summaryrefslogtreecommitdiffstats
path: root/traceroute.cc
blob: 6a302965af927b887472dad382f75ac7db72d4e6 (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
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
/***************************************************************************
 * traceroute.cc -- Parallel multi-protocol traceroute feature             *
 *                                                                         *
 ***********************IMPORTANT NMAP LICENSE TERMS************************
 *
 * The Nmap Security Scanner is (C) 1996-2023 Nmap Software LLC ("The Nmap
 * Project"). Nmap is also a registered trademark of the Nmap Project.
 *
 * This program is distributed under the terms of the Nmap Public Source
 * License (NPSL). The exact license text applying to a particular Nmap
 * release or source code control revision is contained in the LICENSE
 * file distributed with that version of Nmap or source code control
 * revision. More Nmap copyright/legal information is available from
 * https://nmap.org/book/man-legal.html, and further information on the
 * NPSL license itself can be found at https://nmap.org/npsl/ . This
 * header summarizes some key points from the Nmap license, but is no
 * substitute for the actual license text.
 *
 * Nmap is generally free for end users to download and use themselves,
 * including commercial use. It is available from https://nmap.org.
 *
 * The Nmap license generally prohibits companies from using and
 * redistributing Nmap in commercial products, but we sell a special Nmap
 * OEM Edition with a more permissive license and special features for
 * this purpose. See https://nmap.org/oem/
 *
 * If you have received a written Nmap license agreement or contract
 * stating terms other than these (such as an Nmap OEM license), you may
 * choose to use and redistribute Nmap under those terms instead.
 *
 * The official Nmap Windows builds include the Npcap software
 * (https://npcap.com) for packet capture and transmission. It is under
 * separate license terms which forbid redistribution without special
 * permission. So the official Nmap Windows builds may not be redistributed
 * without special permission (such as an Nmap OEM license).
 *
 * Source is provided to this software because we believe users have a
 * right to know exactly what a program is going to do before they run it.
 * This also allows you to audit the software for security holes.
 *
 * Source code also allows you to port Nmap to new platforms, fix bugs, and add
 * new features. You are highly encouraged to submit your changes as a Github PR
 * or by email to the dev@nmap.org mailing list for possible incorporation into
 * the main distribution. Unless you specify otherwise, it is understood that
 * you are offering us very broad rights to use your submissions as described in
 * the Nmap Public Source License Contributor Agreement. This is important
 * because we fund the project by selling licenses with various terms, and also
 * because the inability to relicense code has caused devastating problems for
 * other Free Software projects (such as KDE and NASM).
 *
 * The free version of Nmap is distributed in the hope that it will be
 * useful, but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. Warranties,
 * indemnification and commercial support are all available through the
 * Npcap OEM program--see https://nmap.org/oem/
 *
 ***************************************************************************/

/*
Traceroute for Nmap. This traceroute is faster than a traditional traceroute
because it sends several probes in parallel and detects shared traces.

The algorithm works by sending probes with varying TTL values and waiting for
TTL_EXCEEDED messages. As intermediate hops are discovered, they are entered
into a global hop cache that is shared between targets and across host groups.
When a hop is discovered and is found to be already in the cache, the trace for
that target is linked into the cached trace and there is no need to try lower
TTLs. The process results in the building of a tree of Hop structures.

The order in which probes are sent does not matter to the accuracy of the
algorithm but it does matter to the speed. The sooner a shared trace can be
detected, and the higher the TTL at which it is detected, the fewer probes need
to be sent. The ideal situation is to start sending probes with a TTL equal to
the true distance and count downward from there. In that case it may only be
necessary to send two probes per target: one at the distance of the target to
get a response, and one at distance - 1 to get a cache hit. When the distance
isn't known in advance, the algorithm arbitrarily starts at a TTL of 10 and
counts downward, then counts upward from 11 until it reaches the target. So a
typical trace may look like

TTL 10 -> TTL_EXCEEDED
TTL  9 -> TTL_EXCEEDED
TTL  8 -> TTL_EXCEEDED
TTL  7 -> cache hit
TTL 11 -> TTL_EXCEEDED
TTL 12 -> TTL_EXCEEDED
TTL 13 -> SYN/ACK, or whatever is the target's response to the probe

The output for this host would then say "Hops 1-7 are the same as for ...".

The detection of shared traces rests on the assumption that all paths going
through a router at a certain TTL will be identical up to and including the
router. This assumption is not always true. Even if two targets are each one hop
past router X at TTL 10, packets may follow different paths to each host (and
those paths may even change over time). This traceroute algorithm will be fooled
by such a situation, and will report that the paths are identical up to
router X. The only way to be sure is to do a complete trace for each target
individually.
*/

#include "nmap_dns.h"
#include "nmap_error.h"
#include "nmap_tty.h"
#include "osscan2.h"
#include "payload.h"
#include "timing.h"
#include "NmapOps.h"
#include "Target.h"
#include "tcpip.h"

#include "struct_ip.h"

#ifndef IPPROTO_SCTP
#include "libnetutil/netutil.h"
#endif

#include <dnet.h>

#include <algorithm>
#include <list>
#include <map>
#include <set>
#include <vector>

extern NmapOps o;

/* The highest TTL we go up to if the target itself doesn't respond. */
#define MAX_TTL 30
#define MAX_OUTSTANDING_PROBES 10
#define MAX_RESENDS 2
/* In milliseconds. */
#define PROBE_TIMEOUT 1000
/* If the hop cache (including timed-out hops) is bigger than this after a
   round, the hop is cleared and rebuilt from scratch. */
#define MAX_HOP_CACHE_SIZE 1000

struct Hop;
class HostState;
class Probe;

/* An object of this class is a (TTL, address) pair that uniquely identifies a
   hop. Hops in the hop_cache are indexed by this type. */
struct HopIdent {
  u8 ttl;
  struct sockaddr_storage addr;

  HopIdent(u8 ttl, const struct sockaddr_storage &addr) {
    this->addr = addr;
    this->ttl = ttl;
  }

  bool operator<(const struct HopIdent &other) const {
    if (ttl < other.ttl)
      return true;
    else if (ttl > other.ttl)
      return false;
    else
      return sockaddr_storage_cmp(&addr, &other.addr) < 0;
  }
};

/* A global random token used to distinguish this traceroute's probes from
   those of other traceroutes possibly running on the same machine. */
static u16 global_id;
/* A global cache of known hops, indexed by TTL and address. */
static std::map<struct HopIdent, Hop *> hop_cache;
/* A list of timedout hops, which are not kept in hop_cache, so we can delete
   all hops on occasion. */
/* This would be stack-allocated except for a weird bug on AIX that causes
 * infinite loops when trying to traverse the list. For some reason,
 * dynamically allocating it fixes the bug. */
static std::list<Hop *> *timedout_hops = NULL;
/* The TTL at which we start sending probes if we don't have a distance
   estimate. This is updated after each host group on the assumption that hosts
   across groups will not differ much in distance. Having this closer to the
   true distance makes the trace faster but is not needed for accuracy. */
static u8 initial_ttl = 10;

static struct timeval get_now(struct timeval *now = NULL);
static const char *ss_to_string(const struct sockaddr_storage *ss);

struct Hop {
  Hop *parent;
  struct sockaddr_storage tag;
  /* When addr.ss_family == 0, this hop represents a timeout. */
  struct sockaddr_storage addr;
  u8 ttl;
  float rtt; /* In milliseconds. */
  std::string hostname;

  Hop() {
    this->parent = NULL;
    this->addr.ss_family = 0;
    this->ttl = 0;
    this->rtt = -1.0;
    this->tag.ss_family = 0;
  }

  Hop(u8 ttl, const struct sockaddr_storage &addr, float rtt) {
    this->parent = NULL;
    this->addr = addr;
    this->ttl = ttl;
    this->rtt = rtt;
    this->tag.ss_family = 0;
  }
};

class HostState {
public:
  enum counting_state { COUNTING_DOWN, COUNTING_UP };

  Target *target;
  /* A bitmap of TTLs that have been sent, to avoid duplicates when we switch
     around the order counting up or down. */
  std::vector<bool> sent_ttls;
  u8 current_ttl;
  enum counting_state state;
  /* If nonzero, the known hop distance to the target. */
  int reached_target;
  struct probespec pspec;
  std::list<Probe *> unanswered_probes;
  std::list<Probe *> active_probes;
  std::list<Probe *> pending_resends;
  Hop *hops;

  HostState(Target *target);
  ~HostState();
  bool has_more_probes() const;
  bool is_finished() const;
  bool send_next_probe(int rawsd, eth_t *ethsd);
  void next_ttl();
  void count_up();
  int cancel_probe(std::list<Probe *>::iterator it);
  int cancel_probes_below(u8 ttl);
  int cancel_probes_above(u8 ttl);
  Hop *insert_hop(u8 ttl, const struct sockaddr_storage *addr, float rtt);
  void link_to(Hop *hop);
  double completion_fraction() const;

private:
  void child_parent_ttl(u8 ttl, Hop **child, Hop **parent) const;
  static u8 distance_guess(const Target *target);
  static struct probespec get_probe(const Target *target);
};

class Probe {
private:
  /* This is incremented with each instantiated probe. */
  static u16 token_counter;

  unsigned int num_resends;

public:
  HostState *host;
  struct probespec pspec;
  u8 ttl;
  /* The token is used to match up probe replies. */
  u16 token;
  struct timeval sent_time;

  Probe(HostState *host, struct probespec pspec, u8 ttl);
  virtual ~Probe();
  void send(int rawsd, eth_t *ethsd, struct timeval *now = NULL);
  void resend(int rawsd, eth_t *ethsd, struct timeval *now = NULL);
  bool is_timedout(struct timeval *now = NULL) const;
  bool may_resend() const;
  virtual unsigned char *build_packet(const struct sockaddr_storage *source,
    u32 *len) const = 0;

  static Probe *make(HostState *host, struct probespec pspec, u8 ttl);
};
u16 Probe::token_counter = 0x0000;

class TracerouteState {
public:
  std::list<HostState *> active_hosts;
  /* The next send time for enforcing scan delay. */
  struct timeval next_send_time;

  TracerouteState(std::vector<Target *> &targets);
  ~TracerouteState();

  void send_new_probes();
  void read_replies(long timeout);
  void cull_timeouts();
  void remove_finished_hosts();
  void resolve_hops();
  void transfer_hops();

  double completion_fraction() const;

private:
  eth_t *ethsd;
  int rawsd;
  pcap_t *pd;
  int num_active_probes;

  std::vector<HostState *> hosts;
  std::list<HostState *>::iterator next_sending_host;

  void next_active_host();
  Probe *lookup_probe(const struct sockaddr_storage *target_addr, u16 token);
  void set_host_hop(HostState *host, u8 ttl,
    const struct sockaddr_storage *from_addr, float rtt);
  void set_host_hop_timedout(HostState *host, u8 ttl);
};

static Hop *merge_hops(const struct sockaddr_storage *tag, Hop *a, Hop *b);
static Hop *hop_cache_lookup(u8 ttl, const struct sockaddr_storage *addr);
static void hop_cache_insert(Hop *hop);
static unsigned int hop_cache_size();

HostState::HostState(Target *target) : sent_ttls(MAX_TTL + 1, false) {
  this->target = target;
  current_ttl = MIN(MAX(1, HostState::distance_guess(target)), MAX_TTL);
  state = HostState::COUNTING_DOWN;
  reached_target = 0;
  pspec = HostState::get_probe(target);
  hops = NULL;
}

HostState::~HostState() {
  /* active_probes and pending_resends are subsets of unanswered_probes, so we
     delete the allocated probes in unanswered_probes only. */
  while (!unanswered_probes.empty()) {
    delete *unanswered_probes.begin();
    unanswered_probes.pop_front();
  }
  while (!active_probes.empty())
    active_probes.pop_front();
  while (!pending_resends.empty())
    pending_resends.pop_front();
}

bool HostState::has_more_probes() const {
  /* We are done if we are counting up and
     1. we've reached and exceeded the target, or
     2. we've exceeded MAX_TTL. */
  return !(state == HostState::COUNTING_UP
           && ((reached_target > 0 && current_ttl >= reached_target)
               || current_ttl > MAX_TTL));
}

bool HostState::is_finished() const {
  return !this->has_more_probes()
    && active_probes.empty() && pending_resends.empty();
}

bool HostState::send_next_probe(int rawsd, eth_t *ethsd) {
  Probe *probe;

  /* Do a resend if possible. */
  if (!pending_resends.empty()) {
    probe = pending_resends.front();
    pending_resends.pop_front();
    active_probes.push_back(probe);
    probe->resend(rawsd, ethsd);
    return true;
  }

  this->next_ttl();

  if (!this->has_more_probes())
    return false;

  probe = Probe::make(this, pspec, current_ttl);
  unanswered_probes.push_back(probe);
  active_probes.push_back(probe);
  probe->send(rawsd, ethsd);
  sent_ttls[current_ttl] = true;

  return true;
}

/* Find the next TTL we should send to. */
void HostState::next_ttl() {
  assert(current_ttl > 0);
  if (state == HostState::COUNTING_DOWN) {
    while (current_ttl > 1 && sent_ttls[current_ttl])
      current_ttl--;
    if (current_ttl == 1)
      state = HostState::COUNTING_UP;
  }
  /* Note no "else". */
  if (state == HostState::COUNTING_UP) {
    while (current_ttl <= MAX_TTL && sent_ttls[current_ttl])
      current_ttl++;
  }
}

int HostState::cancel_probe(std::list<Probe *>::iterator it) {
  int count;

  count = active_probes.size();
  active_probes.remove(*it);
  count -= active_probes.size();
  pending_resends.remove(*it);
  delete *it;
  unanswered_probes.erase(it);

  return count;
}

int HostState::cancel_probes_below(u8 ttl) {
  std::list<Probe *>::iterator it, next;
  int count;

  count = 0;
  for (it = unanswered_probes.begin(); it != unanswered_probes.end(); it = next) {
    next = it;
    next++;
    if ((*it)->ttl < ttl)
      count += this->cancel_probe(it);
  }

  return count;
}

int HostState::cancel_probes_above(u8 ttl) {
  std::list<Probe *>::iterator it, next;
  int count;

  count = 0;
  for (it = unanswered_probes.begin(); it != unanswered_probes.end(); it = next) {
    next = it;
    next++;
    if ((*it)->ttl > ttl)
      count += this->cancel_probe(it);
  }

  return count;
}

Hop *HostState::insert_hop(u8 ttl, const struct sockaddr_storage *addr,
  float rtt) {
  Hop *hop, *prev, *p;

  this->child_parent_ttl(ttl, &prev, &p);
  if (p != NULL && p->ttl == ttl) {
    hop = p;
    /* Collision with the same TTL and a different address. */
    if (hop->addr.ss_family == 0) {
      /* Hit a timed-out hop. Fill in the missing address and RTT. */
      hop->addr = *addr;
      hop->rtt = rtt;
    } else {
      if (o.debugging) {
        log_write(LOG_STDOUT, "Found existing %s", ss_to_string(&hop->addr));
        log_write(LOG_STDOUT, " while inserting %s at TTL %d for %s\n",
          ss_to_string(addr), ttl, target->targetipstr());
      }
    }
  } else {
    hop = new Hop(ttl, *addr, rtt);
    hop->parent = p;
    if (prev == NULL) {
      size_t sslen;
      this->hops = hop;
      sslen = sizeof(hop->tag);
      target->TargetSockAddr(&hop->tag, &sslen);
    } else {
      prev->parent = hop;
      hop->tag = prev->tag;
    }
    hop_cache_insert(hop);
  }

  return hop;
}

void HostState::link_to(Hop *hop) {
  Hop *prev, *p;

  this->child_parent_ttl(hop->ttl, &prev, &p);
  if (hop == p) {
    /* Already linked for this host. This can happen a reply for a higher TTL
       results in a merge, and later a reply for a lower TTL comes back. */
    return;
  }
  if (o.debugging > 1) {
    log_write(LOG_STDOUT, "Merging traces below TTL %d for %s",
      hop->ttl, ss_to_string(&hop->tag));
    log_write(LOG_STDOUT, " and %s\n", target->targetipstr());
  }
  hop = merge_hops(&hop->tag, hop, p);
  if (prev == NULL)
    this->hops = hop;
  else
    prev->parent = hop;
}

double HostState::completion_fraction() const {
  unsigned int i, n;

  if (this->is_finished())
    return 1.0;

  n = 0;
  for (i = 0; i < sent_ttls.size(); i++) {
    if (sent_ttls[i])
      n++;
  }

  return (double) n / sent_ttls.size();
}

void HostState::child_parent_ttl(u8 ttl, Hop **child, Hop **parent) const {
  *child = NULL;
  *parent = this->hops;
  while (*parent != NULL && (*parent)->ttl > ttl) {
    *child = *parent;
    *parent = (*parent)->parent;
  }
}

u8 HostState::distance_guess(const Target *target) {
  /* Use the distance from OS detection if we have it. */
  if (target->distance != -1)
    return target->distance;
  else
    /* initial_ttl is a variable with file-level scope. */
    return initial_ttl;
}

/* Get the probe that will be used for the traceroute. This is the
   highest-quality probe found in ping or port scanning, or ICMP echo if no
   responsive probe is known. */
struct probespec HostState::get_probe(const Target *target) {
  struct probespec probe;

  probe = target->pingprobe;
  if (target->af() == AF_INET &&
      (probe.type == PS_TCP || probe.type == PS_UDP || probe.type == PS_SCTP || probe.type == PS_ICMP)) {
    /* Nothing needed. */
  } else if (target->af() == AF_INET6 &&
      (probe.type == PS_TCP || probe.type == PS_UDP || probe.type == PS_SCTP || probe.type == PS_ICMPV6)) {
    /* Nothing needed. */
  } else if (probe.type == PS_PROTO) {
    /* If this is an IP protocol probe, fill in some fields for some common
       protocols. We cheat and store them in the TCP-, UDP-, SCTP- and
       ICMP-specific fields. */
    if (probe.proto == IPPROTO_TCP) {
      probe.pd.tcp.flags = TH_ACK;
      probe.pd.tcp.dport = get_random_u16();
    } else if (probe.proto == IPPROTO_UDP) {
      probe.pd.udp.dport = get_random_u16();
    } else if (probe.proto == IPPROTO_SCTP) {
      probe.pd.sctp.dport = get_random_u16();
    } else if (probe.proto == IPPROTO_ICMP) {
      probe.pd.icmp.type = ICMP_ECHO;
    } else if (probe.proto == IPPROTO_ICMPV6) {
      probe.pd.icmp.type = ICMPV6_ECHO;
    } else {
      fatal("Unknown protocol %d", probe.proto);
    }
  } else {
    /* No responsive probe known? The user probably skipped both ping and
       port scan. Guess ICMP echo as the most likely to get a response. */
    if (target->af() == AF_INET) {
      probe.type = PS_ICMP;
      probe.proto = IPPROTO_ICMP;
      probe.pd.icmp.type = ICMP_ECHO;
      probe.pd.icmp.code = 0;
    } else if (target->af() == AF_INET6) {
      probe.type = PS_ICMPV6;
      probe.proto = IPPROTO_ICMPV6;
      probe.pd.icmp.type = ICMPV6_ECHO;
      probe.pd.icmp.code = 0;
    } else {
      fatal("Unknown address family %d", target->af());
    }
  }

  return probe;
}

Probe::Probe(HostState *host, struct probespec pspec, u8 ttl) {
  this->host = host;
  this->pspec = pspec;
  this->ttl = ttl;
  token = Probe::token_counter++;
  sent_time.tv_sec = 0;
  sent_time.tv_usec = 0;
  num_resends = 0;
}

Probe::~Probe() {
}

void Probe::send(int rawsd, eth_t *ethsd, struct timeval *now) {
  struct eth_nfo eth;
  struct eth_nfo *ethp;
  int decoy;

  /* Set up the Ethernet handle if we're using that. */
  if (ethsd != NULL) {
    memcpy(eth.srcmac, host->target->SrcMACAddress(), 6);
    memcpy(eth.dstmac, host->target->NextHopMACAddress(), 6);
    eth.ethsd = ethsd;
    eth.devname[0] = '\0';
    ethp = &eth;
  } else {
    ethp = NULL;
  }

  for (decoy = 0; decoy < o.numdecoys; decoy++) {
    const struct sockaddr_storage *source;
    unsigned char *packet;
    u32 packetlen;

    if (decoy == o.decoyturn) {
      source = host->target->SourceSockAddr();
      sent_time = get_now(now);
    } else {
      source = &(o.decoys[decoy]);
    }

    packet = this->build_packet(source, &packetlen);
    send_ip_packet(rawsd, ethp, host->target->TargetSockAddr(), packet, packetlen);
    free(packet);
  }
}

void Probe::resend(int rawsd, eth_t *ethsd, struct timeval *now) {
  num_resends++;
  this->send(rawsd, ethsd, now);
}

bool Probe::is_timedout(struct timeval *now) const {
  struct timeval tv;

  tv = get_now(now);

  return TIMEVAL_MSEC_SUBTRACT(tv, sent_time) > PROBE_TIMEOUT;
}

bool Probe::may_resend() const {
  return num_resends < MIN(o.getMaxRetransmissions(), MAX_RESENDS);
}

/* Probe is an abstract class with a missing build_packet method. These concrete
   subclasses implement the method for different probe types. */

class ICMPProbe : public Probe
{
public:
  ICMPProbe(HostState *host, struct probespec pspec, u8 ttl)
  : Probe(host, pspec, ttl) {
  }

  unsigned char *build_packet(const struct sockaddr_storage *source, u32 *len) const {
    const struct sockaddr_in *sin;
    assert(source->ss_family == AF_INET);
    sin = (struct sockaddr_in *) source;
    return build_icmp_raw(&sin->sin_addr, host->target->v4hostip(), ttl,
      0x0000, 0x00, false, NULL, 0, token, global_id,
      pspec.pd.icmp.type, pspec.pd.icmp.code,
      o.extra_payload, o.extra_payload_length, len);
  }
};

class TCPProbe : public Probe
{
public:
  TCPProbe(HostState *host, struct probespec pspec, u8 ttl)
  : Probe(host, pspec, ttl) {
  }
  unsigned char *build_packet(const struct sockaddr_storage *source, u32 *len) const {
    const char *tcpopts;
    int tcpoptslen;
    u32 ack;

    tcpopts = NULL;
    tcpoptslen = 0;
    ack = 0;
    if ((pspec.pd.tcp.flags & TH_SYN) == TH_SYN) {
      /* MSS 1460 bytes. */
      tcpopts = TCP_SYN_PROBE_OPTIONS;
      tcpoptslen = TCP_SYN_PROBE_OPTIONS_LEN;
    } else if ((pspec.pd.tcp.flags & TH_ACK) == TH_ACK) {
      ack = get_random_u32();
    }

    /* For TCP we encode the token in the source port. */
    if (source->ss_family == AF_INET) {
      const struct sockaddr_in *sin = (struct sockaddr_in *) source;
      return build_tcp_raw(&sin->sin_addr, host->target->v4hostip(), ttl,
        get_random_u16(), get_random_u8(), false, NULL, 0,
        token ^ global_id, pspec.pd.tcp.dport, get_random_u32(), ack, 0x00,
        pspec.pd.tcp.flags, get_random_u16(), 0, (const u8 *) tcpopts, tcpoptslen,
        o.extra_payload, o.extra_payload_length, len);
    } else if (source->ss_family == AF_INET6) {
      const struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *) source;
      return build_tcp_raw_ipv6(&sin6->sin6_addr, host->target->v6hostip(),
        0, 0, ttl,
        token ^ global_id, pspec.pd.tcp.dport, get_random_u32(), ack, 0x00,
        pspec.pd.tcp.flags, get_random_u16(), 0, (const u8 *) tcpopts, tcpoptslen,
        o.extra_payload, o.extra_payload_length, len);
    } else {
      fatal("Unknown address family %u in %s.", source->ss_family, __func__);
    }
  }
};

class UDPProbe : public Probe
{
public:
  UDPProbe(HostState *host, struct probespec pspec, u8 ttl)
  : Probe(host, pspec, ttl) {
  }
  unsigned char *build_packet(const struct sockaddr_storage *source, u32 *len) const {
    const u8 *payload;
    int payload_length;

    payload = get_udp_payload(pspec.pd.udp.dport, &payload_length, 0);

    /* For UDP we encode the token in the source port. */
    if (source->ss_family == AF_INET) {
      const struct sockaddr_in *sin = (struct sockaddr_in *) source;
      return build_udp_raw(&sin->sin_addr, host->target->v4hostip(), ttl,
        get_random_u16(), get_random_u8(), false, NULL, 0,
        token ^ global_id, pspec.pd.udp.dport,
        (char *) payload, payload_length, len);
    } else if (source->ss_family == AF_INET6) {
      const struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *) source;
      return build_udp_raw_ipv6(&sin6->sin6_addr, host->target->v6hostip(),
        0, 0, ttl,
        token ^ global_id, pspec.pd.udp.dport,
        (char *) payload, payload_length, len);
    } else {
      fatal("Unknown address family %u in %s.", source->ss_family, __func__);
    }
  }
};

class SCTPProbe : public Probe
{
public:
  SCTPProbe(HostState *host, struct probespec pspec, u8 ttl)
  : Probe(host, pspec, ttl) {
  }
  unsigned char *build_packet(const struct sockaddr_storage *source, u32 *len) const {
    struct sctp_chunkhdr_init chunk;

    sctp_pack_chunkhdr_init(&chunk, SCTP_INIT, 0, sizeof(chunk),
      get_random_u32() /*itag*/, 32768, 10, 2048, get_random_u32() /*itsn*/);

    if (source->ss_family == AF_INET) {
      const struct sockaddr_in *sin = (struct sockaddr_in *) source;
      return build_sctp_raw(&sin->sin_addr, host->target->v4hostip(), ttl,
        get_random_u16(), get_random_u8(), false, NULL, 0,
        token ^ global_id, pspec.pd.sctp.dport, 0UL,
        (char *) &chunk, sizeof(chunk),
        o.extra_payload, o.extra_payload_length, len);
    } else if (source->ss_family == AF_INET6) {
      const struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *) source;
      return build_sctp_raw_ipv6(&sin6->sin6_addr, host->target->v6hostip(),
        0, 0, ttl,
        token ^ global_id, pspec.pd.sctp.dport, 0UL,
        (char *) &chunk, sizeof(chunk),
        o.extra_payload, o.extra_payload_length, len);
    } else {
      fatal("Unknown address family %u in %s.", source->ss_family, __func__);
    }
  }
};

class IPProtoProbe : public Probe
{
public:
  IPProtoProbe(HostState *host, struct probespec pspec, u8 ttl)
  : Probe(host, pspec, ttl) {
  }
  unsigned char *build_packet(const struct sockaddr_storage *source, u32 *len) const {
    /* For IP proto scan the token is put in the IP ID or flow label. */
    if (source->ss_family == AF_INET) {
      const struct sockaddr_in *sin = (struct sockaddr_in *) source;
      return build_ip_raw(&sin->sin_addr, host->target->v4hostip(), pspec.proto, ttl,
        token ^ global_id, get_random_u8(), false, NULL, 0,
        o.extra_payload, o.extra_payload_length, len);
    } else if (source->ss_family == AF_INET6) {
      const struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *) source;
      return build_ipv6_raw(&sin6->sin6_addr, host->target->v6hostip(),
        0, token ^ global_id, pspec.proto, ttl,
        o.extra_payload, o.extra_payload_length, len);
    } else {
      fatal("Unknown address family %u in %s.", source->ss_family, __func__);
    }
  }
};

class ICMPv6Probe : public Probe
{
public:
  ICMPv6Probe(HostState *host, struct probespec pspec, u8 ttl)
  : Probe(host, pspec, ttl) {
  }

  unsigned char *build_packet(const struct sockaddr_storage *source, u32 *len) const {
    const struct sockaddr_in6 *sin6;
    assert(source->ss_family == AF_INET6);
    sin6 = (struct sockaddr_in6 *) source;
    return build_icmpv6_raw(&sin6->sin6_addr, host->target->v6hostip(), 0x00, 0x0000,
      ttl, token, global_id, pspec.pd.icmp.type, pspec.pd.icmp.code,
      o.extra_payload, o.extra_payload_length, len);
  }
};

Probe *Probe::make(HostState *host, struct probespec pspec, u8 ttl)
{
  if (pspec.type == PS_ICMP || (pspec.type == PS_PROTO && pspec.proto == IPPROTO_ICMP))
    return new ICMPProbe(host, pspec, ttl);
  else if (pspec.type == PS_TCP || (pspec.type == PS_PROTO && pspec.proto == IPPROTO_TCP))
    return new TCPProbe(host, pspec, ttl);
  else if (pspec.type == PS_UDP || (pspec.type == PS_PROTO && pspec.proto == IPPROTO_UDP))
    return new UDPProbe(host, pspec, ttl);
  else if (pspec.type == PS_SCTP || (pspec.type == PS_PROTO && pspec.proto == IPPROTO_SCTP))
    return new SCTPProbe(host, pspec, ttl);
  else if (pspec.type == PS_PROTO)
    return new IPProtoProbe(host, pspec, ttl);
  else if (pspec.type == PS_ICMPV6)
    return new ICMPv6Probe(host, pspec, ttl);
  else
    fatal("Unknown probespec type in traceroute");

  return NULL;
}

TracerouteState::TracerouteState(std::vector<Target *> &targets) {
  std::vector<Target *>::const_iterator it;
  char pcap_filter[128];
  int n;

  assert(targets.size() > 0);

  if ((o.sendpref & PACKET_SEND_ETH) && targets[0]->ifType() == devt_ethernet) {
    /* No need to check for g_has_npcap_loopback on WIN32 because devt_loopback
     * is checked earlier. */
    ethsd = eth_open_cached(targets[0]->deviceName());
    if (ethsd == NULL)
      fatal("dnet: failed to open device %s", targets[0]->deviceName());
    rawsd = -1;
  } else {
#ifdef WIN32
    win32_fatal_raw_sockets(targets[0]->deviceName());
#endif
    rawsd = nmap_raw_socket();
    if (rawsd < 0)
      pfatal("traceroute: socket troubles");
    ethsd = NULL;
  }

  /* Assume that all the targets share the same device. */
  if((pd=my_pcap_open_live(targets[0]->deviceName(), 128, o.spoofsource, 2))==NULL)
    fatal("%s", PCAP_OPEN_ERRMSG);
  n = Snprintf(pcap_filter, sizeof(pcap_filter), "(ip or ip6) and dst host %s",
    ss_to_string(targets[0]->SourceSockAddr()));
  assert(n < (int) sizeof(pcap_filter));
  set_pcap_filter(targets[0]->deviceFullName(), pd, pcap_filter);
 if (o.debugging)
   log_write(LOG_STDOUT, "Packet capture filter (device %s): %s\n", targets[0]->deviceFullName(), pcap_filter);
  for (it = targets.begin(); it != targets.end(); it++) {
    HostState *state = new HostState(*it);
    hosts.push_back(state);
    active_hosts.push_back(state);
  }

  num_active_probes = 0;
  next_sending_host = active_hosts.begin();
  next_send_time = get_now();
}

TracerouteState::~TracerouteState() {
  std::vector<HostState *>::iterator it;

  if (rawsd != -1)
    close(rawsd);
  pcap_close(pd);

  for (it = hosts.begin(); it != hosts.end(); it++)
    delete *it;
}

void TracerouteState::next_active_host() {
  assert(next_sending_host != active_hosts.end());
  next_sending_host++;
  if (next_sending_host == active_hosts.end())
    next_sending_host = active_hosts.begin();
}

void TracerouteState::send_new_probes() {
  std::list<HostState *>::const_iterator failed_host;
  struct timeval now;

  now = get_now();

  assert(!active_hosts.empty());
  failed_host = active_hosts.end();
  while (next_sending_host != failed_host
    && num_active_probes < MAX_OUTSTANDING_PROBES
    && !TIMEVAL_BEFORE(now, next_send_time)) {
    if ((*next_sending_host)->send_next_probe(rawsd, ethsd)) {
      num_active_probes++;
      TIMEVAL_MSEC_ADD(next_send_time, next_send_time, o.scan_delay);
      if (TIMEVAL_BEFORE(next_send_time, now))
        next_send_time = now;
      failed_host = active_hosts.end();
    } else if (failed_host == active_hosts.end()) {
      failed_host = next_sending_host;
    }
    next_active_host();
  }
}

static Hop *hop_cache_lookup(u8 ttl, const struct sockaddr_storage *addr) {
  std::map<struct HopIdent, Hop *>::iterator it;
  HopIdent ident(ttl, *addr);

  it = hop_cache.find(ident);
  if (it == hop_cache.end())
    return NULL;
  else
    return it->second;
}

static void hop_cache_insert(Hop *hop) {
  if (hop->addr.ss_family == 0) {
    timedout_hops->push_back(hop);
  } else {
    hop_cache[HopIdent(hop->ttl, hop->addr)] = hop;
  }
}

static unsigned int hop_cache_size() {
  return hop_cache.size() + timedout_hops->size();
}

void traceroute_hop_cache_clear() {
  std::map<struct HopIdent, Hop *>::iterator map_iter;
  std::list<Hop *>::iterator list_iter;

  for (map_iter = hop_cache.begin(); map_iter != hop_cache.end(); map_iter++)
    delete map_iter->second;
  hop_cache.clear();
  if (!timedout_hops) return;
  for (list_iter = timedout_hops->begin(); list_iter != timedout_hops->end(); list_iter++)
    delete *list_iter;
  timedout_hops->clear();
}

/* Merge two hop chains together and return the head of the merged chain. This
   is done when a cache hit finds that two targets share the same intermediate
   hop; rather than doing a full trace for each target, one is linked to the
   other. "Merged" means that, for example, if chain a has an hop at TTL 3 and
   chain b has one for TTL 2, the merged chain will include both. Usually, the
   chains will not have hops at the same TTL (that implies different routes to
   the same host), but see the next paragraph for what we do when that happens.
   Each hop in the merged chain will be tagged with the given tag.

   There are many cases that must be handled correctly by this function: a and b
   may be equal; either may be NULL; a and b may be disjoint chains or may be
   joined somewhere. The biggest difficulty is when both of the chains have a
   hop with the same TTL but a different address. When this happens we
   arbitrarily choose one of the hops to unlink, on the presumption that any
   route through the same intermediate host at a given TTL should be the same
   and that differences aren't meaningful. (This has the same effect as if we
   were to send probes strictly serially, because then there would be no parent
   hops to potentially conflict, even if in fact they would if traced to
   completion.) */
static Hop *merge_hops(const struct sockaddr_storage *tag, Hop *a, Hop *b) {
  Hop head, *p;

  p = &head;

  while (a != NULL && b != NULL && a != b) {
    Hop **next;

    if (a->ttl > b->ttl) {
      next = &a;
    } else if (b->ttl > a->ttl) {
      next = &b;
    } else {
      Hop **discard, *parent;

      if (b->addr.ss_family == 0) {
        next = &a;
        discard = &b;
      } else if (a->addr.ss_family == 0) {
        next = &b;
        discard = &a;
      } else {
        next = &a;
        discard = &b;
        if (o.debugging) {
          log_write(LOG_STDOUT, "Warning: %s", ss_to_string(&(*next)->addr));
          log_write(LOG_STDOUT, " doesn't match %s at TTL %d;",
            ss_to_string(&(*discard)->addr), a->ttl);
          log_write(LOG_STDOUT, " arbitrarily discarding %s\n",
            ss_to_string(&(*discard)->addr));
        }
      }
      parent = (*discard)->parent;
      *discard = parent;
    }
    p->parent = *next;
    p->tag = *tag;
    p = p->parent;
    *next = (*next)->parent;
  }
  /* At most one branch of this is taken, even when a == b. */
  if (a != NULL)
    p->parent = a;
  else if (b != NULL)
    p->parent = b;
  for (; p != NULL; p = p->parent)
    p->tag = *tag;

  return head.parent;
}

/* Record a hop at the given TTL for the given host. This takes care of linking
   the hop into the host's chain as well as into the global hop tree. */
void TracerouteState::set_host_hop(HostState *host, u8 ttl,
  const struct sockaddr_storage *from_addr, float rtt) {
  Hop *hop;

  if (o.debugging > 1) {
    log_write(LOG_STDOUT, "Set hop %s TTL %d to %s RTT %.2f ms\n",
      host->target->targetipstr(), ttl, ss_to_string(from_addr), rtt);
  }

  hop = hop_cache_lookup(ttl, from_addr);
  if (hop == NULL) {
    /* A new hop, never before seen with this address and TTL. Add it to the
       host's chain and to the global cache. */
    hop = host->insert_hop(ttl, from_addr, rtt);
  } else {
    /* An existing hop at this address and TTL. Link this host's chain to it. */
    if (o.debugging > 1) {
      log_write(LOG_STDOUT, "Traceroute cache hit %s TTL %d while tracing",
        ss_to_string(&hop->addr), hop->ttl);
      log_write(LOG_STDOUT, " %s TTL %d\n", host->target->targetipstr(), ttl);
    }

    host->link_to(hop);

    if (host->state == HostState::COUNTING_DOWN) {
      /* Hit the cache going down. Seek to the end of the chain. If we have the
         tag for the last node, we take responsibility for finishing the trace.
         Otherwise, start counting up. */

      while (hop->parent != NULL) {
        hop = hop->parent;
        /* No need to re-probe any merged hops. */
        host->sent_ttls[hop->ttl] = true;
      }
      if (sockaddr_storage_equal(&hop->tag, host->target->TargetSockAddr())) {
        if (o.debugging > 1) {
          log_write(LOG_STDOUT, "%s continuing trace from TTL %d\n",
            host->target->targetipstr(), host->current_ttl);
        }
      } else {
        host->state = HostState::COUNTING_UP;
        num_active_probes -= host->cancel_probes_below(ttl);
      }
    }
  }
}

/* Record that a hop at the given TTL for the given host timed out. */
void TracerouteState::set_host_hop_timedout(HostState *host, u8 ttl) {
  static struct sockaddr_storage EMPTY_ADDR = { 0 };
  host->insert_hop(ttl, &EMPTY_ADDR, -1.0);
}

struct Reply {
  struct timeval rcvdtime;
  struct sockaddr_storage from_addr;
  struct sockaddr_storage target_addr;
  u8 ttl;
  u16 token;
};

static bool parse_encapsulated_reply(const void *ip, unsigned len, Reply *reply) {
  struct abstract_ip_hdr hdr;
  const void *data;

  data = ip_get_data(ip, &len, &hdr);
  if (data == NULL)
    return false;

  if (hdr.version == 4 && hdr.proto == IPPROTO_ICMP) {
    const struct icmp *icmp = (const struct icmp *) data;
    if (len < 8 || ntohs(icmp->icmp_id) != global_id)
      return false;
    reply->token = ntohs(icmp->icmp_seq);
  } else if (hdr.version == 6 && hdr.proto == IPPROTO_ICMPV6) {
    const struct icmpv6_msg_echo *echo = (struct icmpv6_msg_echo *) ((char *) data + sizeof(struct icmpv6_hdr));
    if (len < 8 || ntohs(echo->icmpv6_id) != global_id)
      return false;
    reply->token = ntohs(echo->icmpv6_seq);
  } else if (hdr.proto == IPPROTO_TCP) {
    const struct tcp_hdr *tcp = (const struct tcp_hdr *) data;
    if (len < 2)
      return false;
    reply->token = ntohs(tcp->th_sport) ^ global_id;
  } else if (hdr.proto == IPPROTO_UDP) {
    const struct udp_hdr *udp = (const struct udp_hdr *) data;
    if (len < 2)
      return false;
    reply->token = ntohs(udp->uh_sport) ^ global_id;
  } else if (hdr.proto == IPPROTO_SCTP) {
    const struct sctp_hdr *sctp = (const struct sctp_hdr *) data;
    if (len < 2)
      return false;
    reply->token = ntohs(sctp->sh_sport) ^ global_id;
  } else {
    if (len < 6)
      return false;
    /* Check IP ID for proto scan. */
    reply->token = hdr.ipid ^ global_id;
  }

  reply->target_addr = hdr.dst;

  return true;
}

static bool decode_reply(const void *ip, unsigned int len, Reply *reply) {
  struct abstract_ip_hdr hdr;
  const void *data;

  data = ip_get_data(ip, &len, &hdr);
  if (data == NULL)
    return false;

  reply->from_addr = hdr.src;
  reply->ttl = hdr.ttl;

  if (hdr.version == 4 && hdr.proto == IPPROTO_ICMP) {
    /* ICMP responses comprise all the TTL exceeded messages we expect from all
       probe types, as well as actual replies from ICMP probes. */
    const struct icmp_hdr *icmp = (const struct icmp_hdr *) data;
    if (len < 8)
      return false;
    if ((icmp->icmp_type == ICMP_TIMEXCEED
         && icmp->icmp_code == ICMP_TIMEXCEED_INTRANS)
        || icmp->icmp_type == ICMP_UNREACH) {
      /* Get the encapsulated IP packet. */
      const void *encaps = icmp_get_data(icmp, &len);
      if (encaps == NULL)
        return false;
      return parse_encapsulated_reply(encaps, len, reply);
    } else if (icmp->icmp_type == ICMP_ECHOREPLY
               || icmp->icmp_type == ICMP_MASKREPLY
               || icmp->icmp_type == ICMP_TSTAMPREPLY) {
      /* Need this alternate form of header for icmp_id and icmp_seq. */
      const struct icmp *icmp = (const struct icmp *) data;

      if (ntohs(icmp->icmp_id) != global_id)
        return false;
      reply->token = ntohs(icmp->icmp_seq);
      /* Reply came directly from the target. */
      reply->target_addr = reply->from_addr;
    } else {
      return false;
    }
  } else if (hdr.version == 6 && hdr.proto == IP_PROTO_ICMPV6) {
    /* ICMPv6 responses comprise all the TTL exceeded messages we expect from
       all probe types, as well as actual replies from ICMP probes. */
    const struct icmpv6_hdr *icmpv6 = (const struct icmpv6_hdr *) data;
    if (len < 2)
      return false;
    /* TIMEXCEED, UNREACH */
    if ((icmpv6->icmpv6_type == ICMPV6_TIMEXCEED
         && icmpv6->icmpv6_code == ICMPV6_TIMEXCEED_INTRANS)
        || icmpv6->icmpv6_type == ICMPV6_UNREACH) {
      /* Get the encapsulated IP packet. */
      const void *encaps = icmpv6_get_data(icmpv6, &len);
      if (encaps == NULL)
        return false;
      return parse_encapsulated_reply(encaps, len, reply);
    } else if (icmpv6->icmpv6_type == ICMPV6_ECHOREPLY) {
      /* MASKREPLY, TSTAMPREPLY */
      const struct icmpv6_msg_echo *echo;

      if (len < sizeof(*icmpv6) + 4)
        return false;
      echo = (struct icmpv6_msg_echo *) ((char *) icmpv6 + sizeof(*icmpv6));
      if (ntohs(echo->icmpv6_id) != global_id)
        return false;
      reply->token = ntohs(echo->icmpv6_seq);
      /* Reply came directly from the target. */
      reply->target_addr = reply->from_addr;
    } else {
      return false;
    }
  } else if (hdr.proto == IPPROTO_TCP) {
    const struct tcp_hdr *tcp = (const struct tcp_hdr *) data;
    if (len < 4)
      return false;
    reply->token = ntohs(tcp->th_dport) ^ global_id;
    reply->target_addr = reply->from_addr;
  } else if (hdr.proto == IPPROTO_UDP) {
    const struct udp_hdr *udp = (const struct udp_hdr *) data;
    if (len < 4)
      return false;
    reply->token = ntohs(udp->uh_dport) ^ global_id;
    reply->target_addr = reply->from_addr;
  } else if (hdr.proto == IPPROTO_SCTP) {
    const struct sctp_hdr *sctp = (const struct sctp_hdr *) data;
    if (len < 4)
      return false;
    reply->token = ntohs(sctp->sh_dport) ^ global_id;
    reply->target_addr = reply->from_addr;
  } else {
    return false;
  }

  return true;
}

static bool read_reply(Reply *reply, pcap_t *pd, long timeout) {
  const struct ip *ip;
  unsigned int iplen;
  struct link_header linkhdr;

  ip = (struct ip *) readip_pcap(pd, &iplen, timeout, &reply->rcvdtime, &linkhdr, true);
  if (ip == NULL)
    return false;
  if (ip->ip_v == 4 || ip->ip_v == 6)
    return decode_reply(ip, iplen, reply);
  else
    return false;
}

void TracerouteState::read_replies(long timeout) {
  struct timeval now;
  Reply reply;

  assert(timeout / 1000 <= (long) o.scan_delay);
  timeout = MAX(timeout, 10000);
  now = get_now();

  while (timeout > 0 && read_reply(&reply, pd, timeout)) {
    std::list<Probe *>::iterator it;
    struct timeval oldnow;
    HostState *host;
    Probe *probe;
    float rtt;

    oldnow = now;
    now = get_now();
    timeout -= TIMEVAL_SUBTRACT(now, oldnow);

    probe = this->lookup_probe(&reply.target_addr, reply.token);
    if (probe == NULL)
      continue;
    host = probe->host;

    if (sockaddr_storage_equal(host->target->TargetSockAddr(), &reply.from_addr)) {
      adjust_timeouts2(&probe->sent_time, &reply.rcvdtime, &host->target->to);
      if (host->reached_target == 0 || probe->ttl < host->reached_target)
        host->reached_target = probe->ttl;
      if (host->state == HostState::COUNTING_DOWN) {
        /* If this probe was past the target, skip ahead to what we think the
           actual distance is. */
        int distance = get_initial_ttl_guess(reply.ttl) - reply.ttl + 1;
        if (distance > 0 && distance < host->current_ttl)
          host->current_ttl = MIN(distance, MAX_TTL);
      }
      num_active_probes -= host->cancel_probes_above(probe->ttl);
    }

    rtt = TIMEVAL_SUBTRACT(reply.rcvdtime, probe->sent_time) / 1000.0;
    set_host_hop(host, probe->ttl, &reply.from_addr, rtt);

    it = find(host->unanswered_probes.begin(), host->unanswered_probes.end(), probe);
    num_active_probes -= host->cancel_probe(it);
  }
}

void TracerouteState::cull_timeouts() {
  std::list<HostState *>::iterator host_iter;
  struct timeval now;

  now = get_now();

  for (host_iter = active_hosts.begin(); host_iter != active_hosts.end(); host_iter++) {
    while (!(*host_iter)->active_probes.empty()
           && (*host_iter)->active_probes.front()->is_timedout(&now)) {
      Probe *probe;

      probe = (*host_iter)->active_probes.front();
      if (o.debugging > 1) {
        log_write(LOG_STDOUT, "Traceroute probe to %s TTL %d timed out\n",
          probe->host->target->targetipstr(), probe->ttl);
      }
      set_host_hop_timedout(*host_iter, probe->ttl);
      (*host_iter)->active_probes.pop_front();
      num_active_probes--;
      if (probe->may_resend())
        (*host_iter)->pending_resends.push_front(probe);
    }
  }
}

void TracerouteState::remove_finished_hosts() {
  std::list<HostState *>::iterator it, next;

  for (it = active_hosts.begin(); it != active_hosts.end(); it = next) {
    next = it;
    next++;
    if ((*it)->is_finished()) {
      if (next_sending_host == it)
        next_active_host();
      active_hosts.erase(it);
    }
  }
}

/* Dummy class to use sockaddr_storage as a map key. */
struct lt_sockaddr_storage {
  bool operator()(const struct sockaddr_storage& a, const struct sockaddr_storage& b) const {
    return sockaddr_storage_cmp(&a, &b) < 0;
  }
};

/* Find the reverse-DNS names of the hops. */
void TracerouteState::resolve_hops() {
  std::set<sockaddr_storage, lt_sockaddr_storage> addrs;
  std::set<sockaddr_storage, lt_sockaddr_storage>::iterator addr_iter;
  std::vector<HostState *>::iterator host_iter;
  std::map<sockaddr_storage, const char *, lt_sockaddr_storage> name_map;
  Target **targets;
  Hop *hop;
  int i, n;

  /* First, put all the IPv4 addresses in a set to remove duplicates. This
     re-resolves the addresses of the targets themselves, which is a little
     inefficient. */
  for (host_iter = hosts.begin(); host_iter != hosts.end(); host_iter++) {
    for (hop = (*host_iter)->hops; hop != NULL; hop = hop->parent) {
      if (hop->addr.ss_family != AF_UNSPEC)
        addrs.insert(hop->addr);
    }
  }
  n = addrs.size();
  /* Second, make an array of pointer to Target to suit the interface of
     nmap_mass_rdns. */
  targets = (Target **) safe_malloc(sizeof(*targets) * n);
  i = 0;
  addr_iter = addrs.begin();
  while (i < n) {
    targets[i] = new Target();
    targets[i]->setTargetSockAddr(&*addr_iter, sizeof(*addr_iter));
    targets[i]->flags = HOST_UP;
    i++;
    addr_iter++;
  }
  nmap_mass_rdns(targets, n);
  /* Third, make a map from addresses to names for easy lookup. */
  for (i = 0; i < n; i++) {
    struct sockaddr_storage ss;
    size_t ss_len;
    const char *hostname = targets[i]->HostName();
    if (*hostname == '\0')
      hostname = NULL;
    ss_len = sizeof(ss);
    targets[i]->TargetSockAddr(&ss, &ss_len);
    name_map[ss] = hostname;
  }
  /* Finally, copy the names into the hops. */
  for (host_iter = hosts.begin(); host_iter != hosts.end(); host_iter++) {
    for (hop = (*host_iter)->hops; hop != NULL; hop = hop->parent) {
      if (hop->addr.ss_family != AF_UNSPEC) {
        const char *hostname = name_map[hop->addr];
        if (hostname != NULL)
          hop->hostname = hostname;
      }
    }
  }
  for (i = 0; i < n; i++)
    delete targets[i];
  free(targets);
}

void TracerouteState::transfer_hops() {
  std::vector<HostState *>::iterator it;
  Hop *p;

  for (it = hosts.begin(); it != hosts.end(); it++) {
    for (p = (*it)->hops; p != NULL; p = p->parent) {
      TracerouteHop hop;

      /* Trim excessive hops. */
      if ((*it)->reached_target && p->ttl > (*it)->reached_target)
        continue;

      hop.tag = p->tag;
      if (p->addr.ss_family == 0) {
        hop.timedout = true;
      } else {
        hop.timedout = false;
        hop.rtt = p->rtt;
      }
      hop.name = p->hostname;
      hop.addr = p->addr;
      hop.ttl = p->ttl;
      (*it)->target->traceroute_hops.push_front(hop);
    }

    (*it)->target->traceroute_probespec = (*it)->pspec;

    /* Set the hop distance for OS fingerprints. */
    if ((*it)->reached_target) {
      (*it)->target->distance = (*it)->reached_target;
      (*it)->target->distance_calculation_method = DIST_METHOD_TRACEROUTE;
    }
  }
}

Probe *TracerouteState::lookup_probe(
  const struct sockaddr_storage *target_addr, u16 token) {
  std::list<HostState *>::const_iterator host_iter;
  std::list<Probe *>::iterator probe_iter;

  for (host_iter = active_hosts.begin(); host_iter != active_hosts.end(); host_iter++) {

    if (!sockaddr_storage_equal((*host_iter)->target->TargetSockAddr(), target_addr))
      continue;
    for (probe_iter = (*host_iter)->unanswered_probes.begin();
         probe_iter != (*host_iter)->unanswered_probes.end();
         probe_iter++) {
      if ((*probe_iter)->token == token)
        return *probe_iter;
    }
  }

  return NULL;
}

double TracerouteState::completion_fraction() const {
  std::vector<HostState *>::const_iterator it;
  double sum;

  sum = 0.0;
  for (it = hosts.begin(); it != hosts.end(); it++)
    sum += (*it)->completion_fraction();
  return sum / hosts.size();
}

/* This is a special case of traceroute when all the targets are directly
   connected. Because the distance to each target is known to be 1, we send no
   probes at all, only fill in a TracerouteHop structure. */
static int traceroute_direct(std::vector<Target *> targets) {
  std::vector<Target *>::iterator it;

  for (it = targets.begin(); it != targets.end(); it++) {
    TracerouteHop hop;
    const char *hostname;
    size_t sslen;

    sslen = sizeof(hop.tag);
    (*it)->TargetSockAddr(&hop.tag, &sslen);
    hop.timedout = false;
    hop.rtt = (*it)->to.srtt / 1000.0;
    hostname = (*it)->HostName();
    if (hostname != NULL && hostname[0] != '\0')
      hop.name = hostname;
    hop.addr = hop.tag;
    hop.ttl = 1;
    (*it)->traceroute_hops.push_front(hop);
  }

  return 1;
}

static int traceroute_remote(std::vector<Target *> targets) {
  std::vector<Target *>::iterator target_iter;

  if (targets.empty())
    return 1;

  if (timedout_hops == NULL) {
    timedout_hops = new std::list<Hop *>;
  }

  TracerouteState global_state(targets);

  global_id = get_random_u16();

  ScanProgressMeter SPM("Traceroute");

  o.current_scantype = TRACEROUTE;

  while (!global_state.active_hosts.empty()) {
    struct timeval now;
    long int timeout;

    global_state.send_new_probes();
    now = get_now();
    timeout = TIMEVAL_SUBTRACT(global_state.next_send_time, now);
    global_state.read_replies(timeout);
    global_state.cull_timeouts();
    global_state.remove_finished_hosts();

    if (keyWasPressed())
      SPM.printStats(global_state.completion_fraction(), NULL);
  }

  SPM.endTask(NULL, NULL);

  if (!o.noresolve)
    global_state.resolve_hops();
  /* This puts the hops into the targets known by the global_state. */
  global_state.transfer_hops();

  /* Update initial_ttl to be the highest distance seen in this host group, as
     an estimate for the next. */
  initial_ttl = 0;
  for (target_iter = targets.begin();
       target_iter != targets.end();
       target_iter++) {
    initial_ttl = MAX(initial_ttl, (*target_iter)->traceroute_hops.size());
  }

  if (hop_cache_size() > MAX_HOP_CACHE_SIZE) {
    if (o.debugging) {
      log_write(LOG_STDOUT, "Clearing hop cache that has grown to %d\n",
        hop_cache_size());
    }
    traceroute_hop_cache_clear();
  }

  return 1;
}

int traceroute(std::vector<Target *> &Targets) {
  std::vector<Target *> direct, remote;
  std::vector<Target *>::iterator target_iter;

  /* Separate directly connected targets from remote targets. */
  for (target_iter = Targets.begin();
       target_iter != Targets.end();
       target_iter++) {
    if ((*target_iter)->ifType() == devt_loopback)
      ; /* Ignore */
    else if ((*target_iter)->directlyConnected())
      direct.push_back(*target_iter);
    else
      remote.push_back(*target_iter);
  }

  traceroute_direct(direct);
  traceroute_remote(remote);

  return 1;
}

static struct timeval get_now(struct timeval *now) {
  struct timeval tv;

  if (now != NULL)
    return *now;
  gettimeofday(&tv, NULL);

  return tv;
}

/* Convert the address in ss to a string. The result is returned in a static
   buffer so you can't call this twice in arguments to printf, for example. */
static const char *ss_to_string(const struct sockaddr_storage *ss) {
  return inet_ntop_ez(ss, sizeof(*ss));
}