summaryrefslogtreecommitdiffstats
path: root/zebra/interface.c
blob: 5ce222cc04d7c4319be76a346f09f5ae1e1211b3 (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
2910
2911
2912
2913
2914
2915
2916
2917
2918
2919
2920
2921
2922
2923
2924
2925
2926
2927
2928
2929
2930
2931
2932
2933
2934
2935
2936
2937
2938
2939
2940
2941
2942
2943
2944
2945
2946
2947
2948
2949
2950
2951
2952
2953
2954
2955
2956
2957
2958
2959
2960
2961
2962
2963
2964
2965
2966
2967
2968
2969
2970
2971
2972
2973
2974
2975
2976
2977
2978
2979
2980
2981
2982
2983
2984
2985
2986
2987
2988
2989
2990
2991
2992
2993
2994
2995
2996
2997
2998
2999
3000
3001
3002
3003
3004
3005
3006
3007
3008
3009
3010
3011
3012
3013
3014
3015
3016
3017
3018
3019
3020
3021
3022
3023
3024
3025
3026
3027
3028
3029
3030
3031
3032
3033
3034
3035
3036
3037
3038
3039
3040
3041
3042
3043
3044
3045
3046
3047
3048
3049
3050
3051
3052
3053
3054
3055
3056
3057
3058
3059
3060
3061
3062
3063
3064
3065
3066
3067
3068
3069
3070
3071
3072
3073
3074
3075
3076
3077
3078
3079
3080
3081
3082
3083
3084
3085
3086
3087
3088
3089
3090
3091
3092
3093
3094
3095
3096
3097
3098
3099
3100
3101
3102
3103
3104
3105
3106
3107
3108
3109
3110
3111
3112
3113
3114
3115
3116
3117
3118
3119
3120
3121
3122
3123
3124
3125
3126
3127
3128
3129
3130
3131
3132
3133
3134
3135
3136
3137
3138
3139
3140
3141
3142
3143
3144
3145
3146
3147
3148
3149
3150
3151
3152
3153
3154
3155
3156
3157
3158
3159
3160
3161
3162
3163
3164
3165
3166
3167
3168
3169
3170
3171
3172
3173
3174
3175
3176
3177
3178
3179
3180
3181
3182
3183
3184
3185
3186
3187
3188
3189
3190
3191
3192
3193
3194
3195
3196
3197
3198
3199
3200
3201
3202
3203
3204
3205
3206
3207
3208
3209
3210
3211
3212
3213
3214
3215
3216
3217
3218
3219
3220
3221
3222
3223
3224
3225
3226
3227
3228
3229
3230
3231
3232
3233
3234
3235
3236
3237
3238
3239
3240
3241
3242
3243
3244
3245
3246
3247
3248
3249
3250
3251
3252
3253
3254
3255
3256
3257
3258
3259
3260
3261
3262
3263
3264
3265
3266
3267
3268
3269
3270
3271
3272
3273
3274
3275
3276
3277
3278
3279
3280
3281
3282
3283
3284
3285
3286
3287
3288
3289
3290
3291
3292
3293
3294
3295
3296
3297
3298
3299
3300
3301
3302
3303
3304
3305
3306
3307
3308
3309
3310
3311
3312
3313
3314
3315
3316
3317
3318
3319
3320
3321
3322
3323
3324
3325
3326
3327
3328
3329
3330
3331
3332
3333
3334
3335
3336
3337
3338
3339
3340
3341
3342
3343
3344
3345
3346
3347
3348
3349
3350
3351
3352
3353
3354
3355
3356
3357
3358
3359
3360
3361
3362
3363
3364
3365
3366
3367
3368
3369
3370
3371
3372
3373
3374
3375
3376
3377
3378
3379
3380
3381
3382
3383
3384
3385
3386
3387
3388
3389
3390
3391
3392
3393
3394
3395
3396
3397
3398
3399
3400
3401
3402
3403
3404
3405
3406
3407
3408
3409
3410
3411
3412
3413
3414
3415
3416
3417
3418
3419
3420
3421
3422
3423
3424
3425
3426
3427
3428
3429
3430
3431
3432
3433
3434
3435
3436
3437
3438
3439
3440
3441
3442
3443
3444
3445
3446
3447
3448
3449
3450
3451
3452
3453
3454
3455
3456
3457
3458
3459
3460
3461
3462
3463
3464
3465
3466
3467
3468
3469
3470
3471
3472
3473
3474
3475
3476
3477
3478
3479
3480
3481
3482
3483
3484
3485
3486
3487
3488
3489
3490
3491
3492
3493
3494
3495
3496
3497
3498
3499
3500
3501
3502
3503
3504
3505
3506
3507
3508
3509
3510
3511
3512
3513
3514
3515
3516
3517
3518
3519
3520
3521
3522
3523
3524
3525
3526
3527
3528
3529
3530
3531
3532
3533
3534
3535
3536
3537
3538
3539
3540
3541
3542
3543
3544
3545
3546
3547
3548
3549
3550
3551
3552
3553
3554
3555
3556
3557
3558
3559
3560
3561
3562
3563
3564
3565
3566
3567
3568
3569
3570
3571
3572
3573
3574
3575
3576
3577
3578
3579
3580
3581
3582
3583
3584
3585
3586
3587
3588
3589
3590
3591
3592
3593
3594
3595
3596
3597
3598
3599
3600
3601
3602
3603
3604
3605
3606
3607
3608
3609
3610
3611
3612
3613
3614
3615
3616
3617
3618
3619
3620
3621
3622
3623
3624
3625
3626
3627
3628
3629
3630
3631
3632
3633
3634
3635
3636
3637
3638
3639
3640
3641
3642
3643
3644
3645
3646
3647
3648
3649
3650
3651
3652
3653
3654
3655
3656
3657
3658
3659
3660
3661
3662
3663
3664
3665
3666
3667
3668
3669
3670
3671
3672
3673
3674
3675
3676
3677
3678
3679
3680
3681
3682
3683
3684
3685
3686
3687
3688
3689
3690
3691
3692
3693
3694
3695
3696
3697
3698
3699
3700
3701
3702
3703
3704
3705
3706
3707
3708
3709
3710
3711
3712
3713
3714
3715
3716
3717
3718
3719
3720
3721
3722
3723
3724
3725
3726
3727
3728
3729
3730
3731
3732
3733
3734
3735
3736
3737
3738
3739
3740
3741
3742
3743
3744
3745
3746
3747
3748
3749
3750
3751
3752
3753
3754
3755
3756
3757
3758
3759
3760
3761
3762
3763
3764
3765
3766
3767
3768
3769
3770
3771
3772
3773
3774
3775
3776
3777
3778
3779
3780
3781
3782
3783
3784
3785
3786
3787
3788
3789
3790
3791
3792
3793
3794
3795
3796
3797
3798
3799
3800
3801
3802
3803
3804
3805
3806
3807
3808
3809
3810
3811
3812
3813
3814
3815
3816
3817
3818
3819
3820
3821
3822
3823
3824
3825
3826
3827
3828
3829
3830
3831
3832
3833
3834
3835
3836
3837
3838
3839
3840
3841
3842
3843
3844
3845
3846
3847
3848
3849
3850
3851
3852
3853
3854
3855
3856
3857
3858
3859
3860
3861
3862
3863
3864
3865
3866
3867
3868
3869
3870
3871
3872
3873
3874
3875
3876
3877
3878
3879
3880
3881
3882
3883
3884
3885
3886
3887
3888
3889
3890
3891
3892
3893
3894
3895
3896
3897
3898
3899
3900
3901
3902
3903
3904
3905
3906
3907
3908
3909
3910
3911
3912
3913
3914
3915
3916
3917
3918
3919
3920
3921
3922
3923
3924
3925
3926
3927
3928
3929
3930
3931
3932
3933
3934
3935
3936
3937
3938
3939
3940
3941
3942
3943
3944
3945
3946
3947
3948
3949
3950
3951
3952
3953
3954
3955
3956
3957
3958
3959
3960
3961
3962
3963
3964
3965
3966
3967
3968
3969
3970
3971
3972
3973
3974
3975
3976
3977
3978
3979
// SPDX-License-Identifier: GPL-2.0-or-later
/*
 * Interface function.
 * Copyright (C) 1997, 1999 Kunihiro Ishiguro
 */

#include <zebra.h>

#include "if.h"
#include "lib_errors.h"
#include "vty.h"
#include "sockunion.h"
#include "prefix.h"
#include "command.h"
#include "memory.h"
#include "ioctl.h"
#include "connected.h"
#include "log.h"
#include "zclient.h"
#include "vrf.h"

#include "zebra/rtadv.h"
#include "zebra_ns.h"
#include "zebra_vrf.h"
#include "zebra/interface.h"
#include "zebra/rib.h"
#include "zebra/rt.h"
#include "zebra/zebra_router.h"
#include "zebra/redistribute.h"
#include "zebra/debug.h"
#include "zebra/irdp.h"
#include "zebra/zebra_ptm.h"
#include "zebra/rt_netlink.h"
#include "zebra/if_netlink.h"
#include "zebra/interface.h"
#include "zebra/zebra_vxlan.h"
#include "zebra/zebra_errors.h"
#include "zebra/zebra_evpn_mh.h"

DEFINE_MTYPE_STATIC(ZEBRA, ZINFO, "Zebra Interface Information");

#define ZEBRA_PTM_SUPPORT

DEFINE_HOOK(zebra_if_extra_info, (struct vty * vty, struct interface *ifp),
	    (vty, ifp));

DEFINE_MTYPE(ZEBRA, ZIF_DESC, "Intf desc");

static void if_down_del_nbr_connected(struct interface *ifp);

static void if_zebra_speed_update(struct event *thread)
{
	struct interface *ifp = EVENT_ARG(thread);
	struct zebra_if *zif = ifp->info;
	uint32_t new_speed;
	bool changed = false;
	int error = 0;

	new_speed = kernel_get_speed(ifp, &error);

	/* error may indicate vrf not available or
	 * interfaces not available.
	 * note that loopback & virtual interfaces can return 0 as speed
	 */
	if (error == INTERFACE_SPEED_ERROR_READ)
		return;

	if (new_speed != ifp->speed) {
		zlog_info("%s: %s old speed: %u new speed: %u", __func__,
			  ifp->name, ifp->speed, new_speed);
		ifp->speed = new_speed;
		if_add_update(ifp);
		changed = true;
	}

	if (changed || error == INTERFACE_SPEED_ERROR_UNKNOWN) {
#define SPEED_UPDATE_SLEEP_TIME 5
#define SPEED_UPDATE_COUNT_MAX (4 * 60 / SPEED_UPDATE_SLEEP_TIME)
		/*
		 * Some interfaces never actually have an associated speed
		 * with them ( I am looking at you bridges ).
		 * So instead of iterating forever, let's give the
		 * system 4 minutes to try to figure out the speed
		 * if after that it it's probably never going to become
		 * useful.
		 * Since I don't know all the wonderful types of interfaces
		 * that may come into existence in the future I am going
		 * to not update the system to keep track of that.  This
		 * is far simpler to just stop trying after 4 minutes
		 */
		if (error == INTERFACE_SPEED_ERROR_UNKNOWN &&
		    zif->speed_update_count == SPEED_UPDATE_COUNT_MAX)
			return;

		zif->speed_update_count++;
		event_add_timer(zrouter.master, if_zebra_speed_update, ifp,
				SPEED_UPDATE_SLEEP_TIME, &zif->speed_update);
		event_ignore_late_timer(zif->speed_update);
	}
}

static void zebra_if_node_destroy(route_table_delegate_t *delegate,
				  struct route_table *table,
				  struct route_node *node)
{
	if (node->info)
		list_delete((struct list **)&node->info);
	route_node_destroy(delegate, table, node);
}

static void zebra_if_nhg_dependents_free(struct zebra_if *zebra_if)
{
	nhg_connected_tree_free(&zebra_if->nhg_dependents);
}

static void zebra_if_nhg_dependents_init(struct zebra_if *zebra_if)
{
	nhg_connected_tree_init(&zebra_if->nhg_dependents);
}


route_table_delegate_t zebra_if_table_delegate = {
	.create_node = route_node_create,
	.destroy_node = zebra_if_node_destroy};

/* Called when new interface is added. */
static int if_zebra_new_hook(struct interface *ifp)
{
	struct zebra_if *zebra_if;

	zebra_if = XCALLOC(MTYPE_ZINFO, sizeof(struct zebra_if));
	zebra_if->ifp = ifp;

	zebra_if->multicast = IF_ZEBRA_DATA_UNSPEC;
	zebra_if->mpls_config = IF_ZEBRA_DATA_UNSPEC;
	zebra_if->shutdown = IF_ZEBRA_DATA_UNSPEC;

	zebra_if->link_nsid = NS_UNKNOWN;

	zebra_if_nhg_dependents_init(zebra_if);

	zebra_ptm_if_init(zebra_if);

	ifp->ptm_enable = zebra_ptm_get_enable_state();

	rtadv_if_init(zebra_if);

	zebra_evpn_mh_if_init(zebra_if);

	memset(&zebra_if->neigh_mac[0], 0, 6);

	/* Initialize installed address chains tree. */
	zebra_if->ipv4_subnets =
		route_table_init_with_delegate(&zebra_if_table_delegate);

	ifp->info = zebra_if;

	/*
	 * Some platforms are telling us that the interface is
	 * up and ready to go.  When we check the speed we
	 * sometimes get the wrong value.  Wait a couple
	 * of seconds and ask again.  Hopefully it's all settled
	 * down upon startup.
	 */
	zebra_if->speed_update_count = 0;
	event_add_timer(zrouter.master, if_zebra_speed_update, ifp, 15,
			&zebra_if->speed_update);
	event_ignore_late_timer(zebra_if->speed_update);

	return 0;
}

static void if_down_nhg_dependents(const struct interface *ifp)
{
	struct nhg_connected *rb_node_dep = NULL;
	struct zebra_if *zif = (struct zebra_if *)ifp->info;

	frr_each(nhg_connected_tree, &zif->nhg_dependents, rb_node_dep)
		zebra_nhg_check_valid(rb_node_dep->nhe);
}

static void if_nhg_dependents_release(const struct interface *ifp)
{
	struct nhg_connected *rb_node_dep = NULL;
	struct zebra_if *zif = (struct zebra_if *)ifp->info;

	frr_each(nhg_connected_tree, &zif->nhg_dependents, rb_node_dep) {
		rb_node_dep->nhe->ifp = NULL; /* Null it out */
		zebra_nhg_check_valid(rb_node_dep->nhe);
	}
}

/* Called when interface is deleted. */
static int if_zebra_delete_hook(struct interface *ifp)
{
	struct zebra_if *zebra_if;
	struct zebra_l2info_bond *bond;

	if (ifp->info) {
		zebra_if = ifp->info;

		/* If we set protodown, clear our reason now from the kernel */
		if (ZEBRA_IF_IS_PROTODOWN(zebra_if) && zebra_if->protodown_rc &&
		    !ZEBRA_IF_IS_PROTODOWN_ONLY_EXTERNAL(zebra_if))
			zebra_if_update_protodown_rc(ifp, true,
						     (zebra_if->protodown_rc &
						      ~ZEBRA_PROTODOWN_ALL));

		/* Free installed address chains tree. */
		if (zebra_if->ipv4_subnets)
			route_table_finish(zebra_if->ipv4_subnets);

		rtadv_if_fini(zebra_if);

		bond = &zebra_if->bond_info;
		if (bond && bond->mbr_zifs)
			list_delete(&bond->mbr_zifs);

		zebra_l2_bridge_if_cleanup(ifp);
		zebra_evpn_if_cleanup(zebra_if);
		zebra_evpn_mac_ifp_del(ifp);

		if_nhg_dependents_release(ifp);
		zebra_if_nhg_dependents_free(zebra_if);

		XFREE(MTYPE_ZIF_DESC, zebra_if->desc);

		EVENT_OFF(zebra_if->speed_update);

		XFREE(MTYPE_ZINFO, zebra_if);
	}

	return 0;
}

/* Build the table key */
static void if_build_key(uint32_t ifindex, struct prefix *p)
{
	p->family = AF_INET;
	p->prefixlen = IPV4_MAX_BITLEN;
	p->u.prefix4.s_addr = ifindex;
}

/* Link an interface in a per NS interface tree */
struct interface *if_link_per_ns(struct zebra_ns *ns, struct interface *ifp)
{
	struct prefix p;
	struct route_node *rn;

	if (ifp->ifindex == IFINDEX_INTERNAL)
		return NULL;

	if_build_key(ifp->ifindex, &p);
	rn = route_node_get(ns->if_table, &p);
	if (rn->info) {
		ifp = (struct interface *)rn->info;
		route_unlock_node(rn); /* get */
		return ifp;
	}

	rn->info = ifp;
	ifp->node = rn;

	return ifp;
}

/* Delete a VRF. This is called in vrf_terminate(). */
void if_unlink_per_ns(struct interface *ifp)
{
	if (!ifp->node)
		return;

	ifp->node->info = NULL;
	route_unlock_node(ifp->node);
	ifp->node = NULL;
}

/* Look up an interface by identifier within a NS */
struct interface *if_lookup_by_index_per_ns(struct zebra_ns *ns,
					    uint32_t ifindex)
{
	struct prefix p;
	struct route_node *rn;
	struct interface *ifp = NULL;

	if_build_key(ifindex, &p);
	rn = route_node_lookup(ns->if_table, &p);
	if (rn) {
		ifp = (struct interface *)rn->info;
		route_unlock_node(rn); /* lookup */
	}
	return ifp;
}

/* Look up an interface by name within a NS */
struct interface *if_lookup_by_name_per_ns(struct zebra_ns *ns,
					   const char *ifname)
{
	struct route_node *rn;
	struct interface *ifp;

	for (rn = route_top(ns->if_table); rn; rn = route_next(rn)) {
		ifp = (struct interface *)rn->info;
		if (ifp && strcmp(ifp->name, ifname) == 0) {
			route_unlock_node(rn);
			return (ifp);
		}
	}

	return NULL;
}

struct interface *if_lookup_by_index_per_nsid(ns_id_t ns_id, uint32_t ifindex)
{
	struct zebra_ns *zns;

	zns = zebra_ns_lookup(ns_id);
	return zns ? if_lookup_by_index_per_ns(zns, ifindex) : NULL;
}

const char *ifindex2ifname_per_ns(struct zebra_ns *zns, unsigned int ifindex)
{
	struct interface *ifp;

	return ((ifp = if_lookup_by_index_per_ns(zns, ifindex)) != NULL)
		       ? ifp->name
		       : "unknown";
}

/* Tie an interface address to its derived subnet list of addresses. */
int if_subnet_add(struct interface *ifp, struct connected *ifc)
{
	struct route_node *rn;
	struct zebra_if *zebra_if;
	struct prefix cp;
	struct list *addr_list;

	assert(ifp && ifp->info && ifc);
	zebra_if = ifp->info;

	/* Get address derived subnet node and associated address list, while
	   marking
	   address secondary attribute appropriately. */
	cp = *CONNECTED_PREFIX(ifc);
	apply_mask(&cp);
	rn = route_node_get(zebra_if->ipv4_subnets, &cp);

	if ((addr_list = rn->info))
		SET_FLAG(ifc->flags, ZEBRA_IFA_SECONDARY);
	else {
		UNSET_FLAG(ifc->flags, ZEBRA_IFA_SECONDARY);
		rn->info = addr_list = list_new();
		route_lock_node(rn);
	}

	/* Tie address at the tail of address list. */
	listnode_add(addr_list, ifc);

	/* Return list element count. */
	return (addr_list->count);
}

/* Untie an interface address from its derived subnet list of addresses. */
int if_subnet_delete(struct interface *ifp, struct connected *ifc)
{
	struct route_node *rn;
	struct zebra_if *zebra_if;
	struct list *addr_list;
	struct prefix cp;

	assert(ifp && ifp->info && ifc);
	zebra_if = ifp->info;

	cp = *CONNECTED_PREFIX(ifc);
	apply_mask(&cp);

	/* Get address derived subnet node. */
	rn = route_node_lookup(zebra_if->ipv4_subnets, &cp);
	if (!(rn && rn->info)) {
		flog_warn(EC_ZEBRA_REMOVE_ADDR_UNKNOWN_SUBNET,
			  "Trying to remove an address from an unknown subnet. (please report this bug)");
		return -1;
	}
	route_unlock_node(rn);

	/* Untie address from subnet's address list. */
	addr_list = rn->info;

	/* Deleting an address that is not registered is a bug.
	 * In any case, we shouldn't decrement the lock counter if the address
	 * is unknown. */
	if (!listnode_lookup(addr_list, ifc)) {
		flog_warn(
			EC_ZEBRA_REMOVE_UNREGISTERED_ADDR,
			"Trying to remove an address from a subnet where it is not currently registered. (please report this bug)");
		return -1;
	}

	listnode_delete(addr_list, ifc);
	route_unlock_node(rn);

	/* Return list element count, if not empty. */
	if (addr_list->count) {
		/* If deleted address is primary, mark subsequent one as such
		 * and distribute. */
		if (!CHECK_FLAG(ifc->flags, ZEBRA_IFA_SECONDARY)) {
			ifc = listgetdata(
				(struct listnode *)listhead(addr_list));
			zebra_interface_address_delete_update(ifp, ifc);
			UNSET_FLAG(ifc->flags, ZEBRA_IFA_SECONDARY);
			/* XXX: Linux kernel removes all the secondary addresses
			 * when the primary
			 * address is removed. We could try to work around that,
			 * though this is
			 * non-trivial. */
			zebra_interface_address_add_update(ifp, ifc);
		}

		return addr_list->count;
	}

	/* Otherwise, free list and route node. */
	list_delete(&addr_list);
	rn->info = NULL;
	route_unlock_node(rn);

	return 0;
}

/* if_flags_mangle: A place for hacks that require mangling
 * or tweaking the interface flags.
 *
 * ******************** Solaris flags hacks **************************
 *
 * Solaris IFF_UP flag reflects only the primary interface as the
 * routing socket only sends IFINFO for the primary interface.  Hence
 * ~IFF_UP does not per se imply all the logical interfaces are also
 * down - which we only know of as addresses. Instead we must determine
 * whether the interface really is up or not according to how many
 * addresses are still attached. (Solaris always sends RTM_DELADDR if
 * an interface, logical or not, goes ~IFF_UP).
 *
 * Ie, we mangle IFF_UP to *additionally* reflect whether or not there
 * are addresses left in struct connected, not just the actual underlying
 * IFF_UP flag.
 *
 * We must hence remember the real state of IFF_UP, which we do in
 * struct zebra_if.primary_state.
 *
 * Setting IFF_UP within zebra to administratively shutdown the
 * interface will affect only the primary interface/address on Solaris.
 ************************End Solaris flags hacks ***********************
 */
static void if_flags_mangle(struct interface *ifp, uint64_t *newflags)
{
	return;
}

/* Update the flags field of the ifp with the new flag set provided.
 * Take whatever actions are required for any changes in flags we care
 * about.
 *
 * newflags should be the raw value, as obtained from the OS.
 */
void if_flags_update(struct interface *ifp, uint64_t newflags)
{
	if_flags_mangle(ifp, &newflags);

	if (if_is_no_ptm_operative(ifp)) {
		/* operative -> inoperative? */
		ifp->flags = newflags;
		if (!if_is_operative(ifp))
			if_down(ifp);
	} else {
		/* inoperative -> operative? */
		ifp->flags = newflags;
		if (if_is_operative(ifp))
			if_up(ifp, true);
	}
}

/* Wake up configured address if it is not in current kernel
   address. */
void if_addr_wakeup(struct interface *ifp)
{
	struct connected *ifc;
	struct prefix *p;
	enum zebra_dplane_result dplane_res;

	frr_each_safe (if_connected, ifp->connected, ifc) {
		p = ifc->address;

		if (CHECK_FLAG(ifc->conf, ZEBRA_IFC_CONFIGURED)
		    && !CHECK_FLAG(ifc->conf, ZEBRA_IFC_QUEUED)) {
			/* Address check. */
			if (p->family == AF_INET) {
				if (!if_is_up(ifp)) {
					/* Assume zebra is configured like
					 * following:
					 *
					 *   interface gre0
					 *    ip addr 192.0.2.1/24
					 *   !
					 *
					 * As soon as zebra becomes first aware
					 * that gre0 exists in the
					 * kernel, it will set gre0 up and
					 * configure its addresses.
					 *
					 * (This may happen at startup when the
					 * interface already exists
					 * or during runtime when the interface
					 * is added to the kernel)
					 *
					 * XXX: IRDP code is calling here via
					 * if_add_update - this seems
					 * somewhat weird.
					 * XXX: RUNNING is not a settable flag
					 * on any system
					 * I (paulj) am aware of.
					*/
					if_set_flags(ifp, IFF_UP | IFF_RUNNING);
					if_refresh(ifp);
				}

				dplane_res = dplane_intf_addr_set(ifp, ifc);
				if (dplane_res ==
				    ZEBRA_DPLANE_REQUEST_FAILURE) {
					flog_err_sys(
						EC_ZEBRA_IFACE_ADDR_ADD_FAILED,
						"Can't set interface's address: %s",
						dplane_res2str(dplane_res));
					continue;
				}

				SET_FLAG(ifc->conf, ZEBRA_IFC_QUEUED);
				/* The address will be advertised to zebra
				 * clients when the notification
				 * from the kernel has been received.
				 * It will also be added to the interface's
				 * subnet list then. */
			}
			if (p->family == AF_INET6) {
				if (!if_is_up(ifp)) {
					/* See long comment above */
					if_set_flags(ifp, IFF_UP | IFF_RUNNING);
					if_refresh(ifp);
				}


				dplane_res = dplane_intf_addr_set(ifp, ifc);
				if (dplane_res ==
				    ZEBRA_DPLANE_REQUEST_FAILURE) {
					flog_err_sys(
						EC_ZEBRA_IFACE_ADDR_ADD_FAILED,
						"Can't set interface's address: %s",
						dplane_res2str(dplane_res));
					continue;
				}

				SET_FLAG(ifc->conf, ZEBRA_IFC_QUEUED);
				/* The address will be advertised to zebra
				 * clients when the notification
				 * from the kernel has been received. */
			}
		}
	}
}

/* Handle interface addition */
void if_add_update(struct interface *ifp)
{
	struct zebra_if *if_data;
	struct zebra_ns *zns;
	struct zebra_vrf *zvrf = ifp->vrf->info;

	/* case interface populate before vrf enabled */
	if (zvrf->zns)
		zns = zvrf->zns;
	else
		zns = zebra_ns_lookup(NS_DEFAULT);
	if_link_per_ns(zns, ifp);
	if_data = ifp->info;
	assert(if_data);

	if (if_data->multicast == IF_ZEBRA_DATA_ON)
		if_set_flags(ifp, IFF_MULTICAST);
	else if (if_data->multicast == IF_ZEBRA_DATA_OFF)
		if_unset_flags(ifp, IFF_MULTICAST);

	zebra_ptm_if_set_ptm_state(ifp, if_data);

	zebra_interface_add_update(ifp);

	if (!CHECK_FLAG(ifp->status, ZEBRA_INTERFACE_ACTIVE)) {
		SET_FLAG(ifp->status, ZEBRA_INTERFACE_ACTIVE);

		if (if_data->shutdown == IF_ZEBRA_DATA_ON) {
			if (IS_ZEBRA_DEBUG_KERNEL) {
				zlog_debug(
					"interface %s vrf %s(%u) index %d is shutdown. Won't wake it up.",
					ifp->name, ifp->vrf->name,
					ifp->vrf->vrf_id, ifp->ifindex);
			}

			return;
		}

		if_addr_wakeup(ifp);

		if (if_data->mpls_config == IF_ZEBRA_DATA_ON)
			dplane_intf_mpls_modify_state(ifp, true);
		else if (if_data->mpls_config == IF_ZEBRA_DATA_OFF)
			dplane_intf_mpls_modify_state(ifp, false);

		if (IS_ZEBRA_DEBUG_KERNEL)
			zlog_debug(
				"interface %s vrf %s(%u) index %d becomes active.",
				ifp->name, ifp->vrf->name, ifp->vrf->vrf_id,
				ifp->ifindex);

	} else {
		if (IS_ZEBRA_DEBUG_KERNEL)
			zlog_debug("interface %s vrf %s(%u) index %d is added.",
				   ifp->name, ifp->vrf->name, ifp->vrf->vrf_id,
				   ifp->ifindex);
	}
}

/* Install connected routes corresponding to an interface. */
static void if_install_connected(struct interface *ifp)
{
	struct connected *ifc;

	frr_each (if_connected, ifp->connected, ifc) {
		if (CHECK_FLAG(ifc->conf, ZEBRA_IFC_REAL))
			zebra_interface_address_add_update(ifp, ifc);

		connected_up(ifp, ifc);
	}
}

/* Uninstall connected routes corresponding to an interface. */
static void if_uninstall_connected(struct interface *ifp)
{
	struct connected *ifc;

	frr_each_safe (if_connected, ifp->connected, ifc) {
		zebra_interface_address_delete_update(ifp, ifc);
		connected_down(ifp, ifc);
	}
}

/* Uninstall and delete connected routes corresponding to an interface. */
/* TODO - Check why IPv4 handling here is different from install or if_down */
static void if_delete_connected(struct interface *ifp)
{
	struct connected *ifc, *ifc_next;
	struct prefix cp;
	struct route_node *rn;
	struct zebra_if *zebra_if;

	zebra_if = ifp->info;

	for (ifc = if_connected_first(ifp->connected); ifc; ifc = ifc_next) {
		ifc_next = if_connected_next(ifp->connected, ifc);

		cp = *CONNECTED_PREFIX(ifc);
		apply_mask(&cp);

		if (cp.family == AF_INET
		    && (rn = route_node_lookup(zebra_if->ipv4_subnets, &cp))) {
			struct listnode *anode;
			struct listnode *next;
			struct listnode *first;
			struct list *addr_list;

			route_unlock_node(rn);
			addr_list = (struct list *)rn->info;

			/* Remove addresses, secondaries first. */
			first = listhead(addr_list);
			if (first)
				for (anode = first->next; anode || first;
				     anode = next) {
					if (!anode) {
						anode = first;
						first = NULL;
					}
					next = anode->next;

					ifc = listgetdata(anode);
					connected_down(ifp, ifc);

					/* XXX: We have to send notifications
					 * here explicitly, because we destroy
					 * the ifc before receiving the
					 * notification about the address being
					 * deleted.
					 */
					zebra_interface_address_delete_update(
						ifp, ifc);

					UNSET_FLAG(ifc->conf, ZEBRA_IFC_REAL);
					UNSET_FLAG(ifc->conf, ZEBRA_IFC_QUEUED);

					/* Remove from subnet chain. */
					list_delete_node(addr_list, anode);
					route_unlock_node(rn);

					/* Remove from interface address list
					 * (unconditionally). */
					if (!CHECK_FLAG(ifc->conf,
							ZEBRA_IFC_CONFIGURED)) {
						if (ifc == ifc_next)
							ifc_next = if_connected_next(
								ifp->connected,
								ifc);

						if_connected_del(ifp->connected,
								 ifc);
						connected_free(&ifc);
					}
				}

			/* Free chain list and respective route node. */
			list_delete(&addr_list);
			rn->info = NULL;
			route_unlock_node(rn);
		} else if (cp.family == AF_INET6) {
			connected_down(ifp, ifc);

			zebra_interface_address_delete_update(ifp, ifc);

			UNSET_FLAG(ifc->conf, ZEBRA_IFC_REAL);
			UNSET_FLAG(ifc->conf, ZEBRA_IFC_QUEUED);

			if (!CHECK_FLAG(ifc->conf, ZEBRA_IFC_CONFIGURED)) {
				if_connected_del(ifp->connected, ifc);
				connected_free(&ifc);
			}
		}
	}
}

/* Handle an interface delete event */
void if_delete_update(struct interface **pifp)
{
	struct zebra_if *zif;
	struct interface *ifp = *pifp;

	if (if_is_up(ifp)) {
		flog_err(
			EC_LIB_INTERFACE,
			"interface %s vrf %s(%u) index %d is still up while being deleted.",
			ifp->name, ifp->vrf->name, ifp->vrf->vrf_id,
			ifp->ifindex);
		return;
	}

	if (!CHECK_FLAG(ifp->status, ZEBRA_INTERFACE_ACTIVE))
		return;

	/* Mark interface as inactive */
	UNSET_FLAG(ifp->status, ZEBRA_INTERFACE_ACTIVE);

	if (IS_ZEBRA_DEBUG_KERNEL)
		zlog_debug("interface %s vrf %s(%u) index %d is now inactive.",
			   ifp->name, ifp->vrf->name, ifp->vrf->vrf_id,
			   ifp->ifindex);

	/* Delete connected routes from the kernel. */
	if_delete_connected(ifp);

	/* if the ifp is in a vrf, move it to default so vrf can be deleted if
	 * desired. This operation is not done for netns implementation to avoid
	 * collision with interface with the same name in the default vrf (can
	 * occur with this implementation whereas it is not possible with
	 * vrf-lite).
	 */
	if (ifp->vrf->vrf_id && !vrf_is_backend_netns())
		if_handle_vrf_change(ifp, VRF_DEFAULT);

	/* Send out notification on interface delete. */
	zebra_interface_delete_update(ifp);

	if_unlink_per_ns(ifp);

	/* Update ifindex after distributing the delete message.  This is in
	   case any client needs to have the old value of ifindex available
	   while processing the deletion.  Each client daemon is responsible
	   for setting ifindex to IFINDEX_INTERNAL after processing the
	   interface deletion message. */
	if_set_index(ifp, IFINDEX_INTERNAL);
	ifp->node = NULL;

	UNSET_FLAG(ifp->status, ZEBRA_INTERFACE_VRF_LOOPBACK);

	/* Reset some zebra interface params to default values. */
	zif = ifp->info;
	if (zif) {
		zebra_evpn_if_cleanup(zif);
		zif->zif_type = ZEBRA_IF_OTHER;
		zif->zif_slave_type = ZEBRA_IF_SLAVE_NONE;
		memset(&zif->l2info, 0, sizeof(union zebra_l2if_info));
		memset(&zif->brslave_info, 0,
		       sizeof(struct zebra_l2info_brslave));
		zebra_evpn_mac_ifp_del(ifp);
	}

	if (!ifp->configured) {
		if (IS_ZEBRA_DEBUG_KERNEL)
			zlog_debug("interface %s is being deleted from the system",
				   ifp->name);
		if_delete(pifp);
	}
}

/* VRF change for an interface */
void if_handle_vrf_change(struct interface *ifp, vrf_id_t vrf_id)
{
	vrf_id_t old_vrf_id;

	old_vrf_id = ifp->vrf->vrf_id;

	/* Uninstall connected routes. */
	if_uninstall_connected(ifp);

	/* Delete any IPv4 neighbors created to implement RFC 5549 */
	if_nbr_ipv6ll_to_ipv4ll_neigh_del_all(ifp);

	/* Delete all neighbor addresses learnt through IPv6 RA */
	if_down_del_nbr_connected(ifp);

	/* Send out notification on interface VRF change. */
	/* This is to issue a DELETE, as appropriate. */
	zebra_interface_vrf_update_del(ifp, vrf_id);

	if (if_is_vrf(ifp))
		return;

	/* update VRF */
	if_update_to_new_vrf(ifp, vrf_id);

	/* Send out notification on interface VRF change. */
	/* This is to issue an ADD, if needed. */
	zebra_interface_vrf_update_add(ifp, old_vrf_id);
}

static void ipv6_ll_address_to_mac(struct in6_addr *address, uint8_t *mac)
{
	mac[0] = address->s6_addr[8] ^ 0x02;
	mac[1] = address->s6_addr[9];
	mac[2] = address->s6_addr[10];
	mac[3] = address->s6_addr[13];
	mac[4] = address->s6_addr[14];
	mac[5] = address->s6_addr[15];
}

void if_nbr_mac_to_ipv4ll_neigh_update(struct interface *ifp,
				       char mac[6],
				       struct in6_addr *address,
				       int add)
{
	struct zebra_vrf *zvrf = ifp->vrf->info;
	struct zebra_if *zif = ifp->info;
	char buf[16] = "169.254.0.1";
	struct in_addr ipv4_ll;
	ns_id_t ns_id;

	inet_pton(AF_INET, buf, &ipv4_ll);

	ns_id = zvrf->zns->ns_id;

	/*
	 * Remove and re-add any existing neighbor entry for this address,
	 * since Netlink doesn't currently offer update message types.
	 */
	kernel_neigh_update(0, ifp->ifindex, (void *)&ipv4_ll.s_addr, mac, 6,
			    ns_id, AF_INET, true);

	/* Add new neighbor entry.
	 *
	 * We force installation even if current neighbor entry is the same.
	 * Since this function is used to refresh our MAC entries after an
	 * interface flap, if we don't force in our custom entries with their
	 * state set to PERMANENT or REACHABLE then the kernel will attempt to
	 * resolve our leftover entries, fail, mark them unreachable and then
	 * they'll be useless to us.
	 */
	if (add)
		kernel_neigh_update(add, ifp->ifindex, (void *)&ipv4_ll.s_addr,
				    mac, 6, ns_id, AF_INET, true);

	memcpy(&zif->neigh_mac[0], &mac[0], 6);

	/*
	 * We need to note whether or not we originated a v6
	 * neighbor entry for this interface.  So that when
	 * someone unwisely accidentally deletes this entry
	 * we can shove it back in.
	 */
	zif->v6_2_v4_ll_neigh_entry = !!add;
	memcpy(&zif->v6_2_v4_ll_addr6, address, sizeof(*address));

	zvrf->neigh_updates++;
}

void if_nbr_ipv6ll_to_ipv4ll_neigh_update(struct interface *ifp,
					  struct in6_addr *address, int add)
{

	char mac[6];

	ipv6_ll_address_to_mac(address, (uint8_t *)mac);
	if_nbr_mac_to_ipv4ll_neigh_update(ifp, mac, address, add);
}

static void if_nbr_ipv6ll_to_ipv4ll_neigh_add_all(struct interface *ifp)
{
	if (listhead(ifp->nbr_connected)) {
		struct nbr_connected *nbr_connected;
		struct listnode *node;

		for (ALL_LIST_ELEMENTS_RO(ifp->nbr_connected, node,
					  nbr_connected))
			if_nbr_ipv6ll_to_ipv4ll_neigh_update(
				ifp, &nbr_connected->address->u.prefix6, 1);
	}
}

void if_nbr_ipv6ll_to_ipv4ll_neigh_del_all(struct interface *ifp)
{
	if (listhead(ifp->nbr_connected)) {
		struct nbr_connected *nbr_connected;
		struct listnode *node;

		for (ALL_LIST_ELEMENTS_RO(ifp->nbr_connected, node,
					  nbr_connected))
			if_nbr_ipv6ll_to_ipv4ll_neigh_update(
				ifp, &nbr_connected->address->u.prefix6, 0);
	}
}

static void if_down_del_nbr_connected(struct interface *ifp)
{
	struct nbr_connected *nbr_connected;
	struct listnode *node, *nnode;

	for (ALL_LIST_ELEMENTS(ifp->nbr_connected, node, nnode,
			       nbr_connected)) {
		listnode_delete(ifp->nbr_connected, nbr_connected);
		nbr_connected_free(nbr_connected);
	}
}

void if_nhg_dependents_add(struct interface *ifp, struct nhg_hash_entry *nhe)
{
	if (ifp->info) {
		struct zebra_if *zif = (struct zebra_if *)ifp->info;

		nhg_connected_tree_add_nhe(&zif->nhg_dependents, nhe);
	}
}

void if_nhg_dependents_del(struct interface *ifp, struct nhg_hash_entry *nhe)
{
	if (ifp->info) {
		struct zebra_if *zif = (struct zebra_if *)ifp->info;

		nhg_connected_tree_del_nhe(&zif->nhg_dependents, nhe);
	}
}

unsigned int if_nhg_dependents_count(const struct interface *ifp)
{
	if (ifp->info) {
		struct zebra_if *zif = (struct zebra_if *)ifp->info;

		return nhg_connected_tree_count(&zif->nhg_dependents);
	}

	return 0;
}


bool if_nhg_dependents_is_empty(const struct interface *ifp)
{
	if (ifp->info) {
		struct zebra_if *zif = (struct zebra_if *)ifp->info;

		return nhg_connected_tree_is_empty(&zif->nhg_dependents);
	}

	return false;
}

/* Interface is up. */
void if_up(struct interface *ifp, bool install_connected)
{
	struct zebra_if *zif;
	struct interface *link_if;

	zif = ifp->info;
	zif->up_count++;
	frr_timestamp(2, zif->up_last, sizeof(zif->up_last));

	/* Notify the protocol daemons. */
	if (ifp->ptm_enable && (ifp->ptm_status == ZEBRA_PTM_STATUS_DOWN)) {
		flog_warn(EC_ZEBRA_PTM_NOT_READY,
			  "%s: interface %s hasn't passed ptm check",
			  __func__, ifp->name);
		return;
	}
	zebra_interface_up_update(ifp);

	if_nbr_ipv6ll_to_ipv4ll_neigh_add_all(ifp);

	rtadv_if_up(zif);

	/* Install connected routes to the kernel. */
	if (install_connected)
		if_install_connected(ifp);

	/* Handle interface up for specific types for EVPN. Non-VxLAN interfaces
	 * are checked to see if (remote) neighbor entries need to be installed
	 * on them for ARP suppression.
	 */
	if (IS_ZEBRA_IF_VXLAN(ifp))
		zebra_vxlan_if_up(ifp);
	else if (IS_ZEBRA_IF_BRIDGE(ifp)) {
		link_if = ifp;
		zebra_vxlan_svi_up(ifp, link_if);
	} else if (IS_ZEBRA_IF_VLAN(ifp)) {
		link_if = zif->link;
		if (link_if)
			zebra_vxlan_svi_up(ifp, link_if);
	} else if (IS_ZEBRA_IF_MACVLAN(ifp)) {
		zebra_vxlan_macvlan_up(ifp);
	}

	if (zif->es_info.es)
		zebra_evpn_es_if_oper_state_change(zif, true /*up*/);

	if (zif->flags & ZIF_FLAG_EVPN_MH_UPLINK)
		zebra_evpn_mh_uplink_oper_update(zif);

	event_add_timer(zrouter.master, if_zebra_speed_update, ifp, 0,
			&zif->speed_update);
	event_ignore_late_timer(zif->speed_update);

	if_addr_wakeup(ifp);
}

/* Interface goes down.  We have to manage different behavior of based
   OS. */
void if_down(struct interface *ifp)
{
	struct zebra_if *zif;
	struct interface *link_if;

	zif = ifp->info;
	zif->down_count++;
	frr_timestamp(2, zif->down_last, sizeof(zif->down_last));

	if_down_nhg_dependents(ifp);

	/* Handle interface down for specific types for EVPN. Non-VxLAN
	 * interfaces
	 * are checked to see if (remote) neighbor entries need to be purged
	 * for ARP suppression.
	 */
	if (IS_ZEBRA_IF_VXLAN(ifp))
		zebra_vxlan_if_down(ifp);
	else if (IS_ZEBRA_IF_BRIDGE(ifp)) {
		link_if = ifp;
		zebra_vxlan_svi_down(ifp, link_if);
	} else if (IS_ZEBRA_IF_VLAN(ifp)) {
		link_if = zif->link;
		if (link_if)
			zebra_vxlan_svi_down(ifp, link_if);
	} else if (IS_ZEBRA_IF_MACVLAN(ifp)) {
		zebra_vxlan_macvlan_down(ifp);
	}

	if (zif->es_info.es)
		zebra_evpn_es_if_oper_state_change(zif, false /*up*/);

	if (zif->flags & ZIF_FLAG_EVPN_MH_UPLINK)
		zebra_evpn_mh_uplink_oper_update(zif);

	/* Notify to the protocol daemons. */
	zebra_interface_down_update(ifp);

	/* Uninstall connected routes from the kernel. */
	if_uninstall_connected(ifp);

	if_nbr_ipv6ll_to_ipv4ll_neigh_del_all(ifp);

	/* Delete all neighbor addresses learnt through IPv6 RA */
	if_down_del_nbr_connected(ifp);
}

void if_refresh(struct interface *ifp)
{
#ifndef GNU_LINUX
	if_get_flags(ifp);
#endif
}

void zebra_if_update_link(struct interface *ifp, ifindex_t link_ifindex,
			  ns_id_t ns_id)
{
	struct zebra_if *zif;

	zif = (struct zebra_if *)ifp->info;
	zif->link_nsid = ns_id;
	zif->link_ifindex = link_ifindex;
	zif->link = if_lookup_by_index_per_ns(zebra_ns_lookup(ns_id),
					      link_ifindex);
}

/*
 * during initial link dump kernel does not order lower devices before
 * upper devices so we need to fixup link dependencies at the end of dump
 */
void zebra_if_update_all_links(struct zebra_ns *zns)
{
	struct route_node *rn;
	struct interface *ifp;
	struct zebra_if *zif;

	if (IS_ZEBRA_DEBUG_KERNEL)
		zlog_info("fixup link dependencies");

	for (rn = route_top(zns->if_table); rn; rn = route_next(rn)) {
		ifp = (struct interface *)rn->info;
		if (!ifp)
			continue;
		zif = ifp->info;
		/* update bond-member to bond linkages */
		if ((IS_ZEBRA_IF_BOND_SLAVE(ifp))
		    && (zif->bondslave_info.bond_ifindex != IFINDEX_INTERNAL)
		    && !zif->bondslave_info.bond_if) {
			if (IS_ZEBRA_DEBUG_EVPN_MH_ES || IS_ZEBRA_DEBUG_KERNEL)
				zlog_debug("bond mbr %s map to bond %d",
					   zif->ifp->name,
					   zif->bondslave_info.bond_ifindex);
			zebra_l2_map_slave_to_bond(zif, ifp->vrf->vrf_id);
		}

		/* update SVI linkages */
		if ((zif->link_ifindex != IFINDEX_INTERNAL) && !zif->link) {
			zif->link = if_lookup_by_index_per_nsid(
				zif->link_nsid, zif->link_ifindex);
			if (IS_ZEBRA_DEBUG_KERNEL)
				zlog_debug("interface %s/%d's lower fixup to %s/%d",
						ifp->name, ifp->ifindex,
						zif->link?zif->link->name:"unk",
						zif->link_ifindex);
		}

		/* Update VLAN<=>SVI map */
		if (IS_ZEBRA_IF_VLAN(ifp))
			zebra_evpn_acc_bd_svi_set(zif, NULL,
						  !!if_is_operative(ifp));
	}
}

static bool if_ignore_set_protodown(const struct interface *ifp, bool new_down,
				    uint32_t new_protodown_rc)
{
	struct zebra_if *zif;
	bool old_down, old_set_down, old_unset_down;

	zif = ifp->info;

	/*
	 * FRR does not have enough data to make this request
	 */
	if (ifp->ifindex == IFINDEX_INTERNAL)
		return true;

	/* Current state as we know it */
	old_down = !!(ZEBRA_IF_IS_PROTODOWN(zif));
	old_set_down = !!CHECK_FLAG(zif->flags, ZIF_FLAG_SET_PROTODOWN);
	old_unset_down = !!CHECK_FLAG(zif->flags, ZIF_FLAG_UNSET_PROTODOWN);

	if (new_protodown_rc == zif->protodown_rc) {
		/* Early return if already down & reason bitfield matches */
		if (new_down == old_down) {
			if (IS_ZEBRA_DEBUG_KERNEL)
				zlog_debug(
					"Ignoring request to set protodown %s for interface %s (%u): protodown %s is already set (reason bitfield: old 0x%x new 0x%x)",
					new_down ? "on" : "off", ifp->name,
					ifp->ifindex, new_down ? "on" : "off",
					zif->protodown_rc, new_protodown_rc);

			return true;
		}

		/* Early return if already set queued & reason bitfield matches
		 */
		if (new_down && old_set_down) {
			if (IS_ZEBRA_DEBUG_KERNEL)
				zlog_debug(
					"Ignoring request to set protodown %s for interface %s (%u): protodown %s is already queued to dplane (reason bitfield: old 0x%x new 0x%x)",
					new_down ? "on" : "off", ifp->name,
					ifp->ifindex, new_down ? "on" : "off",
					zif->protodown_rc, new_protodown_rc);

			return true;
		}

		/* Early return if already unset queued & reason bitfield
		 * matches */
		if (!new_down && old_unset_down) {
			if (IS_ZEBRA_DEBUG_KERNEL)
				zlog_debug(
					"Ignoring request to set protodown %s for interface %s (%u): protodown %s is already queued to dplane (reason bitfield: old 0x%x new 0x%x)",
					new_down ? "on" : "off", ifp->name,
					ifp->ifindex, new_down ? "on" : "off",
					zif->protodown_rc, new_protodown_rc);

			return true;
		}
	}

	return false;
}

int zebra_if_update_protodown_rc(struct interface *ifp, bool new_down,
				 uint32_t new_protodown_rc)
{
	struct zebra_if *zif;

	zif = ifp->info;

	/* Check if we already have this state or it's queued */
	if (if_ignore_set_protodown(ifp, new_down, new_protodown_rc))
		return 1;

	zlog_info(
		"Setting protodown %s - interface %s (%u): reason bitfield change from 0x%x --> 0x%x",
		new_down ? "on" : "off", ifp->name, ifp->ifindex,
		zif->protodown_rc, new_protodown_rc);

	zif->protodown_rc = new_protodown_rc;

	if (new_down)
		SET_FLAG(zif->flags, ZIF_FLAG_SET_PROTODOWN);
	else
		SET_FLAG(zif->flags, ZIF_FLAG_UNSET_PROTODOWN);

#ifdef HAVE_NETLINK
	dplane_intf_update(ifp);
#else
	zlog_warn("Protodown is not supported on this platform");
#endif
	return 0;
}

int zebra_if_set_protodown(struct interface *ifp, bool new_down,
			   enum protodown_reasons new_reason)
{
	struct zebra_if *zif;
	uint32_t new_protodown_rc;

	zif = ifp->info;

	if (new_down)
		new_protodown_rc = zif->protodown_rc | new_reason;
	else
		new_protodown_rc = zif->protodown_rc & ~new_reason;

	return zebra_if_update_protodown_rc(ifp, new_down, new_protodown_rc);
}

/*
 * Handle an interface events based on info in a dplane context object.
 * This runs in the main pthread, using the info in the context object to
 * modify an interface.
 */
static void zebra_if_addr_update_ctx(struct zebra_dplane_ctx *ctx,
				     struct interface *ifp)
{
	uint8_t flags = 0;
	const char *label = NULL;
	uint32_t metric = METRIC_MAX;
	const struct prefix *addr, *dest = NULL;
	enum dplane_op_e op;

	if (!ifp)
		return;

	op = dplane_ctx_get_op(ctx);
	addr = dplane_ctx_get_intf_addr(ctx);

	if (IS_ZEBRA_DEBUG_KERNEL)
		zlog_debug("%s: %s: ifindex %s(%u), addr %pFX", __func__,
			   dplane_op2str(dplane_ctx_get_op(ctx)), ifp->name,
			   ifp->ifindex, addr);

	/* Is there a peer or broadcast address? */
	dest = dplane_ctx_get_intf_dest(ctx);
	if (dest->prefixlen == 0)
		dest = NULL;

	if (dplane_ctx_intf_is_connected(ctx))
		SET_FLAG(flags, ZEBRA_IFA_PEER);

	/* Flags. */
	if (dplane_ctx_intf_is_secondary(ctx))
		SET_FLAG(flags, ZEBRA_IFA_SECONDARY);

	if (dplane_ctx_intf_is_noprefixroute(ctx))
		SET_FLAG(flags, ZEBRA_IFA_NOPREFIXROUTE);

	/* Label? */
	if (dplane_ctx_intf_has_label(ctx))
		label = dplane_ctx_get_intf_label(ctx);

	if (label && strcmp(ifp->name, label) == 0)
		label = NULL;

	metric = dplane_ctx_get_intf_metric(ctx);

	/* Register interface address to the interface. */
	if (addr->family == AF_INET) {
		if (op == DPLANE_OP_INTF_ADDR_ADD)
			connected_add_ipv4(
				ifp, flags, &addr->u.prefix4, addr->prefixlen,
				dest ? &dest->u.prefix4 : NULL, label, metric);
		else if (CHECK_FLAG(flags, ZEBRA_IFA_PEER)) {
			/* Delete with a peer address */
			connected_delete_ipv4(ifp, flags, &addr->u.prefix4,
					      addr->prefixlen,
					      &dest->u.prefix4);
		} else
			connected_delete_ipv4(ifp, flags, &addr->u.prefix4,
					      addr->prefixlen, NULL);
	}

	if (addr->family == AF_INET6) {
		if (op == DPLANE_OP_INTF_ADDR_ADD) {
			connected_add_ipv6(ifp, flags, &addr->u.prefix6,
					   dest ? &dest->u.prefix6 : NULL,
					   addr->prefixlen, label, metric);
		} else
			connected_delete_ipv6(ifp, &addr->u.prefix6, NULL,
					      addr->prefixlen);
	}

	/*
	 * Linux kernel does not send route delete on interface down/addr del
	 * so we have to re-process routes it owns (i.e. kernel routes)
	 */
	if (op != DPLANE_OP_INTF_ADDR_ADD)
		rib_update(RIB_UPDATE_KERNEL);
}

static void zebra_if_update_ctx(struct zebra_dplane_ctx *ctx,
				struct interface *ifp)
{
	enum zebra_dplane_result dp_res;
	struct zebra_if *zif;
	bool pd_reason_val;
	bool down;

	if (!ifp) {
		if (IS_ZEBRA_DEBUG_KERNEL)
			zlog_debug("%s: Can't find ifp", __func__);

		return;
	}

	dp_res = dplane_ctx_get_status(ctx);
	pd_reason_val = dplane_ctx_get_intf_pd_reason_val(ctx);
	down = dplane_ctx_intf_is_protodown(ctx);

	if (IS_ZEBRA_DEBUG_KERNEL)
		zlog_debug("%s: %s: if %s(%u) ctx-protodown %s ctx-reason %d",
			   __func__, dplane_op2str(dplane_ctx_get_op(ctx)),
			   ifp->name, ifp->ifindex, down ? "on" : "off",
			   pd_reason_val);

	zif = ifp->info;
	if (!zif) {
		if (IS_ZEBRA_DEBUG_KERNEL)
			zlog_debug("%s: if %s(%u) zebra info pointer is NULL",
				   __func__, ifp->name, ifp->ifindex);
		return;
	}

	if (dp_res != ZEBRA_DPLANE_REQUEST_SUCCESS) {
		if (IS_ZEBRA_DEBUG_KERNEL)
			zlog_debug("%s: if %s(%u) dplane update failed",
				   __func__, ifp->name, ifp->ifindex);
		goto done;
	}

	/* Update our info */
	COND_FLAG(zif->flags, ZIF_FLAG_PROTODOWN, down);

done:
	/* Clear our dplane flags */
	UNSET_FLAG(zif->flags, ZIF_FLAG_SET_PROTODOWN);
	UNSET_FLAG(zif->flags, ZIF_FLAG_UNSET_PROTODOWN);
}

/*
 * Handle netconf change from a dplane context object; runs in the main
 * pthread so it can update zebra data structs.
 */
static void zebra_if_netconf_update_ctx(struct zebra_dplane_ctx *ctx,
					struct interface *ifp,
					ifindex_t ifindex)
{
	struct zebra_if *zif = NULL;
	afi_t afi;
	enum dplane_netconf_status_e mpls, mcast_on, linkdown;
	bool *mcast_set, *linkdown_set;

	if (!ifp && ifindex != -1 && ifindex != -2) {
		if (IS_ZEBRA_DEBUG_KERNEL)
			zlog_debug("%s: Can't find ifp(%u)", __func__, ifindex);

		return;
	}

	afi = dplane_ctx_get_afi(ctx);
	mpls = dplane_ctx_get_netconf_mpls(ctx);
	linkdown = dplane_ctx_get_netconf_linkdown(ctx);
	mcast_on = dplane_ctx_get_netconf_mcast(ctx);

	if (ifindex == DPLANE_NETCONF_IFINDEX_ALL) {
		if (afi == AFI_IP) {
			mcast_set = &zrouter.all_mc_forwardingv4;
			linkdown_set = &zrouter.all_linkdownv4;
		} else {
			mcast_set = &zrouter.all_mc_forwardingv6;
			linkdown_set = &zrouter.all_linkdownv6;
		}
	} else if (ifindex == DPLANE_NETCONF_IFINDEX_DEFAULT) {
		if (afi == AFI_IP) {
			mcast_set = &zrouter.default_mc_forwardingv4;
			linkdown_set = &zrouter.default_linkdownv4;
		} else {
			mcast_set = &zrouter.default_mc_forwardingv6;
			linkdown_set = &zrouter.default_linkdownv6;
		}
	} else {
		zif = ifp->info;
		if (!zif) {
			if (IS_ZEBRA_DEBUG_KERNEL)
				zlog_debug(
					"%s: if %s(%u) zebra info pointer is NULL",
					__func__, ifp ? ifp->name : "(null)",
					ifp ? ifp->ifindex : ifindex);
			return;
		}
		if (afi == AFI_IP) {
			mcast_set = &zif->v4mcast_on;
			linkdown_set = &zif->linkdown;
		} else {
			mcast_set = &zif->v6mcast_on;
			linkdown_set = &zif->linkdownv6;
		}

		/*
		 * mpls netconf data is neither v4 or v6 it's AF_MPLS!
		 */
		if (mpls == DPLANE_NETCONF_STATUS_ENABLED) {
			zif->mpls = true;
			zebra_mpls_turned_on();
		} else if (mpls == DPLANE_NETCONF_STATUS_DISABLED)
			zif->mpls = false;
	}

	if (linkdown == DPLANE_NETCONF_STATUS_ENABLED)
		*linkdown_set = true;
	else if (linkdown == DPLANE_NETCONF_STATUS_DISABLED)
		*linkdown_set = false;

	if (mcast_on == DPLANE_NETCONF_STATUS_ENABLED)
		*mcast_set = true;
	else if (mcast_on == DPLANE_NETCONF_STATUS_DISABLED)
		*mcast_set = false;

	if (IS_ZEBRA_DEBUG_KERNEL)
		zlog_debug(
			"%s: afi: %d if %s, ifindex %d, mpls %s mc_forwarding: %s linkdown %s",
			__func__, afi, ifp ? ifp->name : "Global",
			ifp ? ifp->ifindex : ifindex,
			(zif ? (zif->mpls ? "ON" : "OFF") : "OFF"),
			(*mcast_set ? "ON" : "OFF"),
			(*linkdown_set ? "ON" : "OFF"));
}

static void interface_vrf_change(enum dplane_op_e op, ifindex_t ifindex,
				 const char *name, uint32_t tableid,
				 ns_id_t ns_id)
{
	struct vrf *vrf;
	struct zebra_vrf *zvrf = NULL;

	if (op == DPLANE_OP_INTF_DELETE) {
		if (IS_ZEBRA_DEBUG_DPLANE)
			zlog_debug("DPLANE_OP_INTF_DELETE for VRF %s(%u)", name,
				   ifindex);

		vrf = vrf_lookup_by_id((vrf_id_t)ifindex);
		if (!vrf) {
			flog_warn(EC_ZEBRA_VRF_NOT_FOUND,
				  "%s(%u): vrf not found", name, ifindex);
			return;
		}

		vrf_delete(vrf);
	} else {
		if (IS_ZEBRA_DEBUG_DPLANE)
			zlog_debug(
				"DPLANE_OP_INTF_UPDATE for VRF %s(%u) table %u",
				name, ifindex, tableid);

		if (!vrf_lookup_by_id((vrf_id_t)ifindex)) {
			vrf_id_t exist_id;

			exist_id = zebra_vrf_lookup_by_table(tableid, ns_id);
			if (exist_id != VRF_DEFAULT) {
				vrf = vrf_lookup_by_id(exist_id);

				if (vrf)
					flog_err(EC_ZEBRA_VRF_MISCONFIGURED,
						 "VRF %s id %u table id overlaps existing vrf %s(%d), misconfiguration exiting",
						 name, ifindex, vrf->name,
						 vrf->vrf_id);
				else
					flog_err(EC_ZEBRA_VRF_NOT_FOUND,
						 "VRF %s id %u does not exist",
						 name, ifindex);

				exit(-1);
			}
		}

		vrf = vrf_update((vrf_id_t)ifindex, name);
		if (!vrf) {
			flog_err(EC_LIB_INTERFACE, "VRF %s id %u not created",
				 name, ifindex);
			return;
		}

		/*
		 * This is the only place that we get the actual kernel table_id
		 * being used.  We need it to set the table_id of the routes
		 * we are passing to the kernel.... And to throw some totally
		 * awesome parties. that too.
		 *
		 * At this point we *must* have a zvrf because the vrf_create
		 * callback creates one.  We *must* set the table id
		 * before the vrf_enable because of( at the very least )
		 * static routes being delayed for installation until
		 * during the vrf_enable callbacks.
		 */
		zvrf = (struct zebra_vrf *)vrf->info;
		zvrf->table_id = tableid;

		/* Enable the created VRF. */
		if (!vrf_enable(vrf)) {
			flog_err(EC_LIB_INTERFACE,
				 "Failed to enable VRF %s id %u", name,
				 ifindex);
			return;
		}
	}
}

/*
 *  Note: on netlink systems, there should be a 1-to-1 mapping
 * between interface names and ifindex values.
 */
static void set_ifindex(struct interface *ifp, ifindex_t ifi_index,
			struct zebra_ns *zns)
{
	struct interface *oifp;

	oifp = if_lookup_by_index_per_ns(zns, ifi_index);
	if ((oifp != NULL) && (oifp != ifp)) {
		if (ifi_index == IFINDEX_INTERNAL)
			flog_err(
				EC_LIB_INTERFACE,
				"Netlink is setting interface %s ifindex to reserved internal value %u",
				ifp->name, ifi_index);
		else {
			if (IS_ZEBRA_DEBUG_KERNEL)
				zlog_debug(
					"interface index %d was renamed from %s to %s",
					ifi_index, oifp->name, ifp->name);
			if (if_is_up(oifp))
				flog_err(
					EC_LIB_INTERFACE,
					"interface rename detected on up interface: index %d was renamed from %s to %s, results are uncertain!",
					ifi_index, oifp->name, ifp->name);
			if_delete_update(&oifp);
		}
	}
	if_set_index(ifp, ifi_index);
}

static inline void zebra_if_set_ziftype(struct interface *ifp,
					enum zebra_iftype zif_type,
					enum zebra_slave_iftype zif_slave_type)
{
	struct zebra_if *zif;

	zif = (struct zebra_if *)ifp->info;
	zif->zif_slave_type = zif_slave_type;

	if (zif->zif_type != zif_type) {
		zif->zif_type = zif_type;
		/* If the if_type has been set to bond initialize ES info
		 * against it. XXX - note that we don't handle the case where
		 * a zif changes from bond to non-bond; it is really
		 * an unexpected/error condition.
		 */
		zebra_evpn_if_init(zif);
	}
}

static void interface_update_hw_addr(struct zebra_dplane_ctx *ctx,
				     struct interface *ifp)
{
	int i;

	ifp->hw_addr_len = dplane_ctx_get_ifp_hw_addr_len(ctx);
	memcpy(ifp->hw_addr, dplane_ctx_get_ifp_hw_addr(ctx), ifp->hw_addr_len);

	for (i = 0; i < ifp->hw_addr_len; i++)
		if (ifp->hw_addr[i] != 0)
			break;

	if (i == ifp->hw_addr_len)
		ifp->hw_addr_len = 0;
}

static void interface_update_l2info(struct zebra_dplane_ctx *ctx,
				    struct interface *ifp,
				    enum zebra_iftype zif_type, int add,
				    ns_id_t link_nsid)
{
	const struct zebra_l2info_vxlan *vxlan_info;
	const struct zebra_l2info_gre *gre_info;

	switch (zif_type) {
	case ZEBRA_IF_BRIDGE:
		zebra_l2_bridge_add_update(ifp,
					   dplane_ctx_get_ifp_bridge_info(ctx));
		break;
	case ZEBRA_IF_VLAN:
		zebra_l2_vlanif_update(ifp, dplane_ctx_get_ifp_vlan_info(ctx));
		zebra_evpn_acc_bd_svi_set(ifp->info, NULL,
					  !!if_is_operative(ifp));
		break;
	case ZEBRA_IF_VXLAN:
		vxlan_info = dplane_ctx_get_ifp_vxlan_info(ctx);
		zebra_l2_vxlanif_add_update(ifp, vxlan_info, add);
		if (link_nsid != NS_UNKNOWN && vxlan_info->ifindex_link)
			zebra_if_update_link(ifp, vxlan_info->ifindex_link,
					     link_nsid);
		break;
	case ZEBRA_IF_GRE:
		gre_info = dplane_ctx_get_ifp_gre_info(ctx);
		zebra_l2_greif_add_update(ifp, gre_info, add);
		if (link_nsid != NS_UNKNOWN && gre_info->ifindex_link)
			zebra_if_update_link(ifp, gre_info->ifindex_link,
					     link_nsid);
		break;
	case ZEBRA_IF_OTHER:
	case ZEBRA_IF_VRF:
	case ZEBRA_IF_MACVLAN:
	case ZEBRA_IF_VETH:
	case ZEBRA_IF_BOND:
		break;
	}
}

static bool is_if_protodown_reason_only_frr(uint32_t rc_bitfield)
{
	uint8_t frr_protodown_r_bit = if_netlink_get_frr_protodown_r_bit();

	return (rc_bitfield == (((uint32_t)1) << frr_protodown_r_bit));
}

static void interface_if_protodown(struct interface *ifp, bool protodown,
				   uint32_t rc_bitfield)
{
	struct zebra_if *zif = ifp->info;
	bool old_protodown;

	/*
	 * Set our reason code to note it wasn't us.
	 * If the reason we got from the kernel is ONLY frr though, don't
	 * set it.
	 */
	COND_FLAG(zif->protodown_rc, ZEBRA_PROTODOWN_EXTERNAL,
		  protodown && rc_bitfield &&
			  !is_if_protodown_reason_only_frr(rc_bitfield));


	old_protodown = !!ZEBRA_IF_IS_PROTODOWN(zif);
	if (protodown == old_protodown)
		return;

	if (IS_ZEBRA_DEBUG_EVPN_MH_ES || IS_ZEBRA_DEBUG_DPLANE)
		zlog_debug("interface %s dplane change, protodown %s",
			   ifp->name, protodown ? "on" : "off");

	/* Set protodown, respectively */
	COND_FLAG(zif->flags, ZIF_FLAG_PROTODOWN, protodown);

	if (zebra_evpn_is_es_bond_member(ifp)) {
		/* Check it's not already being sent to the dplane first */
		if (protodown &&
		    CHECK_FLAG(zif->flags, ZIF_FLAG_SET_PROTODOWN)) {
			if (IS_ZEBRA_DEBUG_EVPN_MH_ES || IS_ZEBRA_DEBUG_KERNEL)
				zlog_debug(
					"bond mbr %s protodown on recv'd but already sent protodown on to the dplane",
					ifp->name);
			return;
		}

		if (!protodown &&
		    CHECK_FLAG(zif->flags, ZIF_FLAG_UNSET_PROTODOWN)) {
			if (IS_ZEBRA_DEBUG_EVPN_MH_ES || IS_ZEBRA_DEBUG_KERNEL)
				zlog_debug(
					"bond mbr %s protodown off recv'd but already sent protodown off to the dplane",
					ifp->name);
			return;
		}

		if (IS_ZEBRA_DEBUG_EVPN_MH_ES || IS_ZEBRA_DEBUG_KERNEL)
			zlog_debug(
				"bond mbr %s reinstate protodown %s in the dplane",
				ifp->name, old_protodown ? "on" : "off");

		if (old_protodown)
			SET_FLAG(zif->flags, ZIF_FLAG_SET_PROTODOWN);
		else
			SET_FLAG(zif->flags, ZIF_FLAG_UNSET_PROTODOWN);

		dplane_intf_update(zif->ifp);
	}
}

static void if_sweep_protodown(struct zebra_if *zif)
{
	bool protodown;

	protodown = !!ZEBRA_IF_IS_PROTODOWN(zif);

	if (!protodown)
		return;

	if (IS_ZEBRA_DEBUG_KERNEL)
		zlog_debug("interface %s sweeping protodown %s reason 0x%x",
			   zif->ifp->name, protodown ? "on" : "off",
			   zif->protodown_rc);

	/* Only clear our reason codes, leave external if it was set */
	UNSET_FLAG(zif->protodown_rc, ZEBRA_PROTODOWN_ALL);
	dplane_intf_update(zif->ifp);
}

static void
interface_bridge_vxlan_vlan_vni_map_update(struct zebra_dplane_ctx *ctx,
					   struct interface *ifp)
{
	const struct zebra_vxlan_vni_array *vniarray =
		dplane_ctx_get_ifp_vxlan_vni_array(ctx);
	struct zebra_vxlan_vni vni_start, vni_end;
	struct hash *vni_table = NULL;
	struct zebra_vxlan_vni vni, *vnip;
	vni_t vni_id;
	vlanid_t vid;
	int i;

	if (vniarray == NULL)
		return;

	memset(&vni_start, 0, sizeof(vni_start));
	memset(&vni_end, 0, sizeof(vni_end));

	for (i = 0; i < vniarray->count; i++) {
		uint16_t flags = vniarray->vnis[i].flags;

		if (flags & DPLANE_BRIDGE_VLAN_INFO_RANGE_BEGIN) {
			vni_start = vniarray->vnis[i];
			continue;
		}

		if (flags & DPLANE_BRIDGE_VLAN_INFO_RANGE_END)
			vni_end = vniarray->vnis[i];

		if (!(flags & DPLANE_BRIDGE_VLAN_INFO_RANGE_END)) {
			vni_start = vniarray->vnis[i];
			vni_end = vniarray->vnis[i];
		}

		if (IS_ZEBRA_DEBUG_DPLANE)
			zlog_debug(
				"Vlan-Vni(%d:%d-%d:%d) update for VxLAN IF %s(%u)",
				vni_start.access_vlan, vni_end.access_vlan,
				vni_start.vni, vni_end.vni, ifp->name,
				ifp->ifindex);

		if (!vni_table) {
			vni_table = zebra_vxlan_vni_table_create();
			if (!vni_table)
				return;
		}

		for (vid = vni_start.access_vlan, vni_id = vni_start.vni;
		     vid <= vni_end.access_vlan; vid++, vni_id++) {

			memset(&vni, 0, sizeof(vni));
			vni.vni = vni_id;
			vni.access_vlan = vid;
			vnip = hash_get(vni_table, &vni, zebra_vxlan_vni_alloc);
			if (!vnip)
				return;
		}

		memset(&vni_start, 0, sizeof(vni_start));
		memset(&vni_end, 0, sizeof(vni_end));
	}

	if (vni_table)
		zebra_vxlan_if_vni_table_add_update(ifp, vni_table);
}

static void interface_bridge_vxlan_update(struct zebra_dplane_ctx *ctx,
					  struct interface *ifp)
{
	struct zebra_if *zif = ifp->info;
	const struct zebra_dplane_bridge_vlan_info *bvinfo;

	if (dplane_ctx_get_ifp_no_afspec(ctx))
		return;

	if (IS_ZEBRA_VXLAN_IF_SVD(zif))
		interface_bridge_vxlan_vlan_vni_map_update(ctx, ifp);

	if (dplane_ctx_get_ifp_no_bridge_vlan_info(ctx))
		return;

	bvinfo = dplane_ctx_get_ifp_bridge_vlan_info(ctx);

	if (!(bvinfo->flags & DPLANE_BRIDGE_VLAN_INFO_PVID))
		return;

	if (IS_ZEBRA_DEBUG_DPLANE)
		zlog_debug("Access VLAN %u for VxLAN IF %s(%u)", bvinfo->vid,
			   ifp->name, ifp->ifindex);

	zebra_l2_vxlanif_update_access_vlan(ifp, bvinfo->vid);
}

static void interface_bridge_vlan_update(struct zebra_dplane_ctx *ctx,
					 struct interface *ifp)
{
	struct zebra_if *zif = ifp->info;
	const struct zebra_dplane_bridge_vlan_info_array *bvarray;
	struct zebra_dplane_bridge_vlan_info bvinfo;
	bitfield_t old_vlan_bitmap;
	uint16_t vid_range_start = 0;
	int32_t i;

	/* cache the old bitmap addrs */
	old_vlan_bitmap = zif->vlan_bitmap;
	/* create a new bitmap space for re-eval */
	bf_init(zif->vlan_bitmap, IF_VLAN_BITMAP_MAX);

	/* Could we have multiple bridge vlan infos? */
	bvarray = dplane_ctx_get_ifp_bridge_vlan_info_array(ctx);
	if (!bvarray)
		return;

	for (i = 0; i < bvarray->count; i++) {
		bvinfo = bvarray->array[i];

		if (bvinfo.flags & DPLANE_BRIDGE_VLAN_INFO_RANGE_BEGIN) {
			vid_range_start = bvinfo.vid;
			continue;
		}

		if (!(bvinfo.flags & DPLANE_BRIDGE_VLAN_INFO_RANGE_END))
			vid_range_start = bvinfo.vid;

		zebra_vlan_bitmap_compute(ifp, vid_range_start, bvinfo.vid);
	}

	zebra_vlan_mbr_re_eval(ifp, old_vlan_bitmap);
	bf_free(old_vlan_bitmap);
}

static void interface_bridge_handling(struct zebra_dplane_ctx *ctx,
				      struct interface *ifp,
				      enum zebra_iftype zif_type)
{
	struct zebra_if *zif;

	if (!ifp) {
		zlog_warn("Cannot find bridge if %s(%u)",
			  dplane_ctx_get_ifname(ctx),
			  dplane_ctx_get_ifindex(ctx));
		return;
	}

	if (IS_ZEBRA_IF_VXLAN(ifp))
		return interface_bridge_vxlan_update(ctx, ifp);

	/*
	 * build vlan bitmap associated with this interface if that
	 * device type is interested in the vlans
	 */
	zif = ifp->info;
	if (bf_is_inited(zif->vlan_bitmap))
		interface_bridge_vlan_update(ctx, ifp);
}

static void zebra_if_dplane_ifp_handling(struct zebra_dplane_ctx *ctx)
{
	enum dplane_op_e op = dplane_ctx_get_op(ctx);
	const char *name = dplane_ctx_get_ifname(ctx);
	ns_id_t ns_id = dplane_ctx_get_ns_id(ctx);
	ifindex_t ifindex = dplane_ctx_get_ifindex(ctx);
	ifindex_t bond_ifindex = dplane_ctx_get_ifp_bond_ifindex(ctx);
	uint32_t tableid = dplane_ctx_get_ifp_table_id(ctx);
	enum zebra_iftype zif_type = dplane_ctx_get_ifp_zif_type(ctx);
	struct interface *ifp;
	struct zebra_ns *zns;

	zns = zebra_ns_lookup(ns_id);
	if (!zns) {
		zlog_err("Where is our namespace?");
		return;
	}

	if (IS_ZEBRA_DEBUG_DPLANE)
		zlog_debug("%s for %s(%u)", dplane_op2str(op), name, ifindex);

	ifp = if_lookup_by_name_per_ns(zns, name);
	if (op == DPLANE_OP_INTF_DELETE) {
		/* Delete interface notification from kernel */
		if (ifp == NULL) {
			if (IS_ZEBRA_DEBUG_EVENT)
				zlog_debug(
					"Delete LINK received for unknown interface %s(%u)",
					name, ifindex);
			return;
		}

		if (IS_ZEBRA_IF_BOND(ifp))
			zebra_l2if_update_bond(ifp, false);
		if (IS_ZEBRA_IF_BOND_SLAVE(ifp))
			zebra_l2if_update_bond_slave(ifp, bond_ifindex, false);
		/* Special handling for bridge or VxLAN interfaces. */
		if (IS_ZEBRA_IF_BRIDGE(ifp))
			zebra_l2_bridge_del(ifp);
		else if (IS_ZEBRA_IF_VXLAN(ifp))
			zebra_l2_vxlanif_del(ifp);

		if_delete_update(&ifp);

		if (zif_type == ZEBRA_IF_VRF && !vrf_is_backend_netns())
			interface_vrf_change(op, ifindex, name, tableid, ns_id);
	} else {
		ifindex_t master_ifindex, bridge_ifindex, bond_ifindex,
			link_ifindex;
		enum zebra_slave_iftype zif_slave_type;
		uint8_t bypass;
		uint64_t flags;
		vrf_id_t vrf_id;
		uint32_t mtu;
		ns_id_t link_nsid;
		struct zebra_if *zif;
		bool protodown, protodown_set, startup;
		uint32_t rc_bitfield;
		uint8_t old_hw_addr[INTERFACE_HWADDR_MAX];
		char *desc;
		uint8_t family;

		/* If VRF, create or update the VRF structure itself. */
		if (zif_type == ZEBRA_IF_VRF && !vrf_is_backend_netns())
			interface_vrf_change(op, ifindex, name, tableid, ns_id);

		master_ifindex = dplane_ctx_get_ifp_master_ifindex(ctx);
		zif_slave_type = dplane_ctx_get_ifp_zif_slave_type(ctx);
		bridge_ifindex = dplane_ctx_get_ifp_bridge_ifindex(ctx);
		bond_ifindex = dplane_ctx_get_ifp_bond_ifindex(ctx);
		bypass = dplane_ctx_get_ifp_bypass(ctx);
		flags = dplane_ctx_get_ifp_flags(ctx);
		vrf_id = dplane_ctx_get_ifp_vrf_id(ctx);
		mtu = dplane_ctx_get_ifp_mtu(ctx);
		link_ifindex = dplane_ctx_get_ifp_link_ifindex(ctx);
		link_nsid = dplane_ctx_get_ifp_link_nsid(ctx);
		protodown_set = dplane_ctx_get_ifp_protodown_set(ctx);
		protodown = dplane_ctx_get_ifp_protodown(ctx);
		rc_bitfield = dplane_ctx_get_ifp_rc_bitfield(ctx);
		startup = dplane_ctx_get_ifp_startup(ctx);
		desc = dplane_ctx_get_ifp_desc(ctx);
		family = dplane_ctx_get_ifp_family(ctx);

#ifndef AF_BRIDGE
		/*
		 * Work around to make free bsd happy at the moment
		 */
#define AF_BRIDGE 7
#endif
		if (family == AF_BRIDGE)
			return interface_bridge_handling(ctx, ifp, zif_type);

		if (ifp == NULL ||
		    !CHECK_FLAG(ifp->status, ZEBRA_INTERFACE_ACTIVE)) {
			/* Add interface notification from kernel */
			if (IS_ZEBRA_DEBUG_KERNEL)
				zlog_debug(
					"RTM_NEWLINK ADD for %s(%u) vrf_id %u type %d sl_type %d master %u",
					name, ifindex, vrf_id, zif_type,
					zif_slave_type, master_ifindex);

			if (ifp == NULL) {
				/* unknown interface */
				ifp = if_get_by_name(name, vrf_id, NULL);
			} else {
				/* pre-configured interface, learnt now */
				if (ifp->vrf->vrf_id != vrf_id)
					if_update_to_new_vrf(ifp, vrf_id);
			}

			zif = ifp->info;

			/* Update interface information. */
			set_ifindex(ifp, ifindex, zns);
			ifp->flags = flags;
			ifp->mtu6 = ifp->mtu = mtu;
			ifp->metric = 0;
			ifp->speed = kernel_get_speed(ifp, NULL);
			ifp->ptm_status = ZEBRA_PTM_STATUS_UNKNOWN;
			ifp->txqlen = dplane_ctx_get_intf_txqlen(ctx);

			/* Set interface type */
			zebra_if_set_ziftype(ifp, zif_type, zif_slave_type);
			if (IS_ZEBRA_IF_VRF(ifp))
				SET_FLAG(ifp->status,
					 ZEBRA_INTERFACE_VRF_LOOPBACK);

			/* Update link. */
			zebra_if_update_link(ifp, link_ifindex, link_nsid);

			ifp->ll_type = dplane_ctx_get_ifp_zltype(ctx);
			interface_update_hw_addr(ctx, ifp);

			/* Inform clients, install any configured addresses. */
			if_add_update(ifp);

			/*
			 * Extract and save L2 interface information, take
			 * additional actions.
			 */
			interface_update_l2info(ctx, ifp, zif_type, 1,
						link_nsid);
			if (IS_ZEBRA_IF_BOND(ifp))
				zebra_l2if_update_bond(ifp, true);
			if (IS_ZEBRA_IF_BRIDGE_SLAVE(ifp))
				zebra_l2if_update_bridge_slave(
					ifp, bridge_ifindex, ns_id,
					ZEBRA_BRIDGE_NO_ACTION);
			else if (IS_ZEBRA_IF_BOND_SLAVE(ifp))
				zebra_l2if_update_bond_slave(ifp, bond_ifindex,
							     !!bypass);

			if (protodown_set) {
				interface_if_protodown(ifp, protodown,
						       rc_bitfield);
				if (startup)
					if_sweep_protodown(zif);
			}

			if (IS_ZEBRA_IF_BRIDGE(ifp)) {
				if (IS_ZEBRA_DEBUG_KERNEL)
					zlog_debug(
						"RTM_NEWLINK ADD for %s(%u), vlan-aware %d",
						name, ifp->ifindex,
						IS_ZEBRA_IF_BRIDGE_VLAN_AWARE(
							zif));
			}
		} else if (ifp->vrf->vrf_id != vrf_id) {
			/* VRF change for an interface. */
			if (IS_ZEBRA_DEBUG_KERNEL)
				zlog_debug(
					"RTM_NEWLINK vrf-change for %s(%u) vrf_id %u -> %u",
					name, ifp->ifindex, ifp->vrf->vrf_id,
					vrf_id);

			if_handle_vrf_change(ifp, vrf_id);
		} else {
			bool was_bridge_slave, was_bond_slave;
			uint8_t chgflags = ZEBRA_BRIDGE_NO_ACTION;

			zif = ifp->info;

			/* Interface update. */
			if (IS_ZEBRA_DEBUG_KERNEL)
				zlog_debug(
					"RTM_NEWLINK update for %s(%u) sl_type %d master %u",
					name, ifp->ifindex, zif_slave_type,
					master_ifindex);

			set_ifindex(ifp, ifindex, zns);
			ifp->mtu6 = ifp->mtu = mtu;
			ifp->metric = 0;
			ifp->txqlen = dplane_ctx_get_intf_txqlen(ctx);

			/*
			 * Update interface type - NOTE: Only slave_type can
			 * change.
			 */
			was_bridge_slave = IS_ZEBRA_IF_BRIDGE_SLAVE(ifp);
			was_bond_slave = IS_ZEBRA_IF_BOND_SLAVE(ifp);
			zebra_if_set_ziftype(ifp, zif_type, zif_slave_type);

			memcpy(old_hw_addr, ifp->hw_addr, INTERFACE_HWADDR_MAX);

			/* Update link. */
			zebra_if_update_link(ifp, link_ifindex, link_nsid);

			ifp->ll_type = dplane_ctx_get_ifp_zltype(ctx);
			interface_update_hw_addr(ctx, ifp);

			if (protodown_set)
				interface_if_protodown(ifp, protodown,
						       rc_bitfield);

			if (if_is_no_ptm_operative(ifp)) {
				bool is_up = if_is_operative(ifp);

				ifp->flags = flags;
				if (!if_is_no_ptm_operative(ifp) ||
				    CHECK_FLAG(zif->flags,
					       ZIF_FLAG_PROTODOWN)) {
					if (IS_ZEBRA_DEBUG_KERNEL)
						zlog_debug(
							"Intf %s(%u) has gone DOWN",
							name, ifp->ifindex);
					if_down(ifp);
					rib_update(RIB_UPDATE_KERNEL);
				} else if (if_is_operative(ifp)) {
					bool mac_updated = false;

					/*
					 * Must notify client daemons of new
					 * interface status.
					 */
					if (IS_ZEBRA_DEBUG_KERNEL)
						zlog_debug(
							"Intf %s(%u) PTM up, notifying clients",
							name, ifp->ifindex);
					if_up(ifp, !is_up);

					/*
					 * Update EVPN VNI when SVI MAC change
					 */
					if (memcmp(old_hw_addr, ifp->hw_addr,
						   INTERFACE_HWADDR_MAX))
						mac_updated = true;
					if (IS_ZEBRA_IF_VLAN(ifp) &&
					    mac_updated) {
						struct interface *link_if;

						link_if = if_lookup_by_index_per_ns(
							zebra_ns_lookup(
								NS_DEFAULT),
							link_ifindex);
						if (link_if)
							zebra_vxlan_svi_up(
								ifp, link_if);
					} else if (mac_updated &&
						   IS_ZEBRA_IF_BRIDGE(ifp)) {
						zlog_debug(
							"Intf %s(%u) bridge changed MAC address",
							name, ifp->ifindex);
						chgflags =
							ZEBRA_BRIDGE_MASTER_MAC_CHANGE;
					}
				}
			} else {
				ifp->flags = flags;
				if (if_is_operative(ifp) &&
				    !CHECK_FLAG(zif->flags,
						ZIF_FLAG_PROTODOWN)) {
					if (IS_ZEBRA_DEBUG_KERNEL)
						zlog_debug(
							"Intf %s(%u) has come UP",
							name, ifp->ifindex);
					if_up(ifp, true);
					if (IS_ZEBRA_IF_BRIDGE(ifp))
						chgflags =
							ZEBRA_BRIDGE_MASTER_UP;
				} else {
					if (IS_ZEBRA_DEBUG_KERNEL)
						zlog_debug(
							"Intf %s(%u) has gone DOWN",
							name, ifp->ifindex);
					if_down(ifp);
					rib_update(RIB_UPDATE_KERNEL);
				}
			}

			/*
			 * Extract and save L2 interface information, take
			 * additional actions.
			 */
			interface_update_l2info(ctx, ifp, zif_type, 0,
						link_nsid);
			if (IS_ZEBRA_IF_BRIDGE(ifp))
				zebra_l2if_update_bridge(ifp, chgflags);
			if (IS_ZEBRA_IF_BOND(ifp))
				zebra_l2if_update_bond(ifp, true);
			if (IS_ZEBRA_IF_BRIDGE_SLAVE(ifp) || was_bridge_slave)
				zebra_l2if_update_bridge_slave(
					ifp, bridge_ifindex, ns_id, chgflags);
			else if (IS_ZEBRA_IF_BOND_SLAVE(ifp) || was_bond_slave)
				zebra_l2if_update_bond_slave(ifp, bond_ifindex,
							     !!bypass);
			if (IS_ZEBRA_IF_BRIDGE(ifp)) {
				if (IS_ZEBRA_DEBUG_KERNEL)
					zlog_debug(
						"RTM_NEWLINK update for %s(%u), vlan-aware %d",
						name, ifp->ifindex,
						IS_ZEBRA_IF_BRIDGE_VLAN_AWARE(
							zif));
			}
		}

		zif = ifp->info;
		if (zif) {
			XFREE(MTYPE_ZIF_DESC, zif->desc);
			if (desc[0])
				zif->desc = XSTRDUP(MTYPE_ZIF_DESC, desc);
		}
	}
}

void zebra_if_dplane_result(struct zebra_dplane_ctx *ctx)
{
	struct zebra_ns *zns;
	struct interface *ifp;
	ns_id_t ns_id;
	enum dplane_op_e op;
	enum zebra_dplane_result dp_res;
	ifindex_t ifindex;

	ns_id = dplane_ctx_get_ns_id(ctx);
	dp_res = dplane_ctx_get_status(ctx);
	op = dplane_ctx_get_op(ctx);
	ifindex = dplane_ctx_get_ifindex(ctx);

	if (IS_ZEBRA_DEBUG_DPLANE_DETAIL || IS_ZEBRA_DEBUG_KERNEL)
		zlog_debug("Intf dplane ctx %p, op %s, ifindex (%u), result %s",
			   ctx, dplane_op2str(op), ifindex,
			   dplane_res2str(dp_res));

	zns = zebra_ns_lookup(ns_id);
	if (zns == NULL) {
		/* No ns - deleted maybe? */
		if (IS_ZEBRA_DEBUG_KERNEL)
			zlog_debug("%s: can't find zns id %u", __func__, ns_id);

		return;
	}

	ifp = if_lookup_by_index_per_ns(zns, ifindex);

	if (op == DPLANE_OP_INTF_ADDR_ADD || op == DPLANE_OP_INTF_ADDR_DEL) {
		zebra_if_addr_update_ctx(ctx, ifp);
	} else if (op == DPLANE_OP_INTF_INSTALL ||
		   op == DPLANE_OP_INTF_UPDATE || op == DPLANE_OP_INTF_DELETE) {
		/*
		 * Queued from the dplane means it is something
		 * that we need to handle( create/delete the
		 * interface as needed )
		 */
		if (dp_res == ZEBRA_DPLANE_REQUEST_QUEUED)
			zebra_if_dplane_ifp_handling(ctx);
		else
			zebra_if_update_ctx(ctx, ifp);
	} else if (op == DPLANE_OP_INTF_NETCONFIG) {
		zebra_if_netconf_update_ctx(ctx, ifp, ifindex);
	}
}

/* Dump if address information to vty. */
static void connected_dump_vty(struct vty *vty, json_object *json,
			       struct connected *connected)
{
	struct prefix *p;
	json_object *json_addr = NULL;

	/* Print interface address. */
	p = connected->address;

	if (json) {
		json_addr = json_object_new_object();
		json_object_array_add(json, json_addr);
		json_object_string_addf(json_addr, "address", "%pFX", p);
	} else {
		vty_out(vty, "  %s %pFX", prefix_family_str(p), p);
	}

	/* If there is destination address, print it. */
	if (CONNECTED_PEER(connected) && connected->destination) {
		if (json) {
			json_object_string_addf(json_addr, "peer", "%pFX",
						connected->destination);
		} else {
			vty_out(vty, " peer %pFX", connected->destination);
		}
	}

	if (json)
		json_object_boolean_add(
			json_addr, "secondary",
			CHECK_FLAG(connected->flags, ZEBRA_IFA_SECONDARY));
	else if (CHECK_FLAG(connected->flags, ZEBRA_IFA_SECONDARY))
		vty_out(vty, " secondary");

	if (json)
		json_object_boolean_add(json_addr, "noPrefixRoute",
					CHECK_FLAG(connected->flags, ZEBRA_IFA_NOPREFIXROUTE));
	else if (CHECK_FLAG(connected->flags, ZEBRA_IFA_NOPREFIXROUTE))
		vty_out(vty, " noprefixroute");

	if (json)
		json_object_boolean_add(
			json_addr, "unnumbered",
			CHECK_FLAG(connected->flags, ZEBRA_IFA_UNNUMBERED));
	else if (CHECK_FLAG(connected->flags, ZEBRA_IFA_UNNUMBERED))
		vty_out(vty, " unnumbered");

	if (connected->label) {
		if (json)
			json_object_string_add(json_addr, "label",
					       connected->label);
		else
			vty_out(vty, " %s", connected->label);
	}

	if (!json)
		vty_out(vty, "\n");
}

/* Dump interface neighbor address information to vty. */
static void nbr_connected_dump_vty(struct vty *vty, json_object *json,
				   struct nbr_connected *connected)
{
	struct prefix *p;
	char buf[PREFIX2STR_BUFFER];

	/* Print interface address. */
	p = connected->address;
	if (json)
		json_array_string_add(json, prefix2str(p, buf, sizeof(buf)));
	else
		vty_out(vty, "  %s %pFX\n", prefix_family_str(p), p);
}

static const char *
zebra_zifslavetype_2str(enum zebra_slave_iftype zif_slave_type)
{
	switch (zif_slave_type) {
	case ZEBRA_IF_SLAVE_BRIDGE:
		return "Bridge";
	case ZEBRA_IF_SLAVE_VRF:
		return "Vrf";
	case ZEBRA_IF_SLAVE_BOND:
		return "Bond";
	case ZEBRA_IF_SLAVE_OTHER:
		return "Other";
	case ZEBRA_IF_SLAVE_NONE:
		return "None";
	}
	return "None";
}

static const char *zebra_ziftype_2str(enum zebra_iftype zif_type)
{
	switch (zif_type) {
	case ZEBRA_IF_OTHER:
		return "Other";

	case ZEBRA_IF_BRIDGE:
		return "Bridge";

	case ZEBRA_IF_VLAN:
		return "Vlan";

	case ZEBRA_IF_VXLAN:
		return "Vxlan";

	case ZEBRA_IF_VRF:
		return "VRF";

	case ZEBRA_IF_VETH:
		return "VETH";

	case ZEBRA_IF_BOND:
		return "bond";

	case ZEBRA_IF_MACVLAN:
		return "macvlan";

	case ZEBRA_IF_GRE:
		return "GRE";

	default:
		return "Unknown";
	}
}

/* Interface's brief information print out to vty interface. */
static void ifs_dump_brief_vty(struct vty *vty, struct vrf *vrf)
{
	struct connected *connected;
	struct listnode *node;
	struct route_node *rn;
	struct zebra_if *zebra_if;
	struct prefix *p;
	struct interface *ifp;
	bool print_header = true;

	FOR_ALL_INTERFACES (vrf, ifp) {
		bool first_pfx_printed = false;

		if (print_header) {
			vty_out(vty, "%-16s%-8s%-16s%s\n", "Interface",
				"Status", "VRF", "Addresses");
			vty_out(vty, "%-16s%-8s%-16s%s\n", "---------",
				"------", "---", "---------");
			print_header = false; /* We have at least 1 iface */
		}
		zebra_if = ifp->info;

		vty_out(vty, "%-16s", ifp->name);

		if (if_is_up(ifp))
			vty_out(vty, "%-8s", "up");
		else
			vty_out(vty, "%-8s", "down");

		vty_out(vty, "%-16s", vrf->name);

		for (rn = route_top(zebra_if->ipv4_subnets); rn;
		     rn = route_next(rn)) {
			if (!rn->info)
				continue;
			uint32_t list_size = listcount((struct list *)rn->info);

			for (ALL_LIST_ELEMENTS_RO((struct list *)rn->info, node,
						  connected)) {
				if (!CHECK_FLAG(connected->flags,
						ZEBRA_IFA_SECONDARY)) {
					p = connected->address;
					if (first_pfx_printed) {
						/* padding to prepare row only
						 * for ip addr */
						vty_out(vty, "%-40s", "");
						if (list_size > 1)
							vty_out(vty, "+ ");
						vty_out(vty, "%pFX\n", p);
					} else {
						if (list_size > 1)
							vty_out(vty, "+ ");
						vty_out(vty, "%pFX\n", p);
					}
					first_pfx_printed = true;
					break;
				}
			}
		}

		uint32_t v6_list_size = 0;
		frr_each (if_connected, ifp->connected, connected) {
			if (CHECK_FLAG(connected->conf, ZEBRA_IFC_REAL)
				&& (connected->address->family == AF_INET6))
				v6_list_size++;
		}
		frr_each (if_connected, ifp->connected, connected) {
			if (!CHECK_FLAG(connected->flags, ZEBRA_IFA_SECONDARY) &&
			    (connected->address->family == AF_INET6)) {
				p = connected->address;
				if (first_pfx_printed) {
					vty_out(vty, "%-40s", "");
					if (v6_list_size > 1)
						vty_out(vty, "+ ");
					vty_out(vty, "%pFX\n", p);
				} else {
					if (v6_list_size > 1)
						vty_out(vty, "+ ");
					vty_out(vty, "%pFX\n", p);
				}
				first_pfx_printed = true;
				break;
			}
		}
		if (!first_pfx_printed)
			vty_out(vty, "\n");
	}
	vty_out(vty, "\n");
}

static void ifs_dump_brief_vty_json(json_object *json, struct vrf *vrf)
{
	struct connected *connected;
	struct interface *ifp;

	FOR_ALL_INTERFACES (vrf, ifp) {
		json_object *json_if;
		json_object *json_addrs;

		json_if = json_object_new_object();
		json_object_object_add(json, ifp->name, json_if);

		json_object_string_add(json_if, "status",
				       if_is_up(ifp) ? "up" : "down");
		json_object_string_add(json_if, "vrfName", vrf->name);

		json_addrs = json_object_new_array();
		json_object_object_add(json_if, "addresses", json_addrs);
		frr_each (if_connected, ifp->connected, connected) {
			if (!CHECK_FLAG(connected->flags, ZEBRA_IFA_SECONDARY)) {
				char buf[PREFIX2STR_BUFFER];

				json_array_string_add(
					json_addrs,
					prefix2str(connected->address, buf,
						   sizeof(buf)));
			}
		}
	}
}

const char *zebra_protodown_rc_str(uint32_t protodown_rc, char *pd_buf,
				   uint32_t pd_buf_len)
{
	pd_buf[0] = '\0';
	size_t len;

	strlcat(pd_buf, "(", pd_buf_len);

	if (CHECK_FLAG(protodown_rc, ZEBRA_PROTODOWN_EXTERNAL))
		strlcat(pd_buf, "external,", pd_buf_len);

	if (CHECK_FLAG(protodown_rc, ZEBRA_PROTODOWN_EVPN_STARTUP_DELAY))
		strlcat(pd_buf, "startup-delay,", pd_buf_len);

	if (CHECK_FLAG(protodown_rc, ZEBRA_PROTODOWN_EVPN_UPLINK_DOWN))
		strlcat(pd_buf, "uplinks-down,", pd_buf_len);

	if (CHECK_FLAG(protodown_rc, ZEBRA_PROTODOWN_VRRP))
		strlcat(pd_buf, "vrrp,", pd_buf_len);

	if (CHECK_FLAG(protodown_rc, ZEBRA_PROTODOWN_SHARP))
		strlcat(pd_buf, "sharp,", pd_buf_len);

	len = strnlen(pd_buf, pd_buf_len);

	/* Remove trailing comma */
	if (pd_buf[len - 1] == ',')
		pd_buf[len - 1] = '\0';

	strlcat(pd_buf, ")", pd_buf_len);

	return pd_buf;
}

static inline bool if_is_protodown_applicable(struct interface *ifp)
{
	if (IS_ZEBRA_IF_BOND(ifp))
		return false;

	return true;
}

static void zebra_vxlan_if_vni_dump_vty(struct vty *vty,
					struct zebra_vxlan_vni *vni)
{
	char str[INET6_ADDRSTRLEN];

	vty_out(vty, "  VxLAN Id %u", vni->vni);
	if (vni->access_vlan)
		vty_out(vty, " Access VLAN Id %u\n", vni->access_vlan);

	if (vni->mcast_grp.s_addr != INADDR_ANY)
		vty_out(vty, "  Mcast Group %s",
			inet_ntop(AF_INET, &vni->mcast_grp, str, sizeof(str)));
}

static void zebra_vxlan_if_vni_hash_dump_vty(struct hash_bucket *bucket,
					     void *ctxt)
{
	struct vty *vty;
	struct zebra_vxlan_vni *vni;

	vni = (struct zebra_vxlan_vni *)bucket->data;
	vty = (struct vty *)ctxt;

	zebra_vxlan_if_vni_dump_vty(vty, vni);
}

static void zebra_vxlan_if_dump_vty(struct vty *vty, struct zebra_if *zebra_if)
{
	struct zebra_l2info_vxlan *vxlan_info;
	struct zebra_vxlan_vni_info *vni_info;

	vxlan_info = &zebra_if->l2info.vxl;
	vni_info = &vxlan_info->vni_info;

	if (vxlan_info->vtep_ip.s_addr != INADDR_ANY)
		vty_out(vty, " VTEP IP: %pI4", &vxlan_info->vtep_ip);

	if (vxlan_info->ifindex_link && (vxlan_info->link_nsid != NS_UNKNOWN)) {
		struct interface *ifp;

		ifp = if_lookup_by_index_per_ns(
			zebra_ns_lookup(vxlan_info->link_nsid),
			vxlan_info->ifindex_link);
		vty_out(vty, " Link Interface %s",
			ifp == NULL ? "Unknown" : ifp->name);
	}

	if (IS_ZEBRA_VXLAN_IF_VNI(zebra_if)) {
		zebra_vxlan_if_vni_dump_vty(vty, &vni_info->vni);
	} else {
		hash_iterate(vni_info->vni_table,
			     zebra_vxlan_if_vni_hash_dump_vty, vty);
	}

	vty_out(vty, "\n");
}

/* Interface's information print out to vty interface. */
static void if_dump_vty(struct vty *vty, struct interface *ifp)
{
	struct connected *connected;
	struct nbr_connected *nbr_connected;
	struct listnode *node;
	struct route_node *rn;
	struct zebra_if *zebra_if;
	char pd_buf[ZEBRA_PROTODOWN_RC_STR_LEN];

	zebra_if = ifp->info;

	vty_out(vty, "Interface %s is ", ifp->name);
	if (if_is_up(ifp)) {
		vty_out(vty, "up, line protocol ");

		if (CHECK_FLAG(ifp->status, ZEBRA_INTERFACE_LINKDETECTION)) {
			if (if_is_running(ifp))
				vty_out(vty, "is up\n");
			else
				vty_out(vty, "is down\n");
		} else {
			vty_out(vty, "detection is disabled\n");
		}
	} else {
		vty_out(vty, "down\n");
	}

	vty_out(vty, "  Link ups:   %5u    last: %s\n", zebra_if->up_count,
		zebra_if->up_last[0] ? zebra_if->up_last : "(never)");
	vty_out(vty, "  Link downs: %5u    last: %s\n", zebra_if->down_count,
		zebra_if->down_last[0] ? zebra_if->down_last : "(never)");

	zebra_ptm_show_status(vty, NULL, ifp);

	vty_out(vty, "  vrf: %s\n", ifp->vrf->name);

	if (ifp->desc)
		vty_out(vty, "  Description: %s\n", ifp->desc);
	if (zebra_if->desc)
		vty_out(vty, "  OS Description: %s\n", zebra_if->desc);

	if (ifp->ifindex == IFINDEX_INTERNAL) {
		vty_out(vty, "  pseudo interface\n");
		return;
	} else if (!CHECK_FLAG(ifp->status, ZEBRA_INTERFACE_ACTIVE)) {
		vty_out(vty, "  index %d inactive interface\n", ifp->ifindex);
		return;
	}

	vty_out(vty, "  index %d metric %d mtu %d speed %u txqlen %u",
		ifp->ifindex, ifp->metric, ifp->mtu, ifp->speed, ifp->txqlen);
	if (ifp->mtu6 != ifp->mtu)
		vty_out(vty, "mtu6 %d ", ifp->mtu6);
	vty_out(vty, "\n  flags: %s\n", if_flag_dump(ifp->flags));

	if (zebra_if->mpls)
		vty_out(vty, "  MPLS enabled\n");

	if (zebra_if->linkdown)
		vty_out(vty, "  Ignore all v4 routes with linkdown\n");
	if (zebra_if->linkdownv6)
		vty_out(vty, "  Ignore all v6 routes with linkdown\n");

	if (zebra_if->v4mcast_on)
		vty_out(vty, "  v4 Multicast forwarding is on\n");
	if (zebra_if->v6mcast_on)
		vty_out(vty, "  v6 Multicast forwarding is on\n");

	/* Hardware address. */
	vty_out(vty, "  Type: %s\n", if_link_type_str(ifp->ll_type));
	if (ifp->hw_addr_len != 0) {
		int i;

		vty_out(vty, "  HWaddr: ");
		for (i = 0; i < ifp->hw_addr_len; i++)
			vty_out(vty, "%s%02x", i == 0 ? "" : ":",
				ifp->hw_addr[i]);
		vty_out(vty, "\n");
	}

	/* Bandwidth in Mbps */
	if (ifp->bandwidth != 0) {
		vty_out(vty, "  bandwidth %u Mbps", ifp->bandwidth);
		vty_out(vty, "\n");
	}

	for (rn = route_top(zebra_if->ipv4_subnets); rn; rn = route_next(rn)) {
		if (!rn->info)
			continue;

		for (ALL_LIST_ELEMENTS_RO((struct list *)rn->info, node,
					  connected))
			connected_dump_vty(vty, NULL, connected);
	}

	frr_each (if_connected, ifp->connected, connected) {
		if (connected->address->family == AF_INET6)
			connected_dump_vty(vty, NULL, connected);
	}

	vty_out(vty, "  Interface Type %s\n",
		zebra_ziftype_2str(zebra_if->zif_type));
	vty_out(vty, "  Interface Slave Type %s\n",
		zebra_zifslavetype_2str(zebra_if->zif_slave_type));

	if (IS_ZEBRA_IF_BRIDGE(ifp)) {
		vty_out(vty, "  Bridge VLAN-aware: %s\n",
			IS_ZEBRA_IF_BRIDGE_VLAN_AWARE(zebra_if) ? "yes" : "no");
	} else if (IS_ZEBRA_IF_VLAN(ifp)) {
		struct zebra_l2info_vlan *vlan_info;

		vlan_info = &zebra_if->l2info.vl;
		vty_out(vty, "  VLAN Id %u\n", vlan_info->vid);
	} else if (IS_ZEBRA_IF_VXLAN(ifp)) {
		zebra_vxlan_if_dump_vty(vty, zebra_if);
	} else if (IS_ZEBRA_IF_GRE(ifp)) {
		struct zebra_l2info_gre *gre_info;

		gre_info = &zebra_if->l2info.gre;
		if (gre_info->vtep_ip.s_addr != INADDR_ANY) {
			vty_out(vty, "  VTEP IP: %pI4", &gre_info->vtep_ip);
			if (gre_info->vtep_ip_remote.s_addr != INADDR_ANY)
				vty_out(vty, " , remote %pI4",
					&gre_info->vtep_ip_remote);
			vty_out(vty, "\n");
		}
		if (gre_info->ifindex_link &&
		    (gre_info->link_nsid != NS_UNKNOWN)) {
			struct interface *ifp;

			ifp = if_lookup_by_index_per_ns(
					zebra_ns_lookup(gre_info->link_nsid),
					gre_info->ifindex_link);
			vty_out(vty, "  Link Interface %s\n",
				ifp == NULL ? "Unknown" :
				ifp->name);
		}
	}

	if (IS_ZEBRA_IF_BRIDGE_SLAVE(ifp)) {
		struct zebra_l2info_brslave *br_slave;

		br_slave = &zebra_if->brslave_info;
		if (br_slave->bridge_ifindex != IFINDEX_INTERNAL) {
			if (br_slave->br_if)
				vty_out(vty, "  Master interface: %s\n",
					br_slave->br_if->name);
			else
				vty_out(vty, "  Master ifindex: %u\n",
					br_slave->bridge_ifindex);
		}
	}

	if (IS_ZEBRA_IF_BOND_SLAVE(ifp)) {
		struct zebra_l2info_bondslave *bond_slave;

		bond_slave = &zebra_if->bondslave_info;
		if (bond_slave->bond_ifindex != IFINDEX_INTERNAL) {
			if (bond_slave->bond_if)
				vty_out(vty, "  Master interface: %s\n",
					bond_slave->bond_if->name);
			else
				vty_out(vty, "  Master ifindex: %u\n",
					bond_slave->bond_ifindex);
		}
	}

	if (zebra_if->flags & ZIF_FLAG_LACP_BYPASS)
		vty_out(vty, "  LACP bypass: on\n");

	zebra_evpn_if_es_print(vty, NULL, zebra_if);
	vty_out(vty, "  protodown: %s %s\n",
		(ZEBRA_IF_IS_PROTODOWN(zebra_if)) ? "on" : "off",
		if_is_protodown_applicable(ifp) ? "" : "(n/a)");
	if (zebra_if->protodown_rc)
		vty_out(vty, "  protodown reasons: %s\n",
			zebra_protodown_rc_str(zebra_if->protodown_rc, pd_buf,
					       sizeof(pd_buf)));

	if (zebra_if->link_ifindex != IFINDEX_INTERNAL) {
		if (zebra_if->link)
			vty_out(vty, "  Parent interface: %s\n", zebra_if->link->name);
		else
			vty_out(vty, "  Parent ifindex: %d\n", zebra_if->link_ifindex);
	}

	if (HAS_LINK_PARAMS(ifp)) {
		int i;
		struct if_link_params *iflp = ifp->link_params;
		vty_out(vty, "  Traffic Engineering Link Parameters:\n");
		if (IS_PARAM_SET(iflp, LP_TE_METRIC))
			vty_out(vty, "    TE metric %u\n", iflp->te_metric);
		if (IS_PARAM_SET(iflp, LP_MAX_BW))
			vty_out(vty, "    Maximum Bandwidth %g (Byte/s)\n",
				iflp->max_bw);
		if (IS_PARAM_SET(iflp, LP_MAX_RSV_BW))
			vty_out(vty,
				"    Maximum Reservable Bandwidth %g (Byte/s)\n",
				iflp->max_rsv_bw);
		if (IS_PARAM_SET(iflp, LP_UNRSV_BW)) {
			vty_out(vty,
				"    Unreserved Bandwidth per Class Type in Byte/s:\n");
			for (i = 0; i < MAX_CLASS_TYPE; i += 2)
				vty_out(vty,
					"      [%d]: %g (Bytes/sec),\t[%d]: %g (Bytes/sec)\n",
					i, iflp->unrsv_bw[i], i + 1,
					iflp->unrsv_bw[i + 1]);
		}

		if (IS_PARAM_SET(iflp, LP_ADM_GRP))
			vty_out(vty, "    Administrative Group:%u\n",
				iflp->admin_grp);
		if (IS_PARAM_SET(iflp, LP_DELAY)) {
			vty_out(vty, "    Link Delay Average: %u (micro-sec.)",
				iflp->av_delay);
			if (IS_PARAM_SET(iflp, LP_MM_DELAY)) {
				vty_out(vty, " Min:  %u (micro-sec.)",
					iflp->min_delay);
				vty_out(vty, " Max:  %u (micro-sec.)",
					iflp->max_delay);
			}
			vty_out(vty, "\n");
		}
		if (IS_PARAM_SET(iflp, LP_DELAY_VAR))
			vty_out(vty,
				"    Link Delay Variation %u (micro-sec.)\n",
				iflp->delay_var);
		if (IS_PARAM_SET(iflp, LP_PKT_LOSS))
			vty_out(vty, "    Link Packet Loss %f (in %%)\n",
				(double)iflp->pkt_loss * LOSS_PRECISION);
		if (IS_PARAM_SET(iflp, LP_AVA_BW))
			vty_out(vty, "    Available Bandwidth %g (Byte/s)\n",
				iflp->ava_bw);
		if (IS_PARAM_SET(iflp, LP_RES_BW))
			vty_out(vty, "    Residual Bandwidth %g (Byte/s)\n",
				iflp->res_bw);
		if (IS_PARAM_SET(iflp, LP_USE_BW))
			vty_out(vty, "    Utilized Bandwidth %g (Byte/s)\n",
				iflp->use_bw);
		if (IS_PARAM_SET(iflp, LP_RMT_AS))
			vty_out(vty, "    Neighbor ASBR IP: %pI4 AS: %u \n",
				&iflp->rmt_ip, iflp->rmt_as);
	}

	hook_call(zebra_if_extra_info, vty, ifp);

	if (listhead(ifp->nbr_connected))
		vty_out(vty, "  Neighbor address(s):\n");
	for (ALL_LIST_ELEMENTS_RO(ifp->nbr_connected, node, nbr_connected))
		nbr_connected_dump_vty(vty, NULL, nbr_connected);

#ifdef HAVE_PROC_NET_DEV
	/* Statistics print out using proc file system. */
	vty_out(vty,
		"    %lu input packets (%lu multicast), %lu bytes, %lu dropped\n",
		ifp->stats.rx_packets, ifp->stats.rx_multicast,
		ifp->stats.rx_bytes, ifp->stats.rx_dropped);

	vty_out(vty,
		"    %lu input errors, %lu length, %lu overrun, %lu CRC, %lu frame\n",
		ifp->stats.rx_errors, ifp->stats.rx_length_errors,
		ifp->stats.rx_over_errors, ifp->stats.rx_crc_errors,
		ifp->stats.rx_frame_errors);

	vty_out(vty, "    %lu fifo, %lu missed\n", ifp->stats.rx_fifo_errors,
		ifp->stats.rx_missed_errors);

	vty_out(vty, "    %lu output packets, %lu bytes, %lu dropped\n",
		ifp->stats.tx_packets, ifp->stats.tx_bytes,
		ifp->stats.tx_dropped);

	vty_out(vty,
		"    %lu output errors, %lu aborted, %lu carrier, %lu fifo, %lu heartbeat\n",
		ifp->stats.tx_errors, ifp->stats.tx_aborted_errors,
		ifp->stats.tx_carrier_errors, ifp->stats.tx_fifo_errors,
		ifp->stats.tx_heartbeat_errors);

	vty_out(vty, "    %lu window, %lu collisions\n",
		ifp->stats.tx_window_errors, ifp->stats.collisions);
#endif /* HAVE_PROC_NET_DEV */

#ifdef HAVE_NET_RT_IFLIST
	/* Statistics print out using sysctl (). */
	vty_out(vty,
		"    input packets %llu, bytes %llu, dropped %llu, multicast packets %llu\n",
		(unsigned long long)ifp->stats.ifi_ipackets,
		(unsigned long long)ifp->stats.ifi_ibytes,
		(unsigned long long)ifp->stats.ifi_iqdrops,
		(unsigned long long)ifp->stats.ifi_imcasts);

	vty_out(vty, "    input errors %llu\n",
		(unsigned long long)ifp->stats.ifi_ierrors);

	vty_out(vty,
		"    output packets %llu, bytes %llu, multicast packets %llu\n",
		(unsigned long long)ifp->stats.ifi_opackets,
		(unsigned long long)ifp->stats.ifi_obytes,
		(unsigned long long)ifp->stats.ifi_omcasts);

	vty_out(vty, "    output errors %llu\n",
		(unsigned long long)ifp->stats.ifi_oerrors);

	vty_out(vty, "    collisions %llu\n",
		(unsigned long long)ifp->stats.ifi_collisions);
#endif /* HAVE_NET_RT_IFLIST */
}

static void zebra_vxlan_if_vni_dump_vty_json(json_object *json_if,
					     struct zebra_vxlan_vni *vni)
{
	json_object_int_add(json_if, "vxlanId", vni->vni);
	if (vni->access_vlan)
		json_object_int_add(json_if, "accessVlanId", vni->access_vlan);
	if (vni->mcast_grp.s_addr != INADDR_ANY)
		json_object_string_addf(json_if, "mcastGroup", "%pI4",
					&vni->mcast_grp);
}

static void zebra_vxlan_if_vni_hash_dump_vty_json(struct hash_bucket *bucket,
						  void *ctxt)
{
	json_object *json_if;
	struct zebra_vxlan_vni *vni;

	vni = (struct zebra_vxlan_vni *)bucket->data;
	json_if = (json_object *)ctxt;

	zebra_vxlan_if_vni_dump_vty_json(json_if, vni);
}

static void zebra_vxlan_if_dump_vty_json(json_object *json_if,
					 struct zebra_if *zebra_if)
{
	struct zebra_l2info_vxlan *vxlan_info;
	struct zebra_vxlan_vni_info *vni_info;

	vxlan_info = &zebra_if->l2info.vxl;
	vni_info = &vxlan_info->vni_info;

	if (vxlan_info->vtep_ip.s_addr != INADDR_ANY)
		json_object_string_addf(json_if, "vtepIp", "%pI4",
					&vxlan_info->vtep_ip);

	if (vxlan_info->ifindex_link && (vxlan_info->link_nsid != NS_UNKNOWN)) {
		struct interface *ifp;

		ifp = if_lookup_by_index_per_ns(
			zebra_ns_lookup(vxlan_info->link_nsid),
			vxlan_info->ifindex_link);
		json_object_string_add(json_if, "linkInterface",
				       ifp == NULL ? "Unknown" : ifp->name);
	}
	if (IS_ZEBRA_VXLAN_IF_VNI(zebra_if)) {
		zebra_vxlan_if_vni_dump_vty_json(json_if, &vni_info->vni);
	} else {
		hash_iterate(vni_info->vni_table,
			     zebra_vxlan_if_vni_hash_dump_vty_json, json_if);
	}
}

static void if_dump_vty_json(struct vty *vty, struct interface *ifp,
			     json_object *json)
{
	struct connected *connected;
	struct nbr_connected *nbr_connected;
	struct listnode *node;
	struct route_node *rn;
	struct zebra_if *zebra_if;
	char pd_buf[ZEBRA_PROTODOWN_RC_STR_LEN];
	char buf[BUFSIZ];
	json_object *json_if;
	json_object *json_addrs;

	json_if = json_object_new_object();
	json_object_object_add(json, ifp->name, json_if);

	if (if_is_up(ifp)) {
		json_object_string_add(json_if, "administrativeStatus", "up");

		if (CHECK_FLAG(ifp->status, ZEBRA_INTERFACE_LINKDETECTION)) {
			json_object_string_add(json_if, "operationalStatus",
					       if_is_running(ifp) ? "up"
								  : "down");
			json_object_boolean_add(json_if, "linkDetection", true);
		} else {
			json_object_boolean_add(json_if, "linkDetection",
						false);
		}
	} else {
		json_object_string_add(json_if, "administrativeStatus", "down");
	}

	zebra_if = ifp->info;

	json_object_int_add(json_if, "linkUps", zebra_if->up_count);
	json_object_int_add(json_if, "linkDowns", zebra_if->down_count);
	if (zebra_if->up_last[0])
		json_object_string_add(json_if, "lastLinkUp",
				       zebra_if->up_last);
	if (zebra_if->down_last[0])
		json_object_string_add(json_if, "lastLinkDown",
				       zebra_if->down_last);

	zebra_ptm_show_status(vty, json_if, ifp);

	json_object_string_add(json_if, "vrfName", ifp->vrf->name);

	if (ifp->desc)
		json_object_string_add(json_if, "description", ifp->desc);
	if (zebra_if->desc)
		json_object_string_add(json_if, "OsDescription",
				       zebra_if->desc);

	json_object_boolean_add(json_if, "mplsEnabled", zebra_if->mpls);
	json_object_boolean_add(json_if, "linkDown", zebra_if->linkdown);
	json_object_boolean_add(json_if, "linkDownV6", zebra_if->linkdownv6);
	json_object_boolean_add(json_if, "mcForwardingV4",
				zebra_if->v4mcast_on);
	json_object_boolean_add(json_if, "mcForwardingV6",
				zebra_if->v6mcast_on);

	if (ifp->ifindex == IFINDEX_INTERNAL) {
		json_object_boolean_add(json_if, "pseudoInterface", true);
		return;
	} else if (!CHECK_FLAG(ifp->status, ZEBRA_INTERFACE_ACTIVE)) {
		json_object_int_add(json_if, "index", ifp->ifindex);
		return;
	}

	json_object_boolean_add(json_if, "pseudoInterface", false);
	json_object_int_add(json_if, "index", ifp->ifindex);
	json_object_int_add(json_if, "metric", ifp->metric);
	json_object_int_add(json_if, "mtu", ifp->mtu);
	if (ifp->mtu6 != ifp->mtu)
		json_object_int_add(json_if, "mtu6", ifp->mtu6);
	json_object_int_add(json_if, "speed", ifp->speed);
	json_object_int_add(json_if, "txqlen", ifp->txqlen);
	json_object_string_add(json_if, "flags", if_flag_dump(ifp->flags));

	/* Hardware address. */
	json_object_string_add(json_if, "type", if_link_type_str(ifp->ll_type));
	if (ifp->hw_addr_len != 0) {
		char hwbuf[BUFSIZ];

		hwbuf[0] = '\0';
		for (int i = 0; i < ifp->hw_addr_len; i++) {
			snprintf(buf, sizeof(buf), "%s%02x", i == 0 ? "" : ":",
				 ifp->hw_addr[i]);
			strlcat(hwbuf, buf, sizeof(hwbuf));
		}
		json_object_string_add(json_if, "hardwareAddress", hwbuf);
	}

	/* Bandwidth in Mbps */
	if (ifp->bandwidth != 0)
		json_object_int_add(json_if, "bandwidth", ifp->bandwidth);


	/* IP addresses. */
	json_addrs = json_object_new_array();
	json_object_object_add(json_if, "ipAddresses", json_addrs);

	for (rn = route_top(zebra_if->ipv4_subnets); rn; rn = route_next(rn)) {
		if (!rn->info)
			continue;

		for (ALL_LIST_ELEMENTS_RO((struct list *)rn->info, node,
					  connected))
			connected_dump_vty(vty, json_addrs, connected);
	}

	frr_each (if_connected, ifp->connected, connected) {
		if (connected->address->family == AF_INET6)
			connected_dump_vty(vty, json_addrs, connected);
	}

	json_object_string_add(json_if, "interfaceType",
			       zebra_ziftype_2str(zebra_if->zif_type));
	json_object_string_add(
		json_if, "interfaceSlaveType",
		zebra_zifslavetype_2str(zebra_if->zif_slave_type));

	if (IS_ZEBRA_IF_BRIDGE(ifp)) {
		struct zebra_l2info_bridge *bridge_info;

		bridge_info = &zebra_if->l2info.br;
		json_object_boolean_add(json_if, "bridgeVlanAware",
					bridge_info->bridge.vlan_aware);
	} else if (IS_ZEBRA_IF_VLAN(ifp)) {
		struct zebra_l2info_vlan *vlan_info;

		vlan_info = &zebra_if->l2info.vl;
		json_object_int_add(json_if, "vlanId", vlan_info->vid);
	} else if (IS_ZEBRA_IF_VXLAN(ifp)) {
		zebra_vxlan_if_dump_vty_json(json_if, zebra_if);

	} else if (IS_ZEBRA_IF_GRE(ifp)) {
		struct zebra_l2info_gre *gre_info;

		gre_info = &zebra_if->l2info.gre;
		if (gre_info->vtep_ip.s_addr != INADDR_ANY) {
			json_object_string_addf(json_if, "vtepIp", "%pI4",
						&gre_info->vtep_ip);
			if (gre_info->vtep_ip_remote.s_addr != INADDR_ANY)
				json_object_string_addf(
					json_if, "vtepRemoteIp", "%pI4",
					&gre_info->vtep_ip_remote);
		}
		if (gre_info->ifindex_link
		    && (gre_info->link_nsid != NS_UNKNOWN)) {
			struct interface *ifp;

			ifp = if_lookup_by_index_per_ns(
				zebra_ns_lookup(gre_info->link_nsid),
				gre_info->ifindex_link);
			json_object_string_add(json_if, "linkInterface",
					       ifp == NULL ? "Unknown"
							   : ifp->name);
		}
	}

	if (IS_ZEBRA_IF_BRIDGE_SLAVE(ifp)) {
		struct zebra_l2info_brslave *br_slave;

		br_slave = &zebra_if->brslave_info;
		if (br_slave->bridge_ifindex != IFINDEX_INTERNAL) {
			if (br_slave->br_if)
				json_object_string_add(json_if,
						       "masterInterface",
						       br_slave->br_if->name);
			else
				json_object_int_add(json_if, "masterIfindex",
						    br_slave->bridge_ifindex);
		}
	}

	if (IS_ZEBRA_IF_BOND_SLAVE(ifp)) {
		struct zebra_l2info_bondslave *bond_slave;

		bond_slave = &zebra_if->bondslave_info;
		if (bond_slave->bond_ifindex != IFINDEX_INTERNAL) {
			if (bond_slave->bond_if)
				json_object_string_add(
					json_if, "masterInterface",
					bond_slave->bond_if->name);
			else
				json_object_int_add(json_if, "masterIfindex",
						    bond_slave->bond_ifindex);
		}
	}

	json_object_boolean_add(
		json_if, "lacpBypass",
		CHECK_FLAG(zebra_if->flags, ZIF_FLAG_LACP_BYPASS));

	zebra_evpn_if_es_print(vty, json_if, zebra_if);

	if (if_is_protodown_applicable(ifp)) {
		json_object_string_add(
			json_if, "protodown",
			(ZEBRA_IF_IS_PROTODOWN(zebra_if)) ? "on" : "off");
		if (zebra_if->protodown_rc)
			json_object_string_add(
				json_if, "protodownReason",
				zebra_protodown_rc_str(zebra_if->protodown_rc,
						       pd_buf, sizeof(pd_buf)));
	}

	if (zebra_if->link_ifindex != IFINDEX_INTERNAL) {
		if (zebra_if->link)
			json_object_string_add(json_if, "parentInterface",
					       zebra_if->link->name);
		else
			json_object_int_add(json_if, "parentIfindex",
					    zebra_if->link_ifindex);
	}

	if (HAS_LINK_PARAMS(ifp)) {
		struct if_link_params *iflp = ifp->link_params;
		json_object *json_te;

		json_te = json_object_new_object();
		json_object_object_add(
			json_if, "trafficEngineeringLinkParameters", json_te);

		if (IS_PARAM_SET(iflp, LP_TE_METRIC))
			json_object_int_add(json_te, "teMetric",
					    iflp->te_metric);
		if (IS_PARAM_SET(iflp, LP_MAX_BW))
			json_object_double_add(json_te, "maximumBandwidth",
					       iflp->max_bw);
		if (IS_PARAM_SET(iflp, LP_MAX_RSV_BW))
			json_object_double_add(json_te,
					       "maximumReservableBandwidth",
					       iflp->max_rsv_bw);
		if (IS_PARAM_SET(iflp, LP_UNRSV_BW)) {
			json_object *json_bws;

			json_bws = json_object_new_object();
			json_object_object_add(json_te, "unreservedBandwidth",
					       json_bws);
			for (unsigned int i = 0; i < MAX_CLASS_TYPE; ++i) {
				char buf_ct[64];

				snprintf(buf_ct, sizeof(buf_ct), "classType%u",
					 i);
				json_object_double_add(json_bws, buf_ct,
						       iflp->unrsv_bw[i]);
			}
		}

		if (IS_PARAM_SET(iflp, LP_ADM_GRP))
			json_object_int_add(json_te, "administrativeGroup",
					    iflp->admin_grp);
		if (IS_PARAM_SET(iflp, LP_DELAY)) {
			json_object_int_add(json_te, "linkDelayAverage",
					    iflp->av_delay);
			if (IS_PARAM_SET(iflp, LP_MM_DELAY)) {
				json_object_int_add(json_te, "linkDelayMinimum",
						    iflp->min_delay);
				json_object_int_add(json_te, "linkDelayMaximum",
						    iflp->max_delay);
			}
		}
		if (IS_PARAM_SET(iflp, LP_DELAY_VAR))
			json_object_int_add(json_te, "linkDelayVariation",
					    iflp->delay_var);
		if (IS_PARAM_SET(iflp, LP_PKT_LOSS))
			json_object_double_add(json_te, "linkPacketLoss",
					       (double)iflp->pkt_loss *
						       LOSS_PRECISION);
		if (IS_PARAM_SET(iflp, LP_AVA_BW))
			json_object_double_add(json_te, "availableBandwidth",
					       iflp->ava_bw);
		if (IS_PARAM_SET(iflp, LP_RES_BW))
			json_object_double_add(json_te, "residualBandwidth",
					       iflp->res_bw);
		if (IS_PARAM_SET(iflp, LP_USE_BW))
			json_object_double_add(json_te, "utilizedBandwidth",
					       iflp->use_bw);
		if (IS_PARAM_SET(iflp, LP_RMT_AS))
			json_object_string_addf(json_te, "neighborAsbrIp",
						"%pI4", &iflp->rmt_ip);
		json_object_int_add(json_te, "neighborAsbrAs", iflp->rmt_as);
	}

	if (listhead(ifp->nbr_connected)) {
		json_object *json_nbr_addrs;

		json_nbr_addrs = json_object_new_array();
		json_object_object_add(json_if, "neighborIpAddresses",
				       json_nbr_addrs);

		for (ALL_LIST_ELEMENTS_RO(ifp->nbr_connected, node,
					  nbr_connected))
			nbr_connected_dump_vty(vty, json_nbr_addrs,
					       nbr_connected);
	}

#ifdef HAVE_PROC_NET_DEV
	json_object_int_add(json_if, "inputPackets", stats.rx_packets);
	json_object_int_add(json_if, "inputBytes", ifp->stats.rx_bytes);
	json_object_int_add(json_if, "inputDropped", ifp->stats.rx_dropped);
	json_object_int_add(json_if, "inputMulticastPackets",
			    ifp->stats.rx_multicast);
	json_object_int_add(json_if, "inputErrors", ifp->stats.rx_errors);
	json_object_int_add(json_if, "inputLengthErrors",
			    ifp->stats.rx_length_errors);
	json_object_int_add(json_if, "inputOverrunErrors",
			    ifp->stats.rx_over_errors);
	json_object_int_add(json_if, "inputCrcErrors",
			    ifp->stats.rx_crc_errors);
	json_object_int_add(json_if, "inputFrameErrors",
			    ifp->stats.rx_frame_errors);
	json_object_int_add(json_if, "inputFifoErrors",
			    ifp->stats.rx_fifo_errors);
	json_object_int_add(json_if, "inputMissedErrors",
			    ifp->stats.rx_missed_errors);
	json_object_int_add(json_if, "outputPackets", ifp->stats.tx_packets);
	json_object_int_add(json_if, "outputBytes", ifp->stats.tx_bytes);
	json_object_int_add(json_if, "outputDroppedPackets",
			    ifp->stats.tx_dropped);
	json_object_int_add(json_if, "outputErrors", ifp->stats.tx_errors);
	json_object_int_add(json_if, "outputAbortedErrors",
			    ifp->stats.tx_aborted_errors);
	json_object_int_add(json_if, "outputCarrierErrors",
			    ifp->stats.tx_carrier_errors);
	json_object_int_add(json_if, "outputFifoErrors",
			    ifp->stats.tx_fifo_errors);
	json_object_int_add(json_if, "outputHeartbeatErrors",
			    ifp->stats.tx_heartbeat_errors);
	json_object_int_add(json_if, "outputWindowErrors",
			    ifp->stats.tx_window_errors);
	json_object_int_add(json_if, "collisions", ifp->stats.collisions);
#endif /* HAVE_PROC_NET_DEV */

#ifdef HAVE_NET_RT_IFLIST
	json_object_int_add(json_if, "inputPackets", ifp->stats.ifi_ipackets);
	json_object_int_add(json_if, "inputBytes", ifp->stats.ifi_ibytes);
	json_object_int_add(json_if, "inputDropd", ifp->stats.ifi_iqdrops);
	json_object_int_add(json_if, "inputMulticastPackets",
			    ifp->stats.ifi_imcasts);
	json_object_int_add(json_if, "inputErrors", ifp->stats.ifi_ierrors);
	json_object_int_add(json_if, "outputPackets", ifp->stats.ifi_opackets);
	json_object_int_add(json_if, "outputBytes", ifp->stats.ifi_obytes);
	json_object_int_add(json_if, "outputMulticastPackets",
			    ifp->stats.ifi_omcasts);
	json_object_int_add(json_if, "outputErrors", ifp->stats.ifi_oerrors);
	json_object_int_add(json_if, "collisions", ifp->stats.ifi_collisions);
#endif /* HAVE_NET_RT_IFLIST */
}

static void interface_update_stats(void)
{
#ifdef HAVE_PROC_NET_DEV
	/* If system has interface statistics via proc file system, update
	   statistics. */
	ifstat_update_proc();
#endif /* HAVE_PROC_NET_DEV */
#ifdef HAVE_NET_RT_IFLIST
	ifstat_update_sysctl();
#endif /* HAVE_NET_RT_IFLIST */
}

#include "zebra/interface_clippy.c"
/* Show all interfaces to vty. */
DEFPY(show_interface, show_interface_cmd,
      "show interface vrf NAME$vrf_name [brief$brief] [json$uj]",
      SHOW_STR
      "Interface status and configuration\n"
      VRF_CMD_HELP_STR
      "Interface status and configuration summary\n"
      JSON_STR)
{
	struct vrf *vrf;
	struct interface *ifp;
	json_object *json = NULL;

	interface_update_stats();

	vrf = vrf_lookup_by_name(vrf_name);
	if (!vrf) {
		if (uj)
			vty_out(vty, "{}\n");
		else
			vty_out(vty, "%% VRF %s not found\n", vrf_name);
		return CMD_WARNING;
	}

	if (uj)
		json = json_object_new_object();

	if (brief) {
		if (json)
			ifs_dump_brief_vty_json(json, vrf);
		else
			ifs_dump_brief_vty(vty, vrf);
	} else {
		FOR_ALL_INTERFACES (vrf, ifp) {
			if (json)
				if_dump_vty_json(vty, ifp, json);
			else
				if_dump_vty(vty, ifp);
		}
	}

	if (json)
		vty_json(vty, json);

	return CMD_SUCCESS;
}


/* Show all interfaces to vty. */
DEFPY (show_interface_vrf_all,
       show_interface_vrf_all_cmd,
       "show interface [vrf all] [brief$brief] [json$uj]",
       SHOW_STR
       "Interface status and configuration\n"
       VRF_ALL_CMD_HELP_STR
       "Interface status and configuration summary\n"
       JSON_STR)
{
	struct vrf *vrf;
	struct interface *ifp;
	json_object *json = NULL;

	interface_update_stats();

	if (uj)
		json = json_object_new_object();

	/* All interface print. */
	RB_FOREACH (vrf, vrf_name_head, &vrfs_by_name) {
		if (brief) {
			if (json)
				ifs_dump_brief_vty_json(json, vrf);
			else
				ifs_dump_brief_vty(vty, vrf);
		} else {
			FOR_ALL_INTERFACES (vrf, ifp) {
				if (json)
					if_dump_vty_json(vty, ifp, json);
				else
					if_dump_vty(vty, ifp);
			}
		}
	}

	if (json)
		vty_json(vty, json);

	return CMD_SUCCESS;
}

/* Show specified interface to vty. */

DEFPY (show_interface_name_vrf,
       show_interface_name_vrf_cmd,
       "show interface IFNAME$ifname vrf NAME$vrf_name [json$uj]",
       SHOW_STR
       "Interface status and configuration\n"
       "Interface name\n"
       VRF_CMD_HELP_STR
       JSON_STR)
{
	struct interface *ifp;
	struct vrf *vrf;
	json_object *json = NULL;

	interface_update_stats();

	vrf = vrf_lookup_by_name(vrf_name);
	if (!vrf) {
		if (uj)
			vty_out(vty, "{}\n");
		else
			vty_out(vty, "%% VRF %s not found\n", vrf_name);
		return CMD_WARNING;
	}

	ifp = if_lookup_by_name_vrf(ifname, vrf);
	if (ifp == NULL) {
		if (uj)
			vty_out(vty, "{}\n");
		else
			vty_out(vty, "%% Can't find interface %s\n", ifname);
		return CMD_WARNING;
	}

	if (uj)
		json = json_object_new_object();

	if (json)
		if_dump_vty_json(vty, ifp, json);
	else
		if_dump_vty(vty, ifp);

	if (json)
		vty_json(vty, json);

	return CMD_SUCCESS;
}

/* Show specified interface to vty. */
DEFPY (show_interface_name_vrf_all,
       show_interface_name_vrf_all_cmd,
       "show interface IFNAME$ifname [vrf all] [json$uj]",
       SHOW_STR
       "Interface status and configuration\n"
       "Interface name\n"
       VRF_ALL_CMD_HELP_STR
       JSON_STR)
{
	struct interface *ifp = NULL;
	struct interface *ifptmp;
	struct vrf *vrf;
	json_object *json = NULL;
	int count = 0;

	interface_update_stats();

	RB_FOREACH (vrf, vrf_name_head, &vrfs_by_name) {
		ifptmp = if_lookup_by_name_vrf(ifname, vrf);
		if (ifptmp) {
			ifp = ifptmp;
			count++;
			if (!vrf_is_backend_netns())
				break;
		}
	}

	if (ifp == NULL) {
		if (uj)
			vty_out(vty, "{}\n");
		else
			vty_out(vty, "%% Can't find interface %s\n", ifname);
		return CMD_WARNING;
	}
	if (count > 1) {
		if (uj) {
			vty_out(vty, "{}\n");
		} else {
			vty_out(vty,
				"%% There are multiple interfaces with name %s\n",
				ifname);
			vty_out(vty, "%% You must specify the VRF name\n");
		}
		return CMD_WARNING;
	}

	if (uj)
		json = json_object_new_object();

	if (json)
		if_dump_vty_json(vty, ifp, json);
	else
		if_dump_vty(vty, ifp);

	if (json)
		vty_json(vty, json);

	return CMD_SUCCESS;
}

static void if_show_description(struct vty *vty, struct vrf *vrf)
{
	struct interface *ifp;

	vty_out(vty, "Interface       Status  Protocol  Description\n");
	FOR_ALL_INTERFACES (vrf, ifp) {
		int len;
		struct zebra_if *zif;
		bool intf_desc;

		intf_desc = false;

		len = vty_out(vty, "%s", ifp->name);
		vty_out(vty, "%*s", (16 - len), " ");

		if (if_is_up(ifp)) {
			vty_out(vty, "up      ");
			if (CHECK_FLAG(ifp->status,
				       ZEBRA_INTERFACE_LINKDETECTION)) {
				if (if_is_running(ifp))
					vty_out(vty, "up        ");
				else
					vty_out(vty, "down      ");
			} else {
				vty_out(vty, "unknown   ");
			}
		} else {
			vty_out(vty, "down    down      ");
		}

		if (ifp->desc) {
			intf_desc = true;
			vty_out(vty, "%s", ifp->desc);
		}
		zif = ifp->info;
		if (zif && zif->desc) {
			vty_out(vty, "%s%s",
				intf_desc
					? "\n                                  "
					: "",
				zif->desc);
		}

		vty_out(vty, "\n");
	}
}

DEFUN (show_interface_desc,
       show_interface_desc_cmd,
       "show interface description vrf NAME",
       SHOW_STR
       "Interface status and configuration\n"
       "Interface description\n"
       VRF_CMD_HELP_STR)
{
	struct vrf *vrf;

	vrf = vrf_lookup_by_name(argv[4]->arg);
	if (!vrf) {
		vty_out(vty, "%% VRF %s not found\n", argv[4]->arg);
		return CMD_WARNING;
	}

	if_show_description(vty, vrf);

	return CMD_SUCCESS;
}


DEFUN (show_interface_desc_vrf_all,
       show_interface_desc_vrf_all_cmd,
       "show interface description [vrf all]",
       SHOW_STR
       "Interface status and configuration\n"
       "Interface description\n"
       VRF_ALL_CMD_HELP_STR)
{
	struct vrf *vrf;

	RB_FOREACH (vrf, vrf_name_head, &vrfs_by_name)
		if (!RB_EMPTY(if_name_head, &vrf->ifaces_by_name)) {
			vty_out(vty, "\n\tVRF %s(%u)\n\n", VRF_LOGNAME(vrf),
				vrf->vrf_id);
			if_show_description(vty, vrf);
		}

	return CMD_SUCCESS;
}

void if_arp(struct interface *ifp, bool enable)
{
	int ret;

	if (!CHECK_FLAG(ifp->status, ZEBRA_INTERFACE_ACTIVE))
		return;

	if (enable)
		ret = if_unset_flags(ifp, IFF_NOARP);
	else
		ret = if_set_flags(ifp, IFF_NOARP);

	if (ret < 0) {
		zlog_debug("Can't %sset noarp flag on interface %s",
			   enable ? "" : "un", ifp->name);
		return;
	}

	if_refresh(ifp);
}

int if_multicast_set(struct interface *ifp)
{
	struct zebra_if *if_data;

	if (CHECK_FLAG(ifp->status, ZEBRA_INTERFACE_ACTIVE)) {
		if (if_set_flags(ifp, IFF_MULTICAST) < 0) {
			zlog_debug("Can't set multicast flag on interface %s",
				   ifp->name);
			return -1;
		}
		if_refresh(ifp);
	}
	if_data = ifp->info;
	if_data->multicast = IF_ZEBRA_DATA_ON;

	return 0;
}

int if_multicast_unset(struct interface *ifp)
{
	struct zebra_if *if_data;

	if (CHECK_FLAG(ifp->status, ZEBRA_INTERFACE_ACTIVE)) {
		if (if_unset_flags(ifp, IFF_MULTICAST) < 0) {
			zlog_debug("Can't unset multicast flag on interface %s",
				   ifp->name);
			return -1;
		}
		if_refresh(ifp);
	}
	if_data = ifp->info;
	if_data->multicast = IF_ZEBRA_DATA_OFF;

	return 0;
}

int if_linkdetect(struct interface *ifp, bool detect)
{
	int if_was_operative;

	if_was_operative = if_is_no_ptm_operative(ifp);
	if (detect) {
		SET_FLAG(ifp->status, ZEBRA_INTERFACE_LINKDETECTION);

		/* When linkdetection is enabled, if might come down */
		if (!if_is_no_ptm_operative(ifp) && if_was_operative)
			if_down(ifp);
	} else {
		UNSET_FLAG(ifp->status, ZEBRA_INTERFACE_LINKDETECTION);

		/* Interface may come up after disabling link detection */
		if (if_is_operative(ifp) && !if_was_operative)
			if_up(ifp, true);
	}
	/* FIXME: Will defer status change forwarding if interface
	   does not come down! */
	return 0;
}

int if_shutdown(struct interface *ifp)
{
	struct zebra_if *if_data;

	if (ifp->ifindex != IFINDEX_INTERNAL) {
		/* send RA lifetime of 0 before stopping. rfc4861/6.2.5 */
		rtadv_stop_ra(ifp);
		if (if_unset_flags(ifp, IFF_UP) < 0) {
			zlog_debug("Can't shutdown interface %s", ifp->name);
			return -1;
		}
		if_refresh(ifp);
	}
	if_data = ifp->info;
	if_data->shutdown = IF_ZEBRA_DATA_ON;

	return 0;
}

int if_no_shutdown(struct interface *ifp)
{
	struct zebra_if *if_data;

	if (ifp->ifindex != IFINDEX_INTERNAL) {
		if (if_set_flags(ifp, IFF_UP | IFF_RUNNING) < 0) {
			zlog_debug("Can't up interface %s", ifp->name);
			return -1;
		}
		if_refresh(ifp);

		/* Some addresses (in particular, IPv6 addresses on Linux) get
		 * removed when the interface goes down. They need to be
		 * readded.
		 */
		if_addr_wakeup(ifp);
	}

	if_data = ifp->info;
	if_data->shutdown = IF_ZEBRA_DATA_OFF;

	return 0;
}

void link_param_cmd_set_uint32(struct interface *ifp, uint32_t *field,
			       uint32_t type, uint32_t value)
{
	/* Update field as needed */
	if (IS_PARAM_UNSET(ifp->link_params, type) || *field != value) {
		*field = value;
		SET_PARAM(ifp->link_params, type);
	}
}

void link_param_cmd_set_float(struct interface *ifp, float *field,
			      uint32_t type, float value)
{
	/* Update field as needed */
	if (IS_PARAM_UNSET(ifp->link_params, type) || *field != value) {
		*field = value;
		SET_PARAM(ifp->link_params, type);
	}
}

void link_param_cmd_unset(struct interface *ifp, uint32_t type)
{
	if (ifp->link_params == NULL)
		return;

	/* Unset field */
	UNSET_PARAM(ifp->link_params, type);
}

void if_ip_address_install(struct interface *ifp, struct prefix *prefix,
			   const char *label, struct prefix *pp)
{
	struct zebra_if *if_data;
	struct connected *ifc;

	if_data = ifp->info;

	ifc = connected_check_ptp(ifp, prefix, pp);
	if (!ifc) {
		ifc = connected_new();
		ifc->ifp = ifp;

		/* Address. */
		ifc->address = prefix_new();
		prefix_copy(ifc->address, prefix);

		if (pp) {
			SET_FLAG(ifc->flags, ZEBRA_IFA_PEER);
			ifc->destination = prefix_new();
			prefix_copy(ifc->destination, pp);
		}

		/* Label. */
		if (label)
			ifc->label = XSTRDUP(MTYPE_CONNECTED_LABEL, label);

		/* Add to linked list. */
		if_connected_add_tail(ifp->connected, ifc);
	}

	/* This address is configured from zebra. */
	if (!CHECK_FLAG(ifc->conf, ZEBRA_IFC_CONFIGURED))
		SET_FLAG(ifc->conf, ZEBRA_IFC_CONFIGURED);

	/* In case of this route need to install kernel. */
	if (!CHECK_FLAG(ifc->conf, ZEBRA_IFC_QUEUED) &&
	    CHECK_FLAG(ifp->status, ZEBRA_INTERFACE_ACTIVE) &&
	    !(if_data && if_data->shutdown == IF_ZEBRA_DATA_ON)) {
		/* Some system need to up the interface to set IP address. */
		if (!if_is_up(ifp)) {
			if_set_flags(ifp, IFF_UP | IFF_RUNNING);
			if_refresh(ifp);
		}

		dplane_intf_addr_set(ifp, ifc);

		SET_FLAG(ifc->conf, ZEBRA_IFC_QUEUED);
		/* The address will be advertised to zebra clients when the
		 * notification
		 * from the kernel has been received.
		 * It will also be added to the subnet chain list, then. */
	}
}

void if_ip_address_uninstall(struct interface *ifp, struct prefix *prefix,
			     struct prefix *pp)
{
	struct connected *ifc;

	ifc = connected_check_ptp(ifp, prefix, pp);
	assert(ifc);

	UNSET_FLAG(ifc->conf, ZEBRA_IFC_CONFIGURED);

	/* This is not real address or interface is not active. */
	if (!CHECK_FLAG(ifc->conf, ZEBRA_IFC_QUEUED)
	    || !CHECK_FLAG(ifp->status, ZEBRA_INTERFACE_ACTIVE)) {
		if_connected_del(ifp->connected, ifc);
		connected_free(&ifc);
		return;
	}

	/* This is real route. */
	dplane_intf_addr_unset(ifp, ifc);

	UNSET_FLAG(ifc->conf, ZEBRA_IFC_QUEUED);
}

void if_ipv6_address_install(struct interface *ifp, struct prefix *prefix)
{
	struct zebra_if *if_data;
	struct connected *ifc;

	if_data = ifp->info;

	ifc = connected_check(ifp, prefix);
	if (!ifc) {
		ifc = connected_new();
		ifc->ifp = ifp;

		/* Address. */
		ifc->address = prefix_new();
		prefix_copy(ifc->address, prefix);

		/* Add to linked list. */
		if_connected_add_tail(ifp->connected, ifc);
	}

	/* This address is configured from zebra. */
	if (!CHECK_FLAG(ifc->conf, ZEBRA_IFC_CONFIGURED))
		SET_FLAG(ifc->conf, ZEBRA_IFC_CONFIGURED);

	/* In case of this route need to install kernel. */
	if (!CHECK_FLAG(ifc->conf, ZEBRA_IFC_QUEUED) &&
	    CHECK_FLAG(ifp->status, ZEBRA_INTERFACE_ACTIVE) &&
	    !(if_data && if_data->shutdown == IF_ZEBRA_DATA_ON)) {
		/* Some system need to up the interface to set IP address. */
		if (!if_is_up(ifp)) {
			if_set_flags(ifp, IFF_UP | IFF_RUNNING);
			if_refresh(ifp);
		}

		dplane_intf_addr_set(ifp, ifc);

		SET_FLAG(ifc->conf, ZEBRA_IFC_QUEUED);
		/* The address will be advertised to zebra clients when the
		 * notification
		 * from the kernel has been received. */
	}
}

void if_ipv6_address_uninstall(struct interface *ifp, struct prefix *prefix)
{
	struct connected *ifc;

	ifc = connected_check(ifp, prefix);
	assert(ifc);

	UNSET_FLAG(ifc->conf, ZEBRA_IFC_CONFIGURED);

	/* This is not real address or interface is not active. */
	if (!CHECK_FLAG(ifc->conf, ZEBRA_IFC_QUEUED)
	    || !CHECK_FLAG(ifp->status, ZEBRA_INTERFACE_ACTIVE)) {
		if_connected_del(ifp->connected, ifc);
		connected_free(&ifc);
		return;
	}

	/* This is real route. */
	dplane_intf_addr_unset(ifp, ifc);

	UNSET_FLAG(ifc->conf, ZEBRA_IFC_QUEUED);
}

/* Allocate and initialize interface vector. */
void zebra_if_init(void)
{
	/* Initialize interface and new hook. */
	hook_register_prio(if_add, 0, if_zebra_new_hook);
	hook_register_prio(if_del, 0, if_zebra_delete_hook);

	install_element(VIEW_NODE, &show_interface_cmd);
	install_element(VIEW_NODE, &show_interface_vrf_all_cmd);
	install_element(VIEW_NODE, &show_interface_name_vrf_cmd);
	install_element(VIEW_NODE, &show_interface_name_vrf_all_cmd);

	install_element(ENABLE_NODE, &show_interface_desc_cmd);
	install_element(ENABLE_NODE, &show_interface_desc_vrf_all_cmd);
}