aboutsummaryrefslogtreecommitdiffstats
path: root/pcap-pf.c
blob: 3a59083f4d3829ad105921191ef8e3a0ff02c7e4 (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
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
/*
 * Copyright (c) 1990, 1991, 1992, 1993, 1994, 1995, 1996
 *	The Regents of the University of California.  All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that: (1) source code distributions
 * retain the above copyright notice and this paragraph in its entirety, (2)
 * distributions including binary code include the above copyright notice and
 * this paragraph in its entirety in the documentation or other materials
 * provided with the distribution, and (3) all advertising materials mentioning
 * features or use of this software display the following acknowledgement:
 * ``This product includes software developed by the University of California,
 * Lawrence Berkeley Laboratory and its contributors.'' Neither the name of
 * the University nor the names of its contributors may be used to endorse
 * or promote products derived from this software without specific prior
 * written permission.
 * THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR IMPLIED
 * WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF
 * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.
 *
 * packet filter subroutines for tcpdump
 *	Extraction/creation by Jeffrey Mogul, DECWRL
 */

#ifndef lint
static const char rcsid[] _U_ =
    "@(#) $Header: /tcpdump/master/libpcap/pcap-pf.c,v 1.87 2004-03-23 19:18:06 guy Exp $ (LBL)";
#endif

#ifdef HAVE_CONFIG_H
#include "config.h"
#endif

#include <sys/types.h>
#include <sys/time.h>
#include <sys/timeb.h>
#include <sys/socket.h>
#include <sys/file.h>
#include <sys/ioctl.h>
#include <net/pfilt.h>

struct mbuf;
struct rtentry;
#include <net/if.h>

#include <netinet/in.h>
#include <netinet/in_systm.h>
#include <netinet/ip.h>
#include <netinet/if_ether.h>
#include <netinet/ip_var.h>
#include <netinet/udp.h>
#include <netinet/udp_var.h>
#include <netinet/tcp.h>
#include <netinet/tcpip.h>

#include <ctype.h>
#include <errno.h>
#include <netdb.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>

/*
 * Make "pcap.h" not include "pcap-bpf.h"; we are going to include the
 * native OS version, as we need various BPF ioctls from it.
 */
#define PCAP_DONT_INCLUDE_PCAP_BPF_H
#include <net/bpf.h>

#include "pcap-int.h"

#ifdef HAVE_OS_PROTO_H
#include "os-proto.h"
#endif

static int pcap_setfilter_pf(pcap_t *, struct bpf_program *);

/*
 * BUFSPACE is the size in bytes of the packet read buffer.  Most tcpdump
 * applications aren't going to need more than 200 bytes of packet header
 * and the read shouldn't return more packets than packetfilter's internal
 * queue limit (bounded at 256).
 */
#define BUFSPACE (200 * 256)

static int
pcap_read_pf(pcap_t *pc, int cnt, pcap_handler callback, u_char *user)
{
	register u_char *p, *bp;
	struct bpf_insn *fcode;
	register int cc, n, buflen, inc;
	register struct enstamp *sp;
#ifdef LBL_ALIGN
	struct enstamp stamp;
#endif
#ifdef PCAP_FDDIPAD
	register int pad;
#endif

	fcode = pc->md.use_bpf ? NULL : pc->fcode.bf_insns;
 again:
	cc = pc->cc;
	if (cc == 0) {
		cc = read(pc->fd, (char *)pc->buffer + pc->offset, pc->bufsize);
		if (cc < 0) {
			if (errno == EWOULDBLOCK)
				return (0);
			if (errno == EINVAL &&
			    lseek(pc->fd, 0L, SEEK_CUR) + pc->bufsize < 0) {
				/*
				 * Due to a kernel bug, after 2^31 bytes,
				 * the kernel file offset overflows and
				 * read fails with EINVAL. The lseek()
				 * to 0 will fix things.
				 */
				(void)lseek(pc->fd, 0L, SEEK_SET);
				goto again;
			}
			snprintf(pc->errbuf, sizeof(pc->errbuf), "pf read: %s",
				pcap_strerror(errno));
			return (-1);
		}
		bp = pc->buffer + pc->offset;
	} else
		bp = pc->bp;
	/*
	 * Loop through each packet.
	 */
	n = 0;
#ifdef PCAP_FDDIPAD
	if (pc->linktype == DLT_FDDI)
		pad = pcap_fddipad;
	else
		pad = 0;
#endif
	while (cc > 0) {
		/*
		 * Has "pcap_breakloop()" been called?
		 * If so, return immediately - if we haven't read any
		 * packets, clear the flag and return -2 to indicate
		 * that we were told to break out of the loop, otherwise
		 * leave the flag set, so that the *next* call will break
		 * out of the loop without having read any packets, and
		 * return the number of packets we've processed so far.
		 */
		if (pc->break_loop) {
			if (n == 0) {
				pc->break_loop = 0;
				return (-2);
			} else {
				pc->cc = cc;
				pc->bp = bp;
				return (n);
			}
		}
		if (cc < sizeof(*sp)) {
			snprintf(pc->errbuf, sizeof(pc->errbuf),
			    "pf short read (%d)", cc);
			return (-1);
		}
#ifdef LBL_ALIGN
		if ((long)bp & 3) {
			sp = &stamp;
			memcpy((char *)sp, (char *)bp, sizeof(*sp));
		} else
#endif
			sp = (struct enstamp *)bp;
		if (sp->ens_stamplen != sizeof(*sp)) {
			snprintf(pc->errbuf, sizeof(pc->errbuf),
			    "pf short stamplen (%d)",
			    sp->ens_stamplen);
			return (-1);
		}

		p = bp + sp->ens_stamplen;
		buflen = sp->ens_count;
		if (buflen > pc->snapshot)
			buflen = pc->snapshot;

		/* Calculate inc before possible pad update */
		inc = ENALIGN(buflen + sp->ens_stamplen);
		cc -= inc;
		bp += inc;
#ifdef PCAP_FDDIPAD
		p += pad;
		buflen -= pad;
#endif
		pc->md.TotPkts++;
		pc->md.TotDrops += sp->ens_dropped;
		pc->md.TotMissed = sp->ens_ifoverflows;
		if (pc->md.OrigMissed < 0)
			pc->md.OrigMissed = pc->md.TotMissed;

		/*
		 * Short-circuit evaluation: if using BPF filter
		 * in kernel, no need to do it now.
		 */
		if (fcode == NULL ||
		    bpf_filter(fcode, p, sp->ens_count, buflen)) {
			struct pcap_pkthdr h;
			pc->md.TotAccepted++;
			h.ts = sp->ens_tstamp;
#ifdef PCAP_FDDIPAD
			h.len = sp->ens_count - pad;
#else
			h.len = sp->ens_count;
#endif
			h.caplen = buflen;
			(*callback)(user, &h, p);
			if (++n >= cnt && cnt > 0) {
				pc->cc = cc;
				pc->bp = bp;
				return (n);
			}
		}
	}
	pc->cc = 0;
	return (n);
}

static int
pcap_inject_pf(pcap_t *p, const void *buf, size_t size)
{
	int ret;

	ret = write(p->fd, buf, size);
	if (ret == -1) {
		snprintf(p->errbuf, PCAP_ERRBUF_SIZE, "send: %s",
		    pcap_strerror(errno));
		return (-1);
	}
	return (ret);
}                           

static int
pcap_stats_pf(pcap_t *p, struct pcap_stat *ps)
{

	/*
	 * If packet filtering is being done in the kernel:
	 *
	 *	"ps_recv" counts only packets that passed the filter.
	 *	This does not include packets dropped because we
	 *	ran out of buffer space.  (XXX - perhaps it should,
	 *	by adding "ps_drop" to "ps_recv", for compatibility
	 *	with some other platforms.  On the other hand, on
	 *	some platforms "ps_recv" counts only packets that
	 *	passed the filter, and on others it counts packets
	 *	that didn't pass the filter....)
	 *
	 *	"ps_drop" counts packets that passed the kernel filter
	 *	(if any) but were dropped because the input queue was
	 *	full.
	 *
	 *	"ps_ifdrop" counts packets dropped by the network
	 *	inteface (regardless of whether they would have passed
	 *	the input filter, of course).
	 *
	 * If packet filtering is not being done in the kernel:
	 *
	 *	"ps_recv" counts only packets that passed the filter.
	 *
	 *	"ps_drop" counts packets that were dropped because the
	 *	input queue was full, regardless of whether they passed
	 *	the userland filter.
	 *
	 *	"ps_ifdrop" counts packets dropped by the network
	 *	inteface (regardless of whether they would have passed
	 *	the input filter, of course).
	 *
	 * These statistics don't include packets not yet read from
	 * the kernel by libpcap, but they may include packets not
	 * yet read from libpcap by the application.
	 */
	ps->ps_recv = p->md.TotAccepted;
	ps->ps_drop = p->md.TotDrops;
	ps->ps_ifdrop = p->md.TotMissed - p->md.OrigMissed;
	return (0);
}

static void
pcap_close_pf(pcap_t *p)
{
	if (p->buffer != NULL)
		free(p->buffer);
	if (p->fd >= 0)
		close(p->fd);
}

/*
 * We include the OS's <net/bpf.h>, not our "pcap-bpf.h", so we probably
 * don't get DLT_DOCSIS defined.
 */
#ifndef DLT_DOCSIS
#define DLT_DOCSIS	143
#endif

pcap_t *
pcap_open_live(const char *device, int snaplen, int promisc, int to_ms,
    char *ebuf)
{
	pcap_t *p;
	short enmode;
	int backlog = -1;	/* request the most */
	struct enfilter Filter;
	struct endevp devparams;

	p = (pcap_t *)malloc(sizeof(*p));
	if (p == NULL) {
		snprintf(ebuf, PCAP_ERRBUF_SIZE,
		    "pcap_open_live: %s", pcap_strerror(errno));
		return (0);
	}
	memset(p, 0, sizeof(*p));
	/*
	 * Initially try a read/write open (to allow the inject
	 * method to work).  If that fails due to permission
	 * issues, fall back to read-only.  This allows a
	 * non-root user to be granted specific access to pcap
	 * capabilities via file permissions.
	 *
	 * XXX - we should have an API that has a flag that
	 * controls whether to open read-only or read-write,
	 * so that denial of permission to send (or inability
	 * to send, if sending packets isn't supported on
	 * the device in question) can be indicated at open
	 * time.
	 *
	 * XXX - we assume here that "pfopen()" does not, in fact, modify
	 * its argument, even though it takes a "char *" rather than a
	 * "const char *" as its first argument.  That appears to be
	 * the case, at least on Digital UNIX 4.0.
	 */
	p->fd = pfopen(device, O_RDWR);
	if (p->fd == -1 && errno == EACCES)
		p->fd = pfopen(device, O_RDONLY);
	if (p->fd < 0) {
		snprintf(ebuf, PCAP_ERRBUF_SIZE, "pf open: %s: %s\n\
your system may not be properly configured; see the packetfilter(4) man page\n",
			device, pcap_strerror(errno));
		goto bad;
	}
	p->md.OrigMissed = -1;
	enmode = ENTSTAMP|ENBATCH|ENNONEXCL;
	if (promisc)
		enmode |= ENPROMISC;
	if (ioctl(p->fd, EIOCMBIS, (caddr_t)&enmode) < 0) {
		snprintf(ebuf, PCAP_ERRBUF_SIZE, "EIOCMBIS: %s",
		    pcap_strerror(errno));
		goto bad;
	}
#ifdef	ENCOPYALL
	/* Try to set COPYALL mode so that we see packets to ourself */
	enmode = ENCOPYALL;
	(void)ioctl(p->fd, EIOCMBIS, (caddr_t)&enmode);/* OK if this fails */
#endif
	/* set the backlog */
	if (ioctl(p->fd, EIOCSETW, (caddr_t)&backlog) < 0) {
		snprintf(ebuf, PCAP_ERRBUF_SIZE, "EIOCSETW: %s",
		    pcap_strerror(errno));
		goto bad;
	}
	/* discover interface type */
	if (ioctl(p->fd, EIOCDEVP, (caddr_t)&devparams) < 0) {
		snprintf(ebuf, PCAP_ERRBUF_SIZE, "EIOCDEVP: %s",
		    pcap_strerror(errno));
		goto bad;
	}
	/* HACK: to compile prior to Ultrix 4.2 */
#ifndef	ENDT_FDDI
#define	ENDT_FDDI	4
#endif
	switch (devparams.end_dev_type) {

	case ENDT_10MB:
		p->linktype = DLT_EN10MB;
		p->offset = 2;
		/*
		 * This is (presumably) a real Ethernet capture; give it a
		 * link-layer-type list with DLT_EN10MB and DLT_DOCSIS, so
		 * that an application can let you choose it, in case you're
		 * capturing DOCSIS traffic that a Cisco Cable Modem
		 * Termination System is putting out onto an Ethernet (it
		 * doesn't put an Ethernet header onto the wire, it puts raw
		 * DOCSIS frames out on the wire inside the low-level
		 * Ethernet framing).
		 */
		p->dlt_list = (u_int *) malloc(sizeof(u_int) * 2);
		/*
		 * If that fails, just leave the list empty.
		 */
		if (p->dlt_list != NULL) {
			p->dlt_list[0] = DLT_EN10MB;
			p->dlt_list[1] = DLT_DOCSIS;
			p->dlt_count = 2;
		}
		break;

	case ENDT_FDDI:
		p->linktype = DLT_FDDI;
		break;

#ifdef ENDT_SLIP
	case ENDT_SLIP:
		p->linktype = DLT_SLIP;
		break;
#endif

#ifdef ENDT_PPP
	case ENDT_PPP:
		p->linktype = DLT_PPP;
		break;
#endif

#ifdef ENDT_LOOPBACK
	case ENDT_LOOPBACK:
		/*
		 * It appears to use Ethernet framing, at least on
		 * Digital UNIX 4.0.
		 */
		p->linktype = DLT_EN10MB;
		p->offset = 2;
		break;
#endif

#ifdef ENDT_TRN
	case ENDT_TRN:
		p->linktype = DLT_IEEE802;
		break;
#endif

	default:
		/*
		 * XXX - what about ENDT_IEEE802?  The pfilt.h header
		 * file calls this "IEEE 802 networks (non-Ethernet)",
		 * but that doesn't specify a specific link layer type;
		 * it could be 802.4, or 802.5 (except that 802.5 is
		 * ENDT_TRN), or 802.6, or 802.11, or....  That's why
		 * DLT_IEEE802 was hijacked to mean Token Ring in various
		 * BSDs, and why we went along with that hijacking.
		 *
		 * XXX - what about ENDT_HDLC and ENDT_NULL?
		 * Presumably, as ENDT_OTHER is just "Miscellaneous
		 * framing", there's not much we can do, as that
		 * doesn't specify a particular type of header.
		 */
		snprintf(ebuf, PCAP_ERRBUF_SIZE, "unknown data-link type %u",
		    devparams.end_dev_type);
		goto bad;
	}
	/* set truncation */
#ifdef PCAP_FDDIPAD
	if (p->linktype == DLT_FDDI)
		/* packetfilter includes the padding in the snapshot */
		snaplen += pcap_fddipad;
#endif
	if (ioctl(p->fd, EIOCTRUNCATE, (caddr_t)&snaplen) < 0) {
		snprintf(ebuf, PCAP_ERRBUF_SIZE, "EIOCTRUNCATE: %s",
		    pcap_strerror(errno));
		goto bad;
	}
	p->snapshot = snaplen;
	/* accept all packets */
	memset(&Filter, 0, sizeof(Filter));
	Filter.enf_Priority = 37;	/* anything > 2 */
	Filter.enf_FilterLen = 0;	/* means "always true" */
	if (ioctl(p->fd, EIOCSETF, (caddr_t)&Filter) < 0) {
		snprintf(ebuf, PCAP_ERRBUF_SIZE, "EIOCSETF: %s",
		    pcap_strerror(errno));
		goto bad;
	}

	if (to_ms != 0) {
		struct timeval timeout;
		timeout.tv_sec = to_ms / 1000;
		timeout.tv_usec = (to_ms * 1000) % 1000000;
		if (ioctl(p->fd, EIOCSRTIMEOUT, (caddr_t)&timeout) < 0) {
			snprintf(ebuf, PCAP_ERRBUF_SIZE, "EIOCSRTIMEOUT: %s",
				pcap_strerror(errno));
			goto bad;
		}
	}

	p->bufsize = BUFSPACE;
	p->buffer = (u_char*)malloc(p->bufsize + p->offset);
	if (p->buffer == NULL) {
		strlcpy(ebuf, pcap_strerror(errno), PCAP_ERRBUF_SIZE);
		goto bad;
	}

	/*
	 * "select()" and "poll()" work on packetfilter devices.
	 */
	p->selectable_fd = p->fd;

	p->read_op = pcap_read_pf;
	p->inject_op = pcap_inject_pf;
	p->setfilter_op = pcap_setfilter_pf;
	p->set_datalink_op = NULL;	/* can't change data link type */
	p->getnonblock_op = pcap_getnonblock_fd;
	p->setnonblock_op = pcap_setnonblock_fd;
	p->stats_op = pcap_stats_pf;
	p->close_op = pcap_close_pf;

	return (p);
 bad:
	if (p->fd >= 0)
		close(p->fd);
	/*
	 * Get rid of any link-layer type list we allocated.
	 */
	if (p->dlt_list != NULL)
		free(p->dlt_list);
	free(p);
	return (NULL);
}

int
pcap_platform_finddevs(pcap_if_t **alldevsp, char *errbuf)
{
	return (0);
}

static int
pcap_setfilter_pf(pcap_t *p, struct bpf_program *fp)
{
	struct bpf_version bv;

	/*
	 * See if BIOCVERSION works.  If not, we assume the kernel doesn't
	 * support BPF-style filters (it's not documented in the bpf(7)
	 * or packetfiler(7) man pages, but the code used to fail if
	 * BIOCSETF worked but BIOCVERSION didn't, and I've seen it do
	 * kernel filtering in DU 4.0, so presumably BIOCVERSION works
	 * there, at least).
	 */
	if (ioctl(p->fd, BIOCVERSION, (caddr_t)&bv) >= 0) {
		/*
		 * OK, we have the version of the BPF interpreter;
		 * is it the same major version as us, and the same
		 * or better minor version?
		 */
		if (bv.bv_major == BPF_MAJOR_VERSION &&
		    bv.bv_minor >= BPF_MINOR_VERSION) {
			/*
			 * Yes.  Try to install the filter.
			 */
			if (ioctl(p->fd, BIOCSETF, (caddr_t)fp) < 0) {
				snprintf(p->errbuf, sizeof(p->errbuf),
				    "BIOCSETF: %s", pcap_strerror(errno));
				return (-1);
			}

			/*
			 * OK, that succeeded.  We're doing filtering in
			 * the kernel.  (We assume we don't have a
			 * userland filter installed - that'd require
			 * a previous version check to have failed but
			 * this one to succeed.)
			 *
			 * XXX - this message should be supplied to the
			 * application as a warning of some sort,
			 * except that if it's a GUI application, it's
			 * not clear that it should be displayed in
			 * a window to annoy the user.
			 */
			fprintf(stderr, "tcpdump: Using kernel BPF filter\n");
			p->md.use_bpf = 1;
			return (0);
		}

		/*
		 * We can't use the kernel's BPF interpreter; don't give
		 * up, just log a message and be inefficient.
		 *
		 * XXX - this should really be supplied to the application
		 * as a warning of some sort.
		 */
		fprintf(stderr,
	    "tcpdump: Requires BPF language %d.%d or higher; kernel is %d.%d\n",
		    BPF_MAJOR_VERSION, BPF_MINOR_VERSION,
		    bv.bv_major, bv.bv_minor);
	}

	/*
	 * We couldn't do filtering in the kernel; do it in userland.
	 */
	if (install_bpf_program(p, fp) < 0)
		return (-1);

	/*
	 * XXX - this message should be supplied by the application as
	 * a warning of some sort.
	 */
	fprintf(stderr, "tcpdump: Filtering in user process\n");
	p->md.use_bpf = 0;
	return (0);
}