aboutsummaryrefslogtreecommitdiffstats
path: root/vendor/github.com/google/gopacket/pcap/pcap.go
blob: 1ecdf03544c23ea48f1a5e9797af3648701d277c (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
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
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
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
// Copyright 2012 Google, Inc. All rights reserved.
// Copyright 2009-2011 Andreas Krennmair. All rights reserved.
//
// Use of this source code is governed by a BSD-style license
// that can be found in the LICENSE file in the root of the source
// tree.

package pcap

/*
#cgo solaris LDFLAGS: -L /opt/local/lib -lpcap
#cgo linux LDFLAGS: -lpcap
#cgo dragonfly LDFLAGS: -lpcap
#cgo freebsd LDFLAGS: -lpcap
#cgo openbsd LDFLAGS: -lpcap
#cgo netbsd LDFLAGS: -lpcap
#cgo darwin LDFLAGS: -lpcap
#cgo windows CFLAGS: -I C:/WpdPack/Include
#cgo windows,386 LDFLAGS: -L C:/WpdPack/Lib -lwpcap
#cgo windows,amd64 LDFLAGS: -L C:/WpdPack/Lib/x64 -lwpcap
#include <stdlib.h>
#include <pcap.h>

// Some old versions of pcap don't define this constant.
#ifndef PCAP_NETMASK_UNKNOWN
#define PCAP_NETMASK_UNKNOWN 0xffffffff
#endif

// libpcap doesn't actually export its version in a #define-guardable way,
// so we have to use other defined things to differentiate versions.
// We assume at least libpcap v1.1 at the moment.
// See http://upstream-tracker.org/versions/libpcap.html

#ifndef PCAP_ERROR_TSTAMP_PRECISION_NOTSUP  // < v1.5

int pcap_set_immediate_mode(pcap_t *p, int mode) {
  return PCAP_ERROR;
}

#ifndef PCAP_TSTAMP_HOST  // < v1.2

int pcap_set_tstamp_type(pcap_t* p, int t) { return -1; }
int pcap_list_tstamp_types(pcap_t* p, int** t) { return 0; }
void pcap_free_tstamp_types(int *tstamp_types) {}
const char* pcap_tstamp_type_val_to_name(int t) {
	return "pcap timestamp types not supported";
}
int pcap_tstamp_type_name_to_val(const char* t) {
	return PCAP_ERROR;
}

#endif  // < v1.2
#endif  // < v1.5

#ifndef PCAP_ERROR_PROMISC_PERM_DENIED
#define PCAP_ERROR_PROMISC_PERM_DENIED -11
#endif

// WinPcap doesn't export a pcap_statustostr, so use the less-specific
// pcap_strerror.  Note that linking against something like cygwin libpcap
// may result is less-specific error messages.
#ifdef WIN32
#define pcap_statustostr pcap_strerror

// WinPcap also doesn't export pcap_can_set_rfmon and pcap_set_rfmon,
// as those are handled by separate libraries (airpcap).
// https://www.winpcap.org/docs/docs_412/html/group__wpcapfunc.html
// Stub out those functions here, returning values that indicate rfmon
// setting is unavailable/unsuccessful.
int pcap_can_set_rfmon(pcap_t *p) {
	return 0;
}

int pcap_set_rfmon(pcap_t *p, int rfmon) {
	return PCAP_ERROR;
}
#endif

// Windows, Macs, and Linux all use different time types.  Joy.
#ifdef WIN32
#define gopacket_time_secs_t long
#define gopacket_time_usecs_t long
#elif __APPLE__
#define gopacket_time_secs_t __darwin_time_t
#define gopacket_time_usecs_t __darwin_suseconds_t
#elif __GLIBC__
#define gopacket_time_secs_t __time_t
#define gopacket_time_usecs_t __suseconds_t
#else  // Some form of linux/bsd/etc...
#include <sys/param.h>
#ifdef __OpenBSD__
#define gopacket_time_secs_t u_int32_t
#define gopacket_time_usecs_t u_int32_t
#else
#define gopacket_time_secs_t time_t
#define gopacket_time_usecs_t suseconds_t
#endif
#endif
*/
import "C"

import (
	"errors"
	"fmt"
	"io"
	"net"
	"reflect"
	"runtime"
	"strconv"
	"sync"
	"sync/atomic"
	"syscall"
	"time"
	"unsafe"

	"github.com/google/gopacket"
	"github.com/google/gopacket/layers"
)

const errorBufferSize = 256

// MaxBpfInstructions is the maximum number of BPF instructions supported (BPF_MAXINSNS),
// taken from Linux kernel: include/uapi/linux/bpf_common.h
//
// https://github.com/torvalds/linux/blob/master/include/uapi/linux/bpf_common.h
const MaxBpfInstructions = 4096

// 8 bytes per instruction, max 4096 instructions
const bpfInstructionBufferSize = 8 * MaxBpfInstructions

// Handle provides a connection to a pcap handle, allowing users to read packets
// off the wire (Next), inject packets onto the wire (Inject), and
// perform a number of other functions to affect and understand packet output.
//
// Handles are already pcap_activate'd
type Handle struct {
	// cptr is the handle for the actual pcap C object.
	cptr        *C.pcap_t
	timeout     time.Duration
	device      string
	deviceIndex int
	mu          sync.Mutex
	closeMu     sync.Mutex
	// stop is set to a non-zero value by Handle.Close to signal to
	// getNextBufPtrLocked to stop trying to read packets
	stop uint64

	// Since pointers to these objects are passed into a C function, if
	// they're declared locally then the Go compiler thinks they may have
	// escaped into C-land, so it allocates them on the heap.  This causes a
	// huge memory hit, so to handle that we store them here instead.
	pkthdr *C.struct_pcap_pkthdr
	bufptr *C.u_char
}

// Stats contains statistics on how many packets were handled by a pcap handle,
// and what was done with those packets.
type Stats struct {
	PacketsReceived  int
	PacketsDropped   int
	PacketsIfDropped int
}

// Interface describes a single network interface on a machine.
type Interface struct {
	Name        string
	Description string
	Addresses   []InterfaceAddress
	// TODO: add more elements
}

// Datalink describes the datalink
type Datalink struct {
	Name        string
	Description string
}

// InterfaceAddress describes an address associated with an Interface.
// Currently, it's IPv4/6 specific.
type InterfaceAddress struct {
	IP      net.IP
	Netmask net.IPMask // Netmask may be nil if we were unable to retrieve it.
	// TODO: add broadcast + PtP dst ?
}

// BPF is a compiled filter program, useful for offline packet matching.
type BPF struct {
	orig string
	bpf  _Ctype_struct_bpf_program // takes a finalizer, not overriden by outsiders
}

// BPFInstruction is a byte encoded structure holding a BPF instruction
type BPFInstruction struct {
	Code uint16
	Jt   uint8
	Jf   uint8
	K    uint32
}

// BlockForever causes it to block forever waiting for packets, when passed
// into SetTimeout or OpenLive, while still returning incoming packets to userland relatively
// quickly.
const BlockForever = -time.Millisecond * 10

func timeoutMillis(timeout time.Duration) C.int {
	// Flip sign if necessary.  See package docs on timeout for reasoning behind this.
	if timeout < 0 {
		timeout *= -1
	}
	// Round up
	if timeout != 0 && timeout < time.Millisecond {
		timeout = time.Millisecond
	}
	return C.int(timeout / time.Millisecond)
}

// OpenLive opens a device and returns a *Handle.
// It takes as arguments the name of the device ("eth0"), the maximum size to
// read for each packet (snaplen), whether to put the interface in promiscuous
// mode, and a timeout.
//
// See the package documentation for important details regarding 'timeout'.
func OpenLive(device string, snaplen int32, promisc bool, timeout time.Duration) (handle *Handle, _ error) {
	buf := (*C.char)(C.calloc(errorBufferSize, 1))
	defer C.free(unsafe.Pointer(buf))

	var pro C.int
	if promisc {
		pro = 1
	}
	p := &Handle{timeout: timeout, device: device}

	ifc, err := net.InterfaceByName(device)
	if err != nil {
		// The device wasn't found in the OS, but could be "any"
		// Set index to 0
		p.deviceIndex = 0
	} else {
		p.deviceIndex = ifc.Index
	}

	dev := C.CString(device)
	defer C.free(unsafe.Pointer(dev))

	p.cptr = C.pcap_open_live(dev, C.int(snaplen), pro, timeoutMillis(timeout), buf)
	if p.cptr == nil {
		return nil, errors.New(C.GoString(buf))
	}

	if err := p.openLive(); err != nil {
		C.pcap_close(p.cptr)
		return nil, err
	}

	return p, nil
}

// OpenOffline opens a file and returns its contents as a *Handle.
func OpenOffline(file string) (handle *Handle, err error) {
	buf := (*C.char)(C.calloc(errorBufferSize, 1))
	defer C.free(unsafe.Pointer(buf))
	cf := C.CString(file)
	defer C.free(unsafe.Pointer(cf))

	cptr := C.pcap_open_offline(cf, buf)
	if cptr == nil {
		return nil, errors.New(C.GoString(buf))
	}
	return &Handle{cptr: cptr}, nil
}

// NextError is the return code from a call to Next.
type NextError int32

// NextError implements the error interface.
func (n NextError) Error() string {
	switch n {
	case NextErrorOk:
		return "OK"
	case NextErrorTimeoutExpired:
		return "Timeout Expired"
	case NextErrorReadError:
		return "Read Error"
	case NextErrorNoMorePackets:
		return "No More Packets In File"
	case NextErrorNotActivated:
		return "Not Activated"
	}
	return strconv.Itoa(int(n))
}

// NextError values.
const (
	NextErrorOk             NextError = 1
	NextErrorTimeoutExpired NextError = 0
	NextErrorReadError      NextError = -1
	// NextErrorNoMorePackets is returned when reading from a file (OpenOffline) and
	// EOF is reached.  When this happens, Next() returns io.EOF instead of this.
	NextErrorNoMorePackets NextError = -2
	NextErrorNotActivated  NextError = -3
)

// ReadPacketData returns the next packet read from the pcap handle, along with an error
// code associated with that packet.  If the packet is read successfully, the
// returned error is nil.
func (p *Handle) ReadPacketData() (data []byte, ci gopacket.CaptureInfo, err error) {
	p.mu.Lock()
	err = p.getNextBufPtrLocked(&ci)
	if err == nil {
		data = C.GoBytes(unsafe.Pointer(p.bufptr), C.int(ci.CaptureLength))
	}
	p.mu.Unlock()
	if err == NextErrorTimeoutExpired {
		runtime.Gosched()
	}
	return
}

type activateError C.int

const (
	aeNoError      = 0
	aeActivated    = C.PCAP_ERROR_ACTIVATED
	aePromisc      = C.PCAP_WARNING_PROMISC_NOTSUP
	aeNoSuchDevice = C.PCAP_ERROR_NO_SUCH_DEVICE
	aeDenied       = C.PCAP_ERROR_PERM_DENIED
	aeNotUp        = C.PCAP_ERROR_IFACE_NOT_UP
)

func (a activateError) Error() string {
	switch a {
	case aeNoError:
		return "No Error"
	case aeActivated:
		return "Already Activated"
	case aePromisc:
		return "Cannot set as promisc"
	case aeNoSuchDevice:
		return "No Such Device"
	case aeDenied:
		return "Permission Denied"
	case aeNotUp:
		return "Interface Not Up"
	default:
		return fmt.Sprintf("unknown activated error: %d", a)
	}
}

// getNextBufPtrLocked is shared code for ReadPacketData and
// ZeroCopyReadPacketData.
func (p *Handle) getNextBufPtrLocked(ci *gopacket.CaptureInfo) error {
	if p.cptr == nil {
		return io.EOF
	}

	for atomic.LoadUint64(&p.stop) == 0 {
		// try to read a packet if one is immediately available
		result := NextError(C.pcap_next_ex(p.cptr, &p.pkthdr, &p.bufptr))

		switch result {
		case NextErrorOk:
			// got a packet, set capture info and return
			sec := int64(p.pkthdr.ts.tv_sec)
			// convert micros to nanos
			nanos := int64(p.pkthdr.ts.tv_usec) * 1000

			ci.Timestamp = time.Unix(sec, nanos)
			ci.CaptureLength = int(p.pkthdr.caplen)
			ci.Length = int(p.pkthdr.len)
			ci.InterfaceIndex = p.deviceIndex

			return nil
		case NextErrorNoMorePackets:
			// no more packets, return EOF rather than libpcap-specific error
			return io.EOF
		case NextErrorTimeoutExpired:
			// Negative timeout means to loop forever, instead of actually returning
			// the timeout error.
			if p.timeout < 0 {
				// must have had a timeout... wait before trying again
				p.waitForPacket()
				continue
			}
		default:
			return result
		}
	}

	// stop must be set
	return io.EOF
}

// ZeroCopyReadPacketData reads the next packet off the wire, and returns its data.
// The slice returned by ZeroCopyReadPacketData points to bytes owned by the
// the Handle.  Each call to ZeroCopyReadPacketData invalidates any data previously
// returned by ZeroCopyReadPacketData.  Care must be taken not to keep pointers
// to old bytes when using ZeroCopyReadPacketData... if you need to keep data past
// the next time you call ZeroCopyReadPacketData, use ReadPacketData, which copies
// the bytes into a new buffer for you.
//  data1, _, _ := handle.ZeroCopyReadPacketData()
//  // do everything you want with data1 here, copying bytes out of it if you'd like to keep them around.
//  data2, _, _ := handle.ZeroCopyReadPacketData()  // invalidates bytes in data1
func (p *Handle) ZeroCopyReadPacketData() (data []byte, ci gopacket.CaptureInfo, err error) {
	p.mu.Lock()
	err = p.getNextBufPtrLocked(&ci)
	if err == nil {
		slice := (*reflect.SliceHeader)(unsafe.Pointer(&data))
		slice.Data = uintptr(unsafe.Pointer(p.bufptr))
		slice.Len = ci.CaptureLength
		slice.Cap = ci.CaptureLength
	}
	p.mu.Unlock()
	if err == NextErrorTimeoutExpired {
		runtime.Gosched()
	}
	return
}

// Close closes the underlying pcap handle.
func (p *Handle) Close() {
	p.closeMu.Lock()
	defer p.closeMu.Unlock()

	if p.cptr == nil {
		return
	}

	atomic.StoreUint64(&p.stop, 1)

	// wait for packet reader to stop
	p.mu.Lock()
	defer p.mu.Unlock()

	C.pcap_close(p.cptr)
	p.cptr = nil
}

// Error returns the current error associated with a pcap handle (pcap_geterr).
func (p *Handle) Error() error {
	return errors.New(C.GoString(C.pcap_geterr(p.cptr)))
}

// Stats returns statistics on the underlying pcap handle.
func (p *Handle) Stats() (stat *Stats, err error) {
	var cstats _Ctype_struct_pcap_stat
	if -1 == C.pcap_stats(p.cptr, &cstats) {
		return nil, p.Error()
	}
	return &Stats{
		PacketsReceived:  int(cstats.ps_recv),
		PacketsDropped:   int(cstats.ps_drop),
		PacketsIfDropped: int(cstats.ps_ifdrop),
	}, nil
}

// ListDataLinks obtains a list of all possible data link types supported for an interface.
func (p *Handle) ListDataLinks() (datalinks []Datalink, err error) {
	var dltbuf *C.int

	n := int(C.pcap_list_datalinks(p.cptr, &dltbuf))
	if -1 == n {
		return nil, p.Error()
	}

	defer C.pcap_free_datalinks(dltbuf)

	datalinks = make([]Datalink, n)

	dltArray := (*[100]C.int)(unsafe.Pointer(dltbuf))

	for i := 0; i < n; i++ {
		expr := C.pcap_datalink_val_to_name((*dltArray)[i])
		datalinks[i].Name = C.GoString(expr)

		expr = C.pcap_datalink_val_to_description((*dltArray)[i])
		datalinks[i].Description = C.GoString(expr)
	}

	return datalinks, nil
}

// pcap_compile is NOT thread-safe, so protect it.
var pcapCompileMu sync.Mutex

// compileBPFFilter always returns an allocated _Ctype_struct_bpf_program
// It is the callers responsibility to free the memory again, e.g.
//
//    C.pcap_freecode(&bpf)
//
func (p *Handle) compileBPFFilter(expr string) (_Ctype_struct_bpf_program, error) {
	errorBuf := (*C.char)(C.calloc(errorBufferSize, 1))
	defer C.free(unsafe.Pointer(errorBuf))

	var netp uint32
	var maskp uint32

	// Only do the lookup on network interfaces.
	// No device indicates we're handling a pcap file.
	if len(p.device) > 0 {
		dev := C.CString(p.device)
		defer C.free(unsafe.Pointer(dev))
		if -1 == C.pcap_lookupnet(
			dev,
			(*C.bpf_u_int32)(unsafe.Pointer(&netp)),
			(*C.bpf_u_int32)(unsafe.Pointer(&maskp)),
			errorBuf,
		) {
			// We can't lookup the network, but that could be because the interface
			// doesn't have an IPv4.
		}
	}

	var bpf _Ctype_struct_bpf_program
	cexpr := C.CString(expr)
	defer C.free(unsafe.Pointer(cexpr))

	pcapCompileMu.Lock()
	defer pcapCompileMu.Unlock()
	if -1 == C.pcap_compile(p.cptr, &bpf, cexpr, 1, C.bpf_u_int32(maskp)) {
		return bpf, p.Error()
	}

	return bpf, nil
}

// CompileBPFFilter compiles and returns a BPF filter with given a link type and capture length.
func CompileBPFFilter(linkType layers.LinkType, captureLength int, expr string) ([]BPFInstruction, error) {
	cptr := C.pcap_open_dead(C.int(linkType), C.int(captureLength))
	if cptr == nil {
		return nil, errors.New("error opening dead capture")
	}

	h := Handle{cptr: cptr}
	defer h.Close()
	return h.CompileBPFFilter(expr)
}

// CompileBPFFilter compiles and returns a BPF filter for the pcap handle.
func (p *Handle) CompileBPFFilter(expr string) ([]BPFInstruction, error) {
	bpf, err := p.compileBPFFilter(expr)
	defer C.pcap_freecode(&bpf)
	if err != nil {
		return nil, err
	}

	bpfInsn := (*[bpfInstructionBufferSize]_Ctype_struct_bpf_insn)(unsafe.Pointer(bpf.bf_insns))[0:bpf.bf_len:bpf.bf_len]
	bpfInstruction := make([]BPFInstruction, len(bpfInsn), len(bpfInsn))

	for i, v := range bpfInsn {
		bpfInstruction[i].Code = uint16(v.code)
		bpfInstruction[i].Jt = uint8(v.jt)
		bpfInstruction[i].Jf = uint8(v.jf)
		bpfInstruction[i].K = uint32(v.k)
	}

	return bpfInstruction, nil
}

// SetBPFFilter compiles and sets a BPF filter for the pcap handle.
func (p *Handle) SetBPFFilter(expr string) (err error) {
	bpf, err := p.compileBPFFilter(expr)
	defer C.pcap_freecode(&bpf)
	if err != nil {
		return err
	}

	if -1 == C.pcap_setfilter(p.cptr, &bpf) {
		return p.Error()
	}

	return nil
}

// SetBPFInstructionFilter may be used to apply a filter in BPF asm byte code format.
//
// Simplest way to generate BPF asm byte code is with tcpdump:
//     tcpdump -dd 'udp'
//
// The output may be used directly to add a filter, e.g.:
//     bpfInstructions := []pcap.BpfInstruction{
//			{0x28, 0, 0, 0x0000000c},
//			{0x15, 0, 9, 0x00000800},
//			{0x30, 0, 0, 0x00000017},
//			{0x15, 0, 7, 0x00000006},
//			{0x28, 0, 0, 0x00000014},
//			{0x45, 5, 0, 0x00001fff},
//			{0xb1, 0, 0, 0x0000000e},
//			{0x50, 0, 0, 0x0000001b},
//			{0x54, 0, 0, 0x00000012},
//			{0x15, 0, 1, 0x00000012},
//			{0x6, 0, 0, 0x0000ffff},
//			{0x6, 0, 0, 0x00000000},
//		}
//
// An other posibility is to write the bpf code in bpf asm.
// Documentation: https://www.kernel.org/doc/Documentation/networking/filter.txt
//
// To compile the code use bpf_asm from
// https://github.com/torvalds/linux/tree/master/tools/net
//
// The following command may be used to convert bpf_asm output to c/go struct, usable for SetBPFFilterByte:
// bpf_asm -c tcp.bpf
func (p *Handle) SetBPFInstructionFilter(bpfInstructions []BPFInstruction) (err error) {
	bpf, err := bpfInstructionFilter(bpfInstructions)
	if err != nil {
		return err
	}

	if -1 == C.pcap_setfilter(p.cptr, &bpf) {
		C.pcap_freecode(&bpf)
		return p.Error()
	}

	C.pcap_freecode(&bpf)

	return nil
}
func bpfInstructionFilter(bpfInstructions []BPFInstruction) (bpf _Ctype_struct_bpf_program, err error) {
	if len(bpfInstructions) < 1 {
		return bpf, errors.New("bpfInstructions must not be empty")
	}

	if len(bpfInstructions) > MaxBpfInstructions {
		return bpf, fmt.Errorf("bpfInstructions must not be larger than %d", MaxBpfInstructions)
	}

	bpf.bf_len = C.u_int(len(bpfInstructions))
	cbpfInsns := C.calloc(C.size_t(len(bpfInstructions)), C.size_t(unsafe.Sizeof(bpfInstructions[0])))

	copy((*[bpfInstructionBufferSize]BPFInstruction)(cbpfInsns)[0:len(bpfInstructions)], bpfInstructions)
	bpf.bf_insns = (*_Ctype_struct_bpf_insn)(cbpfInsns)

	return
}

// NewBPF compiles the given string into a new filter program.
//
// BPF filters need to be created from activated handles, because they need to
// know the underlying link type to correctly compile their offsets.
func (p *Handle) NewBPF(expr string) (*BPF, error) {
	bpf := &BPF{orig: expr}
	cexpr := C.CString(expr)
	defer C.free(unsafe.Pointer(cexpr))

	pcapCompileMu.Lock()
	defer pcapCompileMu.Unlock()
	if C.pcap_compile(p.cptr, &bpf.bpf, cexpr /* optimize */, 1, C.PCAP_NETMASK_UNKNOWN) != 0 {
		return nil, p.Error()
	}

	runtime.SetFinalizer(bpf, destroyBPF)
	return bpf, nil
}

// NewBPFInstructionFilter sets the given BPFInstructions as new filter program.
//
// More details see func SetBPFInstructionFilter
//
// BPF filters need to be created from activated handles, because they need to
// know the underlying link type to correctly compile their offsets.
func (p *Handle) NewBPFInstructionFilter(bpfInstructions []BPFInstruction) (*BPF, error) {
	var err error
	bpf := &BPF{orig: "BPF Instruction Filter"}

	bpf.bpf, err = bpfInstructionFilter(bpfInstructions)
	if err != nil {
		return nil, err
	}

	runtime.SetFinalizer(bpf, destroyBPF)
	return bpf, nil
}
func destroyBPF(bpf *BPF) {
	C.pcap_freecode(&bpf.bpf)
}

// String returns the original string this BPF filter was compiled from.
func (b *BPF) String() string {
	return b.orig
}

// Matches returns true if the given packet data matches this filter.
func (b *BPF) Matches(ci gopacket.CaptureInfo, data []byte) bool {
	var hdr C.struct_pcap_pkthdr
	hdr.ts.tv_sec = C.gopacket_time_secs_t(ci.Timestamp.Unix())
	hdr.ts.tv_usec = C.gopacket_time_usecs_t(ci.Timestamp.Nanosecond() / 1000)
	hdr.caplen = C.bpf_u_int32(len(data)) // Trust actual length over ci.Length.
	hdr.len = C.bpf_u_int32(ci.Length)
	dataptr := (*C.u_char)(unsafe.Pointer(&data[0]))
	return C.pcap_offline_filter(&b.bpf, &hdr, dataptr) != 0
}

// Version returns pcap_lib_version.
func Version() string {
	return C.GoString(C.pcap_lib_version())
}

// LinkType returns pcap_datalink, as a layers.LinkType.
func (p *Handle) LinkType() layers.LinkType {
	return layers.LinkType(C.pcap_datalink(p.cptr))
}

// SetLinkType calls pcap_set_datalink on the pcap handle.
func (p *Handle) SetLinkType(dlt layers.LinkType) error {
	if -1 == C.pcap_set_datalink(p.cptr, C.int(dlt)) {
		return p.Error()
	}
	return nil
}

// FindAllDevs attempts to enumerate all interfaces on the current machine.
func FindAllDevs() (ifs []Interface, err error) {
	var buf *C.char
	buf = (*C.char)(C.calloc(errorBufferSize, 1))
	defer C.free(unsafe.Pointer(buf))
	var alldevsp *C.pcap_if_t

	if -1 == C.pcap_findalldevs((**C.pcap_if_t)(&alldevsp), buf) {
		return nil, errors.New(C.GoString(buf))
	}
	defer C.pcap_freealldevs((*C.pcap_if_t)(alldevsp))
	dev := alldevsp
	var i uint32
	for i = 0; dev != nil; dev = (*C.pcap_if_t)(dev.next) {
		i++
	}
	ifs = make([]Interface, i)
	dev = alldevsp
	for j := uint32(0); dev != nil; dev = (*C.pcap_if_t)(dev.next) {
		var iface Interface
		iface.Name = C.GoString(dev.name)
		iface.Description = C.GoString(dev.description)
		iface.Addresses = findalladdresses(dev.addresses)
		// TODO: add more elements
		ifs[j] = iface
		j++
	}
	return
}

func findalladdresses(addresses *_Ctype_struct_pcap_addr) (retval []InterfaceAddress) {
	// TODO - make it support more than IPv4 and IPv6?
	retval = make([]InterfaceAddress, 0, 1)
	for curaddr := addresses; curaddr != nil; curaddr = (*_Ctype_struct_pcap_addr)(curaddr.next) {
		// Strangely, it appears that in some cases, we get a pcap address back from
		// pcap_findalldevs with a nil .addr.  It appears that we can skip over
		// these.
		if curaddr.addr == nil {
			continue
		}
		var a InterfaceAddress
		var err error
		if a.IP, err = sockaddrToIP((*syscall.RawSockaddr)(unsafe.Pointer(curaddr.addr))); err != nil {
			continue
		}
		// To be safe, we'll also check for netmask.
		if curaddr.netmask == nil {
			continue
		}
		if a.Netmask, err = sockaddrToIP((*syscall.RawSockaddr)(unsafe.Pointer(curaddr.netmask))); err != nil {
			// If we got an IP address but we can't get a netmask, just return the IP
			// address.
			a.Netmask = nil
		}
		retval = append(retval, a)
	}
	return
}

func sockaddrToIP(rsa *syscall.RawSockaddr) (IP []byte, err error) {
	switch rsa.Family {
	case syscall.AF_INET:
		pp := (*syscall.RawSockaddrInet4)(unsafe.Pointer(rsa))
		IP = make([]byte, 4)
		for i := 0; i < len(IP); i++ {
			IP[i] = pp.Addr[i]
		}
		return
	case syscall.AF_INET6:
		pp := (*syscall.RawSockaddrInet6)(unsafe.Pointer(rsa))
		IP = make([]byte, 16)
		for i := 0; i < len(IP); i++ {
			IP[i] = pp.Addr[i]
		}
		return
	}
	err = errors.New("Unsupported address type")
	return
}

// WritePacketData calls pcap_sendpacket, injecting the given data into the pcap handle.
func (p *Handle) WritePacketData(data []byte) (err error) {
	if -1 == C.pcap_sendpacket(p.cptr, (*C.u_char)(&data[0]), (C.int)(len(data))) {
		err = p.Error()
	}
	return
}

// Direction is used by Handle.SetDirection.
type Direction uint8

// Direction values for Handle.SetDirection.
const (
	DirectionIn    Direction = C.PCAP_D_IN
	DirectionOut   Direction = C.PCAP_D_OUT
	DirectionInOut Direction = C.PCAP_D_INOUT
)

// SetDirection sets the direction for which packets will be captured.
func (p *Handle) SetDirection(direction Direction) error {
	if direction != DirectionIn && direction != DirectionOut && direction != DirectionInOut {
		return fmt.Errorf("Invalid direction: %v", direction)
	}
	if status := C.pcap_setdirection(p.cptr, (C.pcap_direction_t)(direction)); status < 0 {
		return statusError(status)
	}
	return nil
}

// TimestampSource tells PCAP which type of timestamp to use for packets.
type TimestampSource C.int

// String returns the timestamp type as a human-readable string.
func (t TimestampSource) String() string {
	return C.GoString(C.pcap_tstamp_type_val_to_name(C.int(t)))
}

// TimestampSourceFromString translates a string into a timestamp type, case
// insensitive.
func TimestampSourceFromString(s string) (TimestampSource, error) {
	cs := C.CString(s)
	defer C.free(unsafe.Pointer(cs))
	t := C.pcap_tstamp_type_name_to_val(cs)
	if t < 0 {
		return 0, statusError(t)
	}
	return TimestampSource(t), nil
}

func statusError(status C.int) error {
	return errors.New(C.GoString(C.pcap_statustostr(status)))
}

// InactiveHandle allows you to call pre-pcap_activate functions on your pcap
// handle to set it up just the way you'd like.
type InactiveHandle struct {
	// cptr is the handle for the actual pcap C object.
	cptr        *C.pcap_t
	device      string
	deviceIndex int
	timeout     time.Duration
}

// Activate activates the handle.  The current InactiveHandle becomes invalid
// and all future function calls on it will fail.
func (p *InactiveHandle) Activate() (*Handle, error) {
	err := activateError(C.pcap_activate(p.cptr))
	if err != aeNoError {
		return nil, err
	}
	h := &Handle{
		cptr:        p.cptr,
		timeout:     p.timeout,
		device:      p.device,
		deviceIndex: p.deviceIndex,
	}
	p.cptr = nil
	return h, nil
}

// CleanUp cleans up any stuff left over from a successful or failed building
// of a handle.
func (p *InactiveHandle) CleanUp() {
	if p.cptr != nil {
		C.pcap_close(p.cptr)
	}
}

// NewInactiveHandle creates a new InactiveHandle, which wraps an un-activated PCAP handle.
// Callers of NewInactiveHandle should immediately defer 'CleanUp', as in:
//   inactive := NewInactiveHandle("eth0")
//   defer inactive.CleanUp()
func NewInactiveHandle(device string) (*InactiveHandle, error) {
	buf := (*C.char)(C.calloc(errorBufferSize, 1))
	defer C.free(unsafe.Pointer(buf))
	dev := C.CString(device)
	defer C.free(unsafe.Pointer(dev))

	// Try to get the interface index, but iy could be something like "any"
	// in which case use 0, which doesn't exist in nature
	deviceIndex := 0
	ifc, err := net.InterfaceByName(device)
	if err == nil {
		deviceIndex = ifc.Index
	}

	// This copies a bunch of the pcap_open_live implementation from pcap.c:
	cptr := C.pcap_create(dev, buf)
	if cptr == nil {
		return nil, errors.New(C.GoString(buf))
	}
	return &InactiveHandle{cptr: cptr, device: device, deviceIndex: deviceIndex}, nil
}

// SetSnapLen sets the snap length (max bytes per packet to capture).
func (p *InactiveHandle) SetSnapLen(snaplen int) error {
	if status := C.pcap_set_snaplen(p.cptr, C.int(snaplen)); status < 0 {
		return statusError(status)
	}
	return nil
}

// SetPromisc sets the handle to either be promiscuous (capture packets
// unrelated to this host) or not.
func (p *InactiveHandle) SetPromisc(promisc bool) error {
	var pro C.int
	if promisc {
		pro = 1
	}
	if status := C.pcap_set_promisc(p.cptr, pro); status < 0 {
		return statusError(status)
	}
	return nil
}

// SetTimeout sets the read timeout for the handle.
//
// See the package documentation for important details regarding 'timeout'.
func (p *InactiveHandle) SetTimeout(timeout time.Duration) error {
	if status := C.pcap_set_timeout(p.cptr, timeoutMillis(timeout)); status < 0 {
		return statusError(status)
	}
	p.timeout = timeout
	return nil
}

// SupportedTimestamps returns a list of supported timstamp types for this
// handle.
func (p *InactiveHandle) SupportedTimestamps() (out []TimestampSource) {
	var types *C.int
	n := int(C.pcap_list_tstamp_types(p.cptr, &types))
	defer C.pcap_free_tstamp_types(types)
	typesArray := (*[100]C.int)(unsafe.Pointer(types))
	for i := 0; i < n; i++ {
		out = append(out, TimestampSource((*typesArray)[i]))
	}
	return
}

// SetTimestampSource sets the type of timestamp generator PCAP uses when
// attaching timestamps to packets.
func (p *InactiveHandle) SetTimestampSource(t TimestampSource) error {
	if status := C.pcap_set_tstamp_type(p.cptr, C.int(t)); status < 0 {
		return statusError(status)
	}
	return nil
}

// CannotSetRFMon is returned by SetRFMon if the handle does not allow
// setting RFMon because pcap_can_set_rfmon returns 0.
var CannotSetRFMon = errors.New("Cannot set rfmon for this handle")

// SetRFMon turns on radio monitoring mode, similar to promiscuous mode but for
// wireless networks.  If this mode is enabled, the interface will not need to
// associate with an access point before it can receive traffic.
func (p *InactiveHandle) SetRFMon(monitor bool) error {
	var mon C.int
	if monitor {
		mon = 1
	}
	switch canset := C.pcap_can_set_rfmon(p.cptr); canset {
	case 0:
		return CannotSetRFMon
	case 1:
		// success
	default:
		return statusError(canset)
	}
	if status := C.pcap_set_rfmon(p.cptr, mon); status != 0 {
		return statusError(status)
	}
	return nil
}

// SetBufferSize sets the buffer size (in bytes) of the handle.
func (p *InactiveHandle) SetBufferSize(bufferSize int) error {
	if status := C.pcap_set_buffer_size(p.cptr, C.int(bufferSize)); status < 0 {
		return statusError(status)
	}
	return nil
}

// SetImmediateMode sets (or unsets) the immediate mode of the
// handle. In immediate mode, packets are delivered to the application
// as soon as they arrive.  In other words, this overrides SetTimeout.
func (p *InactiveHandle) SetImmediateMode(mode bool) error {
	var md C.int
	if mode {
		md = 1
	}
	if status := C.pcap_set_immediate_mode(p.cptr, md); status < 0 {
		return statusError(status)
	}
	return nil
}