summaryrefslogtreecommitdiff
path: root/drivers/infiniband/core/device.c
blob: b69e2c4e4d2a40443966836e602641a2767cf796 (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
/*
 * Copyright (c) 2004 Topspin Communications.  All rights reserved.
 * Copyright (c) 2005 Sun Microsystems, Inc. All rights reserved.
 *
 * This software is available to you under a choice of one of two
 * licenses.  You may choose to be licensed under the terms of the GNU
 * General Public License (GPL) Version 2, available from the file
 * COPYING in the main directory of this source tree, or the
 * OpenIB.org BSD license below:
 *
 *     Redistribution and use in source and binary forms, with or
 *     without modification, are permitted provided that the following
 *     conditions are met:
 *
 *      - Redistributions of source code must retain the above
 *        copyright notice, this list of conditions and the following
 *        disclaimer.
 *
 *      - 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.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
 * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
 * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
 * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
 * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 * SOFTWARE.
 */

#include <linux/module.h>
#include <linux/string.h>
#include <linux/errno.h>
#include <linux/kernel.h>
#include <linux/slab.h>
#include <linux/init.h>
#include <linux/netdevice.h>
#include <net/net_namespace.h>
#include <linux/security.h>
#include <linux/notifier.h>
#include <linux/hashtable.h>
#include <rdma/rdma_netlink.h>
#include <rdma/ib_addr.h>
#include <rdma/ib_cache.h>
#include <rdma/rdma_counter.h>

#include "core_priv.h"
#include "restrack.h"

MODULE_AUTHOR("Roland Dreier");
MODULE_DESCRIPTION("core kernel InfiniBand API");
MODULE_LICENSE("Dual BSD/GPL");

struct workqueue_struct *ib_comp_wq;
struct workqueue_struct *ib_comp_unbound_wq;
struct workqueue_struct *ib_wq;
EXPORT_SYMBOL_GPL(ib_wq);
static struct workqueue_struct *ib_unreg_wq;

/*
 * Each of the three rwsem locks (devices, clients, client_data) protects the
 * xarray of the same name. Specifically it allows the caller to assert that
 * the MARK will/will not be changing under the lock, and for devices and
 * clients, that the value in the xarray is still a valid pointer. Change of
 * the MARK is linked to the object state, so holding the lock and testing the
 * MARK also asserts that the contained object is in a certain state.
 *
 * This is used to build a two stage register/unregister flow where objects
 * can continue to be in the xarray even though they are still in progress to
 * register/unregister.
 *
 * The xarray itself provides additional locking, and restartable iteration,
 * which is also relied on.
 *
 * Locks should not be nested, with the exception of client_data, which is
 * allowed to nest under the read side of the other two locks.
 *
 * The devices_rwsem also protects the device name list, any change or
 * assignment of device name must also hold the write side to guarantee unique
 * names.
 */

/*
 * devices contains devices that have had their names assigned. The
 * devices may not be registered. Users that care about the registration
 * status need to call ib_device_try_get() on the device to ensure it is
 * registered, and keep it registered, for the required duration.
 *
 */
static DEFINE_XARRAY_FLAGS(devices, XA_FLAGS_ALLOC);
static DECLARE_RWSEM(devices_rwsem);
#define DEVICE_REGISTERED XA_MARK_1

static u32 highest_client_id;
#define CLIENT_REGISTERED XA_MARK_1
static DEFINE_XARRAY_FLAGS(clients, XA_FLAGS_ALLOC);
static DECLARE_RWSEM(clients_rwsem);

static void ib_client_put(struct ib_client *client)
{
	if (refcount_dec_and_test(&client->uses))
		complete(&client->uses_zero);
}

/*
 * If client_data is registered then the corresponding client must also still
 * be registered.
 */
#define CLIENT_DATA_REGISTERED XA_MARK_1

unsigned int rdma_dev_net_id;

/*
 * A list of net namespaces is maintained in an xarray. This is necessary
 * because we can't get the locking right using the existing net ns list. We
 * would require a init_net callback after the list is updated.
 */
static DEFINE_XARRAY_FLAGS(rdma_nets, XA_FLAGS_ALLOC);
/*
 * rwsem to protect accessing the rdma_nets xarray entries.
 */
static DECLARE_RWSEM(rdma_nets_rwsem);

bool ib_devices_shared_netns = true;
module_param_named(netns_mode, ib_devices_shared_netns, bool, 0444);
MODULE_PARM_DESC(netns_mode,
		 "Share device among net namespaces; default=1 (shared)");
/**
 * rdma_dev_access_netns() - Return whether an rdma device can be accessed
 *			     from a specified net namespace or not.
 * @dev:	Pointer to rdma device which needs to be checked
 * @net:	Pointer to net namesapce for which access to be checked
 *
 * When the rdma device is in shared mode, it ignores the net namespace.
 * When the rdma device is exclusive to a net namespace, rdma device net
 * namespace is checked against the specified one.
 */
bool rdma_dev_access_netns(const struct ib_device *dev, const struct net *net)
{
	return (ib_devices_shared_netns ||
		net_eq(read_pnet(&dev->coredev.rdma_net), net));
}
EXPORT_SYMBOL(rdma_dev_access_netns);

/*
 * xarray has this behavior where it won't iterate over NULL values stored in
 * allocated arrays.  So we need our own iterator to see all values stored in
 * the array. This does the same thing as xa_for_each except that it also
 * returns NULL valued entries if the array is allocating. Simplified to only
 * work on simple xarrays.
 */
static void *xan_find_marked(struct xarray *xa, unsigned long *indexp,
			     xa_mark_t filter)
{
	XA_STATE(xas, xa, *indexp);
	void *entry;

	rcu_read_lock();
	do {
		entry = xas_find_marked(&xas, ULONG_MAX, filter);
		if (xa_is_zero(entry))
			break;
	} while (xas_retry(&xas, entry));
	rcu_read_unlock();

	if (entry) {
		*indexp = xas.xa_index;
		if (xa_is_zero(entry))
			return NULL;
		return entry;
	}
	return XA_ERROR(-ENOENT);
}
#define xan_for_each_marked(xa, index, entry, filter)                          \
	for (index = 0, entry = xan_find_marked(xa, &(index), filter);         \
	     !xa_is_err(entry);                                                \
	     (index)++, entry = xan_find_marked(xa, &(index), filter))

/* RCU hash table mapping netdevice pointers to struct ib_port_data */
static DEFINE_SPINLOCK(ndev_hash_lock);
static DECLARE_HASHTABLE(ndev_hash, 5);

static void free_netdevs(struct ib_device *ib_dev);
static void ib_unregister_work(struct work_struct *work);
static void __ib_unregister_device(struct ib_device *device);
static int ib_security_change(struct notifier_block *nb, unsigned long event,
			      void *lsm_data);
static void ib_policy_change_task(struct work_struct *work);
static DECLARE_WORK(ib_policy_change_work, ib_policy_change_task);

static void __ibdev_printk(const char *level, const struct ib_device *ibdev,
			   struct va_format *vaf)
{
	if (ibdev && ibdev->dev.parent)
		dev_printk_emit(level[1] - '0',
				ibdev->dev.parent,
				"%s %s %s: %pV",
				dev_driver_string(ibdev->dev.parent),
				dev_name(ibdev->dev.parent),
				dev_name(&ibdev->dev),
				vaf);
	else if (ibdev)
		printk("%s%s: %pV",
		       level, dev_name(&ibdev->dev), vaf);
	else
		printk("%s(NULL ib_device): %pV", level, vaf);
}

void ibdev_printk(const char *level, const struct ib_device *ibdev,
		  const char *format, ...)
{
	struct va_format vaf;
	va_list args;

	va_start(args, format);

	vaf.fmt = format;
	vaf.va = &args;

	__ibdev_printk(level, ibdev, &vaf);

	va_end(args);
}
EXPORT_SYMBOL(ibdev_printk);

#define define_ibdev_printk_level(func, level)                  \
void func(const struct ib_device *ibdev, const char *fmt, ...)  \
{                                                               \
	struct va_format vaf;                                   \
	va_list args;                                           \
								\
	va_start(args, fmt);                                    \
								\
	vaf.fmt = fmt;                                          \
	vaf.va = &args;                                         \
								\
	__ibdev_printk(level, ibdev, &vaf);                     \
								\
	va_end(args);                                           \
}                                                               \
EXPORT_SYMBOL(func);

define_ibdev_printk_level(ibdev_emerg, KERN_EMERG);
define_ibdev_printk_level(ibdev_alert, KERN_ALERT);
define_ibdev_printk_level(ibdev_crit, KERN_CRIT);
define_ibdev_printk_level(ibdev_err, KERN_ERR);
define_ibdev_printk_level(ibdev_warn, KERN_WARNING);
define_ibdev_printk_level(ibdev_notice, KERN_NOTICE);
define_ibdev_printk_level(ibdev_info, KERN_INFO);

static struct notifier_block ibdev_lsm_nb = {
	.notifier_call = ib_security_change,
};

static int rdma_dev_change_netns(struct ib_device *device, struct net *cur_net,
				 struct net *net);

/* Pointer to the RCU head at the start of the ib_port_data array */
struct ib_port_data_rcu {
	struct rcu_head rcu_head;
	struct ib_port_data pdata[];
};

static void ib_device_check_mandatory(struct ib_device *device)
{
#define IB_MANDATORY_FUNC(x) { offsetof(struct ib_device_ops, x), #x }
	static const struct {
		size_t offset;
		char  *name;
	} mandatory_table[] = {
		IB_MANDATORY_FUNC(query_device),
		IB_MANDATORY_FUNC(query_port),
		IB_MANDATORY_FUNC(alloc_pd),
		IB_MANDATORY_FUNC(dealloc_pd),
		IB_MANDATORY_FUNC(create_qp),
		IB_MANDATORY_FUNC(modify_qp),
		IB_MANDATORY_FUNC(destroy_qp),
		IB_MANDATORY_FUNC(post_send),
		IB_MANDATORY_FUNC(post_recv),
		IB_MANDATORY_FUNC(create_cq),
		IB_MANDATORY_FUNC(destroy_cq),
		IB_MANDATORY_FUNC(poll_cq),
		IB_MANDATORY_FUNC(req_notify_cq),
		IB_MANDATORY_FUNC(get_dma_mr),
		IB_MANDATORY_FUNC(reg_user_mr),
		IB_MANDATORY_FUNC(dereg_mr),
		IB_MANDATORY_FUNC(get_port_immutable)
	};
	int i;

	device->kverbs_provider = true;
	for (i = 0; i < ARRAY_SIZE(mandatory_table); ++i) {
		if (!*(void **) ((void *) &device->ops +
				 mandatory_table[i].offset)) {
			device->kverbs_provider = false;
			break;
		}
	}
}

/*
 * Caller must perform ib_device_put() to return the device reference count
 * when ib_device_get_by_index() returns valid device pointer.
 */
struct ib_device *ib_device_get_by_index(const struct net *net, u32 index)
{
	struct ib_device *device;

	down_read(&devices_rwsem);
	device = xa_load(&devices, index);
	if (device) {
		if (!rdma_dev_access_netns(device, net)) {
			device = NULL;
			goto out;
		}

		if (!ib_device_try_get(device))
			device = NULL;
	}
out:
	up_read(&devices_rwsem);
	return device;
}

/**
 * ib_device_put - Release IB device reference
 * @device: device whose reference to be released
 *
 * ib_device_put() releases reference to the IB device to allow it to be
 * unregistered and eventually free.
 */
void ib_device_put(struct ib_device *device)
{
	if (refcount_dec_and_test(&device->refcount))
		complete(&device->unreg_completion);
}
EXPORT_SYMBOL(ib_device_put);

static struct ib_device *__ib_device_get_by_name(const char *name)
{
	struct ib_device *device;
	unsigned long index;

	xa_for_each (&devices, index, device)
		if (!strcmp(name, dev_name(&device->dev)))
			return device;

	return NULL;
}

/**
 * ib_device_get_by_name - Find an IB device by name
 * @name: The name to look for
 * @driver_id: The driver ID that must match (RDMA_DRIVER_UNKNOWN matches all)
 *
 * Find and hold an ib_device by its name. The caller must call
 * ib_device_put() on the returned pointer.
 */
struct ib_device *ib_device_get_by_name(const char *name,
					enum rdma_driver_id driver_id)
{
	struct ib_device *device;

	down_read(&devices_rwsem);
	device = __ib_device_get_by_name(name);
	if (device && driver_id != RDMA_DRIVER_UNKNOWN &&
	    device->ops.driver_id != driver_id)
		device = NULL;

	if (device) {
		if (!ib_device_try_get(device))
			device = NULL;
	}
	up_read(&devices_rwsem);
	return device;
}
EXPORT_SYMBOL(ib_device_get_by_name);

static int rename_compat_devs(struct ib_device *device)
{
	struct ib_core_device *cdev;
	unsigned long index;
	int ret = 0;

	mutex_lock(&device->compat_devs_mutex);
	xa_for_each (&device->compat_devs, index, cdev) {
		ret = device_rename(&cdev->dev, dev_name(&device->dev));
		if (ret) {
			dev_warn(&cdev->dev,
				 "Fail to rename compatdev to new name %s\n",
				 dev_name(&device->dev));
			break;
		}
	}
	mutex_unlock(&device->compat_devs_mutex);
	return ret;
}

int ib_device_rename(struct ib_device *ibdev, const char *name)
{
	unsigned long index;
	void *client_data;
	int ret;

	down_write(&devices_rwsem);
	if (!strcmp(name, dev_name(&ibdev->dev))) {
		up_write(&devices_rwsem);
		return 0;
	}

	if (__ib_device_get_by_name(name)) {
		up_write(&devices_rwsem);
		return -EEXIST;
	}

	ret = device_rename(&ibdev->dev, name);
	if (ret) {
		up_write(&devices_rwsem);
		return ret;
	}

	strscpy(ibdev->name, name, IB_DEVICE_NAME_MAX);
	ret = rename_compat_devs(ibdev);

	downgrade_write(&devices_rwsem);
	down_read(&ibdev->client_data_rwsem);
	xan_for_each_marked(&ibdev->client_data, index, client_data,
			    CLIENT_DATA_REGISTERED) {
		struct ib_client *client = xa_load(&clients, index);

		if (!client || !client->rename)
			continue;

		client->rename(ibdev, client_data);
	}
	up_read(&ibdev->client_data_rwsem);
	up_read(&devices_rwsem);
	return 0;
}

int ib_device_set_dim(struct ib_device *ibdev, u8 use_dim)
{
	if (use_dim > 1)
		return -EINVAL;
	ibdev->use_cq_dim = use_dim;

	return 0;
}

static int alloc_name(struct ib_device *ibdev, const char *name)
{
	struct ib_device *device;
	unsigned long index;
	struct ida inuse;
	int rc;
	int i;

	lockdep_assert_held_write(&devices_rwsem);
	ida_init(&inuse);
	xa_for_each (&devices, index, device) {
		char buf[IB_DEVICE_NAME_MAX];

		if (sscanf(dev_name(&device->dev), name, &i) != 1)
			continue;
		if (i < 0 || i >= INT_MAX)
			continue;
		snprintf(buf, sizeof buf, name, i);
		if (strcmp(buf, dev_name(&device->dev)) != 0)
			continue;

		rc = ida_alloc_range(&inuse, i, i, GFP_KERNEL);
		if (rc < 0)
			goto out;
	}

	rc = ida_alloc(&inuse, GFP_KERNEL);
	if (rc < 0)
		goto out;

	rc = dev_set_name(&ibdev->dev, name, rc);
out:
	ida_destroy(&inuse);
	return rc;
}

static void ib_device_release(struct device *device)
{
	struct ib_device *dev = container_of(device, struct ib_device, dev);

	free_netdevs(dev);
	WARN_ON(refcount_read(&dev->refcount));
	if (dev->hw_stats_data)
		ib_device_release_hw_stats(dev->hw_stats_data);
	if (dev->port_data) {
		ib_cache_release_one(dev);
		ib_security_release_port_pkey_list(dev);
		rdma_counter_release(dev);
		kfree_rcu(container_of(dev->port_data, struct ib_port_data_rcu,
				       pdata[0]),
			  rcu_head);
	}

	mutex_destroy(&dev->unregistration_lock);
	mutex_destroy(&dev->compat_devs_mutex);

	xa_destroy(&dev->compat_devs);
	xa_destroy(&dev->client_data);
	kfree_rcu(dev, rcu_head);
}

static int ib_device_uevent(struct device *device,
			    struct kobj_uevent_env *env)
{
	if (add_uevent_var(env, "NAME=%s", dev_name(device)))
		return -ENOMEM;

	/*
	 * It would be nice to pass the node GUID with the event...
	 */

	return 0;
}

static const void *net_namespace(struct device *d)
{
	struct ib_core_device *coredev =
			container_of(d, struct ib_core_device, dev);

	return read_pnet(&coredev->rdma_net);
}

static struct class ib_class = {
	.name    = "infiniband",
	.dev_release = ib_device_release,
	.dev_uevent = ib_device_uevent,
	.ns_type = &net_ns_type_operations,
	.namespace = net_namespace,
};

static void rdma_init_coredev(struct ib_core_device *coredev,
			      struct ib_device *dev, struct net *net)
{
	/* This BUILD_BUG_ON is intended to catch layout change
	 * of union of ib_core_device and device.
	 * dev must be the first element as ib_core and providers
	 * driver uses it. Adding anything in ib_core_device before
	 * device will break this assumption.
	 */
	BUILD_BUG_ON(offsetof(struct ib_device, coredev.dev) !=
		     offsetof(struct ib_device, dev));

	coredev->dev.class = &ib_class;
	coredev->dev.groups = dev->groups;
	device_initialize(&coredev->dev);
	coredev->owner = dev;
	INIT_LIST_HEAD(&coredev->port_list);
	write_pnet(&coredev->rdma_net, net);
}

/**
 * _ib_alloc_device - allocate an IB device struct
 * @size:size of structure to allocate
 *
 * Low-level drivers should use ib_alloc_device() to allocate &struct
 * ib_device.  @size is the size of the structure to be allocated,
 * including any private data used by the low-level driver.
 * ib_dealloc_device() must be used to free structures allocated with
 * ib_alloc_device().
 */
struct ib_device *_ib_alloc_device(size_t size)
{
	struct ib_device *device;
	unsigned int i;

	if (WARN_ON(size < sizeof(struct ib_device)))
		return NULL;

	device = kzalloc(size, GFP_KERNEL);
	if (!device)
		return NULL;

	if (rdma_restrack_init(device)) {
		kfree(device);
		return NULL;
	}

	rdma_init_coredev(&device->coredev, device, &init_net);

	INIT_LIST_HEAD(&device->event_handler_list);
	spin_lock_init(&device->qp_open_list_lock);
	init_rwsem(&device->event_handler_rwsem);
	mutex_init(&device->unregistration_lock);
	/*
	 * client_data needs to be alloc because we don't want our mark to be
	 * destroyed if the user stores NULL in the client data.
	 */
	xa_init_flags(&device->client_data, XA_FLAGS_ALLOC);
	init_rwsem(&device->client_data_rwsem);
	xa_init_flags(&device->compat_devs, XA_FLAGS_ALLOC);
	mutex_init(&device->compat_devs_mutex);
	init_completion(&device->unreg_completion);
	INIT_WORK(&device->unregistration_work, ib_unregister_work);

	spin_lock_init(&device->cq_pools_lock);
	for (i = 0; i < ARRAY_SIZE(device->cq_pools); i++)
		INIT_LIST_HEAD(&device->cq_pools[i]);

	rwlock_init(&device->cache_lock);

	device->uverbs_cmd_mask =
		BIT_ULL(IB_USER_VERBS_CMD_ALLOC_MW) |
		BIT_ULL(IB_USER_VERBS_CMD_ALLOC_PD) |
		BIT_ULL(IB_USER_VERBS_CMD_ATTACH_MCAST) |
		BIT_ULL(IB_USER_VERBS_CMD_CLOSE_XRCD) |
		BIT_ULL(IB_USER_VERBS_CMD_CREATE_AH) |
		BIT_ULL(IB_USER_VERBS_CMD_CREATE_COMP_CHANNEL) |
		BIT_ULL(IB_USER_VERBS_CMD_CREATE_CQ) |
		BIT_ULL(IB_USER_VERBS_CMD_CREATE_QP) |
		BIT_ULL(IB_USER_VERBS_CMD_CREATE_SRQ) |
		BIT_ULL(IB_USER_VERBS_CMD_CREATE_XSRQ) |
		BIT_ULL(IB_USER_VERBS_CMD_DEALLOC_MW) |
		BIT_ULL(IB_USER_VERBS_CMD_DEALLOC_PD) |
		BIT_ULL(IB_USER_VERBS_CMD_DEREG_MR) |
		BIT_ULL(IB_USER_VERBS_CMD_DESTROY_AH) |
		BIT_ULL(IB_USER_VERBS_CMD_DESTROY_CQ) |
		BIT_ULL(IB_USER_VERBS_CMD_DESTROY_QP) |
		BIT_ULL(IB_USER_VERBS_CMD_DESTROY_SRQ) |
		BIT_ULL(IB_USER_VERBS_CMD_DETACH_MCAST) |
		BIT_ULL(IB_USER_VERBS_CMD_GET_CONTEXT) |
		BIT_ULL(IB_USER_VERBS_CMD_MODIFY_QP) |
		BIT_ULL(IB_USER_VERBS_CMD_MODIFY_SRQ) |
		BIT_ULL(IB_USER_VERBS_CMD_OPEN_QP) |
		BIT_ULL(IB_USER_VERBS_CMD_OPEN_XRCD) |
		BIT_ULL(IB_USER_VERBS_CMD_QUERY_DEVICE) |
		BIT_ULL(IB_USER_VERBS_CMD_QUERY_PORT) |
		BIT_ULL(IB_USER_VERBS_CMD_QUERY_QP) |
		BIT_ULL(IB_USER_VERBS_CMD_QUERY_SRQ) |
		BIT_ULL(IB_USER_VERBS_CMD_REG_MR) |
		BIT_ULL(IB_USER_VERBS_CMD_REREG_MR) |
		BIT_ULL(IB_USER_VERBS_CMD_RESIZE_CQ);
	return device;
}
EXPORT_SYMBOL(_ib_alloc_device);

/**
 * ib_dealloc_device - free an IB device struct
 * @device:structure to free
 *
 * Free a structure allocated with ib_alloc_device().
 */
void ib_dealloc_device(struct ib_device *device)
{
	if (device->ops.dealloc_driver)
		device->ops.dealloc_driver(device);

	/*
	 * ib_unregister_driver() requires all devices to remain in the xarray
	 * while their ops are callable. The last op we call is dealloc_driver
	 * above.  This is needed to create a fence on op callbacks prior to
	 * allowing the driver module to unload.
	 */
	down_write(&devices_rwsem);
	if (xa_load(&devices, device->index) == device)
		xa_erase(&devices, device->index);
	up_write(&devices_rwsem);

	/* Expedite releasing netdev references */
	free_netdevs(device);

	WARN_ON(!xa_empty(&device->compat_devs));
	WARN_ON(!xa_empty(&device->client_data));
	WARN_ON(refcount_read(&device->refcount));
	rdma_restrack_clean(device);
	/* Balances with device_initialize */
	put_device(&device->dev);
}
EXPORT_SYMBOL(ib_dealloc_device);

/*
 * add_client_context() and remove_client_context() must be safe against
 * parallel calls on the same device - registration/unregistration of both the
 * device and client can be occurring in parallel.
 *
 * The routines need to be a fence, any caller must not return until the add
 * or remove is fully completed.
 */
static int add_client_context(struct ib_device *device,
			      struct ib_client *client)
{
	int ret = 0;

	if (!device->kverbs_provider && !client->no_kverbs_req)
		return 0;

	down_write(&device->client_data_rwsem);
	/*
	 * So long as the client is registered hold both the client and device
	 * unregistration locks.
	 */
	if (!refcount_inc_not_zero(&client->uses))
		goto out_unlock;
	refcount_inc(&device->refcount);

	/*
	 * Another caller to add_client_context got here first and has already
	 * completely initialized context.
	 */
	if (xa_get_mark(&device->client_data, client->client_id,
		    CLIENT_DATA_REGISTERED))
		goto out;

	ret = xa_err(xa_store(&device->client_data, client->client_id, NULL,
			      GFP_KERNEL));
	if (ret)
		goto out;
	downgrade_write(&device->client_data_rwsem);
	if (client->add) {
		if (client->add(device)) {
			/*
			 * If a client fails to add then the error code is
			 * ignored, but we won't call any more ops on this
			 * client.
			 */
			xa_erase(&device->client_data, client->client_id);
			up_read(&device->client_data_rwsem);
			ib_device_put(device);
			ib_client_put(client);
			return 0;
		}
	}

	/* Readers shall not see a client until add has been completed */
	xa_set_mark(&device->client_data, client->client_id,
		    CLIENT_DATA_REGISTERED);
	up_read(&device->client_data_rwsem);
	return 0;

out:
	ib_device_put(device);
	ib_client_put(client);
out_unlock:
	up_write(&device->client_data_rwsem);
	return ret;
}

static void remove_client_context(struct ib_device *device,
				  unsigned int client_id)
{
	struct ib_client *client;
	void *client_data;

	down_write(&device->client_data_rwsem);
	if (!xa_get_mark(&device->client_data, client_id,
			 CLIENT_DATA_REGISTERED)) {
		up_write(&device->client_data_rwsem);
		return;
	}
	client_data = xa_load(&device->client_data, client_id);
	xa_clear_mark(&device->client_data, client_id, CLIENT_DATA_REGISTERED);
	client = xa_load(&clients, client_id);
	up_write(&device->client_data_rwsem);

	/*
	 * Notice we cannot be holding any exclusive locks when calling the
	 * remove callback as the remove callback can recurse back into any
	 * public functions in this module and thus try for any locks those
	 * functions take.
	 *
	 * For this reason clients and drivers should not call the
	 * unregistration functions will holdling any locks.
	 */
	if (client->remove)
		client->remove(device, client_data);

	xa_erase(&device->client_data, client_id);
	ib_device_put(device);
	ib_client_put(client);
}

static int alloc_port_data(struct ib_device *device)
{
	struct ib_port_data_rcu *pdata_rcu;
	u32 port;

	if (device->port_data)
		return 0;

	/* This can only be called once the physical port range is defined */
	if (WARN_ON(!device->phys_port_cnt))
		return -EINVAL;

	/* Reserve U32_MAX so the logic to go over all the ports is sane */
	if (WARN_ON(device->phys_port_cnt == U32_MAX))
		return -EINVAL;

	/*
	 * device->port_data is indexed directly by the port number to make
	 * access to this data as efficient as possible.
	 *
	 * Therefore port_data is declared as a 1 based array with potential
	 * empty slots at the beginning.
	 */
	pdata_rcu = kzalloc(struct_size(pdata_rcu, pdata,
					rdma_end_port(device) + 1),
			    GFP_KERNEL);
	if (!pdata_rcu)
		return -ENOMEM;
	/*
	 * The rcu_head is put in front of the port data array and the stored
	 * pointer is adjusted since we never need to see that member until
	 * kfree_rcu.
	 */
	device->port_data = pdata_rcu->pdata;

	rdma_for_each_port (device, port) {
		struct ib_port_data *pdata = &device->port_data[port];

		pdata->ib_dev = device;
		spin_lock_init(&pdata->pkey_list_lock);
		INIT_LIST_HEAD(&pdata->pkey_list);
		spin_lock_init(&pdata->netdev_lock);
		INIT_HLIST_NODE(&pdata->ndev_hash_link);
	}
	return 0;
}

static int verify_immutable(const struct ib_device *dev, u32 port)
{
	return WARN_ON(!rdma_cap_ib_mad(dev, port) &&
			    rdma_max_mad_size(dev, port) != 0);
}

static int setup_port_data(struct ib_device *device)
{
	u32 port;
	int ret;

	ret = alloc_port_data(device);
	if (ret)
		return ret;

	rdma_for_each_port (device, port) {
		struct ib_port_data *pdata = &device->port_data[port];

		ret = device->ops.get_port_immutable(device, port,
						     &pdata->immutable);
		if (ret)
			return ret;

		if (verify_immutable(device, port))
			return -EINVAL;
	}
	return 0;
}

/**
 * ib_port_immutable_read() - Read rdma port's immutable data
 * @dev: IB device
 * @port: port number whose immutable data to read. It starts with index 1 and
 *        valid upto including rdma_end_port().
 */
const struct ib_port_immutable*
ib_port_immutable_read(struct ib_device *dev, unsigned int port)
{
	WARN_ON(!rdma_is_port_valid(dev, port));
	return &dev->port_data[port].immutable;
}
EXPORT_SYMBOL(ib_port_immutable_read);

void ib_get_device_fw_str(struct ib_device *dev, char *str)
{
	if (dev->ops.get_dev_fw_str)
		dev->ops.get_dev_fw_str(dev, str);
	else
		str[0] = '\0';
}
EXPORT_SYMBOL(ib_get_device_fw_str);

static void ib_policy_change_task(struct work_struct *work)
{
	struct ib_device *dev;
	unsigned long index;

	down_read(&devices_rwsem);
	xa_for_each_marked (&devices, index, dev, DEVICE_REGISTERED) {
		unsigned int i;

		rdma_for_each_port (dev, i) {
			u64 sp;
			ib_get_cached_subnet_prefix(dev, i, &sp);
			ib_security_cache_change(dev, i, sp);
		}
	}
	up_read(&devices_rwsem);
}

static int ib_security_change(struct notifier_block *nb, unsigned long event,
			      void *lsm_data)
{
	if (event != LSM_POLICY_CHANGE)
		return NOTIFY_DONE;

	schedule_work(&ib_policy_change_work);
	ib_mad_agent_security_change();

	return NOTIFY_OK;
}

static void compatdev_release(struct device *dev)
{
	struct ib_core_device *cdev =
		container_of(dev, struct ib_core_device, dev);

	kfree(cdev);
}

static int add_one_compat_dev(struct ib_device *device,
			      struct rdma_dev_net *rnet)
{
	struct ib_core_device *cdev;
	int ret;

	lockdep_assert_held(&rdma_nets_rwsem);
	if (!ib_devices_shared_netns)
		return 0;

	/*
	 * Create and add compat device in all namespaces other than where it
	 * is currently bound to.
	 */
	if (net_eq(read_pnet(&rnet->net),
		   read_pnet(&device->coredev.rdma_net)))
		return 0;

	/*
	 * The first of init_net() or ib_register_device() to take the
	 * compat_devs_mutex wins and gets to add the device. Others will wait
	 * for completion here.
	 */
	mutex_lock(&device->compat_devs_mutex);
	cdev = xa_load(&device->compat_devs, rnet->id);
	if (cdev) {
		ret = 0;
		goto done;
	}
	ret = xa_reserve(&device->compat_devs, rnet->id, GFP_KERNEL);
	if (ret)
		goto done;

	cdev = kzalloc(sizeof(*cdev), GFP_KERNEL);
	if (!cdev) {
		ret = -ENOMEM;
		goto cdev_err;
	}

	cdev->dev.parent = device->dev.parent;
	rdma_init_coredev(cdev, device, read_pnet(&rnet->net));
	cdev->dev.release = compatdev_release;
	ret = dev_set_name(&cdev->dev, "%s", dev_name(&device->dev));
	if (ret)
		goto add_err;

	ret = device_add(&cdev->dev);
	if (ret)
		goto add_err;
	ret = ib_setup_port_attrs(cdev);
	if (ret)
		goto port_err;

	ret = xa_err(xa_store(&device->compat_devs, rnet->id,
			      cdev, GFP_KERNEL));
	if (ret)
		goto insert_err;

	mutex_unlock(&device->compat_devs_mutex);
	return 0;

insert_err:
	ib_free_port_attrs(cdev);
port_err:
	device_del(&cdev->dev);
add_err:
	put_device(&cdev->dev);
cdev_err:
	xa_release(&device->compat_devs, rnet->id);
done:
	mutex_unlock(&device->compat_devs_mutex);
	return ret;
}

static void remove_one_compat_dev(struct ib_device *device, u32 id)
{
	struct ib_core_device *cdev;

	mutex_lock(&device->compat_devs_mutex);
	cdev = xa_erase(&device->compat_devs, id);
	mutex_unlock(&device->compat_devs_mutex);
	if (cdev) {
		ib_free_port_attrs(cdev);
		device_del(&cdev->dev);
		put_device(&cdev->dev);
	}
}

static void remove_compat_devs(struct ib_device *device)
{
	struct ib_core_device *cdev;
	unsigned long index;

	xa_for_each (&device->compat_devs, index, cdev)
		remove_one_compat_dev(device, index);
}

static int add_compat_devs(struct ib_device *device)
{
	struct rdma_dev_net *rnet;
	unsigned long index;
	int ret = 0;

	lockdep_assert_held(&devices_rwsem);

	down_read(&rdma_nets_rwsem);
	xa_for_each (&rdma_nets, index, rnet) {
		ret = add_one_compat_dev(device, rnet);
		if (ret)
			break;
	}
	up_read(&rdma_nets_rwsem);
	return ret;
}

static void remove_all_compat_devs(void)
{
	struct ib_compat_device *cdev;
	struct ib_device *dev;
	unsigned long index;

	down_read(&devices_rwsem);
	xa_for_each (&devices, index, dev) {
		unsigned long c_index = 0;

		/* Hold nets_rwsem so that any other thread modifying this
		 * system param can sync with this thread.
		 */
		down_read(&rdma_nets_rwsem);
		xa_for_each (&dev->compat_devs, c_index, cdev)
			remove_one_compat_dev(dev, c_index);
		up_read(&rdma_nets_rwsem);
	}
	up_read(&devices_rwsem);
}

static int add_all_compat_devs(void)
{
	struct rdma_dev_net *rnet;
	struct ib_device *dev;
	unsigned long index;
	int ret = 0;

	down_read(&devices_rwsem);
	xa_for_each_marked (&devices, index, dev, DEVICE_REGISTERED) {
		unsigned long net_index = 0;

		/* Hold nets_rwsem so that any other thread modifying this
		 * system param can sync with this thread.
		 */
		down_read(&rdma_nets_rwsem);
		xa_for_each (&rdma_nets, net_index, rnet) {
			ret = add_one_compat_dev(dev, rnet);
			if (ret)
				break;
		}
		up_read(&rdma_nets_rwsem);
	}
	up_read(&devices_rwsem);
	if (ret)
		remove_all_compat_devs();
	return ret;
}

int rdma_compatdev_set(u8 enable)
{
	struct rdma_dev_net *rnet;
	unsigned long index;
	int ret = 0;

	down_write(&rdma_nets_rwsem);
	if (ib_devices_shared_netns == enable) {
		up_write(&rdma_nets_rwsem);
		return 0;
	}

	/* enable/disable of compat devices is not supported
	 * when more than default init_net exists.
	 */
	xa_for_each (&rdma_nets, index, rnet) {
		ret++;
		break;
	}
	if (!ret)
		ib_devices_shared_netns = enable;
	up_write(&rdma_nets_rwsem);
	if (ret)
		return -EBUSY;

	if (enable)
		ret = add_all_compat_devs();
	else
		remove_all_compat_devs();
	return ret;
}

static void rdma_dev_exit_net(struct net *net)
{
	struct rdma_dev_net *rnet = rdma_net_to_dev_net(net);
	struct ib_device *dev;
	unsigned long index;
	int ret;

	down_write(&rdma_nets_rwsem);
	/*
	 * Prevent the ID from being re-used and hide the id from xa_for_each.
	 */
	ret = xa_err(xa_store(&rdma_nets, rnet->id, NULL, GFP_KERNEL));
	WARN_ON(ret);
	up_write(&rdma_nets_rwsem);

	down_read(&devices_rwsem);
	xa_for_each (&devices, index, dev) {
		get_device(&dev->dev);
		/*
		 * Release the devices_rwsem so that pontentially blocking
		 * device_del, doesn't hold the devices_rwsem for too long.
		 */
		up_read(&devices_rwsem);

		remove_one_compat_dev(dev, rnet->id);

		/*
		 * If the real device is in the NS then move it back to init.
		 */
		rdma_dev_change_netns(dev, net, &init_net);

		put_device(&dev->dev);
		down_read(&devices_rwsem);
	}
	up_read(&devices_rwsem);

	rdma_nl_net_exit(rnet);
	xa_erase(&rdma_nets, rnet->id);
}

static __net_init int rdma_dev_init_net(struct net *net)
{
	struct rdma_dev_net *rnet = rdma_net_to_dev_net(net);
	unsigned long index;
	struct ib_device *dev;
	int ret;

	write_pnet(&rnet->net, net);

	ret = rdma_nl_net_init(rnet);
	if (ret)
		return ret;

	/* No need to create any compat devices in default init_net. */
	if (net_eq(net, &init_net))
		return 0;

	ret = xa_alloc(&rdma_nets, &rnet->id, rnet, xa_limit_32b, GFP_KERNEL);
	if (ret) {
		rdma_nl_net_exit(rnet);
		return ret;
	}

	down_read(&devices_rwsem);
	xa_for_each_marked (&devices, index, dev, DEVICE_REGISTERED) {
		/* Hold nets_rwsem so that netlink command cannot change
		 * system configuration for device sharing mode.
		 */
		down_read(&rdma_nets_rwsem);
		ret = add_one_compat_dev(dev, rnet);
		up_read(&rdma_nets_rwsem);
		if (ret)
			break;
	}
	up_read(&devices_rwsem);

	if (ret)
		rdma_dev_exit_net(net);

	return ret;
}

/*
 * Assign the unique string device name and the unique device index. This is
 * undone by ib_dealloc_device.
 */
static int assign_name(struct ib_device *device, const char *name)
{
	static u32 last_id;
	int ret;

	down_write(&devices_rwsem);
	/* Assign a unique name to the device */
	if (strchr(name, '%'))
		ret = alloc_name(device, name);
	else
		ret = dev_set_name(&device->dev, name);
	if (ret)
		goto out;

	if (__ib_device_get_by_name(dev_name(&device->dev))) {
		ret = -ENFILE;
		goto out;
	}
	strscpy(device->name, dev_name(&device->dev), IB_DEVICE_NAME_MAX);

	ret = xa_alloc_cyclic(&devices, &device->index, device, xa_limit_31b,
			&last_id, GFP_KERNEL);
	if (ret > 0)
		ret = 0;

out:
	up_write(&devices_rwsem);
	return ret;
}

/*
 * setup_device() allocates memory and sets up data that requires calling the
 * device ops, this is the only reason these actions are not done during
 * ib_alloc_device. It is undone by ib_dealloc_device().
 */
static int setup_device(struct ib_device *device)
{
	struct ib_udata uhw = {.outlen = 0, .inlen = 0};
	int ret;

	ib_device_check_mandatory(device);

	ret = setup_port_data(device);
	if (ret) {
		dev_warn(&device->dev, "Couldn't create per-port data\n");
		return ret;
	}

	memset(&device->attrs, 0, sizeof(device->attrs));
	ret = device->ops.query_device(device, &device->attrs, &uhw);
	if (ret) {
		dev_warn(&device->dev,
			 "Couldn't query the device attributes\n");
		return ret;
	}

	return 0;
}

static void disable_device(struct ib_device *device)
{
	u32 cid;

	WARN_ON(!refcount_read(&device->refcount));

	down_write(&devices_rwsem);
	xa_clear_mark(&devices, device->index, DEVICE_REGISTERED);
	up_write(&devices_rwsem);

	/*
	 * Remove clients in LIFO order, see assign_client_id. This could be
	 * more efficient if xarray learns to reverse iterate. Since no new
	 * clients can be added to this ib_device past this point we only need
	 * the maximum possible client_id value here.
	 */
	down_read(&clients_rwsem);
	cid = highest_client_id;
	up_read(&clients_rwsem);
	while (cid) {
		cid--;
		remove_client_context(device, cid);
	}

	ib_cq_pool_cleanup(device);

	/* Pairs with refcount_set in enable_device */
	ib_device_put(device);
	wait_for_completion(&device->unreg_completion);

	/*
	 * compat devices must be removed after device refcount drops to zero.
	 * Otherwise init_net() may add more compatdevs after removing compat
	 * devices and before device is disabled.
	 */
	remove_compat_devs(device);
}

/*
 * An enabled device is visible to all clients and to all the public facing
 * APIs that return a device pointer. This always returns with a new get, even
 * if it fails.
 */
static int enable_device_and_get(struct ib_device *device)
{
	struct ib_client *client;
	unsigned long index;
	int ret = 0;

	/*
	 * One ref belongs to the xa and the other belongs to this
	 * thread. This is needed to guard against parallel unregistration.
	 */
	refcount_set(&device->refcount, 2);
	down_write(&devices_rwsem);
	xa_set_mark(&devices, device->index, DEVICE_REGISTERED);

	/*
	 * By using downgrade_write() we ensure that no other thread can clear
	 * DEVICE_REGISTERED while we are completing the client setup.
	 */
	downgrade_write(&devices_rwsem);

	if (device->ops.enable_driver) {
		ret = device->ops.enable_driver(device);
		if (ret)
			goto out;
	}

	down_read(&clients_rwsem);
	xa_for_each_marked (&clients, index, client, CLIENT_REGISTERED) {
		ret = add_client_context(device, client);
		if (ret)
			break;
	}
	up_read(&clients_rwsem);
	if (!ret)
		ret = add_compat_devs(device);
out:
	up_read(&devices_rwsem);
	return ret;
}

static void prevent_dealloc_device(struct ib_device *ib_dev)
{
}

/**
 * ib_register_device - Register an IB device with IB core
 * @device: Device to register
 * @name: unique string device name. This may include a '%' which will
 * 	  cause a unique index to be added to the passed device name.
 * @dma_device: pointer to a DMA-capable device. If %NULL, then the IB
 *	        device will be used. In this case the caller should fully
 *		setup the ibdev for DMA. This usually means using dma_virt_ops.
 *
 * Low-level drivers use ib_register_device() to register their
 * devices with the IB core.  All registered clients will receive a
 * callback for each device that is added. @device must be allocated
 * with ib_alloc_device().
 *
 * If the driver uses ops.dealloc_driver and calls any ib_unregister_device()
 * asynchronously then the device pointer may become freed as soon as this
 * function returns.
 */
int ib_register_device(struct ib_device *device, const char *name,
		       struct device *dma_device)
{
	int ret;

	ret = assign_name(device, name);
	if (ret)
		return ret;

	/*
	 * If the caller does not provide a DMA capable device then the IB core
	 * will set up ib_sge and scatterlist structures that stash the kernel
	 * virtual address into the address field.
	 */
	WARN_ON(dma_device && !dma_device->dma_parms);
	device->dma_device = dma_device;

	ret = setup_device(device);
	if (ret)
		return ret;

	ret = ib_cache_setup_one(device);
	if (ret) {
		dev_warn(&device->dev,
			 "Couldn't set up InfiniBand P_Key/GID cache\n");
		return ret;
	}

	device->groups[0] = &ib_dev_attr_group;
	device->groups[1] = device->ops.device_group;
	ret = ib_setup_device_attrs(device);
	if (ret)
		goto cache_cleanup;

	ib_device_register_rdmacg(device);

	rdma_counter_init(device);

	/*
	 * Ensure that ADD uevent is not fired because it
	 * is too early amd device is not initialized yet.
	 */
	dev_set_uevent_suppress(&device->dev, true);
	ret = device_add(&device->dev);
	if (ret)
		goto cg_cleanup;

	ret = ib_setup_port_attrs(&device->coredev);
	if (ret) {
		dev_warn(&device->dev,
			 "Couldn't register device with driver model\n");
		goto dev_cleanup;
	}

	ret = enable_device_and_get(device);
	if (ret) {
		void (*dealloc_fn)(struct ib_device *);

		/*
		 * If we hit this error flow then we don't want to
		 * automatically dealloc the device since the caller is
		 * expected to call ib_dealloc_device() after
		 * ib_register_device() fails. This is tricky due to the
		 * possibility for a parallel unregistration along with this
		 * error flow. Since we have a refcount here we know any
		 * parallel flow is stopped in disable_device and will see the
		 * special dealloc_driver pointer, causing the responsibility to
		 * ib_dealloc_device() to revert back to this thread.
		 */
		dealloc_fn = device->ops.dealloc_driver;
		device->ops.dealloc_driver = prevent_dealloc_device;
		ib_device_put(device);
		__ib_unregister_device(device);
		device->ops.dealloc_driver = dealloc_fn;
		dev_set_uevent_suppress(&device->dev, false);
		return ret;
	}
	dev_set_uevent_suppress(&device->dev, false);
	/* Mark for userspace that device is ready */
	kobject_uevent(&device->dev.kobj, KOBJ_ADD);
	ib_device_put(device);

	return 0;

dev_cleanup:
	device_del(&device->dev);
cg_cleanup:
	dev_set_uevent_suppress(&device->dev, false);
	ib_device_unregister_rdmacg(device);
cache_cleanup:
	ib_cache_cleanup_one(device);
	return ret;
}
EXPORT_SYMBOL(ib_register_device);

/* Callers must hold a get on the device. */
static void __ib_unregister_device(struct ib_device *ib_dev)
{
	/*
	 * We have a registration lock so that all the calls to unregister are
	 * fully fenced, once any unregister returns the device is truely
	 * unregistered even if multiple callers are unregistering it at the
	 * same time. This also interacts with the registration flow and
	 * provides sane semantics if register and unregister are racing.
	 */
	mutex_lock(&ib_dev->unregistration_lock);
	if (!refcount_read(&ib_dev->refcount))
		goto out;

	disable_device(ib_dev);

	/* Expedite removing unregistered pointers from the hash table */
	free_netdevs(ib_dev);

	ib_free_port_attrs(&ib_dev->coredev);
	device_del(&ib_dev->dev);
	ib_device_unregister_rdmacg(ib_dev);
	ib_cache_cleanup_one(ib_dev);

	/*
	 * Drivers using the new flow may not call ib_dealloc_device except
	 * in error unwind prior to registration success.
	 */
	if (ib_dev->ops.dealloc_driver &&
	    ib_dev->ops.dealloc_driver != prevent_dealloc_device) {
		WARN_ON(kref_read(&ib_dev->dev.kobj.kref) <= 1);
		ib_dealloc_device(ib_dev);
	}
out:
	mutex_unlock(&ib_dev->unregistration_lock);
}

/**
 * ib_unregister_device - Unregister an IB device
 * @ib_dev: The device to unregister
 *
 * Unregister an IB device.  All clients will receive a remove callback.
 *
 * Callers should call this routine only once, and protect against races with
 * registration. Typically it should only be called as part of a remove
 * callback in an implementation of driver core's struct device_driver and
 * related.
 *
 * If ops.dealloc_driver is used then ib_dev will be freed upon return from
 * this function.
 */
void ib_unregister_device(struct ib_device *ib_dev)
{
	get_device(&ib_dev->dev);
	__ib_unregister_device(ib_dev);
	put_device(&ib_dev->dev);
}
EXPORT_SYMBOL(ib_unregister_device);

/**
 * ib_unregister_device_and_put - Unregister a device while holding a 'get'
 * @ib_dev: The device to unregister
 *
 * This is the same as ib_unregister_device(), except it includes an internal
 * ib_device_put() that should match a 'get' obtained by the caller.
 *
 * It is safe to call this routine concurrently from multiple threads while
 * holding the 'get'. When the function returns the device is fully
 * unregistered.
 *
 * Drivers using this flow MUST use the driver_unregister callback to clean up
 * their resources associated with the device and dealloc it.
 */
void ib_unregister_device_and_put(struct ib_device *ib_dev)
{
	WARN_ON(!ib_dev->ops.dealloc_driver);
	get_device(&ib_dev->dev);
	ib_device_put(ib_dev);
	__ib_unregister_device(ib_dev);
	put_device(&ib_dev->dev);
}
EXPORT_SYMBOL(ib_unregister_device_and_put);

/**
 * ib_unregister_driver - Unregister all IB devices for a driver
 * @driver_id: The driver to unregister
 *
 * This implements a fence for device unregistration. It only returns once all
 * devices associated with the driver_id have fully completed their
 * unregistration and returned from ib_unregister_device*().
 *
 * If device's are not yet unregistered it goes ahead and starts unregistering
 * them.
 *
 * This does not block creation of new devices with the given driver_id, that
 * is the responsibility of the caller.
 */
void ib_unregister_driver(enum rdma_driver_id driver_id)
{
	struct ib_device *ib_dev;
	unsigned long index;

	down_read(&devices_rwsem);
	xa_for_each (&devices, index, ib_dev) {
		if (ib_dev->ops.driver_id != driver_id)
			continue;

		get_device(&ib_dev->dev);
		up_read(&devices_rwsem);

		WARN_ON(!ib_dev->ops.dealloc_driver);
		__ib_unregister_device(ib_dev);

		put_device(&ib_dev->dev);
		down_read(&devices_rwsem);
	}
	up_read(&devices_rwsem);
}
EXPORT_SYMBOL(ib_unregister_driver);

static void ib_unregister_work(struct work_struct *work)
{
	struct ib_device *ib_dev =
		container_of(work, struct ib_device, unregistration_work);

	__ib_unregister_device(ib_dev);
	put_device(&ib_dev->dev);
}

/**
 * ib_unregister_device_queued - Unregister a device using a work queue
 * @ib_dev: The device to unregister
 *
 * This schedules an asynchronous unregistration using a WQ for the device. A
 * driver should use this to avoid holding locks while doing unregistration,
 * such as holding the RTNL lock.
 *
 * Drivers using this API must use ib_unregister_driver before module unload
 * to ensure that all scheduled unregistrations have completed.
 */
void ib_unregister_device_queued(struct ib_device *ib_dev)
{
	WARN_ON(!refcount_read(&ib_dev->refcount));
	WARN_ON(!ib_dev->ops.dealloc_driver);
	get_device(&ib_dev->dev);
	if (!queue_work(ib_unreg_wq, &ib_dev->unregistration_work))
		put_device(&ib_dev->dev);
}
EXPORT_SYMBOL(ib_unregister_device_queued);

/*
 * The caller must pass in a device that has the kref held and the refcount
 * released. If the device is in cur_net and still registered then it is moved
 * into net.
 */
static int rdma_dev_change_netns(struct ib_device *device, struct net *cur_net,
				 struct net *net)
{
	int ret2 = -EINVAL;
	int ret;

	mutex_lock(&device->unregistration_lock);

	/*
	 * If a device not under ib_device_get() or if the unregistration_lock
	 * is not held, the namespace can be changed, or it can be unregistered.
	 * Check again under the lock.
	 */
	if (refcount_read(&device->refcount) == 0 ||
	    !net_eq(cur_net, read_pnet(&device->coredev.rdma_net))) {
		ret = -ENODEV;
		goto out;
	}

	kobject_uevent(&device->dev.kobj, KOBJ_REMOVE);
	disable_device(device);

	/*
	 * At this point no one can be using the device, so it is safe to
	 * change the namespace.
	 */
	write_pnet(&device->coredev.rdma_net, net);

	down_read(&devices_rwsem);
	/*
	 * Currently rdma devices are system wide unique. So the device name
	 * is guaranteed free in the new namespace. Publish the new namespace
	 * at the sysfs level.
	 */
	ret = device_rename(&device->dev, dev_name(&device->dev));
	up_read(&devices_rwsem);
	if (ret) {
		dev_warn(&device->dev,
			 "%s: Couldn't rename device after namespace change\n",
			 __func__);
		/* Try and put things back and re-enable the device */
		write_pnet(&device->coredev.rdma_net, cur_net);
	}

	ret2 = enable_device_and_get(device);
	if (ret2) {
		/*
		 * This shouldn't really happen, but if it does, let the user
		 * retry at later point. So don't disable the device.
		 */
		dev_warn(&device->dev,
			 "%s: Couldn't re-enable device after namespace change\n",
			 __func__);
	}
	kobject_uevent(&device->dev.kobj, KOBJ_ADD);

	ib_device_put(device);
out:
	mutex_unlock(&device->unregistration_lock);
	if (ret)
		return ret;
	return ret2;
}

int ib_device_set_netns_put(struct sk_buff *skb,
			    struct ib_device *dev, u32 ns_fd)
{
	struct net *net;
	int ret;

	net = get_net_ns_by_fd(ns_fd);
	if (IS_ERR(net)) {
		ret = PTR_ERR(net);
		goto net_err;
	}

	if (!netlink_ns_capable(skb, net->user_ns, CAP_NET_ADMIN)) {
		ret = -EPERM;
		goto ns_err;
	}

	/*
	 * All the ib_clients, including uverbs, are reset when the namespace is
	 * changed and this cannot be blocked waiting for userspace to do
	 * something, so disassociation is mandatory.
	 */
	if (!dev->ops.disassociate_ucontext || ib_devices_shared_netns) {
		ret = -EOPNOTSUPP;
		goto ns_err;
	}

	get_device(&dev->dev);
	ib_device_put(dev);
	ret = rdma_dev_change_netns(dev, current->nsproxy->net_ns, net);
	put_device(&dev->dev);

	put_net(net);
	return ret;

ns_err:
	put_net(net);
net_err:
	ib_device_put(dev);
	return ret;
}

static struct pernet_operations rdma_dev_net_ops = {
	.init = rdma_dev_init_net,
	.exit = rdma_dev_exit_net,
	.id = &rdma_dev_net_id,
	.size = sizeof(struct rdma_dev_net),
};

static int assign_client_id(struct ib_client *client)
{
	int ret;

	down_write(&clients_rwsem);
	/*
	 * The add/remove callbacks must be called in FIFO/LIFO order. To
	 * achieve this we assign client_ids so they are sorted in
	 * registration order.
	 */
	client->client_id = highest_client_id;
	ret = xa_insert(&clients, client->client_id, client, GFP_KERNEL);
	if (ret)
		goto out;

	highest_client_id++;
	xa_set_mark(&clients, client->client_id, CLIENT_REGISTERED);

out:
	up_write(&clients_rwsem);
	return ret;
}

static void remove_client_id(struct ib_client *client)
{
	down_write(&clients_rwsem);
	xa_erase(&clients, client->client_id);
	for (; highest_client_id; highest_client_id--)
		if (xa_load(&clients, highest_client_id - 1))
			break;
	up_write(&clients_rwsem);
}

/**
 * ib_register_client - Register an IB client
 * @client:Client to register
 *
 * Upper level users of the IB drivers can use ib_register_client() to
 * register callbacks for IB device addition and removal.  When an IB
 * device is added, each registered client's add method will be called
 * (in the order the clients were registered), and when a device is
 * removed, each client's remove method will be called (in the reverse
 * order that clients were registered).  In addition, when
 * ib_register_client() is called, the client will receive an add
 * callback for all devices already registered.
 */
int ib_register_client(struct ib_client *client)
{
	struct ib_device *device;
	unsigned long index;
	int ret;

	refcount_set(&client->uses, 1);
	init_completion(&client->uses_zero);
	ret = assign_client_id(client);
	if (ret)
		return ret;

	down_read(&devices_rwsem);
	xa_for_each_marked (&devices, index, device, DEVICE_REGISTERED) {
		ret = add_client_context(device, client);
		if (ret) {
			up_read(&devices_rwsem);
			ib_unregister_client(client);
			return ret;
		}
	}
	up_read(&devices_rwsem);
	return 0;
}
EXPORT_SYMBOL(ib_register_client);

/**
 * ib_unregister_client - Unregister an IB client
 * @client:Client to unregister
 *
 * Upper level users use ib_unregister_client() to remove their client
 * registration.  When ib_unregister_client() is called, the client
 * will receive a remove callback for each IB device still registered.
 *
 * This is a full fence, once it returns no client callbacks will be called,
 * or are running in another thread.
 */
void ib_unregister_client(struct ib_client *client)
{
	struct ib_device *device;
	unsigned long index;

	down_write(&clients_rwsem);
	ib_client_put(client);
	xa_clear_mark(&clients, client->client_id, CLIENT_REGISTERED);
	up_write(&clients_rwsem);

	/* We do not want to have locks while calling client->remove() */
	rcu_read_lock();
	xa_for_each (&devices, index, device) {
		if (!ib_device_try_get(device))
			continue;
		rcu_read_unlock();

		remove_client_context(device, client->client_id);

		ib_device_put(device);
		rcu_read_lock();
	}
	rcu_read_unlock();

	/*
	 * remove_client_context() is not a fence, it can return even though a
	 * removal is ongoing. Wait until all removals are completed.
	 */
	wait_for_completion(&client->uses_zero);
	remove_client_id(client);
}
EXPORT_SYMBOL(ib_unregister_client);

static int __ib_get_global_client_nl_info(const char *client_name,
					  struct ib_client_nl_info *res)
{
	struct ib_client *client;
	unsigned long index;
	int ret = -ENOENT;

	down_read(&clients_rwsem);
	xa_for_each_marked (&clients, index, client, CLIENT_REGISTERED) {
		if (strcmp(client->name, client_name) != 0)
			continue;
		if (!client->get_global_nl_info) {
			ret = -EOPNOTSUPP;
			break;
		}
		ret = client->get_global_nl_info(res);
		if (WARN_ON(ret == -ENOENT))
			ret = -EINVAL;
		if (!ret && res->cdev)
			get_device(res->cdev);
		break;
	}
	up_read(&clients_rwsem);
	return ret;
}

static int __ib_get_client_nl_info(struct ib_device *ibdev,
				   const char *client_name,
				   struct ib_client_nl_info *res)
{
	unsigned long index;
	void *client_data;
	int ret = -ENOENT;

	down_read(&ibdev->client_data_rwsem);
	xan_for_each_marked (&ibdev->client_data, index, client_data,
			     CLIENT_DATA_REGISTERED) {
		struct ib_client *client = xa_load(&clients, index);

		if (!client || strcmp(client->name, client_name) != 0)
			continue;
		if (!client->get_nl_info) {
			ret = -EOPNOTSUPP;
			break;
		}
		ret = client->get_nl_info(ibdev, client_data, res);
		if (WARN_ON(ret == -ENOENT))
			ret = -EINVAL;

		/*
		 * The cdev is guaranteed valid as long as we are inside the
		 * client_data_rwsem as remove_one can't be called. Keep it
		 * valid for the caller.
		 */
		if (!ret && res->cdev)
			get_device(res->cdev);
		break;
	}
	up_read(&ibdev->client_data_rwsem);

	return ret;
}

/**
 * ib_get_client_nl_info - Fetch the nl_info from a client
 * @ibdev: IB device
 * @client_name: Name of the client
 * @res: Result of the query
 */
int ib_get_client_nl_info(struct ib_device *ibdev, const char *client_name,
			  struct ib_client_nl_info *res)
{
	int ret;

	if (ibdev)
		ret = __ib_get_client_nl_info(ibdev, client_name, res);
	else
		ret = __ib_get_global_client_nl_info(client_name, res);
#ifdef CONFIG_MODULES
	if (ret == -ENOENT) {
		request_module("rdma-client-%s", client_name);
		if (ibdev)
			ret = __ib_get_client_nl_info(ibdev, client_name, res);
		else
			ret = __ib_get_global_client_nl_info(client_name, res);
	}
#endif
	if (ret) {
		if (ret == -ENOENT)
			return -EOPNOTSUPP;
		return ret;
	}

	if (WARN_ON(!res->cdev))
		return -EINVAL;
	return 0;
}

/**
 * ib_set_client_data - Set IB client context
 * @device:Device to set context for
 * @client:Client to set context for
 * @data:Context to set
 *
 * ib_set_client_data() sets client context data that can be retrieved with
 * ib_get_client_data(). This can only be called while the client is
 * registered to the device, once the ib_client remove() callback returns this
 * cannot be called.
 */
void ib_set_client_data(struct ib_device *device, struct ib_client *client,
			void *data)
{
	void *rc;

	if (WARN_ON(IS_ERR(data)))
		data = NULL;

	rc = xa_store(&device->client_data, client->client_id, data,
		      GFP_KERNEL);
	WARN_ON(xa_is_err(rc));
}
EXPORT_SYMBOL(ib_set_client_data);

/**
 * ib_register_event_handler - Register an IB event handler
 * @event_handler:Handler to register
 *
 * ib_register_event_handler() registers an event handler that will be
 * called back when asynchronous IB events occur (as defined in
 * chapter 11 of the InfiniBand Architecture Specification). This
 * callback occurs in workqueue context.
 */
void ib_register_event_handler(struct ib_event_handler *event_handler)
{
	down_write(&event_handler->device->event_handler_rwsem);
	list_add_tail(&event_handler->list,
		      &event_handler->device->event_handler_list);
	up_write(&event_handler->device->event_handler_rwsem);
}
EXPORT_SYMBOL(ib_register_event_handler);

/**
 * ib_unregister_event_handler - Unregister an event handler
 * @event_handler:Handler to unregister
 *
 * Unregister an event handler registered with
 * ib_register_event_handler().
 */
void ib_unregister_event_handler(struct ib_event_handler *event_handler)
{
	down_write(&event_handler->device->event_handler_rwsem);
	list_del(&event_handler->list);
	up_write(&event_handler->device->event_handler_rwsem);
}
EXPORT_SYMBOL(ib_unregister_event_handler);

void ib_dispatch_event_clients(struct ib_event *event)
{
	struct ib_event_handler *handler;

	down_read(&event->device->event_handler_rwsem);

	list_for_each_entry(handler, &event->device->event_handler_list, list)
		handler->handler(handler, event);

	up_read(&event->device->event_handler_rwsem);
}

static int iw_query_port(struct ib_device *device,
			   u32 port_num,
			   struct ib_port_attr *port_attr)
{
	struct in_device *inetdev;
	struct net_device *netdev;

	memset(port_attr, 0, sizeof(*port_attr));

	netdev = ib_device_get_netdev(device, port_num);
	if (!netdev)
		return -ENODEV;

	port_attr->max_mtu = IB_MTU_4096;
	port_attr->active_mtu = ib_mtu_int_to_enum(netdev->mtu);

	if (!netif_carrier_ok(netdev)) {
		port_attr->state = IB_PORT_DOWN;
		port_attr->phys_state = IB_PORT_PHYS_STATE_DISABLED;
	} else {
		rcu_read_lock();
		inetdev = __in_dev_get_rcu(netdev);

		if (inetdev && inetdev->ifa_list) {
			port_attr->state = IB_PORT_ACTIVE;
			port_attr->phys_state = IB_PORT_PHYS_STATE_LINK_UP;
		} else {
			port_attr->state = IB_PORT_INIT;
			port_attr->phys_state =
				IB_PORT_PHYS_STATE_PORT_CONFIGURATION_TRAINING;
		}

		rcu_read_unlock();
	}

	dev_put(netdev);
	return device->ops.query_port(device, port_num, port_attr);
}

static int __ib_query_port(struct ib_device *device,
			   u32 port_num,
			   struct ib_port_attr *port_attr)
{
	int err;

	memset(port_attr, 0, sizeof(*port_attr));

	err = device->ops.query_port(device, port_num, port_attr);
	if (err || port_attr->subnet_prefix)
		return err;

	if (rdma_port_get_link_layer(device, port_num) !=
	    IB_LINK_LAYER_INFINIBAND)
		return 0;

	ib_get_cached_subnet_prefix(device, port_num,
				    &port_attr->subnet_prefix);
	return 0;
}

/**
 * ib_query_port - Query IB port attributes
 * @device:Device to query
 * @port_num:Port number to query
 * @port_attr:Port attributes
 *
 * ib_query_port() returns the attributes of a port through the
 * @port_attr pointer.
 */
int ib_query_port(struct ib_device *device,
		  u32 port_num,
		  struct ib_port_attr *port_attr)
{
	if (!rdma_is_port_valid(device, port_num))
		return -EINVAL;

	if (rdma_protocol_iwarp(device, port_num))
		return iw_query_port(device, port_num, port_attr);
	else
		return __ib_query_port(device, port_num, port_attr);
}
EXPORT_SYMBOL(ib_query_port);

static void add_ndev_hash(struct ib_port_data *pdata)
{
	unsigned long flags;

	might_sleep();

	spin_lock_irqsave(&ndev_hash_lock, flags);
	if (hash_hashed(&pdata->ndev_hash_link)) {
		hash_del_rcu(&pdata->ndev_hash_link);
		spin_unlock_irqrestore(&ndev_hash_lock, flags);
		/*
		 * We cannot do hash_add_rcu after a hash_del_rcu until the
		 * grace period
		 */
		synchronize_rcu();
		spin_lock_irqsave(&ndev_hash_lock, flags);
	}
	if (pdata->netdev)
		hash_add_rcu(ndev_hash, &pdata->ndev_hash_link,
			     (uintptr_t)pdata->netdev);
	spin_unlock_irqrestore(&ndev_hash_lock, flags);
}

/**
 * ib_device_set_netdev - Associate the ib_dev with an underlying net_device
 * @ib_dev: Device to modify
 * @ndev: net_device to affiliate, may be NULL
 * @port: IB port the net_device is connected to
 *
 * Drivers should use this to link the ib_device to a netdev so the netdev
 * shows up in interfaces like ib_enum_roce_netdev. Only one netdev may be
 * affiliated with any port.
 *
 * The caller must ensure that the given ndev is not unregistered or
 * unregistering, and that either the ib_device is unregistered or
 * ib_device_set_netdev() is called with NULL when the ndev sends a
 * NETDEV_UNREGISTER event.
 */
int ib_device_set_netdev(struct ib_device *ib_dev, struct net_device *ndev,
			 u32 port)
{
	struct net_device *old_ndev;
	struct ib_port_data *pdata;
	unsigned long flags;
	int ret;

	/*
	 * Drivers wish to call this before ib_register_driver, so we have to
	 * setup the port data early.
	 */
	ret = alloc_port_data(ib_dev);
	if (ret)
		return ret;

	if (!rdma_is_port_valid(ib_dev, port))
		return -EINVAL;

	pdata = &ib_dev->port_data[port];
	spin_lock_irqsave(&pdata->netdev_lock, flags);
	old_ndev = rcu_dereference_protected(
		pdata->netdev, lockdep_is_held(&pdata->netdev_lock));
	if (old_ndev == ndev) {
		spin_unlock_irqrestore(&pdata->netdev_lock, flags);
		return 0;
	}

	if (ndev)
		dev_hold(ndev);
	rcu_assign_pointer(pdata->netdev, ndev);
	spin_unlock_irqrestore(&pdata->netdev_lock, flags);

	add_ndev_hash(pdata);
	if (old_ndev)
		dev_put(old_ndev);

	return 0;
}
EXPORT_SYMBOL(ib_device_set_netdev);

static void free_netdevs(struct ib_device *ib_dev)
{
	unsigned long flags;
	u32 port;

	if (!ib_dev->port_data)
		return;

	rdma_for_each_port (ib_dev, port) {
		struct ib_port_data *pdata = &ib_dev->port_data[port];
		struct net_device *ndev;

		spin_lock_irqsave(&pdata->netdev_lock, flags);
		ndev = rcu_dereference_protected(
			pdata->netdev, lockdep_is_held(&pdata->netdev_lock));
		if (ndev) {
			spin_lock(&ndev_hash_lock);
			hash_del_rcu(&pdata->ndev_hash_link);
			spin_unlock(&ndev_hash_lock);

			/*
			 * If this is the last dev_put there is still a
			 * synchronize_rcu before the netdev is kfreed, so we
			 * can continue to rely on unlocked pointer
			 * comparisons after the put
			 */
			rcu_assign_pointer(pdata->netdev, NULL);
			dev_put(ndev);
		}
		spin_unlock_irqrestore(&pdata->netdev_lock, flags);
	}
}

struct net_device *ib_device_get_netdev(struct ib_device *ib_dev,
					u32 port)
{
	struct ib_port_data *pdata;
	struct net_device *res;

	if (!rdma_is_port_valid(ib_dev, port))
		return NULL;

	pdata = &ib_dev->port_data[port];

	/*
	 * New drivers should use ib_device_set_netdev() not the legacy
	 * get_netdev().
	 */
	if (ib_dev->ops.get_netdev)
		res = ib_dev->ops.get_netdev(ib_dev, port);
	else {
		spin_lock(&pdata->netdev_lock);
		res = rcu_dereference_protected(
			pdata->netdev, lockdep_is_held(&pdata->netdev_lock));
		if (res)
			dev_hold(res);
		spin_unlock(&pdata->netdev_lock);
	}

	/*
	 * If we are starting to unregister expedite things by preventing
	 * propagation of an unregistering netdev.
	 */
	if (res && res->reg_state != NETREG_REGISTERED) {
		dev_put(res);
		return NULL;
	}

	return res;
}

/**
 * ib_device_get_by_netdev - Find an IB device associated with a netdev
 * @ndev: netdev to locate
 * @driver_id: The driver ID that must match (RDMA_DRIVER_UNKNOWN matches all)
 *
 * Find and hold an ib_device that is associated with a netdev via
 * ib_device_set_netdev(). The caller must call ib_device_put() on the
 * returned pointer.
 */
struct ib_device *ib_device_get_by_netdev(struct net_device *ndev,
					  enum rdma_driver_id driver_id)
{
	struct ib_device *res = NULL;
	struct ib_port_data *cur;

	rcu_read_lock();
	hash_for_each_possible_rcu (ndev_hash, cur, ndev_hash_link,
				    (uintptr_t)ndev) {
		if (rcu_access_pointer(cur->netdev) == ndev &&
		    (driver_id == RDMA_DRIVER_UNKNOWN ||
		     cur->ib_dev->ops.driver_id == driver_id) &&
		    ib_device_try_get(cur->ib_dev)) {
			res = cur->ib_dev;
			break;
		}
	}
	rcu_read_unlock();

	return res;
}
EXPORT_SYMBOL(ib_device_get_by_netdev);

/**
 * ib_enum_roce_netdev - enumerate all RoCE ports
 * @ib_dev : IB device we want to query
 * @filter: Should we call the callback?
 * @filter_cookie: Cookie passed to filter
 * @cb: Callback to call for each found RoCE ports
 * @cookie: Cookie passed back to the callback
 *
 * Enumerates all of the physical RoCE ports of ib_dev
 * which are related to netdevice and calls callback() on each
 * device for which filter() function returns non zero.
 */
void ib_enum_roce_netdev(struct ib_device *ib_dev,
			 roce_netdev_filter filter,
			 void *filter_cookie,
			 roce_netdev_callback cb,
			 void *cookie)
{
	u32 port;

	rdma_for_each_port (ib_dev, port)
		if (rdma_protocol_roce(ib_dev, port)) {
			struct net_device *idev =
				ib_device_get_netdev(ib_dev, port);

			if (filter(ib_dev, port, idev, filter_cookie))
				cb(ib_dev, port, idev, cookie);

			if (idev)
				dev_put(idev);
		}
}

/**
 * ib_enum_all_roce_netdevs - enumerate all RoCE devices
 * @filter: Should we call the callback?
 * @filter_cookie: Cookie passed to filter
 * @cb: Callback to call for each found RoCE ports
 * @cookie: Cookie passed back to the callback
 *
 * Enumerates all RoCE devices' physical ports which are related
 * to netdevices and calls callback() on each device for which
 * filter() function returns non zero.
 */
void ib_enum_all_roce_netdevs(roce_netdev_filter filter,
			      void *filter_cookie,
			      roce_netdev_callback cb,
			      void *cookie)
{
	struct ib_device *dev;
	unsigned long index;

	down_read(&devices_rwsem);
	xa_for_each_marked (&devices, index, dev, DEVICE_REGISTERED)
		ib_enum_roce_netdev(dev, filter, filter_cookie, cb, cookie);
	up_read(&devices_rwsem);
}

/*
 * ib_enum_all_devs - enumerate all ib_devices
 * @cb: Callback to call for each found ib_device
 *
 * Enumerates all ib_devices and calls callback() on each device.
 */
int ib_enum_all_devs(nldev_callback nldev_cb, struct sk_buff *skb,
		     struct netlink_callback *cb)
{
	unsigned long index;
	struct ib_device *dev;
	unsigned int idx = 0;
	int ret = 0;

	down_read(&devices_rwsem);
	xa_for_each_marked (&devices, index, dev, DEVICE_REGISTERED) {
		if (!rdma_dev_access_netns(dev, sock_net(skb->sk)))
			continue;

		ret = nldev_cb(dev, skb, cb, idx);
		if (ret)
			break;
		idx++;
	}
	up_read(&devices_rwsem);
	return ret;
}

/**
 * ib_query_pkey - Get P_Key table entry
 * @device:Device to query
 * @port_num:Port number to query
 * @index:P_Key table index to query
 * @pkey:Returned P_Key
 *
 * ib_query_pkey() fetches the specified P_Key table entry.
 */
int ib_query_pkey(struct ib_device *device,
		  u32 port_num, u16 index, u16 *pkey)
{
	if (!rdma_is_port_valid(device, port_num))
		return -EINVAL;

	if (!device->ops.query_pkey)
		return -EOPNOTSUPP;

	return device->ops.query_pkey(device, port_num, index, pkey);
}
EXPORT_SYMBOL(ib_query_pkey);

/**
 * ib_modify_device - Change IB device attributes
 * @device:Device to modify
 * @device_modify_mask:Mask of attributes to change
 * @device_modify:New attribute values
 *
 * ib_modify_device() changes a device's attributes as specified by
 * the @device_modify_mask and @device_modify structure.
 */
int ib_modify_device(struct ib_device *device,
		     int device_modify_mask,
		     struct ib_device_modify *device_modify)
{
	if (!device->ops.modify_device)
		return -EOPNOTSUPP;

	return device->ops.modify_device(device, device_modify_mask,
					 device_modify);
}
EXPORT_SYMBOL(ib_modify_device);

/**
 * ib_modify_port - Modifies the attributes for the specified port.
 * @device: The device to modify.
 * @port_num: The number of the port to modify.
 * @port_modify_mask: Mask used to specify which attributes of the port
 *   to change.
 * @port_modify: New attribute values for the port.
 *
 * ib_modify_port() changes a port's attributes as specified by the
 * @port_modify_mask and @port_modify structure.
 */
int ib_modify_port(struct ib_device *device,
		   u32 port_num, int port_modify_mask,
		   struct ib_port_modify *port_modify)
{
	int rc;

	if (!rdma_is_port_valid(device, port_num))
		return -EINVAL;

	if (device->ops.modify_port)
		rc = device->ops.modify_port(device, port_num,
					     port_modify_mask,
					     port_modify);
	else if (rdma_protocol_roce(device, port_num) &&
		 ((port_modify->set_port_cap_mask & ~IB_PORT_CM_SUP) == 0 ||
		  (port_modify->clr_port_cap_mask & ~IB_PORT_CM_SUP) == 0))
		rc = 0;
	else
		rc = -EOPNOTSUPP;
	return rc;
}
EXPORT_SYMBOL(ib_modify_port);

/**
 * ib_find_gid - Returns the port number and GID table index where
 *   a specified GID value occurs. Its searches only for IB link layer.
 * @device: The device to query.
 * @gid: The GID value to search for.
 * @port_num: The port number of the device where the GID value was found.
 * @index: The index into the GID table where the GID was found.  This
 *   parameter may be NULL.
 */
int ib_find_gid(struct ib_device *device, union ib_gid *gid,
		u32 *port_num, u16 *index)
{
	union ib_gid tmp_gid;
	u32 port;
	int ret, i;

	rdma_for_each_port (device, port) {
		if (!rdma_protocol_ib(device, port))
			continue;

		for (i = 0; i < device->port_data[port].immutable.gid_tbl_len;
		     ++i) {
			ret = rdma_query_gid(device, port, i, &tmp_gid);
			if (ret)
				continue;

			if (!memcmp(&tmp_gid, gid, sizeof *gid)) {
				*port_num = port;
				if (index)
					*index = i;
				return 0;
			}
		}
	}

	return -ENOENT;
}
EXPORT_SYMBOL(ib_find_gid);

/**
 * ib_find_pkey - Returns the PKey table index where a specified
 *   PKey value occurs.
 * @device: The device to query.
 * @port_num: The port number of the device to search for the PKey.
 * @pkey: The PKey value to search for.
 * @index: The index into the PKey table where the PKey was found.
 */
int ib_find_pkey(struct ib_device *device,
		 u32 port_num, u16 pkey, u16 *index)
{
	int ret, i;
	u16 tmp_pkey;
	int partial_ix = -1;

	for (i = 0; i < device->port_data[port_num].immutable.pkey_tbl_len;
	     ++i) {
		ret = ib_query_pkey(device, port_num, i, &tmp_pkey);
		if (ret)
			return ret;
		if ((pkey & 0x7fff) == (tmp_pkey & 0x7fff)) {
			/* if there is full-member pkey take it.*/
			if (tmp_pkey & 0x8000) {
				*index = i;
				return 0;
			}
			if (partial_ix < 0)
				partial_ix = i;
		}
	}

	/*no full-member, if exists take the limited*/
	if (partial_ix >= 0) {
		*index = partial_ix;
		return 0;
	}
	return -ENOENT;
}
EXPORT_SYMBOL(ib_find_pkey);

/**
 * ib_get_net_dev_by_params() - Return the appropriate net_dev
 * for a received CM request
 * @dev:	An RDMA device on which the request has been received.
 * @port:	Port number on the RDMA device.
 * @pkey:	The Pkey the request came on.
 * @gid:	A GID that the net_dev uses to communicate.
 * @addr:	Contains the IP address that the request specified as its
 *		destination.
 *
 */
struct net_device *ib_get_net_dev_by_params(struct ib_device *dev,
					    u32 port,
					    u16 pkey,
					    const union ib_gid *gid,
					    const struct sockaddr *addr)
{
	struct net_device *net_dev = NULL;
	unsigned long index;
	void *client_data;

	if (!rdma_protocol_ib(dev, port))
		return NULL;

	/*
	 * Holding the read side guarantees that the client will not become
	 * unregistered while we are calling get_net_dev_by_params()
	 */
	down_read(&dev->client_data_rwsem);
	xan_for_each_marked (&dev->client_data, index, client_data,
			     CLIENT_DATA_REGISTERED) {
		struct ib_client *client = xa_load(&clients, index);

		if (!client || !client->get_net_dev_by_params)
			continue;

		net_dev = client->get_net_dev_by_params(dev, port, pkey, gid,
							addr, client_data);
		if (net_dev)
			break;
	}
	up_read(&dev->client_data_rwsem);

	return net_dev;
}
EXPORT_SYMBOL(ib_get_net_dev_by_params);

void ib_set_device_ops(struct ib_device *dev, const struct ib_device_ops *ops)
{
	struct ib_device_ops *dev_ops = &dev->ops;
#define SET_DEVICE_OP(ptr, name)                                               \
	do {                                                                   \
		if (ops->name)                                                 \
			if (!((ptr)->name))				       \
				(ptr)->name = ops->name;                       \
	} while (0)

#define SET_OBJ_SIZE(ptr, name) SET_DEVICE_OP(ptr, size_##name)

	if (ops->driver_id != RDMA_DRIVER_UNKNOWN) {
		WARN_ON(dev_ops->driver_id != RDMA_DRIVER_UNKNOWN &&
			dev_ops->driver_id != ops->driver_id);
		dev_ops->driver_id = ops->driver_id;
	}
	if (ops->owner) {
		WARN_ON(dev_ops->owner && dev_ops->owner != ops->owner);
		dev_ops->owner = ops->owner;
	}
	if (ops->uverbs_abi_ver)
		dev_ops->uverbs_abi_ver = ops->uverbs_abi_ver;

	dev_ops->uverbs_no_driver_id_binding |=
		ops->uverbs_no_driver_id_binding;

	SET_DEVICE_OP(dev_ops, add_gid);
	SET_DEVICE_OP(dev_ops, advise_mr);
	SET_DEVICE_OP(dev_ops, alloc_dm);
	SET_DEVICE_OP(dev_ops, alloc_hw_device_stats);
	SET_DEVICE_OP(dev_ops, alloc_hw_port_stats);
	SET_DEVICE_OP(dev_ops, alloc_mr);
	SET_DEVICE_OP(dev_ops, alloc_mr_integrity);
	SET_DEVICE_OP(dev_ops, alloc_mw);
	SET_DEVICE_OP(dev_ops, alloc_pd);
	SET_DEVICE_OP(dev_ops, alloc_rdma_netdev);
	SET_DEVICE_OP(dev_ops, alloc_ucontext);
	SET_DEVICE_OP(dev_ops, alloc_xrcd);
	SET_DEVICE_OP(dev_ops, attach_mcast);
	SET_DEVICE_OP(dev_ops, check_mr_status);
	SET_DEVICE_OP(dev_ops, counter_alloc_stats);
	SET_DEVICE_OP(dev_ops, counter_bind_qp);
	SET_DEVICE_OP(dev_ops, counter_dealloc);
	SET_DEVICE_OP(dev_ops, counter_unbind_qp);
	SET_DEVICE_OP(dev_ops, counter_update_stats);
	SET_DEVICE_OP(dev_ops, create_ah);
	SET_DEVICE_OP(dev_ops, create_counters);
	SET_DEVICE_OP(dev_ops, create_cq);
	SET_DEVICE_OP(dev_ops, create_flow);
	SET_DEVICE_OP(dev_ops, create_qp);
	SET_DEVICE_OP(dev_ops, create_rwq_ind_table);
	SET_DEVICE_OP(dev_ops, create_srq);
	SET_DEVICE_OP(dev_ops, create_user_ah);
	SET_DEVICE_OP(dev_ops, create_wq);
	SET_DEVICE_OP(dev_ops, dealloc_dm);
	SET_DEVICE_OP(dev_ops, dealloc_driver);
	SET_DEVICE_OP(dev_ops, dealloc_mw);
	SET_DEVICE_OP(dev_ops, dealloc_pd);
	SET_DEVICE_OP(dev_ops, dealloc_ucontext);
	SET_DEVICE_OP(dev_ops, dealloc_xrcd);
	SET_DEVICE_OP(dev_ops, del_gid);
	SET_DEVICE_OP(dev_ops, dereg_mr);
	SET_DEVICE_OP(dev_ops, destroy_ah);
	SET_DEVICE_OP(dev_ops, destroy_counters);
	SET_DEVICE_OP(dev_ops, destroy_cq);
	SET_DEVICE_OP(dev_ops, destroy_flow);
	SET_DEVICE_OP(dev_ops, destroy_flow_action);
	SET_DEVICE_OP(dev_ops, destroy_qp);
	SET_DEVICE_OP(dev_ops, destroy_rwq_ind_table);
	SET_DEVICE_OP(dev_ops, destroy_srq);
	SET_DEVICE_OP(dev_ops, destroy_wq);
	SET_DEVICE_OP(dev_ops, device_group);
	SET_DEVICE_OP(dev_ops, detach_mcast);
	SET_DEVICE_OP(dev_ops, disassociate_ucontext);
	SET_DEVICE_OP(dev_ops, drain_rq);
	SET_DEVICE_OP(dev_ops, drain_sq);
	SET_DEVICE_OP(dev_ops, enable_driver);
	SET_DEVICE_OP(dev_ops, fill_res_cm_id_entry);
	SET_DEVICE_OP(dev_ops, fill_res_cq_entry);
	SET_DEVICE_OP(dev_ops, fill_res_cq_entry_raw);
	SET_DEVICE_OP(dev_ops, fill_res_mr_entry);
	SET_DEVICE_OP(dev_ops, fill_res_mr_entry_raw);
	SET_DEVICE_OP(dev_ops, fill_res_qp_entry);
	SET_DEVICE_OP(dev_ops, fill_res_qp_entry_raw);
	SET_DEVICE_OP(dev_ops, fill_stat_mr_entry);
	SET_DEVICE_OP(dev_ops, get_dev_fw_str);
	SET_DEVICE_OP(dev_ops, get_dma_mr);
	SET_DEVICE_OP(dev_ops, get_hw_stats);
	SET_DEVICE_OP(dev_ops, get_link_layer);
	SET_DEVICE_OP(dev_ops, get_netdev);
	SET_DEVICE_OP(dev_ops, get_numa_node);
	SET_DEVICE_OP(dev_ops, get_port_immutable);
	SET_DEVICE_OP(dev_ops, get_vector_affinity);
	SET_DEVICE_OP(dev_ops, get_vf_config);
	SET_DEVICE_OP(dev_ops, get_vf_guid);
	SET_DEVICE_OP(dev_ops, get_vf_stats);
	SET_DEVICE_OP(dev_ops, iw_accept);
	SET_DEVICE_OP(dev_ops, iw_add_ref);
	SET_DEVICE_OP(dev_ops, iw_connect);
	SET_DEVICE_OP(dev_ops, iw_create_listen);
	SET_DEVICE_OP(dev_ops, iw_destroy_listen);
	SET_DEVICE_OP(dev_ops, iw_get_qp);
	SET_DEVICE_OP(dev_ops, iw_reject);
	SET_DEVICE_OP(dev_ops, iw_rem_ref);
	SET_DEVICE_OP(dev_ops, map_mr_sg);
	SET_DEVICE_OP(dev_ops, map_mr_sg_pi);
	SET_DEVICE_OP(dev_ops, mmap);
	SET_DEVICE_OP(dev_ops, mmap_free);
	SET_DEVICE_OP(dev_ops, modify_ah);
	SET_DEVICE_OP(dev_ops, modify_cq);
	SET_DEVICE_OP(dev_ops, modify_device);
	SET_DEVICE_OP(dev_ops, modify_hw_stat);
	SET_DEVICE_OP(dev_ops, modify_port);
	SET_DEVICE_OP(dev_ops, modify_qp);
	SET_DEVICE_OP(dev_ops, modify_srq);
	SET_DEVICE_OP(dev_ops, modify_wq);
	SET_DEVICE_OP(dev_ops, peek_cq);
	SET_DEVICE_OP(dev_ops, poll_cq);
	SET_DEVICE_OP(dev_ops, port_groups);
	SET_DEVICE_OP(dev_ops, post_recv);
	SET_DEVICE_OP(dev_ops, post_send);
	SET_DEVICE_OP(dev_ops, post_srq_recv);
	SET_DEVICE_OP(dev_ops, process_mad);
	SET_DEVICE_OP(dev_ops, query_ah);
	SET_DEVICE_OP(dev_ops, query_device);
	SET_DEVICE_OP(dev_ops, query_gid);
	SET_DEVICE_OP(dev_ops, query_pkey);
	SET_DEVICE_OP(dev_ops, query_port);
	SET_DEVICE_OP(dev_ops, query_qp);
	SET_DEVICE_OP(dev_ops, query_srq);
	SET_DEVICE_OP(dev_ops, query_ucontext);
	SET_DEVICE_OP(dev_ops, rdma_netdev_get_params);
	SET_DEVICE_OP(dev_ops, read_counters);
	SET_DEVICE_OP(dev_ops, reg_dm_mr);
	SET_DEVICE_OP(dev_ops, reg_user_mr);
	SET_DEVICE_OP(dev_ops, reg_user_mr_dmabuf);
	SET_DEVICE_OP(dev_ops, req_notify_cq);
	SET_DEVICE_OP(dev_ops, rereg_user_mr);
	SET_DEVICE_OP(dev_ops, resize_cq);
	SET_DEVICE_OP(dev_ops, set_vf_guid);
	SET_DEVICE_OP(dev_ops, set_vf_link_state);

	SET_OBJ_SIZE(dev_ops, ib_ah);
	SET_OBJ_SIZE(dev_ops, ib_counters);
	SET_OBJ_SIZE(dev_ops, ib_cq);
	SET_OBJ_SIZE(dev_ops, ib_mw);
	SET_OBJ_SIZE(dev_ops, ib_pd);
	SET_OBJ_SIZE(dev_ops, ib_qp);
	SET_OBJ_SIZE(dev_ops, ib_rwq_ind_table);
	SET_OBJ_SIZE(dev_ops, ib_srq);
	SET_OBJ_SIZE(dev_ops, ib_ucontext);
	SET_OBJ_SIZE(dev_ops, ib_xrcd);
}
EXPORT_SYMBOL(ib_set_device_ops);

#ifdef CONFIG_INFINIBAND_VIRT_DMA
int ib_dma_virt_map_sg(struct ib_device *dev, struct scatterlist *sg, int nents)
{
	struct scatterlist *s;
	int i;

	for_each_sg(sg, s, nents, i) {
		sg_dma_address(s) = (uintptr_t)sg_virt(s);
		sg_dma_len(s) = s->length;
	}
	return nents;
}
EXPORT_SYMBOL(ib_dma_virt_map_sg);
#endif /* CONFIG_INFINIBAND_VIRT_DMA */

static const struct rdma_nl_cbs ibnl_ls_cb_table[RDMA_NL_LS_NUM_OPS] = {
	[RDMA_NL_LS_OP_RESOLVE] = {
		.doit = ib_nl_handle_resolve_resp,
		.flags = RDMA_NL_ADMIN_PERM,
	},
	[RDMA_NL_LS_OP_SET_TIMEOUT] = {
		.doit = ib_nl_handle_set_timeout,
		.flags = RDMA_NL_ADMIN_PERM,
	},
	[RDMA_NL_LS_OP_IP_RESOLVE] = {
		.doit = ib_nl_handle_ip_res_resp,
		.flags = RDMA_NL_ADMIN_PERM,
	},
};

static int __init ib_core_init(void)
{
	int ret = -ENOMEM;

	ib_wq = alloc_workqueue("infiniband", 0, 0);
	if (!ib_wq)
		return -ENOMEM;

	ib_unreg_wq = alloc_workqueue("ib-unreg-wq", WQ_UNBOUND,
				      WQ_UNBOUND_MAX_ACTIVE);
	if (!ib_unreg_wq)
		goto err;

	ib_comp_wq = alloc_workqueue("ib-comp-wq",
			WQ_HIGHPRI | WQ_MEM_RECLAIM | WQ_SYSFS, 0);
	if (!ib_comp_wq)
		goto err_unbound;

	ib_comp_unbound_wq =
		alloc_workqueue("ib-comp-unb-wq",
				WQ_UNBOUND | WQ_HIGHPRI | WQ_MEM_RECLAIM |
				WQ_SYSFS, WQ_UNBOUND_MAX_ACTIVE);
	if (!ib_comp_unbound_wq)
		goto err_comp;

	ret = class_register(&ib_class);
	if (ret) {
		pr_warn("Couldn't create InfiniBand device class\n");
		goto err_comp_unbound;
	}

	rdma_nl_init();

	ret = addr_init();
	if (ret) {
		pr_warn("Couldn't init IB address resolution\n");
		goto err_ibnl;
	}

	ret = ib_mad_init();
	if (ret) {
		pr_warn("Couldn't init IB MAD\n");
		goto err_addr;
	}

	ret = ib_sa_init();
	if (ret) {
		pr_warn("Couldn't init SA\n");
		goto err_mad;
	}

	ret = register_blocking_lsm_notifier(&ibdev_lsm_nb);
	if (ret) {
		pr_warn("Couldn't register LSM notifier. ret %d\n", ret);
		goto err_sa;
	}

	ret = register_pernet_device(&rdma_dev_net_ops);
	if (ret) {
		pr_warn("Couldn't init compat dev. ret %d\n", ret);
		goto err_compat;
	}

	nldev_init();
	rdma_nl_register(RDMA_NL_LS, ibnl_ls_cb_table);
	ret = roce_gid_mgmt_init();
	if (ret) {
		pr_warn("Couldn't init RoCE GID management\n");
		goto err_parent;
	}

	return 0;

err_parent:
	rdma_nl_unregister(RDMA_NL_LS);
	nldev_exit();
	unregister_pernet_device(&rdma_dev_net_ops);
err_compat:
	unregister_blocking_lsm_notifier(&ibdev_lsm_nb);
err_sa:
	ib_sa_cleanup();
err_mad:
	ib_mad_cleanup();
err_addr:
	addr_cleanup();
err_ibnl:
	class_unregister(&ib_class);
err_comp_unbound:
	destroy_workqueue(ib_comp_unbound_wq);
err_comp:
	destroy_workqueue(ib_comp_wq);
err_unbound:
	destroy_workqueue(ib_unreg_wq);
err:
	destroy_workqueue(ib_wq);
	return ret;
}

static void __exit ib_core_cleanup(void)
{
	roce_gid_mgmt_cleanup();
	nldev_exit();
	rdma_nl_unregister(RDMA_NL_LS);
	unregister_pernet_device(&rdma_dev_net_ops);
	unregister_blocking_lsm_notifier(&ibdev_lsm_nb);
	ib_sa_cleanup();
	ib_mad_cleanup();
	addr_cleanup();
	rdma_nl_exit();
	class_unregister(&ib_class);
	destroy_workqueue(ib_comp_unbound_wq);
	destroy_workqueue(ib_comp_wq);
	/* Make sure that any pending umem accounting work is done. */
	destroy_workqueue(ib_wq);
	destroy_workqueue(ib_unreg_wq);
	WARN_ON(!xa_empty(&clients));
	WARN_ON(!xa_empty(&devices));
}

MODULE_ALIAS_RDMA_NETLINK(RDMA_NL_LS, 4);

/* ib core relies on netdev stack to first register net_ns_type_operations
 * ns kobject type before ib_core initialization.
 */
fs_initcall(ib_core_init);
module_exit(ib_core_cleanup);