aboutsummaryrefslogtreecommitdiff
path: root/networking/zcip.c
blob: 3b67a032e93162631a2b70047afcc7c4f4663f3e (plain)
1
2
3
4
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
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
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
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
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
533
534
535
536
537
538
539
540
541
542
543
544
545
546
/* vi: set sw=4 ts=4: */
/*
 * RFC3927 ZeroConf IPv4 Link-Local addressing
 * (see <http://www.zeroconf.org/>)
 *
 * Copyright (C) 2003 by Arthur van Hoff (avh@strangeberry.com)
 * Copyright (C) 2004 by David Brownell
 *
 * Licensed under the GPL v2 or later, see the file LICENSE in this tarball.
 */

/*
 * ZCIP just manages the 169.254.*.* addresses.  That network is not
 * routed at the IP level, though various proxies or bridges can
 * certainly be used.  Its naming is built over multicast DNS.
 */

//#define DEBUG

// TODO:
// - more real-world usage/testing, especially daemon mode
// - kernel packet filters to reduce scheduling noise
// - avoid silent script failures, especially under load...
// - link status monitoring (restart on link-up; stop on link-down)

#include "libbb.h"
#include <syslog.h>
#include <poll.h>
#include <sys/wait.h>
#include <netinet/ether.h>
#include <net/ethernet.h>
#include <net/if.h>
#include <net/if_arp.h>

#include <linux/if_packet.h>
#include <linux/sockios.h>


struct arp_packet {
	struct ether_header hdr;
	struct ether_arp arp;
} ATTRIBUTE_PACKED;

enum {
/* 169.254.0.0 */
	LINKLOCAL_ADDR = 0xa9fe0000,

/* protocol timeout parameters, specified in seconds */
	PROBE_WAIT = 1,
	PROBE_MIN = 1,
	PROBE_MAX = 2,
	PROBE_NUM = 3,
	MAX_CONFLICTS = 10,
	RATE_LIMIT_INTERVAL = 60,
	ANNOUNCE_WAIT = 2,
	ANNOUNCE_NUM = 2,
	ANNOUNCE_INTERVAL = 2,
	DEFEND_INTERVAL = 10
};

/* States during the configuration process. */
enum {
	PROBE = 0,
	RATE_LIMIT_PROBE,
	ANNOUNCE,
	MONITOR,
	DEFEND
};

#define VDBG(fmt,args...) \
	do { } while (0)

/**
 * Pick a random link local IP address on 169.254/16, except that
 * the first and last 256 addresses are reserved.
 */
static void pick(struct in_addr *ip)
{
	unsigned tmp;

	/* use cheaper math than lrand48() mod N */
	do {
		tmp = (lrand48() >> 16) & IN_CLASSB_HOST;
	} while (tmp > (IN_CLASSB_HOST - 0x0200));
	ip->s_addr = htonl((LINKLOCAL_ADDR + 0x0100) + tmp);
}

/* TODO: we need a flag to direct bb_[p]error_msg output to stderr. */

/**
 * Broadcast an ARP packet.
 */
static void arp(int fd, struct sockaddr *saddr, int op,
	const struct ether_addr *source_addr, struct in_addr source_ip,
	const struct ether_addr *target_addr, struct in_addr target_ip)
{
	struct arp_packet p;
	memset(&p, 0, sizeof(p));

	// ether header
	p.hdr.ether_type = htons(ETHERTYPE_ARP);
	memcpy(p.hdr.ether_shost, source_addr, ETH_ALEN);
	memset(p.hdr.ether_dhost, 0xff, ETH_ALEN);

	// arp request
	p.arp.arp_hrd = htons(ARPHRD_ETHER);
	p.arp.arp_pro = htons(ETHERTYPE_IP);
	p.arp.arp_hln = ETH_ALEN;
	p.arp.arp_pln = 4;
	p.arp.arp_op = htons(op);
	memcpy(&p.arp.arp_sha, source_addr, ETH_ALEN);
	memcpy(&p.arp.arp_spa, &source_ip, sizeof(p.arp.arp_spa));
	memcpy(&p.arp.arp_tha, target_addr, ETH_ALEN);
	memcpy(&p.arp.arp_tpa, &target_ip, sizeof(p.arp.arp_tpa));

	// send it
	xsendto(fd, &p, sizeof(p), saddr, sizeof(*saddr));

	// Currently all callers ignore errors, that's why returns are
	// commented out...
	//return 0;
}

/**
 * Run a script. argv[2] is already NULL.
 */
static int run(char *argv[3], const char *intf, struct in_addr *ip)
{
	int status;

	VDBG("%s run %s %s\n", intf, argv[0], argv[1]);

	if (ip) {
		char *addr = inet_ntoa(*ip);
		setenv("ip", addr, 1);
		bb_info_msg("%s %s %s", argv[1], intf, addr);
	}

	status = wait4pid(spawn(argv));
	if (status < 0) {
		bb_perror_msg("%s %s", argv[1], intf);
		return -errno;
	}
	if (status != 0)
		bb_error_msg("script %s %s failed, exitcode=%d", argv[0], argv[1], status);
	return status;
}

/**
 * Return milliseconds of random delay, up to "secs" seconds.
 */
static unsigned ATTRIBUTE_ALWAYS_INLINE ms_rdelay(unsigned secs)
{
	return lrand48() % (secs * 1000);
}

/**
 * main program
 */
int zcip_main(int argc, char **argv);
int zcip_main(int argc, char **argv)
{
	int state = PROBE;
	struct ether_addr eth_addr;
	const char *why;
	int fd;
	char *r_opt;
	unsigned opts;

	/* Ugly trick, but I want these zeroed in one go */
	struct {
		const struct in_addr null_ip;
		const struct ether_addr null_addr;
		struct sockaddr saddr;
		struct in_addr ip;
		struct ifreq ifr;
		char *intf;
		char *script_av[3];
		suseconds_t timeout; // milliseconds
		unsigned conflicts;
		unsigned nprobes;
		unsigned nclaims;
		int ready;
		int verbose;
	} L;
#define null_ip   (L.null_ip  )
#define null_addr (L.null_addr)
#define saddr     (L.saddr    )
#define ip        (L.ip       )
#define ifr       (L.ifr      )
#define intf      (L.intf     )
#define script_av (L.script_av)
#define timeout   (L.timeout  )
#define conflicts (L.conflicts)
#define nprobes   (L.nprobes  )
#define nclaims   (L.nclaims  )
#define ready     (L.ready    )
#define verbose   (L.verbose  )

	memset(&L, 0, sizeof(L));

#define FOREGROUND (opts & 1)
#define QUIT       (opts & 2)
	// parse commandline: prog [options] ifname script
	// exactly 2 args; -v accumulates and implies -f
	opt_complementary = "=2:vv:vf";
	opts = getopt32(argc, argv, "fqr:v", &r_opt, &verbose);
	if (!FOREGROUND) {
		/* Do it early, before all bb_xx_msg calls */
		openlog(applet_name, 0, LOG_DAEMON);
		logmode |= LOGMODE_SYSLOG;
	}
	if (opts & 4) { // -r n.n.n.n
		if (inet_aton(r_opt, &ip) == 0
		 || (ntohl(ip.s_addr) & IN_CLASSB_NET) != LINKLOCAL_ADDR
		) {
			bb_error_msg_and_die("invalid link address");
		}
	}
	// On NOMMU reexec early (or else we will rerun things twice)
#if !BB_MMU
	if (!FOREGROUND)
		bb_daemonize_or_rexec(DAEMON_CHDIR_ROOT, argv);
#endif
	argc -= optind;
	argv += optind;

	intf = argv[0];
	script_av[0] = argv[1];
	setenv("interface", intf, 1);

	// initialize the interface (modprobe, ifup, etc)
	script_av[1] = (char*)"init";
	if (run(script_av, intf, NULL))
		return EXIT_FAILURE;

	// initialize saddr
	//memset(&saddr, 0, sizeof(saddr));
	safe_strncpy(saddr.sa_data, intf, sizeof(saddr.sa_data));

	// open an ARP socket
	fd = xsocket(PF_PACKET, SOCK_PACKET, htons(ETH_P_ARP));
	// bind to the interface's ARP socket
	xbind(fd, &saddr, sizeof(saddr));

	// get the interface's ethernet address
	//memset(&ifr, 0, sizeof(ifr));
	strncpy(ifr.ifr_name, intf, sizeof(ifr.ifr_name));
	if (ioctl(fd, SIOCGIFHWADDR, &ifr) < 0) {
		bb_perror_msg_and_die("get ethernet address");
	}
	memcpy(&eth_addr, &ifr.ifr_hwaddr.sa_data, ETH_ALEN);

	// start with some stable ip address, either a function of
	// the hardware address or else the last address we used.
	// NOTE: the sequence of addresses we try changes only
	// depending on when we detect conflicts.
	// (SVID 3 bogon: who says that "short" is always 16 bits?)
	seed48( (unsigned short*)&ifr.ifr_hwaddr.sa_data );
	if (ip.s_addr == 0)
		pick(&ip);

	// FIXME cases to handle:
	//  - zcip already running!
	//  - link already has local address... just defend/update

	// daemonize now; don't delay system startup
	if (!FOREGROUND) {
#if BB_MMU
		bb_daemonize(DAEMON_CHDIR_ROOT);
#endif
		bb_info_msg("start, interface %s", intf);
	}

	// run the dynamic address negotiation protocol,
	// restarting after address conflicts:
	//  - start with some address we want to try
	//  - short random delay
	//  - arp probes to see if another host else uses it
	//  - arp announcements that we're claiming it
	//  - use it
	//  - defend it, within limits
	while (1) {
		struct pollfd fds[1];
		struct timeval tv1;
		struct arp_packet p;

		int source_ip_conflict = 0;
		int target_ip_conflict = 0;

		fds[0].fd = fd;
		fds[0].events = POLLIN;
		fds[0].revents = 0;

		// poll, being ready to adjust current timeout
		if (!timeout) {
			timeout = ms_rdelay(PROBE_WAIT);
			// FIXME setsockopt(fd, SO_ATTACH_FILTER, ...) to
			// make the kernel filter out all packets except
			// ones we'd care about.
		}
		// set tv1 to the point in time when we timeout
		gettimeofday(&tv1, NULL);
		tv1.tv_usec += (timeout % 1000) * 1000;
		while (tv1.tv_usec > 1000000) {
			tv1.tv_usec -= 1000000;
			tv1.tv_sec++;
		}
		tv1.tv_sec += timeout / 1000;

		VDBG("...wait %ld %s nprobes=%d, nclaims=%d\n",
				timeout, intf, nprobes, nclaims);
		switch (poll(fds, 1, timeout)) {

		// timeout
		case 0:
			VDBG("state = %d\n", state);
			switch (state) {
			case PROBE:
				// timeouts in the PROBE state mean no conflicting ARP packets
				// have been received, so we can progress through the states
				if (nprobes < PROBE_NUM) {
					nprobes++;
					VDBG("probe/%d %s@%s\n",
							nprobes, intf, inet_ntoa(ip));
					arp(fd, &saddr, ARPOP_REQUEST,
							&eth_addr, null_ip,
							&null_addr, ip);
					timeout = PROBE_MIN * 1000;
					timeout += ms_rdelay(PROBE_MAX
							- PROBE_MIN);
				}
				else {
					// Switch to announce state.
					state = ANNOUNCE;
					nclaims = 0;
					VDBG("announce/%d %s@%s\n",
							nclaims, intf, inet_ntoa(ip));
					arp(fd, &saddr, ARPOP_REQUEST,
							&eth_addr, ip,
							&eth_addr, ip);
					timeout = ANNOUNCE_INTERVAL * 1000;
				}
				break;
			case RATE_LIMIT_PROBE:
				// timeouts in the RATE_LIMIT_PROBE state mean no conflicting ARP packets
				// have been received, so we can move immediately to the announce state
				state = ANNOUNCE;
				nclaims = 0;
				VDBG("announce/%d %s@%s\n",
						nclaims, intf, inet_ntoa(ip));
				arp(fd, &saddr, ARPOP_REQUEST,
						&eth_addr, ip,
						&eth_addr, ip);
				timeout = ANNOUNCE_INTERVAL * 1000;
				break;
			case ANNOUNCE:
				// timeouts in the ANNOUNCE state mean no conflicting ARP packets
				// have been received, so we can progress through the states
				if (nclaims < ANNOUNCE_NUM) {
					nclaims++;
					VDBG("announce/%d %s@%s\n",
							nclaims, intf, inet_ntoa(ip));
					arp(fd, &saddr, ARPOP_REQUEST,
							&eth_addr, ip,
							&eth_addr, ip);
					timeout = ANNOUNCE_INTERVAL * 1000;
				}
				else {
					// Switch to monitor state.
					state = MONITOR;
					// link is ok to use earlier
					// FIXME update filters
					script_av[1] = (char*)"config";
					run(script_av, intf, &ip);
					ready = 1;
					conflicts = 0;
					timeout = -1; // Never timeout in the monitor state.

					// NOTE: all other exit paths
					// should deconfig ...
					if (QUIT)
						return EXIT_SUCCESS;
				}
				break;
			case DEFEND:
				// We won!  No ARP replies, so just go back to monitor.
				state = MONITOR;
				timeout = -1;
				conflicts = 0;
				break;
			default:
				// Invalid, should never happen.  Restart the whole protocol.
				state = PROBE;
				pick(&ip);
				timeout = 0;
				nprobes = 0;
				nclaims = 0;
				break;
			} // switch (state)
			break; // case 0 (timeout)
		// packets arriving
		case 1:
			// We need to adjust the timeout in case we didn't receive
			// a conflicting packet.
			if (timeout > 0) {
				struct timeval tv2;

				gettimeofday(&tv2, NULL);
				if (timercmp(&tv1, &tv2, <)) {
					// Current time is greater than the expected timeout time.
					// Should never happen.
					VDBG("missed an expected timeout\n");
					timeout = 0;
				} else {
					VDBG("adjusting timeout\n");
					timersub(&tv1, &tv2, &tv1);
					timeout = 1000 * tv1.tv_sec
							+ tv1.tv_usec / 1000;
				}
			}

			if ((fds[0].revents & POLLIN) == 0) {
				if (fds[0].revents & POLLERR) {
					// FIXME: links routinely go down;
					// this shouldn't necessarily exit.
					bb_error_msg("%s: poll error", intf);
					if (ready) {
						script_av[1] = (char*)"deconfig";
						run(script_av, intf, &ip);
					}
					return EXIT_FAILURE;
				}
				continue;
			}

			// read ARP packet
			if (recv(fd, &p, sizeof(p), 0) < 0) {
				why = "recv";
				goto bad;
			}
			if (p.hdr.ether_type != htons(ETHERTYPE_ARP))
				continue;

#ifdef DEBUG
			{
				struct ether_addr * sha = (struct ether_addr *) p.arp.arp_sha;
				struct ether_addr * tha = (struct ether_addr *) p.arp.arp_tha;
				struct in_addr * spa = (struct in_addr *) p.arp.arp_spa;
				struct in_addr * tpa = (struct in_addr *) p.arp.arp_tpa;
				VDBG("%s recv arp type=%d, op=%d,\n",
					intf, ntohs(p.hdr.ether_type),
					ntohs(p.arp.arp_op));
				VDBG("\tsource=%s %s\n",
					ether_ntoa(sha),
					inet_ntoa(*spa));
				VDBG("\ttarget=%s %s\n",
					ether_ntoa(tha),
					inet_ntoa(*tpa));
			}
#endif
			if (p.arp.arp_op != htons(ARPOP_REQUEST)
					&& p.arp.arp_op != htons(ARPOP_REPLY))
				continue;

			if (memcmp(p.arp.arp_spa, &ip.s_addr, sizeof(struct in_addr)) == 0 &&
				memcmp(&eth_addr, &p.arp.arp_sha, ETH_ALEN) != 0) {
				source_ip_conflict = 1;
			}
			if (memcmp(p.arp.arp_tpa, &ip.s_addr, sizeof(struct in_addr)) == 0 &&
				p.arp.arp_op == htons(ARPOP_REQUEST) &&
				memcmp(&eth_addr, &p.arp.arp_tha, ETH_ALEN) != 0) {
				target_ip_conflict = 1;
			}

			VDBG("state = %d, source ip conflict = %d, target ip conflict = %d\n",
				state, source_ip_conflict, target_ip_conflict);
			switch (state) {
			case PROBE:
			case ANNOUNCE:
				// When probing or announcing, check for source IP conflicts
				// and other hosts doing ARP probes (target IP conflicts).
				if (source_ip_conflict || target_ip_conflict) {
					conflicts++;
					if (conflicts >= MAX_CONFLICTS) {
						VDBG("%s ratelimit\n", intf);
						timeout = RATE_LIMIT_INTERVAL * 1000;
						state = RATE_LIMIT_PROBE;
					}

					// restart the whole protocol
					pick(&ip);
					timeout = 0;
					nprobes = 0;
					nclaims = 0;
				}
				break;
			case MONITOR:
				// If a conflict, we try to defend with a single ARP probe.
				if (source_ip_conflict) {
					VDBG("monitor conflict -- defending\n");
					state = DEFEND;
					timeout = DEFEND_INTERVAL * 1000;
					arp(fd, &saddr,
							ARPOP_REQUEST,
							&eth_addr, ip,
							&eth_addr, ip);
				}
				break;
			case DEFEND:
				// Well, we tried.  Start over (on conflict).
				if (source_ip_conflict) {
					state = PROBE;
					VDBG("defend conflict -- starting over\n");
					ready = 0;
					script_av[1] = (char*)"deconfig";
					run(script_av, intf, &ip);

					// restart the whole protocol
					pick(&ip);
					timeout = 0;
					nprobes = 0;
					nclaims = 0;
				}
				break;
			default:
				// Invalid, should never happen.  Restart the whole protocol.
				VDBG("invalid state -- starting over\n");
				state = PROBE;
				pick(&ip);
				timeout = 0;
				nprobes = 0;
				nclaims = 0;
				break;
			} // switch state

			break; // case 1 (packets arriving)
		default:
			why = "poll";
			goto bad;
		} // switch poll
	}
 bad:
	bb_perror_msg("%s, %s", intf, why);
	return EXIT_FAILURE;
}