aboutsummaryrefslogtreecommitdiffstats
path: root/sys/netinet/tcp_syncache.c
blob: 0475506916848023467c6adda5b085c9480f64a9 (plain) (blame)
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
/*-
 * Copyright (c) 2001 Networks Associates Technology, Inc.
 * All rights reserved.
 *
 * This software was developed for the FreeBSD Project by Jonathan Lemon
 * and NAI Labs, the Security Research Division of Network Associates, Inc.
 * under DARPA/SPAWAR contract N66001-01-C-8035 ("CBOSS"), as part of the
 * DARPA CHATS research program.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 * 3. The name of the author may not be used to endorse or promote
 *    products derived from this software without specific prior written
 *    permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 *
 * $FreeBSD$
 */

#include "opt_inet6.h"
#include "opt_ipsec.h"
#include "opt_mac.h"

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/kernel.h>
#include <sys/sysctl.h>
#include <sys/malloc.h>
#include <sys/mac.h>
#include <sys/mbuf.h>
#include <sys/md5.h>
#include <sys/proc.h>		/* for proc0 declaration */
#include <sys/random.h>
#include <sys/socket.h>
#include <sys/socketvar.h>

#include <net/if.h>
#include <net/route.h>

#include <netinet/in.h>
#include <netinet/in_systm.h>
#include <netinet/ip.h>
#include <netinet/in_var.h>
#include <netinet/in_pcb.h>
#include <netinet/ip_var.h>
#ifdef INET6
#include <netinet/ip6.h>
#include <netinet/icmp6.h>
#include <netinet6/nd6.h>
#include <netinet6/ip6_var.h>
#include <netinet6/in6_pcb.h>
#endif
#include <netinet/tcp.h>
#include <netinet/tcp_fsm.h>
#include <netinet/tcp_seq.h>
#include <netinet/tcp_timer.h>
#include <netinet/tcp_var.h>
#ifdef INET6
#include <netinet6/tcp6_var.h>
#endif

#ifdef IPSEC
#include <netinet6/ipsec.h>
#ifdef INET6
#include <netinet6/ipsec6.h>
#endif
#endif /*IPSEC*/

#ifdef FAST_IPSEC
#include <netipsec/ipsec.h>
#ifdef INET6
#include <netipsec/ipsec6.h>
#endif
#include <netipsec/key.h>
#define	IPSEC
#endif /*FAST_IPSEC*/

#include <machine/in_cksum.h>
#include <vm/uma.h>

static int tcp_syncookies = 1;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, syncookies, CTLFLAG_RW,
    &tcp_syncookies, 0, 
    "Use TCP SYN cookies if the syncache overflows");

static void	 syncache_drop(struct syncache *, struct syncache_head *);
static void	 syncache_free(struct syncache *);
static void	 syncache_insert(struct syncache *, struct syncache_head *);
struct syncache *syncache_lookup(struct in_conninfo *, struct syncache_head **);
static int	 syncache_respond(struct syncache *, struct mbuf *);
static struct 	 socket *syncache_socket(struct syncache *, struct socket *,
		    struct mbuf *m);
static void	 syncache_timer(void *);
static u_int32_t syncookie_generate(struct syncache *);
static struct syncache *syncookie_lookup(struct in_conninfo *,
		    struct tcphdr *, struct socket *);

/*
 * Transmit the SYN,ACK fewer times than TCP_MAXRXTSHIFT specifies.
 * 3 retransmits corresponds to a timeout of (1 + 2 + 4 + 8 == 15) seconds,
 * the odds are that the user has given up attempting to connect by then.
 */
#define SYNCACHE_MAXREXMTS		3

/* Arbitrary values */
#define TCP_SYNCACHE_HASHSIZE		512
#define TCP_SYNCACHE_BUCKETLIMIT	30

struct tcp_syncache {
	struct	syncache_head *hashbase;
	uma_zone_t zone;
	u_int	hashsize;
	u_int	hashmask;
	u_int	bucket_limit;
	u_int	cache_count;
	u_int	cache_limit;
	u_int	rexmt_limit;
	u_int	hash_secret;
	u_int	next_reseed;
	TAILQ_HEAD(, syncache) timerq[SYNCACHE_MAXREXMTS + 1];
	struct	callout tt_timerq[SYNCACHE_MAXREXMTS + 1];
};
static struct tcp_syncache tcp_syncache;

SYSCTL_NODE(_net_inet_tcp, OID_AUTO, syncache, CTLFLAG_RW, 0, "TCP SYN cache");

SYSCTL_INT(_net_inet_tcp_syncache, OID_AUTO, bucketlimit, CTLFLAG_RD,
     &tcp_syncache.bucket_limit, 0, "Per-bucket hash limit for syncache");

SYSCTL_INT(_net_inet_tcp_syncache, OID_AUTO, cachelimit, CTLFLAG_RD,
     &tcp_syncache.cache_limit, 0, "Overall entry limit for syncache");

SYSCTL_INT(_net_inet_tcp_syncache, OID_AUTO, count, CTLFLAG_RD,
     &tcp_syncache.cache_count, 0, "Current number of entries in syncache");

SYSCTL_INT(_net_inet_tcp_syncache, OID_AUTO, hashsize, CTLFLAG_RD,
     &tcp_syncache.hashsize, 0, "Size of TCP syncache hashtable");

SYSCTL_INT(_net_inet_tcp_syncache, OID_AUTO, rexmtlimit, CTLFLAG_RW,
     &tcp_syncache.rexmt_limit, 0, "Limit on SYN/ACK retransmissions");

static MALLOC_DEFINE(M_SYNCACHE, "syncache", "TCP syncache");

#define SYNCACHE_HASH(inc, mask) 					\
	((tcp_syncache.hash_secret ^					\
	  (inc)->inc_faddr.s_addr ^					\
	  ((inc)->inc_faddr.s_addr >> 16) ^ 				\
	  (inc)->inc_fport ^ (inc)->inc_lport) & mask)

#define SYNCACHE_HASH6(inc, mask) 					\
	((tcp_syncache.hash_secret ^					\
	  (inc)->inc6_faddr.s6_addr32[0] ^ 				\
	  (inc)->inc6_faddr.s6_addr32[3] ^ 				\
	  (inc)->inc_fport ^ (inc)->inc_lport) & mask)

#define ENDPTS_EQ(a, b) (						\
	(a)->ie_fport == (b)->ie_fport &&				\
	(a)->ie_lport == (b)->ie_lport &&				\
	(a)->ie_faddr.s_addr == (b)->ie_faddr.s_addr &&			\
	(a)->ie_laddr.s_addr == (b)->ie_laddr.s_addr			\
)

#define ENDPTS6_EQ(a, b) (memcmp(a, b, sizeof(*a)) == 0)

#define SYNCACHE_TIMEOUT(sc, slot) do {				\
	sc->sc_rxtslot = (slot);					\
	sc->sc_rxttime = ticks + TCPTV_RTOBASE * tcp_backoff[(slot)];	\
	TAILQ_INSERT_TAIL(&tcp_syncache.timerq[(slot)], sc, sc_timerq);	\
	if (!callout_active(&tcp_syncache.tt_timerq[(slot)]))		\
		callout_reset(&tcp_syncache.tt_timerq[(slot)],		\
		    TCPTV_RTOBASE * tcp_backoff[(slot)],		\
		    syncache_timer, (void *)((intptr_t)(slot)));	\
} while (0)

static void
syncache_free(struct syncache *sc)
{
	struct rtentry *rt;

	if (sc->sc_ipopts)
		(void) m_free(sc->sc_ipopts);
#ifdef INET6
	if (sc->sc_inc.inc_isipv6)
		rt = sc->sc_route6.ro_rt;
	else
#endif
		rt = sc->sc_route.ro_rt;
	if (rt != NULL) {
		/*
		 * If this is the only reference to a protocol cloned 
		 * route, remove it immediately.
		 */
		if (rt->rt_flags & RTF_WASCLONED &&
		    (sc->sc_flags & SCF_KEEPROUTE) == 0 &&
		    rt->rt_refcnt == 1)
			rtrequest(RTM_DELETE, rt_key(rt),
			    rt->rt_gateway, rt_mask(rt),
			    rt->rt_flags, NULL);
		RTFREE(rt);
	}
	uma_zfree(tcp_syncache.zone, sc);
}

void
syncache_init(void)
{
	int i;

	tcp_syncache.cache_count = 0;
	tcp_syncache.hashsize = TCP_SYNCACHE_HASHSIZE;
	tcp_syncache.bucket_limit = TCP_SYNCACHE_BUCKETLIMIT;
	tcp_syncache.cache_limit =
	    tcp_syncache.hashsize * tcp_syncache.bucket_limit;
	tcp_syncache.rexmt_limit = SYNCACHE_MAXREXMTS;
	tcp_syncache.next_reseed = 0;
	tcp_syncache.hash_secret = arc4random();

        TUNABLE_INT_FETCH("net.inet.tcp.syncache.hashsize",
	    &tcp_syncache.hashsize);
        TUNABLE_INT_FETCH("net.inet.tcp.syncache.cachelimit",
	    &tcp_syncache.cache_limit);
        TUNABLE_INT_FETCH("net.inet.tcp.syncache.bucketlimit",
	    &tcp_syncache.bucket_limit);
	if (!powerof2(tcp_syncache.hashsize)) {
                printf("WARNING: syncache hash size is not a power of 2.\n");
		tcp_syncache.hashsize = 512;	/* safe default */
        }
	tcp_syncache.hashmask = tcp_syncache.hashsize - 1;

	/* Allocate the hash table. */
	MALLOC(tcp_syncache.hashbase, struct syncache_head *,
	    tcp_syncache.hashsize * sizeof(struct syncache_head),
	    M_SYNCACHE, M_WAITOK);

	/* Initialize the hash buckets. */
	for (i = 0; i < tcp_syncache.hashsize; i++) {
		TAILQ_INIT(&tcp_syncache.hashbase[i].sch_bucket);
		tcp_syncache.hashbase[i].sch_length = 0;
	}

	/* Initialize the timer queues. */
	for (i = 0; i <= SYNCACHE_MAXREXMTS; i++) {
		TAILQ_INIT(&tcp_syncache.timerq[i]);
		callout_init(&tcp_syncache.tt_timerq[i], 0);
	}

	/*
	 * Allocate the syncache entries.  Allow the zone to allocate one
	 * more entry than cache limit, so a new entry can bump out an
	 * older one.
	 */
	tcp_syncache.cache_limit -= 1;
	tcp_syncache.zone = uma_zcreate("syncache", sizeof(struct syncache),
	    NULL, NULL, NULL, NULL, UMA_ALIGN_PTR, UMA_ZONE_NOFREE);
	uma_zone_set_max(tcp_syncache.zone, tcp_syncache.cache_limit);
}

static void
syncache_insert(sc, sch)
	struct syncache *sc;
	struct syncache_head *sch;
{
	struct syncache *sc2;
	int s, i;

	/*
	 * Make sure that we don't overflow the per-bucket
	 * limit or the total cache size limit.
	 */
	s = splnet();
	if (sch->sch_length >= tcp_syncache.bucket_limit) {
		/*
		 * The bucket is full, toss the oldest element.
		 */
		sc2 = TAILQ_FIRST(&sch->sch_bucket);
		sc2->sc_tp->ts_recent = ticks;
		syncache_drop(sc2, sch);
		tcpstat.tcps_sc_bucketoverflow++;
	} else if (tcp_syncache.cache_count >= tcp_syncache.cache_limit) {
		/*
		 * The cache is full.  Toss the oldest entry in the
		 * entire cache.  This is the front entry in the
		 * first non-empty timer queue with the largest
		 * timeout value.
		 */
		for (i = SYNCACHE_MAXREXMTS; i >= 0; i--) {
			sc2 = TAILQ_FIRST(&tcp_syncache.timerq[i]);
			if (sc2 != NULL)
				break;
		}
		sc2->sc_tp->ts_recent = ticks;
		syncache_drop(sc2, NULL);
		tcpstat.tcps_sc_cacheoverflow++;
	}

	/* Initialize the entry's timer. */
	SYNCACHE_TIMEOUT(sc, 0);

	/* Put it into the bucket. */
	TAILQ_INSERT_TAIL(&sch->sch_bucket, sc, sc_hash);
	sch->sch_length++;
	tcp_syncache.cache_count++;
	tcpstat.tcps_sc_added++;
	splx(s);
}

static void
syncache_drop(sc, sch)
	struct syncache *sc;
	struct syncache_head *sch;
{
	int s;

	if (sch == NULL) {
#ifdef INET6
		if (sc->sc_inc.inc_isipv6) {
			sch = &tcp_syncache.hashbase[
			    SYNCACHE_HASH6(&sc->sc_inc, tcp_syncache.hashmask)];
		} else
#endif
		{
			sch = &tcp_syncache.hashbase[
			    SYNCACHE_HASH(&sc->sc_inc, tcp_syncache.hashmask)];
		}
	}

	s = splnet();

	TAILQ_REMOVE(&sch->sch_bucket, sc, sc_hash);
	sch->sch_length--;
	tcp_syncache.cache_count--;

	TAILQ_REMOVE(&tcp_syncache.timerq[sc->sc_rxtslot], sc, sc_timerq);
	if (TAILQ_EMPTY(&tcp_syncache.timerq[sc->sc_rxtslot]))
		callout_stop(&tcp_syncache.tt_timerq[sc->sc_rxtslot]);
	splx(s);

	syncache_free(sc);
}

/*
 * Walk the timer queues, looking for SYN,ACKs that need to be retransmitted.
 * If we have retransmitted an entry the maximum number of times, expire it.
 */
static void
syncache_timer(xslot)
	void *xslot;
{
	intptr_t slot = (intptr_t)xslot;
	struct syncache *sc, *nsc;
	struct inpcb *inp;
	int s;

	s = splnet();
	INP_INFO_WLOCK(&tcbinfo);
        if (callout_pending(&tcp_syncache.tt_timerq[slot]) ||
            !callout_active(&tcp_syncache.tt_timerq[slot])) {
		INP_INFO_WUNLOCK(&tcbinfo);
                splx(s);
                return;
        }
        callout_deactivate(&tcp_syncache.tt_timerq[slot]);

        nsc = TAILQ_FIRST(&tcp_syncache.timerq[slot]);
	while (nsc != NULL) {
		if (ticks < nsc->sc_rxttime)
			break;
		sc = nsc;
		inp = sc->sc_tp->t_inpcb;
		if (slot == SYNCACHE_MAXREXMTS ||
		    slot >= tcp_syncache.rexmt_limit ||
		    inp == NULL || inp->inp_gencnt != sc->sc_inp_gencnt) {
			nsc = TAILQ_NEXT(sc, sc_timerq);
			syncache_drop(sc, NULL);
			tcpstat.tcps_sc_stale++;
			continue;
		}
		/*
		 * syncache_respond() may call back into the syncache to
		 * to modify another entry, so do not obtain the next
		 * entry on the timer chain until it has completed.
		 */
		(void) syncache_respond(sc, NULL);
		nsc = TAILQ_NEXT(sc, sc_timerq);
		tcpstat.tcps_sc_retransmitted++;
		TAILQ_REMOVE(&tcp_syncache.timerq[slot], sc, sc_timerq);
		SYNCACHE_TIMEOUT(sc, slot + 1);
	}
	if (nsc != NULL)
		callout_reset(&tcp_syncache.tt_timerq[slot],
		    nsc->sc_rxttime - ticks, syncache_timer, (void *)(slot));
	INP_INFO_WUNLOCK(&tcbinfo);
	splx(s);
}

/*
 * Find an entry in the syncache.
 */
struct syncache *
syncache_lookup(inc, schp)
	struct in_conninfo *inc;
	struct syncache_head **schp;
{
	struct syncache *sc;
	struct syncache_head *sch;
	int s;

#ifdef INET6
	if (inc->inc_isipv6) {
		sch = &tcp_syncache.hashbase[
		    SYNCACHE_HASH6(inc, tcp_syncache.hashmask)];
		*schp = sch;
		s = splnet();
		TAILQ_FOREACH(sc, &sch->sch_bucket, sc_hash) {
			if (ENDPTS6_EQ(&inc->inc_ie, &sc->sc_inc.inc_ie)) {
				splx(s);
				return (sc);
			}
		}
		splx(s);
	} else
#endif
	{
		sch = &tcp_syncache.hashbase[
		    SYNCACHE_HASH(inc, tcp_syncache.hashmask)];
		*schp = sch;
		s = splnet();
		TAILQ_FOREACH(sc, &sch->sch_bucket, sc_hash) {
#ifdef INET6
			if (sc->sc_inc.inc_isipv6)
				continue;
#endif
			if (ENDPTS_EQ(&inc->inc_ie, &sc->sc_inc.inc_ie)) {
				splx(s);
				return (sc);
			}
		}
		splx(s);
	}
	return (NULL);
}

/*
 * This function is called when we get a RST for a
 * non-existent connection, so that we can see if the
 * connection is in the syn cache.  If it is, zap it.
 */
void
syncache_chkrst(inc, th)
	struct in_conninfo *inc;
	struct tcphdr *th;
{
	struct syncache *sc;
	struct syncache_head *sch;

	sc = syncache_lookup(inc, &sch);
	if (sc == NULL)
		return;
	/*
	 * If the RST bit is set, check the sequence number to see
	 * if this is a valid reset segment.
	 * RFC 793 page 37:
	 *   In all states except SYN-SENT, all reset (RST) segments
	 *   are validated by checking their SEQ-fields.  A reset is
	 *   valid if its sequence number is in the window.
	 *
	 *   The sequence number in the reset segment is normally an
	 *   echo of our outgoing acknowlegement numbers, but some hosts
	 *   send a reset with the sequence number at the rightmost edge
	 *   of our receive window, and we have to handle this case.
	 */
	if (SEQ_GEQ(th->th_seq, sc->sc_irs) &&
	    SEQ_LEQ(th->th_seq, sc->sc_irs + sc->sc_wnd)) {
		syncache_drop(sc, sch);
		tcpstat.tcps_sc_reset++;
	}
}

void
syncache_badack(inc)
	struct in_conninfo *inc;
{
	struct syncache *sc;
	struct syncache_head *sch;

	sc = syncache_lookup(inc, &sch);
	if (sc != NULL) {
		syncache_drop(sc, sch);
		tcpstat.tcps_sc_badack++;
	}
}

void
syncache_unreach(inc, th)
	struct in_conninfo *inc;
	struct tcphdr *th;
{
	struct syncache *sc;
	struct syncache_head *sch;

	/* we are called at splnet() here */
	sc = syncache_lookup(inc, &sch);
	if (sc == NULL)
		return;

	/* If the sequence number != sc_iss, then it's a bogus ICMP msg */
	if (ntohl(th->th_seq) != sc->sc_iss)
		return;

	/*
	 * If we've rertransmitted 3 times and this is our second error,
	 * we remove the entry.  Otherwise, we allow it to continue on.
	 * This prevents us from incorrectly nuking an entry during a
	 * spurious network outage.
	 *
	 * See tcp_notify().
	 */
	if ((sc->sc_flags & SCF_UNREACH) == 0 || sc->sc_rxtslot < 3) {
		sc->sc_flags |= SCF_UNREACH;
		return;
	}
	syncache_drop(sc, sch);
	tcpstat.tcps_sc_unreach++;
}

/*
 * Build a new TCP socket structure from a syncache entry.
 */
static struct socket *
syncache_socket(sc, lso, m)
	struct syncache *sc;
	struct socket *lso;
	struct mbuf *m;
{
	struct inpcb *inp = NULL;
	struct socket *so;
	struct tcpcb *tp;

	/*
	 * Ok, create the full blown connection, and set things up
	 * as they would have been set up if we had created the
	 * connection when the SYN arrived.  If we can't create
	 * the connection, abort it.
	 */
	so = sonewconn(lso, SS_ISCONNECTED);
	if (so == NULL) {
		/*
		 * Drop the connection; we will send a RST if the peer
		 * retransmits the ACK,
		 */
		tcpstat.tcps_listendrop++;
		goto abort;
	}
#ifdef MAC
	mac_set_socket_peer_from_mbuf(m, so);
#endif

	inp = sotoinpcb(so);

	/*
	 * Insert new socket into hash list.
	 */
	inp->inp_inc.inc_isipv6 = sc->sc_inc.inc_isipv6;
#ifdef INET6
	if (sc->sc_inc.inc_isipv6) {
		inp->in6p_laddr = sc->sc_inc.inc6_laddr;
	} else {
		inp->inp_vflag &= ~INP_IPV6;
		inp->inp_vflag |= INP_IPV4;
#endif
		inp->inp_laddr = sc->sc_inc.inc_laddr;
#ifdef INET6
	}
#endif
	inp->inp_lport = sc->sc_inc.inc_lport;
	if (in_pcbinshash(inp) != 0) {
		/*
		 * Undo the assignments above if we failed to
		 * put the PCB on the hash lists.
		 */
#ifdef INET6
		if (sc->sc_inc.inc_isipv6)
			inp->in6p_laddr = in6addr_any;
       		else
#endif
			inp->inp_laddr.s_addr = INADDR_ANY;
		inp->inp_lport = 0;
		goto abort;
	}
#ifdef IPSEC
	/* copy old policy into new socket's */
	if (ipsec_copy_policy(sotoinpcb(lso)->inp_sp, inp->inp_sp))
		printf("syncache_expand: could not copy policy\n");
#endif
#ifdef INET6
	if (sc->sc_inc.inc_isipv6) {
		struct inpcb *oinp = sotoinpcb(lso);
		struct in6_addr laddr6;
		struct sockaddr_in6 *sin6;
		/*
		 * Inherit socket options from the listening socket.
		 * Note that in6p_inputopts are not (and should not be)
		 * copied, since it stores previously received options and is
		 * used to detect if each new option is different than the
		 * previous one and hence should be passed to a user.
                 * If we copied in6p_inputopts, a user would not be able to
		 * receive options just after calling the accept system call.
		 */
		inp->inp_flags |= oinp->inp_flags & INP_CONTROLOPTS;
		if (oinp->in6p_outputopts)
			inp->in6p_outputopts =
			    ip6_copypktopts(oinp->in6p_outputopts, M_NOWAIT);
		inp->in6p_route = sc->sc_route6;
		sc->sc_route6.ro_rt = NULL;

		MALLOC(sin6, struct sockaddr_in6 *, sizeof *sin6,
		    M_SONAME, M_NOWAIT | M_ZERO);
		if (sin6 == NULL)
			goto abort;
		sin6->sin6_family = AF_INET6;
		sin6->sin6_len = sizeof(*sin6);
		sin6->sin6_addr = sc->sc_inc.inc6_faddr;
		sin6->sin6_port = sc->sc_inc.inc_fport;
		laddr6 = inp->in6p_laddr;
		if (IN6_IS_ADDR_UNSPECIFIED(&inp->in6p_laddr))
			inp->in6p_laddr = sc->sc_inc.inc6_laddr;
		if (in6_pcbconnect(inp, (struct sockaddr *)sin6, &thread0)) {
			inp->in6p_laddr = laddr6;
			FREE(sin6, M_SONAME);
			goto abort;
		}
		FREE(sin6, M_SONAME);
	} else
#endif
	{
		struct in_addr laddr;
		struct sockaddr_in *sin;

		inp->inp_options = ip_srcroute();
		if (inp->inp_options == NULL) {
			inp->inp_options = sc->sc_ipopts;
			sc->sc_ipopts = NULL;
		}
		inp->inp_route = sc->sc_route;
		sc->sc_route.ro_rt = NULL;

		MALLOC(sin, struct sockaddr_in *, sizeof *sin,
		    M_SONAME, M_NOWAIT | M_ZERO);
		if (sin == NULL)
			goto abort;
		sin->sin_family = AF_INET;
		sin->sin_len = sizeof(*sin);
		sin->sin_addr = sc->sc_inc.inc_faddr;
		sin->sin_port = sc->sc_inc.inc_fport;
		bzero((caddr_t)sin->sin_zero, sizeof(sin->sin_zero));
		laddr = inp->inp_laddr;
		if (inp->inp_laddr.s_addr == INADDR_ANY)
			inp->inp_laddr = sc->sc_inc.inc_laddr;
		if (in_pcbconnect(inp, (struct sockaddr *)sin, &thread0)) {
			inp->inp_laddr = laddr;
			FREE(sin, M_SONAME);
			goto abort;
		}
		FREE(sin, M_SONAME);
	}

	tp = intotcpcb(inp);
	tp->t_state = TCPS_SYN_RECEIVED;
	tp->iss = sc->sc_iss;
	tp->irs = sc->sc_irs;
	tcp_rcvseqinit(tp);
	tcp_sendseqinit(tp);
	tp->snd_wl1 = sc->sc_irs;
	tp->rcv_up = sc->sc_irs + 1;
	tp->rcv_wnd = sc->sc_wnd;
	tp->rcv_adv += tp->rcv_wnd;

	tp->t_flags = sototcpcb(lso)->t_flags & (TF_NOPUSH|TF_NODELAY);
	if (sc->sc_flags & SCF_NOOPT)
		tp->t_flags |= TF_NOOPT;
	if (sc->sc_flags & SCF_WINSCALE) {
		tp->t_flags |= TF_REQ_SCALE|TF_RCVD_SCALE;
		tp->requested_s_scale = sc->sc_requested_s_scale;
		tp->request_r_scale = sc->sc_request_r_scale;
	}
	if (sc->sc_flags & SCF_TIMESTAMP) {
		tp->t_flags |= TF_REQ_TSTMP|TF_RCVD_TSTMP;
		tp->ts_recent = sc->sc_tsrecent;
		tp->ts_recent_age = ticks;
	}
	if (sc->sc_flags & SCF_CC) {
		/*
		 * Initialization of the tcpcb for transaction;
		 *   set SND.WND = SEG.WND,
		 *   initialize CCsend and CCrecv.
		 */
		tp->t_flags |= TF_REQ_CC|TF_RCVD_CC;
		tp->cc_send = sc->sc_cc_send;
		tp->cc_recv = sc->sc_cc_recv;
	}

	tcp_mss(tp, sc->sc_peer_mss);

	/*
	 * If the SYN,ACK was retransmitted, reset cwnd to 1 segment.
	 */
	if (sc->sc_rxtslot != 0)
                tp->snd_cwnd = tp->t_maxseg;
	callout_reset(tp->tt_keep, tcp_keepinit, tcp_timer_keep, tp);

	tcpstat.tcps_accepts++;
	return (so);

abort:
	if (so != NULL)
		(void) soabort(so);
	return (NULL);
}

/*
 * This function gets called when we receive an ACK for a
 * socket in the LISTEN state.  We look up the connection
 * in the syncache, and if its there, we pull it out of
 * the cache and turn it into a full-blown connection in
 * the SYN-RECEIVED state.
 */
int
syncache_expand(inc, th, sop, m)
	struct in_conninfo *inc;
	struct tcphdr *th;
	struct socket **sop;
	struct mbuf *m;
{
	struct syncache *sc;
	struct syncache_head *sch;
	struct socket *so;

	sc = syncache_lookup(inc, &sch);
	if (sc == NULL) {
		/*
		 * There is no syncache entry, so see if this ACK is 
		 * a returning syncookie.  To do this, first:
		 *  A. See if this socket has had a syncache entry dropped in
		 *     the past.  We don't want to accept a bogus syncookie
 		 *     if we've never received a SYN.
		 *  B. check that the syncookie is valid.  If it is, then
		 *     cobble up a fake syncache entry, and return.
		 */
		if (!tcp_syncookies)
			return (0);
		sc = syncookie_lookup(inc, th, *sop);
		if (sc == NULL)
			return (0);
		sch = NULL;
		tcpstat.tcps_sc_recvcookie++;
	}

	/*
	 * If seg contains an ACK, but not for our SYN/ACK, send a RST.
	 */
	if (th->th_ack != sc->sc_iss + 1)
		return (0);

	so = syncache_socket(sc, *sop, m);
	if (so == NULL) {
#if 0
resetandabort:
		/* XXXjlemon check this - is this correct? */
		(void) tcp_respond(NULL, m, m, th,
		    th->th_seq + tlen, (tcp_seq)0, TH_RST|TH_ACK);
#endif
		m_freem(m);			/* XXX only needed for above */
		tcpstat.tcps_sc_aborted++;
	} else {
		sc->sc_flags |= SCF_KEEPROUTE;
		tcpstat.tcps_sc_completed++;
	}
	if (sch == NULL)
		syncache_free(sc);
	else
		syncache_drop(sc, sch);
	*sop = so;
	return (1);
}

/*
 * Given a LISTEN socket and an inbound SYN request, add
 * this to the syn cache, and send back a segment:
 *	<SEQ=ISS><ACK=RCV_NXT><CTL=SYN,ACK>
 * to the source.
 *
 * IMPORTANT NOTE: We do _NOT_ ACK data that might accompany the SYN.
 * Doing so would require that we hold onto the data and deliver it
 * to the application.  However, if we are the target of a SYN-flood
 * DoS attack, an attacker could send data which would eventually
 * consume all available buffer space if it were ACKed.  By not ACKing
 * the data, we avoid this DoS scenario.
 */
int
syncache_add(inc, to, th, sop, m)
	struct in_conninfo *inc;
	struct tcpopt *to;
	struct tcphdr *th;
	struct socket **sop;
	struct mbuf *m;
{
	struct tcpcb *tp;
	struct socket *so;
	struct syncache *sc = NULL;
	struct syncache_head *sch;
	struct mbuf *ipopts = NULL;
	struct rmxp_tao *taop;
	int i, s, win;

	so = *sop;
	tp = sototcpcb(so);

	/*
	 * Remember the IP options, if any.
	 */
#ifdef INET6
	if (!inc->inc_isipv6)
#endif
		ipopts = ip_srcroute();

	/*
	 * See if we already have an entry for this connection.
	 * If we do, resend the SYN,ACK, and reset the retransmit timer.
	 *
	 * XXX
	 * should the syncache be re-initialized with the contents
	 * of the new SYN here (which may have different options?)
	 */
	sc = syncache_lookup(inc, &sch);
	if (sc != NULL) {
		tcpstat.tcps_sc_dupsyn++;
		if (ipopts) {
			/*
			 * If we were remembering a previous source route,
			 * forget it and use the new one we've been given.
			 */
			if (sc->sc_ipopts)
				(void) m_free(sc->sc_ipopts);
			sc->sc_ipopts = ipopts;
		}
		/*
		 * Update timestamp if present.
		 */
		if (sc->sc_flags & SCF_TIMESTAMP)
			sc->sc_tsrecent = to->to_tsval;
		/*
		 * PCB may have changed, pick up new values.
		 */
		sc->sc_tp = tp;
		sc->sc_inp_gencnt = tp->t_inpcb->inp_gencnt;
		if (syncache_respond(sc, m) == 0) {
		        s = splnet();
			TAILQ_REMOVE(&tcp_syncache.timerq[sc->sc_rxtslot],
			    sc, sc_timerq);
			SYNCACHE_TIMEOUT(sc, sc->sc_rxtslot);
		        splx(s);
		 	tcpstat.tcps_sndacks++;
			tcpstat.tcps_sndtotal++;
		}
		*sop = NULL;
		return (1);
	}

	sc = uma_zalloc(tcp_syncache.zone, M_NOWAIT);
	if (sc == NULL) {
		/*
		 * The zone allocator couldn't provide more entries.
		 * Treat this as if the cache was full; drop the oldest 
		 * entry and insert the new one.
		 */
		s = splnet();
		for (i = SYNCACHE_MAXREXMTS; i >= 0; i--) {
			sc = TAILQ_FIRST(&tcp_syncache.timerq[i]);
			if (sc != NULL)
				break;
		}
		sc->sc_tp->ts_recent = ticks;
		syncache_drop(sc, NULL);
		splx(s);
		tcpstat.tcps_sc_zonefail++;
		sc = uma_zalloc(tcp_syncache.zone, M_NOWAIT);
		if (sc == NULL) {
			if (ipopts)
				(void) m_free(ipopts);
			return (0);
		}
	}

	/*
	 * Fill in the syncache values.
	 */
	bzero(sc, sizeof(*sc));
	sc->sc_tp = tp;
	sc->sc_inp_gencnt = tp->t_inpcb->inp_gencnt;
	sc->sc_ipopts = ipopts;
	sc->sc_inc.inc_fport = inc->inc_fport;
	sc->sc_inc.inc_lport = inc->inc_lport;
#ifdef INET6
	sc->sc_inc.inc_isipv6 = inc->inc_isipv6;
	if (inc->inc_isipv6) {
		sc->sc_inc.inc6_faddr = inc->inc6_faddr;
		sc->sc_inc.inc6_laddr = inc->inc6_laddr;
		sc->sc_route6.ro_rt = NULL;
	} else
#endif
	{
		sc->sc_inc.inc_faddr = inc->inc_faddr;
		sc->sc_inc.inc_laddr = inc->inc_laddr;
		sc->sc_route.ro_rt = NULL;
	}
	sc->sc_irs = th->th_seq;
	if (tcp_syncookies)
		sc->sc_iss = syncookie_generate(sc);
	else
		sc->sc_iss = arc4random();

	/* Initial receive window: clip sbspace to [0 .. TCP_MAXWIN] */
	win = sbspace(&so->so_rcv);
	win = imax(win, 0);
	win = imin(win, TCP_MAXWIN);
	sc->sc_wnd = win;

	sc->sc_flags = 0;
	sc->sc_peer_mss = to->to_flags & TOF_MSS ? to->to_mss : 0;
	if (tcp_do_rfc1323) {
		/*
		 * A timestamp received in a SYN makes
		 * it ok to send timestamp requests and replies.
		 */
		if (to->to_flags & TOF_TS) {
			sc->sc_tsrecent = to->to_tsval;
			sc->sc_flags |= SCF_TIMESTAMP;
		}
		if (to->to_flags & TOF_SCALE) {
			int wscale = 0;

			/* Compute proper scaling value from buffer space */
			while (wscale < TCP_MAX_WINSHIFT &&
			    (TCP_MAXWIN << wscale) < so->so_rcv.sb_hiwat)
				wscale++;
			sc->sc_request_r_scale = wscale;
			sc->sc_requested_s_scale = to->to_requested_s_scale;
			sc->sc_flags |= SCF_WINSCALE;
		}
	}
	if (tcp_do_rfc1644) {
		/*
		 * A CC or CC.new option received in a SYN makes
		 * it ok to send CC in subsequent segments.
		 */
		if (to->to_flags & (TOF_CC|TOF_CCNEW)) {
			sc->sc_cc_recv = to->to_cc;
			sc->sc_cc_send = CC_INC(tcp_ccgen);
			sc->sc_flags |= SCF_CC;
		}
	}
	if (tp->t_flags & TF_NOOPT)
		sc->sc_flags = SCF_NOOPT;

	/*
	 * XXX
	 * We have the option here of not doing TAO (even if the segment
	 * qualifies) and instead fall back to a normal 3WHS via the syncache.
	 * This allows us to apply synflood protection to TAO-qualifying SYNs
	 * also. However, there should be a hueristic to determine when to
	 * do this, and is not present at the moment.
	 */

	/*
	 * Perform TAO test on incoming CC (SEG.CC) option, if any.
	 * - compare SEG.CC against cached CC from the same host, if any.
	 * - if SEG.CC > chached value, SYN must be new and is accepted
	 *	immediately: save new CC in the cache, mark the socket
	 *	connected, enter ESTABLISHED state, turn on flag to
	 *	send a SYN in the next segment.
	 *	A virtual advertised window is set in rcv_adv to
	 *	initialize SWS prevention.  Then enter normal segment
	 *	processing: drop SYN, process data and FIN.
	 * - otherwise do a normal 3-way handshake.
	 */
	taop = tcp_gettaocache(&sc->sc_inc);
	if ((to->to_flags & TOF_CC) != 0) {
		if (((tp->t_flags & TF_NOPUSH) != 0) &&
		    sc->sc_flags & SCF_CC && 
		    taop != NULL && taop->tao_cc != 0 &&
		    CC_GT(to->to_cc, taop->tao_cc)) {
			sc->sc_rxtslot = 0;
			so = syncache_socket(sc, *sop, m);
			if (so != NULL) {
				sc->sc_flags |= SCF_KEEPROUTE;
				taop->tao_cc = to->to_cc;
				*sop = so;
			}
			syncache_free(sc);
			return (so != NULL);
		}
	} else {
		/*
		 * No CC option, but maybe CC.NEW: invalidate cached value.
		 */
		if (taop != NULL)
			taop->tao_cc = 0;
	}
	/*
	 * TAO test failed or there was no CC option,
	 *    do a standard 3-way handshake.
	 */
	if (syncache_respond(sc, m) == 0) {
		syncache_insert(sc, sch);
		tcpstat.tcps_sndacks++;
		tcpstat.tcps_sndtotal++;
	} else {
		syncache_free(sc);
		tcpstat.tcps_sc_dropped++;
	}
	*sop = NULL;
	return (1);
}

static int
syncache_respond(sc, m)
	struct syncache *sc;
	struct mbuf *m;
{
	u_int8_t *optp;
	int optlen, error;
	u_int16_t tlen, hlen, mssopt;
	struct ip *ip = NULL;
	struct rtentry *rt;
	struct tcphdr *th;
#ifdef INET6
	struct ip6_hdr *ip6 = NULL;
#endif

#ifdef INET6
	if (sc->sc_inc.inc_isipv6) {
		rt = tcp_rtlookup6(&sc->sc_inc);
		if (rt != NULL)
			mssopt = rt->rt_ifp->if_mtu -
			     (sizeof(struct ip6_hdr) + sizeof(struct tcphdr));
		else 
			mssopt = tcp_v6mssdflt;
		hlen = sizeof(struct ip6_hdr);
	} else
#endif
	{
		rt = tcp_rtlookup(&sc->sc_inc);
		if (rt != NULL)
			mssopt = rt->rt_ifp->if_mtu -
			     (sizeof(struct ip) + sizeof(struct tcphdr));
		else 
			mssopt = tcp_mssdflt;
		hlen = sizeof(struct ip);
	}

	/* Compute the size of the TCP options. */
	if (sc->sc_flags & SCF_NOOPT) {
		optlen = 0;
	} else {
		optlen = TCPOLEN_MAXSEG +
		    ((sc->sc_flags & SCF_WINSCALE) ? 4 : 0) +
		    ((sc->sc_flags & SCF_TIMESTAMP) ? TCPOLEN_TSTAMP_APPA : 0) +
		    ((sc->sc_flags & SCF_CC) ? TCPOLEN_CC_APPA * 2 : 0);
	}
	tlen = hlen + sizeof(struct tcphdr) + optlen;

	/*
	 * XXX
	 * assume that the entire packet will fit in a header mbuf
	 */
	KASSERT(max_linkhdr + tlen <= MHLEN, ("syncache: mbuf too small"));

	/*
	 * XXX shouldn't this reuse the mbuf if possible ?
	 * Create the IP+TCP header from scratch.
	 */
	if (m)
		m_freem(m);

	m = m_gethdr(M_DONTWAIT, MT_HEADER);
	if (m == NULL)
		return (ENOBUFS);
	m->m_data += max_linkhdr;
	m->m_len = tlen;
	m->m_pkthdr.len = tlen;
	m->m_pkthdr.rcvif = NULL;
#ifdef MAC
	mac_create_mbuf_from_socket(sc->sc_tp->t_inpcb->inp_socket, m);
#endif

#ifdef INET6
	if (sc->sc_inc.inc_isipv6) {
		ip6 = mtod(m, struct ip6_hdr *);
		ip6->ip6_vfc = IPV6_VERSION;
		ip6->ip6_nxt = IPPROTO_TCP;
		ip6->ip6_src = sc->sc_inc.inc6_laddr;
		ip6->ip6_dst = sc->sc_inc.inc6_faddr;
		ip6->ip6_plen = htons(tlen - hlen);
		/* ip6_hlim is set after checksum */
		/* ip6_flow = ??? */

		th = (struct tcphdr *)(ip6 + 1);
	} else
#endif
	{
		ip = mtod(m, struct ip *);
		ip->ip_v = IPVERSION;
		ip->ip_hl = sizeof(struct ip) >> 2;
		ip->ip_len = tlen;
		ip->ip_id = 0;
		ip->ip_off = 0;
		ip->ip_sum = 0;
		ip->ip_p = IPPROTO_TCP;
		ip->ip_src = sc->sc_inc.inc_laddr;
		ip->ip_dst = sc->sc_inc.inc_faddr;
		ip->ip_ttl = sc->sc_tp->t_inpcb->inp_ip_ttl;   /* XXX */
		ip->ip_tos = sc->sc_tp->t_inpcb->inp_ip_tos;   /* XXX */

		/*
		 * See if we should do MTU discovery.  Route lookups are expensive,
		 * so we will only unset the DF bit if:
		 *
		 *	1) path_mtu_discovery is disabled
		 *	2) the SCF_UNREACH flag has been set
		 */
		if (path_mtu_discovery
		    && ((sc->sc_flags & SCF_UNREACH) == 0)) {
		       ip->ip_off |= IP_DF;
		}

		th = (struct tcphdr *)(ip + 1);
	}
	th->th_sport = sc->sc_inc.inc_lport;
	th->th_dport = sc->sc_inc.inc_fport;

	th->th_seq = htonl(sc->sc_iss);
	th->th_ack = htonl(sc->sc_irs + 1);
	th->th_off = (sizeof(struct tcphdr) + optlen) >> 2;
	th->th_x2 = 0;
	th->th_flags = TH_SYN|TH_ACK;
	th->th_win = htons(sc->sc_wnd);
	th->th_urp = 0;

	/* Tack on the TCP options. */
	if (optlen == 0)
		goto no_options;
	optp = (u_int8_t *)(th + 1);
	*optp++ = TCPOPT_MAXSEG;
	*optp++ = TCPOLEN_MAXSEG;
	*optp++ = (mssopt >> 8) & 0xff;
	*optp++ = mssopt & 0xff;

	if (sc->sc_flags & SCF_WINSCALE) {
		*((u_int32_t *)optp) = htonl(TCPOPT_NOP << 24 |
		    TCPOPT_WINDOW << 16 | TCPOLEN_WINDOW << 8 |
		    sc->sc_request_r_scale);
		optp += 4;
	}

	if (sc->sc_flags & SCF_TIMESTAMP) {
		u_int32_t *lp = (u_int32_t *)(optp);

		/* Form timestamp option as shown in appendix A of RFC 1323. */
		*lp++ = htonl(TCPOPT_TSTAMP_HDR);
		*lp++ = htonl(ticks);
		*lp   = htonl(sc->sc_tsrecent);
		optp += TCPOLEN_TSTAMP_APPA;
	}

	/*
         * Send CC and CC.echo if we received CC from our peer.
         */
        if (sc->sc_flags & SCF_CC) {
		u_int32_t *lp = (u_int32_t *)(optp);

		*lp++ = htonl(TCPOPT_CC_HDR(TCPOPT_CC));
		*lp++ = htonl(sc->sc_cc_send);
		*lp++ = htonl(TCPOPT_CC_HDR(TCPOPT_CCECHO));
		*lp   = htonl(sc->sc_cc_recv);
		optp += TCPOLEN_CC_APPA * 2;
	}
no_options:

#ifdef INET6
	if (sc->sc_inc.inc_isipv6) {
		struct route_in6 *ro6 = &sc->sc_route6;

		th->th_sum = 0;
		th->th_sum = in6_cksum(m, IPPROTO_TCP, hlen, tlen - hlen);
		ip6->ip6_hlim = in6_selecthlim(NULL,
		    ro6->ro_rt ? ro6->ro_rt->rt_ifp : NULL);
		error = ip6_output(m, NULL, ro6, 0, NULL, NULL,
				sc->sc_tp->t_inpcb);
	} else
#endif
	{
        	th->th_sum = in_pseudo(ip->ip_src.s_addr, ip->ip_dst.s_addr,
		    htons(tlen - hlen + IPPROTO_TCP));
		m->m_pkthdr.csum_flags = CSUM_TCP;
		m->m_pkthdr.csum_data = offsetof(struct tcphdr, th_sum);
		error = ip_output(m, sc->sc_ipopts, &sc->sc_route, 0, NULL,
				sc->sc_tp->t_inpcb);
	}
	return (error);
}

/*
 * cookie layers:
 *
 *	|. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .|
 *	| peer iss                                                      |
 *	| MD5(laddr,faddr,lport,fport,secret)             |. . . . . . .|
 *	|                     0                       |(A)|             |
 * (A): peer mss index
 */

/*
 * The values below are chosen to minimize the size of the tcp_secret
 * table, as well as providing roughly a 16 second lifetime for the cookie.
 */

#define SYNCOOKIE_WNDBITS	5	/* exposed bits for window indexing */
#define SYNCOOKIE_TIMESHIFT	1	/* scale ticks to window time units */

#define SYNCOOKIE_WNDMASK	((1 << SYNCOOKIE_WNDBITS) - 1)
#define SYNCOOKIE_NSECRETS	(1 << SYNCOOKIE_WNDBITS)
#define SYNCOOKIE_TIMEOUT \
    (hz * (1 << SYNCOOKIE_WNDBITS) / (1 << SYNCOOKIE_TIMESHIFT))
#define SYNCOOKIE_DATAMASK 	((3 << SYNCOOKIE_WNDBITS) | SYNCOOKIE_WNDMASK)

static struct {
	u_int32_t	ts_secbits[4];
	u_int		ts_expire;
} tcp_secret[SYNCOOKIE_NSECRETS];

static int tcp_msstab[] = { 0, 536, 1460, 8960 };

static MD5_CTX syn_ctx;

#define MD5Add(v)	MD5Update(&syn_ctx, (u_char *)&v, sizeof(v))

struct md5_add {
	u_int32_t laddr, faddr;
	u_int32_t secbits[4];
	u_int16_t lport, fport;
};

#ifdef CTASSERT
CTASSERT(sizeof(struct md5_add) == 28);
#endif

/*
 * Consider the problem of a recreated (and retransmitted) cookie.  If the
 * original SYN was accepted, the connection is established.  The second 
 * SYN is inflight, and if it arrives with an ISN that falls within the 
 * receive window, the connection is killed.  
 *
 * However, since cookies have other problems, this may not be worth
 * worrying about.
 */

static u_int32_t
syncookie_generate(struct syncache *sc)
{
	u_int32_t md5_buffer[4];
	u_int32_t data;
	int idx, i;
	struct md5_add add;

	idx = ((ticks << SYNCOOKIE_TIMESHIFT) / hz) & SYNCOOKIE_WNDMASK;
	if (tcp_secret[idx].ts_expire < ticks) {
		for (i = 0; i < 4; i++)
			tcp_secret[idx].ts_secbits[i] = arc4random();
		tcp_secret[idx].ts_expire = ticks + SYNCOOKIE_TIMEOUT;
	}
	for (data = sizeof(tcp_msstab) / sizeof(int) - 1; data > 0; data--)
		if (tcp_msstab[data] <= sc->sc_peer_mss)
			break;
	data = (data << SYNCOOKIE_WNDBITS) | idx;
	data ^= sc->sc_irs;				/* peer's iss */
	MD5Init(&syn_ctx);
#ifdef INET6
	if (sc->sc_inc.inc_isipv6) {
		MD5Add(sc->sc_inc.inc6_laddr);
		MD5Add(sc->sc_inc.inc6_faddr);
		add.laddr = 0;
		add.faddr = 0;
	} else
#endif
	{
		add.laddr = sc->sc_inc.inc_laddr.s_addr;
		add.faddr = sc->sc_inc.inc_faddr.s_addr;
	}
	add.lport = sc->sc_inc.inc_lport;
	add.fport = sc->sc_inc.inc_fport;
	add.secbits[0] = tcp_secret[idx].ts_secbits[0];
	add.secbits[1] = tcp_secret[idx].ts_secbits[1];
	add.secbits[2] = tcp_secret[idx].ts_secbits[2];
	add.secbits[3] = tcp_secret[idx].ts_secbits[3];
	MD5Add(add);
	MD5Final((u_char *)&md5_buffer, &syn_ctx);
	data ^= (md5_buffer[0] & ~SYNCOOKIE_WNDMASK);
	return (data);
}

static struct syncache *
syncookie_lookup(inc, th, so)
	struct in_conninfo *inc;
	struct tcphdr *th;
	struct socket *so;
{
	u_int32_t md5_buffer[4];
	struct syncache *sc;
	u_int32_t data;
	int wnd, idx;
	struct md5_add add;

	data = (th->th_ack - 1) ^ (th->th_seq - 1);	/* remove ISS */
	idx = data & SYNCOOKIE_WNDMASK;
	if (tcp_secret[idx].ts_expire < ticks ||
	    sototcpcb(so)->ts_recent + SYNCOOKIE_TIMEOUT < ticks)
		return (NULL);
	MD5Init(&syn_ctx);
#ifdef INET6
	if (inc->inc_isipv6) {
		MD5Add(inc->inc6_laddr);
		MD5Add(inc->inc6_faddr);
		add.laddr = 0;
		add.faddr = 0;
	} else
#endif
	{
		add.laddr = inc->inc_laddr.s_addr;
		add.faddr = inc->inc_faddr.s_addr;
	}
	add.lport = inc->inc_lport;
	add.fport = inc->inc_fport;
	add.secbits[0] = tcp_secret[idx].ts_secbits[0];
	add.secbits[1] = tcp_secret[idx].ts_secbits[1];
	add.secbits[2] = tcp_secret[idx].ts_secbits[2];
	add.secbits[3] = tcp_secret[idx].ts_secbits[3];
	MD5Add(add);
	MD5Final((u_char *)&md5_buffer, &syn_ctx);
	data ^= md5_buffer[0];
	if ((data & ~SYNCOOKIE_DATAMASK) != 0)
		return (NULL);
	data = data >> SYNCOOKIE_WNDBITS;

	sc = uma_zalloc(tcp_syncache.zone, M_NOWAIT);
	if (sc == NULL)
		return (NULL);
	/*
	 * Fill in the syncache values.
	 * XXX duplicate code from syncache_add
	 */
	sc->sc_ipopts = NULL;
	sc->sc_inc.inc_fport = inc->inc_fport;
	sc->sc_inc.inc_lport = inc->inc_lport;
#ifdef INET6
	sc->sc_inc.inc_isipv6 = inc->inc_isipv6;
	if (inc->inc_isipv6) {
		sc->sc_inc.inc6_faddr = inc->inc6_faddr;
		sc->sc_inc.inc6_laddr = inc->inc6_laddr;
		sc->sc_route6.ro_rt = NULL;
	} else
#endif
	{
		sc->sc_inc.inc_faddr = inc->inc_faddr;
		sc->sc_inc.inc_laddr = inc->inc_laddr;
		sc->sc_route.ro_rt = NULL;
	}
	sc->sc_irs = th->th_seq - 1;
	sc->sc_iss = th->th_ack - 1;
	wnd = sbspace(&so->so_rcv);
	wnd = imax(wnd, 0);
	wnd = imin(wnd, TCP_MAXWIN);
	sc->sc_wnd = wnd;
	sc->sc_flags = 0;
	sc->sc_rxtslot = 0;
	sc->sc_peer_mss = tcp_msstab[data];
	return (sc);
}