aboutsummaryrefslogtreecommitdiff
path: root/sys/dev/netmap/netmap.c
blob: 05b369fad4fd8eef9d3cad133f3a955a61e1f1cd (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
/*
 * Copyright (C) 2011-2012 Matteo Landi, Luigi Rizzo. All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 *   1. Redistributions of source code must retain the above copyright
 *      notice, this list of conditions and the following disclaimer.
 *   2. Redistributions in binary form must reproduce the above copyright
 *      notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 *
 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 */

#define NM_BRIDGE

/*
 * This module supports memory mapped access to network devices,
 * see netmap(4).
 *
 * The module uses a large, memory pool allocated by the kernel
 * and accessible as mmapped memory by multiple userspace threads/processes.
 * The memory pool contains packet buffers and "netmap rings",
 * i.e. user-accessible copies of the interface's queues.
 *
 * Access to the network card works like this:
 * 1. a process/thread issues one or more open() on /dev/netmap, to create
 *    select()able file descriptor on which events are reported.
 * 2. on each descriptor, the process issues an ioctl() to identify
 *    the interface that should report events to the file descriptor.
 * 3. on each descriptor, the process issues an mmap() request to
 *    map the shared memory region within the process' address space.
 *    The list of interesting queues is indicated by a location in
 *    the shared memory region.
 * 4. using the functions in the netmap(4) userspace API, a process
 *    can look up the occupation state of a queue, access memory buffers,
 *    and retrieve received packets or enqueue packets to transmit.
 * 5. using some ioctl()s the process can synchronize the userspace view
 *    of the queue with the actual status in the kernel. This includes both
 *    receiving the notification of new packets, and transmitting new
 *    packets on the output interface.
 * 6. select() or poll() can be used to wait for events on individual
 *    transmit or receive queues (or all queues for a given interface).
 */

#ifdef linux
#include "bsd_glue.h"
static netdev_tx_t linux_netmap_start(struct sk_buff *skb, struct net_device *dev);
#endif /* linux */

#ifdef __APPLE__
#include "osx_glue.h"
#endif /* __APPLE__ */

#ifdef __FreeBSD__
#include <sys/cdefs.h> /* prerequisite */
__FBSDID("$FreeBSD$");

#include <sys/types.h>
#include <sys/module.h>
#include <sys/errno.h>
#include <sys/param.h>	/* defines used in kernel.h */
#include <sys/jail.h>
#include <sys/kernel.h>	/* types used in module initialization */
#include <sys/conf.h>	/* cdevsw struct */
#include <sys/uio.h>	/* uio struct */
#include <sys/sockio.h>
#include <sys/socketvar.h>	/* struct socket */
#include <sys/malloc.h>
#include <sys/mman.h>	/* PROT_EXEC */
#include <sys/poll.h>
#include <sys/proc.h>
#include <vm/vm.h>	/* vtophys */
#include <vm/pmap.h>	/* vtophys */
#include <sys/socket.h> /* sockaddrs */
#include <machine/bus.h>
#include <sys/selinfo.h>
#include <sys/sysctl.h>
#include <net/if.h>
#include <net/bpf.h>		/* BIOCIMMEDIATE */
#include <net/vnet.h>
#include <machine/bus.h>	/* bus_dmamap_* */

MALLOC_DEFINE(M_NETMAP, "netmap", "Network memory map");
#endif /* __FreeBSD__ */

#include <net/netmap.h>
#include <dev/netmap/netmap_kern.h>

/*
 * lock and unlock for the netmap memory allocator
 */
#define NMA_LOCK()	mtx_lock(&nm_mem->nm_mtx);
#define NMA_UNLOCK()	mtx_unlock(&nm_mem->nm_mtx);
struct netmap_mem_d;
static struct netmap_mem_d *nm_mem;	/* Our memory allocator. */

u_int netmap_total_buffers;
char *netmap_buffer_base;	/* address of an invalid buffer */

/* user-controlled variables */
int netmap_verbose;

static int netmap_no_timestamp; /* don't timestamp on rxsync */

SYSCTL_NODE(_dev, OID_AUTO, netmap, CTLFLAG_RW, 0, "Netmap args");
SYSCTL_INT(_dev_netmap, OID_AUTO, verbose,
    CTLFLAG_RW, &netmap_verbose, 0, "Verbose mode");
SYSCTL_INT(_dev_netmap, OID_AUTO, no_timestamp,
    CTLFLAG_RW, &netmap_no_timestamp, 0, "no_timestamp");
u_int netmap_buf_size = 2048;
TUNABLE_INT("hw.netmap.buf_size", (u_int *)&netmap_buf_size);
SYSCTL_INT(_dev_netmap, OID_AUTO, buf_size,
    CTLFLAG_RD, &netmap_buf_size, 0, "Size of packet buffers");
int netmap_mitigate = 1;
SYSCTL_INT(_dev_netmap, OID_AUTO, mitigate, CTLFLAG_RW, &netmap_mitigate, 0, "");
int netmap_no_pendintr = 1;
SYSCTL_INT(_dev_netmap, OID_AUTO, no_pendintr,
    CTLFLAG_RW, &netmap_no_pendintr, 0, "Always look for new received packets.");

int netmap_drop = 0;	/* debugging */
int netmap_flags = 0;	/* debug flags */
int netmap_copy = 0;	/* debugging, copy content */

SYSCTL_INT(_dev_netmap, OID_AUTO, drop, CTLFLAG_RW, &netmap_drop, 0 , "");
SYSCTL_INT(_dev_netmap, OID_AUTO, flags, CTLFLAG_RW, &netmap_flags, 0 , "");
SYSCTL_INT(_dev_netmap, OID_AUTO, copy, CTLFLAG_RW, &netmap_copy, 0 , "");

#ifdef NM_BRIDGE /* support for netmap bridge */

/*
 * system parameters.
 *
 * All switched ports have prefix NM_NAME.
 * The switch has a max of NM_BDG_MAXPORTS ports (often stored in a bitmap,
 * so a practical upper bound is 64).
 * Each tx ring is read-write, whereas rx rings are readonly (XXX not done yet).
 * The virtual interfaces use per-queue lock instead of core lock.
 * In the tx loop, we aggregate traffic in batches to make all operations
 * faster. The batch size is NM_BDG_BATCH
 */
#define	NM_NAME			"vale"	/* prefix for the interface */
#define NM_BDG_MAXPORTS		16	/* up to 64 ? */
#define NM_BRIDGE_RINGSIZE	1024	/* in the device */
#define NM_BDG_HASH		1024	/* forwarding table entries */
#define NM_BDG_BATCH		1024	/* entries in the forwarding buffer */
#define	NM_BRIDGES		4	/* number of bridges */
int netmap_bridge = NM_BDG_BATCH; /* bridge batch size */
SYSCTL_INT(_dev_netmap, OID_AUTO, bridge, CTLFLAG_RW, &netmap_bridge, 0 , "");

#ifdef linux
#define	ADD_BDG_REF(ifp)	(NA(ifp)->if_refcount++)
#define	DROP_BDG_REF(ifp)	(NA(ifp)->if_refcount-- <= 1)
#else /* !linux */
#define	ADD_BDG_REF(ifp)	(ifp)->if_refcount++
#define	DROP_BDG_REF(ifp)	refcount_release(&(ifp)->if_refcount)
#ifdef __FreeBSD__
#include <sys/endian.h>
#include <sys/refcount.h>
#endif /* __FreeBSD__ */
#define prefetch(x)	__builtin_prefetch(x)
#endif /* !linux */

static void bdg_netmap_attach(struct ifnet *ifp);
static int bdg_netmap_reg(struct ifnet *ifp, int onoff);
/* per-tx-queue entry */
struct nm_bdg_fwd {	/* forwarding entry for a bridge */
	void *buf;
	uint64_t dst;	/* dst mask */
	uint32_t src;	/* src index ? */
	uint16_t len;	/* src len */
};

struct nm_hash_ent {
	uint64_t	mac;	/* the top 2 bytes are the epoch */
	uint64_t	ports;
};

/*
 * Interfaces for a bridge are all in ports[].
 * The array has fixed size, an empty entry does not terminate
 * the search.
 */
struct nm_bridge {
	struct ifnet *bdg_ports[NM_BDG_MAXPORTS];
	int n_ports;
	uint64_t act_ports;
	int freelist;	/* first buffer index */
	NM_SELINFO_T si;	/* poll/select wait queue */
	NM_LOCK_T bdg_lock;	/* protect the selinfo ? */

	/* the forwarding table, MAC+ports */
	struct nm_hash_ent ht[NM_BDG_HASH];

	int namelen;	/* 0 means free */
	char basename[IFNAMSIZ];
};

struct nm_bridge nm_bridges[NM_BRIDGES];

#define BDG_LOCK(b)	mtx_lock(&(b)->bdg_lock)
#define BDG_UNLOCK(b)	mtx_unlock(&(b)->bdg_lock)

/*
 * NA(ifp)->bdg_port	port index
 */

// XXX only for multiples of 64 bytes, non overlapped.
static inline void
pkt_copy(void *_src, void *_dst, int l)
{
        uint64_t *src = _src;
        uint64_t *dst = _dst;
        if (unlikely(l >= 1024)) {
                bcopy(src, dst, l);
                return;
        }
        for (; likely(l > 0); l-=64) {
                *dst++ = *src++;
                *dst++ = *src++;
                *dst++ = *src++;
                *dst++ = *src++;
                *dst++ = *src++;
                *dst++ = *src++;
                *dst++ = *src++;
                *dst++ = *src++;
        }
}

/*
 * locate a bridge among the existing ones.
 * a ':' in the name terminates the bridge name. Otherwise, just NM_NAME.
 * We assume that this is called with a name of at least NM_NAME chars.
 */
static struct nm_bridge *
nm_find_bridge(const char *name)
{
	int i, l, namelen, e;
	struct nm_bridge *b = NULL;

	namelen = strlen(NM_NAME);	/* base length */
	l = strlen(name);		/* actual length */
	for (i = namelen + 1; i < l; i++) {
		if (name[i] == ':') {
			namelen = i;
			break;
		}
	}
	if (namelen >= IFNAMSIZ)
		namelen = IFNAMSIZ;
	ND("--- prefix is '%.*s' ---", namelen, name);

	/* use the first entry for locking */
	BDG_LOCK(nm_bridges); // XXX do better
	for (e = -1, i = 1; i < NM_BRIDGES; i++) {
		b = nm_bridges + i;
		if (b->namelen == 0)
			e = i;	/* record empty slot */
		else if (strncmp(name, b->basename, namelen) == 0) {
			ND("found '%.*s' at %d", namelen, name, i);
			break;
		}
	}
	if (i == NM_BRIDGES) { /* all full */
		if (e == -1) { /* no empty slot */
			b = NULL;
		} else {
			b = nm_bridges + e;
			strncpy(b->basename, name, namelen);
			b->namelen = namelen;
		}
	}
	BDG_UNLOCK(nm_bridges);
	return b;
}
#endif /* NM_BRIDGE */

/*------------- memory allocator -----------------*/
#ifdef NETMAP_MEM2
#include "netmap_mem2.c"
#else /* !NETMAP_MEM2 */
#include "netmap_mem1.c"
#endif /* !NETMAP_MEM2 */
/*------------ end of memory allocator ----------*/

/* Structure associated to each thread which registered an interface. */
struct netmap_priv_d {
	struct netmap_if *np_nifp;	/* netmap interface descriptor. */

	struct ifnet	*np_ifp;	/* device for which we hold a reference */
	int		np_ringid;	/* from the ioctl */
	u_int		np_qfirst, np_qlast;	/* range of rings to scan */
	uint16_t	np_txpoll;
};


/*
 * File descriptor's private data destructor.
 *
 * Call nm_register(ifp,0) to stop netmap mode on the interface and
 * revert to normal operation. We expect that np_ifp has not gone.
 */
static void
netmap_dtor_locked(void *data)
{
	struct netmap_priv_d *priv = data;
	struct ifnet *ifp = priv->np_ifp;
	struct netmap_adapter *na = NA(ifp);
	struct netmap_if *nifp = priv->np_nifp;

	na->refcount--;
	if (na->refcount <= 0) {	/* last instance */
		u_int i, j, lim;

		D("deleting last netmap instance for %s", ifp->if_xname);
		/*
		 * there is a race here with *_netmap_task() and
		 * netmap_poll(), which don't run under NETMAP_REG_LOCK.
		 * na->refcount == 0 && na->ifp->if_capenable & IFCAP_NETMAP
		 * (aka NETMAP_DELETING(na)) are a unique marker that the
		 * device is dying.
		 * Before destroying stuff we sleep a bit, and then complete
		 * the job. NIOCREG should realize the condition and
		 * loop until they can continue; the other routines
		 * should check the condition at entry and quit if
		 * they cannot run.
		 */
		na->nm_lock(ifp, NETMAP_REG_UNLOCK, 0);
		tsleep(na, 0, "NIOCUNREG", 4);
		na->nm_lock(ifp, NETMAP_REG_LOCK, 0);
		na->nm_register(ifp, 0); /* off, clear IFCAP_NETMAP */
		/* Wake up any sleeping threads. netmap_poll will
		 * then return POLLERR
		 */
		for (i = 0; i < na->num_tx_rings + 1; i++)
			selwakeuppri(&na->tx_rings[i].si, PI_NET);
		for (i = 0; i < na->num_rx_rings + 1; i++)
			selwakeuppri(&na->rx_rings[i].si, PI_NET);
		selwakeuppri(&na->tx_si, PI_NET);
		selwakeuppri(&na->rx_si, PI_NET);
		/* release all buffers */
		NMA_LOCK();
		for (i = 0; i < na->num_tx_rings + 1; i++) {
			struct netmap_ring *ring = na->tx_rings[i].ring;
			lim = na->tx_rings[i].nkr_num_slots;
			for (j = 0; j < lim; j++)
				netmap_free_buf(nifp, ring->slot[j].buf_idx);
			/* knlist_destroy(&na->tx_rings[i].si.si_note); */
			mtx_destroy(&na->tx_rings[i].q_lock);
		}
		for (i = 0; i < na->num_rx_rings + 1; i++) {
			struct netmap_ring *ring = na->rx_rings[i].ring;
			lim = na->rx_rings[i].nkr_num_slots;
			for (j = 0; j < lim; j++)
				netmap_free_buf(nifp, ring->slot[j].buf_idx);
			/* knlist_destroy(&na->rx_rings[i].si.si_note); */
			mtx_destroy(&na->rx_rings[i].q_lock);
		}
		/* XXX kqueue(9) needed; these will mirror knlist_init. */
		/* knlist_destroy(&na->tx_si.si_note); */
		/* knlist_destroy(&na->rx_si.si_note); */
		NMA_UNLOCK();
		netmap_free_rings(na);
		wakeup(na);
	}
	netmap_if_free(nifp);
}

static void
nm_if_rele(struct ifnet *ifp)
{
#ifndef NM_BRIDGE
	if_rele(ifp);
#else /* NM_BRIDGE */
	int i, full;
	struct nm_bridge *b;

	if (strncmp(ifp->if_xname, NM_NAME, sizeof(NM_NAME) - 1)) {
		if_rele(ifp);
		return;
	}
	if (!DROP_BDG_REF(ifp))
		return;
	b = ifp->if_bridge;
	BDG_LOCK(nm_bridges);
	BDG_LOCK(b);
	ND("want to disconnect %s from the bridge", ifp->if_xname);
	full = 0;
	for (i = 0; i < NM_BDG_MAXPORTS; i++) {
		if (b->bdg_ports[i] == ifp) {
			b->bdg_ports[i] = NULL;
			bzero(ifp, sizeof(*ifp));
			free(ifp, M_DEVBUF);
			break;
		} 
		else if (b->bdg_ports[i] != NULL)
			full = 1;
	}
	BDG_UNLOCK(b);
	if (full == 0) {
		ND("freeing bridge %d", b - nm_bridges);
		b->namelen = 0;
	}
	BDG_UNLOCK(nm_bridges);
	if (i == NM_BDG_MAXPORTS)
		D("ouch, cannot find ifp to remove");
#endif /* NM_BRIDGE */
}

static void
netmap_dtor(void *data)
{
	struct netmap_priv_d *priv = data;
	struct ifnet *ifp = priv->np_ifp;
	struct netmap_adapter *na = NA(ifp);

	na->nm_lock(ifp, NETMAP_REG_LOCK, 0);
	netmap_dtor_locked(data);
	na->nm_lock(ifp, NETMAP_REG_UNLOCK, 0);

	nm_if_rele(ifp);
	bzero(priv, sizeof(*priv));	/* XXX for safety */
	free(priv, M_DEVBUF);
}


/*
 * mmap(2) support for the "netmap" device.
 *
 * Expose all the memory previously allocated by our custom memory
 * allocator: this way the user has only to issue a single mmap(2), and
 * can work on all the data structures flawlessly.
 *
 * Return 0 on success, -1 otherwise.
 */

#ifdef __FreeBSD__
static int
netmap_mmap(__unused struct cdev *dev,
#if __FreeBSD_version < 900000
		vm_offset_t offset, vm_paddr_t *paddr, int nprot
#else
		vm_ooffset_t offset, vm_paddr_t *paddr, int nprot,
		__unused vm_memattr_t *memattr
#endif
	)
{
	if (nprot & PROT_EXEC)
		return (-1);	// XXX -1 or EINVAL ?

	ND("request for offset 0x%x", (uint32_t)offset);
	*paddr = netmap_ofstophys(offset);

	return (0);
}
#endif /* __FreeBSD__ */


/*
 * Handlers for synchronization of the queues from/to the host.
 *
 * netmap_sync_to_host() passes packets up. We are called from a
 * system call in user process context, and the only contention
 * can be among multiple user threads erroneously calling
 * this routine concurrently. In principle we should not even
 * need to lock.
 */
static void
netmap_sync_to_host(struct netmap_adapter *na)
{
	struct netmap_kring *kring = &na->tx_rings[na->num_tx_rings];
	struct netmap_ring *ring = kring->ring;
	struct mbuf *head = NULL, *tail = NULL, *m;
	u_int k, n, lim = kring->nkr_num_slots - 1;

	k = ring->cur;
	if (k > lim) {
		netmap_ring_reinit(kring);
		return;
	}
	// na->nm_lock(na->ifp, NETMAP_CORE_LOCK, 0);

	/* Take packets from hwcur to cur and pass them up.
	 * In case of no buffers we give up. At the end of the loop,
	 * the queue is drained in all cases.
	 */
	for (n = kring->nr_hwcur; n != k;) {
		struct netmap_slot *slot = &ring->slot[n];

		n = (n == lim) ? 0 : n + 1;
		if (slot->len < 14 || slot->len > NETMAP_BUF_SIZE) {
			D("bad pkt at %d len %d", n, slot->len);
			continue;
		}
		m = m_devget(NMB(slot), slot->len, 0, na->ifp, NULL);

		if (m == NULL)
			break;
		if (tail)
			tail->m_nextpkt = m;
		else
			head = m;
		tail = m;
		m->m_nextpkt = NULL;
	}
	kring->nr_hwcur = k;
	kring->nr_hwavail = ring->avail = lim;
	// na->nm_lock(na->ifp, NETMAP_CORE_UNLOCK, 0);

	/* send packets up, outside the lock */
	while ((m = head) != NULL) {
		head = head->m_nextpkt;
		m->m_nextpkt = NULL;
		if (netmap_verbose & NM_VERB_HOST)
			D("sending up pkt %p size %d", m, MBUF_LEN(m));
		NM_SEND_UP(na->ifp, m);
	}
}

/*
 * rxsync backend for packets coming from the host stack.
 * They have been put in the queue by netmap_start() so we
 * need to protect access to the kring using a lock.
 *
 * This routine also does the selrecord if called from the poll handler
 * (we know because td != NULL).
 *
 * NOTE: on linux, selrecord() is defined as a macro and uses pwait
 *     as an additional hidden argument.
 */
static void
netmap_sync_from_host(struct netmap_adapter *na, struct thread *td, void *pwait)
{
	struct netmap_kring *kring = &na->rx_rings[na->num_rx_rings];
	struct netmap_ring *ring = kring->ring;
	u_int j, n, lim = kring->nkr_num_slots;
	u_int k = ring->cur, resvd = ring->reserved;

	(void)pwait;	/* disable unused warnings */
	na->nm_lock(na->ifp, NETMAP_CORE_LOCK, 0);
	if (k >= lim) {
		netmap_ring_reinit(kring);
		return;
	}
	/* new packets are already set in nr_hwavail */
	/* skip past packets that userspace has released */
	j = kring->nr_hwcur;
	if (resvd > 0) {
		if (resvd + ring->avail >= lim + 1) {
			D("XXX invalid reserve/avail %d %d", resvd, ring->avail);
			ring->reserved = resvd = 0; // XXX panic...
		}
		k = (k >= resvd) ? k - resvd : k + lim - resvd;
        }
	if (j != k) {
		n = k >= j ? k - j : k + lim - j;
		kring->nr_hwavail -= n;
		kring->nr_hwcur = k;
	}
	k = ring->avail = kring->nr_hwavail - resvd;
	if (k == 0 && td)
		selrecord(td, &kring->si);
	if (k && (netmap_verbose & NM_VERB_HOST))
		D("%d pkts from stack", k);
	na->nm_lock(na->ifp, NETMAP_CORE_UNLOCK, 0);
}


/*
 * get a refcounted reference to an interface.
 * Return ENXIO if the interface does not exist, EINVAL if netmap
 * is not supported by the interface.
 * If successful, hold a reference.
 */
static int
get_ifp(const char *name, struct ifnet **ifp)
{
#ifdef NM_BRIDGE
	struct ifnet *iter = NULL;

	do {
		struct nm_bridge *b;
		int i, l, cand = -1;

		if (strncmp(name, NM_NAME, sizeof(NM_NAME) - 1))
			break;
		b = nm_find_bridge(name);
		if (b == NULL) {
			D("no bridges available for '%s'", name);
			return (ENXIO);
		}
		/* XXX locking */
		BDG_LOCK(b);
		/* lookup in the local list of ports */
		for (i = 0; i < NM_BDG_MAXPORTS; i++) {
			iter = b->bdg_ports[i];
			if (iter == NULL) {
				if (cand == -1)
					cand = i; /* potential insert point */
				continue;
			}
			if (!strcmp(iter->if_xname, name)) {
				ADD_BDG_REF(iter);
				ND("found existing interface");
				BDG_UNLOCK(b);
				break;
			}
		}
		if (i < NM_BDG_MAXPORTS) /* already unlocked */
			break;
		if (cand == -1) {
			D("bridge full, cannot create new port");
no_port:
			BDG_UNLOCK(b);
			*ifp = NULL;
			return EINVAL;
		}
		ND("create new bridge port %s", name);
		/* space for forwarding list after the ifnet */
		l = sizeof(*iter) +
			 sizeof(struct nm_bdg_fwd)*NM_BDG_BATCH ;
		iter = malloc(l, M_DEVBUF, M_NOWAIT | M_ZERO);
		if (!iter)
			goto no_port;
		strcpy(iter->if_xname, name);
		bdg_netmap_attach(iter);
		b->bdg_ports[cand] = iter;
		iter->if_bridge = b;
		ADD_BDG_REF(iter);
		BDG_UNLOCK(b);
		ND("attaching virtual bridge %p", b);
	} while (0);
	*ifp = iter;
	if (! *ifp)
#endif /* NM_BRIDGE */
	*ifp = ifunit_ref(name);
	if (*ifp == NULL)
		return (ENXIO);
	/* can do this if the capability exists and if_pspare[0]
	 * points to the netmap descriptor.
	 */
	if ((*ifp)->if_capabilities & IFCAP_NETMAP && NA(*ifp))
		return 0;	/* valid pointer, we hold the refcount */
	nm_if_rele(*ifp);
	return EINVAL;	// not NETMAP capable
}


/*
 * Error routine called when txsync/rxsync detects an error.
 * Can't do much more than resetting cur = hwcur, avail = hwavail.
 * Return 1 on reinit.
 *
 * This routine is only called by the upper half of the kernel.
 * It only reads hwcur (which is changed only by the upper half, too)
 * and hwavail (which may be changed by the lower half, but only on
 * a tx ring and only to increase it, so any error will be recovered
 * on the next call). For the above, we don't strictly need to call
 * it under lock.
 */
int
netmap_ring_reinit(struct netmap_kring *kring)
{
	struct netmap_ring *ring = kring->ring;
	u_int i, lim = kring->nkr_num_slots - 1;
	int errors = 0;

	D("called for %s", kring->na->ifp->if_xname);
	if (ring->cur > lim)
		errors++;
	for (i = 0; i <= lim; i++) {
		u_int idx = ring->slot[i].buf_idx;
		u_int len = ring->slot[i].len;
		if (idx < 2 || idx >= netmap_total_buffers) {
			if (!errors++)
				D("bad buffer at slot %d idx %d len %d ", i, idx, len);
			ring->slot[i].buf_idx = 0;
			ring->slot[i].len = 0;
		} else if (len > NETMAP_BUF_SIZE) {
			ring->slot[i].len = 0;
			if (!errors++)
				D("bad len %d at slot %d idx %d",
					len, i, idx);
		}
	}
	if (errors) {
		int pos = kring - kring->na->tx_rings;
		int n = kring->na->num_tx_rings + 1;

		D("total %d errors", errors);
		errors++;
		D("%s %s[%d] reinit, cur %d -> %d avail %d -> %d",
			kring->na->ifp->if_xname,
			pos < n ?  "TX" : "RX", pos < n ? pos : pos - n,
			ring->cur, kring->nr_hwcur,
			ring->avail, kring->nr_hwavail);
		ring->cur = kring->nr_hwcur;
		ring->avail = kring->nr_hwavail;
	}
	return (errors ? 1 : 0);
}


/*
 * Set the ring ID. For devices with a single queue, a request
 * for all rings is the same as a single ring.
 */
static int
netmap_set_ringid(struct netmap_priv_d *priv, u_int ringid)
{
	struct ifnet *ifp = priv->np_ifp;
	struct netmap_adapter *na = NA(ifp);
	u_int i = ringid & NETMAP_RING_MASK;
	/* initially (np_qfirst == np_qlast) we don't want to lock */
	int need_lock = (priv->np_qfirst != priv->np_qlast);
	int lim = na->num_rx_rings;

	if (na->num_tx_rings > lim)
		lim = na->num_tx_rings;
	if ( (ringid & NETMAP_HW_RING) && i >= lim) {
		D("invalid ring id %d", i);
		return (EINVAL);
	}
	if (need_lock)
		na->nm_lock(ifp, NETMAP_CORE_LOCK, 0);
	priv->np_ringid = ringid;
	if (ringid & NETMAP_SW_RING) {
		priv->np_qfirst = NETMAP_SW_RING;
		priv->np_qlast = 0;
	} else if (ringid & NETMAP_HW_RING) {
		priv->np_qfirst = i;
		priv->np_qlast = i + 1;
	} else {
		priv->np_qfirst = 0;
		priv->np_qlast = NETMAP_HW_RING ;
	}
	priv->np_txpoll = (ringid & NETMAP_NO_TX_POLL) ? 0 : 1;
	if (need_lock)
		na->nm_lock(ifp, NETMAP_CORE_UNLOCK, 0);
	if (ringid & NETMAP_SW_RING)
		D("ringid %s set to SW RING", ifp->if_xname);
	else if (ringid & NETMAP_HW_RING)
		D("ringid %s set to HW RING %d", ifp->if_xname,
			priv->np_qfirst);
	else
		D("ringid %s set to all %d HW RINGS", ifp->if_xname, lim);
	return 0;
}

/*
 * ioctl(2) support for the "netmap" device.
 *
 * Following a list of accepted commands:
 * - NIOCGINFO
 * - SIOCGIFADDR	just for convenience
 * - NIOCREGIF
 * - NIOCUNREGIF
 * - NIOCTXSYNC
 * - NIOCRXSYNC
 *
 * Return 0 on success, errno otherwise.
 */
static int
netmap_ioctl(struct cdev *dev, u_long cmd, caddr_t data,
	int fflag, struct thread *td)
{
	struct netmap_priv_d *priv = NULL;
	struct ifnet *ifp;
	struct nmreq *nmr = (struct nmreq *) data;
	struct netmap_adapter *na;
	int error;
	u_int i, lim;
	struct netmap_if *nifp;

	(void)dev;	/* UNUSED */
	(void)fflag;	/* UNUSED */
#ifdef linux
#define devfs_get_cdevpriv(pp)				\
	({ *(struct netmap_priv_d **)pp = ((struct file *)td)->private_data; 	\
		(*pp ? 0 : ENOENT); })

/* devfs_set_cdevpriv cannot fail on linux */
#define devfs_set_cdevpriv(p, fn)				\
	({ ((struct file *)td)->private_data = p; (p ? 0 : EINVAL); })


#define devfs_clear_cdevpriv()	do {				\
		netmap_dtor(priv); ((struct file *)td)->private_data = 0;	\
	} while (0)
#endif /* linux */

	CURVNET_SET(TD_TO_VNET(td));

	error = devfs_get_cdevpriv((void **)&priv);
	if (error != ENOENT && error != 0) {
		CURVNET_RESTORE();
		return (error);
	}

	error = 0;	/* Could be ENOENT */
	nmr->nr_name[sizeof(nmr->nr_name) - 1] = '\0';	/* truncate name */
	switch (cmd) {
	case NIOCGINFO:		/* return capabilities etc */
		/* memsize is always valid */
		nmr->nr_memsize = nm_mem->nm_totalsize;
		nmr->nr_offset = 0;
		nmr->nr_rx_rings = nmr->nr_tx_rings = 0;
		nmr->nr_rx_slots = nmr->nr_tx_slots = 0;
		if (nmr->nr_version != NETMAP_API) {
			D("API mismatch got %d have %d",
				nmr->nr_version, NETMAP_API);
			nmr->nr_version = NETMAP_API;
			error = EINVAL;
			break;
		}
		if (nmr->nr_name[0] == '\0')	/* just get memory info */
			break;
		error = get_ifp(nmr->nr_name, &ifp); /* get a refcount */
		if (error)
			break;
		na = NA(ifp); /* retrieve netmap_adapter */
		nmr->nr_rx_rings = na->num_rx_rings;
		nmr->nr_tx_rings = na->num_tx_rings;
		nmr->nr_rx_slots = na->num_rx_desc;
		nmr->nr_tx_slots = na->num_tx_desc;
		nm_if_rele(ifp);	/* return the refcount */
		break;

	case NIOCREGIF:
		if (nmr->nr_version != NETMAP_API) {
			nmr->nr_version = NETMAP_API;
			error = EINVAL;
			break;
		}
		if (priv != NULL) {	/* thread already registered */
			error = netmap_set_ringid(priv, nmr->nr_ringid);
			break;
		}
		/* find the interface and a reference */
		error = get_ifp(nmr->nr_name, &ifp); /* keep reference */
		if (error)
			break;
		na = NA(ifp); /* retrieve netmap adapter */
		/*
		 * Allocate the private per-thread structure.
		 * XXX perhaps we can use a blocking malloc ?
		 */
		priv = malloc(sizeof(struct netmap_priv_d), M_DEVBUF,
			      M_NOWAIT | M_ZERO);
		if (priv == NULL) {
			error = ENOMEM;
			nm_if_rele(ifp);   /* return the refcount */
			break;
		}

		for (i = 10; i > 0; i--) {
			na->nm_lock(ifp, NETMAP_REG_LOCK, 0);
			if (!NETMAP_DELETING(na))
				break;
			na->nm_lock(ifp, NETMAP_REG_UNLOCK, 0);
			tsleep(na, 0, "NIOCREGIF", hz/10);
		}
		if (i == 0) {
			D("too many NIOCREGIF attempts, give up");
			error = EINVAL;
			free(priv, M_DEVBUF);
			nm_if_rele(ifp);	/* return the refcount */
			break;
		}

		priv->np_ifp = ifp;	/* store the reference */
		error = netmap_set_ringid(priv, nmr->nr_ringid);
		if (error)
			goto error;
		priv->np_nifp = nifp = netmap_if_new(nmr->nr_name, na);
		if (nifp == NULL) { /* allocation failed */
			error = ENOMEM;
		} else if (ifp->if_capenable & IFCAP_NETMAP) {
			/* was already set */
		} else {
			/* Otherwise set the card in netmap mode
			 * and make it use the shared buffers.
			 */
			for (i = 0 ; i < na->num_tx_rings + 1; i++)
				mtx_init(&na->tx_rings[i].q_lock, "nm_txq_lock", MTX_NETWORK_LOCK, MTX_DEF);
			for (i = 0 ; i < na->num_rx_rings + 1; i++) {
				mtx_init(&na->rx_rings[i].q_lock, "nm_rxq_lock", MTX_NETWORK_LOCK, MTX_DEF);
			}
			error = na->nm_register(ifp, 1); /* mode on */
			if (error)
				netmap_dtor_locked(priv);
		}

		if (error) {	/* reg. failed, release priv and ref */
error:
			na->nm_lock(ifp, NETMAP_REG_UNLOCK, 0);
			nm_if_rele(ifp);	/* return the refcount */
			bzero(priv, sizeof(*priv));
			free(priv, M_DEVBUF);
			break;
		}

		na->nm_lock(ifp, NETMAP_REG_UNLOCK, 0);
		error = devfs_set_cdevpriv(priv, netmap_dtor);

		if (error != 0) {
			/* could not assign the private storage for the
			 * thread, call the destructor explicitly.
			 */
			netmap_dtor(priv);
			break;
		}

		/* return the offset of the netmap_if object */
		nmr->nr_rx_rings = na->num_rx_rings;
		nmr->nr_tx_rings = na->num_tx_rings;
		nmr->nr_rx_slots = na->num_rx_desc;
		nmr->nr_tx_slots = na->num_tx_desc;
		nmr->nr_memsize = nm_mem->nm_totalsize;
		nmr->nr_offset = netmap_if_offset(nifp);
		break;

	case NIOCUNREGIF:
		if (priv == NULL) {
			error = ENXIO;
			break;
		}

		/* the interface is unregistered inside the
		   destructor of the private data. */
		devfs_clear_cdevpriv();
		break;

	case NIOCTXSYNC:
        case NIOCRXSYNC:
		if (priv == NULL) {
			error = ENXIO;
			break;
		}
		ifp = priv->np_ifp;	/* we have a reference */
		na = NA(ifp); /* retrieve netmap adapter */
		if (priv->np_qfirst == NETMAP_SW_RING) { /* host rings */
			if (cmd == NIOCTXSYNC)
				netmap_sync_to_host(na);
			else
				netmap_sync_from_host(na, NULL, NULL);
			break;
		}
		/* find the last ring to scan */
		lim = priv->np_qlast;
		if (lim == NETMAP_HW_RING)
			lim = (cmd == NIOCTXSYNC) ?
			    na->num_tx_rings : na->num_rx_rings;

		for (i = priv->np_qfirst; i < lim; i++) {
			if (cmd == NIOCTXSYNC) {
				struct netmap_kring *kring = &na->tx_rings[i];
				if (netmap_verbose & NM_VERB_TXSYNC)
					D("pre txsync ring %d cur %d hwcur %d",
					    i, kring->ring->cur,
					    kring->nr_hwcur);
				na->nm_txsync(ifp, i, 1 /* do lock */);
				if (netmap_verbose & NM_VERB_TXSYNC)
					D("post txsync ring %d cur %d hwcur %d",
					    i, kring->ring->cur,
					    kring->nr_hwcur);
			} else {
				na->nm_rxsync(ifp, i, 1 /* do lock */);
				microtime(&na->rx_rings[i].ring->ts);
			}
		}

		break;

#ifdef __FreeBSD__
	case BIOCIMMEDIATE:
	case BIOCGHDRCMPLT:
	case BIOCSHDRCMPLT:
	case BIOCSSEESENT:
		D("ignore BIOCIMMEDIATE/BIOCSHDRCMPLT/BIOCSHDRCMPLT/BIOCSSEESENT");
		break;

	default:	/* allow device-specific ioctls */
	    {
		struct socket so;
		bzero(&so, sizeof(so));
		error = get_ifp(nmr->nr_name, &ifp); /* keep reference */
		if (error)
			break;
		so.so_vnet = ifp->if_vnet;
		// so->so_proto not null.
		error = ifioctl(&so, cmd, data, td);
		nm_if_rele(ifp);
		break;
	    }

#else /* linux */
	default:
		error = EOPNOTSUPP;
#endif /* linux */
	}

	CURVNET_RESTORE();
	return (error);
}


/*
 * select(2) and poll(2) handlers for the "netmap" device.
 *
 * Can be called for one or more queues.
 * Return true the event mask corresponding to ready events.
 * If there are no ready events, do a selrecord on either individual
 * selfd or on the global one.
 * Device-dependent parts (locking and sync of tx/rx rings)
 * are done through callbacks.
 *
 * On linux, arguments are really pwait, the poll table, and 'td' is struct file *
 * The first one is remapped to pwait as selrecord() uses the name as an
 * hidden argument.
 */
static int
netmap_poll(struct cdev *dev, int events, struct thread *td)
{
	struct netmap_priv_d *priv = NULL;
	struct netmap_adapter *na;
	struct ifnet *ifp;
	struct netmap_kring *kring;
	u_int core_lock, i, check_all, want_tx, want_rx, revents = 0;
	u_int lim_tx, lim_rx;
	enum {NO_CL, NEED_CL, LOCKED_CL }; /* see below */
	void *pwait = dev;	/* linux compatibility */

	(void)pwait;

	if (devfs_get_cdevpriv((void **)&priv) != 0 || priv == NULL)
		return POLLERR;

	ifp = priv->np_ifp;
	// XXX check for deleting() ?
	if ( (ifp->if_capenable & IFCAP_NETMAP) == 0)
		return POLLERR;

	if (netmap_verbose & 0x8000)
		D("device %s events 0x%x", ifp->if_xname, events);
	want_tx = events & (POLLOUT | POLLWRNORM);
	want_rx = events & (POLLIN | POLLRDNORM);

	na = NA(ifp); /* retrieve netmap adapter */

	lim_tx = na->num_tx_rings;
	lim_rx = na->num_rx_rings;
	/* how many queues we are scanning */
	if (priv->np_qfirst == NETMAP_SW_RING) {
		if (priv->np_txpoll || want_tx) {
			/* push any packets up, then we are always ready */
			kring = &na->tx_rings[lim_tx];
			netmap_sync_to_host(na);
			revents |= want_tx;
		}
		if (want_rx) {
			kring = &na->rx_rings[lim_rx];
			if (kring->ring->avail == 0)
				netmap_sync_from_host(na, td, dev);
			if (kring->ring->avail > 0) {
				revents |= want_rx;
			}
		}
		return (revents);
	}

	/*
	 * check_all is set if the card has more than one queue and
	 * the client is polling all of them. If true, we sleep on
	 * the "global" selfd, otherwise we sleep on individual selfd
	 * (we can only sleep on one of them per direction).
	 * The interrupt routine in the driver should always wake on
	 * the individual selfd, and also on the global one if the card
	 * has more than one ring.
	 *
	 * If the card has only one lock, we just use that.
	 * If the card has separate ring locks, we just use those
	 * unless we are doing check_all, in which case the whole
	 * loop is wrapped by the global lock.
	 * We acquire locks only when necessary: if poll is called
	 * when buffers are available, we can just return without locks.
	 *
	 * rxsync() is only called if we run out of buffers on a POLLIN.
	 * txsync() is called if we run out of buffers on POLLOUT, or
	 * there are pending packets to send. The latter can be disabled
	 * passing NETMAP_NO_TX_POLL in the NIOCREG call.
	 */
	check_all = (priv->np_qlast == NETMAP_HW_RING) && (lim_tx > 1 || lim_rx > 1);

	/*
	 * core_lock indicates what to do with the core lock.
	 * The core lock is used when either the card has no individual
	 * locks, or it has individual locks but we are cheking all
	 * rings so we need the core lock to avoid missing wakeup events.
	 *
	 * It has three possible states:
	 * NO_CL	we don't need to use the core lock, e.g.
	 *		because we are protected by individual locks.
	 * NEED_CL	we need the core lock. In this case, when we
	 *		call the lock routine, move to LOCKED_CL
	 *		to remember to release the lock once done.
	 * LOCKED_CL	core lock is set, so we need to release it.
	 */
	core_lock = (check_all || !na->separate_locks) ? NEED_CL : NO_CL;
#ifdef NM_BRIDGE
	/* the bridge uses separate locks */
	if (na->nm_register == bdg_netmap_reg) {
		ND("not using core lock for %s", ifp->if_xname);
		core_lock = NO_CL;
	}
#endif /* NM_BRIDGE */
	if (priv->np_qlast != NETMAP_HW_RING) {
		lim_tx = lim_rx = priv->np_qlast;
	}

	/*
	 * We start with a lock free round which is good if we have
	 * data available. If this fails, then lock and call the sync
	 * routines.
	 */
	for (i = priv->np_qfirst; want_rx && i < lim_rx; i++) {
		kring = &na->rx_rings[i];
		if (kring->ring->avail > 0) {
			revents |= want_rx;
			want_rx = 0;	/* also breaks the loop */
		}
	}
	for (i = priv->np_qfirst; want_tx && i < lim_tx; i++) {
		kring = &na->tx_rings[i];
		if (kring->ring->avail > 0) {
			revents |= want_tx;
			want_tx = 0;	/* also breaks the loop */
		}
	}

	/*
	 * If we to push packets out (priv->np_txpoll) or want_tx is
	 * still set, we do need to run the txsync calls (on all rings,
	 * to avoid that the tx rings stall).
	 */
	if (priv->np_txpoll || want_tx) {
		for (i = priv->np_qfirst; i < lim_tx; i++) {
			kring = &na->tx_rings[i];
			/*
			 * Skip the current ring if want_tx == 0
			 * (we have already done a successful sync on
			 * a previous ring) AND kring->cur == kring->hwcur
			 * (there are no pending transmissions for this ring).
			 */
			if (!want_tx && kring->ring->cur == kring->nr_hwcur)
				continue;
			if (core_lock == NEED_CL) {
				na->nm_lock(ifp, NETMAP_CORE_LOCK, 0);
				core_lock = LOCKED_CL;
			}
			if (na->separate_locks)
				na->nm_lock(ifp, NETMAP_TX_LOCK, i);
			if (netmap_verbose & NM_VERB_TXSYNC)
				D("send %d on %s %d",
					kring->ring->cur,
					ifp->if_xname, i);
			if (na->nm_txsync(ifp, i, 0 /* no lock */))
				revents |= POLLERR;

			/* Check avail/call selrecord only if called with POLLOUT */
			if (want_tx) {
				if (kring->ring->avail > 0) {
					/* stop at the first ring. We don't risk
					 * starvation.
					 */
					revents |= want_tx;
					want_tx = 0;
				} else if (!check_all)
					selrecord(td, &kring->si);
			}
			if (na->separate_locks)
				na->nm_lock(ifp, NETMAP_TX_UNLOCK, i);
		}
	}

	/*
	 * now if want_rx is still set we need to lock and rxsync.
	 * Do it on all rings because otherwise we starve.
	 */
	if (want_rx) {
		for (i = priv->np_qfirst; i < lim_rx; i++) {
			kring = &na->rx_rings[i];
			if (core_lock == NEED_CL) {
				na->nm_lock(ifp, NETMAP_CORE_LOCK, 0);
				core_lock = LOCKED_CL;
			}
			if (na->separate_locks)
				na->nm_lock(ifp, NETMAP_RX_LOCK, i);

			if (na->nm_rxsync(ifp, i, 0 /* no lock */))
				revents |= POLLERR;
			if (netmap_no_timestamp == 0 ||
					kring->ring->flags & NR_TIMESTAMP) {
				microtime(&kring->ring->ts);
			}

			if (kring->ring->avail > 0)
				revents |= want_rx;
			else if (!check_all)
				selrecord(td, &kring->si);
			if (na->separate_locks)
				na->nm_lock(ifp, NETMAP_RX_UNLOCK, i);
		}
	}
	if (check_all && revents == 0) { /* signal on the global queue */
		if (want_tx)
			selrecord(td, &na->tx_si);
		if (want_rx)
			selrecord(td, &na->rx_si);
	}
	if (core_lock == LOCKED_CL)
		na->nm_lock(ifp, NETMAP_CORE_UNLOCK, 0);

	return (revents);
}

/*------- driver support routines ------*/

/*
 * default lock wrapper.
 */
static void
netmap_lock_wrapper(struct ifnet *dev, int what, u_int queueid)
{
	struct netmap_adapter *na = NA(dev);

	switch (what) {
#ifdef linux	/* some system do not need lock on register */
	case NETMAP_REG_LOCK:
	case NETMAP_REG_UNLOCK:
		break;
#endif /* linux */

	case NETMAP_CORE_LOCK:
		mtx_lock(&na->core_lock);
		break;

	case NETMAP_CORE_UNLOCK:
		mtx_unlock(&na->core_lock);
		break;

	case NETMAP_TX_LOCK:
		mtx_lock(&na->tx_rings[queueid].q_lock);
		break;

	case NETMAP_TX_UNLOCK:
		mtx_unlock(&na->tx_rings[queueid].q_lock);
		break;

	case NETMAP_RX_LOCK:
		mtx_lock(&na->rx_rings[queueid].q_lock);
		break;

	case NETMAP_RX_UNLOCK:
		mtx_unlock(&na->rx_rings[queueid].q_lock);
		break;
	}
}


/*
 * Initialize a ``netmap_adapter`` object created by driver on attach.
 * We allocate a block of memory with room for a struct netmap_adapter
 * plus two sets of N+2 struct netmap_kring (where N is the number
 * of hardware rings):
 * krings	0..N-1	are for the hardware queues.
 * kring	N	is for the host stack queue
 * kring	N+1	is only used for the selinfo for all queues.
 * Return 0 on success, ENOMEM otherwise.
 *
 * By default the receive and transmit adapter ring counts are both initialized
 * to num_queues.  na->num_tx_rings can be set for cards with different tx/rx
 * setups.
 */
int
netmap_attach(struct netmap_adapter *na, int num_queues)
{
	int n, size;
	void *buf;
	struct ifnet *ifp = na->ifp;

	if (ifp == NULL) {
		D("ifp not set, giving up");
		return EINVAL;
	}
	/* clear other fields ? */
	na->refcount = 0;
	if (na->num_tx_rings == 0)
		na->num_tx_rings = num_queues;
	na->num_rx_rings = num_queues;
	/* on each direction we have N+1 resources
	 * 0..n-1	are the hardware rings
	 * n		is the ring attached to the stack.
	 */
	n = na->num_rx_rings + na->num_tx_rings + 2;
	size = sizeof(*na) + n * sizeof(struct netmap_kring);

	buf = malloc(size, M_DEVBUF, M_NOWAIT | M_ZERO);
	if (buf) {
		WNA(ifp) = buf;
		na->tx_rings = (void *)((char *)buf + sizeof(*na));
		na->rx_rings = na->tx_rings + na->num_tx_rings + 1;
		bcopy(na, buf, sizeof(*na));
		ifp->if_capabilities |= IFCAP_NETMAP;

		na = buf;
		/* Core lock initialized here.  Others are initialized after
		 * netmap_if_new.
		 */
		mtx_init(&na->core_lock, "netmap core lock", MTX_NETWORK_LOCK,
		    MTX_DEF);
		if (na->nm_lock == NULL) {
			ND("using default locks for %s", ifp->if_xname);
			na->nm_lock = netmap_lock_wrapper;
		}
	}
#ifdef linux
	if (ifp->netdev_ops) {
		D("netdev_ops %p", ifp->netdev_ops);
		/* prepare a clone of the netdev ops */
		na->nm_ndo = *ifp->netdev_ops;
	}
	na->nm_ndo.ndo_start_xmit = linux_netmap_start;
#endif
	D("%s for %s", buf ? "ok" : "failed", ifp->if_xname);

	return (buf ? 0 : ENOMEM);
}


/*
 * Free the allocated memory linked to the given ``netmap_adapter``
 * object.
 */
void
netmap_detach(struct ifnet *ifp)
{
	struct netmap_adapter *na = NA(ifp);

	if (!na)
		return;

	mtx_destroy(&na->core_lock);

	bzero(na, sizeof(*na));
	WNA(ifp) = NULL;
	free(na, M_DEVBUF);
}


/*
 * Intercept packets from the network stack and pass them
 * to netmap as incoming packets on the 'software' ring.
 * We are not locked when called.
 */
int
netmap_start(struct ifnet *ifp, struct mbuf *m)
{
	struct netmap_adapter *na = NA(ifp);
	struct netmap_kring *kring = &na->rx_rings[na->num_rx_rings];
	u_int i, len = MBUF_LEN(m);
	u_int error = EBUSY, lim = kring->nkr_num_slots - 1;
	struct netmap_slot *slot;

	if (netmap_verbose & NM_VERB_HOST)
		D("%s packet %d len %d from the stack", ifp->if_xname,
			kring->nr_hwcur + kring->nr_hwavail, len);
	na->nm_lock(ifp, NETMAP_CORE_LOCK, 0);
	if (kring->nr_hwavail >= lim) {
		if (netmap_verbose)
			D("stack ring %s full\n", ifp->if_xname);
		goto done;	/* no space */
	}
	if (len > NETMAP_BUF_SIZE) {
		D("drop packet size %d > %d", len, NETMAP_BUF_SIZE);
		goto done;	/* too long for us */
	}

	/* compute the insert position */
	i = kring->nr_hwcur + kring->nr_hwavail;
	if (i > lim)
		i -= lim + 1;
	slot = &kring->ring->slot[i];
	m_copydata(m, 0, len, NMB(slot));
	slot->len = len;
	kring->nr_hwavail++;
	if (netmap_verbose  & NM_VERB_HOST)
		D("wake up host ring %s %d", na->ifp->if_xname, na->num_rx_rings);
	selwakeuppri(&kring->si, PI_NET);
	error = 0;
done:
	na->nm_lock(ifp, NETMAP_CORE_UNLOCK, 0);

	/* release the mbuf in either cases of success or failure. As an
	 * alternative, put the mbuf in a free list and free the list
	 * only when really necessary.
	 */
	m_freem(m);

	return (error);
}


/*
 * netmap_reset() is called by the driver routines when reinitializing
 * a ring. The driver is in charge of locking to protect the kring.
 * If netmap mode is not set just return NULL.
 */
struct netmap_slot *
netmap_reset(struct netmap_adapter *na, enum txrx tx, int n,
	u_int new_cur)
{
	struct netmap_kring *kring;
	int new_hwofs, lim;

	if (na == NULL)
		return NULL;	/* no netmap support here */
	if (!(na->ifp->if_capenable & IFCAP_NETMAP))
		return NULL;	/* nothing to reinitialize */

	if (tx == NR_TX) {
		kring = na->tx_rings + n;
		new_hwofs = kring->nr_hwcur - new_cur;
	} else {
		kring = na->rx_rings + n;
		new_hwofs = kring->nr_hwcur + kring->nr_hwavail - new_cur;
	}
	lim = kring->nkr_num_slots - 1;
	if (new_hwofs > lim)
		new_hwofs -= lim + 1;

	/* Alwayws set the new offset value and realign the ring. */
	kring->nkr_hwofs = new_hwofs;
	if (tx == NR_TX)
		kring->nr_hwavail = kring->nkr_num_slots - 1;
	D("new hwofs %d on %s %s[%d]",
			kring->nkr_hwofs, na->ifp->if_xname,
			tx == NR_TX ? "TX" : "RX", n);

#if 0 // def linux
	/* XXX check that the mappings are correct */
	/* need ring_nr, adapter->pdev, direction */
	buffer_info->dma = dma_map_single(&pdev->dev, addr, adapter->rx_buffer_len, DMA_FROM_DEVICE);
	if (dma_mapping_error(&adapter->pdev->dev, buffer_info->dma)) {
		D("error mapping rx netmap buffer %d", i);
		// XXX fix error handling
	}

#endif /* linux */
	/*
	 * Wakeup on the individual and global lock
	 * We do the wakeup here, but the ring is not yet reconfigured.
	 * However, we are under lock so there are no races.
	 */
	selwakeuppri(&kring->si, PI_NET);
	selwakeuppri(tx == NR_TX ? &na->tx_si : &na->rx_si, PI_NET);
	return kring->ring->slot;
}


/*
 * Default functions to handle rx/tx interrupts
 * we have 4 cases:
 * 1 ring, single lock:
 *	lock(core); wake(i=0); unlock(core)
 * N rings, single lock:
 *	lock(core); wake(i); wake(N+1) unlock(core)
 * 1 ring, separate locks: (i=0)
 *	lock(i); wake(i); unlock(i)
 * N rings, separate locks:
 *	lock(i); wake(i); unlock(i); lock(core) wake(N+1) unlock(core)
 * work_done is non-null on the RX path.
 */
int
netmap_rx_irq(struct ifnet *ifp, int q, int *work_done)
{
	struct netmap_adapter *na;
	struct netmap_kring *r;
	NM_SELINFO_T *main_wq;

	if (!(ifp->if_capenable & IFCAP_NETMAP))
		return 0;
	na = NA(ifp);
	if (work_done) { /* RX path */
		r = na->rx_rings + q;
		r->nr_kflags |= NKR_PENDINTR;
		main_wq = (na->num_rx_rings > 1) ? &na->rx_si : NULL;
	} else { /* tx path */
		r = na->tx_rings + q;
		main_wq = (na->num_tx_rings > 1) ? &na->tx_si : NULL;
		work_done = &q; /* dummy */
	}
	if (na->separate_locks) {
		mtx_lock(&r->q_lock);
		selwakeuppri(&r->si, PI_NET);
		mtx_unlock(&r->q_lock);
		if (main_wq) {
			mtx_lock(&na->core_lock);
			selwakeuppri(main_wq, PI_NET);
			mtx_unlock(&na->core_lock);
		}
	} else {
		mtx_lock(&na->core_lock);
		selwakeuppri(&r->si, PI_NET);
		if (main_wq)
			selwakeuppri(main_wq, PI_NET);
		mtx_unlock(&na->core_lock);
	}
	*work_done = 1; /* do not fire napi again */
	return 1;
}


#ifdef linux	/* linux-specific routines */

/*
 * Remap linux arguments into the FreeBSD call.
 * - pwait is the poll table, passed as 'dev';
 *   If pwait == NULL someone else already woke up before. We can report
 *   events but they are filtered upstream.
 *   If pwait != NULL, then pwait->key contains the list of events.
 * - events is computed from pwait as above.
 * - file is passed as 'td';
 */
static u_int
linux_netmap_poll(struct file * file, struct poll_table_struct *pwait)
{
#if LINUX_VERSION_CODE < KERNEL_VERSION(3,4,0)
	int events = pwait ? pwait->key : POLLIN | POLLOUT;
#else /* in 3.4.0 field 'key' was renamed to '_key' */
	int events = pwait ? pwait->_key : POLLIN | POLLOUT;
#endif
	return netmap_poll((void *)pwait, events, (void *)file);
}

static int
linux_netmap_mmap(struct file *f, struct vm_area_struct *vma)
{
	int lut_skip, i, j;
	int user_skip = 0;
	struct lut_entry *l_entry;
	const struct netmap_obj_pool *p[] = {
		nm_mem->nm_if_pool,
		nm_mem->nm_ring_pool,
		nm_mem->nm_buf_pool };
	/*
	 * vma->vm_start: start of mapping user address space
	 * vma->vm_end: end of the mapping user address space
	 */

	(void)f;	/* UNUSED */
	// XXX security checks

	for (i = 0; i < 3; i++) {  /* loop through obj_pools */
		/*
		 * In each pool memory is allocated in clusters
		 * of size _clustsize , each containing clustentries
		 * entries. For each object k we already store the
		 * vtophys malling in lut[k] so we use that, scanning
		 * the lut[] array in steps of clustentries,
		 * and we map each cluster (not individual pages,
		 * it would be overkill).
		 */
		for (lut_skip = 0, j = 0; j < p[i]->_numclusters; j++) {
			l_entry = &p[i]->lut[lut_skip];
			if (remap_pfn_range(vma, vma->vm_start + user_skip,
					l_entry->paddr >> PAGE_SHIFT, p[i]->_clustsize,
					vma->vm_page_prot))
				return -EAGAIN; // XXX check return value
			lut_skip += p[i]->clustentries;
			user_skip += p[i]->_clustsize;
		}
	}

	return 0;
}

static netdev_tx_t
linux_netmap_start(struct sk_buff *skb, struct net_device *dev)
{
	netmap_start(dev, skb);
	return (NETDEV_TX_OK);
}


#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,37)	// XXX was 38
#define LIN_IOCTL_NAME	.ioctl
int
linux_netmap_ioctl(struct inode *inode, struct file *file, u_int cmd, u_long data /* arg */)
#else
#define LIN_IOCTL_NAME	.unlocked_ioctl
long
linux_netmap_ioctl(struct file *file, u_int cmd, u_long data /* arg */)
#endif
{
	int ret;
	struct nmreq nmr;
	bzero(&nmr, sizeof(nmr));

	if (data && copy_from_user(&nmr, (void *)data, sizeof(nmr) ) != 0)
		return -EFAULT;
	ret = netmap_ioctl(NULL, cmd, (caddr_t)&nmr, 0, (void *)file);
	if (data && copy_to_user((void*)data, &nmr, sizeof(nmr) ) != 0)
		return -EFAULT;
	return -ret;
}


static int
netmap_release(struct inode *inode, struct file *file)
{
	(void)inode;	/* UNUSED */
	if (file->private_data)
		netmap_dtor(file->private_data);
	return (0);
}


static struct file_operations netmap_fops = {
    .mmap = linux_netmap_mmap,
    LIN_IOCTL_NAME = linux_netmap_ioctl,
    .poll = linux_netmap_poll,
    .release = netmap_release,
};

static struct miscdevice netmap_cdevsw = {	/* same name as FreeBSD */
	MISC_DYNAMIC_MINOR,
	"netmap",
	&netmap_fops,
};

static int netmap_init(void);
static void netmap_fini(void);

/* Errors have negative values on linux */
static int linux_netmap_init(void)
{
	return -netmap_init();
}

module_init(linux_netmap_init);
module_exit(netmap_fini);
/* export certain symbols to other modules */
EXPORT_SYMBOL(netmap_attach);		// driver attach routines
EXPORT_SYMBOL(netmap_detach);		// driver detach routines
EXPORT_SYMBOL(netmap_ring_reinit);	// ring init on error
EXPORT_SYMBOL(netmap_buffer_lut);
EXPORT_SYMBOL(netmap_total_buffers);	// index check
EXPORT_SYMBOL(netmap_buffer_base);
EXPORT_SYMBOL(netmap_reset);		// ring init routines
EXPORT_SYMBOL(netmap_buf_size);
EXPORT_SYMBOL(netmap_rx_irq);		// default irq handler
EXPORT_SYMBOL(netmap_no_pendintr);	// XXX mitigation - should go away


MODULE_AUTHOR("http://info.iet.unipi.it/~luigi/netmap/");
MODULE_DESCRIPTION("The netmap packet I/O framework");
MODULE_LICENSE("Dual BSD/GPL"); /* the code here is all BSD. */

#else /* __FreeBSD__ */

static struct cdevsw netmap_cdevsw = {
	.d_version = D_VERSION,
	.d_name = "netmap",
	.d_mmap = netmap_mmap,
	.d_ioctl = netmap_ioctl,
	.d_poll = netmap_poll,
};
#endif /* __FreeBSD__ */

#ifdef NM_BRIDGE
/*
 *---- support for virtual bridge -----
 */

/* ----- FreeBSD if_bridge hash function ------- */

/*
 * The following hash function is adapted from "Hash Functions" by Bob Jenkins
 * ("Algorithm Alley", Dr. Dobbs Journal, September 1997).
 *
 * http://www.burtleburtle.net/bob/hash/spooky.html
 */
#define mix(a, b, c)                                                    \
do {                                                                    \
        a -= b; a -= c; a ^= (c >> 13);                                 \
        b -= c; b -= a; b ^= (a << 8);                                  \
        c -= a; c -= b; c ^= (b >> 13);                                 \
        a -= b; a -= c; a ^= (c >> 12);                                 \
        b -= c; b -= a; b ^= (a << 16);                                 \
        c -= a; c -= b; c ^= (b >> 5);                                  \
        a -= b; a -= c; a ^= (c >> 3);                                  \
        b -= c; b -= a; b ^= (a << 10);                                 \
        c -= a; c -= b; c ^= (b >> 15);                                 \
} while (/*CONSTCOND*/0)

static __inline uint32_t
nm_bridge_rthash(const uint8_t *addr)
{
        uint32_t a = 0x9e3779b9, b = 0x9e3779b9, c = 0; // hask key

        b += addr[5] << 8;
        b += addr[4];
        a += addr[3] << 24;
        a += addr[2] << 16;
        a += addr[1] << 8;
        a += addr[0];

        mix(a, b, c);
#define BRIDGE_RTHASH_MASK	(NM_BDG_HASH-1)
        return (c & BRIDGE_RTHASH_MASK);
}

#undef mix


static int
bdg_netmap_reg(struct ifnet *ifp, int onoff)
{
	int i, err = 0;
	struct nm_bridge *b = ifp->if_bridge;

	BDG_LOCK(b);
	if (onoff) {
		/* the interface must be already in the list.
		 * only need to mark the port as active
		 */
		ND("should attach %s to the bridge", ifp->if_xname);
		for (i=0; i < NM_BDG_MAXPORTS; i++)
			if (b->bdg_ports[i] == ifp)
				break;
		if (i == NM_BDG_MAXPORTS) {
			D("no more ports available");
			err = EINVAL;
			goto done;
		}
		ND("setting %s in netmap mode", ifp->if_xname);
		ifp->if_capenable |= IFCAP_NETMAP;
		NA(ifp)->bdg_port = i;
		b->act_ports |= (1<<i);
		b->bdg_ports[i] = ifp;
	} else {
		/* should be in the list, too -- remove from the mask */
		ND("removing %s from netmap mode", ifp->if_xname);
		ifp->if_capenable &= ~IFCAP_NETMAP;
		i = NA(ifp)->bdg_port;
		b->act_ports &= ~(1<<i);
	}
done:
	BDG_UNLOCK(b);
	return err;
}


static int
nm_bdg_flush(struct nm_bdg_fwd *ft, int n, struct ifnet *ifp)
{
	int i, ifn;
	uint64_t all_dst, dst;
	uint32_t sh, dh;
	uint64_t mysrc = 1 << NA(ifp)->bdg_port;
	uint64_t smac, dmac;
	struct netmap_slot *slot;
	struct nm_bridge *b = ifp->if_bridge;

	ND("prepare to send %d packets, act_ports 0x%x", n, b->act_ports);
	/* only consider valid destinations */
	all_dst = (b->act_ports & ~mysrc);
	/* first pass: hash and find destinations */
	for (i = 0; likely(i < n); i++) {
		uint8_t *buf = ft[i].buf;
		dmac = le64toh(*(uint64_t *)(buf)) & 0xffffffffffff;
		smac = le64toh(*(uint64_t *)(buf + 4));
		smac >>= 16;
		if (unlikely(netmap_verbose)) {
		    uint8_t *s = buf+6, *d = buf;
		    D("%d len %4d %02x:%02x:%02x:%02x:%02x:%02x -> %02x:%02x:%02x:%02x:%02x:%02x",
			i,
			ft[i].len,
			s[0], s[1], s[2], s[3], s[4], s[5],
			d[0], d[1], d[2], d[3], d[4], d[5]);
		}
		/*
		 * The hash is somewhat expensive, there might be some
		 * worthwhile optimizations here.
		 */
		if ((buf[6] & 1) == 0) { /* valid src */
		    	uint8_t *s = buf+6;
			sh = nm_bridge_rthash(buf+6); // XXX hash of source
			/* update source port forwarding entry */
			b->ht[sh].mac = smac;	/* XXX expire ? */
			b->ht[sh].ports = mysrc;
			if (netmap_verbose)
			    D("src %02x:%02x:%02x:%02x:%02x:%02x on port %d",
				s[0], s[1], s[2], s[3], s[4], s[5], NA(ifp)->bdg_port);
		}
		dst = 0;
		if ( (buf[0] & 1) == 0) { /* unicast */
		    	uint8_t *d = buf;
			dh = nm_bridge_rthash(buf); // XXX hash of dst
			if (b->ht[dh].mac == dmac) {	/* found dst */
				dst = b->ht[dh].ports;
				if (netmap_verbose)
				    D("dst %02x:%02x:%02x:%02x:%02x:%02x to port %x",
					d[0], d[1], d[2], d[3], d[4], d[5], (uint32_t)(dst >> 16));
			}
		}
		if (dst == 0)
			dst = all_dst;
		dst &= all_dst; /* only consider valid ports */
		if (unlikely(netmap_verbose))
			D("pkt goes to ports 0x%x", (uint32_t)dst);
		ft[i].dst = dst;
	}

	/* second pass, scan interfaces and forward */
	all_dst = (b->act_ports & ~mysrc);
	for (ifn = 0; all_dst; ifn++) {
		struct ifnet *dst_ifp = b->bdg_ports[ifn];
		struct netmap_adapter *na;
		struct netmap_kring *kring;
		struct netmap_ring *ring;
		int j, lim, sent, locked;

		if (!dst_ifp)
			continue;
		ND("scan port %d %s", ifn, dst_ifp->if_xname);
		dst = 1 << ifn;
		if ((dst & all_dst) == 0)	/* skip if not set */
			continue;
		all_dst &= ~dst;	/* clear current node */
		na = NA(dst_ifp);

		ring = NULL;
		kring = NULL;
		lim = sent = locked = 0;
		/* inside, scan slots */
		for (i = 0; likely(i < n); i++) {
			if ((ft[i].dst & dst) == 0)
				continue;	/* not here */
			if (!locked) {
				kring = &na->rx_rings[0];
				ring = kring->ring;
				lim = kring->nkr_num_slots - 1;
				na->nm_lock(dst_ifp, NETMAP_RX_LOCK, 0);
				locked = 1;
			}
			if (unlikely(kring->nr_hwavail >= lim)) {
				if (netmap_verbose)
					D("rx ring full on %s", ifp->if_xname);
				break;
			}
			j = kring->nr_hwcur + kring->nr_hwavail;
			if (j > lim)
				j -= kring->nkr_num_slots;
			slot = &ring->slot[j];
			ND("send %d %d bytes at %s:%d", i, ft[i].len, dst_ifp->if_xname, j);
			pkt_copy(ft[i].buf, NMB(slot), ft[i].len);
			slot->len = ft[i].len;
			kring->nr_hwavail++;
			sent++;
		}
		if (locked) {
			ND("sent %d on %s", sent, dst_ifp->if_xname);
			if (sent)
				selwakeuppri(&kring->si, PI_NET);
			na->nm_lock(dst_ifp, NETMAP_RX_UNLOCK, 0);
		}
	}
	return 0;
}

/*
 * main dispatch routine
 */
static int
bdg_netmap_txsync(struct ifnet *ifp, u_int ring_nr, int do_lock)
{
	struct netmap_adapter *na = NA(ifp);
	struct netmap_kring *kring = &na->tx_rings[ring_nr];
	struct netmap_ring *ring = kring->ring;
	int i, j, k, lim = kring->nkr_num_slots - 1;
	struct nm_bdg_fwd *ft = (struct nm_bdg_fwd *)(ifp + 1);
	int ft_i;	/* position in the forwarding table */

	k = ring->cur;
	if (k > lim)
		return netmap_ring_reinit(kring);
	if (do_lock)
		na->nm_lock(ifp, NETMAP_TX_LOCK, ring_nr);

	if (netmap_bridge <= 0) { /* testing only */
		j = k; // used all
		goto done;
	}
	if (netmap_bridge > NM_BDG_BATCH)
		netmap_bridge = NM_BDG_BATCH;

	ft_i = 0;	/* start from 0 */
	for (j = kring->nr_hwcur; likely(j != k); j = unlikely(j == lim) ? 0 : j+1) {
		struct netmap_slot *slot = &ring->slot[j];
		int len = ft[ft_i].len = slot->len;
		char *buf = ft[ft_i].buf = NMB(slot);

		prefetch(buf);
		if (unlikely(len < 14))
			continue;
		if (unlikely(++ft_i == netmap_bridge))
			ft_i = nm_bdg_flush(ft, ft_i, ifp);
	}
	if (ft_i)
		ft_i = nm_bdg_flush(ft, ft_i, ifp);
	/* count how many packets we sent */
	i = k - j;
	if (i < 0)
		i += kring->nkr_num_slots;
	kring->nr_hwavail = kring->nkr_num_slots - 1 - i;
	if (j != k)
		D("early break at %d/ %d, avail %d", j, k, kring->nr_hwavail);

done:
	kring->nr_hwcur = j;
	ring->avail = kring->nr_hwavail;
	if (do_lock)
		na->nm_lock(ifp, NETMAP_TX_UNLOCK, ring_nr);

	if (netmap_verbose)
		D("%s ring %d lock %d", ifp->if_xname, ring_nr, do_lock);
	return 0;
}

static int
bdg_netmap_rxsync(struct ifnet *ifp, u_int ring_nr, int do_lock)
{
	struct netmap_adapter *na = NA(ifp);
	struct netmap_kring *kring = &na->rx_rings[ring_nr];
	struct netmap_ring *ring = kring->ring;
	u_int j, n, lim = kring->nkr_num_slots - 1;
	u_int k = ring->cur, resvd = ring->reserved;

	ND("%s ring %d lock %d avail %d",
		ifp->if_xname, ring_nr, do_lock, kring->nr_hwavail);

	if (k > lim)
		return netmap_ring_reinit(kring);
	if (do_lock)
		na->nm_lock(ifp, NETMAP_RX_LOCK, ring_nr);

	/* skip past packets that userspace has released */
	j = kring->nr_hwcur;    /* netmap ring index */
	if (resvd > 0) {
		if (resvd + ring->avail >= lim + 1) {
			D("XXX invalid reserve/avail %d %d", resvd, ring->avail);
			ring->reserved = resvd = 0; // XXX panic...
		}
		k = (k >= resvd) ? k - resvd : k + lim + 1 - resvd;
	}

	if (j != k) { /* userspace has released some packets. */
		n = k - j;
		if (n < 0)
			n += kring->nkr_num_slots;
		ND("userspace releases %d packets", n);
                for (n = 0; likely(j != k); n++) {
                        struct netmap_slot *slot = &ring->slot[j];
                        void *addr = NMB(slot);

                        if (addr == netmap_buffer_base) { /* bad buf */
                                if (do_lock)
                                        na->nm_lock(ifp, NETMAP_RX_UNLOCK, ring_nr);
                                return netmap_ring_reinit(kring);
                        }
			/* decrease refcount for buffer */

			slot->flags &= ~NS_BUF_CHANGED;
                        j = unlikely(j == lim) ? 0 : j + 1;
                }
                kring->nr_hwavail -= n;
                kring->nr_hwcur = k;
        }
        /* tell userspace that there are new packets */
        ring->avail = kring->nr_hwavail - resvd;

	if (do_lock)
		na->nm_lock(ifp, NETMAP_RX_UNLOCK, ring_nr);
	return 0;
}

static void
bdg_netmap_attach(struct ifnet *ifp)
{
	struct netmap_adapter na;

	ND("attaching virtual bridge");
	bzero(&na, sizeof(na));

	na.ifp = ifp;
	na.separate_locks = 1;
	na.num_tx_desc = NM_BRIDGE_RINGSIZE;
	na.num_rx_desc = NM_BRIDGE_RINGSIZE;
	na.nm_txsync = bdg_netmap_txsync;
	na.nm_rxsync = bdg_netmap_rxsync;
	na.nm_register = bdg_netmap_reg;
	netmap_attach(&na, 1);
}

#endif /* NM_BRIDGE */

static struct cdev *netmap_dev; /* /dev/netmap character device. */


/*
 * Module loader.
 *
 * Create the /dev/netmap device and initialize all global
 * variables.
 *
 * Return 0 on success, errno on failure.
 */
static int
netmap_init(void)
{
	int error;

	error = netmap_memory_init();
	if (error != 0) {
		printf("netmap: unable to initialize the memory allocator.\n");
		return (error);
	}
	printf("netmap: loaded module with %d Mbytes\n",
		(int)(nm_mem->nm_totalsize >> 20));
	netmap_dev = make_dev(&netmap_cdevsw, 0, UID_ROOT, GID_WHEEL, 0660,
			      "netmap");

#ifdef NM_BRIDGE
	{
	int i;
	for (i = 0; i < NM_BRIDGES; i++)
		mtx_init(&nm_bridges[i].bdg_lock, "bdg lock", "bdg_lock", MTX_DEF);
	}
#endif
	return (error);
}


/*
 * Module unloader.
 *
 * Free all the memory, and destroy the ``/dev/netmap`` device.
 */
static void
netmap_fini(void)
{
	destroy_dev(netmap_dev);
	netmap_memory_fini();
	printf("netmap: unloaded module.\n");
}


#ifdef __FreeBSD__
/*
 * Kernel entry point.
 *
 * Initialize/finalize the module and return.
 *
 * Return 0 on success, errno on failure.
 */
static int
netmap_loader(__unused struct module *module, int event, __unused void *arg)
{
	int error = 0;

	switch (event) {
	case MOD_LOAD:
		error = netmap_init();
		break;

	case MOD_UNLOAD:
		netmap_fini();
		break;

	default:
		error = EOPNOTSUPP;
		break;
	}

	return (error);
}


DEV_MODULE(netmap, netmap_loader, NULL);
#endif /* __FreeBSD__ */