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
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
|
/*-
* Copyright (c) 1982, 1986, 1988, 1990, 1993, 1995
* The Regents of the University of California. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that 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.
* 4. Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS 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 REGENTS 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.
*
* @(#)tcp_subr.c 8.2 (Berkeley) 5/24/95
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include "opt_compat.h"
#include "opt_inet.h"
#include "opt_inet6.h"
#include "opt_ipsec.h"
#include "opt_mac.h"
#include "opt_tcpdebug.h"
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/callout.h>
#include <sys/kernel.h>
#include <sys/sysctl.h>
#include <sys/malloc.h>
#include <sys/mbuf.h>
#ifdef INET6
#include <sys/domain.h>
#endif
#include <sys/priv.h>
#include <sys/proc.h>
#include <sys/socket.h>
#include <sys/socketvar.h>
#include <sys/protosw.h>
#include <sys/random.h>
#include <sys/vimage.h>
#include <vm/uma.h>
#include <net/route.h>
#include <net/if.h>
#include <netinet/in.h>
#include <netinet/in_systm.h>
#include <netinet/ip.h>
#ifdef INET6
#include <netinet/ip6.h>
#endif
#include <netinet/in_pcb.h>
#ifdef INET6
#include <netinet6/in6_pcb.h>
#endif
#include <netinet/in_var.h>
#include <netinet/ip_var.h>
#ifdef INET6
#include <netinet6/ip6_var.h>
#include <netinet6/scope6_var.h>
#include <netinet6/nd6.h>
#endif
#include <netinet/ip_icmp.h>
#include <netinet/tcp.h>
#include <netinet/tcp_fsm.h>
#include <netinet/tcp_seq.h>
#include <netinet/tcp_timer.h>
#include <netinet/tcp_var.h>
#include <netinet/tcp_syncache.h>
#include <netinet/tcp_offload.h>
#ifdef INET6
#include <netinet6/tcp6_var.h>
#endif
#include <netinet/tcpip.h>
#ifdef TCPDEBUG
#include <netinet/tcp_debug.h>
#endif
#include <netinet6/ip6protosw.h>
#ifdef IPSEC
#include <netipsec/ipsec.h>
#include <netipsec/xform.h>
#ifdef INET6
#include <netipsec/ipsec6.h>
#endif
#include <netipsec/key.h>
#endif /*IPSEC*/
#include <machine/in_cksum.h>
#include <sys/md5.h>
#include <security/mac/mac_framework.h>
int tcp_mssdflt = TCP_MSS;
SYSCTL_INT(_net_inet_tcp, TCPCTL_MSSDFLT, mssdflt, CTLFLAG_RW,
&tcp_mssdflt, 0, "Default TCP Maximum Segment Size");
#ifdef INET6
int tcp_v6mssdflt = TCP6_MSS;
SYSCTL_INT(_net_inet_tcp, TCPCTL_V6MSSDFLT, v6mssdflt,
CTLFLAG_RW, &tcp_v6mssdflt , 0,
"Default TCP Maximum Segment Size for IPv6");
#endif
/*
* Minimum MSS we accept and use. This prevents DoS attacks where
* we are forced to a ridiculous low MSS like 20 and send hundreds
* of packets instead of one. The effect scales with the available
* bandwidth and quickly saturates the CPU and network interface
* with packet generation and sending. Set to zero to disable MINMSS
* checking. This setting prevents us from sending too small packets.
*/
int tcp_minmss = TCP_MINMSS;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, minmss, CTLFLAG_RW,
&tcp_minmss , 0, "Minmum TCP Maximum Segment Size");
int tcp_do_rfc1323 = 1;
SYSCTL_INT(_net_inet_tcp, TCPCTL_DO_RFC1323, rfc1323, CTLFLAG_RW,
&tcp_do_rfc1323, 0, "Enable rfc1323 (high performance TCP) extensions");
static int tcp_log_debug = 0;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, log_debug, CTLFLAG_RW,
&tcp_log_debug, 0, "Log errors caused by incoming TCP segments");
static int tcp_tcbhashsize = 0;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, tcbhashsize, CTLFLAG_RDTUN,
&tcp_tcbhashsize, 0, "Size of TCP control-block hashtable");
static int do_tcpdrain = 1;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, do_tcpdrain, CTLFLAG_RW,
&do_tcpdrain, 0,
"Enable tcp_drain routine for extra help when low on mbufs");
SYSCTL_INT(_net_inet_tcp, OID_AUTO, pcbcount, CTLFLAG_RD,
&tcbinfo.ipi_count, 0, "Number of active PCBs");
static int icmp_may_rst = 1;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, icmp_may_rst, CTLFLAG_RW,
&icmp_may_rst, 0,
"Certain ICMP unreachable messages may abort connections in SYN_SENT");
static int tcp_isn_reseed_interval = 0;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, isn_reseed_interval, CTLFLAG_RW,
&tcp_isn_reseed_interval, 0, "Seconds between reseeding of ISN secret");
/*
* TCP bandwidth limiting sysctls. Note that the default lower bound of
* 1024 exists only for debugging. A good production default would be
* something like 6100.
*/
SYSCTL_NODE(_net_inet_tcp, OID_AUTO, inflight, CTLFLAG_RW, 0,
"TCP inflight data limiting");
static int tcp_inflight_enable = 1;
SYSCTL_INT(_net_inet_tcp_inflight, OID_AUTO, enable, CTLFLAG_RW,
&tcp_inflight_enable, 0, "Enable automatic TCP inflight data limiting");
static int tcp_inflight_debug = 0;
SYSCTL_INT(_net_inet_tcp_inflight, OID_AUTO, debug, CTLFLAG_RW,
&tcp_inflight_debug, 0, "Debug TCP inflight calculations");
static int tcp_inflight_rttthresh;
SYSCTL_PROC(_net_inet_tcp_inflight, OID_AUTO, rttthresh, CTLTYPE_INT|CTLFLAG_RW,
&tcp_inflight_rttthresh, 0, sysctl_msec_to_ticks, "I",
"RTT threshold below which inflight will deactivate itself");
static int tcp_inflight_min = 6144;
SYSCTL_INT(_net_inet_tcp_inflight, OID_AUTO, min, CTLFLAG_RW,
&tcp_inflight_min, 0, "Lower-bound for TCP inflight window");
static int tcp_inflight_max = TCP_MAXWIN << TCP_MAX_WINSHIFT;
SYSCTL_INT(_net_inet_tcp_inflight, OID_AUTO, max, CTLFLAG_RW,
&tcp_inflight_max, 0, "Upper-bound for TCP inflight window");
static int tcp_inflight_stab = 20;
SYSCTL_INT(_net_inet_tcp_inflight, OID_AUTO, stab, CTLFLAG_RW,
&tcp_inflight_stab, 0, "Inflight Algorithm Stabilization 20 = 2 packets");
uma_zone_t sack_hole_zone;
static struct inpcb *tcp_notify(struct inpcb *, int);
static void tcp_isn_tick(void *);
/*
* Target size of TCP PCB hash tables. Must be a power of two.
*
* Note that this can be overridden by the kernel environment
* variable net.inet.tcp.tcbhashsize
*/
#ifndef TCBHASHSIZE
#define TCBHASHSIZE 512
#endif
/*
* XXX
* Callouts should be moved into struct tcp directly. They are currently
* separate because the tcpcb structure is exported to userland for sysctl
* parsing purposes, which do not know about callouts.
*/
struct tcpcb_mem {
struct tcpcb tcb;
struct tcp_timer tt;
};
static uma_zone_t tcpcb_zone;
MALLOC_DEFINE(M_TCPLOG, "tcplog", "TCP address and flags print buffers");
struct callout isn_callout;
static struct mtx isn_mtx;
#define ISN_LOCK_INIT() mtx_init(&isn_mtx, "isn_mtx", NULL, MTX_DEF)
#define ISN_LOCK() mtx_lock(&isn_mtx)
#define ISN_UNLOCK() mtx_unlock(&isn_mtx)
/*
* TCP initialization.
*/
static void
tcp_zone_change(void *tag)
{
uma_zone_set_max(V_tcbinfo.ipi_zone, maxsockets);
uma_zone_set_max(tcpcb_zone, maxsockets);
tcp_tw_zone_change();
}
static int
tcp_inpcb_init(void *mem, int size, int flags)
{
struct inpcb *inp = mem;
INP_LOCK_INIT(inp, "inp", "tcpinp");
return (0);
}
void
tcp_init(void)
{
int hashsize = TCBHASHSIZE;
tcp_delacktime = TCPTV_DELACK;
tcp_keepinit = TCPTV_KEEP_INIT;
tcp_keepidle = TCPTV_KEEP_IDLE;
tcp_keepintvl = TCPTV_KEEPINTVL;
tcp_maxpersistidle = TCPTV_KEEP_IDLE;
tcp_msl = TCPTV_MSL;
tcp_rexmit_min = TCPTV_MIN;
if (tcp_rexmit_min < 1)
tcp_rexmit_min = 1;
tcp_rexmit_slop = TCPTV_CPU_VAR;
V_tcp_inflight_rttthresh = TCPTV_INFLIGHT_RTTTHRESH;
tcp_finwait2_timeout = TCPTV_FINWAIT2_TIMEOUT;
INP_INFO_LOCK_INIT(&V_tcbinfo, "tcp");
LIST_INIT(&V_tcb);
V_tcbinfo.ipi_listhead = &V_tcb;
TUNABLE_INT_FETCH("net.inet.tcp.tcbhashsize", &hashsize);
if (!powerof2(hashsize)) {
printf("WARNING: TCB hash size not a power of 2\n");
hashsize = 512; /* safe default */
}
tcp_tcbhashsize = hashsize;
V_tcbinfo.ipi_hashbase = hashinit(hashsize, M_PCB,
&V_tcbinfo.ipi_hashmask);
V_tcbinfo.ipi_porthashbase = hashinit(hashsize, M_PCB,
&V_tcbinfo.ipi_porthashmask);
V_tcbinfo.ipi_zone = uma_zcreate("inpcb", sizeof(struct inpcb),
NULL, NULL, tcp_inpcb_init, NULL, UMA_ALIGN_PTR, UMA_ZONE_NOFREE);
uma_zone_set_max(V_tcbinfo.ipi_zone, maxsockets);
#ifdef INET6
#define TCP_MINPROTOHDR (sizeof(struct ip6_hdr) + sizeof(struct tcphdr))
#else /* INET6 */
#define TCP_MINPROTOHDR (sizeof(struct tcpiphdr))
#endif /* INET6 */
if (max_protohdr < TCP_MINPROTOHDR)
max_protohdr = TCP_MINPROTOHDR;
if (max_linkhdr + TCP_MINPROTOHDR > MHLEN)
panic("tcp_init");
#undef TCP_MINPROTOHDR
/*
* These have to be type stable for the benefit of the timers.
*/
tcpcb_zone = uma_zcreate("tcpcb", sizeof(struct tcpcb_mem),
NULL, NULL, NULL, NULL, UMA_ALIGN_PTR, UMA_ZONE_NOFREE);
uma_zone_set_max(tcpcb_zone, maxsockets);
tcp_tw_init();
syncache_init();
tcp_hc_init();
tcp_reass_init();
ISN_LOCK_INIT();
callout_init(&isn_callout, CALLOUT_MPSAFE);
tcp_isn_tick(NULL);
EVENTHANDLER_REGISTER(shutdown_pre_sync, tcp_fini, NULL,
SHUTDOWN_PRI_DEFAULT);
sack_hole_zone = uma_zcreate("sackhole", sizeof(struct sackhole),
NULL, NULL, NULL, NULL, UMA_ALIGN_PTR, UMA_ZONE_NOFREE);
EVENTHANDLER_REGISTER(maxsockets_change, tcp_zone_change, NULL,
EVENTHANDLER_PRI_ANY);
}
void
tcp_fini(void *xtp)
{
callout_stop(&isn_callout);
}
/*
* Fill in the IP and TCP headers for an outgoing packet, given the tcpcb.
* tcp_template used to store this data in mbufs, but we now recopy it out
* of the tcpcb each time to conserve mbufs.
*/
void
tcpip_fillheaders(struct inpcb *inp, void *ip_ptr, void *tcp_ptr)
{
struct tcphdr *th = (struct tcphdr *)tcp_ptr;
INP_WLOCK_ASSERT(inp);
#ifdef INET6
if ((inp->inp_vflag & INP_IPV6) != 0) {
struct ip6_hdr *ip6;
ip6 = (struct ip6_hdr *)ip_ptr;
ip6->ip6_flow = (ip6->ip6_flow & ~IPV6_FLOWINFO_MASK) |
(inp->in6p_flowinfo & IPV6_FLOWINFO_MASK);
ip6->ip6_vfc = (ip6->ip6_vfc & ~IPV6_VERSION_MASK) |
(IPV6_VERSION & IPV6_VERSION_MASK);
ip6->ip6_nxt = IPPROTO_TCP;
ip6->ip6_plen = sizeof(struct tcphdr);
ip6->ip6_src = inp->in6p_laddr;
ip6->ip6_dst = inp->in6p_faddr;
} else
#endif
{
struct ip *ip;
ip = (struct ip *)ip_ptr;
ip->ip_v = IPVERSION;
ip->ip_hl = 5;
ip->ip_tos = inp->inp_ip_tos;
ip->ip_len = 0;
ip->ip_id = 0;
ip->ip_off = 0;
ip->ip_ttl = inp->inp_ip_ttl;
ip->ip_sum = 0;
ip->ip_p = IPPROTO_TCP;
ip->ip_src = inp->inp_laddr;
ip->ip_dst = inp->inp_faddr;
}
th->th_sport = inp->inp_lport;
th->th_dport = inp->inp_fport;
th->th_seq = 0;
th->th_ack = 0;
th->th_x2 = 0;
th->th_off = 5;
th->th_flags = 0;
th->th_win = 0;
th->th_urp = 0;
th->th_sum = 0; /* in_pseudo() is called later for ipv4 */
}
/*
* Create template to be used to send tcp packets on a connection.
* Allocates an mbuf and fills in a skeletal tcp/ip header. The only
* use for this function is in keepalives, which use tcp_respond.
*/
struct tcptemp *
tcpip_maketemplate(struct inpcb *inp)
{
struct tcptemp *t;
t = malloc(sizeof(*t), M_TEMP, M_NOWAIT);
if (t == NULL)
return (NULL);
tcpip_fillheaders(inp, (void *)&t->tt_ipgen, (void *)&t->tt_t);
return (t);
}
/*
* Send a single message to the TCP at address specified by
* the given TCP/IP header. If m == NULL, then we make a copy
* of the tcpiphdr at ti and send directly to the addressed host.
* This is used to force keep alive messages out using the TCP
* template for a connection. If flags are given then we send
* a message back to the TCP which originated the * segment ti,
* and discard the mbuf containing it and any other attached mbufs.
*
* In any case the ack and sequence number of the transmitted
* segment are as specified by the parameters.
*
* NOTE: If m != NULL, then ti must point to *inside* the mbuf.
*/
void
tcp_respond(struct tcpcb *tp, void *ipgen, struct tcphdr *th, struct mbuf *m,
tcp_seq ack, tcp_seq seq, int flags)
{
int tlen;
int win = 0;
struct ip *ip;
struct tcphdr *nth;
#ifdef INET6
struct ip6_hdr *ip6;
int isipv6;
#endif /* INET6 */
int ipflags = 0;
struct inpcb *inp;
KASSERT(tp != NULL || m != NULL, ("tcp_respond: tp and m both NULL"));
#ifdef INET6
isipv6 = ((struct ip *)ipgen)->ip_v == 6;
ip6 = ipgen;
#endif /* INET6 */
ip = ipgen;
if (tp != NULL) {
inp = tp->t_inpcb;
KASSERT(inp != NULL, ("tcp control block w/o inpcb"));
INP_WLOCK_ASSERT(inp);
} else
inp = NULL;
if (tp != NULL) {
if (!(flags & TH_RST)) {
win = sbspace(&inp->inp_socket->so_rcv);
if (win > (long)TCP_MAXWIN << tp->rcv_scale)
win = (long)TCP_MAXWIN << tp->rcv_scale;
}
}
if (m == NULL) {
m = m_gethdr(M_DONTWAIT, MT_DATA);
if (m == NULL)
return;
tlen = 0;
m->m_data += max_linkhdr;
#ifdef INET6
if (isipv6) {
bcopy((caddr_t)ip6, mtod(m, caddr_t),
sizeof(struct ip6_hdr));
ip6 = mtod(m, struct ip6_hdr *);
nth = (struct tcphdr *)(ip6 + 1);
} else
#endif /* INET6 */
{
bcopy((caddr_t)ip, mtod(m, caddr_t), sizeof(struct ip));
ip = mtod(m, struct ip *);
nth = (struct tcphdr *)(ip + 1);
}
bcopy((caddr_t)th, (caddr_t)nth, sizeof(struct tcphdr));
flags = TH_ACK;
} else {
/*
* reuse the mbuf.
* XXX MRT We inherrit the FIB, which is lucky.
*/
m_freem(m->m_next);
m->m_next = NULL;
m->m_data = (caddr_t)ipgen;
/* m_len is set later */
tlen = 0;
#define xchg(a,b,type) { type t; t=a; a=b; b=t; }
#ifdef INET6
if (isipv6) {
xchg(ip6->ip6_dst, ip6->ip6_src, struct in6_addr);
nth = (struct tcphdr *)(ip6 + 1);
} else
#endif /* INET6 */
{
xchg(ip->ip_dst.s_addr, ip->ip_src.s_addr, n_long);
nth = (struct tcphdr *)(ip + 1);
}
if (th != nth) {
/*
* this is usually a case when an extension header
* exists between the IPv6 header and the
* TCP header.
*/
nth->th_sport = th->th_sport;
nth->th_dport = th->th_dport;
}
xchg(nth->th_dport, nth->th_sport, n_short);
#undef xchg
}
#ifdef INET6
if (isipv6) {
ip6->ip6_flow = 0;
ip6->ip6_vfc = IPV6_VERSION;
ip6->ip6_nxt = IPPROTO_TCP;
ip6->ip6_plen = htons((u_short)(sizeof (struct tcphdr) +
tlen));
tlen += sizeof (struct ip6_hdr) + sizeof (struct tcphdr);
} else
#endif
{
tlen += sizeof (struct tcpiphdr);
ip->ip_len = tlen;
ip->ip_ttl = V_ip_defttl;
if (V_path_mtu_discovery)
ip->ip_off |= IP_DF;
}
m->m_len = tlen;
m->m_pkthdr.len = tlen;
m->m_pkthdr.rcvif = NULL;
#ifdef MAC
if (inp != NULL) {
/*
* Packet is associated with a socket, so allow the
* label of the response to reflect the socket label.
*/
INP_WLOCK_ASSERT(inp);
mac_inpcb_create_mbuf(inp, m);
} else {
/*
* Packet is not associated with a socket, so possibly
* update the label in place.
*/
mac_netinet_tcp_reply(m);
}
#endif
nth->th_seq = htonl(seq);
nth->th_ack = htonl(ack);
nth->th_x2 = 0;
nth->th_off = sizeof (struct tcphdr) >> 2;
nth->th_flags = flags;
if (tp != NULL)
nth->th_win = htons((u_short) (win >> tp->rcv_scale));
else
nth->th_win = htons((u_short)win);
nth->th_urp = 0;
#ifdef INET6
if (isipv6) {
nth->th_sum = 0;
nth->th_sum = in6_cksum(m, IPPROTO_TCP,
sizeof(struct ip6_hdr),
tlen - sizeof(struct ip6_hdr));
ip6->ip6_hlim = in6_selecthlim(tp != NULL ? tp->t_inpcb :
NULL, NULL);
} else
#endif /* INET6 */
{
nth->th_sum = in_pseudo(ip->ip_src.s_addr, ip->ip_dst.s_addr,
htons((u_short)(tlen - sizeof(struct ip) + ip->ip_p)));
m->m_pkthdr.csum_flags = CSUM_TCP;
m->m_pkthdr.csum_data = offsetof(struct tcphdr, th_sum);
}
#ifdef TCPDEBUG
if (tp == NULL || (inp->inp_socket->so_options & SO_DEBUG))
tcp_trace(TA_OUTPUT, 0, tp, mtod(m, void *), th, 0);
#endif
#ifdef INET6
if (isipv6)
(void) ip6_output(m, NULL, NULL, ipflags, NULL, NULL, inp);
else
#endif /* INET6 */
(void) ip_output(m, NULL, NULL, ipflags, NULL, inp);
}
/*
* Create a new TCP control block, making an
* empty reassembly queue and hooking it to the argument
* protocol control block. The `inp' parameter must have
* come from the zone allocator set up in tcp_init().
*/
struct tcpcb *
tcp_newtcpcb(struct inpcb *inp)
{
struct tcpcb_mem *tm;
struct tcpcb *tp;
#ifdef INET6
int isipv6 = (inp->inp_vflag & INP_IPV6) != 0;
#endif /* INET6 */
tm = uma_zalloc(tcpcb_zone, M_NOWAIT | M_ZERO);
if (tm == NULL)
return (NULL);
tp = &tm->tcb;
tp->t_timers = &tm->tt;
/* LIST_INIT(&tp->t_segq); */ /* XXX covered by M_ZERO */
tp->t_maxseg = tp->t_maxopd =
#ifdef INET6
isipv6 ? V_tcp_v6mssdflt :
#endif /* INET6 */
V_tcp_mssdflt;
/* Set up our timeouts. */
callout_init(&tp->t_timers->tt_rexmt, CALLOUT_MPSAFE);
callout_init(&tp->t_timers->tt_persist, CALLOUT_MPSAFE);
callout_init(&tp->t_timers->tt_keep, CALLOUT_MPSAFE);
callout_init(&tp->t_timers->tt_2msl, CALLOUT_MPSAFE);
callout_init(&tp->t_timers->tt_delack, CALLOUT_MPSAFE);
if (V_tcp_do_rfc1323)
tp->t_flags = (TF_REQ_SCALE|TF_REQ_TSTMP);
if (V_tcp_do_sack)
tp->t_flags |= TF_SACK_PERMIT;
TAILQ_INIT(&tp->snd_holes);
tp->t_inpcb = inp; /* XXX */
/*
* Init srtt to TCPTV_SRTTBASE (0), so we can tell that we have no
* rtt estimate. Set rttvar so that srtt + 4 * rttvar gives
* reasonable initial retransmit time.
*/
tp->t_srtt = TCPTV_SRTTBASE;
tp->t_rttvar = ((TCPTV_RTOBASE - TCPTV_SRTTBASE) << TCP_RTTVAR_SHIFT) / 4;
tp->t_rttmin = tcp_rexmit_min;
tp->t_rxtcur = TCPTV_RTOBASE;
tp->snd_cwnd = TCP_MAXWIN << TCP_MAX_WINSHIFT;
tp->snd_bwnd = TCP_MAXWIN << TCP_MAX_WINSHIFT;
tp->snd_ssthresh = TCP_MAXWIN << TCP_MAX_WINSHIFT;
tp->t_rcvtime = ticks;
tp->t_bw_rtttime = ticks;
/*
* IPv4 TTL initialization is necessary for an IPv6 socket as well,
* because the socket may be bound to an IPv6 wildcard address,
* which may match an IPv4-mapped IPv6 address.
*/
inp->inp_ip_ttl = V_ip_defttl;
inp->inp_ppcb = tp;
return (tp); /* XXX */
}
/*
* Drop a TCP connection, reporting
* the specified error. If connection is synchronized,
* then send a RST to peer.
*/
struct tcpcb *
tcp_drop(struct tcpcb *tp, int errno)
{
struct socket *so = tp->t_inpcb->inp_socket;
INP_INFO_WLOCK_ASSERT(&V_tcbinfo);
INP_WLOCK_ASSERT(tp->t_inpcb);
if (TCPS_HAVERCVDSYN(tp->t_state)) {
tp->t_state = TCPS_CLOSED;
(void) tcp_output_reset(tp);
V_tcpstat.tcps_drops++;
} else
V_tcpstat.tcps_conndrops++;
if (errno == ETIMEDOUT && tp->t_softerror)
errno = tp->t_softerror;
so->so_error = errno;
return (tcp_close(tp));
}
void
tcp_discardcb(struct tcpcb *tp)
{
struct tseg_qent *q;
struct inpcb *inp = tp->t_inpcb;
struct socket *so = inp->inp_socket;
#ifdef INET6
int isipv6 = (inp->inp_vflag & INP_IPV6) != 0;
#endif /* INET6 */
INP_WLOCK_ASSERT(inp);
/*
* Make sure that all of our timers are stopped before we
* delete the PCB.
*/
callout_stop(&tp->t_timers->tt_rexmt);
callout_stop(&tp->t_timers->tt_persist);
callout_stop(&tp->t_timers->tt_keep);
callout_stop(&tp->t_timers->tt_2msl);
callout_stop(&tp->t_timers->tt_delack);
/*
* If we got enough samples through the srtt filter,
* save the rtt and rttvar in the routing entry.
* 'Enough' is arbitrarily defined as 4 rtt samples.
* 4 samples is enough for the srtt filter to converge
* to within enough % of the correct value; fewer samples
* and we could save a bogus rtt. The danger is not high
* as tcp quickly recovers from everything.
* XXX: Works very well but needs some more statistics!
*/
if (tp->t_rttupdated >= 4) {
struct hc_metrics_lite metrics;
u_long ssthresh;
bzero(&metrics, sizeof(metrics));
/*
* Update the ssthresh always when the conditions below
* are satisfied. This gives us better new start value
* for the congestion avoidance for new connections.
* ssthresh is only set if packet loss occured on a session.
*
* XXXRW: 'so' may be NULL here, and/or socket buffer may be
* being torn down. Ideally this code would not use 'so'.
*/
ssthresh = tp->snd_ssthresh;
if (ssthresh != 0 && ssthresh < so->so_snd.sb_hiwat / 2) {
/*
* convert the limit from user data bytes to
* packets then to packet data bytes.
*/
ssthresh = (ssthresh + tp->t_maxseg / 2) / tp->t_maxseg;
if (ssthresh < 2)
ssthresh = 2;
ssthresh *= (u_long)(tp->t_maxseg +
#ifdef INET6
(isipv6 ? sizeof (struct ip6_hdr) +
sizeof (struct tcphdr) :
#endif
sizeof (struct tcpiphdr)
#ifdef INET6
)
#endif
);
} else
ssthresh = 0;
metrics.rmx_ssthresh = ssthresh;
metrics.rmx_rtt = tp->t_srtt;
metrics.rmx_rttvar = tp->t_rttvar;
/* XXX: This wraps if the pipe is more than 4 Gbit per second */
metrics.rmx_bandwidth = tp->snd_bandwidth;
metrics.rmx_cwnd = tp->snd_cwnd;
metrics.rmx_sendpipe = 0;
metrics.rmx_recvpipe = 0;
tcp_hc_update(&inp->inp_inc, &metrics);
}
/* free the reassembly queue, if any */
while ((q = LIST_FIRST(&tp->t_segq)) != NULL) {
LIST_REMOVE(q, tqe_q);
m_freem(q->tqe_m);
uma_zfree(tcp_reass_zone, q);
tp->t_segqlen--;
V_tcp_reass_qsize--;
}
/* Disconnect offload device, if any. */
tcp_offload_detach(tp);
tcp_free_sackholes(tp);
inp->inp_ppcb = NULL;
tp->t_inpcb = NULL;
uma_zfree(tcpcb_zone, tp);
}
/*
* Attempt to close a TCP control block, marking it as dropped, and freeing
* the socket if we hold the only reference.
*/
struct tcpcb *
tcp_close(struct tcpcb *tp)
{
struct inpcb *inp = tp->t_inpcb;
struct socket *so;
INP_INFO_WLOCK_ASSERT(&V_tcbinfo);
INP_WLOCK_ASSERT(inp);
/* Notify any offload devices of listener close */
if (tp->t_state == TCPS_LISTEN)
tcp_offload_listen_close(tp);
in_pcbdrop(inp);
V_tcpstat.tcps_closed++;
KASSERT(inp->inp_socket != NULL, ("tcp_close: inp_socket NULL"));
so = inp->inp_socket;
soisdisconnected(so);
if (inp->inp_vflag & INP_SOCKREF) {
KASSERT(so->so_state & SS_PROTOREF,
("tcp_close: !SS_PROTOREF"));
inp->inp_vflag &= ~INP_SOCKREF;
INP_WUNLOCK(inp);
ACCEPT_LOCK();
SOCK_LOCK(so);
so->so_state &= ~SS_PROTOREF;
sofree(so);
return (NULL);
}
return (tp);
}
void
tcp_drain(void)
{
if (do_tcpdrain) {
struct inpcb *inpb;
struct tcpcb *tcpb;
struct tseg_qent *te;
/*
* Walk the tcpbs, if existing, and flush the reassembly queue,
* if there is one...
* XXX: The "Net/3" implementation doesn't imply that the TCP
* reassembly queue should be flushed, but in a situation
* where we're really low on mbufs, this is potentially
* usefull.
*/
INP_INFO_RLOCK(&V_tcbinfo);
LIST_FOREACH(inpb, V_tcbinfo.ipi_listhead, inp_list) {
if (inpb->inp_vflag & INP_TIMEWAIT)
continue;
INP_WLOCK(inpb);
if ((tcpb = intotcpcb(inpb)) != NULL) {
while ((te = LIST_FIRST(&tcpb->t_segq))
!= NULL) {
LIST_REMOVE(te, tqe_q);
m_freem(te->tqe_m);
uma_zfree(tcp_reass_zone, te);
tcpb->t_segqlen--;
V_tcp_reass_qsize--;
}
tcp_clean_sackreport(tcpb);
}
INP_WUNLOCK(inpb);
}
INP_INFO_RUNLOCK(&V_tcbinfo);
}
}
/*
* Notify a tcp user of an asynchronous error;
* store error as soft error, but wake up user
* (for now, won't do anything until can select for soft error).
*
* Do not wake up user since there currently is no mechanism for
* reporting soft errors (yet - a kqueue filter may be added).
*/
static struct inpcb *
tcp_notify(struct inpcb *inp, int error)
{
struct tcpcb *tp;
INP_INFO_WLOCK_ASSERT(&V_tcbinfo);
INP_WLOCK_ASSERT(inp);
if ((inp->inp_vflag & INP_TIMEWAIT) ||
(inp->inp_vflag & INP_DROPPED))
return (inp);
tp = intotcpcb(inp);
KASSERT(tp != NULL, ("tcp_notify: tp == NULL"));
/*
* Ignore some errors if we are hooked up.
* If connection hasn't completed, has retransmitted several times,
* and receives a second error, give up now. This is better
* than waiting a long time to establish a connection that
* can never complete.
*/
if (tp->t_state == TCPS_ESTABLISHED &&
(error == EHOSTUNREACH || error == ENETUNREACH ||
error == EHOSTDOWN)) {
return (inp);
} else if (tp->t_state < TCPS_ESTABLISHED && tp->t_rxtshift > 3 &&
tp->t_softerror) {
tp = tcp_drop(tp, error);
if (tp != NULL)
return (inp);
else
return (NULL);
} else {
tp->t_softerror = error;
return (inp);
}
#if 0
wakeup( &so->so_timeo);
sorwakeup(so);
sowwakeup(so);
#endif
}
static int
tcp_pcblist(SYSCTL_HANDLER_ARGS)
{
int error, i, m, n, pcb_count;
struct inpcb *inp, **inp_list;
inp_gen_t gencnt;
struct xinpgen xig;
/*
* The process of preparing the TCB list is too time-consuming and
* resource-intensive to repeat twice on every request.
*/
if (req->oldptr == NULL) {
m = syncache_pcbcount();
n = V_tcbinfo.ipi_count;
req->oldidx = 2 * (sizeof xig)
+ ((m + n) + n/8) * sizeof(struct xtcpcb);
return (0);
}
if (req->newptr != NULL)
return (EPERM);
/*
* OK, now we're committed to doing something.
*/
INP_INFO_RLOCK(&V_tcbinfo);
gencnt = V_tcbinfo.ipi_gencnt;
n = V_tcbinfo.ipi_count;
INP_INFO_RUNLOCK(&V_tcbinfo);
m = syncache_pcbcount();
error = sysctl_wire_old_buffer(req, 2 * (sizeof xig)
+ (n + m) * sizeof(struct xtcpcb));
if (error != 0)
return (error);
xig.xig_len = sizeof xig;
xig.xig_count = n + m;
xig.xig_gen = gencnt;
xig.xig_sogen = so_gencnt;
error = SYSCTL_OUT(req, &xig, sizeof xig);
if (error)
return (error);
error = syncache_pcblist(req, m, &pcb_count);
if (error)
return (error);
inp_list = malloc(n * sizeof *inp_list, M_TEMP, M_WAITOK);
if (inp_list == NULL)
return (ENOMEM);
INP_INFO_RLOCK(&V_tcbinfo);
for (inp = LIST_FIRST(V_tcbinfo.ipi_listhead), i = 0;
inp != NULL && i < n; inp = LIST_NEXT(inp, inp_list)) {
INP_RLOCK(inp);
if (inp->inp_gencnt <= gencnt) {
/*
* XXX: This use of cr_cansee(), introduced with
* TCP state changes, is not quite right, but for
* now, better than nothing.
*/
if (inp->inp_vflag & INP_TIMEWAIT) {
if (intotw(inp) != NULL)
error = cr_cansee(req->td->td_ucred,
intotw(inp)->tw_cred);
else
error = EINVAL; /* Skip this inp. */
} else
error = cr_canseesocket(req->td->td_ucred,
inp->inp_socket);
if (error == 0)
inp_list[i++] = inp;
}
INP_RUNLOCK(inp);
}
INP_INFO_RUNLOCK(&V_tcbinfo);
n = i;
error = 0;
for (i = 0; i < n; i++) {
inp = inp_list[i];
INP_RLOCK(inp);
if (inp->inp_gencnt <= gencnt) {
struct xtcpcb xt;
void *inp_ppcb;
bzero(&xt, sizeof(xt));
xt.xt_len = sizeof xt;
/* XXX should avoid extra copy */
bcopy(inp, &xt.xt_inp, sizeof *inp);
inp_ppcb = inp->inp_ppcb;
if (inp_ppcb == NULL)
bzero((char *) &xt.xt_tp, sizeof xt.xt_tp);
else if (inp->inp_vflag & INP_TIMEWAIT) {
bzero((char *) &xt.xt_tp, sizeof xt.xt_tp);
xt.xt_tp.t_state = TCPS_TIME_WAIT;
} else
bcopy(inp_ppcb, &xt.xt_tp, sizeof xt.xt_tp);
if (inp->inp_socket != NULL)
sotoxsocket(inp->inp_socket, &xt.xt_socket);
else {
bzero(&xt.xt_socket, sizeof xt.xt_socket);
xt.xt_socket.xso_protocol = IPPROTO_TCP;
}
xt.xt_inp.inp_gencnt = inp->inp_gencnt;
INP_RUNLOCK(inp);
error = SYSCTL_OUT(req, &xt, sizeof xt);
} else
INP_RUNLOCK(inp);
}
if (!error) {
/*
* Give the user an updated idea of our state.
* If the generation differs from what we told
* her before, she knows that something happened
* while we were processing this request, and it
* might be necessary to retry.
*/
INP_INFO_RLOCK(&V_tcbinfo);
xig.xig_gen = V_tcbinfo.ipi_gencnt;
xig.xig_sogen = so_gencnt;
xig.xig_count = V_tcbinfo.ipi_count + pcb_count;
INP_INFO_RUNLOCK(&V_tcbinfo);
error = SYSCTL_OUT(req, &xig, sizeof xig);
}
free(inp_list, M_TEMP);
return (error);
}
SYSCTL_PROC(_net_inet_tcp, TCPCTL_PCBLIST, pcblist, CTLFLAG_RD, 0, 0,
tcp_pcblist, "S,xtcpcb", "List of active TCP connections");
static int
tcp_getcred(SYSCTL_HANDLER_ARGS)
{
struct xucred xuc;
struct sockaddr_in addrs[2];
struct inpcb *inp;
int error;
error = priv_check(req->td, PRIV_NETINET_GETCRED);
if (error)
return (error);
error = SYSCTL_IN(req, addrs, sizeof(addrs));
if (error)
return (error);
INP_INFO_RLOCK(&V_tcbinfo);
inp = in_pcblookup_hash(&V_tcbinfo, addrs[1].sin_addr,
addrs[1].sin_port, addrs[0].sin_addr, addrs[0].sin_port, 0, NULL);
if (inp != NULL) {
INP_RLOCK(inp);
INP_INFO_RUNLOCK(&V_tcbinfo);
if (inp->inp_socket == NULL)
error = ENOENT;
if (error == 0)
error = cr_canseesocket(req->td->td_ucred,
inp->inp_socket);
if (error == 0)
cru2x(inp->inp_socket->so_cred, &xuc);
INP_RUNLOCK(inp);
} else {
INP_INFO_RUNLOCK(&V_tcbinfo);
error = ENOENT;
}
if (error == 0)
error = SYSCTL_OUT(req, &xuc, sizeof(struct xucred));
return (error);
}
SYSCTL_PROC(_net_inet_tcp, OID_AUTO, getcred,
CTLTYPE_OPAQUE|CTLFLAG_RW|CTLFLAG_PRISON, 0, 0,
tcp_getcred, "S,xucred", "Get the xucred of a TCP connection");
#ifdef INET6
static int
tcp6_getcred(SYSCTL_HANDLER_ARGS)
{
struct xucred xuc;
struct sockaddr_in6 addrs[2];
struct inpcb *inp;
int error, mapped = 0;
error = priv_check(req->td, PRIV_NETINET_GETCRED);
if (error)
return (error);
error = SYSCTL_IN(req, addrs, sizeof(addrs));
if (error)
return (error);
if ((error = sa6_embedscope(&addrs[0], V_ip6_use_defzone)) != 0 ||
(error = sa6_embedscope(&addrs[1], V_ip6_use_defzone)) != 0) {
return (error);
}
if (IN6_IS_ADDR_V4MAPPED(&addrs[0].sin6_addr)) {
if (IN6_IS_ADDR_V4MAPPED(&addrs[1].sin6_addr))
mapped = 1;
else
return (EINVAL);
}
INP_INFO_RLOCK(&V_tcbinfo);
if (mapped == 1)
inp = in_pcblookup_hash(&V_tcbinfo,
*(struct in_addr *)&addrs[1].sin6_addr.s6_addr[12],
addrs[1].sin6_port,
*(struct in_addr *)&addrs[0].sin6_addr.s6_addr[12],
addrs[0].sin6_port,
0, NULL);
else
inp = in6_pcblookup_hash(&V_tcbinfo,
&addrs[1].sin6_addr, addrs[1].sin6_port,
&addrs[0].sin6_addr, addrs[0].sin6_port, 0, NULL);
if (inp != NULL) {
INP_RLOCK(inp);
INP_INFO_RUNLOCK(&V_tcbinfo);
if (inp->inp_socket == NULL)
error = ENOENT;
if (error == 0)
error = cr_canseesocket(req->td->td_ucred,
inp->inp_socket);
if (error == 0)
cru2x(inp->inp_socket->so_cred, &xuc);
INP_RUNLOCK(inp);
} else {
INP_INFO_RUNLOCK(&V_tcbinfo);
error = ENOENT;
}
if (error == 0)
error = SYSCTL_OUT(req, &xuc, sizeof(struct xucred));
return (error);
}
SYSCTL_PROC(_net_inet6_tcp6, OID_AUTO, getcred,
CTLTYPE_OPAQUE|CTLFLAG_RW|CTLFLAG_PRISON, 0, 0,
tcp6_getcred, "S,xucred", "Get the xucred of a TCP6 connection");
#endif
void
tcp_ctlinput(int cmd, struct sockaddr *sa, void *vip)
{
struct ip *ip = vip;
struct tcphdr *th;
struct in_addr faddr;
struct inpcb *inp;
struct tcpcb *tp;
struct inpcb *(*notify)(struct inpcb *, int) = tcp_notify;
struct icmp *icp;
struct in_conninfo inc;
tcp_seq icmp_tcp_seq;
int mtu;
faddr = ((struct sockaddr_in *)sa)->sin_addr;
if (sa->sa_family != AF_INET || faddr.s_addr == INADDR_ANY)
return;
if (cmd == PRC_MSGSIZE)
notify = tcp_mtudisc;
else if (V_icmp_may_rst && (cmd == PRC_UNREACH_ADMIN_PROHIB ||
cmd == PRC_UNREACH_PORT || cmd == PRC_TIMXCEED_INTRANS) && ip)
notify = tcp_drop_syn_sent;
/*
* Redirects don't need to be handled up here.
*/
else if (PRC_IS_REDIRECT(cmd))
return;
/*
* Source quench is depreciated.
*/
else if (cmd == PRC_QUENCH)
return;
/*
* Hostdead is ugly because it goes linearly through all PCBs.
* XXX: We never get this from ICMP, otherwise it makes an
* excellent DoS attack on machines with many connections.
*/
else if (cmd == PRC_HOSTDEAD)
ip = NULL;
else if ((unsigned)cmd >= PRC_NCMDS || inetctlerrmap[cmd] == 0)
return;
if (ip != NULL) {
icp = (struct icmp *)((caddr_t)ip
- offsetof(struct icmp, icmp_ip));
th = (struct tcphdr *)((caddr_t)ip
+ (ip->ip_hl << 2));
INP_INFO_WLOCK(&V_tcbinfo);
inp = in_pcblookup_hash(&V_tcbinfo, faddr, th->th_dport,
ip->ip_src, th->th_sport, 0, NULL);
if (inp != NULL) {
INP_WLOCK(inp);
if (!(inp->inp_vflag & INP_TIMEWAIT) &&
!(inp->inp_vflag & INP_DROPPED) &&
!(inp->inp_socket == NULL)) {
icmp_tcp_seq = htonl(th->th_seq);
tp = intotcpcb(inp);
if (SEQ_GEQ(icmp_tcp_seq, tp->snd_una) &&
SEQ_LT(icmp_tcp_seq, tp->snd_max)) {
if (cmd == PRC_MSGSIZE) {
/*
* MTU discovery:
* If we got a needfrag set the MTU
* in the route to the suggested new
* value (if given) and then notify.
*/
bzero(&inc, sizeof(inc));
inc.inc_flags = 0; /* IPv4 */
inc.inc_faddr = faddr;
inc.inc_fibnum =
inp->inp_inc.inc_fibnum;
mtu = ntohs(icp->icmp_nextmtu);
/*
* If no alternative MTU was
* proposed, try the next smaller
* one. ip->ip_len has already
* been swapped in icmp_input().
*/
if (!mtu)
mtu = ip_next_mtu(ip->ip_len,
1);
if (mtu < max(296, V_tcp_minmss
+ sizeof(struct tcpiphdr)))
mtu = 0;
if (!mtu)
mtu = V_tcp_mssdflt
+ sizeof(struct tcpiphdr);
/*
* Only cache the the MTU if it
* is smaller than the interface
* or route MTU. tcp_mtudisc()
* will do right thing by itself.
*/
if (mtu <= tcp_maxmtu(&inc, NULL))
tcp_hc_updatemtu(&inc, mtu);
}
inp = (*notify)(inp, inetctlerrmap[cmd]);
}
}
if (inp != NULL)
INP_WUNLOCK(inp);
} else {
inc.inc_fport = th->th_dport;
inc.inc_lport = th->th_sport;
inc.inc_faddr = faddr;
inc.inc_laddr = ip->ip_src;
#ifdef INET6
inc.inc_isipv6 = 0;
#endif
syncache_unreach(&inc, th);
}
INP_INFO_WUNLOCK(&V_tcbinfo);
} else
in_pcbnotifyall(&V_tcbinfo, faddr, inetctlerrmap[cmd], notify);
}
#ifdef INET6
void
tcp6_ctlinput(int cmd, struct sockaddr *sa, void *d)
{
struct tcphdr th;
struct inpcb *(*notify)(struct inpcb *, int) = tcp_notify;
struct ip6_hdr *ip6;
struct mbuf *m;
struct ip6ctlparam *ip6cp = NULL;
const struct sockaddr_in6 *sa6_src = NULL;
int off;
struct tcp_portonly {
u_int16_t th_sport;
u_int16_t th_dport;
} *thp;
if (sa->sa_family != AF_INET6 ||
sa->sa_len != sizeof(struct sockaddr_in6))
return;
if (cmd == PRC_MSGSIZE)
notify = tcp_mtudisc;
else if (!PRC_IS_REDIRECT(cmd) &&
((unsigned)cmd >= PRC_NCMDS || inet6ctlerrmap[cmd] == 0))
return;
/* Source quench is depreciated. */
else if (cmd == PRC_QUENCH)
return;
/* if the parameter is from icmp6, decode it. */
if (d != NULL) {
ip6cp = (struct ip6ctlparam *)d;
m = ip6cp->ip6c_m;
ip6 = ip6cp->ip6c_ip6;
off = ip6cp->ip6c_off;
sa6_src = ip6cp->ip6c_src;
} else {
m = NULL;
ip6 = NULL;
off = 0; /* fool gcc */
sa6_src = &sa6_any;
}
if (ip6 != NULL) {
struct in_conninfo inc;
/*
* XXX: We assume that when IPV6 is non NULL,
* M and OFF are valid.
*/
/* check if we can safely examine src and dst ports */
if (m->m_pkthdr.len < off + sizeof(*thp))
return;
bzero(&th, sizeof(th));
m_copydata(m, off, sizeof(*thp), (caddr_t)&th);
in6_pcbnotify(&V_tcbinfo, sa, th.th_dport,
(struct sockaddr *)ip6cp->ip6c_src,
th.th_sport, cmd, NULL, notify);
inc.inc_fport = th.th_dport;
inc.inc_lport = th.th_sport;
inc.inc6_faddr = ((struct sockaddr_in6 *)sa)->sin6_addr;
inc.inc6_laddr = ip6cp->ip6c_src->sin6_addr;
inc.inc_isipv6 = 1;
INP_INFO_WLOCK(&V_tcbinfo);
syncache_unreach(&inc, &th);
INP_INFO_WUNLOCK(&V_tcbinfo);
} else
in6_pcbnotify(&V_tcbinfo, sa, 0, (const struct sockaddr *)sa6_src,
0, cmd, NULL, notify);
}
#endif /* INET6 */
/*
* Following is where TCP initial sequence number generation occurs.
*
* There are two places where we must use initial sequence numbers:
* 1. In SYN-ACK packets.
* 2. In SYN packets.
*
* All ISNs for SYN-ACK packets are generated by the syncache. See
* tcp_syncache.c for details.
*
* The ISNs in SYN packets must be monotonic; TIME_WAIT recycling
* depends on this property. In addition, these ISNs should be
* unguessable so as to prevent connection hijacking. To satisfy
* the requirements of this situation, the algorithm outlined in
* RFC 1948 is used, with only small modifications.
*
* Implementation details:
*
* Time is based off the system timer, and is corrected so that it
* increases by one megabyte per second. This allows for proper
* recycling on high speed LANs while still leaving over an hour
* before rollover.
*
* As reading the *exact* system time is too expensive to be done
* whenever setting up a TCP connection, we increment the time
* offset in two ways. First, a small random positive increment
* is added to isn_offset for each connection that is set up.
* Second, the function tcp_isn_tick fires once per clock tick
* and increments isn_offset as necessary so that sequence numbers
* are incremented at approximately ISN_BYTES_PER_SECOND. The
* random positive increments serve only to ensure that the same
* exact sequence number is never sent out twice (as could otherwise
* happen when a port is recycled in less than the system tick
* interval.)
*
* net.inet.tcp.isn_reseed_interval controls the number of seconds
* between seeding of isn_secret. This is normally set to zero,
* as reseeding should not be necessary.
*
* Locking of the global variables isn_secret, isn_last_reseed, isn_offset,
* isn_offset_old, and isn_ctx is performed using the TCP pcbinfo lock. In
* general, this means holding an exclusive (write) lock.
*/
#define ISN_BYTES_PER_SECOND 1048576
#define ISN_STATIC_INCREMENT 4096
#define ISN_RANDOM_INCREMENT (4096 - 1)
static u_char isn_secret[32];
static int isn_last_reseed;
static u_int32_t isn_offset, isn_offset_old;
static MD5_CTX isn_ctx;
tcp_seq
tcp_new_isn(struct tcpcb *tp)
{
u_int32_t md5_buffer[4];
tcp_seq new_isn;
INP_WLOCK_ASSERT(tp->t_inpcb);
ISN_LOCK();
/* Seed if this is the first use, reseed if requested. */
if ((V_isn_last_reseed == 0) || ((V_tcp_isn_reseed_interval > 0) &&
(((u_int)V_isn_last_reseed + (u_int)V_tcp_isn_reseed_interval*hz)
< (u_int)ticks))) {
read_random(&V_isn_secret, sizeof(V_isn_secret));
V_isn_last_reseed = ticks;
}
/* Compute the md5 hash and return the ISN. */
MD5Init(&V_isn_ctx);
MD5Update(&V_isn_ctx, (u_char *) &tp->t_inpcb->inp_fport, sizeof(u_short));
MD5Update(&V_isn_ctx, (u_char *) &tp->t_inpcb->inp_lport, sizeof(u_short));
#ifdef INET6
if ((tp->t_inpcb->inp_vflag & INP_IPV6) != 0) {
MD5Update(&V_isn_ctx, (u_char *) &tp->t_inpcb->in6p_faddr,
sizeof(struct in6_addr));
MD5Update(&V_isn_ctx, (u_char *) &tp->t_inpcb->in6p_laddr,
sizeof(struct in6_addr));
} else
#endif
{
MD5Update(&V_isn_ctx, (u_char *) &tp->t_inpcb->inp_faddr,
sizeof(struct in_addr));
MD5Update(&V_isn_ctx, (u_char *) &tp->t_inpcb->inp_laddr,
sizeof(struct in_addr));
}
MD5Update(&V_isn_ctx, (u_char *) &V_isn_secret, sizeof(V_isn_secret));
MD5Final((u_char *) &md5_buffer, &V_isn_ctx);
new_isn = (tcp_seq) md5_buffer[0];
V_isn_offset += ISN_STATIC_INCREMENT +
(arc4random() & ISN_RANDOM_INCREMENT);
new_isn += V_isn_offset;
ISN_UNLOCK();
return (new_isn);
}
/*
* Increment the offset to the next ISN_BYTES_PER_SECOND / 100 boundary
* to keep time flowing at a relatively constant rate. If the random
* increments have already pushed us past the projected offset, do nothing.
*/
static void
tcp_isn_tick(void *xtp)
{
u_int32_t projected_offset;
ISN_LOCK();
projected_offset = V_isn_offset_old + ISN_BYTES_PER_SECOND / 100;
if (SEQ_GT(projected_offset, V_isn_offset))
V_isn_offset = projected_offset;
V_isn_offset_old = V_isn_offset;
callout_reset(&isn_callout, hz/100, tcp_isn_tick, NULL);
ISN_UNLOCK();
}
/*
* When a specific ICMP unreachable message is received and the
* connection state is SYN-SENT, drop the connection. This behavior
* is controlled by the icmp_may_rst sysctl.
*/
struct inpcb *
tcp_drop_syn_sent(struct inpcb *inp, int errno)
{
struct tcpcb *tp;
INP_INFO_WLOCK_ASSERT(&V_tcbinfo);
INP_WLOCK_ASSERT(inp);
if ((inp->inp_vflag & INP_TIMEWAIT) ||
(inp->inp_vflag & INP_DROPPED))
return (inp);
tp = intotcpcb(inp);
if (tp->t_state != TCPS_SYN_SENT)
return (inp);
tp = tcp_drop(tp, errno);
if (tp != NULL)
return (inp);
else
return (NULL);
}
/*
* When `need fragmentation' ICMP is received, update our idea of the MSS
* based on the new value in the route. Also nudge TCP to send something,
* since we know the packet we just sent was dropped.
* This duplicates some code in the tcp_mss() function in tcp_input.c.
*/
struct inpcb *
tcp_mtudisc(struct inpcb *inp, int errno)
{
struct tcpcb *tp;
struct socket *so = inp->inp_socket;
u_int maxmtu;
u_int romtu;
int mss;
#ifdef INET6
int isipv6;
#endif /* INET6 */
INP_WLOCK_ASSERT(inp);
if ((inp->inp_vflag & INP_TIMEWAIT) ||
(inp->inp_vflag & INP_DROPPED))
return (inp);
tp = intotcpcb(inp);
KASSERT(tp != NULL, ("tcp_mtudisc: tp == NULL"));
#ifdef INET6
isipv6 = (tp->t_inpcb->inp_vflag & INP_IPV6) != 0;
#endif
maxmtu = tcp_hc_getmtu(&inp->inp_inc); /* IPv4 and IPv6 */
romtu =
#ifdef INET6
isipv6 ? tcp_maxmtu6(&inp->inp_inc, NULL) :
#endif /* INET6 */
tcp_maxmtu(&inp->inp_inc, NULL);
if (!maxmtu)
maxmtu = romtu;
else
maxmtu = min(maxmtu, romtu);
if (!maxmtu) {
tp->t_maxopd = tp->t_maxseg =
#ifdef INET6
isipv6 ? V_tcp_v6mssdflt :
#endif /* INET6 */
V_tcp_mssdflt;
return (inp);
}
mss = maxmtu -
#ifdef INET6
(isipv6 ? sizeof(struct ip6_hdr) + sizeof(struct tcphdr) :
#endif /* INET6 */
sizeof(struct tcpiphdr)
#ifdef INET6
)
#endif /* INET6 */
;
/*
* XXX - The above conditional probably violates the TCP
* spec. The problem is that, since we don't know the
* other end's MSS, we are supposed to use a conservative
* default. But, if we do that, then MTU discovery will
* never actually take place, because the conservative
* default is much less than the MTUs typically seen
* on the Internet today. For the moment, we'll sweep
* this under the carpet.
*
* The conservative default might not actually be a problem
* if the only case this occurs is when sending an initial
* SYN with options and data to a host we've never talked
* to before. Then, they will reply with an MSS value which
* will get recorded and the new parameters should get
* recomputed. For Further Study.
*/
if (tp->t_maxopd <= mss)
return (inp);
tp->t_maxopd = mss;
if ((tp->t_flags & (TF_REQ_TSTMP|TF_NOOPT)) == TF_REQ_TSTMP &&
(tp->t_flags & TF_RCVD_TSTMP) == TF_RCVD_TSTMP)
mss -= TCPOLEN_TSTAMP_APPA;
#if (MCLBYTES & (MCLBYTES - 1)) == 0
if (mss > MCLBYTES)
mss &= ~(MCLBYTES-1);
#else
if (mss > MCLBYTES)
mss = mss / MCLBYTES * MCLBYTES;
#endif
if (so->so_snd.sb_hiwat < mss)
mss = so->so_snd.sb_hiwat;
tp->t_maxseg = mss;
V_tcpstat.tcps_mturesent++;
tp->t_rtttime = 0;
tp->snd_nxt = tp->snd_una;
tcp_free_sackholes(tp);
tp->snd_recover = tp->snd_max;
if (tp->t_flags & TF_SACK_PERMIT)
EXIT_FASTRECOVERY(tp);
tcp_output_send(tp);
return (inp);
}
/*
* Look-up the routing entry to the peer of this inpcb. If no route
* is found and it cannot be allocated, then return NULL. This routine
* is called by TCP routines that access the rmx structure and by tcp_mss
* to get the interface MTU.
*/
u_long
tcp_maxmtu(struct in_conninfo *inc, int *flags)
{
struct route sro;
struct sockaddr_in *dst;
struct ifnet *ifp;
u_long maxmtu = 0;
KASSERT(inc != NULL, ("tcp_maxmtu with NULL in_conninfo pointer"));
bzero(&sro, sizeof(sro));
if (inc->inc_faddr.s_addr != INADDR_ANY) {
dst = (struct sockaddr_in *)&sro.ro_dst;
dst->sin_family = AF_INET;
dst->sin_len = sizeof(*dst);
dst->sin_addr = inc->inc_faddr;
in_rtalloc_ign(&sro, RTF_CLONING, inc->inc_fibnum);
}
if (sro.ro_rt != NULL) {
ifp = sro.ro_rt->rt_ifp;
if (sro.ro_rt->rt_rmx.rmx_mtu == 0)
maxmtu = ifp->if_mtu;
else
maxmtu = min(sro.ro_rt->rt_rmx.rmx_mtu, ifp->if_mtu);
/* Report additional interface capabilities. */
if (flags != NULL) {
if (ifp->if_capenable & IFCAP_TSO4 &&
ifp->if_hwassist & CSUM_TSO)
*flags |= CSUM_TSO;
}
RTFREE(sro.ro_rt);
}
return (maxmtu);
}
#ifdef INET6
u_long
tcp_maxmtu6(struct in_conninfo *inc, int *flags)
{
struct route_in6 sro6;
struct ifnet *ifp;
u_long maxmtu = 0;
KASSERT(inc != NULL, ("tcp_maxmtu6 with NULL in_conninfo pointer"));
bzero(&sro6, sizeof(sro6));
if (!IN6_IS_ADDR_UNSPECIFIED(&inc->inc6_faddr)) {
sro6.ro_dst.sin6_family = AF_INET6;
sro6.ro_dst.sin6_len = sizeof(struct sockaddr_in6);
sro6.ro_dst.sin6_addr = inc->inc6_faddr;
rtalloc_ign((struct route *)&sro6, RTF_CLONING);
}
if (sro6.ro_rt != NULL) {
ifp = sro6.ro_rt->rt_ifp;
if (sro6.ro_rt->rt_rmx.rmx_mtu == 0)
maxmtu = IN6_LINKMTU(sro6.ro_rt->rt_ifp);
else
maxmtu = min(sro6.ro_rt->rt_rmx.rmx_mtu,
IN6_LINKMTU(sro6.ro_rt->rt_ifp));
/* Report additional interface capabilities. */
if (flags != NULL) {
if (ifp->if_capenable & IFCAP_TSO6 &&
ifp->if_hwassist & CSUM_TSO)
*flags |= CSUM_TSO;
}
RTFREE(sro6.ro_rt);
}
return (maxmtu);
}
#endif /* INET6 */
#ifdef IPSEC
/* compute ESP/AH header size for TCP, including outer IP header. */
size_t
ipsec_hdrsiz_tcp(struct tcpcb *tp)
{
struct inpcb *inp;
struct mbuf *m;
size_t hdrsiz;
struct ip *ip;
#ifdef INET6
struct ip6_hdr *ip6;
#endif
struct tcphdr *th;
if ((tp == NULL) || ((inp = tp->t_inpcb) == NULL))
return (0);
MGETHDR(m, M_DONTWAIT, MT_DATA);
if (!m)
return (0);
#ifdef INET6
if ((inp->inp_vflag & INP_IPV6) != 0) {
ip6 = mtod(m, struct ip6_hdr *);
th = (struct tcphdr *)(ip6 + 1);
m->m_pkthdr.len = m->m_len =
sizeof(struct ip6_hdr) + sizeof(struct tcphdr);
tcpip_fillheaders(inp, ip6, th);
hdrsiz = ipsec6_hdrsiz(m, IPSEC_DIR_OUTBOUND, inp);
} else
#endif /* INET6 */
{
ip = mtod(m, struct ip *);
th = (struct tcphdr *)(ip + 1);
m->m_pkthdr.len = m->m_len = sizeof(struct tcpiphdr);
tcpip_fillheaders(inp, ip, th);
hdrsiz = ipsec4_hdrsiz(m, IPSEC_DIR_OUTBOUND, inp);
}
m_free(m);
return (hdrsiz);
}
#endif /* IPSEC */
/*
* TCP BANDWIDTH DELAY PRODUCT WINDOW LIMITING
*
* This code attempts to calculate the bandwidth-delay product as a
* means of determining the optimal window size to maximize bandwidth,
* minimize RTT, and avoid the over-allocation of buffers on interfaces and
* routers. This code also does a fairly good job keeping RTTs in check
* across slow links like modems. We implement an algorithm which is very
* similar (but not meant to be) TCP/Vegas. The code operates on the
* transmitter side of a TCP connection and so only effects the transmit
* side of the connection.
*
* BACKGROUND: TCP makes no provision for the management of buffer space
* at the end points or at the intermediate routers and switches. A TCP
* stream, whether using NewReno or not, will eventually buffer as
* many packets as it is able and the only reason this typically works is
* due to the fairly small default buffers made available for a connection
* (typicaly 16K or 32K). As machines use larger windows and/or window
* scaling it is now fairly easy for even a single TCP connection to blow-out
* all available buffer space not only on the local interface, but on
* intermediate routers and switches as well. NewReno makes a misguided
* attempt to 'solve' this problem by waiting for an actual failure to occur,
* then backing off, then steadily increasing the window again until another
* failure occurs, ad-infinitum. This results in terrible oscillation that
* is only made worse as network loads increase and the idea of intentionally
* blowing out network buffers is, frankly, a terrible way to manage network
* resources.
*
* It is far better to limit the transmit window prior to the failure
* condition being achieved. There are two general ways to do this: First
* you can 'scan' through different transmit window sizes and locate the
* point where the RTT stops increasing, indicating that you have filled the
* pipe, then scan backwards until you note that RTT stops decreasing, then
* repeat ad-infinitum. This method works in principle but has severe
* implementation issues due to RTT variances, timer granularity, and
* instability in the algorithm which can lead to many false positives and
* create oscillations as well as interact badly with other TCP streams
* implementing the same algorithm.
*
* The second method is to limit the window to the bandwidth delay product
* of the link. This is the method we implement. RTT variances and our
* own manipulation of the congestion window, bwnd, can potentially
* destabilize the algorithm. For this reason we have to stabilize the
* elements used to calculate the window. We do this by using the minimum
* observed RTT, the long term average of the observed bandwidth, and
* by adding two segments worth of slop. It isn't perfect but it is able
* to react to changing conditions and gives us a very stable basis on
* which to extend the algorithm.
*/
void
tcp_xmit_bandwidth_limit(struct tcpcb *tp, tcp_seq ack_seq)
{
u_long bw;
u_long bwnd;
int save_ticks;
INP_WLOCK_ASSERT(tp->t_inpcb);
/*
* If inflight_enable is disabled in the middle of a tcp connection,
* make sure snd_bwnd is effectively disabled.
*/
if (V_tcp_inflight_enable == 0 ||
tp->t_rttlow < V_tcp_inflight_rttthresh) {
tp->snd_bwnd = TCP_MAXWIN << TCP_MAX_WINSHIFT;
tp->snd_bandwidth = 0;
return;
}
/*
* Figure out the bandwidth. Due to the tick granularity this
* is a very rough number and it MUST be averaged over a fairly
* long period of time. XXX we need to take into account a link
* that is not using all available bandwidth, but for now our
* slop will ramp us up if this case occurs and the bandwidth later
* increases.
*
* Note: if ticks rollover 'bw' may wind up negative. We must
* effectively reset t_bw_rtttime for this case.
*/
save_ticks = ticks;
if ((u_int)(save_ticks - tp->t_bw_rtttime) < 1)
return;
bw = (int64_t)(ack_seq - tp->t_bw_rtseq) * hz /
(save_ticks - tp->t_bw_rtttime);
tp->t_bw_rtttime = save_ticks;
tp->t_bw_rtseq = ack_seq;
if (tp->t_bw_rtttime == 0 || (int)bw < 0)
return;
bw = ((int64_t)tp->snd_bandwidth * 15 + bw) >> 4;
tp->snd_bandwidth = bw;
/*
* Calculate the semi-static bandwidth delay product, plus two maximal
* segments. The additional slop puts us squarely in the sweet
* spot and also handles the bandwidth run-up case and stabilization.
* Without the slop we could be locking ourselves into a lower
* bandwidth.
*
* Situations Handled:
* (1) Prevents over-queueing of packets on LANs, especially on
* high speed LANs, allowing larger TCP buffers to be
* specified, and also does a good job preventing
* over-queueing of packets over choke points like modems
* (at least for the transmit side).
*
* (2) Is able to handle changing network loads (bandwidth
* drops so bwnd drops, bandwidth increases so bwnd
* increases).
*
* (3) Theoretically should stabilize in the face of multiple
* connections implementing the same algorithm (this may need
* a little work).
*
* (4) Stability value (defaults to 20 = 2 maximal packets) can
* be adjusted with a sysctl but typically only needs to be
* on very slow connections. A value no smaller then 5
* should be used, but only reduce this default if you have
* no other choice.
*/
#define USERTT ((tp->t_srtt + tp->t_rttbest) / 2)
bwnd = (int64_t)bw * USERTT / (hz << TCP_RTT_SHIFT) + V_tcp_inflight_stab * tp->t_maxseg / 10;
#undef USERTT
if (tcp_inflight_debug > 0) {
static int ltime;
if ((u_int)(ticks - ltime) >= hz / tcp_inflight_debug) {
ltime = ticks;
printf("%p bw %ld rttbest %d srtt %d bwnd %ld\n",
tp,
bw,
tp->t_rttbest,
tp->t_srtt,
bwnd
);
}
}
if ((long)bwnd < V_tcp_inflight_min)
bwnd = V_tcp_inflight_min;
if (bwnd > V_tcp_inflight_max)
bwnd = V_tcp_inflight_max;
if ((long)bwnd < tp->t_maxseg * 2)
bwnd = tp->t_maxseg * 2;
tp->snd_bwnd = bwnd;
}
#ifdef TCP_SIGNATURE
/*
* Callback function invoked by m_apply() to digest TCP segment data
* contained within an mbuf chain.
*/
static int
tcp_signature_apply(void *fstate, void *data, u_int len)
{
MD5Update(fstate, (u_char *)data, len);
return (0);
}
/*
* Compute TCP-MD5 hash of a TCPv4 segment. (RFC2385)
*
* Parameters:
* m pointer to head of mbuf chain
* off0 offset to TCP header within the mbuf chain
* len length of TCP segment data, excluding options
* optlen length of TCP segment options
* buf pointer to storage for computed MD5 digest
* direction direction of flow (IPSEC_DIR_INBOUND or OUTBOUND)
*
* We do this over ip, tcphdr, segment data, and the key in the SADB.
* When called from tcp_input(), we can be sure that th_sum has been
* zeroed out and verified already.
*
* This function is for IPv4 use only. Calling this function with an
* IPv6 packet in the mbuf chain will yield undefined results.
*
* Return 0 if successful, otherwise return -1.
*
* XXX The key is retrieved from the system's PF_KEY SADB, by keying a
* search with the destination IP address, and a 'magic SPI' to be
* determined by the application. This is hardcoded elsewhere to 1179
* right now. Another branch of this code exists which uses the SPD to
* specify per-application flows but it is unstable.
*/
int
tcp_signature_compute(struct mbuf *m, int off0, int len, int optlen,
u_char *buf, u_int direction)
{
union sockaddr_union dst;
struct ippseudo ippseudo;
MD5_CTX ctx;
int doff;
struct ip *ip;
struct ipovly *ipovly;
struct secasvar *sav;
struct tcphdr *th;
u_short savecsum;
KASSERT(m != NULL, ("NULL mbuf chain"));
KASSERT(buf != NULL, ("NULL signature pointer"));
/* Extract the destination from the IP header in the mbuf. */
ip = mtod(m, struct ip *);
bzero(&dst, sizeof(union sockaddr_union));
dst.sa.sa_len = sizeof(struct sockaddr_in);
dst.sa.sa_family = AF_INET;
dst.sin.sin_addr = (direction == IPSEC_DIR_INBOUND) ?
ip->ip_src : ip->ip_dst;
/* Look up an SADB entry which matches the address of the peer. */
sav = KEY_ALLOCSA(&dst, IPPROTO_TCP, htonl(TCP_SIG_SPI));
if (sav == NULL) {
printf("%s: SADB lookup failed for %s\n", __func__,
inet_ntoa(dst.sin.sin_addr));
return (EINVAL);
}
MD5Init(&ctx);
ipovly = (struct ipovly *)ip;
th = (struct tcphdr *)((u_char *)ip + off0);
doff = off0 + sizeof(struct tcphdr) + optlen;
/*
* Step 1: Update MD5 hash with IP pseudo-header.
*
* XXX The ippseudo header MUST be digested in network byte order,
* or else we'll fail the regression test. Assume all fields we've
* been doing arithmetic on have been in host byte order.
* XXX One cannot depend on ipovly->ih_len here. When called from
* tcp_output(), the underlying ip_len member has not yet been set.
*/
ippseudo.ippseudo_src = ipovly->ih_src;
ippseudo.ippseudo_dst = ipovly->ih_dst;
ippseudo.ippseudo_pad = 0;
ippseudo.ippseudo_p = IPPROTO_TCP;
ippseudo.ippseudo_len = htons(len + sizeof(struct tcphdr) + optlen);
MD5Update(&ctx, (char *)&ippseudo, sizeof(struct ippseudo));
/*
* Step 2: Update MD5 hash with TCP header, excluding options.
* The TCP checksum must be set to zero.
*/
savecsum = th->th_sum;
th->th_sum = 0;
MD5Update(&ctx, (char *)th, sizeof(struct tcphdr));
th->th_sum = savecsum;
/*
* Step 3: Update MD5 hash with TCP segment data.
* Use m_apply() to avoid an early m_pullup().
*/
if (len > 0)
m_apply(m, doff, len, tcp_signature_apply, &ctx);
/*
* Step 4: Update MD5 hash with shared secret.
*/
MD5Update(&ctx, sav->key_auth->key_data, _KEYLEN(sav->key_auth));
MD5Final(buf, &ctx);
key_sa_recordxfer(sav, m);
KEY_FREESAV(&sav);
return (0);
}
#endif /* TCP_SIGNATURE */
static int
sysctl_drop(SYSCTL_HANDLER_ARGS)
{
/* addrs[0] is a foreign socket, addrs[1] is a local one. */
struct sockaddr_storage addrs[2];
struct inpcb *inp;
struct tcpcb *tp;
struct tcptw *tw;
struct sockaddr_in *fin, *lin;
#ifdef INET6
struct sockaddr_in6 *fin6, *lin6;
struct in6_addr f6, l6;
#endif
int error;
inp = NULL;
fin = lin = NULL;
#ifdef INET6
fin6 = lin6 = NULL;
#endif
error = 0;
if (req->oldptr != NULL || req->oldlen != 0)
return (EINVAL);
if (req->newptr == NULL)
return (EPERM);
if (req->newlen < sizeof(addrs))
return (ENOMEM);
error = SYSCTL_IN(req, &addrs, sizeof(addrs));
if (error)
return (error);
switch (addrs[0].ss_family) {
#ifdef INET6
case AF_INET6:
fin6 = (struct sockaddr_in6 *)&addrs[0];
lin6 = (struct sockaddr_in6 *)&addrs[1];
if (fin6->sin6_len != sizeof(struct sockaddr_in6) ||
lin6->sin6_len != sizeof(struct sockaddr_in6))
return (EINVAL);
if (IN6_IS_ADDR_V4MAPPED(&fin6->sin6_addr)) {
if (!IN6_IS_ADDR_V4MAPPED(&lin6->sin6_addr))
return (EINVAL);
in6_sin6_2_sin_in_sock((struct sockaddr *)&addrs[0]);
in6_sin6_2_sin_in_sock((struct sockaddr *)&addrs[1]);
fin = (struct sockaddr_in *)&addrs[0];
lin = (struct sockaddr_in *)&addrs[1];
break;
}
error = sa6_embedscope(fin6, V_ip6_use_defzone);
if (error)
return (error);
error = sa6_embedscope(lin6, V_ip6_use_defzone);
if (error)
return (error);
break;
#endif
case AF_INET:
fin = (struct sockaddr_in *)&addrs[0];
lin = (struct sockaddr_in *)&addrs[1];
if (fin->sin_len != sizeof(struct sockaddr_in) ||
lin->sin_len != sizeof(struct sockaddr_in))
return (EINVAL);
break;
default:
return (EINVAL);
}
INP_INFO_WLOCK(&V_tcbinfo);
switch (addrs[0].ss_family) {
#ifdef INET6
case AF_INET6:
inp = in6_pcblookup_hash(&V_tcbinfo, &f6, fin6->sin6_port,
&l6, lin6->sin6_port, 0, NULL);
break;
#endif
case AF_INET:
inp = in_pcblookup_hash(&V_tcbinfo, fin->sin_addr,
fin->sin_port, lin->sin_addr, lin->sin_port, 0, NULL);
break;
}
if (inp != NULL) {
INP_WLOCK(inp);
if (inp->inp_vflag & INP_TIMEWAIT) {
/*
* XXXRW: There currently exists a state where an
* inpcb is present, but its timewait state has been
* discarded. For now, don't allow dropping of this
* type of inpcb.
*/
tw = intotw(inp);
if (tw != NULL)
tcp_twclose(tw, 0);
else
INP_WUNLOCK(inp);
} else if (!(inp->inp_vflag & INP_DROPPED) &&
!(inp->inp_socket->so_options & SO_ACCEPTCONN)) {
tp = intotcpcb(inp);
tp = tcp_drop(tp, ECONNABORTED);
if (tp != NULL)
INP_WUNLOCK(inp);
} else
INP_WUNLOCK(inp);
} else
error = ESRCH;
INP_INFO_WUNLOCK(&V_tcbinfo);
return (error);
}
SYSCTL_PROC(_net_inet_tcp, TCPCTL_DROP, drop,
CTLTYPE_STRUCT|CTLFLAG_WR|CTLFLAG_SKIP, NULL,
0, sysctl_drop, "", "Drop TCP connection");
/*
* Generate a standardized TCP log line for use throughout the
* tcp subsystem. Memory allocation is done with M_NOWAIT to
* allow use in the interrupt context.
*
* NB: The caller MUST free(s, M_TCPLOG) the returned string.
* NB: The function may return NULL if memory allocation failed.
*
* Due to header inclusion and ordering limitations the struct ip
* and ip6_hdr pointers have to be passed as void pointers.
*/
char *
tcp_log_addrs(struct in_conninfo *inc, struct tcphdr *th, void *ip4hdr,
const void *ip6hdr)
{
char *s, *sp;
size_t size;
struct ip *ip;
#ifdef INET6
const struct ip6_hdr *ip6;
ip6 = (const struct ip6_hdr *)ip6hdr;
#endif /* INET6 */
ip = (struct ip *)ip4hdr;
/*
* The log line looks like this:
* "TCP: [1.2.3.4]:50332 to [1.2.3.4]:80 tcpflags 0x2<SYN>"
*/
size = sizeof("TCP: []:12345 to []:12345 tcpflags 0x2<>") +
sizeof(PRINT_TH_FLAGS) + 1 +
#ifdef INET6
2 * INET6_ADDRSTRLEN;
#else
2 * INET_ADDRSTRLEN;
#endif /* INET6 */
/* Is logging enabled? */
if (tcp_log_debug == 0 && tcp_log_in_vain == 0)
return (NULL);
s = malloc(size, M_TCPLOG, M_ZERO|M_NOWAIT);
if (s == NULL)
return (NULL);
strcat(s, "TCP: [");
sp = s + strlen(s);
if (inc && inc->inc_isipv6 == 0) {
inet_ntoa_r(inc->inc_faddr, sp);
sp = s + strlen(s);
sprintf(sp, "]:%i to [", ntohs(inc->inc_fport));
sp = s + strlen(s);
inet_ntoa_r(inc->inc_laddr, sp);
sp = s + strlen(s);
sprintf(sp, "]:%i", ntohs(inc->inc_lport));
#ifdef INET6
} else if (inc) {
ip6_sprintf(sp, &inc->inc6_faddr);
sp = s + strlen(s);
sprintf(sp, "]:%i to [", ntohs(inc->inc_fport));
sp = s + strlen(s);
ip6_sprintf(sp, &inc->inc6_laddr);
sp = s + strlen(s);
sprintf(sp, "]:%i", ntohs(inc->inc_lport));
} else if (ip6 && th) {
ip6_sprintf(sp, &ip6->ip6_src);
sp = s + strlen(s);
sprintf(sp, "]:%i to [", ntohs(th->th_sport));
sp = s + strlen(s);
ip6_sprintf(sp, &ip6->ip6_dst);
sp = s + strlen(s);
sprintf(sp, "]:%i", ntohs(th->th_dport));
#endif /* INET6 */
} else if (ip && th) {
inet_ntoa_r(ip->ip_src, sp);
sp = s + strlen(s);
sprintf(sp, "]:%i to [", ntohs(th->th_sport));
sp = s + strlen(s);
inet_ntoa_r(ip->ip_dst, sp);
sp = s + strlen(s);
sprintf(sp, "]:%i", ntohs(th->th_dport));
} else {
free(s, M_TCPLOG);
return (NULL);
}
sp = s + strlen(s);
if (th)
sprintf(sp, " tcpflags 0x%b", th->th_flags, PRINT_TH_FLAGS);
if (*(s + size - 1) != '\0')
panic("%s: string too long", __func__);
return (s);
}
|