summaryrefslogtreecommitdiff
path: root/src/common/socket.c
blob: c74cd4661d6f2e96faf8c8ed8027f7a4f7bcdfc3 (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
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
/**
 * This file is part of Hercules.
 * http://herc.ws - http://github.com/HerculesWS/Hercules
 *
 * Copyright (C) 2012-2016  Hercules Dev Team
 * Copyright (C)  Athena Dev Teams
 *
 * Hercules is free software: you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation, either version 3 of the License, or
 * (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program.  If not, see <http://www.gnu.org/licenses/>.
 */
#define HERCULES_CORE

#include "config/core.h" // SHOW_SERVER_STATS
#include "socket.h"

#include "common/HPM.h"
#include "common/cbasetypes.h"
#include "common/conf.h"
#include "common/db.h"
#include "common/memmgr.h"
#include "common/mmo.h"
#include "common/nullpo.h"
#include "common/showmsg.h"
#include "common/strlib.h"
#include "common/timer.h"

#include <stdio.h>
#include <stdlib.h>
#include <sys/types.h>

#ifdef SOCKET_EPOLL
#include <sys/epoll.h>
#endif  // SOCKET_EPOLL

#ifdef WIN32
#	include "common/winapi.h"
#else  // WIN32
#	include <arpa/inet.h>
#	include <errno.h>
#	include <net/if.h>
#	include <netdb.h>
#if defined __linux__ || defined __linux
#       include <linux/tcp.h>
#else  // defined __linux__ || defined __linux
#	include <netinet/in.h>
#	include <netinet/tcp.h>
#endif  // defined __linux__ || defined __linux
#	include <sys/ioctl.h>
#	include <sys/socket.h>
#	include <sys/time.h>
#	include <unistd.h>

#ifndef SIOCGIFCONF
#	include <sys/sockio.h> // SIOCGIFCONF on Solaris, maybe others? [Shinomori]
#endif  // SIOCGIFCONF
#ifndef FIONBIO
#	include <sys/filio.h> // FIONBIO on Solaris [FlavioJS]
#endif  // FIONBIO

#ifdef HAVE_SETRLIMIT
#	include <sys/resource.h>
#endif  // HAVE_SETRLIMIT
#endif  // WIN32

/**
 * Socket Interface Source
 **/
struct socket_interface sockt_s;
struct socket_interface *sockt;

struct socket_data **session;

const char *SOCKET_CONF_FILENAME = "conf/common/socket.conf";

#ifdef SEND_SHORTLIST
// Add a fd to the shortlist so that it'll be recognized as a fd that needs
// sending done on it.
void send_shortlist_add_fd(int fd);
// Do pending network sends (and eof handling) from the shortlist.
void send_shortlist_do_sends(void);
#endif  // SEND_SHORTLIST

/////////////////////////////////////////////////////////////////////
#if defined(WIN32)
/////////////////////////////////////////////////////////////////////
// windows portability layer

typedef int socklen_t;

#define sErrno WSAGetLastError()
#define S_ENOTSOCK WSAENOTSOCK
#define S_EWOULDBLOCK WSAEWOULDBLOCK
#define S_EINTR WSAEINTR
#define S_ECONNABORTED WSAECONNABORTED

#define SHUT_RD   SD_RECEIVE
#define SHUT_WR   SD_SEND
#define SHUT_RDWR SD_BOTH

// global array of sockets (emulating linux)
// fd is the position in the array
static SOCKET sock_arr[FD_SETSIZE];
static int sock_arr_len = 0;

/// Returns the socket associated with the target fd.
///
/// @param fd Target fd.
/// @return Socket
#define fd2sock(fd) sock_arr[fd]

/// Returns the first fd associated with the socket.
/// Returns -1 if the socket is not found.
///
/// @param s Socket
/// @return Fd or -1
int sock2fd(SOCKET s)
{
	int fd;

	// search for the socket
	for( fd = 1; fd < sock_arr_len; ++fd )
		if( sock_arr[fd] == s )
			break;// found the socket
	if( fd == sock_arr_len )
		return -1;// not found
	return fd;
}

/// Inserts the socket into the global array of sockets.
/// Returns a new fd associated with the socket.
/// If there are too many sockets it closes the socket, sets an error and
//  returns -1 instead.
/// Since fd 0 is reserved, it returns values in the range [1,FD_SETSIZE[.
///
/// @param s Socket
/// @return New fd or -1
int sock2newfd(SOCKET s)
{
	int fd;

	// find an empty position
	for( fd = 1; fd < sock_arr_len; ++fd )
		if( sock_arr[fd] == INVALID_SOCKET )
			break;// empty position
	if( fd == ARRAYLENGTH(sock_arr) )
	{// too many sockets
		closesocket(s);
		WSASetLastError(WSAEMFILE);
		return -1;
	}
	sock_arr[fd] = s;
	if( sock_arr_len <= fd )
		sock_arr_len = fd+1;
	return fd;
}

int sAccept(int fd, struct sockaddr* addr, int* addrlen)
{
	SOCKET s;

	// accept connection
	s = accept(fd2sock(fd), addr, addrlen);
	if( s == INVALID_SOCKET )
		return -1;// error
	return sock2newfd(s);
}

int sClose(int fd)
{
	int ret = closesocket(fd2sock(fd));
	fd2sock(fd) = INVALID_SOCKET;
	return ret;
}

int sSocket(int af, int type, int protocol)
{
	SOCKET s;

	// create socket
	s = socket(af,type,protocol);
	if( s == INVALID_SOCKET )
		return -1;// error
	return sock2newfd(s);
}

char* sErr(int code)
{
	static char sbuf[512];
	// strerror does not handle socket codes
	if( FormatMessage(FORMAT_MESSAGE_FROM_SYSTEM|FORMAT_MESSAGE_IGNORE_INSERTS, NULL,
			code, MAKELANGID(LANG_ENGLISH, SUBLANG_DEFAULT), (LPTSTR)&sbuf, sizeof(sbuf), NULL) == 0 )
		snprintf(sbuf, sizeof(sbuf), "unknown error");
	return sbuf;
}

#define sBind(fd,name,namelen)                      bind(fd2sock(fd),(name),(namelen))
#define sConnect(fd,name,namelen)                   connect(fd2sock(fd),(name),(namelen))
#define sIoctl(fd,cmd,argp)                         ioctlsocket(fd2sock(fd),(cmd),(argp))
#define sListen(fd,backlog)                         listen(fd2sock(fd),(backlog))
#define sRecv(fd,buf,len,flags)                     recv(fd2sock(fd),(buf),(len),(flags))
#define sSelect                                     select
#define sSend(fd,buf,len,flags)                     send(fd2sock(fd),(buf),(len),(flags))
#define sSetsockopt(fd,level,optname,optval,optlen) setsockopt(fd2sock(fd),(level),(optname),(optval),(optlen))
#define sShutdown(fd,how)                           shutdown(fd2sock(fd),(how))
#define sFD_SET(fd,set)                             FD_SET(fd2sock(fd),(set))
#define sFD_CLR(fd,set)                             FD_CLR(fd2sock(fd),(set))
#define sFD_ISSET(fd,set)                           FD_ISSET(fd2sock(fd),(set))
#define sFD_ZERO                                    FD_ZERO

/////////////////////////////////////////////////////////////////////
#else  // defined(WIN32)
/////////////////////////////////////////////////////////////////////
// nix portability layer

#define SOCKET_ERROR (-1)

#define sErrno errno
#define S_ENOTSOCK EBADF
#define S_EWOULDBLOCK EAGAIN
#define S_EINTR EINTR
#define S_ECONNABORTED ECONNABORTED

#define sAccept accept
#define sClose close
#define sSocket socket
#define sErr strerror

#define sBind bind
#define sConnect connect
#define sIoctl ioctl
#define sListen listen
#define sRecv recv
#define sSelect select
#define sSend send
#define sSetsockopt setsockopt
#define sShutdown shutdown
#define sFD_SET FD_SET
#define sFD_CLR FD_CLR
#define sFD_ISSET FD_ISSET
#define sFD_ZERO FD_ZERO

/////////////////////////////////////////////////////////////////////
#endif  // defined(WIN32)
/////////////////////////////////////////////////////////////////////

#ifndef MSG_NOSIGNAL
	#define MSG_NOSIGNAL 0
#endif  // MSG_NOSIGNAL

#ifndef SOCKET_EPOLL
// Select based Event Dispatcher:
fd_set readfds;

#else  // SOCKET_EPOLL
// Epoll based Event Dispatcher:
static int epoll_maxevents = (FD_SETSIZE / 2);
static int epfd = SOCKET_ERROR;
static struct epoll_event epevent;
static struct epoll_event *epevents = NULL;

#endif  // SOCKET_EPOLL

// Maximum packet size in bytes, which the client is able to handle.
// Larger packets cause a buffer overflow and stack corruption.
#if PACKETVER >= 20131223
static size_t socket_max_client_packet = 0xFFFF;
#else  // PACKETVER >= 20131223
static size_t socket_max_client_packet = 0x6000;
#endif  // PACKETVER >= 20131223

#ifdef SHOW_SERVER_STATS
// Data I/O statistics
static size_t socket_data_i = 0, socket_data_ci = 0, socket_data_qi = 0;
static size_t socket_data_o = 0, socket_data_co = 0, socket_data_qo = 0;
static time_t socket_data_last_tick = 0;
#endif  // SHOW_SERVER_STATS

// initial recv buffer size (this will also be the max. size)
// biggest known packet: S 0153 <len>.w <emblem data>.?B -> 24x24 256 color .bmp (0153 + len.w + 1618/1654/1756 bytes)
#define RFIFO_SIZE (2*1024)
// initial send buffer size (will be resized as needed)
#define WFIFO_SIZE (16*1024)

// Maximum size of pending data in the write fifo. (for non-server connections)
// The connection is closed if it goes over the limit.
#define WFIFO_MAX (1*1024*1024)

#ifdef SEND_SHORTLIST
int send_shortlist_array[FD_SETSIZE];// we only support FD_SETSIZE sockets, limit the array to that
int send_shortlist_count = 0;// how many fd's are in the shortlist
uint32 send_shortlist_set[(FD_SETSIZE+31)/32];// to know if specific fd's are already in the shortlist
#endif  // SEND_SHORTLIST

static int create_session(int fd, RecvFunc func_recv, SendFunc func_send, ParseFunc func_parse);

#ifndef MINICORE
	int ip_rules = 1;
	static int connect_check(uint32 ip);
#endif  // MINICORE

const char* error_msg(void)
{
	static char buf[512];
	int code = sErrno;
	snprintf(buf, sizeof(buf), "error %d: %s", code, sErr(code));
	return buf;
}

/*======================================
 * CORE : Default processing functions
 *--------------------------------------*/
int null_recv(int fd) { return 0; }
int null_send(int fd) { return 0; }
int null_parse(int fd) { return 0; }

ParseFunc default_func_parse = null_parse;

void set_defaultparse(ParseFunc defaultparse)
{
	default_func_parse = defaultparse;
}

/*======================================
 * CORE : Socket options
 *--------------------------------------*/
void set_nonblocking(int fd, unsigned long yes)
{
	// FIONBIO Use with a nonzero argp parameter to enable the nonblocking mode of socket s.
	// The argp parameter is zero if nonblocking is to be disabled.
	if( sIoctl(fd, FIONBIO, &yes) != 0 )
		ShowError("set_nonblocking: Failed to set socket #%d to non-blocking mode (%s) - Please report this!!!\n", fd, error_msg());
}

/**
 * Sets the options for a socket.
 *
 * @param fd  The socket descriptor
 * @param opt Optional, additional options to set (Can be NULL).
 */
void setsocketopts(int fd, struct hSockOpt *opt)
{
#if defined(WIN32)
	BOOL yes = TRUE;
#else // not WIN32
	int yes = 1;
#endif // WIN32
	struct linger lopt = { 0 };

	// Note: We cast the fourth argument to (char *) because, while in UNIX
	// it takes a const void *, in Windows it takes a const char *.
#if !defined(WIN32)
	// set SO_REAUSEADDR to true, unix only. on windows this option causes
	// the previous owner of the socket to give up, which is not desirable
	// in most cases, neither compatible with unix.
	if (sSetsockopt(fd, SOL_SOCKET, SO_REUSEADDR, (char *)&yes, sizeof(yes)))
		ShowWarning("setsocketopts: Unable to set SO_REUSEADDR mode for connection #%d!\n", fd);
#ifdef SO_REUSEPORT
	if (sSetsockopt(fd, SOL_SOCKET, SO_REUSEPORT, (char *)&yes, sizeof(yes)))
		ShowWarning("setsocketopts: Unable to set SO_REUSEPORT mode for connection #%d!\n", fd);
#endif // SO_REUSEPORT
#endif // WIN32

	// Set the socket into no-delay mode; otherwise packets get delayed for up to 200ms, likely creating server-side lag.
	// The RO protocol is mainly single-packet request/response, plus the FIFO model already does packet grouping anyway.
	if (sSetsockopt(fd, IPPROTO_TCP, TCP_NODELAY, (char *)&yes, sizeof(yes)))
		ShowWarning("setsocketopts: Unable to set TCP_NODELAY mode for connection #%d!\n", fd);

	if (opt && opt->setTimeo) {
#if defined(WIN32)
		DWORD timeout = 5000; // https://msdn.microsoft.com/en-us/library/windows/desktop/ms740476(v=vs.85).aspx
#else // not WIN32
		struct timeval timeout = { 0 };
		timeout.tv_sec = 5;
#endif // WIN32

		if (sSetsockopt(fd, SOL_SOCKET, SO_RCVTIMEO, (char *)&timeout, sizeof(timeout)))
			ShowWarning("setsocketopts: Unable to set SO_RCVTIMEO for connection #%d!\n", fd);
		if (sSetsockopt(fd, SOL_SOCKET, SO_SNDTIMEO, (char *)&timeout, sizeof(timeout)))
			ShowWarning("setsocketopts: Unable to set SO_SNDTIMEO for connection #%d!\n", fd);
	}

	// force the socket into no-wait, graceful-close mode (should be the default, but better make sure)
	//(http://msdn.microsoft.com/library/default.asp?url=/library/en-us/winsock/winsock/closesocket_2.asp)
	lopt.l_onoff = 0; // SO_DONTLINGER
	lopt.l_linger = 0; // Do not care
	if (sSetsockopt(fd, SOL_SOCKET, SO_LINGER, (char *)&lopt, sizeof(lopt)))
		ShowWarning("setsocketopts: Unable to set SO_LINGER mode for connection #%d!\n", fd);

#ifdef TCP_THIN_LINEAR_TIMEOUTS
	if (sSetsockopt(fd, IPPROTO_TCP, TCP_THIN_LINEAR_TIMEOUTS, (char *)&yes, sizeof(yes)))
		ShowWarning("setsocketopts: Unable to set TCP_THIN_LINEAR_TIMEOUTS mode for connection #%d!\n", fd);
#endif  // TCP_THIN_LINEAR_TIMEOUTS
#ifdef TCP_THIN_DUPACK
	if (sSetsockopt(fd, IPPROTO_TCP, TCP_THIN_DUPACK, (char *)&yes, sizeof(yes)))
		ShowWarning("setsocketopts: Unable to set TCP_THIN_DUPACK mode for connection #%d!\n", fd);
#endif  // TCP_THIN_DUPACK
}

/*======================================
 * CORE : Socket Sub Function
 *--------------------------------------*/
void set_eof(int fd)
{
	if (sockt->session_is_active(fd)) {
#ifdef SEND_SHORTLIST
		// Add this socket to the shortlist for eof handling.
		send_shortlist_add_fd(fd);
#endif  // SEND_SHORTLIST
		sockt->session[fd]->flag.eof = 1;
	}
}

int recv_to_fifo(int fd)
{
	ssize_t len;

	if (!sockt->session_is_active(fd))
		return -1;

	len = sRecv(fd, (char *) sockt->session[fd]->rdata + sockt->session[fd]->rdata_size, (int)RFIFOSPACE(fd), 0);

	if( len == SOCKET_ERROR )
	{//An exception has occurred
		if( sErrno != S_EWOULDBLOCK ) {
			//ShowDebug("recv_to_fifo: %s, closing connection #%d\n", error_msg(), fd);
			sockt->eof(fd);
		}
		return 0;
	}

	if( len == 0 )
	{//Normal connection end.
		sockt->eof(fd);
		return 0;
	}

	sockt->session[fd]->rdata_size += len;
	sockt->session[fd]->rdata_tick = sockt->last_tick;
#ifdef SHOW_SERVER_STATS
	socket_data_i += len;
	socket_data_qi += len;
	if (!sockt->session[fd]->flag.server)
	{
		socket_data_ci += len;
	}
#endif  // SHOW_SERVER_STATS
	return 0;
}

int send_from_fifo(int fd)
{
	ssize_t len;

	if (!sockt->session_is_valid(fd))
		return -1;

	if( sockt->session[fd]->wdata_size == 0 )
		return 0; // nothing to send

	len = sSend(fd, (const char *) sockt->session[fd]->wdata, (int)sockt->session[fd]->wdata_size, MSG_NOSIGNAL);

	if( len == SOCKET_ERROR )
	{ //An exception has occurred
		if( sErrno != S_EWOULDBLOCK ) {
			//ShowDebug("send_from_fifo: %s, ending connection #%d\n", error_msg(), fd);
#ifdef SHOW_SERVER_STATS
			socket_data_qo -= sockt->session[fd]->wdata_size;
#endif  // SHOW_SERVER_STATS
			sockt->session[fd]->wdata_size = 0; //Clear the send queue as we can't send anymore. [Skotlex]
			sockt->eof(fd);
		}
		return 0;
	}

	if( len > 0 )
	{
		// some data could not be transferred?
		// shift unsent data to the beginning of the queue
		if( (size_t)len < sockt->session[fd]->wdata_size )
			memmove(sockt->session[fd]->wdata, sockt->session[fd]->wdata + len, sockt->session[fd]->wdata_size - len);

		sockt->session[fd]->wdata_size -= len;
#ifdef SHOW_SERVER_STATS
		socket_data_o += len;
		socket_data_qo -= len;
		if (!sockt->session[fd]->flag.server)
		{
			socket_data_co += len;
		}
#endif  // SHOW_SERVER_STATS
	}

	return 0;
}

/// Best effort - there's no warranty that the data will be sent.
void flush_fifo(int fd)
{
	if(sockt->session[fd] != NULL)
		sockt->session[fd]->func_send(fd);
}

void flush_fifos(void)
{
	int i;
	for(i = 1; i < sockt->fd_max; i++)
		sockt->flush(i);
}

/*======================================
 * CORE : Connection functions
 *--------------------------------------*/
int connect_client(int listen_fd)
{
	int fd;
	struct sockaddr_in client_address;
	socklen_t len;

	len = sizeof(client_address);

	fd = sAccept(listen_fd, (struct sockaddr*)&client_address, &len);
	if ( fd == -1 ) {
		ShowError("connect_client: accept failed (%s)!\n", error_msg());
		return -1;
	}
	if( fd == 0 ) { // reserved
		ShowError("connect_client: Socket #0 is reserved - Please report this!!!\n");
		sClose(fd);
		return -1;
	}
	if( fd >= FD_SETSIZE ) { // socket number too big
		ShowError("connect_client: New socket #%d is greater than can we handle! Increase the value of FD_SETSIZE (currently %d) for your OS to fix this!\n", fd, FD_SETSIZE);
		sClose(fd);
		return -1;
	}

	setsocketopts(fd,NULL);
	sockt->set_nonblocking(fd, 1);

#ifndef MINICORE
	if( ip_rules && !connect_check(ntohl(client_address.sin_addr.s_addr)) ) {
		sockt->close(fd);
		return -1;
	}
#endif  // MINICORE

#ifndef SOCKET_EPOLL
	// Select Based Event Dispatcher
	sFD_SET(fd,&readfds);

#else  // SOCKET_EPOLL
	// Epoll based Event Dispatcher
	epevent.data.fd = fd;
	epevent.events = EPOLLIN;

	if(epoll_ctl(epfd, EPOLL_CTL_ADD, fd, &epevent) == SOCKET_ERROR){
		ShowError("connect_client: New Socket #%d failed to add to epoll event dispatcher: %s\n", fd, error_msg());
		sClose(fd);
		return -1;
	}

#endif  // SOCKET_EPOLL

	if( sockt->fd_max <= fd ) sockt->fd_max = fd + 1;

	create_session(fd, recv_to_fifo, send_from_fifo, default_func_parse);
	sockt->session[fd]->client_addr = ntohl(client_address.sin_addr.s_addr);

	return fd;
}

int make_listen_bind(uint32 ip, uint16 port)
{
	struct sockaddr_in server_address = { 0 };
	int fd;
	int result;

	fd = sSocket(AF_INET, SOCK_STREAM, 0);

	if( fd == -1 ) {
		ShowError("make_listen_bind: socket creation failed (%s)!\n", error_msg());
		exit(EXIT_FAILURE);
	}
	if( fd == 0 ) { // reserved
		ShowError("make_listen_bind: Socket #0 is reserved - Please report this!!!\n");
		sClose(fd);
		return -1;
	}
	if( fd >= FD_SETSIZE ) { // socket number too big
		ShowError("make_listen_bind: New socket #%d is greater than can we handle! Increase the value of FD_SETSIZE (currently %d) for your OS to fix this!\n", fd, FD_SETSIZE);
		sClose(fd);
		return -1;
	}

	setsocketopts(fd,NULL);
	sockt->set_nonblocking(fd, 1);

	server_address.sin_family      = AF_INET;
	server_address.sin_addr.s_addr = htonl(ip);
	server_address.sin_port        = htons(port);

	result = sBind(fd, (struct sockaddr*)&server_address, sizeof(server_address));
	if( result == SOCKET_ERROR ) {
		ShowError("make_listen_bind: bind failed (socket #%d, %s)!\n", fd, error_msg());
		exit(EXIT_FAILURE);
	}
	result = sListen(fd,5);
	if( result == SOCKET_ERROR ) {
		ShowError("make_listen_bind: listen failed (socket #%d, %s)!\n", fd, error_msg());
		exit(EXIT_FAILURE);
	}


#ifndef SOCKET_EPOLL
	// Select Based Event Dispatcher
	sFD_SET(fd,&readfds);

#else  // SOCKET_EPOLL
	// Epoll based Event Dispatcher
	epevent.data.fd = fd;
	epevent.events = EPOLLIN;

	if(epoll_ctl(epfd, EPOLL_CTL_ADD, fd, &epevent) == SOCKET_ERROR){
		ShowError("make_listen_bind: failed to add listener socket #%d to epoll event dispatcher: %s\n", fd, error_msg());
		sClose(fd);
		exit(EXIT_FAILURE);
	}

#endif  // SOCKET_EPOLL

	if(sockt->fd_max <= fd) sockt->fd_max = fd + 1;


	create_session(fd, connect_client, null_send, null_parse);
	sockt->session[fd]->client_addr = 0; // just listens
	sockt->session[fd]->rdata_tick = 0; // disable timeouts on this socket

	return fd;
}

int make_connection(uint32 ip, uint16 port, struct hSockOpt *opt)
{
	struct sockaddr_in remote_address = { 0 };
	int fd;
	int result;

	fd = sSocket(AF_INET, SOCK_STREAM, 0);

	if (fd == -1) {
		ShowError("make_connection: socket creation failed (%s)!\n", error_msg());
		return -1;
	}
	if( fd == 0 ) {// reserved
		ShowError("make_connection: Socket #0 is reserved - Please report this!!!\n");
		sClose(fd);
		return -1;
	}
	if( fd >= FD_SETSIZE ) {// socket number too big
		ShowError("make_connection: New socket #%d is greater than can we handle! Increase the value of FD_SETSIZE (currently %d) for your OS to fix this!\n", fd, FD_SETSIZE);
		sClose(fd);
		return -1;
	}

	setsocketopts(fd,opt);

	remote_address.sin_family      = AF_INET;
	remote_address.sin_addr.s_addr = htonl(ip);
	remote_address.sin_port        = htons(port);

	if( !( opt && opt->silent ) )
		ShowStatus("Connecting to %u.%u.%u.%u:%i\n", CONVIP(ip), port);

	result = sConnect(fd, (struct sockaddr *)(&remote_address), sizeof(struct sockaddr_in));
	if( result == SOCKET_ERROR ) {
		if( !( opt && opt->silent ) )
			ShowError("make_connection: connect failed (socket #%d, %s)!\n", fd, error_msg());
		sockt->close(fd);
		return -1;
	}
	//Now the socket can be made non-blocking. [Skotlex]
	sockt->set_nonblocking(fd, 1);


#ifndef SOCKET_EPOLL
	// Select Based Event Dispatcher
	sFD_SET(fd,&readfds);

#else  // SOCKET_EPOLL
	// Epoll based Event Dispatcher
	epevent.data.fd = fd;
	epevent.events = EPOLLIN;

	if(epoll_ctl(epfd, EPOLL_CTL_ADD, fd, &epevent) == SOCKET_ERROR){
		ShowError("make_connection: failed to add socket #%d to epoll event dispatcher: %s\n", fd, error_msg());
		sClose(fd);
		return -1;
	}

#endif  // SOCKET_EPOLL

	if(sockt->fd_max <= fd) sockt->fd_max = fd + 1;

	create_session(fd, recv_to_fifo, send_from_fifo, default_func_parse);
	sockt->session[fd]->client_addr = ntohl(remote_address.sin_addr.s_addr);

	return fd;
}

static int create_session(int fd, RecvFunc func_recv, SendFunc func_send, ParseFunc func_parse)
{
	CREATE(sockt->session[fd], struct socket_data, 1);
	CREATE(sockt->session[fd]->rdata, unsigned char, RFIFO_SIZE);
	CREATE(sockt->session[fd]->wdata, unsigned char, WFIFO_SIZE);
	sockt->session[fd]->max_rdata  = RFIFO_SIZE;
	sockt->session[fd]->max_wdata  = WFIFO_SIZE;
	sockt->session[fd]->func_recv  = func_recv;
	sockt->session[fd]->func_send  = func_send;
	sockt->session[fd]->func_parse = func_parse;
	sockt->session[fd]->rdata_tick = sockt->last_tick;
	sockt->session[fd]->session_data = NULL;
	sockt->session[fd]->hdata = NULL;
	return 0;
}

static void delete_session(int fd)
{
	if (sockt->session_is_valid(fd)) {
#ifdef SHOW_SERVER_STATS
		socket_data_qi -= sockt->session[fd]->rdata_size - sockt->session[fd]->rdata_pos;
		socket_data_qo -= sockt->session[fd]->wdata_size;
#endif  // SHOW_SERVER_STATS
		aFree(sockt->session[fd]->rdata);
		aFree(sockt->session[fd]->wdata);
		if( sockt->session[fd]->session_data )
			aFree(sockt->session[fd]->session_data);
		HPM->data_store_destroy(&sockt->session[fd]->hdata);
		aFree(sockt->session[fd]);
		sockt->session[fd] = NULL;
	}
}

int realloc_fifo(int fd, unsigned int rfifo_size, unsigned int wfifo_size)
{
	if (!sockt->session_is_valid(fd))
		return 0;

	if( sockt->session[fd]->max_rdata != rfifo_size && sockt->session[fd]->rdata_size < rfifo_size) {
		RECREATE(sockt->session[fd]->rdata, unsigned char, rfifo_size);
		sockt->session[fd]->max_rdata  = rfifo_size;
	}

	if( sockt->session[fd]->max_wdata != wfifo_size && sockt->session[fd]->wdata_size < wfifo_size) {
		RECREATE(sockt->session[fd]->wdata, unsigned char, wfifo_size);
		sockt->session[fd]->max_wdata  = wfifo_size;
	}
	return 0;
}

int realloc_writefifo(int fd, size_t addition)
{
	size_t newsize;

	if (!sockt->session_is_valid(fd)) // might not happen
		return 0;

	if (sockt->session[fd]->wdata_size + addition  > sockt->session[fd]->max_wdata) {
		// grow rule; grow in multiples of WFIFO_SIZE
		newsize = WFIFO_SIZE;
		while( sockt->session[fd]->wdata_size + addition > newsize ) newsize += WFIFO_SIZE;
	} else if (sockt->session[fd]->max_wdata >= (size_t)2*(sockt->session[fd]->flag.server?FIFOSIZE_SERVERLINK:WFIFO_SIZE)
	       && (sockt->session[fd]->wdata_size+addition)*4 < sockt->session[fd]->max_wdata
	) {
		// shrink rule, shrink by 2 when only a quarter of the fifo is used, don't shrink below nominal size.
		newsize = sockt->session[fd]->max_wdata / 2;
	} else {
		// no change
		return 0;
	}

	RECREATE(sockt->session[fd]->wdata, unsigned char, newsize);
	sockt->session[fd]->max_wdata  = newsize;

	return 0;
}

/// advance the RFIFO cursor (marking 'len' bytes as processed)
int rfifoskip(int fd, size_t len)
{
	struct socket_data *s;

	if (!sockt->session_is_active(fd))
		return 0;

	s = sockt->session[fd];

	if (s->rdata_size < s->rdata_pos + len) {
		ShowError("RFIFOSKIP: skipped past end of read buffer! Adjusting from %"PRIuS" to %"PRIuS" (session #%d)\n", len, RFIFOREST(fd), fd);
		len = RFIFOREST(fd);
	}

	s->rdata_pos = s->rdata_pos + len;
#ifdef SHOW_SERVER_STATS
	socket_data_qi -= len;
#endif  // SHOW_SERVER_STATS
	return 0;
}

/// advance the WFIFO cursor (marking 'len' bytes for sending)
int wfifoset(int fd, size_t len)
{
	size_t newreserve;
	struct socket_data* s;

	if (!sockt->session_is_valid(fd))
		return 0;
	s = sockt->session[fd];
	if (s == NULL || s->wdata == NULL)
		return 0;

	// we have written len bytes to the buffer already before calling WFIFOSET
	if (s->wdata_size+len > s->max_wdata) {
		// actually there was a buffer overflow already
		uint32 ip = s->client_addr;
		ShowFatalError("WFIFOSET: Write Buffer Overflow. Connection %d (%u.%u.%u.%u) has written %u bytes on a %u/%u bytes buffer.\n", fd, CONVIP(ip), (unsigned int)len, (unsigned int)s->wdata_size, (unsigned int)s->max_wdata);
		ShowDebug("Likely command that caused it: 0x%x\n", (*(uint16*)(s->wdata + s->wdata_size)));
		// no other chance, make a better fifo model
		exit(EXIT_FAILURE);
	}

	if( len > 0xFFFF )
	{
		// dynamic packets allow up to UINT16_MAX bytes (<packet_id>.W <packet_len>.W ...)
		// all known fixed-size packets are within this limit, so use the same limit
		ShowFatalError("WFIFOSET: Packet 0x%x is too big. (len=%u, max=%u)\n", (*(uint16*)(s->wdata + s->wdata_size)), (unsigned int)len, 0xFFFFU);
		exit(EXIT_FAILURE);
	}
	else if( len == 0 )
	{
		// abuses the fact, that the code that did WFIFOHEAD(fd,0), already wrote
		// the packet type into memory, even if it could have overwritten vital data
		// this can happen when a new packet was added on map-server, but packet len table was not updated
		ShowWarning("WFIFOSET: Attempted to send zero-length packet, most likely 0x%04x (please report this).\n", WFIFOW(fd,0));
		return 0;
	}

	if( !s->flag.server ) {

		if (len > socket_max_client_packet) { // see declaration of socket_max_client_packet for details
			ShowError("WFIFOSET: Dropped too large client packet 0x%04x (length=%"PRIuS", max=%"PRIuS").\n",
			          WFIFOW(fd,0), len, socket_max_client_packet);
			return 0;
		}

	}
	s->wdata_size += len;
#ifdef SHOW_SERVER_STATS
	socket_data_qo += len;
#endif  // SHOW_SERVER_STATS
	//If the interserver has 200% of its normal size full, flush the data.
	if( s->flag.server && s->wdata_size >= 2*FIFOSIZE_SERVERLINK )
		sockt->flush(fd);

	// always keep a WFIFO_SIZE reserve in the buffer
	// For inter-server connections, let the reserve be 1/4th of the link size.
	newreserve = s->flag.server ? FIFOSIZE_SERVERLINK / 4 : WFIFO_SIZE;

	// readjust the buffer to include the chosen reserve
	sockt->realloc_writefifo(fd, newreserve);

#ifdef SEND_SHORTLIST
	send_shortlist_add_fd(fd);
#endif  // SEND_SHORTLIST

	return 0;
}

int do_sockets(int next)
{
#ifndef SOCKET_EPOLL
	fd_set rfd;
	struct timeval timeout;
#endif  // SOCKET_EPOLL
	int ret,i;

	// PRESEND Timers are executed before do_sendrecv and can send packets and/or set sessions to eof.
	// Send remaining data and process client-side disconnects here.
#ifdef SEND_SHORTLIST
	send_shortlist_do_sends();
#else  // SEND_SHORTLIST
	for (i = 1; i < sockt->fd_max; i++) {
		if (sockt->session[i] == NULL)
			continue;

		if (sockt->session[i]->wdata_size > 0)
			sockt->session[i]->func_send(i);
	}
#endif  // SEND_SHORTLIST

#ifndef SOCKET_EPOLL
	// Select based Event Dispatcher:

	// can timeout until the next tick
	timeout.tv_sec  = next/1000;
	timeout.tv_usec = next%1000*1000;

	memcpy(&rfd, &readfds, sizeof(rfd));
	ret = sSelect(sockt->fd_max, &rfd, NULL, NULL, &timeout);

	if( ret == SOCKET_ERROR )
	{
		if( sErrno != S_EINTR )
		{
			ShowFatalError("do_sockets: select() failed, %s!\n", error_msg());
			exit(EXIT_FAILURE);
		}
		return 0; // interrupted by a signal, just loop and try again
	}
#else  // SOCKET_EPOLL
	// Epoll based Event Dispatcher

	ret = epoll_wait(epfd, epevents, epoll_maxevents, next);
	if(ret == SOCKET_ERROR)
	{
		if( sErrno != S_EINTR )
		{
			ShowFatalError("do_sockets: epoll_wait() failed, %s!\n", error_msg());
			exit(EXIT_FAILURE);
		}
		return 0; // interrupted by a signal, just loop and try again
	}
#endif  // SOCKET_EPOLL

	sockt->last_tick = time(NULL);

#if defined(WIN32)
	// on windows, enumerating all members of the fd_set is way faster if we access the internals
	for( i = 0; i < (int)rfd.fd_count; ++i )
	{
		int fd = sock2fd(rfd.fd_array[i]);
		if( sockt->session[fd] )
			sockt->session[fd]->func_recv(fd);
	}
#elif defined(SOCKET_EPOLL)
	// epoll based selection

	for( i = 0; i < ret; i++ )
	{
		struct epoll_event *it = &epevents[i];
		struct socket_data *sock = sockt->session[ it->data.fd ];

		if(!sock)
			continue;

		if ((it->events & EPOLLERR) ||
			(it->events & EPOLLHUP) ||
			(!(it->events & EPOLLIN)))
		{
			// Got Error on this connection
			sockt->eof( it->data.fd );

		} else if (it->events & EPOLLIN) {
			// data wainting
			sock->func_recv( it->data.fd );

		}

	}

#else  // defined(SOCKET_EPOLL)
	// otherwise assume that the fd_set is a bit-array and enumerate it in a standard way
	for( i = 1; ret && i < sockt->fd_max; ++i )
	{
		if(sFD_ISSET(i,&rfd) && sockt->session[i])
		{
			sockt->session[i]->func_recv(i);
			--ret;
		}
	}
#endif  // defined(SOCKET_EPOLL)

	// POSTSEND Send remaining data and handle eof sessions.
#ifdef SEND_SHORTLIST
	send_shortlist_do_sends();
#else  // SEND_SHORTLIST
	for (i = 1; i < sockt->fd_max; i++)
	{
		if(!sockt->session[i])
			continue;

		if(sockt->session[i]->wdata_size)
			sockt->session[i]->func_send(i);

		if (sockt->session[i]->flag.eof) { //func_send can't free a session, this is safe.
			//Finally, even if there is no data to parse, connections signaled eof should be closed, so we call parse_func [Skotlex]
			sockt->session[i]->func_parse(i); //This should close the session immediately.
		}
	}
#endif  // SEND_SHORTLIST

	// parse input data on each socket
	for(i = 1; i < sockt->fd_max; i++)
	{
		if(!sockt->session[i])
			continue;

		if (sockt->session[i]->rdata_tick && DIFF_TICK(sockt->last_tick, sockt->session[i]->rdata_tick) > sockt->stall_time) {
			if( sockt->session[i]->flag.server ) {/* server is special */
				if( sockt->session[i]->flag.ping != 2 )/* only update if necessary otherwise it'd resend the ping unnecessarily */
					sockt->session[i]->flag.ping = 1;
			} else {
				ShowInfo("Session #%d timed out\n", i);
				sockt->eof(i);
			}
		}

		sockt->session[i]->func_parse(i);

		if(!sockt->session[i])
			continue;

		RFIFOFLUSH(i);
		// after parse, check client's RFIFO size to know if there is an invalid packet (too big and not parsed)
		if (sockt->session[i]->rdata_size == sockt->session[i]->max_rdata) {
			sockt->eof(i);
			continue;
		}
	}

#ifdef SHOW_SERVER_STATS
	if (sockt->last_tick != socket_data_last_tick)
	{
		char buf[1024];

		sprintf(buf, "In: %.03f kB/s (%.03f kB/s, Q: %.03f kB) | Out: %.03f kB/s (%.03f kB/s, Q: %.03f kB) | RAM: %.03f MB", socket_data_i/1024., socket_data_ci/1024., socket_data_qi/1024., socket_data_o/1024., socket_data_co/1024., socket_data_qo/1024., iMalloc->usage()/1024.);
#ifdef _WIN32
		SetConsoleTitle(buf);
#else  // _WIN32
		ShowMessage("\033[s\033[1;1H\033[2K%s\033[u", buf);
#endif  // _WIN32
		socket_data_last_tick = sockt->last_tick;
		socket_data_i = socket_data_ci = 0;
		socket_data_o = socket_data_co = 0;
	}
#endif  // SHOW_SERVER_STATS

	return 0;
}

//////////////////////////////
#ifndef MINICORE
//////////////////////////////
// IP rules and DDoS protection

struct connect_history {
	uint32 ip;
	int64 tick;
	int count;
	unsigned ddos : 1;
};

struct access_control {
	uint32 ip;
	uint32 mask;
};

VECTOR_STRUCT_DECL(access_control_list, struct access_control);

enum aco {
	ACO_DENY_ALLOW,
	ACO_ALLOW_DENY,
	ACO_MUTUAL_FAILURE
};

static struct access_control_list access_allow;
static struct access_control_list access_deny;
static int access_order    = ACO_DENY_ALLOW;
static int access_debug    = 0;
static int ddos_count      = 10;
static int ddos_interval   = 3*1000;
static int ddos_autoreset  = 10*60*1000;
struct DBMap *connect_history = NULL;

static int connect_check_(uint32 ip);

/// Verifies if the IP can connect. (with debug info)
/// @see connect_check_()
static int connect_check(uint32 ip)
{
	int result = connect_check_(ip);
	if( access_debug ) {
		ShowInfo("connect_check: Connection from %u.%u.%u.%u %s\n", CONVIP(ip),result ? "allowed." : "denied!");
	}
	return result;
}

/// Verifies if the IP can connect.
///  0      : Connection Rejected
///  1 or 2 : Connection Accepted
static int connect_check_(uint32 ip)
{
	struct connect_history *hist = NULL;
	int i;
	int is_allowip = 0;
	int is_denyip = 0;
	int connect_ok = 0;

	// Search the allow list
	for (i = 0; i < VECTOR_LENGTH(access_allow); ++i) {
		struct access_control *entry = &VECTOR_INDEX(access_allow, i);
		if (SUBNET_MATCH(ip, entry->ip, entry->mask)) {
			if (access_debug) {
				ShowInfo("connect_check: Found match from allow list:%u.%u.%u.%u IP:%u.%u.%u.%u Mask:%u.%u.%u.%u\n",
					CONVIP(ip),
					CONVIP(entry->ip),
					CONVIP(entry->mask));
			}
			is_allowip = 1;
			break;
		}
	}
	// Search the deny list
	for (i = 0; i < VECTOR_LENGTH(access_deny); ++i) {
		struct access_control *entry = &VECTOR_INDEX(access_deny, i);
		if (SUBNET_MATCH(ip, entry->ip, entry->mask)) {
			if (access_debug) {
				ShowInfo("connect_check: Found match from deny list:%u.%u.%u.%u IP:%u.%u.%u.%u Mask:%u.%u.%u.%u\n",
					CONVIP(ip),
					CONVIP(entry->ip),
					CONVIP(entry->mask));
			}
			is_denyip = 1;
			break;
		}
	}
	// Decide connection status
	//  0 : Reject
	//  1 : Accept
	//  2 : Unconditional Accept (accepts even if flagged as DDoS)
	switch(access_order) {
		case ACO_DENY_ALLOW:
		default:
			if( is_denyip )
				connect_ok = 0; // Reject
			else if( is_allowip )
				connect_ok = 2; // Unconditional Accept
			else
				connect_ok = 1; // Accept
			break;
		case ACO_ALLOW_DENY:
			if( is_allowip )
				connect_ok = 2; // Unconditional Accept
			else if( is_denyip )
				connect_ok = 0; // Reject
			else
				connect_ok = 1; // Accept
			break;
		case ACO_MUTUAL_FAILURE:
			if( is_allowip && !is_denyip )
				connect_ok = 2; // Unconditional Accept
			else
				connect_ok = 0; // Reject
			break;
	}

	// Inspect connection history
	if( ( hist = uidb_get(connect_history, ip)) ) { //IP found
		if( hist->ddos ) {// flagged as DDoS
			return (connect_ok == 2 ? 1 : 0);
		} else if( DIFF_TICK(timer->gettick(),hist->tick) < ddos_interval ) {// connection within ddos_interval
				hist->tick = timer->gettick();
				if( ++hist->count >= ddos_count ) {// DDoS attack detected
					hist->ddos = 1;
					ShowWarning("connect_check: DDoS Attack detected from %u.%u.%u.%u!\n", CONVIP(ip));
					return (connect_ok == 2 ? 1 : 0);
				}
				return connect_ok;
		} else {// not within ddos_interval, clear data
			hist->tick  = timer->gettick();
			hist->count = 0;
			return connect_ok;
		}
	}
	// IP not found, add to history
	CREATE(hist, struct connect_history, 1);
	hist->ip   = ip;
	hist->tick = timer->gettick();
	uidb_put(connect_history, ip, hist);
	return connect_ok;
}

/// Timer function.
/// Deletes old connection history records.
static int connect_check_clear(int tid, int64 tick, int id, intptr_t data)
{
	int clear = 0;
	int list  = 0;
	struct connect_history *hist = NULL;
	struct DBIterator *iter;

	if( !db_size(connect_history) )
		return 0;

	iter = db_iterator(connect_history);

	for( hist = dbi_first(iter); dbi_exists(iter); hist = dbi_next(iter) ){
		if( (!hist->ddos && DIFF_TICK(tick,hist->tick) > ddos_interval*3) ||
			(hist->ddos && DIFF_TICK(tick,hist->tick) > ddos_autoreset) )
			{// Remove connection history
				uidb_remove(connect_history, hist->ip);
				clear++;
			}
		list++;
	}
	dbi_destroy(iter);

	if( access_debug ){
		ShowInfo("connect_check_clear: Cleared %d of %d from IP list.\n", clear, list);
	}

	return list;
}

/// Parses the ip address and mask and puts it into acc.
/// Returns 1 is successful, 0 otherwise.
int access_ipmask(const char *str, struct access_control *acc)
{
	uint32 ip;
	uint32 mask;

	nullpo_ret(str);
	nullpo_ret(acc);

	if( strcmp(str,"all") == 0 ) {
		ip   = 0;
		mask = 0;
	} else {
		unsigned int a[4];
		unsigned int m[4];
		int n;
		if( ((n=sscanf(str,"%u.%u.%u.%u/%u.%u.%u.%u",a,a+1,a+2,a+3,m,m+1,m+2,m+3)) != 8 && // not an ip + standard mask
				(n=sscanf(str,"%u.%u.%u.%u/%u",a,a+1,a+2,a+3,m)) != 5 && // not an ip + bit mask
				(n=sscanf(str,"%u.%u.%u.%u",a,a+1,a+2,a+3)) != 4 ) || // not an ip
				a[0] > 255 || a[1] > 255 || a[2] > 255 || a[3] > 255 || // invalid ip
				(n == 8 && (m[0] > 255 || m[1] > 255 || m[2] > 255 || m[3] > 255)) || // invalid standard mask
				(n == 5 && m[0] > 32) ){ // invalid bit mask
			return 0;
		}
		ip = MAKEIP(a[0],a[1],a[2],a[3]);
		if( n == 8 )
		{// standard mask
			mask = MAKEIP(m[0],m[1],m[2],m[3]);
		} else if( n == 5 )
		{// bit mask
			mask = 0;
			while( m[0] ){
				mask = (mask >> 1) | 0x80000000;
				--m[0];
			}
		} else
		{// just this ip
			mask = 0xFFFFFFFF;
		}
	}
	if( access_debug ){
		ShowInfo("access_ipmask: Loaded IP:%u.%u.%u.%u mask:%u.%u.%u.%u\n", CONVIP(ip), CONVIP(mask));
	}
	acc->ip   = ip;
	acc->mask = mask;
	return 1;
}

/**
 * Adds an entry to the access list.
 *
 * @param setting     The setting to read from.
 * @param list_name   The list name (used in error messages).
 * @param access_list The access list to edit.
 *
 * @retval false in case of failure
 */
bool access_list_add(struct config_setting_t *setting, const char *list_name, struct access_control_list *access_list)
{
	const char *temp = NULL;
	int i, setting_length;

	nullpo_retr(false, setting);
	nullpo_retr(false, list_name);
	nullpo_retr(false, access_list);

	if ((setting_length = libconfig->setting_length(setting)) <= 0)
		return false;

	VECTOR_ENSURE(*access_list, setting_length, 1);
	for (i = 0; i < setting_length; i++) {
		struct access_control acc;
		if ((temp = libconfig->setting_get_string_elem(setting, i)) == NULL) {
			continue;
		}

		if (!access_ipmask(temp, &acc)) {
			ShowError("access_list_add: Invalid ip or ip range %s '%d'!\n", list_name, i);
			continue;
		}
		VECTOR_PUSH(*access_list, acc);
	}

	return true;
}

//////////////////////////////
#endif  // MINICORE
//////////////////////////////

/**
 * Reads 'socket_configuration/ip_rules' and initializes required variables.
 *
 * @param filename Path to configuration file (used in error and warning messages).
 * @param config   The current config being parsed.
 * @param imported Whether the current config is imported from another file.
 *
 * @retval false in case of error.
 */
bool socket_config_read_iprules(const char *filename, struct config_t *config, bool imported)
{
#ifndef MINICORE
	struct config_setting_t *setting = NULL;
	const char *temp = NULL;

	nullpo_retr(false, filename);
	nullpo_retr(false, config);

	if ((setting = libconfig->lookup(config, "socket_configuration/ip_rules")) == NULL) {
		if (imported)
			return true;
		ShowError("socket_config_read: socket_configuration/ip_rules was not found in %s!\n", filename);
		return false;
	}
	libconfig->setting_lookup_bool(setting, "enable", &ip_rules);

	if (!ip_rules)
		return true;

	if (libconfig->setting_lookup_string(setting, "order", &temp) == CONFIG_TRUE) {
		if (strcmpi(temp, "deny,allow" ) == 0) {
			access_order = ACO_DENY_ALLOW;
		} else if (strcmpi(temp, "allow, deny") == 0) {
			access_order = ACO_ALLOW_DENY;
		} else if (strcmpi(temp, "mutual-failure") == 0) {
			access_order = ACO_MUTUAL_FAILURE;
		} else {
			ShowWarning("socket_config_read: invalid value '%s' for socket_configuration/ip_rules/order.\n", temp);
		}
	}

	if ((setting = libconfig->lookup(config, "socket_configuration/ip_rules/allow_list")) == NULL) {
		if (!imported)
			ShowError("socket_config_read: socket_configuration/ip_rules/allow_list was not found in %s!\n", filename);
	} else {
		access_list_add(setting, "allow_list", &access_allow);
	}

	if ((setting = libconfig->lookup(config, "socket_configuration/ip_rules/deny_list")) == NULL) {
		if (!imported)
			ShowError("socket_config_read: socket_configuration/ip_rules/deny_list was not found in %s!\n", filename);
	} else {
		access_list_add(setting, "deny_list", &access_deny);
	}
#endif // ! MINICORE

	return true;
}

/**
 * Reads 'socket_configuration/ddos' and initializes required variables.
 *
 * @param filename Path to configuration file (used in error and warning messages).
 * @param config   The current config being parsed.
 * @param imported Whether the current config is imported from another file.
 *
 * @retval false in case of error.
 */
bool socket_config_read_ddos(const char *filename, struct config_t *config, bool imported)
{
#ifndef MINICORE
	struct config_setting_t *setting = NULL;

	nullpo_retr(false, filename);
	nullpo_retr(false, config);

	if ((setting = libconfig->lookup(config, "socket_configuration/ddos")) == NULL) {
		if (imported)
			return true;
		ShowError("socket_config_read: socket_configuration/ddos was not found in %s!\n", filename);
		return false;
	}

	libconfig->setting_lookup_int(setting, "interval", &ddos_interval);
	libconfig->setting_lookup_int(setting, "count", &ddos_count);
	libconfig->setting_lookup_int(setting, "autoreset", &ddos_autoreset);

#endif // ! MINICORE
	return true;
}

/**
 * Reads 'socket_configuration' and initializes required variables.
 *
 * @param filename Path to configuration file.
 * @param imported Whether the current config is imported from another file.
 *
 * @retval false in case of error.
 */
bool socket_config_read(const char *filename, bool imported)
{
	struct config_t config;
	struct config_setting_t *setting = NULL;
	const char *import;
	int i32 = 0;
	bool retval = true;

	nullpo_retr(false, filename);

	if (!libconfig->load_file(&config, filename))
		return false;

	if ((setting = libconfig->lookup(&config, "socket_configuration")) == NULL) {
		libconfig->destroy(&config);
		if (imported)
			return true;
		ShowError("socket_config_read: socket_configuration was not found in %s!\n", filename);
		return false;
	}

	if (libconfig->setting_lookup_int(setting, "stall_time", &i32) == CONFIG_TRUE) {
		if (i32 < 3)
			i32 = 3; /* a minimum is required in order to refrain from killing itself */
		sockt->stall_time = i32;
	}

#ifdef SOCKET_EPOLL
	if (libconfig->setting_lookup_int(setting, "epoll_maxevents", &i32) == CONFIG_TRUE) {
		if (i32 < 16)
			i32 = 16; // minimum that seems to be useful
		epoll_maxevents = i32;
	}
#endif  // SOCKET_EPOLL

#ifndef MINICORE
	{
		uint32 ui32 = 0;
		libconfig->setting_lookup_bool(setting, "debug", &access_debug);
		if (libconfig->setting_lookup_uint32(setting, "socket_max_client_packet", &ui32) == CONFIG_TRUE) {
			socket_max_client_packet = ui32;
		}
	}

	if (!socket_config_read_iprules(filename, &config, imported))
		retval = false;
	if (!socket_config_read_ddos(filename, &config, imported))
		retval = false;
#endif // MINICORE

	// import should overwrite any previous configuration, so it should be called last
	if (libconfig->lookup_string(&config, "import", &import) == CONFIG_TRUE) {
		if (strcmp(import, filename) == 0 || strcmp(import, SOCKET_CONF_FILENAME) == 0) {
			ShowWarning("socket_config_read: Loop detected! Skipping 'import'...\n");
		} else {
			if (!socket_config_read(import, true))
				retval = false;
		}
	}

	libconfig->destroy(&config);
	return retval;
}

void socket_final(void)
{
	int i;
#ifndef MINICORE
	if( connect_history )
		db_destroy(connect_history);
	VECTOR_CLEAR(access_allow);
	VECTOR_CLEAR(access_deny);
#endif  // MINICORE

	for( i = 1; i < sockt->fd_max; i++ )
		if(sockt->session[i])
			sockt->close(i);

	// sockt->session[0]
	aFree(sockt->session[0]->rdata);
	aFree(sockt->session[0]->wdata);
	aFree(sockt->session[0]);

	aFree(sockt->session);

	VECTOR_CLEAR(sockt->lan_subnets);
	VECTOR_CLEAR(sockt->allowed_ips);
	VECTOR_CLEAR(sockt->trusted_ips);

#ifdef SOCKET_EPOLL
	if(epfd != SOCKET_ERROR){
		close(epfd);
		epfd = SOCKET_ERROR;
	}
	if(epevents != NULL){
		aFree(epevents);
		epevents = NULL;
	}
#endif  // SOCKET_EPOLL

}

/// Closes a socket.
void socket_close(int fd)
{
	if( fd <= 0 ||fd >= FD_SETSIZE )
		return;// invalid

	sockt->flush(fd); // Try to send what's left (although it might not succeed since it's a nonblocking socket)

#ifndef SOCKET_EPOLL
	// Select based Event Dispatcher
	sFD_CLR(fd, &readfds);// this needs to be done before closing the socket
#else  // SOCKET_EPOLL
	// Epoll based Event Dispatcher
	epevent.data.fd = fd;
	epevent.events = EPOLLIN;
	epoll_ctl(epfd, EPOLL_CTL_DEL, fd, &epevent); // removing the socket from epoll when it's being closed is not required but recommended
#endif  // SOCKET_EPOLL

	sShutdown(fd, SHUT_RDWR); // Disallow further reads/writes
	sClose(fd); // We don't really care if these closing functions return an error, we are just shutting down and not reusing this socket.
	if (sockt->session[fd]) delete_session(fd);
}

/// Retrieve local ips in host byte order.
/// Uses loopback is no address is found.
int socket_getips(uint32* ips, int max)
{
	int num = 0;

	if( ips == NULL || max <= 0 )
		return 0;

#ifdef WIN32
	{
		char fullhost[255];

		// XXX This should look up the local IP addresses in the registry
		// instead of calling gethostbyname. However, the way IP addresses
		// are stored in the registry is annoyingly complex, so I'll leave
		// this as T.B.D. [Meruru]
		if (gethostname(fullhost, sizeof(fullhost)) == SOCKET_ERROR) {
			ShowError("socket_getips: No hostname defined!\n");
			return 0;
		} else {
			u_long** a;
			struct hostent *hent =gethostbyname(fullhost);
			if( hent == NULL ){
				ShowError("socket_getips: Cannot resolve our own hostname to an IP address\n");
				return 0;
			}
			a = (u_long**)hent->h_addr_list;
			for (; num < max && a[num] != NULL; ++num)
				ips[num] = (uint32)ntohl(*a[num]);
		}
	}
#else // not WIN32
	{
		int fd;
		char buf[2*16*sizeof(struct ifreq)];
		struct ifconf ic;
		u_long ad;

		fd = sSocket(AF_INET, SOCK_STREAM, 0);
		if (fd == -1) {
			ShowError("socket_getips: Unable to create a socket!\n");
			return 0;
		}

		memset(buf, 0x00, sizeof(buf));

		// The ioctl call will fail with Invalid Argument if there are more
		// interfaces than will fit in the buffer
		ic.ifc_len = sizeof(buf);
		ic.ifc_buf = buf;
		if (sIoctl(fd, SIOCGIFCONF, &ic) == -1) {
			ShowError("socket_getips: SIOCGIFCONF failed!\n");
			sClose(fd);
			return 0;
		} else {
			int pos;
			for (pos = 0; pos < ic.ifc_len && num < max; ) {
				struct ifreq *ir = (struct ifreq*)(buf+pos);
				struct sockaddr_in *a = (struct sockaddr_in*) &(ir->ifr_addr);
				if (a->sin_family == AF_INET) {
					ad = ntohl(a->sin_addr.s_addr);
					if (ad != INADDR_LOOPBACK && ad != INADDR_ANY)
						ips[num++] = (uint32)ad;
				}
	#if (defined(BSD) && BSD >= 199103) || defined(_AIX) || defined(__APPLE__)
				pos += ir->ifr_addr.sa_len + sizeof(ir->ifr_name);
	#else// not AIX or APPLE
				pos += sizeof(struct ifreq);
	#endif//not AIX or APPLE
			}
		}
		sClose(fd);
	}
#endif // not W32

	// Use loopback if no ips are found
	if( num == 0 )
		ips[num++] = (uint32)INADDR_LOOPBACK;

	return num;
}

void socket_init(void)
{
	uint64 rlim_cur = FD_SETSIZE;

#ifdef WIN32
	{// Start up windows networking
		WSADATA wsaData;
		WORD wVersionRequested = MAKEWORD(2, 0);
		if( WSAStartup(wVersionRequested, &wsaData) != 0 )
		{
			ShowError("socket_init: WinSock not available!\n");
			return;
		}
		if( LOBYTE(wsaData.wVersion) != 2 || HIBYTE(wsaData.wVersion) != 0 )
		{
			ShowError("socket_init: WinSock version mismatch (2.0 or compatible required)!\n");
			return;
		}
	}
#elif defined(HAVE_SETRLIMIT) && !defined(CYGWIN)
	// NOTE: getrlimit and setrlimit have bogus behavior in cygwin.
	//       "Number of fds is virtually unlimited in cygwin" (sys/param.h)
	{// set socket limit to FD_SETSIZE
		struct rlimit rlp;
		if( 0 == getrlimit(RLIMIT_NOFILE, &rlp) )
		{
			rlp.rlim_cur = FD_SETSIZE;
			if( 0 != setrlimit(RLIMIT_NOFILE, &rlp) )
			{// failed, try setting the maximum too (permission to change system limits is required)
				rlp.rlim_max = FD_SETSIZE;
				if( 0 != setrlimit(RLIMIT_NOFILE, &rlp) )
				{// failed
					const char *errmsg = error_msg();
					int rlim_ori;
					// set to maximum allowed
					getrlimit(RLIMIT_NOFILE, &rlp);
					rlim_ori = (int)rlp.rlim_cur;
					rlp.rlim_cur = rlp.rlim_max;
					setrlimit(RLIMIT_NOFILE, &rlp);
					// report limit
					getrlimit(RLIMIT_NOFILE, &rlp);
					rlim_cur = rlp.rlim_cur;
					ShowWarning("socket_init: failed to set socket limit to %d, setting to maximum allowed (original limit=%d, current limit=%d, maximum allowed=%d, %s).\n", FD_SETSIZE, rlim_ori, (int)rlp.rlim_cur, (int)rlp.rlim_max, errmsg);
				}
			}
		}
	}
#endif  // defined(HAVE_SETRLIMIT) && !defined(CYGWIN)

#ifndef MINICORE
	VECTOR_INIT(access_allow);
	VECTOR_INIT(access_deny);
#endif // ! MINICORE

	// Get initial local ips
	sockt->naddr_ = sockt->getips(sockt->addr_,16);

	socket_config_read(SOCKET_CONF_FILENAME, false);

#ifndef SOCKET_EPOLL
	// Select based Event Dispatcher:
	sFD_ZERO(&readfds);
	ShowInfo("Server uses '" CL_WHITE "select" CL_RESET "' as event dispatcher\n");

#else  // SOCKET_EPOLL
	// Epoll based Event Dispatcher:
	epfd = epoll_create(FD_SETSIZE); // 2.6.8 or newer ignores the expected socket amount argument
	if(epfd == SOCKET_ERROR){
		ShowError("Failed to Create Epoll Event Dispatcher: %s\n", error_msg());
		exit(EXIT_FAILURE);
	}

	memset(&epevent, 0x00, sizeof(struct epoll_event));
	epevents = aCalloc(epoll_maxevents, sizeof(struct epoll_event));

	ShowInfo("Server uses '" CL_WHITE "epoll" CL_RESET "' with up to " CL_WHITE "%d" CL_RESET " events per cycle as event dispatcher\n", epoll_maxevents);

#endif  // SOCKET_EPOLL

#if defined(SEND_SHORTLIST)
	memset(send_shortlist_set, 0, sizeof(send_shortlist_set));
#endif  // defined(SEND_SHORTLIST)

	CREATE(sockt->session, struct socket_data *, FD_SETSIZE);

	// initialize last send-receive tick
	sockt->last_tick = time(NULL);

	// sockt->session[0] is now currently used for disconnected sessions of the map server, and as such,
	// should hold enough buffer (it is a vacuum so to speak) as it is never flushed. [Skotlex]
	create_session(0, null_recv, null_send, null_parse);

#ifndef MINICORE
	// Delete old connection history every 5 minutes
	connect_history = uidb_alloc(DB_OPT_RELEASE_DATA);
	timer->add_func_list(connect_check_clear, "connect_check_clear");
	timer->add_interval(timer->gettick()+1000, connect_check_clear, 0, 0, 5*60*1000);
#endif  // MINICORE

	ShowInfo("Server supports up to '"CL_WHITE"%"PRIu64""CL_RESET"' concurrent connections.\n", rlim_cur);
}

bool session_is_valid(int fd)
{
	return ( fd > 0 && fd < FD_SETSIZE && sockt->session[fd] != NULL );
}

bool session_is_active(int fd)
{
	return ( sockt->session_is_valid(fd) && !sockt->session[fd]->flag.eof );
}

// Resolves hostname into a numeric ip.
uint32 host2ip(const char *hostname)
{
	struct hostent* h;
	nullpo_ret(hostname);
	h = gethostbyname(hostname);
	return (h != NULL) ? ntohl(*(uint32*)h->h_addr) : 0;
}

/**
 * Converts a numeric ip into a dot-formatted string.
 *
 * @param ip     Numeric IP to convert.
 * @param ip_str Output buffer, optional (if provided, must have size greater or equal to 16).
 *
 * @return A pointer to the output string.
 */
const char *ip2str(uint32 ip, char *ip_str)
{
	struct in_addr addr;
	addr.s_addr = htonl(ip);
	return (ip_str == NULL) ? inet_ntoa(addr) : strncpy(ip_str, inet_ntoa(addr), 16);
}

// Converts a dot-formatted ip string into a numeric ip.
uint32 str2ip(const char* ip_str)
{
	return ntohl(inet_addr(ip_str));
}

// Reorders bytes from network to little endian (Windows).
// Necessary for sending port numbers to the RO client until Gravity notices that they forgot ntohs() calls.
uint16 ntows(uint16 netshort)
{
	return ((netshort & 0xFF) << 8) | ((netshort & 0xFF00) >> 8);
}

/* [Ind/Hercules] - socket_datasync */
void socket_datasync(int fd, bool send)
{
	struct {
		unsigned int length;/* short is not enough for some */
	} data_list[] = {
		{ sizeof(struct mmo_charstatus) },
		{ sizeof(struct quest) },
		{ sizeof(struct item) },
		{ sizeof(struct point) },
		{ sizeof(struct s_skill) },
		{ sizeof(struct status_change_data) },
		{ sizeof(struct storage_data) },
		{ sizeof(struct guild_storage) },
		{ sizeof(struct s_pet) },
		{ sizeof(struct s_mercenary) },
		{ sizeof(struct s_homunculus) },
		{ sizeof(struct s_elemental) },
		{ sizeof(struct s_friend) },
		{ sizeof(struct mail_message) },
		{ sizeof(struct mail_data) },
		{ sizeof(struct party_member) },
		{ sizeof(struct party) },
		{ sizeof(struct guild_member) },
		{ sizeof(struct guild_position) },
		{ sizeof(struct guild_alliance) },
		{ sizeof(struct guild_expulsion) },
		{ sizeof(struct guild_skill) },
		{ sizeof(struct guild) },
		{ sizeof(struct guild_castle) },
		{ sizeof(struct fame_list) },
		{ PACKETVER },
	};
	unsigned short i;
	unsigned int alen = ARRAYLENGTH(data_list);
	if( send ) {
		unsigned short p_len = ( alen * 4 ) + 4;
		WFIFOHEAD(fd, p_len);

		WFIFOW(fd, 0) = 0x2b0a;
		WFIFOW(fd, 2) = p_len;

		for( i = 0; i < alen; i++ ) {
			WFIFOL(fd, 4 + ( i * 4 ) ) = data_list[i].length;
		}

		WFIFOSET(fd, p_len);
	} else {
		for( i = 0; i < alen; i++ ) {
			if( RFIFOL(fd, 4 + (i * 4) ) != data_list[i].length ) {
				/* force the other to go wrong too so both are taken down */
				WFIFOHEAD(fd, 8);
				WFIFOW(fd, 0) = 0x2b0a;
				WFIFOW(fd, 2) = 8;
				WFIFOL(fd, 4) = 0;
				WFIFOSET(fd, 8);
				sockt->flush(fd);
				/* shut down */
				ShowFatalError("Servers are out of sync! recompile from scratch (%d)\n",i);
				exit(EXIT_FAILURE);
			}
		}
	}
}

#ifdef SEND_SHORTLIST
// Add a fd to the shortlist so that it'll be recognized as a fd that needs
// sending or eof handling.
void send_shortlist_add_fd(int fd)
{
	int i;
	int bit;

	if (!sockt->session_is_valid(fd))
		return;// out of range

	i = fd/32;
	bit = fd%32;

	if( (send_shortlist_set[i]>>bit)&1 )
		return;// already in the list

	if (send_shortlist_count >= ARRAYLENGTH(send_shortlist_array)) {
		ShowDebug("send_shortlist_add_fd: shortlist is full, ignoring... (fd=%d shortlist.count=%d shortlist.length=%d)\n",
		          fd, send_shortlist_count, ARRAYLENGTH(send_shortlist_array));
		return;
	}

	// set the bit
	send_shortlist_set[i] |= 1<<bit;
	// Add to the end of the shortlist array.
	send_shortlist_array[send_shortlist_count++] = fd;
}

// Do pending network sends and eof handling from the shortlist.
void send_shortlist_do_sends(void)
{
	int i;

	for( i = send_shortlist_count-1; i >= 0; --i )
	{
		int fd = send_shortlist_array[i];
		int idx = fd/32;
		int bit = fd%32;

		// Remove fd from shortlist, move the last fd to the current position
		--send_shortlist_count;
		send_shortlist_array[i] = send_shortlist_array[send_shortlist_count];
		send_shortlist_array[send_shortlist_count] = 0;

		if( fd <= 0 || fd >= FD_SETSIZE )
		{
			ShowDebug("send_shortlist_do_sends: fd is out of range, corrupted memory? (fd=%d)\n", fd);
			continue;
		}
		if( ((send_shortlist_set[idx]>>bit)&1) == 0 )
		{
			ShowDebug("send_shortlist_do_sends: fd is not set, why is it in the shortlist? (fd=%d)\n", fd);
			continue;
		}
		send_shortlist_set[idx]&=~(1<<bit);// unset fd
		// If this session still exists, perform send operations on it and
		// check for the eof state.
		if( sockt->session[fd] )
		{
			// Send data
			if( sockt->session[fd]->wdata_size )
				sockt->session[fd]->func_send(fd);

			// If it's been marked as eof, call the parse func on it so that
			// the socket will be immediately closed.
			if( sockt->session[fd]->flag.eof )
				sockt->session[fd]->func_parse(fd);

			// If the session still exists, is not eof and has things left to
			// be sent from it we'll re-add it to the shortlist.
			if( sockt->session[fd] && !sockt->session[fd]->flag.eof && sockt->session[fd]->wdata_size )
				send_shortlist_add_fd(fd);
		}
	}
}
#endif  // SEND_SHORTLIST

/**
 * Checks whether the given IP comes from LAN or WAN.
 *
 * @param[in]  ip   IP address to check.
 * @param[out] info Verbose output, if requested. Filled with the matching entry. Ignored if NULL.
 * @retval 0 if it is a WAN IP.
 * @return the appropriate LAN server address to send, if it is a LAN IP.
 */
uint32 socket_lan_subnet_check(uint32 ip, struct s_subnet *info)
{
	int i;
	ARR_FIND(0, VECTOR_LENGTH(sockt->lan_subnets), i, SUBNET_MATCH(ip, VECTOR_INDEX(sockt->lan_subnets, i).ip, VECTOR_INDEX(sockt->lan_subnets, i).mask));
	if (i != VECTOR_LENGTH(sockt->lan_subnets)) {
		if (info) {
			info->ip = VECTOR_INDEX(sockt->lan_subnets, i).ip;
			info->mask = VECTOR_INDEX(sockt->lan_subnets, i).mask;
		}
		return VECTOR_INDEX(sockt->lan_subnets, i).ip;
	}
	if (info) {
		info->ip = info->mask = 0;
	}
	return 0;
}

/**
 * Checks whether the given IP is allowed to connect as a server.
 *
 * @param ip IP address to check.
 * @retval true if we allow server connections from the given IP.
 * @retval false otherwise.
 */
bool socket_allowed_ip_check(uint32 ip)
{
	int i;
	ARR_FIND(0, VECTOR_LENGTH(sockt->allowed_ips), i, SUBNET_MATCH(ip, VECTOR_INDEX(sockt->allowed_ips, i).ip, VECTOR_INDEX(sockt->allowed_ips, i).mask));
	if (i != VECTOR_LENGTH(sockt->allowed_ips))
		return true;
	return sockt->trusted_ip_check(ip); // If an address is trusted, it's automatically also allowed.
}

/**
 * Checks whether the given IP is trusted and can skip ipban checks.
 *
 * @param ip IP address to check.
 * @retval true if we trust the given IP.
 * @retval false otherwise.
 */
bool socket_trusted_ip_check(uint32 ip)
{
	int i;
	ARR_FIND(0, VECTOR_LENGTH(sockt->trusted_ips), i, SUBNET_MATCH(ip, VECTOR_INDEX(sockt->trusted_ips, i).ip, VECTOR_INDEX(sockt->trusted_ips, i).mask));
	if (i != VECTOR_LENGTH(sockt->trusted_ips))
		return true;
	return false;
}

/**
 * Helper function to read a list of network.conf values.
 *
 * Entries will be appended to the variable-size array pointed to by list/count.
 *
 * @param[in]     t         The list to parse.
 * @param[in,out] list      Vector to append to. Must not be NULL (but the vector may be empty).
 * @param[in]     filename  Current filename, for output/logging reasons.
 * @param[in]     groupname Current group name, for output/logging reasons.
 * @return The amount of entries read, zero in case of errors.
 */
int socket_net_config_read_sub(struct config_setting_t *t, struct s_subnet_vector *list, const char *filename, const char *groupname)
{
	int i, len;
	char ipbuf[64], maskbuf[64];

	nullpo_retr(0, list);

	if (t == NULL)
		return 0;

	len = libconfig->setting_length(t);

	VECTOR_ENSURE(*list, len, 1);
	for (i = 0; i < len; ++i) {
		const char *subnet = libconfig->setting_get_string_elem(t, i);
		struct s_subnet *entry = NULL;

		if (sscanf(subnet, "%63[^:]:%63[^:]", ipbuf, maskbuf) != 2) {
			ShowWarning("Invalid IP:Subnet entry in configuration file %s: '%s' (%s)\n", filename, subnet, groupname);
			continue;
		}
		VECTOR_PUSHZEROED(*list);
		entry = &VECTOR_LAST(*list);
		entry->ip = sockt->str2ip(ipbuf);
		entry->mask = sockt->str2ip(maskbuf);
	}
	return (int)VECTOR_LENGTH(*list);
}

/**
 * Reads the network configuration file.
 *
 * @param filename The filename to read from.
 */
void socket_net_config_read(const char *filename)
{
	struct config_t network_config;
	int i;
	nullpo_retv(filename);

	if (!libconfig->load_file(&network_config, filename)) {
		ShowError("LAN Support configuration file is not found: '%s'. This server won't be able to accept connections from any servers.\n", filename);
		return;
	}

	VECTOR_CLEAR(sockt->lan_subnets);
	if (sockt->net_config_read_sub(libconfig->lookup(&network_config, "lan_subnets"), &sockt->lan_subnets, filename, "lan_subnets") > 0)
		ShowStatus("Read information about %d LAN subnets.\n", (int)VECTOR_LENGTH(sockt->lan_subnets));

	VECTOR_CLEAR(sockt->trusted_ips);
	if (sockt->net_config_read_sub(libconfig->lookup(&network_config, "trusted"), &sockt->trusted_ips, filename, "trusted") > 0)
		ShowStatus("Read information about %d trusted IP ranges.\n", (int)VECTOR_LENGTH(sockt->trusted_ips));
	ARR_FIND(0, VECTOR_LENGTH(sockt->trusted_ips), i, SUBNET_MATCH(0, VECTOR_INDEX(sockt->trusted_ips, i).ip, VECTOR_INDEX(sockt->trusted_ips, i).mask));
	if (i != VECTOR_LENGTH(sockt->trusted_ips)) {
		ShowError("Using a wildcard IP range in the trusted server IPs is NOT RECOMMENDED.\n");
		ShowNotice("Please edit your '%s' trusted list to fit your network configuration.\n", filename);
	}

	VECTOR_CLEAR(sockt->allowed_ips);
	if (sockt->net_config_read_sub(libconfig->lookup(&network_config, "allowed"), &sockt->allowed_ips, filename, "allowed") > 0)
		ShowStatus("Read information about %d allowed server IP ranges.\n", (int)VECTOR_LENGTH(sockt->allowed_ips));
	if (VECTOR_LENGTH(sockt->allowed_ips) + VECTOR_LENGTH(sockt->trusted_ips) == 0) {
		ShowError("No allowed server IP ranges configured. This server won't be able to accept connections from any char servers.\n");
	}
	ARR_FIND(0, VECTOR_LENGTH(sockt->allowed_ips), i, SUBNET_MATCH(0, VECTOR_INDEX(sockt->allowed_ips, i).ip, VECTOR_INDEX(sockt->allowed_ips, i).mask));
#ifndef BUILDBOT
	if (i != VECTOR_LENGTH(sockt->allowed_ips)) {
		ShowWarning("Using a wildcard IP range in the allowed server IPs is NOT RECOMMENDED.\n");
		ShowNotice("Please edit your '%s' allowed list to fit your network configuration.\n", filename);
	}
#endif  // BUILDBOT
	libconfig->destroy(&network_config);
	return;
}

void socket_defaults(void)
{
	sockt = &sockt_s;

	sockt->fd_max = 0;
	/* */
	sockt->stall_time = 60;
	sockt->last_tick = 0;
	/* */
	memset(&sockt->addr_, 0, sizeof(sockt->addr_));
	sockt->naddr_ = 0;
	/* */
	VECTOR_INIT(sockt->lan_subnets);
	VECTOR_INIT(sockt->allowed_ips);
	VECTOR_INIT(sockt->trusted_ips);

	sockt->init = socket_init;
	sockt->final = socket_final;
	/* */
	sockt->perform = do_sockets;
	/* */
	sockt->datasync = socket_datasync;
	/* */
	sockt->make_listen_bind = make_listen_bind;
	sockt->make_connection = make_connection;
	sockt->realloc_fifo = realloc_fifo;
	sockt->realloc_writefifo = realloc_writefifo;
	sockt->wfifoset = wfifoset;
	sockt->rfifoskip = rfifoskip;
	sockt->close = socket_close;
	/* */
	sockt->session_is_valid = session_is_valid;
	sockt->session_is_active = session_is_active;
	/* */
	sockt->flush = flush_fifo;
	sockt->flush_fifos = flush_fifos;
	sockt->set_nonblocking = set_nonblocking;
	sockt->set_defaultparse = set_defaultparse;
	sockt->host2ip = host2ip;
	sockt->ip2str = ip2str;
	sockt->str2ip = str2ip;
	sockt->ntows = ntows;
	sockt->getips = socket_getips;
	sockt->eof = set_eof;

	sockt->lan_subnet_check = socket_lan_subnet_check;
	sockt->allowed_ip_check = socket_allowed_ip_check;
	sockt->trusted_ip_check = socket_trusted_ip_check;
	sockt->net_config_read_sub = socket_net_config_read_sub;
	sockt->net_config_read = socket_net_config_read;
}