_device *dev;
1031 	char buff[IFNAMSIZ];
1032 	u16 ident_len;
1033 	u8 status;
1034 
1035 	if (!READ_ONCE(net->ipv4.sysctl_icmp_echo_enable_probe))
1036 		return false;
1037 
1038 	/* We currently only support probing interfaces on the proxy node
1039 	 * Check to ensure L-bit is set
1040 	 */
1041 	if (!(ntohs(icmphdr->un.echo.sequence) & 1))
1042 		return false;
1043 	/* Clear status bits in reply message */
1044 	icmphdr->un.echo.sequence &= htons(0xFF00);
1045 	if (icmphdr->type == ICMP_EXT_ECHO)
1046 		icmphdr->type = ICMP_EXT_ECHOREPLY;
1047 	else
1048 		icmphdr->type = ICMPV6_EXT_ECHO_REPLY;
1049 	ext_hdr = skb_header_pointer(skb, 0, sizeof(_ext_hdr), &_ext_hdr);
1050 	/* Size of iio is class_type dependent.
1051 	 * Only check header here and assign length based on ctype in the switch statement
1052 	 */
1053 	iio = skb_header_pointer(skb, sizeof(_ext_hdr), sizeof(iio->extobj_hdr), &_iio);
1054 	if (!ext_hdr || !iio)
1055 		goto send_mal_query;
1056 	if (ntohs(iio->extobj_hdr.length) <= sizeof(iio->extobj_hdr) ||
1057 	    ntohs(iio->extobj_hdr.length) > sizeof(_iio))
1058 		goto send_mal_query;
1059 	ident_len = ntohs(iio->extobj_hdr.length) - sizeof(iio->extobj_hdr);
1060 	iio = skb_header_pointer(skb, sizeof(_ext_hdr),
1061 				 sizeof(iio->extobj_hdr) + ident_len, &_iio);
1062 	if (!iio)
1063 		goto send_mal_query;
1064 
1065 	status = 0;
1066 	dev = NULL;
1067 	switch (iio->extobj_hdr.class_type) {
1068 	case ICMP_EXT_ECHO_CTYPE_NAME:
1069 		if (ident_len >= IFNAMSIZ)
1070 			goto send_mal_query;
1071 		memset(buff, 0, sizeof(buff));
1072 		memcpy(buff, &iio->ident.name, ident_len);
1073 		dev = dev_get_by_name(net, buff);
1074 		break;
1075 	case ICMP_EXT_ECHO_CTYPE_INDEX:
1076 		if (ident_len != sizeof(iio->ident.ifindex))
1077 			goto send_mal_query;
1078 		dev = dev_get_by_index(net, ntohl(iio->ident.ifindex));
1079 		break;
1080 	case ICMP_EXT_ECHO_CTYPE_ADDR:
1081 		if (ident_len < sizeof(iio->ident.addr.ctype3_hdr) ||
1082 		    ident_len != sizeof(iio->ident.addr.ctype3_hdr) +
1083 				 iio->ident.addr.ctype3_hdr.addrlen)
1084 			goto send_mal_query;
1085 		switch (ntohs(iio->ident.addr.ctype3_hdr.afi)) {
1086 		case ICMP_AFI_IP:
1087 			if (iio->ident.addr.ctype3_hdr.addrlen != sizeof(struct in_addr))
1088 				goto send_mal_query;
1089 			dev = ip_dev_find(net, iio->ident.addr.ip_addr.ipv4_addr);
1090 			break;
1091 #if IS_ENABLED(CONFIG_IPV6)
1092 		case ICMP_AFI_IP6:
1093 			if (iio->ident.addr.ctype3_hdr.addrlen != sizeof(struct in6_addr))
1094 				goto send_mal_query;
1095 			dev = ipv6_stub->ipv6_dev_find(net, &iio->ident.addr.ip_addr.ipv6_addr, dev);
1096 			dev_hold(dev);
1097 			break;
1098 #endif
1099 		default:
1100 			goto send_mal_query;
1101 		}
1102 		break;
1103 	default:
1104 		goto send_mal_query;
1105 	}
1106 	if (!dev) {
1107 		icmphdr->code = ICMP_EXT_CODE_NO_IF;
1108 		return true;
1109 	}
1110 	/* Fill bits in reply message */
1111 	if (dev->flags & IFF_UP)
1112 		status |= ICMP_EXT_ECHOREPLY_ACTIVE;
1113 	if (__in_dev_get_rcu(dev) && __in_dev_get_rcu(dev)->ifa_list)
1114 		status |= ICMP_EXT_ECHOREPLY_IPV4;
1115 	if (!list_empty(&rcu_dereference(dev->ip6_ptr)->addr_list))
1116 		status |= ICMP_EXT_ECHOREPLY_IPV6;
1117 	dev_put(dev);
1118 	icmphdr->un.echo.sequence |= htons(status);
1119 	return true;
1120 send_mal_query:
1121 	icmphdr->code = ICMP_EXT_CODE_MAL_QUERY;
1122 	return true;
1123 }
1124 EXPORT_SYMBOL_GPL(icmp_build_probe);
1125 
1126 /*
1127  *	Handle ICMP Timestamp requests.
1128  *	RFC 1122: 3.2.2.8 MAY implement ICMP timestamp requests.
1129  *		  SHOULD be in the kernel for minimum random latency.
1130  *		  MUST be accurate to a few minutes.
1131  *		  MUST be updated at least at 15Hz.
1132  */
icmp_timestamp(struct sk_buff * skb)1133 static enum skb_drop_reason icmp_timestamp(struct sk_buff *skb)
1134 {
1135 	struct icmp_bxm icmp_param;
1136 	/*
1137 	 *	Too short.
1138 	 */
1139 	if (skb->len < 4)
1140 		goto out_err;
1141 
1142 	/*
1143 	 *	Fill in the current time as ms since midnight UT:
1144 	 */
1145 	icmp_param.data.times[1] = inet_current_timestamp();
1146 	icmp_param.data.times[2] = icmp_param.data.times[1];
1147 
1148 	BUG_ON(skb_copy_bits(skb, 0, &icmp_param.data.times[0], 4));
1149 
1150 	icmp_param.data.icmph	   = *icmp_hdr(skb);
1151 	icmp_param.data.icmph.type = ICMP_TIMESTAMPREPLY;
1152 	icmp_param.data.icmph.code = 0;
1153 	icmp_param.skb		   = skb;
1154 	icmp_param.offset	   = 0;
1155 	icmp_param.data_len	   = 0;
1156 	icmp_param.head_len	   = sizeof(struct icmphdr) + 12;
1157 	icmp_reply(&icmp_param, skb);
1158 	return SKB_NOT_DROPPED_YET;
1159 
1160 out_err:
1161 	__ICMP_INC_STATS(dev_net(skb_dst(skb)->dev), ICMP_MIB_INERRORS);
1162 	return SKB_DROP_REASON_PKT_TOO_SMALL;
1163 }
1164 
icmp_discard(struct sk_buff * skb)1165 static enum skb_drop_reason icmp_discard(struct sk_buff *skb)
1166 {
1167 	/* pretend it was a success */
1168 	return SKB_NOT_DROPPED_YET;
1169 }
1170 
1171 /*
1172  *	Deal with incoming ICMP packets.
1173  */
icmp_rcv(struct sk_buff * skb)1174 int icmp_rcv(struct sk_buff *skb)
1175 {
1176 	enum skb_drop_reason reason = SKB_DROP_REASON_NOT_SPECIFIED;
1177 	struct rtable *rt = skb_rtable(skb);
1178 	struct net *net = dev_net(rt->dst.dev);
1179 	struct icmphdr *icmph;
1180 
1181 	if (!xfrm4_policy_check(NULL, XFRM_POLICY_IN, skb)) {
1182 		struct sec_path *sp = skb_sec_path(skb);
1183 		int nh;
1184 
1185 		if (!(sp && sp->xvec[sp->len - 1]->props.flags &
1186 				 XFRM_STATE_ICMP)) {
1187 			reason = SKB_DROP_REASON_XFRM_POLICY;
1188 			goto drop;
1189 		}
1190 
1191 		if (!pskb_may_pull(skb, sizeof(*icmph) + sizeof(struct iphdr)))
1192 			goto drop;
1193 
1194 		nh = skb_network_offset(skb);
1195 		skb_set_network_header(skb, sizeof(*icmph));
1196 
1197 		if (!xfrm4_policy_check_reverse(NULL, XFRM_POLICY_IN,
1198 						skb)) {
1199 			reason = SKB_DROP_REASON_XFRM_POLICY;
1200 			goto drop;
1201 		}
1202 
1203 		skb_set_network_header(skb, nh);
1204 	}
1205 
1206 	__ICMP_INC_STATS(net, ICMP_MIB_INMSGS);
1207 
1208 	if (skb_checksum_simple_validate(skb))
1209 		goto csum_error;
1210 
1211 	if (!pskb_pull(skb, sizeof(*icmph)))
1212 		goto error;
1213 
1214 	icmph = icmp_hdr(skb);
1215 
1216 	ICMPMSGIN_INC_STATS(net, icmph->type);
1217 
1218 	/* Check for ICMP Extended Echo (PROBE) messages */
1219 	if (icmph->type == ICMP_EXT_ECHO) {
1220 		/* We can't use icmp_pointers[].handler() because it is an array of
1221 		 * size NR_ICMP_TYPES + 1 (19 elements) and PROBE has code 42.
1222 		 */
1223 		reason = icmp_echo(skb);
1224 		goto reason_check;
1225 	}
1226 
1227 	if (icmph->type == ICMP_EXT_ECHOREPLY) {
1228 		reason = ping_rcv(skb);
1229 		goto reason_check;
1230 	}
1231 
1232 	/*
1233 	 *	18 is the highest 'known' ICMP type. Anything else is a mystery
1234 	 *
1235 	 *	RFC 1122: 3.2.2  Unknown ICMP messages types MUST be silently
1236 	 *		  discarded.
1237 	 */
1238 	if (icmph->type > NR_ICMP_TYPES) {
1239 		reason = SKB_DROP_REASON_UNHANDLED_PROTO;
1240 		goto error;
1241 	}
1242 
1243 	/*
1244 	 *	Parse the ICMP message
1245 	 */
1246 
1247 	if (rt->rt_flags & (RTCF_BROADCAST | RTCF_MULTICAST)) {
1248 		/*
1249 		 *	RFC 1122: 3.2.2.6 An ICMP_ECHO to broadcast MAY be
1250 		 *	  silently ignored (we let user decide with a sysctl).
1251 		 *	RFC 1122: 3.2.2.8 An ICMP_TIMESTAMP MAY be silently
1252 		 *	  discarded if to broadcast/multicast.
1253 		 */
1254 		if ((icmph->type == ICMP_ECHO ||
1255 		     icmph->type == ICMP_TIMESTAMP) &&
1256 		    READ_ONCE(net->ipv4.sysctl_icmp_echo_ignore_broadcasts)) {
1257 			reason = SKB_DROP_REASON_INVALID_PROTO;
1258 			goto error;
1259 		}
1260 		if (icmph->type != ICMP_ECHO &&
1261 		    icmph->type != ICMP_TIMESTAMP &&
1262 		    icmph->type != ICMP_ADDRESS &&
1263 		    icmph->type != ICMP_ADDRESSREPLY) {
1264 			reason = SKB_DROP_REASON_INVALID_PROTO;
1265 			goto error;
1266 		}
1267 	}
1268 
1269 	reason = icmp_pointers[icmph->type].handler(skb);
1270 reason_check:
1271 	if (!reason)  {
1272 		consume_skb(skb);
1273 		return NET_RX_SUCCESS;
1274 	}
1275 
1276 drop:
1277 	kfree_skb_reason(skb, reason);
1278 	return NET_RX_DROP;
1279 csum_error:
1280 	reason = SKB_DROP_REASON_ICMP_CSUM;
1281 	__ICMP_INC_STATS(net, ICMP_MIB_CSUMERRORS);
1282 error:
1283 	__ICMP_INC_STATS(net, ICMP_MIB_INERRORS);
1284 	goto drop;
1285 }
1286 
ip_icmp_error_rfc4884_validate(const struct sk_buff * skb,int off)1287 static bool ip_icmp_error_rfc4884_validate(const struct sk_buff *skb, int off)
1288 {
1289 	struct icmp_extobj_hdr *objh, _objh;
1290 	struct icmp_ext_hdr *exth, _exth;
1291 	u16 olen;
1292 
1293 	exth = skb_header_pointer(skb, off, sizeof(_exth), &_exth);
1294 	if (!exth)
1295 		return false;
1296 	if (exth->version != 2)
1297 		return true;
1298 
1299 	if (exth->checksum &&
1300 	    csum_fold(skb_checksum(skb, off, skb->len - off, 0)))
1301 		return false;
1302 
1303 	off += sizeof(_exth);
1304 	while (off < skb->len) {
1305 		objh = skb_header_pointer(skb, off, sizeof(_objh), &_objh);
1306 		if (!objh)
1307 			return false;
1308 
1309 		olen = ntohs(objh->length);
1310 		if (olen < sizeof(_objh))
1311 			return false;
1312 
1313 		off += olen;
1314 		if (off > skb->len)
1315 			return false;
1316 	}
1317 
1318 	return true;
1319 }
1320 
ip_icmp_error_rfc4884(const struct sk_buff * skb,struct sock_ee_data_rfc4884 * out,int thlen,int off)1321 void ip_icmp_error_rfc4884(const struct sk_buff *skb,
1322 			   struct sock_ee_data_rfc4884 *out,
1323 			   int thlen, int off)
1324 {
1325 	int hlen;
1326 
1327 	/* original datagram headers: end of icmph to payload (skb->data) */
1328 	hlen = -skb_transport_offset(skb) - thlen;
1329 
1330 	/* per rfc 4884: minimal datagram length of 128 bytes */
1331 	if (off < 128 || off < hlen)
1332 		return;
1333 
1334 	/* kernel has stripped headers: return payload offset in bytes */
1335 	off -= hlen;
1336 	if (off + sizeof(struct icmp_ext_hdr) > skb->len)
1337 		return;
1338 
1339 	out->len = off;
1340 
1341 	if (!ip_icmp_error_rfc4884_validate(skb, off))
1342 		out->flags |= SO_EE_RFC4884_FLAG_INVALID;
1343 }
1344 EXPORT_SYMBOL_GPL(ip_icmp_error_rfc4884);
1345 
icmp_err(struct sk_buff * skb,u32 info)1346 int icmp_err(struct sk_buff *skb, u32 info)
1347 {
1348 	struct iphdr *iph = (struct iphdr *)skb->data;
1349 	int offset = iph->ihl<<2;
1350 	struct icmphdr *icmph = (struct icmphdr *)(skb->data + offset);
1351 	int type = icmp_hdr(skb)->type;
1352 	int code = icmp_hdr(skb)->code;
1353 	struct net *net = dev_net(skb->dev);
1354 
1355 	/*
1356 	 * Use ping_err to handle all icmp errors except those
1357 	 * triggered by ICMP_ECHOREPLY which sent from kernel.
1358 	 */
1359 	if (icmph->type != ICMP_ECHOREPLY) {
1360 		ping_err(skb, offset, info);
1361 		return 0;
1362 	}
1363 
1364 	if (type == ICMP_DEST_UNREACH && code == ICMP_FRAG_NEEDED)
1365 		ipv4_update_pmtu(skb, net, info, 0, IPPROTO_ICMP);
1366 	else if (type == ICMP_REDIRECT)
1367 		ipv4_redirect(skb, net, 0, IPPROTO_ICMP);
1368 
1369 	return 0;
1370 }
1371 
1372 /*
1373  *	This table is the definition of how we handle ICMP.
1374  */
1375 static const struct icmp_control icmp_pointers[NR_ICMP_TYPES + 1] = {
1376 	[ICMP_ECHOREPLY] = {
1377 		.handler = ping_rcv,
1378 	},
1379 	[1] = {
1380 		.handler = icmp_discard,
1381 		.error = 1,
1382 	},
1383 	[2] = {
1384 		.handler = icmp_discard,
1385 		.error = 1,
1386 	},
1387 	[ICMP_DEST_UNREACH] = {
1388 		.handler = icmp_unreach,
1389 		.error = 1,
1390 	},
1391 	[ICMP_SOURCE_QUENCH] = {
1392 		.handler = icmp_unreach,
1393 		.error = 1,
1394 	},
1395 	[ICMP_REDIRECT] = {
1396 		.handler = icmp_redirect,
1397 		.error = 1,
1398 	},
1399 	[6] = {
1400 		.handler = icmp_discard,
1401 		.error = 1,
1402 	},
1403 	[7] = {
1404 		.handler = icmp_discard,
1405 		.error = 1,
1406 	},
1407 	[ICMP_ECHO] = {
1408 		.handler = icmp_echo,
1409 	},
1410 	[9] = {
1411 		.handler = icmp_discard,
1412 		.error = 1,
1413 	},
1414 	[10] = {
1415 		.handler = icmp_discard,
1416 		.error = 1,
1417 	},
1418 	[ICMP_TIME_EXCEEDED] = {
1419 		.handler = icmp_unreach,
1420 		.error = 1,
1421 	},
1422 	[ICMP_PARAMETERPROB] = {
1423 		.handler = icmp_unreach,
1424 		.error = 1,
1425 	},
1426 	[ICMP_TIMESTAMP] = {
1427 		.handler = icmp_timestamp,
1428 	},
1429 	[ICMP_TIMESTAMPREPLY] = {
1430 		.handler = icmp_discard,
1431 	},
1432 	[ICMP_INFO_REQUEST] = {
1433 		.handler = icmp_discard,
1434 	},
1435 	[ICMP_INFO_REPLY] = {
1436 		.handler = icmp_discard,
1437 	},
1438 	[ICMP_ADDRESS] = {
1439 		.handler = icmp_discard,
1440 	},
1441 	[ICMP_ADDRESSREPLY] = {
1442 		.handler = icmp_discard,
1443 	},
1444 };
1445 
icmp_sk_init(struct net * net)1446 static int __net_init icmp_sk_init(struct net *net)
1447 {
1448 	/* Control parameters for ECHO replies. */
1449 	net->ipv4.sysctl_icmp_echo_ignore_all = 0;
1450 	net->ipv4.sysctl_icmp_echo_enable_probe = 0;
1451 	net->ipv4.sysctl_icmp_echo_ignore_broadcasts = 1;
1452 
1453 	/* Control parameter - ignore bogus broadcast responses? */
1454 	net->ipv4.sysctl_icmp_ignore_bogus_error_responses = 1;
1455 
1456 	/*
1457 	 * 	Configurable global rate limit.
1458 	 *
1459 	 *	ratelimit defines tokens/packet consumed for dst->rate_token
1460 	 *	bucket ratemask defines which icmp types are ratelimited by
1461 	 *	setting	it's bit position.
1462 	 *
1463 	 *	default:
1464 	 *	dest unreachable (3), source quench (4),
1465 	 *	time exceeded (11), parameter problem (12)
1466 	 */
1467 
1468 	net->ipv4.sysctl_icmp_ratelimit = 1 * HZ;
1469 	net->ipv4.sysctl_icmp_ratemask = 0x1818;
1470 	net->ipv4.sysctl_icmp_errors_use_inbound_ifaddr = 0;
1471 
1472 	return 0;
1473 }
1474 
1475 static struct pernet_operations __net_initdata icmp_sk_ops = {
1476        .init = icmp_sk_init,
1477 };
1478 
icmp_init(void)1479 int __init icmp_init(void)
1480 {
1481 	int err, i;
1482 
1483 	for_each_possible_cpu(i) {
1484 		struct sock *sk;
1485 
1486 		err = inet_ctl_sock_create(&sk, PF_INET,
1487 					   SOCK_RAW, IPPROTO_ICMP, &init_net);
1488 		if (err < 0)
1489 			return err;
1490 
1491 		per_cpu(ipv4_icmp_sk, i) = sk;
1492 
1493 		/* Enough space for 2 64K ICMP packets, including
1494 		 * sk_buff/skb_shared_info struct overhead.
1495 		 */
1496 		sk->sk_sndbuf =	2 * SKB_TRUESIZE(64 * 1024);
1497 
1498 		/*
1499 		 * Speedup sock_wfree()
1500 		 */
1501 		sock_set_flag(sk, SOCK_USE_WRITE_QUEUE);
1502 		inet_sk(sk)->pmtudisc = IP_PMTUDISC_DONT;
1503 	}
1504 	return register_pernet_subsys(&icmp_sk_ops);
1505 }
1506