summaryrefslogtreecommitdiffstats
path: root/sys/netinet/in_mcast.c
blob: 1f658da6d838e4d8467c248aeab91d8d710213ef (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
2358
2359
2360
2361
2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
2381
2382
2383
2384
2385
2386
2387
2388
2389
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
2411
2412
2413
2414
2415
2416
2417
2418
2419
2420
2421
2422
2423
2424
2425
2426
2427
2428
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
2462
2463
2464
2465
2466
2467
2468
2469
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
2525
2526
2527
2528
2529
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
2584
2585
2586
2587
2588
2589
2590
2591
2592
2593
2594
2595
2596
2597
2598
2599
2600
2601
2602
2603
2604
2605
2606
2607
2608
2609
2610
2611
2612
2613
2614
2615
2616
2617
2618
2619
2620
2621
2622
2623
2624
2625
2626
2627
2628
2629
2630
2631
2632
2633
2634
2635
2636
2637
2638
2639
2640
2641
2642
2643
2644
2645
2646
2647
2648
2649
2650
2651
2652
2653
2654
2655
2656
2657
2658
2659
2660
2661
2662
2663
2664
2665
2666
2667
2668
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
2686
2687
2688
2689
2690
2691
2692
2693
2694
2695
2696
2697
2698
2699
2700
2701
2702
2703
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
2714
2715
2716
2717
2718
2719
2720
2721
2722
2723
2724
2725
2726
2727
2728
2729
2730
2731
2732
2733
2734
2735
2736
2737
2738
2739
2740
2741
2742
2743
2744
2745
2746
2747
2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
2758
2759
2760
2761
2762
2763
2764
2765
2766
2767
2768
2769
2770
2771
2772
2773
2774
2775
2776
2777
2778
2779
2780
2781
2782
2783
2784
2785
2786
2787
2788
2789
2790
2791
2792
2793
2794
2795
2796
2797
2798
2799
2800
2801
2802
2803
2804
2805
2806
2807
2808
2809
2810
2811
2812
2813
2814
2815
2816
2817
2818
2819
2820
2821
2822
2823
2824
2825
2826
2827
2828
2829
2830
2831
2832
2833
2834
2835
2836
2837
2838
2839
2840
2841
2842
2843
2844
2845
2846
2847
2848
2849
2850
2851
2852
2853
2854
2855
2856
2857
2858
2859
2860
2861
2862
2863
2864
2865
2866
2867
2868
2869
2870
2871
2872
2873
2874
2875
2876
2877
2878
2879
2880
2881
2882
2883
2884
2885
2886
2887
2888
2889
2890
2891
2892
2893
2894
2895
2896
2897
2898
2899
2900
2901
2902
2903
2904
2905
2906
2907
2908
2909
/*-
 * Copyright (c) 2007-2009 Bruce Simpson.
 * Copyright (c) 2005 Robert N. M. Watson.
 * 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.
 * 3. The name of the author may not be used to endorse or promote
 *    products derived from this software without specific prior written
 *    permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 */

/*
 * IPv4 multicast socket, group, and socket option processing module.
 */

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

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/kernel.h>
#include <sys/malloc.h>
#include <sys/mbuf.h>
#include <sys/protosw.h>
#include <sys/socket.h>
#include <sys/socketvar.h>
#include <sys/protosw.h>
#include <sys/sysctl.h>
#include <sys/ktr.h>
#include <sys/tree.h>

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

#include <netinet/in.h>
#include <netinet/in_systm.h>
#include <netinet/in_pcb.h>
#include <netinet/in_var.h>
#include <netinet/ip_var.h>
#include <netinet/igmp_var.h>

#ifndef KTR_IGMPV3
#define KTR_IGMPV3 KTR_INET
#endif

#ifndef __SOCKUNION_DECLARED
union sockunion {
	struct sockaddr_storage	ss;
	struct sockaddr		sa;
	struct sockaddr_dl	sdl;
	struct sockaddr_in	sin;
};
typedef union sockunion sockunion_t;
#define __SOCKUNION_DECLARED
#endif /* __SOCKUNION_DECLARED */

static MALLOC_DEFINE(M_INMFILTER, "in_mfilter",
    "IPv4 multicast PCB-layer source filter");
static MALLOC_DEFINE(M_IPMADDR, "in_multi", "IPv4 multicast group");
static MALLOC_DEFINE(M_IPMOPTS, "ip_moptions", "IPv4 multicast options");
static MALLOC_DEFINE(M_IPMSOURCE, "ip_msource",
    "IPv4 multicast IGMP-layer source filter");

/*
 * Locking:
 * - Lock order is: Giant, INP_WLOCK, IN_MULTI_LOCK, IGMP_LOCK, IF_ADDR_LOCK.
 * - The IF_ADDR_LOCK is implicitly taken by inm_lookup() earlier, however
 *   it can be taken by code in net/if.c also.
 * - ip_moptions and in_mfilter are covered by the INP_WLOCK.
 *
 * struct in_multi is covered by IN_MULTI_LOCK. There isn't strictly
 * any need for in_multi itself to be virtualized -- it is bound to an ifp
 * anyway no matter what happens.
 */
struct mtx in_multi_mtx;
MTX_SYSINIT(in_multi_mtx, &in_multi_mtx, "in_multi_mtx", MTX_DEF);

/*
 * Functions with non-static linkage defined in this file should be
 * declared in in_var.h:
 *  imo_multi_filter()
 *  in_addmulti()
 *  in_delmulti()
 *  in_joingroup()
 *  in_joingroup_locked()
 *  in_leavegroup()
 *  in_leavegroup_locked()
 * and ip_var.h:
 *  inp_freemoptions()
 *  inp_getmoptions()
 *  inp_setmoptions()
 *
 * XXX: Both carp and pf need to use the legacy (*,G) KPIs in_addmulti()
 * and in_delmulti().
 */
static void	imf_commit(struct in_mfilter *);
static int	imf_get_source(struct in_mfilter *imf,
		    const struct sockaddr_in *psin,
		    struct in_msource **);
static struct in_msource *
		imf_graft(struct in_mfilter *, const uint8_t,
		    const struct sockaddr_in *);
static void	imf_leave(struct in_mfilter *);
static int	imf_prune(struct in_mfilter *, const struct sockaddr_in *);
static void	imf_purge(struct in_mfilter *);
static void	imf_rollback(struct in_mfilter *);
static void	imf_reap(struct in_mfilter *);
static int	imo_grow(struct ip_moptions *);
static size_t	imo_match_group(const struct ip_moptions *,
		    const struct ifnet *, const struct sockaddr *);
static struct in_msource *
		imo_match_source(const struct ip_moptions *, const size_t,
		    const struct sockaddr *);
static void	ims_merge(struct ip_msource *ims,
		    const struct in_msource *lims, const int rollback);
static int	in_getmulti(struct ifnet *, const struct in_addr *,
		    struct in_multi **);
static int	inm_get_source(struct in_multi *inm, const in_addr_t haddr,
		    const int noalloc, struct ip_msource **pims);
static int	inm_is_ifp_detached(const struct in_multi *);
static int	inm_merge(struct in_multi *, /*const*/ struct in_mfilter *);
static void	inm_purge(struct in_multi *);
static void	inm_reap(struct in_multi *);
static struct ip_moptions *
		inp_findmoptions(struct inpcb *);
static int	inp_get_source_filters(struct inpcb *, struct sockopt *);
static int	inp_join_group(struct inpcb *, struct sockopt *);
static int	inp_leave_group(struct inpcb *, struct sockopt *);
static struct ifnet *
		inp_lookup_mcast_ifp(const struct inpcb *,
		    const struct sockaddr_in *, const struct in_addr);
static int	inp_block_unblock_source(struct inpcb *, struct sockopt *);
static int	inp_set_multicast_if(struct inpcb *, struct sockopt *);
static int	inp_set_source_filters(struct inpcb *, struct sockopt *);
static int	sysctl_ip_mcast_filters(SYSCTL_HANDLER_ARGS);

SYSCTL_NODE(_net_inet_ip, OID_AUTO, mcast, CTLFLAG_RW, 0, "IPv4 multicast");

static u_long in_mcast_maxgrpsrc = IP_MAX_GROUP_SRC_FILTER;
SYSCTL_ULONG(_net_inet_ip_mcast, OID_AUTO, maxgrpsrc,
    CTLFLAG_RW | CTLFLAG_TUN, &in_mcast_maxgrpsrc, 0,
    "Max source filters per group");
TUNABLE_ULONG("net.inet.ip.mcast.maxgrpsrc", &in_mcast_maxgrpsrc);

static u_long in_mcast_maxsocksrc = IP_MAX_SOCK_SRC_FILTER;
SYSCTL_ULONG(_net_inet_ip_mcast, OID_AUTO, maxsocksrc,
    CTLFLAG_RW | CTLFLAG_TUN, &in_mcast_maxsocksrc, 0,
    "Max source filters per socket");
TUNABLE_ULONG("net.inet.ip.mcast.maxsocksrc", &in_mcast_maxsocksrc);

int in_mcast_loop = IP_DEFAULT_MULTICAST_LOOP;
SYSCTL_INT(_net_inet_ip_mcast, OID_AUTO, loop, CTLFLAG_RW | CTLFLAG_TUN,
    &in_mcast_loop, 0, "Loopback multicast datagrams by default");
TUNABLE_INT("net.inet.ip.mcast.loop", &in_mcast_loop);

SYSCTL_NODE(_net_inet_ip_mcast, OID_AUTO, filters,
    CTLFLAG_RD | CTLFLAG_MPSAFE, sysctl_ip_mcast_filters,
    "Per-interface stack-wide source filters");

/*
 * Inline function which wraps assertions for a valid ifp.
 * The ifnet layer will set the ifma's ifp pointer to NULL if the ifp
 * is detached.
 */
static int __inline
inm_is_ifp_detached(const struct in_multi *inm)
{
	struct ifnet *ifp;

	KASSERT(inm->inm_ifma != NULL, ("%s: no ifma", __func__));
	ifp = inm->inm_ifma->ifma_ifp;
	if (ifp != NULL) {
		/*
		 * Sanity check that netinet's notion of ifp is the
		 * same as net's.
		 */
		KASSERT(inm->inm_ifp == ifp, ("%s: bad ifp", __func__));
	}

	return (ifp == NULL);
}

/*
 * Initialize an in_mfilter structure to a known state at t0, t1
 * with an empty source filter list.
 */
static __inline void
imf_init(struct in_mfilter *imf, const int st0, const int st1)
{
	memset(imf, 0, sizeof(struct in_mfilter));
	RB_INIT(&imf->imf_sources);
	imf->imf_st[0] = st0;
	imf->imf_st[1] = st1;
}

/*
 * Resize the ip_moptions vector to the next power-of-two minus 1.
 * May be called with locks held; do not sleep.
 */
static int
imo_grow(struct ip_moptions *imo)
{
	struct in_multi		**nmships;
	struct in_multi		**omships;
	struct in_mfilter	 *nmfilters;
	struct in_mfilter	 *omfilters;
	size_t			  idx;
	size_t			  newmax;
	size_t			  oldmax;

	nmships = NULL;
	nmfilters = NULL;
	omships = imo->imo_membership;
	omfilters = imo->imo_mfilters;
	oldmax = imo->imo_max_memberships;
	newmax = ((oldmax + 1) * 2) - 1;

	if (newmax <= IP_MAX_MEMBERSHIPS) {
		nmships = (struct in_multi **)realloc(omships,
		    sizeof(struct in_multi *) * newmax, M_IPMOPTS, M_NOWAIT);
		nmfilters = (struct in_mfilter *)realloc(omfilters,
		    sizeof(struct in_mfilter) * newmax, M_INMFILTER, M_NOWAIT);
		if (nmships != NULL && nmfilters != NULL) {
			/* Initialize newly allocated source filter heads. */
			for (idx = oldmax; idx < newmax; idx++) {
				imf_init(&nmfilters[idx], MCAST_UNDEFINED,
				    MCAST_EXCLUDE);
			}
			imo->imo_max_memberships = newmax;
			imo->imo_membership = nmships;
			imo->imo_mfilters = nmfilters;
		}
	}

	if (nmships == NULL || nmfilters == NULL) {
		if (nmships != NULL)
			free(nmships, M_IPMOPTS);
		if (nmfilters != NULL)
			free(nmfilters, M_INMFILTER);
		return (ETOOMANYREFS);
	}

	return (0);
}

/*
 * Find an IPv4 multicast group entry for this ip_moptions instance
 * which matches the specified group, and optionally an interface.
 * Return its index into the array, or -1 if not found.
 */
static size_t
imo_match_group(const struct ip_moptions *imo, const struct ifnet *ifp,
    const struct sockaddr *group)
{
	const struct sockaddr_in *gsin;
	struct in_multi	**pinm;
	int		  idx;
	int		  nmships;

	gsin = (const struct sockaddr_in *)group;

	/* The imo_membership array may be lazy allocated. */
	if (imo->imo_membership == NULL || imo->imo_num_memberships == 0)
		return (-1);

	nmships = imo->imo_num_memberships;
	pinm = &imo->imo_membership[0];
	for (idx = 0; idx < nmships; idx++, pinm++) {
		if (*pinm == NULL)
			continue;
		if ((ifp == NULL || ((*pinm)->inm_ifp == ifp)) &&
		    in_hosteq((*pinm)->inm_addr, gsin->sin_addr)) {
			break;
		}
	}
	if (idx >= nmships)
		idx = -1;

	return (idx);
}

/*
 * Find an IPv4 multicast source entry for this imo which matches
 * the given group index for this socket, and source address.
 *
 * NOTE: This does not check if the entry is in-mode, merely if
 * it exists, which may not be the desired behaviour.
 */
static struct in_msource *
imo_match_source(const struct ip_moptions *imo, const size_t gidx,
    const struct sockaddr *src)
{
	struct ip_msource	 find;
	struct in_mfilter	*imf;
	struct ip_msource	*ims;
	const sockunion_t	*psa;

	KASSERT(src->sa_family == AF_INET, ("%s: !AF_INET", __func__));
	KASSERT(gidx != -1 && gidx < imo->imo_num_memberships,
	    ("%s: invalid index %d\n", __func__, (int)gidx));

	/* The imo_mfilters array may be lazy allocated. */
	if (imo->imo_mfilters == NULL)
		return (NULL);
	imf = &imo->imo_mfilters[gidx];

	/* Source trees are keyed in host byte order. */
	psa = (const sockunion_t *)src;
	find.ims_haddr = ntohl(psa->sin.sin_addr.s_addr);
	ims = RB_FIND(ip_msource_tree, &imf->imf_sources, &find);

	return ((struct in_msource *)ims);
}

/*
 * Perform filtering for multicast datagrams on a socket by group and source.
 *
 * Returns 0 if a datagram should be allowed through, or various error codes
 * if the socket was not a member of the group, or the source was muted, etc.
 */
int
imo_multi_filter(const struct ip_moptions *imo, const struct ifnet *ifp,
    const struct sockaddr *group, const struct sockaddr *src)
{
	size_t gidx;
	struct in_msource *ims;
	int mode;

	KASSERT(ifp != NULL, ("%s: null ifp", __func__));

	gidx = imo_match_group(imo, ifp, group);
	if (gidx == -1)
		return (MCAST_NOTGMEMBER);

	/*
	 * Check if the source was included in an (S,G) join.
	 * Allow reception on exclusive memberships by default,
	 * reject reception on inclusive memberships by default.
	 * Exclude source only if an in-mode exclude filter exists.
	 * Include source only if an in-mode include filter exists.
	 * NOTE: We are comparing group state here at IGMP t1 (now)
	 * with socket-layer t0 (since last downcall).
	 */
	mode = imo->imo_mfilters[gidx].imf_st[1];
	ims = imo_match_source(imo, gidx, src);

	if ((ims == NULL && mode == MCAST_INCLUDE) ||
	    (ims != NULL && ims->imsl_st[0] != mode))
		return (MCAST_NOTSMEMBER);

	return (MCAST_PASS);
}

/*
 * Find and return a reference to an in_multi record for (ifp, group),
 * and bump its reference count.
 * If one does not exist, try to allocate it, and update link-layer multicast
 * filters on ifp to listen for group.
 * Assumes the IN_MULTI lock is held across the call.
 * Return 0 if successful, otherwise return an appropriate error code.
 */
static int
in_getmulti(struct ifnet *ifp, const struct in_addr *group,
    struct in_multi **pinm)
{
	struct sockaddr_in	 gsin;
	struct ifmultiaddr	*ifma;
	struct in_ifinfo	*ii;
	struct in_multi		*inm;
	int error;

	IN_MULTI_LOCK_ASSERT();

	ii = (struct in_ifinfo *)ifp->if_afdata[AF_INET];

	inm = inm_lookup(ifp, *group);
	if (inm != NULL) {
		/*
		 * If we already joined this group, just bump the
		 * refcount and return it.
		 */
		KASSERT(inm->inm_refcount >= 1,
		    ("%s: bad refcount %d", __func__, inm->inm_refcount));
		++inm->inm_refcount;
		*pinm = inm;
		return (0);
	}

	memset(&gsin, 0, sizeof(gsin));
	gsin.sin_family = AF_INET;
	gsin.sin_len = sizeof(struct sockaddr_in);
	gsin.sin_addr = *group;

	/*
	 * Check if a link-layer group is already associated
	 * with this network-layer group on the given ifnet.
	 */
	error = if_addmulti(ifp, (struct sockaddr *)&gsin, &ifma);
	if (error != 0)
		return (error);

	/* XXX ifma_protospec must be covered by IF_ADDR_LOCK */
	IF_ADDR_LOCK(ifp);

	/*
	 * If something other than netinet is occupying the link-layer
	 * group, print a meaningful error message and back out of
	 * the allocation.
	 * Otherwise, bump the refcount on the existing network-layer
	 * group association and return it.
	 */
	if (ifma->ifma_protospec != NULL) {
		inm = (struct in_multi *)ifma->ifma_protospec;
#ifdef INVARIANTS
		KASSERT(ifma->ifma_addr != NULL, ("%s: no ifma_addr",
		    __func__));
		KASSERT(ifma->ifma_addr->sa_family == AF_INET,
		    ("%s: ifma not AF_INET", __func__));
		KASSERT(inm != NULL, ("%s: no ifma_protospec", __func__));
		if (inm->inm_ifma != ifma || inm->inm_ifp != ifp ||
		    !in_hosteq(inm->inm_addr, *group))
			panic("%s: ifma %p is inconsistent with %p (%s)",
			    __func__, ifma, inm, inet_ntoa(*group));
#endif
		++inm->inm_refcount;
		*pinm = inm;
		IF_ADDR_UNLOCK(ifp);
		return (0);
	}

	IF_ADDR_LOCK_ASSERT(ifp);

	/*
	 * A new in_multi record is needed; allocate and initialize it.
	 * We DO NOT perform an IGMP join as the in_ layer may need to
	 * push an initial source list down to IGMP to support SSM.
	 *
	 * The initial source filter state is INCLUDE, {} as per the RFC.
	 */
	inm = malloc(sizeof(*inm), M_IPMADDR, M_NOWAIT | M_ZERO);
	if (inm == NULL) {
		if_delmulti_ifma(ifma);
		IF_ADDR_UNLOCK(ifp);
		return (ENOMEM);
	}
	inm->inm_addr = *group;
	inm->inm_ifp = ifp;
	inm->inm_igi = ii->ii_igmp;
	inm->inm_ifma = ifma;
	inm->inm_refcount = 1;
	inm->inm_state = IGMP_NOT_MEMBER;

	/*
	 * Pending state-changes per group are subject to a bounds check.
	 */
	IFQ_SET_MAXLEN(&inm->inm_scq, IGMP_MAX_STATE_CHANGES);

	inm->inm_st[0].iss_fmode = MCAST_UNDEFINED;
	inm->inm_st[1].iss_fmode = MCAST_UNDEFINED;
	RB_INIT(&inm->inm_srcs);

	ifma->ifma_protospec = inm;

	*pinm = inm;

	IF_ADDR_UNLOCK(ifp);
	return (0);
}

/*
 * Drop a reference to an in_multi record.
 *
 * If the refcount drops to 0, free the in_multi record and
 * delete the underlying link-layer membership.
 */
void
inm_release_locked(struct in_multi *inm)
{
	struct ifmultiaddr *ifma;

	IN_MULTI_LOCK_ASSERT();

	CTR2(KTR_IGMPV3, "%s: refcount is %d", __func__, inm->inm_refcount);

	if (--inm->inm_refcount > 0) {
		CTR2(KTR_IGMPV3, "%s: refcount is now %d", __func__,
		    inm->inm_refcount);
		return;
	}

	CTR2(KTR_IGMPV3, "%s: freeing inm %p", __func__, inm);

	ifma = inm->inm_ifma;

	/* XXX this access is not covered by IF_ADDR_LOCK */
	CTR2(KTR_IGMPV3, "%s: purging ifma %p", __func__, ifma);
	KASSERT(ifma->ifma_protospec == inm,
	    ("%s: ifma_protospec != inm", __func__));
	ifma->ifma_protospec = NULL;

	inm_purge(inm);

	free(inm, M_IPMADDR);

	if_delmulti_ifma(ifma);
}

/*
 * Clear recorded source entries for a group.
 * Used by the IGMP code. Caller must hold the IN_MULTI lock.
 * FIXME: Should reap.
 */
void
inm_clear_recorded(struct in_multi *inm)
{
	struct ip_msource	*ims;

	IN_MULTI_LOCK_ASSERT();

	RB_FOREACH(ims, ip_msource_tree, &inm->inm_srcs) {
		if (ims->ims_stp) {
			ims->ims_stp = 0;
			--inm->inm_st[1].iss_rec;
		}
	}
	KASSERT(inm->inm_st[1].iss_rec == 0,
	    ("%s: iss_rec %d not 0", __func__, inm->inm_st[1].iss_rec));
}

/*
 * Record a source as pending for a Source-Group IGMPv3 query.
 * This lives here as it modifies the shared tree.
 *
 * inm is the group descriptor.
 * naddr is the address of the source to record in network-byte order.
 *
 * If the net.inet.igmp.sgalloc sysctl is non-zero, we will
 * lazy-allocate a source node in response to an SG query.
 * Otherwise, no allocation is performed. This saves some memory
 * with the trade-off that the source will not be reported to the
 * router if joined in the window between the query response and
 * the group actually being joined on the local host.
 *
 * VIMAGE: XXX: Currently the igmp_sgalloc feature has been removed.
 * This turns off the allocation of a recorded source entry if
 * the group has not been joined.
 *
 * Return 0 if the source didn't exist or was already marked as recorded.
 * Return 1 if the source was marked as recorded by this function.
 * Return <0 if any error occured (negated errno code).
 */
int
inm_record_source(struct in_multi *inm, const in_addr_t naddr)
{
	struct ip_msource	 find;
	struct ip_msource	*ims, *nims;

	IN_MULTI_LOCK_ASSERT();

	find.ims_haddr = ntohl(naddr);
	ims = RB_FIND(ip_msource_tree, &inm->inm_srcs, &find);
	if (ims && ims->ims_stp)
		return (0);
	if (ims == NULL) {
		if (inm->inm_nsrc == in_mcast_maxgrpsrc)
			return (-ENOSPC);
		nims = malloc(sizeof(struct ip_msource), M_IPMSOURCE,
		    M_NOWAIT | M_ZERO);
		if (nims == NULL)
			return (-ENOMEM);
		nims->ims_haddr = find.ims_haddr;
		RB_INSERT(ip_msource_tree, &inm->inm_srcs, nims);
		++inm->inm_nsrc;
		ims = nims;
	}

	/*
	 * Mark the source as recorded and update the recorded
	 * source count.
	 */
	++ims->ims_stp;
	++inm->inm_st[1].iss_rec;

	return (1);
}

/*
 * Return a pointer to an in_msource owned by an in_mfilter,
 * given its source address.
 * Lazy-allocate if needed. If this is a new entry its filter state is
 * undefined at t0.
 *
 * imf is the filter set being modified.
 * haddr is the source address in *host* byte-order.
 *
 * SMPng: May be called with locks held; malloc must not block.
 */
static int
imf_get_source(struct in_mfilter *imf, const struct sockaddr_in *psin,
    struct in_msource **plims)
{
	struct ip_msource	 find;
	struct ip_msource	*ims, *nims;
	struct in_msource	*lims;
	int			 error;

	error = 0;
	ims = NULL;
	lims = NULL;

	/* key is host byte order */
	find.ims_haddr = ntohl(psin->sin_addr.s_addr);
	ims = RB_FIND(ip_msource_tree, &imf->imf_sources, &find);
	lims = (struct in_msource *)ims;
	if (lims == NULL) {
		if (imf->imf_nsrc == in_mcast_maxsocksrc)
			return (ENOSPC);
		nims = malloc(sizeof(struct in_msource), M_INMFILTER,
		    M_NOWAIT | M_ZERO);
		if (nims == NULL)
			return (ENOMEM);
		lims = (struct in_msource *)nims;
		lims->ims_haddr = find.ims_haddr;
		lims->imsl_st[0] = MCAST_UNDEFINED;
		RB_INSERT(ip_msource_tree, &imf->imf_sources, nims);
		++imf->imf_nsrc;
	}

	*plims = lims;

	return (error);
}

/*
 * Graft a source entry into an existing socket-layer filter set,
 * maintaining any required invariants and checking allocations.
 *
 * The source is marked as being in the new filter mode at t1.
 *
 * Return the pointer to the new node, otherwise return NULL.
 */
static struct in_msource *
imf_graft(struct in_mfilter *imf, const uint8_t st1,
    const struct sockaddr_in *psin)
{
	struct ip_msource	*nims;
	struct in_msource	*lims;

	nims = malloc(sizeof(struct in_msource), M_INMFILTER,
	    M_NOWAIT | M_ZERO);
	if (nims == NULL)
		return (NULL);
	lims = (struct in_msource *)nims;
	lims->ims_haddr = ntohl(psin->sin_addr.s_addr);
	lims->imsl_st[0] = MCAST_UNDEFINED;
	lims->imsl_st[1] = st1;
	RB_INSERT(ip_msource_tree, &imf->imf_sources, nims);
	++imf->imf_nsrc;

	return (lims);
}

/*
 * Prune a source entry from an existing socket-layer filter set,
 * maintaining any required invariants and checking allocations.
 *
 * The source is marked as being left at t1, it is not freed.
 *
 * Return 0 if no error occurred, otherwise return an errno value.
 */
static int
imf_prune(struct in_mfilter *imf, const struct sockaddr_in *psin)
{
	struct ip_msource	 find;
	struct ip_msource	*ims;
	struct in_msource	*lims;

	/* key is host byte order */
	find.ims_haddr = ntohl(psin->sin_addr.s_addr);
	ims = RB_FIND(ip_msource_tree, &imf->imf_sources, &find);
	if (ims == NULL)
		return (ENOENT);
	lims = (struct in_msource *)ims;
	lims->imsl_st[1] = MCAST_UNDEFINED;
	return (0);
}

/*
 * Revert socket-layer filter set deltas at t1 to t0 state.
 */
static void
imf_rollback(struct in_mfilter *imf)
{
	struct ip_msource	*ims, *tims;
	struct in_msource	*lims;

	RB_FOREACH_SAFE(ims, ip_msource_tree, &imf->imf_sources, tims) {
		lims = (struct in_msource *)ims;
		if (lims->imsl_st[0] == lims->imsl_st[1]) {
			/* no change at t1 */
			continue;
		} else if (lims->imsl_st[0] != MCAST_UNDEFINED) {
			/* revert change to existing source at t1 */
			lims->imsl_st[1] = lims->imsl_st[0];
		} else {
			/* revert source added t1 */
			CTR2(KTR_IGMPV3, "%s: free ims %p", __func__, ims);
			RB_REMOVE(ip_msource_tree, &imf->imf_sources, ims);
			free(ims, M_INMFILTER);
			imf->imf_nsrc--;
		}
	}
	imf->imf_st[1] = imf->imf_st[0];
}

/*
 * Mark socket-layer filter set as INCLUDE {} at t1.
 */
static void
imf_leave(struct in_mfilter *imf)
{
	struct ip_msource	*ims;
	struct in_msource	*lims;

	RB_FOREACH(ims, ip_msource_tree, &imf->imf_sources) {
		lims = (struct in_msource *)ims;
		lims->imsl_st[1] = MCAST_UNDEFINED;
	}
	imf->imf_st[1] = MCAST_INCLUDE;
}

/*
 * Mark socket-layer filter set deltas as committed.
 */
static void
imf_commit(struct in_mfilter *imf)
{
	struct ip_msource	*ims;
	struct in_msource	*lims;

	RB_FOREACH(ims, ip_msource_tree, &imf->imf_sources) {
		lims = (struct in_msource *)ims;
		lims->imsl_st[0] = lims->imsl_st[1];
	}
	imf->imf_st[0] = imf->imf_st[1];
}

/*
 * Reap unreferenced sources from socket-layer filter set.
 */
static void
imf_reap(struct in_mfilter *imf)
{
	struct ip_msource	*ims, *tims;
	struct in_msource	*lims;

	RB_FOREACH_SAFE(ims, ip_msource_tree, &imf->imf_sources, tims) {
		lims = (struct in_msource *)ims;
		if ((lims->imsl_st[0] == MCAST_UNDEFINED) &&
		    (lims->imsl_st[1] == MCAST_UNDEFINED)) {
			CTR2(KTR_IGMPV3, "%s: free lims %p", __func__, ims);
			RB_REMOVE(ip_msource_tree, &imf->imf_sources, ims);
			free(ims, M_INMFILTER);
			imf->imf_nsrc--;
		}
	}
}

/*
 * Purge socket-layer filter set.
 */
static void
imf_purge(struct in_mfilter *imf)
{
	struct ip_msource	*ims, *tims;

	RB_FOREACH_SAFE(ims, ip_msource_tree, &imf->imf_sources, tims) {
		CTR2(KTR_IGMPV3, "%s: free ims %p", __func__, ims);
		RB_REMOVE(ip_msource_tree, &imf->imf_sources, ims);
		free(ims, M_INMFILTER);
		imf->imf_nsrc--;
	}
	imf->imf_st[0] = imf->imf_st[1] = MCAST_UNDEFINED;
	KASSERT(RB_EMPTY(&imf->imf_sources),
	    ("%s: imf_sources not empty", __func__));
}

/*
 * Look up a source filter entry for a multicast group.
 *
 * inm is the group descriptor to work with.
 * haddr is the host-byte-order IPv4 address to look up.
 * noalloc may be non-zero to suppress allocation of sources.
 * *pims will be set to the address of the retrieved or allocated source.
 *
 * SMPng: NOTE: may be called with locks held.
 * Return 0 if successful, otherwise return a non-zero error code.
 */
static int
inm_get_source(struct in_multi *inm, const in_addr_t haddr,
    const int noalloc, struct ip_msource **pims)
{
	struct ip_msource	 find;
	struct ip_msource	*ims, *nims;
#ifdef KTR
	struct in_addr ia;
#endif

	find.ims_haddr = haddr;
	ims = RB_FIND(ip_msource_tree, &inm->inm_srcs, &find);
	if (ims == NULL && !noalloc) {
		if (inm->inm_nsrc == in_mcast_maxgrpsrc)
			return (ENOSPC);
		nims = malloc(sizeof(struct ip_msource), M_IPMSOURCE,
		    M_NOWAIT | M_ZERO);
		if (nims == NULL)
			return (ENOMEM);
		nims->ims_haddr = haddr;
		RB_INSERT(ip_msource_tree, &inm->inm_srcs, nims);
		++inm->inm_nsrc;
		ims = nims;
#ifdef KTR
		ia.s_addr = htonl(haddr);
		CTR3(KTR_IGMPV3, "%s: allocated %s as %p", __func__,
		    inet_ntoa(ia), ims);
#endif
	}

	*pims = ims;
	return (0);
}

/*
 * Merge socket-layer source into IGMP-layer source.
 * If rollback is non-zero, perform the inverse of the merge.
 */
static void
ims_merge(struct ip_msource *ims, const struct in_msource *lims,
    const int rollback)
{
	int n = rollback ? -1 : 1;
#ifdef KTR
	struct in_addr ia;

	ia.s_addr = htonl(ims->ims_haddr);
#endif

	if (lims->imsl_st[0] == MCAST_EXCLUDE) {
		CTR3(KTR_IGMPV3, "%s: t1 ex -= %d on %s",
		    __func__, n, inet_ntoa(ia));
		ims->ims_st[1].ex -= n;
	} else if (lims->imsl_st[0] == MCAST_INCLUDE) {
		CTR3(KTR_IGMPV3, "%s: t1 in -= %d on %s",
		    __func__, n, inet_ntoa(ia));
		ims->ims_st[1].in -= n;
	}

	if (lims->imsl_st[1] == MCAST_EXCLUDE) {
		CTR3(KTR_IGMPV3, "%s: t1 ex += %d on %s",
		    __func__, n, inet_ntoa(ia));
		ims->ims_st[1].ex += n;
	} else if (lims->imsl_st[1] == MCAST_INCLUDE) {
		CTR3(KTR_IGMPV3, "%s: t1 in += %d on %s",
		    __func__, n, inet_ntoa(ia));
		ims->ims_st[1].in += n;
	}
}

/*
 * Atomically update the global in_multi state, when a membership's
 * filter list is being updated in any way.
 *
 * imf is the per-inpcb-membership group filter pointer.
 * A fake imf may be passed for in-kernel consumers.
 *
 * XXX This is a candidate for a set-symmetric-difference style loop
 * which would eliminate the repeated lookup from root of ims nodes,
 * as they share the same key space.
 *
 * If any error occurred this function will back out of refcounts
 * and return a non-zero value.
 */
static int
inm_merge(struct in_multi *inm, /*const*/ struct in_mfilter *imf)
{
	struct ip_msource	*ims, *nims;
	struct in_msource	*lims;
	int			 schanged, error;
	int			 nsrc0, nsrc1;

	schanged = 0;
	error = 0;
	nsrc1 = nsrc0 = 0;

	/*
	 * Update the source filters first, as this may fail.
	 * Maintain count of in-mode filters at t0, t1. These are
	 * used to work out if we transition into ASM mode or not.
	 * Maintain a count of source filters whose state was
	 * actually modified by this operation.
	 */
	RB_FOREACH(ims, ip_msource_tree, &imf->imf_sources) {
		lims = (struct in_msource *)ims;
		if (lims->imsl_st[0] == imf->imf_st[0]) nsrc0++;
		if (lims->imsl_st[1] == imf->imf_st[1]) nsrc1++;
		if (lims->imsl_st[0] == lims->imsl_st[1]) continue;
		error = inm_get_source(inm, lims->ims_haddr, 0, &nims);
		++schanged;
		if (error)
			break;
		ims_merge(nims, lims, 0);
	}
	if (error) {
		struct ip_msource *bims;

		RB_FOREACH_REVERSE_FROM(ims, ip_msource_tree, nims) {
			lims = (struct in_msource *)ims;
			if (lims->imsl_st[0] == lims->imsl_st[1])
				continue;
			(void)inm_get_source(inm, lims->ims_haddr, 1, &bims);
			if (bims == NULL)
				continue;
			ims_merge(bims, lims, 1);
		}
		goto out_reap;
	}

	CTR3(KTR_IGMPV3, "%s: imf filters in-mode: %d at t0, %d at t1",
	    __func__, nsrc0, nsrc1);

	/* Handle transition between INCLUDE {n} and INCLUDE {} on socket. */
	if (imf->imf_st[0] == imf->imf_st[1] &&
	    imf->imf_st[1] == MCAST_INCLUDE) {
		if (nsrc1 == 0) {
			CTR1(KTR_IGMPV3, "%s: --in on inm at t1", __func__);
			--inm->inm_st[1].iss_in;
		}
	}

	/* Handle filter mode transition on socket. */
	if (imf->imf_st[0] != imf->imf_st[1]) {
		CTR3(KTR_IGMPV3, "%s: imf transition %d to %d",
		    __func__, imf->imf_st[0], imf->imf_st[1]);

		if (imf->imf_st[0] == MCAST_EXCLUDE) {
			CTR1(KTR_IGMPV3, "%s: --ex on inm at t1", __func__);
			--inm->inm_st[1].iss_ex;
		} else if (imf->imf_st[0] == MCAST_INCLUDE) {
			CTR1(KTR_IGMPV3, "%s: --in on inm at t1", __func__);
			--inm->inm_st[1].iss_in;
		}

		if (imf->imf_st[1] == MCAST_EXCLUDE) {
			CTR1(KTR_IGMPV3, "%s: ex++ on inm at t1", __func__);
			inm->inm_st[1].iss_ex++;
		} else if (imf->imf_st[1] == MCAST_INCLUDE && nsrc1 > 0) {
			CTR1(KTR_IGMPV3, "%s: in++ on inm at t1", __func__);
			inm->inm_st[1].iss_in++;
		}
	}

	/*
	 * Track inm filter state in terms of listener counts.
	 * If there are any exclusive listeners, stack-wide
	 * membership is exclusive.
	 * Otherwise, if only inclusive listeners, stack-wide is inclusive.
	 * If no listeners remain, state is undefined at t1,
	 * and the IGMP lifecycle for this group should finish.
	 */
	if (inm->inm_st[1].iss_ex > 0) {
		CTR1(KTR_IGMPV3, "%s: transition to EX", __func__);
		inm->inm_st[1].iss_fmode = MCAST_EXCLUDE;
	} else if (inm->inm_st[1].iss_in > 0) {
		CTR1(KTR_IGMPV3, "%s: transition to IN", __func__);
		inm->inm_st[1].iss_fmode = MCAST_INCLUDE;
	} else {
		CTR1(KTR_IGMPV3, "%s: transition to UNDEF", __func__);
		inm->inm_st[1].iss_fmode = MCAST_UNDEFINED;
	}

	/* Decrement ASM listener count on transition out of ASM mode. */
	if (imf->imf_st[0] == MCAST_EXCLUDE && nsrc0 == 0) {
		if ((imf->imf_st[1] != MCAST_EXCLUDE) ||
		    (imf->imf_st[1] == MCAST_EXCLUDE && nsrc1 > 0))
			CTR1(KTR_IGMPV3, "%s: --asm on inm at t1", __func__);
			--inm->inm_st[1].iss_asm;
	}

	/* Increment ASM listener count on transition to ASM mode. */
	if (imf->imf_st[1] == MCAST_EXCLUDE && nsrc1 == 0) {
		CTR1(KTR_IGMPV3, "%s: asm++ on inm at t1", __func__);
		inm->inm_st[1].iss_asm++;
	}

	CTR3(KTR_IGMPV3, "%s: merged imf %p to inm %p", __func__, imf, inm);
	inm_print(inm);

out_reap:
	if (schanged > 0) {
		CTR1(KTR_IGMPV3, "%s: sources changed; reaping", __func__);
		inm_reap(inm);
	}
	return (error);
}

/*
 * Mark an in_multi's filter set deltas as committed.
 * Called by IGMP after a state change has been enqueued.
 */
void
inm_commit(struct in_multi *inm)
{
	struct ip_msource	*ims;

	CTR2(KTR_IGMPV3, "%s: commit inm %p", __func__, inm);
	CTR1(KTR_IGMPV3, "%s: pre commit:", __func__);
	inm_print(inm);

	RB_FOREACH(ims, ip_msource_tree, &inm->inm_srcs) {
		ims->ims_st[0] = ims->ims_st[1];
	}
	inm->inm_st[0] = inm->inm_st[1];
}

/*
 * Reap unreferenced nodes from an in_multi's filter set.
 */
static void
inm_reap(struct in_multi *inm)
{
	struct ip_msource	*ims, *tims;

	RB_FOREACH_SAFE(ims, ip_msource_tree, &inm->inm_srcs, tims) {
		if (ims->ims_st[0].ex > 0 || ims->ims_st[0].in > 0 ||
		    ims->ims_st[1].ex > 0 || ims->ims_st[1].in > 0 ||
		    ims->ims_stp != 0)
			continue;
		CTR2(KTR_IGMPV3, "%s: free ims %p", __func__, ims);
		RB_REMOVE(ip_msource_tree, &inm->inm_srcs, ims);
		free(ims, M_IPMSOURCE);
		inm->inm_nsrc--;
	}
}

/*
 * Purge all source nodes from an in_multi's filter set.
 */
static void
inm_purge(struct in_multi *inm)
{
	struct ip_msource	*ims, *tims;

	RB_FOREACH_SAFE(ims, ip_msource_tree, &inm->inm_srcs, tims) {
		CTR2(KTR_IGMPV3, "%s: free ims %p", __func__, ims);
		RB_REMOVE(ip_msource_tree, &inm->inm_srcs, ims);
		free(ims, M_IPMSOURCE);
		inm->inm_nsrc--;
	}
}

/*
 * Join a multicast group; unlocked entry point.
 *
 * SMPng: XXX: in_joingroup() is called from in_control() when Giant
 * is not held. Fortunately, ifp is unlikely to have been detached
 * at this point, so we assume it's OK to recurse.
 */
int
in_joingroup(struct ifnet *ifp, const struct in_addr *gina,
    /*const*/ struct in_mfilter *imf, struct in_multi **pinm)
{
	int error;

	IN_MULTI_LOCK();
	error = in_joingroup_locked(ifp, gina, imf, pinm);
	IN_MULTI_UNLOCK();

	return (error);
}

/*
 * Join a multicast group; real entry point.
 *
 * Only preserves atomicity at inm level.
 * NOTE: imf argument cannot be const due to sys/tree.h limitations.
 *
 * If the IGMP downcall fails, the group is not joined, and an error
 * code is returned.
 */
int
in_joingroup_locked(struct ifnet *ifp, const struct in_addr *gina,
    /*const*/ struct in_mfilter *imf, struct in_multi **pinm)
{
	struct in_mfilter	 timf;
	struct in_multi		*inm;
	int			 error;

	IN_MULTI_LOCK_ASSERT();

	CTR4(KTR_IGMPV3, "%s: join %s on %p(%s))", __func__,
	    inet_ntoa(*gina), ifp, ifp->if_xname);

	error = 0;
	inm = NULL;

	/*
	 * If no imf was specified (i.e. kernel consumer),
	 * fake one up and assume it is an ASM join.
	 */
	if (imf == NULL) {
		imf_init(&timf, MCAST_UNDEFINED, MCAST_EXCLUDE);
		imf = &timf;
	}

	error = in_getmulti(ifp, gina, &inm);
	if (error) {
		CTR1(KTR_IGMPV3, "%s: in_getmulti() failure", __func__);
		return (error);
	}

	CTR1(KTR_IGMPV3, "%s: merge inm state", __func__);
	error = inm_merge(inm, imf);
	if (error) {
		CTR1(KTR_IGMPV3, "%s: failed to merge inm state", __func__);
		goto out_inm_release;
	}

	CTR1(KTR_IGMPV3, "%s: doing igmp downcall", __func__);
	error = igmp_change_state(inm);
	if (error) {
		CTR1(KTR_IGMPV3, "%s: failed to update source", __func__);
		goto out_inm_release;
	}

out_inm_release:
	if (error) {
		CTR2(KTR_IGMPV3, "%s: dropping ref on %p", __func__, inm);
		inm_release_locked(inm);
	} else {
		*pinm = inm;
	}

	return (error);
}

/*
 * Leave a multicast group; unlocked entry point.
 */
int
in_leavegroup(struct in_multi *inm, /*const*/ struct in_mfilter *imf)
{
	struct ifnet *ifp;
	int error;

	ifp = inm->inm_ifp;

	IN_MULTI_LOCK();
	error = in_leavegroup_locked(inm, imf);
	IN_MULTI_UNLOCK();

	return (error);
}

/*
 * Leave a multicast group; real entry point.
 * All source filters will be expunged.
 *
 * Only preserves atomicity at inm level.
 *
 * Holding the write lock for the INP which contains imf
 * is highly advisable. We can't assert for it as imf does not
 * contain a back-pointer to the owning inp.
 *
 * Note: This is not the same as inm_release(*) as this function also
 * makes a state change downcall into IGMP.
 */
int
in_leavegroup_locked(struct in_multi *inm, /*const*/ struct in_mfilter *imf)
{
	struct in_mfilter	 timf;
	int			 error;

	error = 0;

	IN_MULTI_LOCK_ASSERT();

	CTR5(KTR_IGMPV3, "%s: leave inm %p, %s/%s, imf %p", __func__,
	    inm, inet_ntoa(inm->inm_addr),
	    (inm_is_ifp_detached(inm) ? "null" : inm->inm_ifp->if_xname),
	    imf);

	/*
	 * If no imf was specified (i.e. kernel consumer),
	 * fake one up and assume it is an ASM join.
	 */
	if (imf == NULL) {
		imf_init(&timf, MCAST_EXCLUDE, MCAST_UNDEFINED);
		imf = &timf;
	}

	/*
	 * Begin state merge transaction at IGMP layer.
	 *
	 * As this particular invocation should not cause any memory
	 * to be allocated, and there is no opportunity to roll back
	 * the transaction, it MUST NOT fail.
	 */
	CTR1(KTR_IGMPV3, "%s: merge inm state", __func__);
	error = inm_merge(inm, imf);
	KASSERT(error == 0, ("%s: failed to merge inm state", __func__));

	CTR1(KTR_IGMPV3, "%s: doing igmp downcall", __func__);
	error = igmp_change_state(inm);
	if (error)
		CTR1(KTR_IGMPV3, "%s: failed igmp downcall", __func__);

	CTR2(KTR_IGMPV3, "%s: dropping ref on %p", __func__, inm);
	inm_release_locked(inm);

	return (error);
}

/*#ifndef BURN_BRIDGES*/
/*
 * Join an IPv4 multicast group in (*,G) exclusive mode.
 * The group must be a 224.0.0.0/24 link-scope group.
 * This KPI is for legacy kernel consumers only.
 */
struct in_multi *
in_addmulti(struct in_addr *ap, struct ifnet *ifp)
{
	struct in_multi *pinm;
	int error;

	KASSERT(IN_LOCAL_GROUP(ntohl(ap->s_addr)),
	    ("%s: %s not in 224.0.0.0/24", __func__, inet_ntoa(*ap)));

	error = in_joingroup(ifp, ap, NULL, &pinm);
	if (error != 0)
		pinm = NULL;

	return (pinm);
}

/*
 * Leave an IPv4 multicast group, assumed to be in exclusive (*,G) mode.
 * This KPI is for legacy kernel consumers only.
 */
void
in_delmulti(struct in_multi *inm)
{

	(void)in_leavegroup(inm, NULL);
}
/*#endif*/

/*
 * Block or unblock an ASM multicast source on an inpcb.
 * This implements the delta-based API described in RFC 3678.
 *
 * The delta-based API applies only to exclusive-mode memberships.
 * An IGMP downcall will be performed.
 *
 * SMPng: NOTE: Must take Giant as a join may create a new ifma.
 *
 * Return 0 if successful, otherwise return an appropriate error code.
 */
static int
inp_block_unblock_source(struct inpcb *inp, struct sockopt *sopt)
{
	struct group_source_req		 gsr;
	sockunion_t			*gsa, *ssa;
	struct ifnet			*ifp;
	struct in_mfilter		*imf;
	struct ip_moptions		*imo;
	struct in_msource		*ims;
	struct in_multi			*inm;
	size_t				 idx;
	uint16_t			 fmode;
	int				 error, doblock;

	ifp = NULL;
	error = 0;
	doblock = 0;

	memset(&gsr, 0, sizeof(struct group_source_req));
	gsa = (sockunion_t *)&gsr.gsr_group;
	ssa = (sockunion_t *)&gsr.gsr_source;

	switch (sopt->sopt_name) {
	case IP_BLOCK_SOURCE:
	case IP_UNBLOCK_SOURCE: {
		struct ip_mreq_source	 mreqs;

		error = sooptcopyin(sopt, &mreqs,
		    sizeof(struct ip_mreq_source),
		    sizeof(struct ip_mreq_source));
		if (error)
			return (error);

		gsa->sin.sin_family = AF_INET;
		gsa->sin.sin_len = sizeof(struct sockaddr_in);
		gsa->sin.sin_addr = mreqs.imr_multiaddr;

		ssa->sin.sin_family = AF_INET;
		ssa->sin.sin_len = sizeof(struct sockaddr_in);
		ssa->sin.sin_addr = mreqs.imr_sourceaddr;

		if (!in_nullhost(mreqs.imr_interface))
			INADDR_TO_IFP(mreqs.imr_interface, ifp);

		if (sopt->sopt_name == IP_BLOCK_SOURCE)
			doblock = 1;

		CTR3(KTR_IGMPV3, "%s: imr_interface = %s, ifp = %p",
		    __func__, inet_ntoa(mreqs.imr_interface), ifp);
		break;
	    }

	case MCAST_BLOCK_SOURCE:
	case MCAST_UNBLOCK_SOURCE:
		error = sooptcopyin(sopt, &gsr,
		    sizeof(struct group_source_req),
		    sizeof(struct group_source_req));
		if (error)
			return (error);

		if (gsa->sin.sin_family != AF_INET ||
		    gsa->sin.sin_len != sizeof(struct sockaddr_in))
			return (EINVAL);

		if (ssa->sin.sin_family != AF_INET ||
		    ssa->sin.sin_len != sizeof(struct sockaddr_in))
			return (EINVAL);

		if (gsr.gsr_interface == 0 || V_if_index < gsr.gsr_interface)
			return (EADDRNOTAVAIL);

		ifp = ifnet_byindex(gsr.gsr_interface);

		if (sopt->sopt_name == MCAST_BLOCK_SOURCE)
			doblock = 1;
		break;

	default:
		CTR2(KTR_IGMPV3, "%s: unknown sopt_name %d",
		    __func__, sopt->sopt_name);
		return (EOPNOTSUPP);
		break;
	}

	if (!IN_MULTICAST(ntohl(gsa->sin.sin_addr.s_addr)))
		return (EINVAL);

	/*
	 * Check if we are actually a member of this group.
	 */
	imo = inp_findmoptions(inp);
	idx = imo_match_group(imo, ifp, &gsa->sa);
	if (idx == -1 || imo->imo_mfilters == NULL) {
		error = EADDRNOTAVAIL;
		goto out_inp_locked;
	}

	KASSERT(imo->imo_mfilters != NULL,
	    ("%s: imo_mfilters not allocated", __func__));
	imf = &imo->imo_mfilters[idx];
	inm = imo->imo_membership[idx];

	/*
	 * Attempting to use the delta-based API on an
	 * non exclusive-mode membership is an error.
	 */
	fmode = imf->imf_st[0];
	if (fmode != MCAST_EXCLUDE) {
		error = EINVAL;
		goto out_inp_locked;
	}

	/*
	 * Deal with error cases up-front:
	 *  Asked to block, but already blocked; or
	 *  Asked to unblock, but nothing to unblock.
	 * If adding a new block entry, allocate it.
	 */
	ims = imo_match_source(imo, idx, &ssa->sa);
	if ((ims != NULL && doblock) || (ims == NULL && !doblock)) {
		CTR3(KTR_IGMPV3, "%s: source %s %spresent", __func__,
		    inet_ntoa(ssa->sin.sin_addr), doblock ? "" : "not ");
		error = EADDRNOTAVAIL;
		goto out_inp_locked;
	}

	INP_WLOCK_ASSERT(inp);

	/*
	 * Begin state merge transaction at socket layer.
	 */
	if (doblock) {
		CTR2(KTR_IGMPV3, "%s: %s source", __func__, "block");
		ims = imf_graft(imf, fmode, &ssa->sin);
		if (ims == NULL)
			error = ENOMEM;
	} else {
		CTR2(KTR_IGMPV3, "%s: %s source", __func__, "allow");
		error = imf_prune(imf, &ssa->sin);
	}

	if (error) {
		CTR1(KTR_IGMPV3, "%s: merge imf state failed", __func__);
		goto out_imf_rollback;
	}

	/*
	 * Begin state merge transaction at IGMP layer.
	 */
	IN_MULTI_LOCK();

	CTR1(KTR_IGMPV3, "%s: merge inm state", __func__);
	error = inm_merge(inm, imf);
	if (error) {
		CTR1(KTR_IGMPV3, "%s: failed to merge inm state", __func__);
		goto out_imf_rollback;
	}

	CTR1(KTR_IGMPV3, "%s: doing igmp downcall", __func__);
	error = igmp_change_state(inm);
	if (error)
		CTR1(KTR_IGMPV3, "%s: failed igmp downcall", __func__);

	IN_MULTI_UNLOCK();

out_imf_rollback:
	if (error)
		imf_rollback(imf);
	else
		imf_commit(imf);

	imf_reap(imf);

out_inp_locked:
	INP_WUNLOCK(inp);
	return (error);
}

/*
 * Given an inpcb, return its multicast options structure pointer.  Accepts
 * an unlocked inpcb pointer, but will return it locked.  May sleep.
 *
 * SMPng: NOTE: Potentially calls malloc(M_WAITOK) with Giant held.
 * SMPng: NOTE: Returns with the INP write lock held.
 */
static struct ip_moptions *
inp_findmoptions(struct inpcb *inp)
{
	struct ip_moptions	 *imo;
	struct in_multi		**immp;
	struct in_mfilter	 *imfp;
	size_t			  idx;

	INP_WLOCK(inp);
	if (inp->inp_moptions != NULL)
		return (inp->inp_moptions);

	INP_WUNLOCK(inp);

	imo = malloc(sizeof(*imo), M_IPMOPTS, M_WAITOK);
	immp = malloc(sizeof(*immp) * IP_MIN_MEMBERSHIPS, M_IPMOPTS,
	    M_WAITOK | M_ZERO);
	imfp = malloc(sizeof(struct in_mfilter) * IP_MIN_MEMBERSHIPS,
	    M_INMFILTER, M_WAITOK);

	imo->imo_multicast_ifp = NULL;
	imo->imo_multicast_addr.s_addr = INADDR_ANY;
	imo->imo_multicast_vif = -1;
	imo->imo_multicast_ttl = IP_DEFAULT_MULTICAST_TTL;
	imo->imo_multicast_loop = in_mcast_loop;
	imo->imo_num_memberships = 0;
	imo->imo_max_memberships = IP_MIN_MEMBERSHIPS;
	imo->imo_membership = immp;

	/* Initialize per-group source filters. */
	for (idx = 0; idx < IP_MIN_MEMBERSHIPS; idx++)
		imf_init(&imfp[idx], MCAST_UNDEFINED, MCAST_EXCLUDE);
	imo->imo_mfilters = imfp;

	INP_WLOCK(inp);
	if (inp->inp_moptions != NULL) {
		free(imfp, M_INMFILTER);
		free(immp, M_IPMOPTS);
		free(imo, M_IPMOPTS);
		return (inp->inp_moptions);
	}
	inp->inp_moptions = imo;
	return (imo);
}

/*
 * Discard the IP multicast options (and source filters).
 *
 * SMPng: NOTE: assumes INP write lock is held.
 */
void
inp_freemoptions(struct ip_moptions *imo)
{
	struct in_mfilter	*imf;
	size_t			 idx, nmships;

	KASSERT(imo != NULL, ("%s: ip_moptions is NULL", __func__));

	nmships = imo->imo_num_memberships;
	for (idx = 0; idx < nmships; ++idx) {
		imf = imo->imo_mfilters ? &imo->imo_mfilters[idx] : NULL;
		if (imf)
			imf_leave(imf);
		(void)in_leavegroup(imo->imo_membership[idx], imf);
		if (imf)
			imf_purge(imf);
	}

	if (imo->imo_mfilters)
		free(imo->imo_mfilters, M_INMFILTER);
	free(imo->imo_membership, M_IPMOPTS);
	free(imo, M_IPMOPTS);
}

/*
 * Atomically get source filters on a socket for an IPv4 multicast group.
 * Called with INP lock held; returns with lock released.
 */
static int
inp_get_source_filters(struct inpcb *inp, struct sockopt *sopt)
{
	struct __msfilterreq	 msfr;
	sockunion_t		*gsa;
	struct ifnet		*ifp;
	struct ip_moptions	*imo;
	struct in_mfilter	*imf;
	struct ip_msource	*ims;
	struct in_msource	*lims;
	struct sockaddr_in	*psin;
	struct sockaddr_storage	*ptss;
	struct sockaddr_storage	*tss;
	int			 error;
	size_t			 idx, nsrcs, ncsrcs;

	INP_WLOCK_ASSERT(inp);

	imo = inp->inp_moptions;
	KASSERT(imo != NULL, ("%s: null ip_moptions", __func__));

	INP_WUNLOCK(inp);

	error = sooptcopyin(sopt, &msfr, sizeof(struct __msfilterreq),
	    sizeof(struct __msfilterreq));
	if (error)
		return (error);

	if (msfr.msfr_ifindex == 0 || V_if_index < msfr.msfr_ifindex)
		return (EINVAL);

	ifp = ifnet_byindex(msfr.msfr_ifindex);
	if (ifp == NULL)
		return (EINVAL);

	INP_WLOCK(inp);

	/*
	 * Lookup group on the socket.
	 */
	gsa = (sockunion_t *)&msfr.msfr_group;
	idx = imo_match_group(imo, ifp, &gsa->sa);
	if (idx == -1 || imo->imo_mfilters == NULL) {
		INP_WUNLOCK(inp);
		return (EADDRNOTAVAIL);
	}
	imf = &imo->imo_mfilters[idx];

	/*
	 * Ignore memberships which are in limbo.
	 */
	if (imf->imf_st[1] == MCAST_UNDEFINED) {
		INP_WUNLOCK(inp);
		return (EAGAIN);
	}
	msfr.msfr_fmode = imf->imf_st[1];

	/*
	 * If the user specified a buffer, copy out the source filter
	 * entries to userland gracefully.
	 * We only copy out the number of entries which userland
	 * has asked for, but we always tell userland how big the
	 * buffer really needs to be.
	 */
	tss = NULL;
	if (msfr.msfr_srcs != NULL && msfr.msfr_nsrcs > 0) {
		tss = malloc(sizeof(struct sockaddr_storage) * msfr.msfr_nsrcs,
		    M_TEMP, M_NOWAIT | M_ZERO);
		if (tss == NULL) {
			INP_WUNLOCK(inp);
			return (ENOBUFS);
		}
	}

	/*
	 * Count number of sources in-mode at t0.
	 * If buffer space exists and remains, copy out source entries.
	 */
	nsrcs = msfr.msfr_nsrcs;
	ncsrcs = 0;
	ptss = tss;
	RB_FOREACH(ims, ip_msource_tree, &imf->imf_sources) {
		lims = (struct in_msource *)ims;
		if (lims->imsl_st[0] == MCAST_UNDEFINED ||
		    lims->imsl_st[0] != imf->imf_st[0])
			continue;
		++ncsrcs;
		if (tss != NULL && nsrcs > 0) {
			psin = (struct sockaddr_in *)ptss;
			psin->sin_family = AF_INET;
			psin->sin_len = sizeof(struct sockaddr_in);
			psin->sin_addr.s_addr = htonl(lims->ims_haddr);
			psin->sin_port = 0;
			++ptss;
			--nsrcs;
		}
	}

	INP_WUNLOCK(inp);

	if (tss != NULL) {
		error = copyout(tss, msfr.msfr_srcs,
		    sizeof(struct sockaddr_storage) * msfr.msfr_nsrcs);
		free(tss, M_TEMP);
		if (error)
			return (error);
	}

	msfr.msfr_nsrcs = ncsrcs;
	error = sooptcopyout(sopt, &msfr, sizeof(struct __msfilterreq));

	return (error);
}

/*
 * Return the IP multicast options in response to user getsockopt().
 */
int
inp_getmoptions(struct inpcb *inp, struct sockopt *sopt)
{
	struct ip_mreqn		 mreqn;
	struct ip_moptions	*imo;
	struct ifnet		*ifp;
	struct in_ifaddr	*ia;
	int			 error, optval;
	u_char			 coptval;

	INP_WLOCK(inp);
	imo = inp->inp_moptions;
	/*
	 * If socket is neither of type SOCK_RAW or SOCK_DGRAM,
	 * or is a divert socket, reject it.
	 */
	if (inp->inp_socket->so_proto->pr_protocol == IPPROTO_DIVERT ||
	    (inp->inp_socket->so_proto->pr_type != SOCK_RAW &&
	    inp->inp_socket->so_proto->pr_type != SOCK_DGRAM)) {
		INP_WUNLOCK(inp);
		return (EOPNOTSUPP);
	}

	error = 0;
	switch (sopt->sopt_name) {
	case IP_MULTICAST_VIF:
		if (imo != NULL)
			optval = imo->imo_multicast_vif;
		else
			optval = -1;
		INP_WUNLOCK(inp);
		error = sooptcopyout(sopt, &optval, sizeof(int));
		break;

	case IP_MULTICAST_IF:
		memset(&mreqn, 0, sizeof(struct ip_mreqn));
		if (imo != NULL) {
			ifp = imo->imo_multicast_ifp;
			if (!in_nullhost(imo->imo_multicast_addr)) {
				mreqn.imr_address = imo->imo_multicast_addr;
			} else if (ifp != NULL) {
				mreqn.imr_ifindex = ifp->if_index;
				IFP_TO_IA(ifp, ia);
				if (ia != NULL) {
					mreqn.imr_address =
					    IA_SIN(ia)->sin_addr;
					ifa_free(&ia->ia_ifa);
				}
			}
		}
		INP_WUNLOCK(inp);
		if (sopt->sopt_valsize == sizeof(struct ip_mreqn)) {
			error = sooptcopyout(sopt, &mreqn,
			    sizeof(struct ip_mreqn));
		} else {
			error = sooptcopyout(sopt, &mreqn.imr_address,
			    sizeof(struct in_addr));
		}
		break;

	case IP_MULTICAST_TTL:
		if (imo == 0)
			optval = coptval = IP_DEFAULT_MULTICAST_TTL;
		else
			optval = coptval = imo->imo_multicast_ttl;
		INP_WUNLOCK(inp);
		if (sopt->sopt_valsize == sizeof(u_char))
			error = sooptcopyout(sopt, &coptval, sizeof(u_char));
		else
			error = sooptcopyout(sopt, &optval, sizeof(int));
		break;

	case IP_MULTICAST_LOOP:
		if (imo == 0)
			optval = coptval = IP_DEFAULT_MULTICAST_LOOP;
		else
			optval = coptval = imo->imo_multicast_loop;
		INP_WUNLOCK(inp);
		if (sopt->sopt_valsize == sizeof(u_char))
			error = sooptcopyout(sopt, &coptval, sizeof(u_char));
		else
			error = sooptcopyout(sopt, &optval, sizeof(int));
		break;

	case IP_MSFILTER:
		if (imo == NULL) {
			error = EADDRNOTAVAIL;
			INP_WUNLOCK(inp);
		} else {
			error = inp_get_source_filters(inp, sopt);
		}
		break;

	default:
		INP_WUNLOCK(inp);
		error = ENOPROTOOPT;
		break;
	}

	INP_UNLOCK_ASSERT(inp);

	return (error);
}

/*
 * Look up the ifnet to use for a multicast group membership,
 * given the IPv4 address of an interface, and the IPv4 group address.
 *
 * This routine exists to support legacy multicast applications
 * which do not understand that multicast memberships are scoped to
 * specific physical links in the networking stack, or which need
 * to join link-scope groups before IPv4 addresses are configured.
 *
 * If inp is non-NULL, use this socket's current FIB number for any
 * required FIB lookup.
 * If ina is INADDR_ANY, look up the group address in the unicast FIB,
 * and use its ifp; usually, this points to the default next-hop.
 *
 * If the FIB lookup fails, attempt to use the first non-loopback
 * interface with multicast capability in the system as a
 * last resort. The legacy IPv4 ASM API requires that we do
 * this in order to allow groups to be joined when the routing
 * table has not yet been populated during boot.
 *
 * Returns NULL if no ifp could be found.
 *
 * SMPng: TODO: Acquire the appropriate locks for INADDR_TO_IFP.
 * FUTURE: Implement IPv4 source-address selection.
 */
static struct ifnet *
inp_lookup_mcast_ifp(const struct inpcb *inp,
    const struct sockaddr_in *gsin, const struct in_addr ina)
{
	struct ifnet *ifp;

	KASSERT(gsin->sin_family == AF_INET, ("%s: not AF_INET", __func__));
	KASSERT(IN_MULTICAST(ntohl(gsin->sin_addr.s_addr)),
	    ("%s: not multicast", __func__));

	ifp = NULL;
	if (!in_nullhost(ina)) {
		INADDR_TO_IFP(ina, ifp);
	} else {
		struct route ro;

		ro.ro_rt = NULL;
		memcpy(&ro.ro_dst, gsin, sizeof(struct sockaddr_in));
		in_rtalloc_ign(&ro, 0, inp ? inp->inp_inc.inc_fibnum : 0);
		if (ro.ro_rt != NULL) {
			ifp = ro.ro_rt->rt_ifp;
			KASSERT(ifp != NULL, ("%s: null ifp", __func__));
			RTFREE(ro.ro_rt);
		} else {
			struct in_ifaddr *ia;
			struct ifnet *mifp;

			mifp = NULL;
			IN_IFADDR_RLOCK();
			TAILQ_FOREACH(ia, &V_in_ifaddrhead, ia_link) {
				mifp = ia->ia_ifp;
				if (!(mifp->if_flags & IFF_LOOPBACK) &&
				     (mifp->if_flags & IFF_MULTICAST)) {
					ifp = mifp;
					break;
				}
			}
			IN_IFADDR_RUNLOCK();
		}
	}

	return (ifp);
}

/*
 * Join an IPv4 multicast group, possibly with a source.
 */
static int
inp_join_group(struct inpcb *inp, struct sockopt *sopt)
{
	struct group_source_req		 gsr;
	sockunion_t			*gsa, *ssa;
	struct ifnet			*ifp;
	struct in_mfilter		*imf;
	struct ip_moptions		*imo;
	struct in_multi			*inm;
	struct in_msource		*lims;
	size_t				 idx;
	int				 error, is_new;

	ifp = NULL;
	imf = NULL;
	lims = NULL;
	error = 0;
	is_new = 0;

	memset(&gsr, 0, sizeof(struct group_source_req));
	gsa = (sockunion_t *)&gsr.gsr_group;
	gsa->ss.ss_family = AF_UNSPEC;
	ssa = (sockunion_t *)&gsr.gsr_source;
	ssa->ss.ss_family = AF_UNSPEC;

	switch (sopt->sopt_name) {
	case IP_ADD_MEMBERSHIP:
	case IP_ADD_SOURCE_MEMBERSHIP: {
		struct ip_mreq_source	 mreqs;

		if (sopt->sopt_name == IP_ADD_MEMBERSHIP) {
			error = sooptcopyin(sopt, &mreqs,
			    sizeof(struct ip_mreq),
			    sizeof(struct ip_mreq));
			/*
			 * Do argument switcharoo from ip_mreq into
			 * ip_mreq_source to avoid using two instances.
			 */
			mreqs.imr_interface = mreqs.imr_sourceaddr;
			mreqs.imr_sourceaddr.s_addr = INADDR_ANY;
		} else if (sopt->sopt_name == IP_ADD_SOURCE_MEMBERSHIP) {
			error = sooptcopyin(sopt, &mreqs,
			    sizeof(struct ip_mreq_source),
			    sizeof(struct ip_mreq_source));
		}
		if (error)
			return (error);

		gsa->sin.sin_family = AF_INET;
		gsa->sin.sin_len = sizeof(struct sockaddr_in);
		gsa->sin.sin_addr = mreqs.imr_multiaddr;

		if (sopt->sopt_name == IP_ADD_SOURCE_MEMBERSHIP) {
			ssa->sin.sin_family = AF_INET;
			ssa->sin.sin_len = sizeof(struct sockaddr_in);
			ssa->sin.sin_addr = mreqs.imr_sourceaddr;
		}

		if (!IN_MULTICAST(ntohl(gsa->sin.sin_addr.s_addr)))
			return (EINVAL);

		ifp = inp_lookup_mcast_ifp(inp, &gsa->sin,
		    mreqs.imr_interface);
		CTR3(KTR_IGMPV3, "%s: imr_interface = %s, ifp = %p",
		    __func__, inet_ntoa(mreqs.imr_interface), ifp);
		break;
	}

	case MCAST_JOIN_GROUP:
	case MCAST_JOIN_SOURCE_GROUP:
		if (sopt->sopt_name == MCAST_JOIN_GROUP) {
			error = sooptcopyin(sopt, &gsr,
			    sizeof(struct group_req),
			    sizeof(struct group_req));
		} else if (sopt->sopt_name == MCAST_JOIN_SOURCE_GROUP) {
			error = sooptcopyin(sopt, &gsr,
			    sizeof(struct group_source_req),
			    sizeof(struct group_source_req));
		}
		if (error)
			return (error);

		if (gsa->sin.sin_family != AF_INET ||
		    gsa->sin.sin_len != sizeof(struct sockaddr_in))
			return (EINVAL);

		/*
		 * Overwrite the port field if present, as the sockaddr
		 * being copied in may be matched with a binary comparison.
		 */
		gsa->sin.sin_port = 0;
		if (sopt->sopt_name == MCAST_JOIN_SOURCE_GROUP) {
			if (ssa->sin.sin_family != AF_INET ||
			    ssa->sin.sin_len != sizeof(struct sockaddr_in))
				return (EINVAL);
			ssa->sin.sin_port = 0;
		}

		if (!IN_MULTICAST(ntohl(gsa->sin.sin_addr.s_addr)))
			return (EINVAL);

		if (gsr.gsr_interface == 0 || V_if_index < gsr.gsr_interface)
			return (EADDRNOTAVAIL);
		ifp = ifnet_byindex(gsr.gsr_interface);
		break;

	default:
		CTR2(KTR_IGMPV3, "%s: unknown sopt_name %d",
		    __func__, sopt->sopt_name);
		return (EOPNOTSUPP);
		break;
	}

	if (ifp == NULL || (ifp->if_flags & IFF_MULTICAST) == 0)
		return (EADDRNOTAVAIL);

	imo = inp_findmoptions(inp);
	idx = imo_match_group(imo, ifp, &gsa->sa);
	if (idx == -1) {
		is_new = 1;
	} else {
		inm = imo->imo_membership[idx];
		imf = &imo->imo_mfilters[idx];
		if (ssa->ss.ss_family != AF_UNSPEC) {
			/*
			 * MCAST_JOIN_SOURCE_GROUP on an exclusive membership
			 * is an error. On an existing inclusive membership,
			 * it just adds the source to the filter list.
			 */
			if (imf->imf_st[1] != MCAST_INCLUDE) {
				error = EINVAL;
				goto out_inp_locked;
			}
			/*
			 * Throw out duplicates.
			 *
			 * XXX FIXME: This makes a naive assumption that
			 * even if entries exist for *ssa in this imf,
			 * they will be rejected as dupes, even if they
			 * are not valid in the current mode (in-mode).
			 *
			 * in_msource is transactioned just as for anything
			 * else in SSM -- but note naive use of inm_graft()
			 * below for allocating new filter entries.
			 *
			 * This is only an issue if someone mixes the
			 * full-state SSM API with the delta-based API,
			 * which is discouraged in the relevant RFCs.
			 */
			lims = imo_match_source(imo, idx, &ssa->sa);
			if (lims != NULL /*&&
			    lims->imsl_st[1] == MCAST_INCLUDE*/) {
				error = EADDRNOTAVAIL;
				goto out_inp_locked;
			}
		} else {
			/*
			 * MCAST_JOIN_GROUP alone, on any existing membership,
			 * is rejected, to stop the same inpcb tying up
			 * multiple refs to the in_multi.
			 * On an existing inclusive membership, this is also
			 * an error; if you want to change filter mode,
			 * you must use the userland API setsourcefilter().
			 * XXX We don't reject this for imf in UNDEFINED
			 * state at t1, because allocation of a filter
			 * is atomic with allocation of a membership.
			 */
			error = EINVAL;
			goto out_inp_locked;
		}
	}

	/*
	 * Begin state merge transaction at socket layer.
	 */
	INP_WLOCK_ASSERT(inp);

	if (is_new) {
		if (imo->imo_num_memberships == imo->imo_max_memberships) {
			error = imo_grow(imo);
			if (error)
				goto out_inp_locked;
		}
		/*
		 * Allocate the new slot upfront so we can deal with
		 * grafting the new source filter in same code path
		 * as for join-source on existing membership.
		 */
		idx = imo->imo_num_memberships;
		imo->imo_membership[idx] = NULL;
		imo->imo_num_memberships++;
		KASSERT(imo->imo_mfilters != NULL,
		    ("%s: imf_mfilters vector was not allocated", __func__));
		imf = &imo->imo_mfilters[idx];
		KASSERT(RB_EMPTY(&imf->imf_sources),
		    ("%s: imf_sources not empty", __func__));
	}

	/*
	 * Graft new source into filter list for this inpcb's
	 * membership of the group. The in_multi may not have
	 * been allocated yet if this is a new membership, however,
	 * the in_mfilter slot will be allocated and must be initialized.
	 *
	 * Note: Grafting of exclusive mode filters doesn't happen
	 * in this path.
	 * XXX: Should check for non-NULL lims (node exists but may
	 * not be in-mode) for interop with full-state API.
	 */
	if (ssa->ss.ss_family != AF_UNSPEC) {
		/* Membership starts in IN mode */
		if (is_new) {
			CTR1(KTR_IGMPV3, "%s: new join w/source", __func__);
			imf_init(imf, MCAST_UNDEFINED, MCAST_INCLUDE);
		} else {
			CTR2(KTR_IGMPV3, "%s: %s source", __func__, "allow");
		}
		lims = imf_graft(imf, MCAST_INCLUDE, &ssa->sin);
		if (lims == NULL) {
			CTR1(KTR_IGMPV3, "%s: merge imf state failed",
			    __func__);
			error = ENOMEM;
			goto out_imo_free;
		}
	} else {
		/* No address specified; Membership starts in EX mode */
		if (is_new) {
			CTR1(KTR_IGMPV3, "%s: new join w/o source", __func__);
			imf_init(imf, MCAST_UNDEFINED, MCAST_EXCLUDE);
		}
	}

	/*
	 * Begin state merge transaction at IGMP layer.
	 */
	IN_MULTI_LOCK();

	if (is_new) {
		error = in_joingroup_locked(ifp, &gsa->sin.sin_addr, imf,
		    &inm);
		if (error)
			goto out_imo_free;
		imo->imo_membership[idx] = inm;
	} else {
		CTR1(KTR_IGMPV3, "%s: merge inm state", __func__);
		error = inm_merge(inm, imf);
		if (error) {
			CTR1(KTR_IGMPV3, "%s: failed to merge inm state",
			    __func__);
			goto out_imf_rollback;
		}
		CTR1(KTR_IGMPV3, "%s: doing igmp downcall", __func__);
		error = igmp_change_state(inm);
		if (error) {
			CTR1(KTR_IGMPV3, "%s: failed igmp downcall",
			    __func__);
			goto out_imf_rollback;
		}
	}

	IN_MULTI_UNLOCK();

out_imf_rollback:
	INP_WLOCK_ASSERT(inp);
	if (error) {
		imf_rollback(imf);
		if (is_new)
			imf_purge(imf);
		else
			imf_reap(imf);
	} else {
		imf_commit(imf);
	}

out_imo_free:
	if (error && is_new) {
		imo->imo_membership[idx] = NULL;
		--imo->imo_num_memberships;
	}

out_inp_locked:
	INP_WUNLOCK(inp);
	return (error);
}

/*
 * Leave an IPv4 multicast group on an inpcb, possibly with a source.
 */
static int
inp_leave_group(struct inpcb *inp, struct sockopt *sopt)
{
	struct group_source_req		 gsr;
	struct ip_mreq_source		 mreqs;
	sockunion_t			*gsa, *ssa;
	struct ifnet			*ifp;
	struct in_mfilter		*imf;
	struct ip_moptions		*imo;
	struct in_msource		*ims;
	struct in_multi			*inm;
	size_t				 idx;
	int				 error, is_final;

	ifp = NULL;
	error = 0;
	is_final = 1;

	memset(&gsr, 0, sizeof(struct group_source_req));
	gsa = (sockunion_t *)&gsr.gsr_group;
	gsa->ss.ss_family = AF_UNSPEC;
	ssa = (sockunion_t *)&gsr.gsr_source;
	ssa->ss.ss_family = AF_UNSPEC;

	switch (sopt->sopt_name) {
	case IP_DROP_MEMBERSHIP:
	case IP_DROP_SOURCE_MEMBERSHIP:
		if (sopt->sopt_name == IP_DROP_MEMBERSHIP) {
			error = sooptcopyin(sopt, &mreqs,
			    sizeof(struct ip_mreq),
			    sizeof(struct ip_mreq));
			/*
			 * Swap interface and sourceaddr arguments,
			 * as ip_mreq and ip_mreq_source are laid
			 * out differently.
			 */
			mreqs.imr_interface = mreqs.imr_sourceaddr;
			mreqs.imr_sourceaddr.s_addr = INADDR_ANY;
		} else if (sopt->sopt_name == IP_DROP_SOURCE_MEMBERSHIP) {
			error = sooptcopyin(sopt, &mreqs,
			    sizeof(struct ip_mreq_source),
			    sizeof(struct ip_mreq_source));
		}
		if (error)
			return (error);

		gsa->sin.sin_family = AF_INET;
		gsa->sin.sin_len = sizeof(struct sockaddr_in);
		gsa->sin.sin_addr = mreqs.imr_multiaddr;

		if (sopt->sopt_name == IP_DROP_SOURCE_MEMBERSHIP) {
			ssa->sin.sin_family = AF_INET;
			ssa->sin.sin_len = sizeof(struct sockaddr_in);
			ssa->sin.sin_addr = mreqs.imr_sourceaddr;
		}

		if (!in_nullhost(gsa->sin.sin_addr))
			INADDR_TO_IFP(mreqs.imr_interface, ifp);

		CTR3(KTR_IGMPV3, "%s: imr_interface = %s, ifp = %p",
		    __func__, inet_ntoa(mreqs.imr_interface), ifp);

		break;

	case MCAST_LEAVE_GROUP:
	case MCAST_LEAVE_SOURCE_GROUP:
		if (sopt->sopt_name == MCAST_LEAVE_GROUP) {
			error = sooptcopyin(sopt, &gsr,
			    sizeof(struct group_req),
			    sizeof(struct group_req));
		} else if (sopt->sopt_name == MCAST_LEAVE_SOURCE_GROUP) {
			error = sooptcopyin(sopt, &gsr,
			    sizeof(struct group_source_req),
			    sizeof(struct group_source_req));
		}
		if (error)
			return (error);

		if (gsa->sin.sin_family != AF_INET ||
		    gsa->sin.sin_len != sizeof(struct sockaddr_in))
			return (EINVAL);

		if (sopt->sopt_name == MCAST_LEAVE_SOURCE_GROUP) {
			if (ssa->sin.sin_family != AF_INET ||
			    ssa->sin.sin_len != sizeof(struct sockaddr_in))
				return (EINVAL);
		}

		if (gsr.gsr_interface == 0 || V_if_index < gsr.gsr_interface)
			return (EADDRNOTAVAIL);

		ifp = ifnet_byindex(gsr.gsr_interface);
		break;

	default:
		CTR2(KTR_IGMPV3, "%s: unknown sopt_name %d",
		    __func__, sopt->sopt_name);
		return (EOPNOTSUPP);
		break;
	}

	if (!IN_MULTICAST(ntohl(gsa->sin.sin_addr.s_addr)))
		return (EINVAL);

	if (ifp == NULL)
		return (EADDRNOTAVAIL);

	/*
	 * Find the membership in the membership array.
	 */
	imo = inp_findmoptions(inp);
	idx = imo_match_group(imo, ifp, &gsa->sa);
	if (idx == -1) {
		error = EADDRNOTAVAIL;
		goto out_inp_locked;
	}
	inm = imo->imo_membership[idx];
	imf = &imo->imo_mfilters[idx];

	if (ssa->ss.ss_family != AF_UNSPEC)
		is_final = 0;

	/*
	 * Begin state merge transaction at socket layer.
	 */
	INP_WLOCK_ASSERT(inp);

	/*
	 * If we were instructed only to leave a given source, do so.
	 * MCAST_LEAVE_SOURCE_GROUP is only valid for inclusive memberships.
	 */
	if (is_final) {
		imf_leave(imf);
	} else {
		if (imf->imf_st[0] == MCAST_EXCLUDE) {
			error = EADDRNOTAVAIL;
			goto out_inp_locked;
		}
		ims = imo_match_source(imo, idx, &ssa->sa);
		if (ims == NULL) {
			CTR3(KTR_IGMPV3, "%s: source %s %spresent", __func__,
			    inet_ntoa(ssa->sin.sin_addr), "not ");
			error = EADDRNOTAVAIL;
			goto out_inp_locked;
		}
		CTR2(KTR_IGMPV3, "%s: %s source", __func__, "block");
		error = imf_prune(imf, &ssa->sin);
		if (error) {
			CTR1(KTR_IGMPV3, "%s: merge imf state failed",
			    __func__);
			goto out_inp_locked;
		}
	}

	/*
	 * Begin state merge transaction at IGMP layer.
	 */
	IN_MULTI_LOCK();

	if (is_final) {
		/*
		 * Give up the multicast address record to which
		 * the membership points.
		 */
		(void)in_leavegroup_locked(inm, imf);
	} else {
		CTR1(KTR_IGMPV3, "%s: merge inm state", __func__);
		error = inm_merge(inm, imf);
		if (error) {
			CTR1(KTR_IGMPV3, "%s: failed to merge inm state",
			    __func__);
			goto out_imf_rollback;
		}

		CTR1(KTR_IGMPV3, "%s: doing igmp downcall", __func__);
		error = igmp_change_state(inm);
		if (error) {
			CTR1(KTR_IGMPV3, "%s: failed igmp downcall",
			    __func__);
		}
	}

	IN_MULTI_UNLOCK();

out_imf_rollback:
	if (error)
		imf_rollback(imf);
	else
		imf_commit(imf);

	imf_reap(imf);

	if (is_final) {
		/* Remove the gap in the membership and filter array. */
		for (++idx; idx < imo->imo_num_memberships; ++idx) {
			imo->imo_membership[idx-1] = imo->imo_membership[idx];
			imo->imo_mfilters[idx-1] = imo->imo_mfilters[idx];
		}
		imo->imo_num_memberships--;
	}

out_inp_locked:
	INP_WUNLOCK(inp);
	return (error);
}

/*
 * Select the interface for transmitting IPv4 multicast datagrams.
 *
 * Either an instance of struct in_addr or an instance of struct ip_mreqn
 * may be passed to this socket option. An address of INADDR_ANY or an
 * interface index of 0 is used to remove a previous selection.
 * When no interface is selected, one is chosen for every send.
 */
static int
inp_set_multicast_if(struct inpcb *inp, struct sockopt *sopt)
{
	struct in_addr		 addr;
	struct ip_mreqn		 mreqn;
	struct ifnet		*ifp;
	struct ip_moptions	*imo;
	int			 error;

	if (sopt->sopt_valsize == sizeof(struct ip_mreqn)) {
		/*
		 * An interface index was specified using the
		 * Linux-derived ip_mreqn structure.
		 */
		error = sooptcopyin(sopt, &mreqn, sizeof(struct ip_mreqn),
		    sizeof(struct ip_mreqn));
		if (error)
			return (error);

		if (mreqn.imr_ifindex < 0 || V_if_index < mreqn.imr_ifindex)
			return (EINVAL);

		if (mreqn.imr_ifindex == 0) {
			ifp = NULL;
		} else {
			ifp = ifnet_byindex(mreqn.imr_ifindex);
			if (ifp == NULL)
				return (EADDRNOTAVAIL);
		}
	} else {
		/*
		 * An interface was specified by IPv4 address.
		 * This is the traditional BSD usage.
		 */
		error = sooptcopyin(sopt, &addr, sizeof(struct in_addr),
		    sizeof(struct in_addr));
		if (error)
			return (error);
		if (in_nullhost(addr)) {
			ifp = NULL;
		} else {
			INADDR_TO_IFP(addr, ifp);
			if (ifp == NULL)
				return (EADDRNOTAVAIL);
		}
		CTR3(KTR_IGMPV3, "%s: ifp = %p, addr = %s", __func__, ifp,
		    inet_ntoa(addr));
	}

	/* Reject interfaces which do not support multicast. */
	if (ifp != NULL && (ifp->if_flags & IFF_MULTICAST) == 0)
		return (EOPNOTSUPP);

	imo = inp_findmoptions(inp);
	imo->imo_multicast_ifp = ifp;
	imo->imo_multicast_addr.s_addr = INADDR_ANY;
	INP_WUNLOCK(inp);

	return (0);
}

/*
 * Atomically set source filters on a socket for an IPv4 multicast group.
 *
 * SMPng: NOTE: Potentially calls malloc(M_WAITOK) with Giant held.
 */
static int
inp_set_source_filters(struct inpcb *inp, struct sockopt *sopt)
{
	struct __msfilterreq	 msfr;
	sockunion_t		*gsa;
	struct ifnet		*ifp;
	struct in_mfilter	*imf;
	struct ip_moptions	*imo;
	struct in_multi		*inm;
	size_t			 idx;
	int			 error;

	error = sooptcopyin(sopt, &msfr, sizeof(struct __msfilterreq),
	    sizeof(struct __msfilterreq));
	if (error)
		return (error);

	if (msfr.msfr_nsrcs > in_mcast_maxsocksrc)
		return (ENOBUFS);

	if ((msfr.msfr_fmode != MCAST_EXCLUDE &&
	     msfr.msfr_fmode != MCAST_INCLUDE))
		return (EINVAL);

	if (msfr.msfr_group.ss_family != AF_INET ||
	    msfr.msfr_group.ss_len != sizeof(struct sockaddr_in))
		return (EINVAL);

	gsa = (sockunion_t *)&msfr.msfr_group;
	if (!IN_MULTICAST(ntohl(gsa->sin.sin_addr.s_addr)))
		return (EINVAL);

	gsa->sin.sin_port = 0;	/* ignore port */

	if (msfr.msfr_ifindex == 0 || V_if_index < msfr.msfr_ifindex)
		return (EADDRNOTAVAIL);

	ifp = ifnet_byindex(msfr.msfr_ifindex);
	if (ifp == NULL)
		return (EADDRNOTAVAIL);

	/*
	 * Take the INP write lock.
	 * Check if this socket is a member of this group.
	 */
	imo = inp_findmoptions(inp);
	idx = imo_match_group(imo, ifp, &gsa->sa);
	if (idx == -1 || imo->imo_mfilters == NULL) {
		error = EADDRNOTAVAIL;
		goto out_inp_locked;
	}
	inm = imo->imo_membership[idx];
	imf = &imo->imo_mfilters[idx];

	/*
	 * Begin state merge transaction at socket layer.
	 */
	INP_WLOCK_ASSERT(inp);

	imf->imf_st[1] = msfr.msfr_fmode;

	/*
	 * Apply any new source filters, if present.
	 * Make a copy of the user-space source vector so
	 * that we may copy them with a single copyin. This
	 * allows us to deal with page faults up-front.
	 */
	if (msfr.msfr_nsrcs > 0) {
		struct in_msource	*lims;
		struct sockaddr_in	*psin;
		struct sockaddr_storage	*kss, *pkss;
		int			 i;

		INP_WUNLOCK(inp);
 
		CTR2(KTR_IGMPV3, "%s: loading %lu source list entries",
		    __func__, (unsigned long)msfr.msfr_nsrcs);
		kss = malloc(sizeof(struct sockaddr_storage) * msfr.msfr_nsrcs,
		    M_TEMP, M_WAITOK);
		error = copyin(msfr.msfr_srcs, kss,
		    sizeof(struct sockaddr_storage) * msfr.msfr_nsrcs);
		if (error) {
			free(kss, M_TEMP);
			return (error);
		}

		INP_WLOCK(inp);

		/*
		 * Mark all source filters as UNDEFINED at t1.
		 * Restore new group filter mode, as imf_leave()
		 * will set it to INCLUDE.
		 */
		imf_leave(imf);
		imf->imf_st[1] = msfr.msfr_fmode;

		/*
		 * Update socket layer filters at t1, lazy-allocating
		 * new entries. This saves a bunch of memory at the
		 * cost of one RB_FIND() per source entry; duplicate
		 * entries in the msfr_nsrcs vector are ignored.
		 * If we encounter an error, rollback transaction.
		 *
		 * XXX This too could be replaced with a set-symmetric
		 * difference like loop to avoid walking from root
		 * every time, as the key space is common.
		 */
		for (i = 0, pkss = kss; i < msfr.msfr_nsrcs; i++, pkss++) {
			psin = (struct sockaddr_in *)pkss;
			if (psin->sin_family != AF_INET) {
				error = EAFNOSUPPORT;
				break;
			}
			if (psin->sin_len != sizeof(struct sockaddr_in)) {
				error = EINVAL;
				break;
			}
			error = imf_get_source(imf, psin, &lims);
			if (error)
				break;
			lims->imsl_st[1] = imf->imf_st[1];
		}
		free(kss, M_TEMP);
	}

	if (error)
		goto out_imf_rollback;

	INP_WLOCK_ASSERT(inp);
	IN_MULTI_LOCK();

	/*
	 * Begin state merge transaction at IGMP layer.
	 */
	CTR1(KTR_IGMPV3, "%s: merge inm state", __func__);
	error = inm_merge(inm, imf);
	if (error) {
		CTR1(KTR_IGMPV3, "%s: failed to merge inm state", __func__);
		goto out_imf_rollback;
	}

	CTR1(KTR_IGMPV3, "%s: doing igmp downcall", __func__);
	error = igmp_change_state(inm);
	if (error)
		CTR1(KTR_IGMPV3, "%s: failed igmp downcall", __func__);

	IN_MULTI_UNLOCK();

out_imf_rollback:
	if (error)
		imf_rollback(imf);
	else
		imf_commit(imf);

	imf_reap(imf);

out_inp_locked:
	INP_WUNLOCK(inp);
	return (error);
}

/*
 * Set the IP multicast options in response to user setsockopt().
 *
 * Many of the socket options handled in this function duplicate the
 * functionality of socket options in the regular unicast API. However,
 * it is not possible to merge the duplicate code, because the idempotence
 * of the IPv4 multicast part of the BSD Sockets API must be preserved;
 * the effects of these options must be treated as separate and distinct.
 *
 * SMPng: XXX: Unlocked read of inp_socket believed OK.
 * FUTURE: The IP_MULTICAST_VIF option may be eliminated if MROUTING
 * is refactored to no longer use vifs.
 */
int
inp_setmoptions(struct inpcb *inp, struct sockopt *sopt)
{
	struct ip_moptions	*imo;
	int			 error;

	error = 0;

	/*
	 * If socket is neither of type SOCK_RAW or SOCK_DGRAM,
	 * or is a divert socket, reject it.
	 */
	if (inp->inp_socket->so_proto->pr_protocol == IPPROTO_DIVERT ||
	    (inp->inp_socket->so_proto->pr_type != SOCK_RAW &&
	     inp->inp_socket->so_proto->pr_type != SOCK_DGRAM))
		return (EOPNOTSUPP);

	switch (sopt->sopt_name) {
	case IP_MULTICAST_VIF: {
		int vifi;
		/*
		 * Select a multicast VIF for transmission.
		 * Only useful if multicast forwarding is active.
		 */
		if (legal_vif_num == NULL) {
			error = EOPNOTSUPP;
			break;
		}
		error = sooptcopyin(sopt, &vifi, sizeof(int), sizeof(int));
		if (error)
			break;
		if (!legal_vif_num(vifi) && (vifi != -1)) {
			error = EINVAL;
			break;
		}
		imo = inp_findmoptions(inp);
		imo->imo_multicast_vif = vifi;
		INP_WUNLOCK(inp);
		break;
	}

	case IP_MULTICAST_IF:
		error = inp_set_multicast_if(inp, sopt);
		break;

	case IP_MULTICAST_TTL: {
		u_char ttl;

		/*
		 * Set the IP time-to-live for outgoing multicast packets.
		 * The original multicast API required a char argument,
		 * which is inconsistent with the rest of the socket API.
		 * We allow either a char or an int.
		 */
		if (sopt->sopt_valsize == sizeof(u_char)) {
			error = sooptcopyin(sopt, &ttl, sizeof(u_char),
			    sizeof(u_char));
			if (error)
				break;
		} else {
			u_int ittl;

			error = sooptcopyin(sopt, &ittl, sizeof(u_int),
			    sizeof(u_int));
			if (error)
				break;
			if (ittl > 255) {
				error = EINVAL;
				break;
			}
			ttl = (u_char)ittl;
		}
		imo = inp_findmoptions(inp);
		imo->imo_multicast_ttl = ttl;
		INP_WUNLOCK(inp);
		break;
	}

	case IP_MULTICAST_LOOP: {
		u_char loop;

		/*
		 * Set the loopback flag for outgoing multicast packets.
		 * Must be zero or one.  The original multicast API required a
		 * char argument, which is inconsistent with the rest
		 * of the socket API.  We allow either a char or an int.
		 */
		if (sopt->sopt_valsize == sizeof(u_char)) {
			error = sooptcopyin(sopt, &loop, sizeof(u_char),
			    sizeof(u_char));
			if (error)
				break;
		} else {
			u_int iloop;

			error = sooptcopyin(sopt, &iloop, sizeof(u_int),
					    sizeof(u_int));
			if (error)
				break;
			loop = (u_char)iloop;
		}
		imo = inp_findmoptions(inp);
		imo->imo_multicast_loop = !!loop;
		INP_WUNLOCK(inp);
		break;
	}

	case IP_ADD_MEMBERSHIP:
	case IP_ADD_SOURCE_MEMBERSHIP:
	case MCAST_JOIN_GROUP:
	case MCAST_JOIN_SOURCE_GROUP:
		error = inp_join_group(inp, sopt);
		break;

	case IP_DROP_MEMBERSHIP:
	case IP_DROP_SOURCE_MEMBERSHIP:
	case MCAST_LEAVE_GROUP:
	case MCAST_LEAVE_SOURCE_GROUP:
		error = inp_leave_group(inp, sopt);
		break;

	case IP_BLOCK_SOURCE:
	case IP_UNBLOCK_SOURCE:
	case MCAST_BLOCK_SOURCE:
	case MCAST_UNBLOCK_SOURCE:
		error = inp_block_unblock_source(inp, sopt);
		break;

	case IP_MSFILTER:
		error = inp_set_source_filters(inp, sopt);
		break;

	default:
		error = EOPNOTSUPP;
		break;
	}

	INP_UNLOCK_ASSERT(inp);

	return (error);
}

/*
 * Expose IGMP's multicast filter mode and source list(s) to userland,
 * keyed by (ifindex, group).
 * The filter mode is written out as a uint32_t, followed by
 * 0..n of struct in_addr.
 * For use by ifmcstat(8).
 * SMPng: NOTE: unlocked read of ifindex space.
 */
static int
sysctl_ip_mcast_filters(SYSCTL_HANDLER_ARGS)
{
	struct in_addr			 src, group;
	struct ifnet			*ifp;
	struct ifmultiaddr		*ifma;
	struct in_multi			*inm;
	struct ip_msource		*ims;
	int				*name;
	int				 retval;
	u_int				 namelen;
	uint32_t			 fmode, ifindex;

	name = (int *)arg1;
	namelen = arg2;

	if (req->newptr != NULL)
		return (EPERM);

	if (namelen != 2)
		return (EINVAL);

	ifindex = name[0];
	if (ifindex <= 0 || ifindex > V_if_index) {
		CTR2(KTR_IGMPV3, "%s: ifindex %u out of range",
		    __func__, ifindex);
		return (ENOENT);
	}

	group.s_addr = name[1];
	if (!IN_MULTICAST(ntohl(group.s_addr))) {
		CTR2(KTR_IGMPV3, "%s: group %s is not multicast",
		    __func__, inet_ntoa(group));
		return (EINVAL);
	}

	ifp = ifnet_byindex(ifindex);
	if (ifp == NULL) {
		CTR2(KTR_IGMPV3, "%s: no ifp for ifindex %u",
		    __func__, ifindex);
		return (ENOENT);
	}

	retval = sysctl_wire_old_buffer(req,
	    sizeof(uint32_t) + (in_mcast_maxgrpsrc * sizeof(struct in_addr)));
	if (retval)
		return (retval);

	IN_MULTI_LOCK();

	IF_ADDR_LOCK(ifp);
	TAILQ_FOREACH(ifma, &ifp->if_multiaddrs, ifma_link) {
		if (ifma->ifma_addr->sa_family != AF_INET ||
		    ifma->ifma_protospec == NULL)
			continue;
		inm = (struct in_multi *)ifma->ifma_protospec;
		if (!in_hosteq(inm->inm_addr, group))
			continue;
		fmode = inm->inm_st[1].iss_fmode;
		retval = SYSCTL_OUT(req, &fmode, sizeof(uint32_t));
		if (retval != 0)
			break;
		RB_FOREACH(ims, ip_msource_tree, &inm->inm_srcs) {
#ifdef KTR
			struct in_addr ina;
			ina.s_addr = htonl(ims->ims_haddr);
			CTR2(KTR_IGMPV3, "%s: visit node %s", __func__,
			    inet_ntoa(ina));
#endif
			/*
			 * Only copy-out sources which are in-mode.
			 */
			if (fmode != ims_get_mode(inm, ims, 1)) {
				CTR1(KTR_IGMPV3, "%s: skip non-in-mode",
				    __func__);
				continue;
			}
			src.s_addr = htonl(ims->ims_haddr);
			retval = SYSCTL_OUT(req, &src, sizeof(struct in_addr));
			if (retval != 0)
				break;
		}
	}
	IF_ADDR_UNLOCK(ifp);

	IN_MULTI_UNLOCK();

	return (retval);
}

#ifdef KTR

static const char *inm_modestrs[] = { "un", "in", "ex" };

static const char *
inm_mode_str(const int mode)
{

	if (mode >= MCAST_UNDEFINED && mode <= MCAST_EXCLUDE)
		return (inm_modestrs[mode]);
	return ("??");
}

static const char *inm_statestrs[] = {
	"not-member",
	"silent",
	"idle",
	"lazy",
	"sleeping",
	"awakening",
	"query-pending",
	"sg-query-pending",
	"leaving"
};

static const char *
inm_state_str(const int state)
{

	if (state >= IGMP_NOT_MEMBER && state <= IGMP_LEAVING_MEMBER)
		return (inm_statestrs[state]);
	return ("??");
}

/*
 * Dump an in_multi structure to the console.
 */
void
inm_print(const struct in_multi *inm)
{
	int t;

	if ((ktr_mask & KTR_IGMPV3) == 0)
		return;

	printf("%s: --- begin inm %p ---\n", __func__, inm);
	printf("addr %s ifp %p(%s) ifma %p\n",
	    inet_ntoa(inm->inm_addr),
	    inm->inm_ifp,
	    inm->inm_ifp->if_xname,
	    inm->inm_ifma);
	printf("timer %u state %s refcount %u scq.len %u\n",
	    inm->inm_timer,
	    inm_state_str(inm->inm_state),
	    inm->inm_refcount,
	    inm->inm_scq.ifq_len);
	printf("igi %p nsrc %lu sctimer %u scrv %u\n",
	    inm->inm_igi,
	    inm->inm_nsrc,
	    inm->inm_sctimer,
	    inm->inm_scrv);
	for (t = 0; t < 2; t++) {
		printf("t%d: fmode %s asm %u ex %u in %u rec %u\n", t,
		    inm_mode_str(inm->inm_st[t].iss_fmode),
		    inm->inm_st[t].iss_asm,
		    inm->inm_st[t].iss_ex,
		    inm->inm_st[t].iss_in,
		    inm->inm_st[t].iss_rec);
	}
	printf("%s: --- end inm %p ---\n", __func__, inm);
}

#else /* !KTR */

void
inm_print(const struct in_multi *inm)
{

}

#endif /* KTR */

RB_GENERATE(ip_msource_tree, ip_msource, ims_link, ip_msource_cmp);
OpenPOWER on IntegriCloud