aboutsummaryrefslogtreecommitdiffstats
path: root/src/plugins/nat/nat_ipfix_logging.h
blob: a5cdb1a3c1aa4422717fe892600357596fa1bcc8 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
/*
 * nat_ipfix_logging.h - NAT Events IPFIX logging
 *
 * Copyright (c) 2016 Cisco and/or its affiliates.
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at:
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */
#ifndef __included_nat_ipfix_logging_h__
#define __included_nat_ipfix_logging_h__

#include <nat/nat.h>

typedef enum {
  NAT_ADDRESSES_EXHAUTED = 3,
  NAT44_SESSION_CREATE = 4,
  NAT44_SESSION_DELETE = 5,
  NAT64_SESSION_CREATE = 6,
  NAT64_SESSION_DELETE = 7,
  NAT64_BIB_CREATE = 10,
  NAT64_BIB_DELETE = 11,
  NAT_PORTS_EXHAUSTED = 12,
  QUOTA_EXCEEDED = 13,
} nat_event_t;

typedef enum {
  MAX_SESSION_ENTRIES = 1,
  MAX_BIB_ENTRIES = 2,
  MAX_ENTRIES_PER_USER = 3,
  MAX_FRAGMENTS_PENDING_REASSEMBLY = 5,
  MAX_FRAGMENTS_PENDING_REASSEMBLY_IP6,
} quota_exceed_event_t;

typedef struct {

  /** ipfix buffers under construction */
  vlib_buffer_t *nat44_session_buffer;
  vlib_buffer_t *addr_exhausted_buffer;
  vlib_buffer_t *max_entries_per_user_buffer;
  vlib_buffer_t *max_sessions_buffer;
  vlib_buffer_t *max_bibs_buffer;
  vlib_buffer_t *max_frags_ip4_buffer;
  vlib_buffer_t *max_frags_ip6_buffer;
  vlib_buffer_t *nat64_bib_buffer;
  vlib_buffer_t *nat64_ses_buffer;

  /** frames containing ipfix buffers */
  vlib_frame_t *nat44_session_frame;
  vlib_frame_t *addr_exhausted_frame;
  vlib_frame_t *max_entries_per_user_frame;
  vlib_frame_t *max_sessions_frame;
  vlib_frame_t *max_bibs_frame;
  vlib_frame_t *max_frags_ip4_frame;
  vlib_frame_t *max_frags_ip6_frame;
  vlib_frame_t *nat64_bib_frame;
  vlib_frame_t *nat64_ses_frame;

  /** next record offset */
  u32 nat44_session_next_record_offset;
  u32 addr_exhausted_next_record_offset;
  u32 max_entries_per_user_next_record_offset;
  u32 max_sessions_next_record_offset;
  u32 max_bibs_next_record_offset;
  u32 max_frags_ip4_next_record_offset;
  u32 max_frags_ip6_next_record_offset;
  u32 nat64_bib_next_record_offset;
  u32 nat64_ses_next_record_offset;

} snat_ipfix_per_thread_data_t;

typedef struct {
  /** NAT plugin IPFIX logging enabled */
  u8 enabled;

  /** Time reference pair */
  u64 milisecond_time_0;
  f64 vlib_time_0;

  /* Per thread data */
  snat_ipfix_per_thread_data_t *per_thread_data;

  /** template IDs */
  u16 nat44_session_template_id;
  u16 addr_exhausted_template_id;
  u16 max_entries_per_user_template_id;
  u16 max_sessions_template_id;
  u16 max_bibs_template_id;
  u16 max_frags_ip4_template_id;
  u16 max_frags_ip6_template_id;
  u16 nat64_bib_template_id;
  u16 nat64_ses_template_id;

  /** stream index */
  u32 stream_index;

  /** vector of worker vlib mains */
  vlib_main_t **worker_vms;

  /** nat data callbacks call counter */
  u16 call_counter;

} snat_ipfix_logging_main_t;

extern snat_ipfix_logging_main_t snat_ipfix_logging_main;

void snat_ipfix_logging_init (vlib_main_t * vm);
int snat_ipfix_logging_enable_disable (int enable, u32 domain_id, u16 src_port);
void snat_ipfix_logging_nat44_ses_create (u32 thread_index, u32 src_ip,
                                          u32 nat_src_ip,
                                          snat_protocol_t snat_proto,
                                          u16 src_port, u16 nat_src_port,
                                          u32 vrf_id);
void snat_ipfix_logging_nat44_ses_delete (u32 thread_index, u32 src_ip,
                                          u32 nat_src_ip,
                                          snat_protocol_t snat_proto,
                                          u16 src_port, u16 nat_src_port,
                                          u32 vrf_id);
void snat_ipfix_logging_addresses_exhausted(u32 thread_index, u32 pool_id);
void snat_ipfix_logging_max_entries_per_user(u32 thread_index,
                                             u32 limit, u32 src_ip);
void nat_ipfix_logging_max_sessions(u32 thread_index, u32 limit);
void nat_ipfix_logging_max_bibs(u32 thread_index, u32 limit);
void nat_ipfix_logging_max_fragments_ip4(u32 thread_index,
                                         u32 limit, ip4_address_t * src);
void nat_ipfix_logging_max_fragments_ip6(u32 thread_index,
                                         u32 limit, ip6_address_t * src);
void nat_ipfix_logging_nat64_session(u32 thread_index,
                                     ip6_address_t * src_ip,
                                     ip4_address_t * nat_src_ip, u8 proto,
                                     u16 src_port, u16 nat_src_port,
                                     ip6_address_t * dst_ip,
                                     ip4_address_t * nat_dst_ip,
                                     u16 dst_port, u16 nat_dst_port,
                                     u32 vrf_id, u8 is_create);
void nat_ipfix_logging_nat64_bib(u32 thread_index,
                                 ip6_address_t * src_ip,
                                 ip4_address_t * nat_src_ip, u8 proto,
                                 u16 src_port, u16 nat_src_port,
                                 u32 vrf_id, u8 is_create);

#endif /* __included_nat_ipfix_logging_h__ */
923'>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 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568 2569 2570 2571 2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 2662 2663 2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 2957 2958 2959 2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978 2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030 3031 3032 3033 3034 3035 3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062 3063 3064 3065 3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111 3112 3113 3114 3115 3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141 3142 3143 3144 3145 3146 3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 3196 3197 3198 3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 3209 3210 3211 3212 3213 3214 3215 3216 3217 3218 3219 3220 3221 3222 3223 3224 3225 3226 3227 3228 3229 3230 3231 3232 3233 3234 3235 3236 3237 3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264 3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288 3289 3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 3300
/*
 * Copyright (c) 2016-2019 Cisco and/or its affiliates.
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at:
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */

#include <vppinfra/sparse_vec.h>
#include <vnet/fib/ip4_fib.h>
#include <vnet/fib/ip6_fib.h>
#include <vnet/tcp/tcp.h>
#include <vnet/tcp/tcp_inlines.h>
#include <vnet/session/session.h>
#include <math.h>

static char *tcp_error_strings[] = {
#define tcp_error(n,s) s,
#include <vnet/tcp/tcp_error.def>
#undef tcp_error
};

/* All TCP nodes have the same outgoing arcs */
#define foreach_tcp_state_next                  \
  _ (DROP4, "ip4-drop")                         \
  _ (DROP6, "ip6-drop")                         \
  _ (TCP4_OUTPUT, "tcp4-output")                \
  _ (TCP6_OUTPUT, "tcp6-output")

typedef enum _tcp_established_next
{
#define _(s,n) TCP_ESTABLISHED_NEXT_##s,
  foreach_tcp_state_next
#undef _
    TCP_ESTABLISHED_N_NEXT,
} tcp_established_next_t;

typedef enum _tcp_rcv_process_next
{
#define _(s,n) TCP_RCV_PROCESS_NEXT_##s,
  foreach_tcp_state_next
#undef _
    TCP_RCV_PROCESS_N_NEXT,
} tcp_rcv_process_next_t;

typedef enum _tcp_syn_sent_next
{
#define _(s,n) TCP_SYN_SENT_NEXT_##s,
  foreach_tcp_state_next
#undef _
    TCP_SYN_SENT_N_NEXT,
} tcp_syn_sent_next_t;

typedef enum _tcp_listen_next
{
#define _(s,n) TCP_LISTEN_NEXT_##s,
  foreach_tcp_state_next
#undef _
    TCP_LISTEN_N_NEXT,
} tcp_listen_next_t;

/* Generic, state independent indices */
typedef enum _tcp_state_next
{
#define _(s,n) TCP_NEXT_##s,
  foreach_tcp_state_next
#undef _
    TCP_STATE_N_NEXT,
} tcp_state_next_t;

#define tcp_next_output(is_ip4) (is_ip4 ? TCP_NEXT_TCP4_OUTPUT          \
                                        : TCP_NEXT_TCP6_OUTPUT)

#define tcp_next_drop(is_ip4) (is_ip4 ? TCP_NEXT_DROP4                  \
                                      : TCP_NEXT_DROP6)

/**
 * Validate segment sequence number. As per RFC793:
 *
 * Segment Receive Test
 *      Length  Window
 *      ------- -------  -------------------------------------------
 *      0       0       SEG.SEQ = RCV.NXT
 *      0       >0      RCV.NXT =< SEG.SEQ < RCV.NXT+RCV.WND
 *      >0      0       not acceptable
 *      >0      >0      RCV.NXT =< SEG.SEQ < RCV.NXT+RCV.WND
 *                      or RCV.NXT =< SEG.SEQ+SEG.LEN-1 < RCV.NXT+RCV.WND
 *
 * This ultimately consists in checking if segment falls within the window.
 * The one important difference compared to RFC793 is that we use rcv_las,
 * or the rcv_nxt at last ack sent instead of rcv_nxt since that's the
 * peer's reference when computing our receive window.
 *
 * This:
 *  seq_leq (end_seq, tc->rcv_las + tc->rcv_wnd) && seq_geq (seq, tc->rcv_las)
 * however, is too strict when we have retransmits. Instead we just check that
 * the seq is not beyond the right edge and that the end of the segment is not
 * less than the left edge.
 *
 * N.B. rcv_nxt and rcv_wnd are both updated in this node if acks are sent, so
 * use rcv_nxt in the right edge window test instead of rcv_las.
 *
 */
always_inline u8
tcp_segment_in_rcv_wnd (tcp_connection_t * tc, u32 seq, u32 end_seq)
{
  return (seq_geq (end_seq, tc->rcv_las)
	  && seq_leq (seq, tc->rcv_nxt + tc->rcv_wnd));
}

/**
 * RFC1323: Check against wrapped sequence numbers (PAWS). If we have
 * timestamp to echo and it's less than tsval_recent, drop segment
 * but still send an ACK in order to retain TCP's mechanism for detecting
 * and recovering from half-open connections
 *
 * Or at least that's what the theory says. It seems that this might not work
 * very well with packet reordering and fast retransmit. XXX
 */
always_inline int
tcp_segment_check_paws (tcp_connection_t * tc)
{
  return tcp_opts_tstamp (&tc->rcv_opts)
    && timestamp_lt (tc->rcv_opts.tsval, tc->tsval_recent);
}

/**
 * Update tsval recent
 */
always_inline void
tcp_update_timestamp (tcp_connection_t * tc, u32 seq, u32 seq_end)
{
  /*
   * RFC1323: If Last.ACK.sent falls within the range of sequence numbers
   * of an incoming segment:
   *    SEG.SEQ <= Last.ACK.sent < SEG.SEQ + SEG.LEN
   * then the TSval from the segment is copied to TS.Recent;
   * otherwise, the TSval is ignored.
   */
  if (tcp_opts_tstamp (&tc->rcv_opts) && seq_leq (seq, tc->rcv_las)
      && seq_leq (tc->rcv_las, seq_end))
    {
      ASSERT (timestamp_leq (tc->tsval_recent, tc->rcv_opts.tsval));
      tc->tsval_recent = tc->rcv_opts.tsval;
      tc->tsval_recent_age = tcp_time_now_w_thread (tc->c_thread_index);
    }
}

static void
tcp_handle_rst (tcp_connection_t * tc)
{
  switch (tc->rst_state)
    {
    case TCP_STATE_SYN_RCVD:
      /* Cleanup everything. App wasn't notified yet */
      session_transport_delete_notify (&tc->connection);
      tcp_connection_cleanup (tc);
      break;
    case TCP_STATE_SYN_SENT:
      session_stream_connect_notify (&tc->connection, 1 /* fail */ );
      tcp_connection_cleanup (tc);
      break;
    case TCP_STATE_ESTABLISHED:
      session_transport_reset_notify (&tc->connection);
      session_transport_closed_notify (&tc->connection);
      break;
    case TCP_STATE_CLOSE_WAIT:
    case TCP_STATE_FIN_WAIT_1:
    case TCP_STATE_FIN_WAIT_2:
    case TCP_STATE_CLOSING:
    case TCP_STATE_LAST_ACK:
      session_transport_closed_notify (&tc->connection);
      break;
    case TCP_STATE_CLOSED:
    case TCP_STATE_TIME_WAIT:
      break;
    default:
      TCP_DBG ("reset state: %u", tc->state);
    }
}

static void
tcp_program_reset_ntf (tcp_worker_ctx_t * wrk, tcp_connection_t * tc)
{
  if (!tcp_disconnect_pending (tc))
    {
      tc->rst_state = tc->state;
      vec_add1 (wrk->pending_resets, tc->c_c_index);
      tcp_disconnect_pending_on (tc);
    }
}

/**
 * Handle reset packet
 *
 * Programs disconnect/reset notification that should be sent
 * later by calling @ref tcp_handle_disconnects
 */
static void
tcp_rcv_rst (tcp_worker_ctx_t * wrk, tcp_connection_t * tc)
{
  TCP_EVT (TCP_EVT_RST_RCVD, tc);
  switch (tc->state)
    {
    case TCP_STATE_SYN_RCVD:
      tcp_program_reset_ntf (wrk, tc);
      tcp_connection_set_state (tc, TCP_STATE_CLOSED);
      break;
    case TCP_STATE_SYN_SENT:
      /* Do not program ntf because the connection is half-open */
      tcp_handle_rst (tc);
      break;
    case TCP_STATE_ESTABLISHED:
      tcp_connection_timers_reset (tc);
      tcp_cong_recovery_off (tc);
      tcp_program_reset_ntf (wrk, tc);
      tcp_connection_set_state (tc, TCP_STATE_CLOSED);
      tcp_program_cleanup (wrk, tc);
      break;
    case TCP_STATE_CLOSE_WAIT:
    case TCP_STATE_FIN_WAIT_1:
    case TCP_STATE_FIN_WAIT_2:
    case TCP_STATE_CLOSING:
    case TCP_STATE_LAST_ACK:
      tcp_connection_timers_reset (tc);
      tcp_cong_recovery_off (tc);
      tcp_program_reset_ntf (wrk, tc);
      /* Make sure we mark the session as closed. In some states we may
       * be still trying to send data */
      tcp_connection_set_state (tc, TCP_STATE_CLOSED);
      tcp_program_cleanup (wrk, tc);
      break;
    case TCP_STATE_CLOSED:
    case TCP_STATE_TIME_WAIT:
      break;
    default:
      TCP_DBG ("reset state: %u", tc->state);
    }
}

/**
 * Validate incoming segment as per RFC793 p. 69 and RFC1323 p. 19
 *
 * It first verifies if segment has a wrapped sequence number (PAWS) and then
 * does the processing associated to the first four steps (ignoring security
 * and precedence): sequence number, rst bit and syn bit checks.
 *
 * @return 0 if segments passes validation.
 */
static int
tcp_segment_validate (tcp_worker_ctx_t * wrk, tcp_connection_t * tc0,
		      vlib_buffer_t * b0, tcp_header_t * th0, u32 * error0)
{
  /* We could get a burst of RSTs interleaved with acks */
  if (PREDICT_FALSE (tc0->state == TCP_STATE_CLOSED))
    {
      tcp_send_reset (tc0);
      *error0 = TCP_ERROR_CONNECTION_CLOSED;
      goto error;
    }

  if (PREDICT_FALSE (!tcp_ack (th0) && !tcp_rst (th0) && !tcp_syn (th0)))
    {
      *error0 = TCP_ERROR_SEGMENT_INVALID;
      goto error;
    }

  if (PREDICT_FALSE (tcp_options_parse (th0, &tc0->rcv_opts, 0)))
    {
      *error0 = TCP_ERROR_OPTIONS;
      goto error;
    }

  if (PREDICT_FALSE (tcp_segment_check_paws (tc0)))
    {
      *error0 = TCP_ERROR_PAWS;
      TCP_EVT (TCP_EVT_PAWS_FAIL, tc0, vnet_buffer (b0)->tcp.seq_number,
	       vnet_buffer (b0)->tcp.seq_end);

      /* If it just so happens that a segment updates tsval_recent for a
       * segment over 24 days old, invalidate tsval_recent. */
      if (timestamp_lt (tc0->tsval_recent_age + TCP_PAWS_IDLE,
			tcp_time_now_w_thread (tc0->c_thread_index)))
	{
	  tc0->tsval_recent = tc0->rcv_opts.tsval;
	  clib_warning ("paws failed: 24-day old segment");
	}
      /* Drop after ack if not rst. Resets can fail paws check as per
       * RFC 7323 sec. 5.2: When an <RST> segment is received, it MUST NOT
       * be subjected to the PAWS check by verifying an acceptable value in
       * SEG.TSval */
      else if (!tcp_rst (th0))
	{
	  tcp_program_ack (tc0);
	  TCP_EVT (TCP_EVT_DUPACK_SENT, tc0, vnet_buffer (b0)->tcp);
	  goto error;
	}
    }

  /* 1st: check sequence number */
  if (!tcp_segment_in_rcv_wnd (tc0, vnet_buffer (b0)->tcp.seq_number,
			       vnet_buffer (b0)->tcp.seq_end))
    {
      /* SYN/SYN-ACK retransmit */
      if (tcp_syn (th0)
	  && vnet_buffer (b0)->tcp.seq_number == tc0->rcv_nxt - 1)
	{
	  tcp_options_parse (th0, &tc0->rcv_opts, 1);
	  if (tc0->state == TCP_STATE_SYN_RCVD)
	    {
	      tcp_send_synack (tc0);
	      TCP_EVT (TCP_EVT_SYN_RCVD, tc0, 0);
	      *error0 = TCP_ERROR_SYNS_RCVD;
	    }
	  else
	    {
	      tcp_program_ack (tc0);
	      TCP_EVT (TCP_EVT_SYNACK_RCVD, tc0);
	      *error0 = TCP_ERROR_SYN_ACKS_RCVD;
	    }
	  goto error;
	}

      /* If our window is 0 and the packet is in sequence, let it pass
       * through for ack processing. It should be dropped later. */
      if (tc0->rcv_wnd < tc0->snd_mss
	  && tc0->rcv_nxt == vnet_buffer (b0)->tcp.seq_number)
	goto check_reset;

      /* If we entered recovery and peer did so as well, there's a chance that
       * dup acks won't be acceptable on either end because seq_end may be less
       * than rcv_las. This can happen if acks are lost in both directions. */
      if (tcp_in_recovery (tc0)
	  && seq_geq (vnet_buffer (b0)->tcp.seq_number,
		      tc0->rcv_las - tc0->rcv_wnd)
	  && seq_leq (vnet_buffer (b0)->tcp.seq_end,
		      tc0->rcv_nxt + tc0->rcv_wnd))
	goto check_reset;

      *error0 = TCP_ERROR_RCV_WND;

      /* If we advertised a zero rcv_wnd and the segment is in the past or the
       * next one that we expect, it is probably a window probe */
      if ((tc0->flags & TCP_CONN_ZERO_RWND_SENT)
	  && seq_lt (vnet_buffer (b0)->tcp.seq_end,
		     tc0->rcv_las + tc0->rcv_opts.mss))
	*error0 = TCP_ERROR_ZERO_RWND;

      tc0->errors.below_data_wnd += seq_lt (vnet_buffer (b0)->tcp.seq_end,
					    tc0->rcv_las);

      /* If not RST, send dup ack */
      if (!tcp_rst (th0))
	{
	  tcp_program_dupack (tc0);
	  TCP_EVT (TCP_EVT_DUPACK_SENT, tc0, vnet_buffer (b0)->tcp);
	}
      goto error;

    check_reset:
      ;
    }

  /* 2nd: check the RST bit */
  if (PREDICT_FALSE (tcp_rst (th0)))
    {
      tcp_rcv_rst (wrk, tc0);
      *error0 = TCP_ERROR_RST_RCVD;
      goto error;
    }

  /* 3rd: check security and precedence (skip) */

  /* 4th: check the SYN bit (in window) */
  if (PREDICT_FALSE (tcp_syn (th0)))
    {
      /* As per RFC5961 send challenge ack instead of reset */
      tcp_program_ack (tc0);
      *error0 = TCP_ERROR_SPURIOUS_SYN;
      goto error;
    }

  /* If segment in window, save timestamp */
  tcp_update_timestamp (tc0, vnet_buffer (b0)->tcp.seq_number,
			vnet_buffer (b0)->tcp.seq_end);
  return 0;

error:
  return -1;
}

always_inline int
tcp_rcv_ack_no_cc (tcp_connection_t * tc, vlib_buffer_t * b, u32 * error)
{
  /* SND.UNA =< SEG.ACK =< SND.NXT */
  if (!(seq_leq (tc->snd_una, vnet_buffer (b)->tcp.ack_number)
	&& seq_leq (vnet_buffer (b)->tcp.ack_number, tc->snd_nxt)))
    {
      if (seq_leq (vnet_buffer (b)->tcp.ack_number, tc->snd_una_max)
	  && seq_gt (vnet_buffer (b)->tcp.ack_number, tc->snd_una))
	{
	  tc->snd_nxt = vnet_buffer (b)->tcp.ack_number;
	  goto acceptable;
	}
      *error = TCP_ERROR_ACK_INVALID;
      return -1;
    }

acceptable:
  tc->bytes_acked = vnet_buffer (b)->tcp.ack_number - tc->snd_una;
  tc->snd_una = vnet_buffer (b)->tcp.ack_number;
  *error = TCP_ERROR_ACK_OK;
  return 0;
}

/**
 * Compute smoothed RTT as per VJ's '88 SIGCOMM and RFC6298
 *
 * Note that although the original article, srtt and rttvar are scaled
 * to minimize round-off errors, here we don't. Instead, we rely on
 * better precision time measurements.
 */
static void
tcp_estimate_rtt (tcp_connection_t * tc, u32 mrtt)
{
  int err, diff;

  if (tc->srtt != 0)
    {
      err = mrtt - tc->srtt;

      /* XXX Drop in RTT results in RTTVAR increase and bigger RTO.
       * The increase should be bound */
      tc->srtt = clib_max ((int) tc->srtt + (err >> 3), 1);
      diff = (clib_abs (err) - (int) tc->rttvar) >> 2;
      tc->rttvar = clib_max ((int) tc->rttvar + diff, 1);
    }
  else
    {
      /* First measurement. */
      tc->srtt = mrtt;
      tc->rttvar = mrtt >> 1;
    }
}

static inline void
tcp_estimate_rtt_us (tcp_connection_t * tc, f64 mrtt)
{
  tc->mrtt_us = tc->mrtt_us + (mrtt - tc->mrtt_us) * 0.125;
}

/**
 * Update rtt estimate
 *
 * We have potentially three sources of rtt measurements:
 *
 * TSOPT	difference between current and echoed timestamp. It has ms
 * 	    	precision and can be computed per ack
 * ACK timing	one sequence number is tracked per rtt with us (micro second)
 * 		precision.
 * rate sample	if enabled, all outstanding bytes are tracked with us
 * 		precision. Every ack and sack are a rtt sample
 *
 * Middle boxes are known to fiddle with TCP options so we give higher
 * priority to ACK timing.
 *
 * For now, rate sample rtts are only used under congestion.
 */
static int
tcp_update_rtt (tcp_connection_t * tc, tcp_rate_sample_t * rs, u32 ack)
{
  u32 mrtt = 0;

  /* Karn's rule, part 1. Don't use retransmitted segments to estimate
   * RTT because they're ambiguous. */
  if (tcp_in_cong_recovery (tc))
    {
      /* Accept rtt estimates for samples that have not been retransmitted */
      if (!(tc->cfg_flags & TCP_CFG_F_RATE_SAMPLE)
	  || (rs->flags & TCP_BTS_IS_RXT))
	goto done;
      if (rs->rtt_time)
	tcp_estimate_rtt_us (tc, rs->rtt_time);
      mrtt = rs->rtt_time * THZ;
      goto estimate_rtt;
    }

  if (tc->rtt_ts && seq_geq (ack, tc->rtt_seq))
    {
      f64 sample = tcp_time_now_us (tc->c_thread_index) - tc->rtt_ts;
      tcp_estimate_rtt_us (tc, sample);
      mrtt = clib_max ((u32) (sample * THZ), 1);
      /* Allow measuring of a new RTT */
      tc->rtt_ts = 0;
    }
  /* As per RFC7323 TSecr can be used for RTTM only if the segment advances
   * snd_una, i.e., the left side of the send window:
   * seq_lt (tc->snd_una, ack). This is a condition for calling update_rtt */
  else if (tcp_opts_tstamp (&tc->rcv_opts) && tc->rcv_opts.tsecr)
    {
      u32 now = tcp_tstamp (tc);
      mrtt = clib_max (now - tc->rcv_opts.tsecr, 1);
    }

estimate_rtt:

  /* Ignore dubious measurements */
  if (mrtt == 0 || mrtt > TCP_RTT_MAX)
    goto done;

  tcp_estimate_rtt (tc, mrtt);

done:

  /* If we got here something must've been ACKed so make sure boff is 0,
   * even if mrtt is not valid since we update the rto lower */
  tc->rto_boff = 0;
  tcp_update_rto (tc);

  return 0;
}

static void
tcp_estimate_initial_rtt (tcp_connection_t * tc)
{
  u8 thread_index = vlib_num_workers ()? 1 : 0;
  int mrtt;

  if (tc->rtt_ts)
    {
      tc->mrtt_us = tcp_time_now_us (thread_index) - tc->rtt_ts;
      tc->mrtt_us = clib_max (tc->mrtt_us, 0.0001);
      mrtt = clib_max ((u32) (tc->mrtt_us * THZ), 1);
      tc->rtt_ts = 0;
    }
  else
    {
      mrtt = tcp_time_now_w_thread (thread_index) - tc->rcv_opts.tsecr;
      mrtt = clib_max (mrtt, 1);
      /* Due to retransmits we don't know the initial mrtt */
      if (tc->rto_boff && mrtt > 1 * THZ)
	mrtt = 1 * THZ;
      tc->mrtt_us = (f64) mrtt *TCP_TICK;
    }

  if (mrtt > 0 && mrtt < TCP_RTT_MAX)
    tcp_estimate_rtt (tc, mrtt);
  tcp_update_rto (tc);
}

/**
 * Dequeue bytes for connections that have received acks in last burst
 */
static void
tcp_handle_postponed_dequeues (tcp_worker_ctx_t * wrk)
{
  u32 thread_index = wrk->vm->thread_index;
  u32 *pending_deq_acked;
  tcp_connection_t *tc;
  int i;

  if (!vec_len (wrk->pending_deq_acked))
    return;

  pending_deq_acked = wrk->pending_deq_acked;
  for (i = 0; i < vec_len (pending_deq_acked); i++)
    {
      tc = tcp_connection_get (pending_deq_acked[i], thread_index);
      tc->flags &= ~TCP_CONN_DEQ_PENDING;

      if (PREDICT_FALSE (!tc->burst_acked))
	continue;

      /* Dequeue the newly ACKed bytes */
      session_tx_fifo_dequeue_drop (&tc->connection, tc->burst_acked);
      tcp_validate_txf_size (tc, tc->snd_una_max - tc->snd_una);

      if (PREDICT_FALSE (tc->flags & TCP_CONN_PSH_PENDING))
	{
	  if (seq_leq (tc->psh_seq, tc->snd_una))
	    tc->flags &= ~TCP_CONN_PSH_PENDING;
	}

      if (tcp_is_descheduled (tc))
	tcp_reschedule (tc);

      /* If everything has been acked, stop retransmit timer
       * otherwise update. */
      tcp_retransmit_timer_update (&wrk->timer_wheel, tc);

      /* Update pacer based on our new cwnd estimate */
      tcp_connection_tx_pacer_update (tc);

      tc->burst_acked = 0;
    }
  _vec_len (wrk->pending_deq_acked) = 0;
}

static void
tcp_program_dequeue (tcp_worker_ctx_t * wrk, tcp_connection_t * tc)
{
  if (!(tc->flags & TCP_CONN_DEQ_PENDING))
    {
      vec_add1 (wrk->pending_deq_acked, tc->c_c_index);
      tc->flags |= TCP_CONN_DEQ_PENDING;
    }
  tc->burst_acked += tc->bytes_acked;
}

/**
 * Try to update snd_wnd based on feedback received from peer.
 *
 * If successful, and new window is 'effectively' 0, activate persist
 * timer.
 */
static void
tcp_update_snd_wnd (tcp_connection_t * tc, u32 seq, u32 ack, u32 snd_wnd)
{
  /* If (SND.WL1 < SEG.SEQ or (SND.WL1 = SEG.SEQ and SND.WL2 =< SEG.ACK)), set
   * SND.WND <- SEG.WND, set SND.WL1 <- SEG.SEQ, and set SND.WL2 <- SEG.ACK */
  if (seq_lt (tc->snd_wl1, seq)
      || (tc->snd_wl1 == seq && seq_leq (tc->snd_wl2, ack)))
    {
      tc->snd_wnd = snd_wnd;
      tc->snd_wl1 = seq;
      tc->snd_wl2 = ack;
      TCP_EVT (TCP_EVT_SND_WND, tc);

      if (PREDICT_FALSE (tc->snd_wnd < tc->snd_mss))
	{
	  /* Set persist timer if not set and we just got 0 wnd */
	  if (!tcp_timer_is_active (tc, TCP_TIMER_PERSIST)
	      && !tcp_timer_is_active (tc, TCP_TIMER_RETRANSMIT))
	    {
	      tcp_worker_ctx_t *wrk = tcp_get_worker (tc->c_thread_index);
	      tcp_persist_timer_set (&wrk->timer_wheel, tc);
	    }
	}
      else
	{
	  if (PREDICT_FALSE (tcp_timer_is_active (tc, TCP_TIMER_PERSIST)))
	    {
	      tcp_worker_ctx_t *wrk = tcp_get_worker (tc->c_thread_index);
	      tcp_persist_timer_reset (&wrk->timer_wheel, tc);
	    }

	  if (PREDICT_FALSE (tcp_is_descheduled (tc)))
	    tcp_reschedule (tc);

	  if (PREDICT_FALSE (!tcp_in_recovery (tc) && tc->rto_boff > 0))
	    {
	      tc->rto_boff = 0;
	      tcp_update_rto (tc);
	    }
	}
    }
}

/**
 * Init loss recovery/fast recovery.
 *
 * Triggered by dup acks as opposed to timer timeout. Note that cwnd is
 * updated in @ref tcp_cc_handle_event after fast retransmit
 */
static void
tcp_cc_init_congestion (tcp_connection_t * tc)
{
  tcp_fastrecovery_on (tc);
  tc->snd_congestion = tc->snd_nxt;
  tc->cwnd_acc_bytes = 0;
  tc->snd_rxt_bytes = 0;
  tc->rxt_delivered = 0;
  tc->prr_delivered = 0;
  tc->prr_start = tc->snd_una;
  tc->prev_ssthresh = tc->ssthresh;
  tc->prev_cwnd = tc->cwnd;

  tc->snd_rxt_ts = tcp_tstamp (tc);
  tcp_cc_congestion (tc);

  /* Post retransmit update cwnd to ssthresh and account for the
   * three segments that have left the network and should've been
   * buffered at the receiver XXX */
  if (!tcp_opts_sack_permitted (&tc->rcv_opts))
    tc->cwnd += 3 * tc->snd_mss;

  tc->fr_occurences += 1;
  TCP_EVT (TCP_EVT_CC_EVT, tc, 4);
}

static void
tcp_cc_congestion_undo (tcp_connection_t * tc)
{
  tc->cwnd = tc->prev_cwnd;
  tc->ssthresh = tc->prev_ssthresh;
  tcp_cc_undo_recovery (tc);
  ASSERT (tc->rto_boff == 0);
  TCP_EVT (TCP_EVT_CC_EVT, tc, 5);
}

static inline u8
tcp_cc_is_spurious_timeout_rxt (tcp_connection_t * tc)
{
  return (tcp_in_recovery (tc) && tc->rto_boff == 1
	  && tc->snd_rxt_ts
	  && tcp_opts_tstamp (&tc->rcv_opts)
	  && timestamp_lt (tc->rcv_opts.tsecr, tc->snd_rxt_ts));
}

static inline u8
tcp_cc_is_spurious_retransmit (tcp_connection_t * tc)
{
  return (tcp_cc_is_spurious_timeout_rxt (tc));
}

static inline u8
tcp_should_fastrecover_sack (tcp_connection_t * tc)
{
  return (tc->sack_sb.lost_bytes
	  || ((TCP_DUPACK_THRESHOLD - 1) * tc->snd_mss
	      < tc->sack_sb.sacked_bytes));
}

static inline u8
tcp_should_fastrecover (tcp_connection_t * tc, u8 has_sack)
{
  if (!has_sack)
    {
      /* If of of the two conditions lower hold, reset dupacks because
       * we're probably after timeout (RFC6582 heuristics).
       * If Cumulative ack does not cover more than congestion threshold,
       * and:
       * 1) The following doesn't hold: The congestion window is greater
       *    than SMSS bytes and the difference between highest_ack
       *    and prev_highest_ack is at most 4*SMSS bytes
       * 2) Echoed timestamp in the last non-dup ack does not equal the
       *    stored timestamp
       */
      if (seq_leq (tc->snd_una, tc->snd_congestion)
	  && ((!(tc->cwnd > tc->snd_mss
		 && tc->bytes_acked <= 4 * tc->snd_mss))
	      || (tc->rcv_opts.tsecr != tc->tsecr_last_ack)))
	{
	  tc->rcv_dupacks = 0;
	  return 0;
	}
    }
  return ((tc->rcv_dupacks == TCP_DUPACK_THRESHOLD)
	  || tcp_should_fastrecover_sack (tc));
}

static int
tcp_cc_recover (tcp_connection_t * tc)
{
  sack_scoreboard_hole_t *hole;
  u8 is_spurious = 0;

  ASSERT (tcp_in_cong_recovery (tc));

  if (tcp_cc_is_spurious_retransmit (tc))
    {
      tcp_cc_congestion_undo (tc);
      is_spurious = 1;
    }

  tcp_connection_tx_pacer_reset (tc, tc->cwnd, 0 /* start bucket */ );
  tc->rcv_dupacks = 0;

  /* Previous recovery left us congested. Continue sending as part
   * of the current recovery event with an updated snd_congestion */
  if (tc->sack_sb.sacked_bytes)
    {
      tc->snd_congestion = tc->snd_nxt;
      tcp_program_retransmit (tc);
      return is_spurious;
    }

  tc->rxt_delivered = 0;
  tc->snd_rxt_bytes = 0;
  tc->snd_rxt_ts = 0;
  tc->prr_delivered = 0;
  tc->rtt_ts = 0;
  tc->flags &= ~TCP_CONN_RXT_PENDING;

  hole = scoreboard_first_hole (&tc->sack_sb);
  if (hole && hole->start == tc->snd_una && hole->end == tc->snd_nxt)
    scoreboard_clear (&tc->sack_sb);

  if (!tcp_in_recovery (tc) && !is_spurious)
    tcp_cc_recovered (tc);

  tcp_fastrecovery_off (tc);
  tcp_fastrecovery_first_off (tc);
  tcp_recovery_off (tc);
  TCP_EVT (TCP_EVT_CC_EVT, tc, 3);

  ASSERT (tc->rto_boff == 0);
  ASSERT (!tcp_in_cong_recovery (tc));
  ASSERT (tcp_scoreboard_is_sane_post_recovery (tc));

  return is_spurious;
}

static void
tcp_cc_update (tcp_connection_t * tc, tcp_rate_sample_t * rs)
{
  ASSERT (!tcp_in_cong_recovery (tc) || tcp_is_lost_fin (tc));

  /* Congestion avoidance */
  tcp_cc_rcv_ack (tc, rs);

  /* If a cumulative ack, make sure dupacks is 0 */
  tc->rcv_dupacks = 0;

  /* When dupacks hits the threshold we only enter fast retransmit if
   * cumulative ack covers more than snd_congestion. Should snd_una
   * wrap this test may fail under otherwise valid circumstances.
   * Therefore, proactively update snd_congestion when wrap detected. */
  if (PREDICT_FALSE
      (seq_leq (tc->snd_congestion, tc->snd_una - tc->bytes_acked)
       && seq_gt (tc->snd_congestion, tc->snd_una)))
    tc->snd_congestion = tc->snd_una - 1;
}

/**
 * One function to rule them all ... and in the darkness bind them
 */
static void
tcp_cc_handle_event (tcp_connection_t * tc, tcp_rate_sample_t * rs,
		     u32 is_dack)
{
  u8 has_sack = tcp_opts_sack_permitted (&tc->rcv_opts);

  /* If reneging, wait for timer based retransmits */
  if (PREDICT_FALSE (tcp_is_lost_fin (tc) || tc->sack_sb.is_reneging))
    return;

  /*
   * If not in recovery, figure out if we should enter
   */
  if (!tcp_in_cong_recovery (tc))
    {
      ASSERT (is_dack);

      tc->rcv_dupacks++;
      TCP_EVT (TCP_EVT_DUPACK_RCVD, tc, 1);
      tcp_cc_rcv_cong_ack (tc, TCP_CC_DUPACK, rs);

      if (tcp_should_fastrecover (tc, has_sack))
	{
	  tcp_cc_init_congestion (tc);

	  if (has_sack)
	    scoreboard_init_rxt (&tc->sack_sb, tc->snd_una);

	  tcp_connection_tx_pacer_reset (tc, tc->cwnd, 0 /* start bucket */ );
	  tcp_program_retransmit (tc);
	}

      return;
    }

  /*
   * Already in recovery
   */

  /*
   * Process (re)transmit feedback. Output path uses this to decide how much
   * more data to release into the network
   */
  if (has_sack)
    {
      if (!tc->bytes_acked && tc->sack_sb.rxt_sacked)
	tcp_fastrecovery_first_on (tc);

      tc->rxt_delivered += tc->sack_sb.rxt_sacked;
      tc->prr_delivered += tc->bytes_acked + tc->sack_sb.last_sacked_bytes
	- tc->sack_sb.last_bytes_delivered;
    }
  else
    {
      if (is_dack)
	{
	  tc->rcv_dupacks += 1;
	  TCP_EVT (TCP_EVT_DUPACK_RCVD, tc, 1);
	}
      tc->rxt_delivered = clib_min (tc->rxt_delivered + tc->bytes_acked,
				    tc->snd_rxt_bytes);
      if (is_dack)
	tc->prr_delivered += clib_min (tc->snd_mss,
				       tc->snd_nxt - tc->snd_una);
      else
	tc->prr_delivered += tc->bytes_acked - clib_min (tc->bytes_acked,
							 tc->snd_mss *
							 tc->rcv_dupacks);

      /* If partial ack, assume that the first un-acked segment was lost */
      if (tc->bytes_acked || tc->rcv_dupacks == TCP_DUPACK_THRESHOLD)
	tcp_fastrecovery_first_on (tc);
    }

  /*
   * See if we can exit and stop retransmitting
   */
  if (seq_geq (tc->snd_una, tc->snd_congestion))
    {
      /* If spurious return, we've already updated everything */
      if (tcp_cc_recover (tc))
	{
	  tc->tsecr_last_ack = tc->rcv_opts.tsecr;
	  return;
	}

      /* Treat as congestion avoidance ack */
      tcp_cc_rcv_ack (tc, rs);
      return;
    }

  tcp_program_retransmit (tc);

  /*
   * Notify cc of the event
   */

  if (!tc->bytes_acked)
    {
      tcp_cc_rcv_cong_ack (tc, TCP_CC_DUPACK, rs);
      return;
    }

  /* RFC6675: If the incoming ACK is a cumulative acknowledgment,
   * reset dupacks to 0. Also needed if in congestion recovery */
  tc->rcv_dupacks = 0;

  if (tcp_in_recovery (tc))
    tcp_cc_rcv_ack (tc, rs);
  else
    tcp_cc_rcv_cong_ack (tc, TCP_CC_PARTIALACK, rs);
}

static void
tcp_handle_old_ack (tcp_connection_t * tc, tcp_rate_sample_t * rs)
{
  if (!tcp_in_cong_recovery (tc))
    return;

  if (tcp_opts_sack_permitted (&tc->rcv_opts))
    tcp_rcv_sacks (tc, tc->snd_una);

  tc->bytes_acked = 0;

  if (tc->cfg_flags & TCP_CFG_F_RATE_SAMPLE)
    tcp_bt_sample_delivery_rate (tc, rs);

  tcp_cc_handle_event (tc, rs, 1);
}

/**
 * Check if duplicate ack as per RFC5681 Sec. 2
 */
always_inline u8
tcp_ack_is_dupack (tcp_connection_t * tc, vlib_buffer_t * b, u32 prev_snd_wnd,
		   u32 prev_snd_una)
{
  return ((vnet_buffer (b)->tcp.ack_number == prev_snd_una)
	  && seq_gt (tc->snd_nxt, tc->snd_una)
	  && (vnet_buffer (b)->tcp.seq_end == vnet_buffer (b)->tcp.seq_number)
	  && (prev_snd_wnd == tc->snd_wnd));
}

/**
 * Checks if ack is a congestion control event.
 */
static u8
tcp_ack_is_cc_event (tcp_connection_t * tc, vlib_buffer_t * b,
		     u32 prev_snd_wnd, u32 prev_snd_una, u8 * is_dack)
{
  /* Check if ack is duplicate. Per RFC 6675, ACKs that SACK new data are
   * defined to be 'duplicate' as well */
  *is_dack = tc->sack_sb.last_sacked_bytes
    || tcp_ack_is_dupack (tc, b, prev_snd_wnd, prev_snd_una);

  return (*is_dack || tcp_in_cong_recovery (tc));
}

/**
 * Process incoming ACK
 */
static int
tcp_rcv_ack (tcp_worker_ctx_t * wrk, tcp_connection_t * tc, vlib_buffer_t * b,
	     tcp_header_t * th, u32 * error)
{
  u32 prev_snd_wnd, prev_snd_una;
  tcp_rate_sample_t rs = { 0 };
  u8 is_dack;

  TCP_EVT (TCP_EVT_CC_STAT, tc);

  /* If the ACK acks something not yet sent (SEG.ACK > SND.NXT) */
  if (PREDICT_FALSE (seq_gt (vnet_buffer (b)->tcp.ack_number, tc->snd_nxt)))
    {
      /* We've probably entered recovery and the peer still has some
       * of the data we've sent. Update snd_nxt and accept the ack */
      if (seq_leq (vnet_buffer (b)->tcp.ack_number, tc->snd_una_max)
	  && seq_gt (vnet_buffer (b)->tcp.ack_number, tc->snd_una))
	{
	  tc->snd_nxt = vnet_buffer (b)->tcp.ack_number;
	  goto process_ack;
	}

      tc->errors.above_ack_wnd += 1;
      *error = TCP_ERROR_ACK_FUTURE;
      TCP_EVT (TCP_EVT_ACK_RCV_ERR, tc, 0, vnet_buffer (b)->tcp.ack_number);
      return -1;
    }

  /* If old ACK, probably it's an old dupack */
  if (PREDICT_FALSE (seq_lt (vnet_buffer (b)->tcp.ack_number, tc->snd_una)))
    {
      tc->errors.below_ack_wnd += 1;
      *error = TCP_ERROR_ACK_OLD;
      TCP_EVT (TCP_EVT_ACK_RCV_ERR, tc, 1, vnet_buffer (b)->tcp.ack_number);

      if (seq_lt (vnet_buffer (b)->tcp.ack_number, tc->snd_una - tc->rcv_wnd))
	return -1;

      tcp_handle_old_ack (tc, &rs);

      /* Don't drop yet */
      return 0;
    }

process_ack:

  /*
   * Looks okay, process feedback
   */

  if (tcp_opts_sack_permitted (&tc->rcv_opts))
    tcp_rcv_sacks (tc, vnet_buffer (b)->tcp.ack_number);

  prev_snd_wnd = tc->snd_wnd;
  prev_snd_una = tc->snd_una;
  tcp_update_snd_wnd (tc, vnet_buffer (b)->tcp.seq_number,
		      vnet_buffer (b)->tcp.ack_number,
		      clib_net_to_host_u16 (th->window) << tc->snd_wscale);
  tc->bytes_acked = vnet_buffer (b)->tcp.ack_number - tc->snd_una;
  tc->snd_una = vnet_buffer (b)->tcp.ack_number;
  tcp_validate_txf_size (tc, tc->bytes_acked);

  if (tc->cfg_flags & TCP_CFG_F_RATE_SAMPLE)
    tcp_bt_sample_delivery_rate (tc, &rs);

  if (tc->bytes_acked + tc->sack_sb.last_sacked_bytes)
    {
      tcp_update_rtt (tc, &rs, vnet_buffer (b)->tcp.ack_number);
      if (tc->bytes_acked)
	tcp_program_dequeue (wrk, tc);
    }

  TCP_EVT (TCP_EVT_ACK_RCVD, tc);

  /*
   * Check if we have congestion event
   */

  if (tcp_ack_is_cc_event (tc, b, prev_snd_wnd, prev_snd_una, &is_dack))
    {
      tcp_cc_handle_event (tc, &rs, is_dack);
      tc->dupacks_in += is_dack;
      if (!tcp_in_cong_recovery (tc))
	{
	  *error = TCP_ERROR_ACK_OK;
	  return 0;
	}
      *error = TCP_ERROR_ACK_DUP;
      if (vnet_buffer (b)->tcp.data_len || tcp_is_fin (th))
	return 0;
      return -1;
    }

  /*
   * Update congestion control (slow start/congestion avoidance)
   */
  tcp_cc_update (tc, &rs);
  *error = TCP_ERROR_ACK_OK;
  return 0;
}

static void
tcp_program_disconnect (tcp_worker_ctx_t * wrk, tcp_connection_t * tc)
{
  if (!tcp_disconnect_pending (tc))
    {
      vec_add1 (wrk->pending_disconnects, tc->c_c_index);
      tcp_disconnect_pending_on (tc);
    }
}

static void
tcp_handle_disconnects (tcp_worker_ctx_t * wrk)
{
  u32 thread_index, *pending_disconnects, *pending_resets;
  tcp_connection_t *tc;
  int i;

  if (vec_len (wrk->pending_disconnects))
    {
      thread_index = wrk->vm->thread_index;
      pending_disconnects = wrk->pending_disconnects;
      for (i = 0; i < vec_len (pending_disconnects); i++)
	{
	  tc = tcp_connection_get (pending_disconnects[i], thread_index);
	  tcp_disconnect_pending_off (tc);
	  session_transport_closing_notify (&tc->connection);
	}
      _vec_len (wrk->pending_disconnects) = 0;
    }

  if (vec_len (wrk->pending_resets))
    {
      thread_index = wrk->vm->thread_index;
      pending_resets = wrk->pending_resets;
      for (i = 0; i < vec_len (pending_resets); i++)
	{
	  tc = tcp_connection_get (pending_resets[i], thread_index);
	  tcp_disconnect_pending_off (tc);
	  tcp_handle_rst (tc);
	}
      _vec_len (wrk->pending_resets) = 0;
    }
}

static void
tcp_rcv_fin (tcp_worker_ctx_t * wrk, tcp_connection_t * tc, vlib_buffer_t * b,
	     u32 * error)
{
  /* Reject out-of-order fins */
  if (vnet_buffer (b)->tcp.seq_end != tc->rcv_nxt)
    return;

  /* Account for the FIN and send ack */
  tc->rcv_nxt += 1;
  tc->flags |= TCP_CONN_FINRCVD;
  tcp_program_ack (tc);
  /* Enter CLOSE-WAIT and notify session. To avoid lingering
   * in CLOSE-WAIT, set timer (reuse WAITCLOSE). */
  tcp_connection_set_state (tc, TCP_STATE_CLOSE_WAIT);
  tcp_program_disconnect (wrk, tc);
  tcp_timer_update (&wrk->timer_wheel, tc, TCP_TIMER_WAITCLOSE,
		    tcp_cfg.closewait_time);
  TCP_EVT (TCP_EVT_FIN_RCVD, tc);
  *error = TCP_ERROR_FIN_RCVD;
}

/** Enqueue data for delivery to application */
static int
tcp_session_enqueue_data (tcp_connection_t * tc, vlib_buffer_t * b,
			  u16 data_len)
{
  int written, error = TCP_ERROR_ENQUEUED;

  ASSERT (seq_geq (vnet_buffer (b)->tcp.seq_number, tc->rcv_nxt));
  ASSERT (data_len);
  written = session_enqueue_stream_connection (&tc->connection, b, 0,
					       1 /* queue event */ , 1);
  tc->bytes_in += written;

  TCP_EVT (TCP_EVT_INPUT, tc, 0, data_len, written);

  /* Update rcv_nxt */
  if (PREDICT_TRUE (written == data_len))
    {
      tc->rcv_nxt += written;
    }
  /* If more data written than expected, account for out-of-order bytes. */
  else if (written > data_len)
    {
      tc->rcv_nxt += written;
      TCP_EVT (TCP_EVT_CC_INPUT, tc, data_len, written);
    }
  else if (written > 0)
    {
      /* We've written something but FIFO is probably full now */
      tc->rcv_nxt += written;
      error = TCP_ERROR_PARTIALLY_ENQUEUED;
    }
  else
    {
      /* Packet made it through for ack processing */
      if (tc->rcv_wnd < tc->snd_mss)
	return TCP_ERROR_ZERO_RWND;

      return TCP_ERROR_FIFO_FULL;
    }

  /* Update SACK list if need be */
  if (tcp_opts_sack_permitted (&tc->rcv_opts))
    {
      /* Remove SACK blocks that have been delivered */
      tcp_update_sack_list (tc, tc->rcv_nxt, tc->rcv_nxt);
    }

  return error;
}

/** Enqueue out-of-order data */
static int
tcp_session_enqueue_ooo (tcp_connection_t * tc, vlib_buffer_t * b,
			 u16 data_len)
{
  session_t *s0;
  int rv, offset;

  ASSERT (seq_gt (vnet_buffer (b)->tcp.seq_number, tc->rcv_nxt));
  ASSERT (data_len);

  /* Enqueue out-of-order data with relative offset */
  rv = session_enqueue_stream_connection (&tc->connection, b,
					  vnet_buffer (b)->tcp.seq_number -
					  tc->rcv_nxt, 0 /* queue event */ ,
					  0);

  /* Nothing written */
  if (rv)
    {
      TCP_EVT (TCP_EVT_INPUT, tc, 1, data_len, 0);
      return TCP_ERROR_FIFO_FULL;
    }

  TCP_EVT (TCP_EVT_INPUT, tc, 1, data_len, data_len);
  tc->bytes_in += data_len;

  /* Update SACK list if in use */
  if (tcp_opts_sack_permitted (&tc->rcv_opts))
    {
      ooo_segment_t *newest;
      u32 start, end;

      s0 = session_get (tc->c_s_index, tc->c_thread_index);

      /* Get the newest segment from the fifo */
      newest = svm_fifo_newest_ooo_segment (s0->rx_fifo);
      if (newest)
	{
	  offset = ooo_segment_offset_prod (s0->rx_fifo, newest);
	  ASSERT (offset <= vnet_buffer (b)->tcp.seq_number - tc->rcv_nxt);
	  start = tc->rcv_nxt + offset;
	  end = start + ooo_segment_length (s0->rx_fifo, newest);
	  tcp_update_sack_list (tc, start, end);
	  svm_fifo_newest_ooo_segment_reset (s0->rx_fifo);
	  TCP_EVT (TCP_EVT_CC_SACKS, tc);
	}
    }

  return TCP_ERROR_ENQUEUED_OOO;
}

static int
tcp_buffer_discard_bytes (vlib_buffer_t * b, u32 n_bytes_to_drop)
{
  u32 discard, first = b->current_length;
  vlib_main_t *vm = vlib_get_main ();

  /* Handle multi-buffer segments */
  if (n_bytes_to_drop > b->current_length)
    {
      if (!(b->flags & VLIB_BUFFER_NEXT_PRESENT))
	return -1;
      do
	{
	  discard = clib_min (n_bytes_to_drop, b->current_length);
	  vlib_buffer_advance (b, discard);
	  b = vlib_get_buffer (vm, b->next_buffer);
	  n_bytes_to_drop -= discard;
	}
      while (n_bytes_to_drop);
      if (n_bytes_to_drop > first)
	b->total_length_not_including_first_buffer -= n_bytes_to_drop - first;
    }
  else
    vlib_buffer_advance (b, n_bytes_to_drop);
  vnet_buffer (b)->tcp.data_len -= n_bytes_to_drop;
  return 0;
}

/**
 * Receive buffer for connection and handle acks
 *
 * It handles both in order or out-of-order data.
 */
static int
tcp_segment_rcv (tcp_worker_ctx_t * wrk, tcp_connection_t * tc,
		 vlib_buffer_t * b)
{
  u32 error, n_bytes_to_drop, n_data_bytes;

  vlib_buffer_advance (b, vnet_buffer (b)->tcp.data_offset);
  n_data_bytes = vnet_buffer (b)->tcp.data_len;
  ASSERT (n_data_bytes);
  tc->data_segs_in += 1;

  /* Handle out-of-order data */
  if (PREDICT_FALSE (vnet_buffer (b)->tcp.seq_number != tc->rcv_nxt))
    {
      /* Old sequence numbers allowed through because they overlapped
       * the rx window */
      if (seq_lt (vnet_buffer (b)->tcp.seq_number, tc->rcv_nxt))
	{
	  /* Completely in the past (possible retransmit). Ack
	   * retransmissions since we may not have any data to send */
	  if (seq_leq (vnet_buffer (b)->tcp.seq_end, tc->rcv_nxt))
	    {
	      tcp_program_dupack (tc);
	      tc->errors.below_data_wnd++;
	      error = TCP_ERROR_SEGMENT_OLD;
	      goto done;
	    }

	  /* Chop off the bytes in the past and see if what is left
	   * can be enqueued in order */
	  n_bytes_to_drop = tc->rcv_nxt - vnet_buffer (b)->tcp.seq_number;
	  n_data_bytes -= n_bytes_to_drop;
	  vnet_buffer (b)->tcp.seq_number = tc->rcv_nxt;
	  if (tcp_buffer_discard_bytes (b, n_bytes_to_drop))
	    {
	      error = TCP_ERROR_SEGMENT_OLD;
	      goto done;
	    }
	  goto in_order;
	}

      /* RFC2581: Enqueue and send DUPACK for fast retransmit */
      error = tcp_session_enqueue_ooo (tc, b, n_data_bytes);
      tcp_program_dupack (tc);
      TCP_EVT (TCP_EVT_DUPACK_SENT, tc, vnet_buffer (b)->tcp);
      tc->errors.above_data_wnd += seq_gt (vnet_buffer (b)->tcp.seq_end,
					   tc->rcv_las + tc->rcv_wnd);
      goto done;
    }

in_order:

  /* In order data, enqueue. Fifo figures out by itself if any out-of-order
   * segments can be enqueued after fifo tail offset changes. */
  error = tcp_session_enqueue_data (tc, b, n_data_bytes);
  tcp_program_ack (tc);

done:
  return error;
}

typedef struct
{
  tcp_header_t tcp_header;
  tcp_connection_t tcp_connection;
} tcp_rx_trace_t;

static u8 *
format_tcp_rx_trace (u8 * s, va_list * args)
{
  CLIB_UNUSED (vlib_main_t * vm) = va_arg (*args, vlib_main_t *);
  CLIB_UNUSED (vlib_node_t * node) = va_arg (*args, vlib_node_t *);
  tcp_rx_trace_t *t = va_arg (*args, tcp_rx_trace_t *);
  tcp_connection_t *tc = &t->tcp_connection;
  u32 indent = format_get_indent (s);

  s = format (s, "%U state %U\n%U%U", format_tcp_connection_id, tc,
	      format_tcp_state, tc->state, format_white_space, indent,
	      format_tcp_header, &t->tcp_header, 128);

  return s;
}

static u8 *
format_tcp_rx_trace_short (u8 * s, va_list * args)
{
  CLIB_UNUSED (vlib_main_t * vm) = va_arg (*args, vlib_main_t *);
  CLIB_UNUSED (vlib_node_t * node) = va_arg (*args, vlib_node_t *);
  tcp_rx_trace_t *t = va_arg (*args, tcp_rx_trace_t *);

  s = format (s, "%d -> %d (%U)",
	      clib_net_to_host_u16 (t->tcp_header.dst_port),
	      clib_net_to_host_u16 (t->tcp_header.src_port), format_tcp_state,
	      t->tcp_connection.state);

  return s;
}

static void
tcp_set_rx_trace_data (tcp_rx_trace_t * t0, tcp_connection_t * tc0,
		       tcp_header_t * th0, vlib_buffer_t * b0, u8 is_ip4)
{
  if (tc0)
    {
      clib_memcpy_fast (&t0->tcp_connection, tc0,
			sizeof (t0->tcp_connection));
    }
  else
    {
      th0 = tcp_buffer_hdr (b0);
    }
  clib_memcpy_fast (&t0->tcp_header, th0, sizeof (t0->tcp_header));
}

static void
tcp_established_trace_frame (vlib_main_t * vm, vlib_node_runtime_t * node,
			     vlib_frame_t * frame, u8 is_ip4)
{
  u32 *from, n_left;

  n_left = frame->n_vectors;
  from = vlib_frame_vector_args (frame);

  while (n_left >= 1)
    {
      tcp_connection_t *tc0;
      tcp_rx_trace_t *t0;
      tcp_header_t *th0;
      vlib_buffer_t *b0;
      u32 bi0;

      bi0 = from[0];
      b0 = vlib_get_buffer (vm, bi0);

      if (b0->flags & VLIB_BUFFER_IS_TRACED)
	{
	  t0 = vlib_add_trace (vm, node, b0, sizeof (*t0));
	  tc0 = tcp_connection_get (vnet_buffer (b0)->tcp.connection_index,
				    vm->thread_index);
	  th0 = tcp_buffer_hdr (b0);
	  tcp_set_rx_trace_data (t0, tc0, th0, b0, is_ip4);
	}

      from += 1;
      n_left -= 1;
    }
}

always_inline void
tcp_node_inc_counter_i (vlib_main_t * vm, u32 tcp4_node, u32 tcp6_node,
			u8 is_ip4, u32 evt, u32 val)
{
  if (is_ip4)
    vlib_node_increment_counter (vm, tcp4_node, evt, val);
  else
    vlib_node_increment_counter (vm, tcp6_node, evt, val);
}

#define tcp_maybe_inc_counter(node_id, err, count)			\
{									\
  if (next0 != tcp_next_drop (is_ip4))					\
    tcp_node_inc_counter_i (vm, tcp4_##node_id##_node.index,		\
                            tcp6_##node_id##_node.index, is_ip4, err, 	\
			    1);						\
}
#define tcp_inc_counter(node_id, err, count)				\
  tcp_node_inc_counter_i (vm, tcp4_##node_id##_node.index,		\
	                   tcp6_##node_id##_node.index, is_ip4,		\
	                   err, count)
#define tcp_maybe_inc_err_counter(cnts, err)				\
{									\
  cnts[err] += (next0 != tcp_next_drop (is_ip4));			\
}
#define tcp_inc_err_counter(cnts, err, val)				\
{									\
  cnts[err] += val;							\
}
#define tcp_store_err_counters(node_id, cnts)				\
{									\
  int i;								\
  for (i = 0; i < TCP_N_ERROR; i++)					\
    if (cnts[i])							\
      tcp_inc_counter(node_id, i, cnts[i]);				\
}


always_inline uword
tcp46_established_inline (vlib_main_t * vm, vlib_node_runtime_t * node,
			  vlib_frame_t * frame, int is_ip4)
{
  u32 thread_index = vm->thread_index, errors = 0;
  tcp_worker_ctx_t *wrk = tcp_get_worker (thread_index);
  u32 n_left_from, *from, *first_buffer;
  u16 err_counters[TCP_N_ERROR] = { 0 };

  if (node->flags & VLIB_NODE_FLAG_TRACE)
    tcp_established_trace_frame (vm, node, frame, is_ip4);

  first_buffer = from = vlib_frame_vector_args (frame);
  n_left_from = frame->n_vectors;

  while (n_left_from > 0)
    {
      u32 bi0, error0 = TCP_ERROR_ACK_OK;
      vlib_buffer_t *b0;
      tcp_header_t *th0;
      tcp_connection_t *tc0;

      if (n_left_from > 1)
	{
	  vlib_buffer_t *pb;
	  pb = vlib_get_buffer (vm, from[1]);
	  vlib_prefetch_buffer_header (pb, LOAD);
	  CLIB_PREFETCH (pb->data, 2 * CLIB_CACHE_LINE_BYTES, LOAD);
	}

      bi0 = from[0];
      from += 1;
      n_left_from -= 1;

      b0 = vlib_get_buffer (vm, bi0);
      tc0 = tcp_connection_get (vnet_buffer (b0)->tcp.connection_index,
				thread_index);

      if (PREDICT_FALSE (tc0 == 0))
	{
	  error0 = TCP_ERROR_INVALID_CONNECTION;
	  goto done;
	}

      th0 = tcp_buffer_hdr (b0);

      /* TODO header prediction fast path */

      /* 1-4: check SEQ, RST, SYN */
      if (PREDICT_FALSE (tcp_segment_validate (wrk, tc0, b0, th0, &error0)))
	{
	  TCP_EVT (TCP_EVT_SEG_INVALID, tc0, vnet_buffer (b0)->tcp);
	  goto done;
	}

      /* 5: check the ACK field  */
      if (PREDICT_FALSE (tcp_rcv_ack (wrk, tc0, b0, th0, &error0)))
	goto done;

      /* 6: check the URG bit TODO */

      /* 7: process the segment text */
      if (vnet_buffer (b0)->tcp.data_len)
	error0 = tcp_segment_rcv (wrk, tc0, b0);

      /* 8: check the FIN bit */
      if (PREDICT_FALSE (tcp_is_fin (th0)))
	tcp_rcv_fin (wrk, tc0, b0, &error0);

    done:
      tcp_inc_err_counter (err_counters, error0, 1);
    }

  errors = session_main_flush_enqueue_events (TRANSPORT_PROTO_TCP,
					      thread_index);
  err_counters[TCP_ERROR_MSG_QUEUE_FULL] = errors;
  tcp_store_err_counters (established, err_counters);
  tcp_handle_postponed_dequeues (wrk);
  tcp_handle_disconnects (wrk);
  vlib_buffer_free (vm, first_buffer, frame->n_vectors);

  return frame->n_vectors;
}

VLIB_NODE_FN (tcp4_established_node) (vlib_main_t * vm,
				      vlib_node_runtime_t * node,
				      vlib_frame_t * from_frame)
{
  return tcp46_established_inline (vm, node, from_frame, 1 /* is_ip4 */ );
}

VLIB_NODE_FN (tcp6_established_node) (vlib_main_t * vm,
				      vlib_node_runtime_t * node,
				      vlib_frame_t * from_frame)
{
  return tcp46_established_inline (vm, node, from_frame, 0 /* is_ip4 */ );
}

/* *INDENT-OFF* */
VLIB_REGISTER_NODE (tcp4_established_node) =
{
  .name = "tcp4-established",
  /* Takes a vector of packets. */
  .vector_size = sizeof (u32),
  .n_errors = TCP_N_ERROR,
  .error_strings = tcp_error_strings,
  .n_next_nodes = TCP_ESTABLISHED_N_NEXT,
  .next_nodes =
  {
#define _(s,n) [TCP_ESTABLISHED_NEXT_##s] = n,
    foreach_tcp_state_next
#undef _
  },
  .format_trace = format_tcp_rx_trace_short,
};
/* *INDENT-ON* */

/* *INDENT-OFF* */
VLIB_REGISTER_NODE (tcp6_established_node) =
{
  .name = "tcp6-established",
  /* Takes a vector of packets. */
  .vector_size = sizeof (u32),
  .n_errors = TCP_N_ERROR,
  .error_strings = tcp_error_strings,
  .n_next_nodes = TCP_ESTABLISHED_N_NEXT,
  .next_nodes =
  {
#define _(s,n) [TCP_ESTABLISHED_NEXT_##s] = n,
    foreach_tcp_state_next
#undef _
  },
  .format_trace = format_tcp_rx_trace_short,
};
/* *INDENT-ON* */


static u8
tcp_lookup_is_valid (tcp_connection_t * tc, vlib_buffer_t * b,
		     tcp_header_t * hdr)
{
  transport_connection_t *tmp = 0;
  u64 handle;

  if (!tc)
    return 1;

  /* Proxy case */
  if (tc->c_lcl_port == 0 && tc->state == TCP_STATE_LISTEN)
    return 1;

  u8 is_ip_valid = 0, val_l, val_r;

  if (tc->connection.is_ip4)
    {
      ip4_header_t *ip4_hdr = (ip4_header_t *) vlib_buffer_get_current (b);

      val_l = !ip4_address_compare (&ip4_hdr->dst_address,
				    &tc->connection.lcl_ip.ip4);
      val_l = val_l || ip_is_zero (&tc->connection.lcl_ip, 1);
      val_r = !ip4_address_compare (&ip4_hdr->src_address,
				    &tc->connection.rmt_ip.ip4);
      val_r = val_r || tc->state == TCP_STATE_LISTEN;
      is_ip_valid = val_l && val_r;
    }
  else
    {
      ip6_header_t *ip6_hdr = (ip6_header_t *) vlib_buffer_get_current (b);

      val_l = !ip6_address_compare (&ip6_hdr->dst_address,
				    &tc->connection.lcl_ip.ip6);
      val_l = val_l || ip_is_zero (&tc->connection.lcl_ip, 0);
      val_r = !ip6_address_compare (&ip6_hdr->src_address,
				    &tc->connection.rmt_ip.ip6);
      val_r = val_r || tc->state == TCP_STATE_LISTEN;
      is_ip_valid = val_l && val_r;
    }

  u8 is_valid = (tc->c_lcl_port == hdr->dst_port
		 && (tc->state == TCP_STATE_LISTEN
		     || tc->c_rmt_port == hdr->src_port) && is_ip_valid);

  if (!is_valid)
    {
      handle = session_lookup_half_open_handle (&tc->connection);
      tmp = session_lookup_half_open_connection (handle & 0xFFFFFFFF,
						 tc->c_proto, tc->c_is_ip4);

      if (tmp)
	{
	  if (tmp->lcl_port == hdr->dst_port
	      && tmp->rmt_port == hdr->src_port)
	    {
	      TCP_DBG ("half-open is valid!");
	      is_valid = 1;
	    }
	}
    }
  return is_valid;
}

/**
 * Lookup transport connection
 */
static tcp_connection_t *
tcp_lookup_connection (u32 fib_index, vlib_buffer_t * b, u8 thread_index,
		       u8 is_ip4)
{
  tcp_header_t *tcp;
  transport_connection_t *tconn;
  tcp_connection_t *tc;
  u8 is_filtered = 0;
  if (is_ip4)
    {
      ip4_header_t *ip4;
      ip4 = vlib_buffer_get_current (b);
      tcp = ip4_next_header (ip4);
      tconn = session_lookup_connection_wt4 (fib_index,
					     &ip4->dst_address,
					     &ip4->src_address,
					     tcp->dst_port,
					     tcp->src_port,
					     TRANSPORT_PROTO_TCP,
					     thread_index, &is_filtered);
      tc = tcp_get_connection_from_transport (tconn);
      ASSERT (tcp_lookup_is_valid (tc, b, tcp));
    }
  else
    {
      ip6_header_t *ip6;
      ip6 = vlib_buffer_get_current (b);
      tcp = ip6_next_header (ip6);
      tconn = session_lookup_connection_wt6 (fib_index,
					     &ip6->dst_address,
					     &ip6->src_address,
					     tcp->dst_port,
					     tcp->src_port,
					     TRANSPORT_PROTO_TCP,
					     thread_index, &is_filtered);
      tc = tcp_get_connection_from_transport (tconn);
      ASSERT (tcp_lookup_is_valid (tc, b, tcp));
    }
  return tc;
}

static tcp_connection_t *
tcp_lookup_listener (vlib_buffer_t * b, u32 fib_index, int is_ip4)
{
  session_t *s;

  if (is_ip4)
    {
      ip4_header_t *ip4 = vlib_buffer_get_current (b);
      tcp_header_t *tcp = tcp_buffer_hdr (b);
      s = session_lookup_listener4 (fib_index,
				    &ip4->dst_address,
				    tcp->dst_port, TRANSPORT_PROTO_TCP, 1);
    }
  else
    {
      ip6_header_t *ip6 = vlib_buffer_get_current (b);
      tcp_header_t *tcp = tcp_buffer_hdr (b);
      s = session_lookup_listener6 (fib_index,
				    &ip6->dst_address,
				    tcp->dst_port, TRANSPORT_PROTO_TCP, 1);

    }
  if (PREDICT_TRUE (s != 0))
    return tcp_get_connection_from_transport (transport_get_listener
					      (TRANSPORT_PROTO_TCP,
					       s->connection_index));
  else
    return 0;
}

always_inline void
tcp_check_tx_offload (tcp_connection_t * tc, int is_ipv4)
{
  vnet_main_t *vnm = vnet_get_main ();
  const dpo_id_t *dpo;
  const load_balance_t *lb;
  vnet_hw_interface_t *hw_if;
  u32 sw_if_idx, lb_idx;

  if (is_ipv4)
    {
      ip4_address_t *dst_addr = &(tc->c_rmt_ip.ip4);
      lb_idx = ip4_fib_forwarding_lookup (tc->c_fib_index, dst_addr);
    }
  else
    {
      ip6_address_t *dst_addr = &(tc->c_rmt_ip.ip6);
      lb_idx = ip6_fib_table_fwding_lookup (tc->c_fib_index, dst_addr);
    }

  lb = load_balance_get (lb_idx);
  if (PREDICT_FALSE (lb->lb_n_buckets > 1))
    return;
  dpo = load_balance_get_bucket_i (lb, 0);

  sw_if_idx = dpo_get_urpf (dpo);
  if (PREDICT_FALSE (sw_if_idx == ~0))
    return;

  hw_if = vnet_get_sup_hw_interface (vnm, sw_if_idx);
  if (hw_if->flags & VNET_HW_INTERFACE_FLAG_SUPPORTS_GSO)
    tc->cfg_flags |= TCP_CFG_F_TSO;
}

always_inline uword
tcp46_syn_sent_inline (vlib_main_t * vm, vlib_node_runtime_t * node,
		       vlib_frame_t * from_frame, int is_ip4)
{
  u32 n_left_from, *from, *first_buffer, errors = 0;
  u32 my_thread_index = vm->thread_index;
  tcp_worker_ctx_t *wrk = tcp_get_worker (my_thread_index);

  from = first_buffer = vlib_frame_vector_args (from_frame);
  n_left_from = from_frame->n_vectors;

  while (n_left_from > 0)
    {
      u32 bi0, ack0, seq0, error0 = TCP_ERROR_NONE;
      tcp_connection_t *tc0, *new_tc0;
      tcp_header_t *tcp0 = 0;
      tcp_rx_trace_t *t0;
      vlib_buffer_t *b0;

      bi0 = from[0];
      from += 1;
      n_left_from -= 1;

      b0 = vlib_get_buffer (vm, bi0);
      tc0 =
	tcp_half_open_connection_get (vnet_buffer (b0)->tcp.connection_index);
      if (PREDICT_FALSE (tc0 == 0))
	{
	  error0 = TCP_ERROR_INVALID_CONNECTION;
	  goto drop;
	}

      /* Half-open completed recently but the connection was't removed
       * yet by the owning thread */
      if (PREDICT_FALSE (tc0->flags & TCP_CONN_HALF_OPEN_DONE))
	{
	  /* Make sure the connection actually exists */
	  ASSERT (tcp_lookup_connection (tc0->c_fib_index, b0,
					 my_thread_index, is_ip4));
	  error0 = TCP_ERROR_SPURIOUS_SYN_ACK;
	  goto drop;
	}

      ack0 = vnet_buffer (b0)->tcp.ack_number;
      seq0 = vnet_buffer (b0)->tcp.seq_number;
      tcp0 = tcp_buffer_hdr (b0);

      /* Crude check to see if the connection handle does not match
       * the packet. Probably connection just switched to established */
      if (PREDICT_FALSE (tcp0->dst_port != tc0->c_lcl_port
			 || tcp0->src_port != tc0->c_rmt_port))
	{
	  error0 = TCP_ERROR_INVALID_CONNECTION;
	  goto drop;
	}

      if (PREDICT_FALSE (!tcp_ack (tcp0) && !tcp_rst (tcp0)
			 && !tcp_syn (tcp0)))
	{
	  error0 = TCP_ERROR_SEGMENT_INVALID;
	  goto drop;
	}

      /* SYNs consume sequence numbers */
      vnet_buffer (b0)->tcp.seq_end += tcp_is_syn (tcp0);

      /*
       *  1. check the ACK bit
       */

      /*
       *   If the ACK bit is set
       *     If SEG.ACK =< ISS, or SEG.ACK > SND.NXT, send a reset (unless
       *     the RST bit is set, if so drop the segment and return)
       *       <SEQ=SEG.ACK><CTL=RST>
       *     and discard the segment.  Return.
       *     If SND.UNA =< SEG.ACK =< SND.NXT then the ACK is acceptable.
       */
      if (tcp_ack (tcp0))
	{
	  if (seq_leq (ack0, tc0->iss) || seq_gt (ack0, tc0->snd_nxt))
	    {
	      if (!tcp_rst (tcp0))
		tcp_send_reset_w_pkt (tc0, b0, my_thread_index, is_ip4);
	      error0 = TCP_ERROR_RCV_WND;
	      goto drop;
	    }

	  /* Make sure ACK is valid */
	  if (seq_gt (tc0->snd_una, ack0))
	    {
	      error0 = TCP_ERROR_ACK_INVALID;
	      goto drop;
	    }
	}

      /*
       * 2. check the RST bit
       */

      if (tcp_rst (tcp0))
	{
	  /* If ACK is acceptable, signal client that peer is not
	   * willing to accept connection and drop connection*/
	  if (tcp_ack (tcp0))
	    tcp_rcv_rst (wrk, tc0);
	  error0 = TCP_ERROR_RST_RCVD;
	  goto drop;
	}

      /*
       * 3. check the security and precedence (skipped)
       */

      /*
       * 4. check the SYN bit
       */

      /* No SYN flag. Drop. */
      if (!tcp_syn (tcp0))
	{
	  error0 = TCP_ERROR_SEGMENT_INVALID;
	  goto drop;
	}

      /* Parse options */
      if (tcp_options_parse (tcp0, &tc0->rcv_opts, 1))
	{
	  error0 = TCP_ERROR_OPTIONS;
	  goto drop;
	}

      /* Valid SYN or SYN-ACK. Move connection from half-open pool to
       * current thread pool. */
      new_tc0 = tcp_connection_alloc_w_base (my_thread_index, tc0);
      new_tc0->rcv_nxt = vnet_buffer (b0)->tcp.seq_end;
      new_tc0->irs = seq0;
      new_tc0->timers[TCP_TIMER_RETRANSMIT_SYN] = TCP_TIMER_HANDLE_INVALID;
      new_tc0->sw_if_index = vnet_buffer (b0)->sw_if_index[VLIB_RX];

      if (tcp_opts_tstamp (&new_tc0->rcv_opts))
	{
	  new_tc0->tsval_recent = new_tc0->rcv_opts.tsval;
	  new_tc0->tsval_recent_age = tcp_time_now ();
	}

      if (tcp_opts_wscale (&new_tc0->rcv_opts))
	new_tc0->snd_wscale = new_tc0->rcv_opts.wscale;
      else
	new_tc0->rcv_wscale = 0;

      new_tc0->snd_wnd = clib_net_to_host_u16 (tcp0->window)
	<< new_tc0->snd_wscale;
      new_tc0->snd_wl1 = seq0;
      new_tc0->snd_wl2 = ack0;

      tcp_connection_init_vars (new_tc0);

      /* SYN-ACK: See if we can switch to ESTABLISHED state */
      if (PREDICT_TRUE (tcp_ack (tcp0)))
	{
	  /* Our SYN is ACKed: we have iss < ack = snd_una */

	  /* TODO Dequeue acknowledged segments if we support Fast Open */
	  new_tc0->snd_una = ack0;
	  new_tc0->state = TCP_STATE_ESTABLISHED;

	  /* Make sure las is initialized for the wnd computation */
	  new_tc0->rcv_las = new_tc0->rcv_nxt;

	  /* Notify app that we have connection. If session layer can't
	   * allocate session send reset */
	  if (session_stream_connect_notify (&new_tc0->connection,
					     SESSION_E_NONE))
	    {
	      tcp_send_reset_w_pkt (new_tc0, b0, my_thread_index, is_ip4);
	      tcp_connection_cleanup (new_tc0);
	      error0 = TCP_ERROR_CREATE_SESSION_FAIL;
	      goto cleanup_ho;
	    }

	  new_tc0->tx_fifo_size =
	    transport_tx_fifo_size (&new_tc0->connection);
	  /* Update rtt with the syn-ack sample */
	  tcp_estimate_initial_rtt (new_tc0);
	  TCP_EVT (TCP_EVT_SYNACK_RCVD, new_tc0);
	  error0 = TCP_ERROR_SYN_ACKS_RCVD;
	}
      /* SYN: Simultaneous open. Change state to SYN-RCVD and send SYN-ACK */
      else
	{
	  new_tc0->state = TCP_STATE_SYN_RCVD;

	  /* Notify app that we have connection */
	  if (session_stream_connect_notify (&new_tc0->connection,
					     SESSION_E_NONE))
	    {
	      tcp_connection_cleanup (new_tc0);
	      tcp_send_reset_w_pkt (tc0, b0, my_thread_index, is_ip4);
	      TCP_EVT (TCP_EVT_RST_SENT, tc0);
	      error0 = TCP_ERROR_CREATE_SESSION_FAIL;
	      goto cleanup_ho;
	    }

	  new_tc0->tx_fifo_size =
	    transport_tx_fifo_size (&new_tc0->connection);
	  new_tc0->rtt_ts = 0;
	  tcp_init_snd_vars (new_tc0);
	  tcp_send_synack (new_tc0);
	  error0 = TCP_ERROR_SYNS_RCVD;
	  goto cleanup_ho;
	}

      if (!(new_tc0->cfg_flags & TCP_CFG_F_NO_TSO))
	tcp_check_tx_offload (new_tc0, is_ip4);

      /* Read data, if any */
      if (PREDICT_FALSE (vnet_buffer (b0)->tcp.data_len))
	{
	  clib_warning ("rcvd data in syn-sent");
	  error0 = tcp_segment_rcv (wrk, new_tc0, b0);
	  if (error0 == TCP_ERROR_ACK_OK)
	    error0 = TCP_ERROR_SYN_ACKS_RCVD;
	}
      else
	{
	  /* Send ack now instead of programming it because connection was
	   * just established and it's not optional. */
	  tcp_send_ack (new_tc0);
	}

    cleanup_ho:

      /* If this is not the owning thread, wait for syn retransmit to
       * expire and cleanup then */
      if (tcp_half_open_connection_cleanup (tc0))
	tc0->flags |= TCP_CONN_HALF_OPEN_DONE;

    drop:

      tcp_inc_counter (syn_sent, error0, 1);
      if (PREDICT_FALSE ((b0->flags & VLIB_BUFFER_IS_TRACED) && tcp0 != 0))
	{
	  t0 = vlib_add_trace (vm, node, b0, sizeof (*t0));
	  clib_memcpy_fast (&t0->tcp_header, tcp0, sizeof (t0->tcp_header));
	  clib_memcpy_fast (&t0->tcp_connection, tc0,
			    sizeof (t0->tcp_connection));
	}
    }

  errors = session_main_flush_enqueue_events (TRANSPORT_PROTO_TCP,
					      my_thread_index);
  tcp_inc_counter (syn_sent, TCP_ERROR_MSG_QUEUE_FULL, errors);
  vlib_buffer_free (vm, first_buffer, from_frame->n_vectors);
  tcp_handle_disconnects (wrk);

  return from_frame->n_vectors;
}

VLIB_NODE_FN (tcp4_syn_sent_node) (vlib_main_t * vm,
				   vlib_node_runtime_t * node,
				   vlib_frame_t * from_frame)
{
  return tcp46_syn_sent_inline (vm, node, from_frame, 1 /* is_ip4 */ );
}

VLIB_NODE_FN (tcp6_syn_sent_node) (vlib_main_t * vm,
				   vlib_node_runtime_t * node,
				   vlib_frame_t * from_frame)
{
  return tcp46_syn_sent_inline (vm, node, from_frame, 0 /* is_ip4 */ );
}

/* *INDENT-OFF* */
VLIB_REGISTER_NODE (tcp4_syn_sent_node) =
{
  .name = "tcp4-syn-sent",
  /* Takes a vector of packets. */
  .vector_size = sizeof (u32),
  .n_errors = TCP_N_ERROR,
  .error_strings = tcp_error_strings,
  .n_next_nodes = TCP_SYN_SENT_N_NEXT,
  .next_nodes =
  {
#define _(s,n) [TCP_SYN_SENT_NEXT_##s] = n,
    foreach_tcp_state_next
#undef _
  },
  .format_trace = format_tcp_rx_trace_short,
};
/* *INDENT-ON* */

/* *INDENT-OFF* */
VLIB_REGISTER_NODE (tcp6_syn_sent_node) =
{
  .name = "tcp6-syn-sent",
  /* Takes a vector of packets. */
  .vector_size = sizeof (u32),
  .n_errors = TCP_N_ERROR,
  .error_strings = tcp_error_strings,
  .n_next_nodes = TCP_SYN_SENT_N_NEXT,
  .next_nodes =
  {
#define _(s,n) [TCP_SYN_SENT_NEXT_##s] = n,
    foreach_tcp_state_next
#undef _
  },
  .format_trace = format_tcp_rx_trace_short,
};
/* *INDENT-ON* */

/**
 * Handles reception for all states except LISTEN, SYN-SENT and ESTABLISHED
 * as per RFC793 p. 64
 */
always_inline uword
tcp46_rcv_process_inline (vlib_main_t * vm, vlib_node_runtime_t * node,
			  vlib_frame_t * from_frame, int is_ip4)
{
  u32 thread_index = vm->thread_index, errors = 0, *first_buffer;
  tcp_worker_ctx_t *wrk = tcp_get_worker (thread_index);
  u32 n_left_from, *from, max_dequeue;

  from = first_buffer = vlib_frame_vector_args (from_frame);
  n_left_from = from_frame->n_vectors;

  while (n_left_from > 0)
    {
      u32 bi0, error0 = TCP_ERROR_NONE;
      tcp_header_t *tcp0 = 0;
      tcp_connection_t *tc0;
      vlib_buffer_t *b0;
      u8 is_fin0;

      bi0 = from[0];
      from += 1;
      n_left_from -= 1;

      b0 = vlib_get_buffer (vm, bi0);
      tc0 = tcp_connection_get (vnet_buffer (b0)->tcp.connection_index,
				thread_index);
      if (PREDICT_FALSE (tc0 == 0))
	{
	  error0 = TCP_ERROR_INVALID_CONNECTION;
	  goto drop;
	}

      tcp0 = tcp_buffer_hdr (b0);
      is_fin0 = tcp_is_fin (tcp0);

      if (CLIB_DEBUG)
	{
	  if (!(tc0->connection.flags & TRANSPORT_CONNECTION_F_NO_LOOKUP))
	    {
	      tcp_connection_t *tmp;
	      tmp = tcp_lookup_connection (tc0->c_fib_index, b0, thread_index,
					   is_ip4);
	      if (tmp->state != tc0->state)
		{
		  if (tc0->state != TCP_STATE_CLOSED)
		    clib_warning ("state changed");
		  goto drop;
		}
	    }
	}

      /*
       * Special treatment for CLOSED
       */
      if (PREDICT_FALSE (tc0->state == TCP_STATE_CLOSED))
	{
	  error0 = TCP_ERROR_CONNECTION_CLOSED;
	  goto drop;
	}

      /*
       * For all other states (except LISTEN)
       */

      /* 1-4: check SEQ, RST, SYN */
      if (PREDICT_FALSE (tcp_segment_validate (wrk, tc0, b0, tcp0, &error0)))
	goto drop;

      /* 5: check the ACK field  */
      switch (tc0->state)
	{
	case TCP_STATE_SYN_RCVD:

	  /* Make sure the segment is exactly right */
	  if (tc0->rcv_nxt != vnet_buffer (b0)->tcp.seq_number || is_fin0)
	    {
	      tcp_send_reset_w_pkt (tc0, b0, thread_index, is_ip4);
	      error0 = TCP_ERROR_SEGMENT_INVALID;
	      goto drop;
	    }

	  /*
	   * If the segment acknowledgment is not acceptable, form a
	   * reset segment,
	   *  <SEQ=SEG.ACK><CTL=RST>
	   * and send it.
	   */
	  if (tcp_rcv_ack_no_cc (tc0, b0, &error0))
	    {
	      tcp_send_reset_w_pkt (tc0, b0, thread_index, is_ip4);
	      error0 = TCP_ERROR_SEGMENT_INVALID;
	      goto drop;
	    }

	  /* Update rtt and rto */
	  tcp_estimate_initial_rtt (tc0);
	  tcp_connection_tx_pacer_update (tc0);

	  /* Switch state to ESTABLISHED */
	  tc0->state = TCP_STATE_ESTABLISHED;
	  TCP_EVT (TCP_EVT_STATE_CHANGE, tc0);

	  if (!(tc0->cfg_flags & TCP_CFG_F_NO_TSO))
	    tcp_check_tx_offload (tc0, is_ip4);

	  /* Initialize session variables */
	  tc0->snd_una = vnet_buffer (b0)->tcp.ack_number;
	  tc0->snd_wnd = clib_net_to_host_u16 (tcp0->window)
	    << tc0->rcv_opts.wscale;
	  tc0->snd_wl1 = vnet_buffer (b0)->tcp.seq_number;
	  tc0->snd_wl2 = vnet_buffer (b0)->tcp.ack_number;

	  /* Reset SYN-ACK retransmit and SYN_RCV establish timers */
	  tcp_retransmit_timer_reset (&wrk->timer_wheel, tc0);
	  if (session_stream_accept_notify (&tc0->connection))
	    {
	      error0 = TCP_ERROR_MSG_QUEUE_FULL;
	      tcp_send_reset (tc0);
	      session_transport_delete_notify (&tc0->connection);
	      tcp_connection_cleanup (tc0);
	      goto drop;
	    }
	  error0 = TCP_ERROR_ACK_OK;
	  break;
	case TCP_STATE_ESTABLISHED:
	  /* We can get packets in established state here because they
	   * were enqueued before state change */
	  if (tcp_rcv_ack (wrk, tc0, b0, tcp0, &error0))
	    goto drop;

	  break;
	case TCP_STATE_FIN_WAIT_1:
	  /* In addition to the processing for the ESTABLISHED state, if
	   * our FIN is now acknowledged then enter FIN-WAIT-2 and
	   * continue processing in that state. */
	  if (tcp_rcv_ack (wrk, tc0, b0, tcp0, &error0))
	    goto drop;

	  /* Still have to send the FIN */
	  if (tc0->flags & TCP_CONN_FINPNDG)
	    {
	      /* TX fifo finally drained */
	      max_dequeue = transport_max_tx_dequeue (&tc0->connection);
	      if (max_dequeue <= tc0->burst_acked)
		tcp_send_fin (tc0);
	      /* If a fin was received and data was acked extend wait */
	      else if ((tc0->flags & TCP_CONN_FINRCVD) && tc0->bytes_acked)
		tcp_timer_update (&wrk->timer_wheel, tc0, TCP_TIMER_WAITCLOSE,
				  tcp_cfg.closewait_time);
	    }
	  /* If FIN is ACKed */
	  else if (tc0->snd_una == tc0->snd_nxt)
	    {
	      /* Stop all retransmit timers because we have nothing more
	       * to send. */
	      tcp_connection_timers_reset (tc0);

	      /* We already have a FIN but didn't transition to CLOSING
	       * because of outstanding tx data. Close the connection. */
	      if (tc0->flags & TCP_CONN_FINRCVD)
		{
		  tcp_connection_set_state (tc0, TCP_STATE_CLOSED);
		  session_transport_closed_notify (&tc0->connection);
		  tcp_program_cleanup (wrk, tc0);
		  goto drop;
		}

	      tcp_connection_set_state (tc0, TCP_STATE_FIN_WAIT_2);
	      /* Enable waitclose because we're willing to wait for peer's
	       * FIN but not indefinitely. */
	      tcp_timer_set (&wrk->timer_wheel, tc0, TCP_TIMER_WAITCLOSE,
			     tcp_cfg.finwait2_time);

	      /* Don't try to deq the FIN acked */
	      if (tc0->burst_acked > 1)
		session_tx_fifo_dequeue_drop (&tc0->connection,
					      tc0->burst_acked - 1);
	      tc0->burst_acked = 0;
	    }
	  break;
	case TCP_STATE_FIN_WAIT_2:
	  /* In addition to the processing for the ESTABLISHED state, if
	   * the retransmission queue is empty, the user's CLOSE can be
	   * acknowledged ("ok") but do not delete the TCB. */
	  if (tcp_rcv_ack_no_cc (tc0, b0, &error0))
	    goto drop;
	  tc0->burst_acked = 0;
	  break;
	case TCP_STATE_CLOSE_WAIT:
	  /* Do the same processing as for the ESTABLISHED state. */
	  if (tcp_rcv_ack (wrk, tc0, b0, tcp0, &error0))
	    goto drop;

	  if (!(tc0->flags & TCP_CONN_FINPNDG))
	    break;

	  /* Still have outstanding tx data */
	  max_dequeue = transport_max_tx_dequeue (&tc0->connection);
	  if (max_dequeue > tc0->burst_acked)
	    break;

	  tcp_send_fin (tc0);
	  tcp_connection_timers_reset (tc0);
	  tcp_connection_set_state (tc0, TCP_STATE_LAST_ACK);
	  tcp_timer_set (&wrk->timer_wheel, tc0, TCP_TIMER_WAITCLOSE,
			 tcp_cfg.lastack_time);
	  break;
	case TCP_STATE_CLOSING:
	  /* In addition to the processing for the ESTABLISHED state, if
	   * the ACK acknowledges our FIN then enter the TIME-WAIT state,
	   * otherwise ignore the segment. */
	  if (tcp_rcv_ack_no_cc (tc0, b0, &error0))
	    goto drop;

	  if (tc0->snd_una != tc0->snd_nxt)
	    goto drop;

	  tcp_connection_timers_reset (tc0);
	  tcp_connection_set_state (tc0, TCP_STATE_TIME_WAIT);
	  tcp_timer_set (&wrk->timer_wheel, tc0, TCP_TIMER_WAITCLOSE,
			 tcp_cfg.timewait_time);
	  session_transport_closed_notify (&tc0->connection);
	  goto drop;

	  break;
	case TCP_STATE_LAST_ACK:
	  /* The only thing that [should] arrive in this state is an
	   * acknowledgment of our FIN. If our FIN is now acknowledged,
	   * delete the TCB, enter the CLOSED state, and return. */

	  if (tcp_rcv_ack_no_cc (tc0, b0, &error0))
	    goto drop;

	  /* Apparently our ACK for the peer's FIN was lost */
	  if (is_fin0 && tc0->snd_una != tc0->snd_nxt)
	    {
	      tcp_send_fin (tc0);
	      goto drop;
	    }

	  tcp_connection_set_state (tc0, TCP_STATE_CLOSED);
	  session_transport_closed_notify (&tc0->connection);

	  /* Don't free the connection from the data path since
	   * we can't ensure that we have no packets already enqueued
	   * to output. Rely instead on the waitclose timer */
	  tcp_connection_timers_reset (tc0);
	  tcp_program_cleanup (tcp_get_worker (tc0->c_thread_index), tc0);

	  goto drop;

	  break;
	case TCP_STATE_TIME_WAIT:
	  /* The only thing that can arrive in this state is a
	   * retransmission of the remote FIN. Acknowledge it, and restart
	   * the 2 MSL timeout. */

	  if (tcp_rcv_ack_no_cc (tc0, b0, &error0))
	    goto drop;

	  if (!is_fin0)
	    goto drop;

	  tcp_program_ack (tc0);
	  tcp_timer_update (&wrk->timer_wheel, tc0, TCP_TIMER_WAITCLOSE,
			    tcp_cfg.timewait_time);
	  goto drop;

	  break;
	default:
	  ASSERT (0);
	}

      /* 6: check the URG bit TODO */

      /* 7: process the segment text */
      switch (tc0->state)
	{
	case TCP_STATE_ESTABLISHED:
	case TCP_STATE_FIN_WAIT_1:
	case TCP_STATE_FIN_WAIT_2:
	  if (vnet_buffer (b0)->tcp.data_len)
	    error0 = tcp_segment_rcv (wrk, tc0, b0);
	  break;
	case TCP_STATE_CLOSE_WAIT:
	case TCP_STATE_CLOSING:
	case TCP_STATE_LAST_ACK:
	case TCP_STATE_TIME_WAIT:
	  /* This should not occur, since a FIN has been received from the
	   * remote side.  Ignore the segment text. */
	  break;
	}

      /* 8: check the FIN bit */
      if (!is_fin0)
	goto drop;

      TCP_EVT (TCP_EVT_FIN_RCVD, tc0);

      switch (tc0->state)
	{
	case TCP_STATE_ESTABLISHED:
	  /* Account for the FIN and send ack */
	  tc0->rcv_nxt += 1;
	  tcp_program_ack (tc0);
	  tcp_connection_set_state (tc0, TCP_STATE_CLOSE_WAIT);
	  tcp_program_disconnect (wrk, tc0);
	  tcp_timer_update (&wrk->timer_wheel, tc0, TCP_TIMER_WAITCLOSE,
			    tcp_cfg.closewait_time);
	  break;
	case TCP_STATE_SYN_RCVD:
	  /* Send FIN-ACK, enter LAST-ACK and because the app was not
	   * notified yet, set a cleanup timer instead of relying on
	   * disconnect notify and the implicit close call. */
	  tcp_connection_timers_reset (tc0);
	  tc0->rcv_nxt += 1;
	  tcp_send_fin (tc0);
	  tcp_connection_set_state (tc0, TCP_STATE_LAST_ACK);
	  tcp_timer_set (&wrk->timer_wheel, tc0, TCP_TIMER_WAITCLOSE,
			 tcp_cfg.lastack_time);
	  break;
	case TCP_STATE_CLOSE_WAIT:
	case TCP_STATE_CLOSING:
	case TCP_STATE_LAST_ACK:
	  /* move along .. */
	  break;
	case TCP_STATE_FIN_WAIT_1:
	  tc0->rcv_nxt += 1;

	  if (tc0->flags & TCP_CONN_FINPNDG)
	    {
	      /* If data is outstanding, stay in FIN_WAIT_1 and try to finish
	       * sending it. Since we already received a fin, do not wait
	       * for too long. */
	      tc0->flags |= TCP_CONN_FINRCVD;
	      tcp_timer_update (&wrk->timer_wheel, tc0, TCP_TIMER_WAITCLOSE,
				tcp_cfg.closewait_time);
	    }
	  else
	    {
	      tcp_connection_set_state (tc0, TCP_STATE_CLOSING);
	      tcp_program_ack (tc0);
	      /* Wait for ACK for our FIN but not forever */
	      tcp_timer_update (&wrk->timer_wheel, tc0, TCP_TIMER_WAITCLOSE,
				tcp_cfg.closing_time);
	    }
	  break;
	case TCP_STATE_FIN_WAIT_2:
	  /* Got FIN, send ACK! Be more aggressive with resource cleanup */
	  tc0->rcv_nxt += 1;
	  tcp_connection_set_state (tc0, TCP_STATE_TIME_WAIT);
	  tcp_connection_timers_reset (tc0);
	  tcp_timer_set (&wrk->timer_wheel, tc0, TCP_TIMER_WAITCLOSE,
			 tcp_cfg.timewait_time);
	  tcp_program_ack (tc0);
	  session_transport_closed_notify (&tc0->connection);
	  break;
	case TCP_STATE_TIME_WAIT:
	  /* Remain in the TIME-WAIT state. Restart the time-wait
	   * timeout.
	   */
	  tcp_timer_update (&wrk->timer_wheel, tc0, TCP_TIMER_WAITCLOSE,
			    tcp_cfg.timewait_time);
	  break;
	}
      error0 = TCP_ERROR_FIN_RCVD;

    drop:

      tcp_inc_counter (rcv_process, error0, 1);
      if (PREDICT_FALSE (b0->flags & VLIB_BUFFER_IS_TRACED))
	{
	  tcp_rx_trace_t *t0 = vlib_add_trace (vm, node, b0, sizeof (*t0));
	  tcp_set_rx_trace_data (t0, tc0, tcp0, b0, is_ip4);
	}
    }

  errors = session_main_flush_enqueue_events (TRANSPORT_PROTO_TCP,
					      thread_index);
  tcp_inc_counter (rcv_process, TCP_ERROR_MSG_QUEUE_FULL, errors);
  tcp_handle_postponed_dequeues (wrk);
  tcp_handle_disconnects (wrk);
  vlib_buffer_free (vm, first_buffer, from_frame->n_vectors);

  return from_frame->n_vectors;
}

VLIB_NODE_FN (tcp4_rcv_process_node) (vlib_main_t * vm,
				      vlib_node_runtime_t * node,
				      vlib_frame_t * from_frame)
{
  return tcp46_rcv_process_inline (vm, node, from_frame, 1 /* is_ip4 */ );
}

VLIB_NODE_FN (tcp6_rcv_process_node) (vlib_main_t * vm,
				      vlib_node_runtime_t * node,
				      vlib_frame_t * from_frame)
{
  return tcp46_rcv_process_inline (vm, node, from_frame, 0 /* is_ip4 */ );
}

/* *INDENT-OFF* */
VLIB_REGISTER_NODE (tcp4_rcv_process_node) =
{
  .name = "tcp4-rcv-process",
  /* Takes a vector of packets. */
  .vector_size = sizeof (u32),
  .n_errors = TCP_N_ERROR,
  .error_strings = tcp_error_strings,
  .n_next_nodes = TCP_RCV_PROCESS_N_NEXT,
  .next_nodes =
  {
#define _(s,n) [TCP_RCV_PROCESS_NEXT_##s] = n,
    foreach_tcp_state_next
#undef _
  },
  .format_trace = format_tcp_rx_trace_short,
};
/* *INDENT-ON* */

/* *INDENT-OFF* */
VLIB_REGISTER_NODE (tcp6_rcv_process_node) =
{
  .name = "tcp6-rcv-process",
  /* Takes a vector of packets. */
  .vector_size = sizeof (u32),
  .n_errors = TCP_N_ERROR,
  .error_strings = tcp_error_strings,
  .n_next_nodes = TCP_RCV_PROCESS_N_NEXT,
  .next_nodes =
  {
#define _(s,n) [TCP_RCV_PROCESS_NEXT_##s] = n,
    foreach_tcp_state_next
#undef _
  },
  .format_trace = format_tcp_rx_trace_short,
};
/* *INDENT-ON* */

/**
 * LISTEN state processing as per RFC 793 p. 65
 */
always_inline uword
tcp46_listen_inline (vlib_main_t * vm, vlib_node_runtime_t * node,
		     vlib_frame_t * from_frame, int is_ip4)
{
  u32 n_left_from, *from, n_syns = 0, *first_buffer;
  u32 thread_index = vm->thread_index;

  from = first_buffer = vlib_frame_vector_args (from_frame);
  n_left_from = from_frame->n_vectors;

  while (n_left_from > 0)
    {
      u32 bi, error = TCP_ERROR_NONE;
      tcp_connection_t *lc, *child;
      vlib_buffer_t *b;

      bi = from[0];
      from += 1;
      n_left_from -= 1;

      b = vlib_get_buffer (vm, bi);

      lc = tcp_listener_get (vnet_buffer (b)->tcp.connection_index);
      if (PREDICT_FALSE (lc == 0))
	{
	  tcp_connection_t *tc;
	  tc = tcp_connection_get (vnet_buffer (b)->tcp.connection_index,
				   thread_index);
	  if (tc->state != TCP_STATE_TIME_WAIT)
	    {
	      error = TCP_ERROR_CREATE_EXISTS;
	      goto done;
	    }
	  lc = tcp_lookup_listener (b, tc->c_fib_index, is_ip4);
	  /* clean up the old session */
	  tcp_connection_del (tc);
	}

      /* Make sure connection wasn't just created */
      child = tcp_lookup_connection (lc->c_fib_index, b, thread_index,
				     is_ip4);
      if (PREDICT_FALSE (child->state != TCP_STATE_LISTEN))
	{
	  error = TCP_ERROR_CREATE_EXISTS;
	  goto done;
	}

      /* Create child session. For syn-flood protection use filter */

      /* 1. first check for an RST: handled in dispatch */
      /* if (tcp_rst (th0))
         goto drop;
       */

      /* 2. second check for an ACK: handled in dispatch */
      /* if (tcp_ack (th0))
         {
         tcp_send_reset (b0, is_ip4);
         goto drop;
         }
       */

      /* 3. check for a SYN (did that already) */

      /* Create child session and send SYN-ACK */
      child = tcp_connection_alloc (thread_index);

      if (tcp_options_parse (tcp_buffer_hdr (b), &child->rcv_opts, 1))
	{
	  error = TCP_ERROR_OPTIONS;
	  tcp_connection_free (child);
	  goto done;
	}

      tcp_init_w_buffer (child, b, is_ip4);

      child->state = TCP_STATE_SYN_RCVD;
      child->c_fib_index = lc->c_fib_index;
      child->cc_algo = lc->cc_algo;
      tcp_connection_init_vars (child);
      child->rto = TCP_RTO_MIN;

      if (session_stream_accept (&child->connection, lc->c_s_index,
				 lc->c_thread_index, 0 /* notify */ ))
	{
	  tcp_connection_cleanup (child);
	  error = TCP_ERROR_CREATE_SESSION_FAIL;
	  goto done;
	}

      child->tx_fifo_size = transport_tx_fifo_size (&child->connection);

      tcp_send_synack (child);

      TCP_EVT (TCP_EVT_SYN_RCVD, child, 1);

    done:

      if (PREDICT_FALSE (b->flags & VLIB_BUFFER_IS_TRACED))
	{
	  tcp_rx_trace_t *t;
	  t = vlib_add_trace (vm, node, b, sizeof (*t));
	  clib_memcpy_fast (&t->tcp_header, tcp_buffer_hdr (b),
			    sizeof (t->tcp_header));
	  clib_memcpy_fast (&t->tcp_connection, lc,
			    sizeof (t->tcp_connection));
	}

      n_syns += (error == TCP_ERROR_NONE);
    }

  tcp_inc_counter (listen, TCP_ERROR_SYNS_RCVD, n_syns);
  vlib_buffer_free (vm, first_buffer, from_frame->n_vectors);

  return from_frame->n_vectors;
}

VLIB_NODE_FN (tcp4_listen_node) (vlib_main_t * vm, vlib_node_runtime_t * node,
				 vlib_frame_t * from_frame)
{
  return tcp46_listen_inline (vm, node, from_frame, 1 /* is_ip4 */ );
}

VLIB_NODE_FN (tcp6_listen_node) (vlib_main_t * vm, vlib_node_runtime_t * node,
				 vlib_frame_t * from_frame)
{
  return tcp46_listen_inline (vm, node, from_frame, 0 /* is_ip4 */ );
}

/* *INDENT-OFF* */
VLIB_REGISTER_NODE (tcp4_listen_node) =
{
  .name = "tcp4-listen",
  /* Takes a vector of packets. */
  .vector_size = sizeof (u32),
  .n_errors = TCP_N_ERROR,
  .error_strings = tcp_error_strings,
  .n_next_nodes = TCP_LISTEN_N_NEXT,
  .next_nodes =
  {
#define _(s,n) [TCP_LISTEN_NEXT_##s] = n,
    foreach_tcp_state_next
#undef _
  },
  .format_trace = format_tcp_rx_trace_short,
};
/* *INDENT-ON* */

/* *INDENT-OFF* */
VLIB_REGISTER_NODE (tcp6_listen_node) =
{
  .name = "tcp6-listen",
  /* Takes a vector of packets. */
  .vector_size = sizeof (u32),
  .n_errors = TCP_N_ERROR,
  .error_strings = tcp_error_strings,
  .n_next_nodes = TCP_LISTEN_N_NEXT,
  .next_nodes =
  {
#define _(s,n) [TCP_LISTEN_NEXT_##s] = n,
    foreach_tcp_state_next
#undef _
  },
  .format_trace = format_tcp_rx_trace_short,
};
/* *INDENT-ON* */

typedef enum _tcp_input_next
{
  TCP_INPUT_NEXT_DROP,
  TCP_INPUT_NEXT_LISTEN,
  TCP_INPUT_NEXT_RCV_PROCESS,
  TCP_INPUT_NEXT_SYN_SENT,
  TCP_INPUT_NEXT_ESTABLISHED,
  TCP_INPUT_NEXT_RESET,
  TCP_INPUT_NEXT_PUNT,
  TCP_INPUT_N_NEXT
} tcp_input_next_t;

#define foreach_tcp4_input_next                 \
  _ (DROP, "ip4-drop")                          \
  _ (LISTEN, "tcp4-listen")                     \
  _ (RCV_PROCESS, "tcp4-rcv-process")           \
  _ (SYN_SENT, "tcp4-syn-sent")                 \
  _ (ESTABLISHED, "tcp4-established")		\
  _ (RESET, "tcp4-reset")			\
  _ (PUNT, "ip4-punt")

#define foreach_tcp6_input_next                 \
  _ (DROP, "ip6-drop")                          \
  _ (LISTEN, "tcp6-listen")                     \
  _ (RCV_PROCESS, "tcp6-rcv-process")           \
  _ (SYN_SENT, "tcp6-syn-sent")                 \
  _ (ESTABLISHED, "tcp6-established")		\
  _ (RESET, "tcp6-reset")			\
  _ (PUNT, "ip6-punt")

#define filter_flags (TCP_FLAG_SYN|TCP_FLAG_ACK|TCP_FLAG_RST|TCP_FLAG_FIN)

static void
tcp_input_trace_frame (vlib_main_t * vm, vlib_node_runtime_t * node,
		       vlib_buffer_t ** bs, u32 n_bufs, u8 is_ip4)
{
  tcp_connection_t *tc;
  tcp_header_t *tcp;
  tcp_rx_trace_t *t;
  int i;

  for (i = 0; i < n_bufs; i++)
    {
      if (bs[i]->flags & VLIB_BUFFER_IS_TRACED)
	{
	  t = vlib_add_trace (vm, node, bs[i], sizeof (*t));
	  tc = tcp_connection_get (vnet_buffer (bs[i])->tcp.connection_index,
				   vm->thread_index);
	  tcp = vlib_buffer_get_current (bs[i]);
	  tcp_set_rx_trace_data (t, tc, tcp, bs[i], is_ip4);
	}
    }
}

static void
tcp_input_set_error_next (tcp_main_t * tm, u16 * next, u32 * error, u8 is_ip4)
{
  if (*error == TCP_ERROR_FILTERED || *error == TCP_ERROR_WRONG_THREAD)
    {
      *next = TCP_INPUT_NEXT_DROP;
    }
  else if ((is_ip4 && tm->punt_unknown4) || (!is_ip4 && tm->punt_unknown6))
    {
      *next = TCP_INPUT_NEXT_PUNT;
      *error = TCP_ERROR_PUNT;
    }
  else
    {
      *next = TCP_INPUT_NEXT_RESET;
      *error = TCP_ERROR_NO_LISTENER;
    }
}

static inline void
tcp_input_dispatch_buffer (tcp_main_t * tm, tcp_connection_t * tc,
			   vlib_buffer_t * b, u16 * next,
			   vlib_node_runtime_t * error_node)
{
  tcp_header_t *tcp;
  u32 error;
  u8 flags;

  tcp = tcp_buffer_hdr (b);
  flags = tcp->flags & filter_flags;
  *next = tm->dispatch_table[tc->state][flags].next;
  error = tm->dispatch_table[tc->state][flags].error;
  tc->segs_in += 1;

  if (PREDICT_FALSE (error != TCP_ERROR_NONE))
    {
      b->error = error_node->errors[error];
      if (error == TCP_ERROR_DISPATCH)
	clib_warning ("tcp conn %u disp error state %U flags %U",
		      tc->c_c_index, format_tcp_state, tc->state,
		      format_tcp_flags, (int) flags);
    }
}

always_inline uword
tcp46_input_inline (vlib_main_t * vm, vlib_node_runtime_t * node,
		    vlib_frame_t * frame, int is_ip4, u8 is_nolookup)
{
  u32 n_left_from, *from, thread_index = vm->thread_index;
  tcp_main_t *tm = vnet_get_tcp_main ();
  vlib_buffer_t *bufs[VLIB_FRAME_SIZE], **b;
  u16 nexts[VLIB_FRAME_SIZE], *next;

  tcp_set_time_now (tcp_get_worker (thread_index));

  from = vlib_frame_vector_args (frame);
  n_left_from = frame->n_vectors;
  vlib_get_buffers (vm, from, bufs, n_left_from);

  b = bufs;
  next = nexts;

  while (n_left_from >= 4)
    {
      u32 error0 = TCP_ERROR_NO_LISTENER, error1 = TCP_ERROR_NO_LISTENER;
      tcp_connection_t *tc0, *tc1;

      {
	vlib_prefetch_buffer_header (b[2], STORE);
	CLIB_PREFETCH (b[2]->data, 2 * CLIB_CACHE_LINE_BYTES, LOAD);

	vlib_prefetch_buffer_header (b[3], STORE);
	CLIB_PREFETCH (b[3]->data, 2 * CLIB_CACHE_LINE_BYTES, LOAD);
      }

      next[0] = next[1] = TCP_INPUT_NEXT_DROP;

      tc0 = tcp_input_lookup_buffer (b[0], thread_index, &error0, is_ip4,
				     is_nolookup);
      tc1 = tcp_input_lookup_buffer (b[1], thread_index, &error1, is_ip4,
				     is_nolookup);

      if (PREDICT_TRUE (!tc0 + !tc1 == 0))
	{
	  ASSERT (tcp_lookup_is_valid (tc0, b[0], tcp_buffer_hdr (b[0])));
	  ASSERT (tcp_lookup_is_valid (tc1, b[1], tcp_buffer_hdr (b[1])));

	  vnet_buffer (b[0])->tcp.connection_index = tc0->c_c_index;
	  vnet_buffer (b[1])->tcp.connection_index = tc1->c_c_index;

	  tcp_input_dispatch_buffer (tm, tc0, b[0], &next[0], node);
	  tcp_input_dispatch_buffer (tm, tc1, b[1], &next[1], node);
	}
      else
	{
	  if (PREDICT_TRUE (tc0 != 0))
	    {
	      ASSERT (tcp_lookup_is_valid (tc0, b[0], tcp_buffer_hdr (b[0])));
	      vnet_buffer (b[0])->tcp.connection_index = tc0->c_c_index;
	      tcp_input_dispatch_buffer (tm, tc0, b[0], &next[0], node);
	    }
	  else
	    {
	      tcp_input_set_error_next (tm, &next[0], &error0, is_ip4);
	      b[0]->error = node->errors[error0];
	    }

	  if (PREDICT_TRUE (tc1 != 0))
	    {
	      ASSERT (tcp_lookup_is_valid (tc1, b[1], tcp_buffer_hdr (b[1])));
	      vnet_buffer (b[1])->tcp.connection_index = tc1->c_c_index;
	      tcp_input_dispatch_buffer (tm, tc1, b[1], &next[1], node);
	    }
	  else
	    {
	      tcp_input_set_error_next (tm, &next[1], &error1, is_ip4);
	      b[1]->error = node->errors[error1];
	    }
	}

      b += 2;
      next += 2;
      n_left_from -= 2;
    }
  while (n_left_from > 0)
    {
      tcp_connection_t *tc0;
      u32 error0 = TCP_ERROR_NO_LISTENER;

      if (n_left_from > 1)
	{
	  vlib_prefetch_buffer_header (b[1], STORE);
	  CLIB_PREFETCH (b[1]->data, 2 * CLIB_CACHE_LINE_BYTES, LOAD);
	}

      next[0] = TCP_INPUT_NEXT_DROP;
      tc0 = tcp_input_lookup_buffer (b[0], thread_index, &error0, is_ip4,
				     is_nolookup);
      if (PREDICT_TRUE (tc0 != 0))
	{
	  ASSERT (tcp_lookup_is_valid (tc0, b[0], tcp_buffer_hdr (b[0])));
	  vnet_buffer (b[0])->tcp.connection_index = tc0->c_c_index;
	  tcp_input_dispatch_buffer (tm, tc0, b[0], &next[0], node);
	}
      else
	{
	  tcp_input_set_error_next (tm, &next[0], &error0, is_ip4);
	  b[0]->error = node->errors[error0];
	}

      b += 1;
      next += 1;
      n_left_from -= 1;
    }

  if (PREDICT_FALSE (node->flags & VLIB_NODE_FLAG_TRACE))
    tcp_input_trace_frame (vm, node, bufs, frame->n_vectors, is_ip4);

  vlib_buffer_enqueue_to_next (vm, node, from, nexts, frame->n_vectors);
  return frame->n_vectors;
}

VLIB_NODE_FN (tcp4_input_nolookup_node) (vlib_main_t * vm,
					 vlib_node_runtime_t * node,
					 vlib_frame_t * from_frame)
{
  return tcp46_input_inline (vm, node, from_frame, 1 /* is_ip4 */ ,
			     1 /* is_nolookup */ );
}

VLIB_NODE_FN (tcp6_input_nolookup_node) (vlib_main_t * vm,
					 vlib_node_runtime_t * node,
					 vlib_frame_t * from_frame)
{
  return tcp46_input_inline (vm, node, from_frame, 0 /* is_ip4 */ ,
			     1 /* is_nolookup */ );
}

/* *INDENT-OFF* */
VLIB_REGISTER_NODE (tcp4_input_nolookup_node) =
{
  .name = "tcp4-input-nolookup",
  /* Takes a vector of packets. */
  .vector_size = sizeof (u32),
  .n_errors = TCP_N_ERROR,
  .error_strings = tcp_error_strings,
  .n_next_nodes = TCP_INPUT_N_NEXT,
  .next_nodes =
  {
#define _(s,n) [TCP_INPUT_NEXT_##s] = n,
    foreach_tcp4_input_next
#undef _
  },
  .format_buffer = format_tcp_header,
  .format_trace = format_tcp_rx_trace,
};
/* *INDENT-ON* */

/* *INDENT-OFF* */
VLIB_REGISTER_NODE (tcp6_input_nolookup_node) =
{
  .name = "tcp6-input-nolookup",
  /* Takes a vector of packets. */
  .vector_size = sizeof (u32),
  .n_errors = TCP_N_ERROR,
  .error_strings = tcp_error_strings,
  .n_next_nodes = TCP_INPUT_N_NEXT,
  .next_nodes =
  {
#define _(s,n) [TCP_INPUT_NEXT_##s] = n,
    foreach_tcp6_input_next
#undef _
  },
  .format_buffer = format_tcp_header,
  .format_trace = format_tcp_rx_trace,
};
/* *INDENT-ON* */

VLIB_NODE_FN (tcp4_input_node) (vlib_main_t * vm, vlib_node_runtime_t * node,
				vlib_frame_t * from_frame)
{
  return tcp46_input_inline (vm, node, from_frame, 1 /* is_ip4 */ ,
			     0 /* is_nolookup */ );
}

VLIB_NODE_FN (tcp6_input_node) (vlib_main_t * vm, vlib_node_runtime_t * node,
				vlib_frame_t * from_frame)
{
  return tcp46_input_inline (vm, node, from_frame, 0 /* is_ip4 */ ,
			     0 /* is_nolookup */ );
}

/* *INDENT-OFF* */
VLIB_REGISTER_NODE (tcp4_input_node) =
{
  .name = "tcp4-input",
  /* Takes a vector of packets. */
  .vector_size = sizeof (u32),
  .n_errors = TCP_N_ERROR,
  .error_strings = tcp_error_strings,
  .n_next_nodes = TCP_INPUT_N_NEXT,
  .next_nodes =
  {
#define _(s,n) [TCP_INPUT_NEXT_##s] = n,
    foreach_tcp4_input_next
#undef _
  },
  .format_buffer = format_tcp_header,
  .format_trace = format_tcp_rx_trace,
};
/* *INDENT-ON* */

/* *INDENT-OFF* */
VLIB_REGISTER_NODE (tcp6_input_node) =
{
  .name = "tcp6-input",
  /* Takes a vector of packets. */
  .vector_size = sizeof (u32),
  .n_errors = TCP_N_ERROR,
  .error_strings = tcp_error_strings,
  .n_next_nodes = TCP_INPUT_N_NEXT,
  .next_nodes =
  {
#define _(s,n) [TCP_INPUT_NEXT_##s] = n,
    foreach_tcp6_input_next
#undef _
  },
  .format_buffer = format_tcp_header,
  .format_trace = format_tcp_rx_trace,
};
/* *INDENT-ON* */

#ifndef CLIB_MARCH_VARIANT
static void
tcp_dispatch_table_init (tcp_main_t * tm)
{
  int i, j;
  for (i = 0; i < ARRAY_LEN (tm->dispatch_table); i++)
    for (j = 0; j < ARRAY_LEN (tm->dispatch_table[i]); j++)
      {
	tm->dispatch_table[i][j].next = TCP_INPUT_NEXT_DROP;
	tm->dispatch_table[i][j].error = TCP_ERROR_DISPATCH;
      }

#define _(t,f,n,e)                                           	\
do {                                                       	\
    tm->dispatch_table[TCP_STATE_##t][f].next = (n);         	\
    tm->dispatch_table[TCP_STATE_##t][f].error = (e);        	\
} while (0)

  /* RFC 793: In LISTEN if RST drop and if ACK return RST */
  _(LISTEN, 0, TCP_INPUT_NEXT_DROP, TCP_ERROR_SEGMENT_INVALID);
  _(LISTEN, TCP_FLAG_ACK, TCP_INPUT_NEXT_RESET, TCP_ERROR_ACK_INVALID);
  _(LISTEN, TCP_FLAG_RST, TCP_INPUT_NEXT_DROP, TCP_ERROR_INVALID_CONNECTION);
  _(LISTEN, TCP_FLAG_SYN, TCP_INPUT_NEXT_LISTEN, TCP_ERROR_NONE);
  _(LISTEN, TCP_FLAG_SYN | TCP_FLAG_ACK, TCP_INPUT_NEXT_RESET,
    TCP_ERROR_ACK_INVALID);
  _(LISTEN, TCP_FLAG_SYN | TCP_FLAG_RST, TCP_INPUT_NEXT_DROP,
    TCP_ERROR_SEGMENT_INVALID);
  _(LISTEN, TCP_FLAG_SYN | TCP_FLAG_RST | TCP_FLAG_ACK, TCP_INPUT_NEXT_DROP,
    TCP_ERROR_SEGMENT_INVALID);
  _(LISTEN, TCP_FLAG_RST | TCP_FLAG_ACK, TCP_INPUT_NEXT_DROP,
    TCP_ERROR_INVALID_CONNECTION);
  _(LISTEN, TCP_FLAG_FIN, TCP_INPUT_NEXT_RESET, TCP_ERROR_SEGMENT_INVALID);
  _(LISTEN, TCP_FLAG_FIN | TCP_FLAG_ACK, TCP_INPUT_NEXT_RESET,
    TCP_ERROR_SEGMENT_INVALID);
  _(LISTEN, TCP_FLAG_FIN | TCP_FLAG_RST, TCP_INPUT_NEXT_DROP,
    TCP_ERROR_SEGMENT_INVALID);
  _(LISTEN, TCP_FLAG_FIN | TCP_FLAG_RST | TCP_FLAG_ACK, TCP_INPUT_NEXT_DROP,
    TCP_ERROR_SEGMENT_INVALID);
  _(LISTEN, TCP_FLAG_FIN | TCP_FLAG_SYN, TCP_INPUT_NEXT_DROP,
    TCP_ERROR_SEGMENT_INVALID);
  _(LISTEN, TCP_FLAG_FIN | TCP_FLAG_SYN | TCP_FLAG_ACK, TCP_INPUT_NEXT_DROP,
    TCP_ERROR_SEGMENT_INVALID);
  _(LISTEN, TCP_FLAG_FIN | TCP_FLAG_SYN | TCP_FLAG_RST, TCP_INPUT_NEXT_DROP,
    TCP_ERROR_SEGMENT_INVALID);
  _(LISTEN, TCP_FLAG_FIN | TCP_FLAG_SYN | TCP_FLAG_RST | TCP_FLAG_ACK,
    TCP_INPUT_NEXT_DROP, TCP_ERROR_SEGMENT_INVALID);
  /* ACK for for a SYN-ACK -> tcp-rcv-process. */
  _(SYN_RCVD, TCP_FLAG_ACK, TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(SYN_RCVD, TCP_FLAG_RST, TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(SYN_RCVD, TCP_FLAG_RST | TCP_FLAG_ACK, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(SYN_RCVD, TCP_FLAG_SYN, TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(SYN_RCVD, TCP_FLAG_SYN | TCP_FLAG_ACK, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(SYN_RCVD, TCP_FLAG_SYN | TCP_FLAG_RST, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(SYN_RCVD, TCP_FLAG_SYN | TCP_FLAG_RST | TCP_FLAG_ACK,
    TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(SYN_RCVD, TCP_FLAG_FIN, TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(SYN_RCVD, TCP_FLAG_FIN | TCP_FLAG_ACK, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(SYN_RCVD, TCP_FLAG_FIN | TCP_FLAG_RST, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(SYN_RCVD, TCP_FLAG_FIN | TCP_FLAG_RST | TCP_FLAG_ACK,
    TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(SYN_RCVD, TCP_FLAG_FIN | TCP_FLAG_SYN, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(SYN_RCVD, TCP_FLAG_FIN | TCP_FLAG_SYN | TCP_FLAG_RST,
    TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(SYN_RCVD, TCP_FLAG_FIN | TCP_FLAG_SYN | TCP_FLAG_ACK,
    TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(SYN_RCVD, TCP_FLAG_FIN | TCP_FLAG_SYN | TCP_FLAG_RST | TCP_FLAG_ACK,
    TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(SYN_RCVD, 0, TCP_INPUT_NEXT_DROP, TCP_ERROR_SEGMENT_INVALID);
  /* SYN-ACK for a SYN */
  _(SYN_SENT, TCP_FLAG_SYN | TCP_FLAG_ACK, TCP_INPUT_NEXT_SYN_SENT,
    TCP_ERROR_NONE);
  _(SYN_SENT, TCP_FLAG_ACK, TCP_INPUT_NEXT_SYN_SENT, TCP_ERROR_NONE);
  _(SYN_SENT, TCP_FLAG_RST, TCP_INPUT_NEXT_SYN_SENT, TCP_ERROR_NONE);
  _(SYN_SENT, TCP_FLAG_RST | TCP_FLAG_ACK, TCP_INPUT_NEXT_SYN_SENT,
    TCP_ERROR_NONE);
  _(SYN_SENT, TCP_FLAG_FIN, TCP_INPUT_NEXT_SYN_SENT, TCP_ERROR_NONE);
  _(SYN_SENT, TCP_FLAG_FIN | TCP_FLAG_ACK, TCP_INPUT_NEXT_SYN_SENT,
    TCP_ERROR_NONE);
  /* ACK for for established connection -> tcp-established. */
  _(ESTABLISHED, TCP_FLAG_ACK, TCP_INPUT_NEXT_ESTABLISHED, TCP_ERROR_NONE);
  /* FIN for for established connection -> tcp-established. */
  _(ESTABLISHED, TCP_FLAG_FIN, TCP_INPUT_NEXT_ESTABLISHED, TCP_ERROR_NONE);
  _(ESTABLISHED, TCP_FLAG_FIN | TCP_FLAG_ACK, TCP_INPUT_NEXT_ESTABLISHED,
    TCP_ERROR_NONE);
  _(ESTABLISHED, TCP_FLAG_FIN | TCP_FLAG_RST, TCP_INPUT_NEXT_ESTABLISHED,
    TCP_ERROR_NONE);
  _(ESTABLISHED, TCP_FLAG_FIN | TCP_FLAG_RST | TCP_FLAG_ACK,
    TCP_INPUT_NEXT_ESTABLISHED, TCP_ERROR_NONE);
  _(ESTABLISHED, TCP_FLAG_FIN | TCP_FLAG_SYN, TCP_INPUT_NEXT_ESTABLISHED,
    TCP_ERROR_NONE);
  _(ESTABLISHED, TCP_FLAG_FIN | TCP_FLAG_SYN | TCP_FLAG_ACK,
    TCP_INPUT_NEXT_ESTABLISHED, TCP_ERROR_NONE);
  _(ESTABLISHED, TCP_FLAG_FIN | TCP_FLAG_SYN | TCP_FLAG_RST,
    TCP_INPUT_NEXT_ESTABLISHED, TCP_ERROR_NONE);
  _(ESTABLISHED, TCP_FLAG_FIN | TCP_FLAG_SYN | TCP_FLAG_RST | TCP_FLAG_ACK,
    TCP_INPUT_NEXT_ESTABLISHED, TCP_ERROR_NONE);
  _(ESTABLISHED, TCP_FLAG_RST, TCP_INPUT_NEXT_ESTABLISHED, TCP_ERROR_NONE);
  _(ESTABLISHED, TCP_FLAG_RST | TCP_FLAG_ACK, TCP_INPUT_NEXT_ESTABLISHED,
    TCP_ERROR_NONE);
  _(ESTABLISHED, TCP_FLAG_SYN, TCP_INPUT_NEXT_ESTABLISHED, TCP_ERROR_NONE);
  _(ESTABLISHED, TCP_FLAG_SYN | TCP_FLAG_ACK, TCP_INPUT_NEXT_ESTABLISHED,
    TCP_ERROR_NONE);
  _(ESTABLISHED, TCP_FLAG_SYN | TCP_FLAG_RST, TCP_INPUT_NEXT_ESTABLISHED,
    TCP_ERROR_NONE);
  _(ESTABLISHED, TCP_FLAG_SYN | TCP_FLAG_RST | TCP_FLAG_ACK,
    TCP_INPUT_NEXT_ESTABLISHED, TCP_ERROR_NONE);
  _(ESTABLISHED, 0, TCP_INPUT_NEXT_DROP, TCP_ERROR_SEGMENT_INVALID);
  /* ACK or FIN-ACK to our FIN */
  _(FIN_WAIT_1, TCP_FLAG_ACK, TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(FIN_WAIT_1, TCP_FLAG_ACK | TCP_FLAG_FIN, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  /* FIN in reply to our FIN from the other side */
  _(FIN_WAIT_1, 0, TCP_INPUT_NEXT_DROP, TCP_ERROR_SEGMENT_INVALID);
  _(FIN_WAIT_1, TCP_FLAG_FIN, TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(FIN_WAIT_1, TCP_FLAG_FIN | TCP_FLAG_SYN, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(FIN_WAIT_1, TCP_FLAG_FIN | TCP_FLAG_SYN | TCP_FLAG_ACK,
    TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(FIN_WAIT_1, TCP_FLAG_FIN | TCP_FLAG_SYN | TCP_FLAG_RST,
    TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(FIN_WAIT_1, TCP_FLAG_FIN | TCP_FLAG_SYN | TCP_FLAG_RST | TCP_FLAG_ACK,
    TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(FIN_WAIT_1, TCP_FLAG_FIN | TCP_FLAG_RST, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(FIN_WAIT_1, TCP_FLAG_FIN | TCP_FLAG_RST | TCP_FLAG_ACK,
    TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(FIN_WAIT_1, TCP_FLAG_SYN, TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(FIN_WAIT_1, TCP_FLAG_SYN | TCP_FLAG_RST, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(FIN_WAIT_1, TCP_FLAG_SYN | TCP_FLAG_ACK, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(FIN_WAIT_1, TCP_FLAG_SYN | TCP_FLAG_RST | TCP_FLAG_ACK,
    TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(FIN_WAIT_1, TCP_FLAG_RST, TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(FIN_WAIT_1, TCP_FLAG_RST | TCP_FLAG_ACK, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(CLOSING, 0, TCP_INPUT_NEXT_DROP, TCP_ERROR_SEGMENT_INVALID);
  _(CLOSING, TCP_FLAG_ACK, TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(CLOSING, TCP_FLAG_SYN, TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(CLOSING, TCP_FLAG_SYN | TCP_FLAG_RST, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(CLOSING, TCP_FLAG_SYN | TCP_FLAG_ACK, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(CLOSING, TCP_FLAG_SYN | TCP_FLAG_RST | TCP_FLAG_ACK,
    TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(CLOSING, TCP_FLAG_RST, TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(CLOSING, TCP_FLAG_RST | TCP_FLAG_ACK, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(CLOSING, TCP_FLAG_FIN, TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(CLOSING, TCP_FLAG_FIN | TCP_FLAG_ACK, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(CLOSING, TCP_FLAG_FIN | TCP_FLAG_RST, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(CLOSING, TCP_FLAG_FIN | TCP_FLAG_RST | TCP_FLAG_ACK,
    TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(CLOSING, TCP_FLAG_FIN | TCP_FLAG_SYN, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(CLOSING, TCP_FLAG_FIN | TCP_FLAG_SYN | TCP_FLAG_ACK,
    TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(CLOSING, TCP_FLAG_FIN | TCP_FLAG_SYN | TCP_FLAG_RST,
    TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(CLOSING, TCP_FLAG_FIN | TCP_FLAG_SYN | TCP_FLAG_RST | TCP_FLAG_ACK,
    TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  /* FIN confirming that the peer (app) has closed */
  _(FIN_WAIT_2, TCP_FLAG_FIN, TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(FIN_WAIT_2, TCP_FLAG_ACK, TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(FIN_WAIT_2, TCP_FLAG_FIN | TCP_FLAG_ACK, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(FIN_WAIT_2, TCP_FLAG_RST, TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(FIN_WAIT_2, TCP_FLAG_RST | TCP_FLAG_ACK, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(CLOSE_WAIT, TCP_FLAG_ACK, TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(CLOSE_WAIT, TCP_FLAG_FIN | TCP_FLAG_ACK, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(CLOSE_WAIT, TCP_FLAG_RST, TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(CLOSE_WAIT, TCP_FLAG_RST | TCP_FLAG_ACK, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(LAST_ACK, 0, TCP_INPUT_NEXT_DROP, TCP_ERROR_SEGMENT_INVALID);
  _(LAST_ACK, TCP_FLAG_ACK, TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(LAST_ACK, TCP_FLAG_FIN, TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(LAST_ACK, TCP_FLAG_FIN | TCP_FLAG_ACK, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(LAST_ACK, TCP_FLAG_FIN | TCP_FLAG_SYN, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(LAST_ACK, TCP_FLAG_FIN | TCP_FLAG_SYN | TCP_FLAG_ACK,
    TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(LAST_ACK, TCP_FLAG_FIN | TCP_FLAG_RST, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(LAST_ACK, TCP_FLAG_FIN | TCP_FLAG_RST | TCP_FLAG_ACK,
    TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(LAST_ACK, TCP_FLAG_FIN | TCP_FLAG_SYN | TCP_FLAG_RST,
    TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(LAST_ACK, TCP_FLAG_FIN | TCP_FLAG_SYN | TCP_FLAG_RST | TCP_FLAG_ACK,
    TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(LAST_ACK, TCP_FLAG_RST, TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(LAST_ACK, TCP_FLAG_RST | TCP_FLAG_ACK, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(LAST_ACK, TCP_FLAG_SYN, TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(LAST_ACK, TCP_FLAG_SYN | TCP_FLAG_ACK, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(LAST_ACK, TCP_FLAG_SYN | TCP_FLAG_RST, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(LAST_ACK, TCP_FLAG_SYN | TCP_FLAG_RST | TCP_FLAG_ACK,
    TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(TIME_WAIT, TCP_FLAG_SYN, TCP_INPUT_NEXT_LISTEN, TCP_ERROR_NONE);
  _(TIME_WAIT, TCP_FLAG_FIN, TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(TIME_WAIT, TCP_FLAG_FIN | TCP_FLAG_ACK, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(TIME_WAIT, TCP_FLAG_RST, TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  _(TIME_WAIT, TCP_FLAG_RST | TCP_FLAG_ACK, TCP_INPUT_NEXT_RCV_PROCESS,
    TCP_ERROR_NONE);
  _(TIME_WAIT, TCP_FLAG_ACK, TCP_INPUT_NEXT_RCV_PROCESS, TCP_ERROR_NONE);
  /* RFC793 CLOSED: An incoming segment containing a RST is discarded. An
   * incoming segment not containing a RST causes a RST to be sent in
   * response.*/
  _(CLOSED, TCP_FLAG_RST, TCP_INPUT_NEXT_DROP, TCP_ERROR_CONNECTION_CLOSED);
  _(CLOSED, TCP_FLAG_RST | TCP_FLAG_ACK, TCP_INPUT_NEXT_DROP,
    TCP_ERROR_CONNECTION_CLOSED);
  _(CLOSED, TCP_FLAG_ACK, TCP_INPUT_NEXT_RESET, TCP_ERROR_CONNECTION_CLOSED);
  _(CLOSED, TCP_FLAG_SYN, TCP_INPUT_NEXT_RESET, TCP_ERROR_CONNECTION_CLOSED);
  _(CLOSED, TCP_FLAG_FIN | TCP_FLAG_ACK, TCP_INPUT_NEXT_RESET,
    TCP_ERROR_CONNECTION_CLOSED);
#undef _
}

static clib_error_t *
tcp_input_init (vlib_main_t * vm)
{
  clib_error_t *error = 0;
  tcp_main_t *tm = vnet_get_tcp_main ();

  if ((error = vlib_call_init_function (vm, tcp_init)))
    return error;

  /* Initialize dispatch table. */
  tcp_dispatch_table_init (tm);

  return error;
}

VLIB_INIT_FUNCTION (tcp_input_init);

#endif /* CLIB_MARCH_VARIANT */

/*
 * fd.io coding-style-patch-verification: ON
 *
 * Local Variables:
 * eval: (c-set-style "gnu")
 * End:
 */