aboutsummaryrefslogtreecommitdiffstats
path: root/src/plugins/lb/lb.c
blob: 42103eeef2aef0d05263da8298de3cf2e60a87c7 (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
/*
 * Copyright (c) 2016 Cisco and/or its affiliates.
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at:
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */

#include <lb/lb.h>
#include <vnet/plugin/plugin.h>
#include <vpp/app/version.h>
#include <vnet/api_errno.h>
#include <vnet/udp/udp.h>

//GC runs at most once every so many seconds
#define LB_GARBAGE_RUN 60

//After so many seconds. It is assumed that inter-core race condition will not occur.
#define LB_CONCURRENCY_TIMEOUT 10

lb_main_t lb_main;

#define lb_get_writer_lock() do {} while(__sync_lock_test_and_set (lb_main.writer_lock, 1))
#define lb_put_writer_lock() lb_main.writer_lock[0] = 0

static void lb_as_stack (lb_as_t *as);


const static char * const lb_dpo_gre4_ip4[] = { "lb4-gre4" , NULL };
const static char * const lb_dpo_gre4_ip6[] = { "lb6-gre4" , NULL };
const static char* const * const lb_dpo_gre4_nodes[DPO_PROTO_NUM] =
    {
        [DPO_PROTO_IP4]  = lb_dpo_gre4_ip4,
        [DPO_PROTO_IP6]  = lb_dpo_gre4_ip6,
    };

const static char * const lb_dpo_gre6_ip4[] = { "lb4-gre6" , NULL };
const static char * const lb_dpo_gre6_ip6[] = { "lb6-gre6" , NULL };
const static char* const * const lb_dpo_gre6_nodes[DPO_PROTO_NUM] =
    {
        [DPO_PROTO_IP4]  = lb_dpo_gre6_ip4,
        [DPO_PROTO_IP6]  = lb_dpo_gre6_ip6,
    };

const static char * const lb_dpo_gre4_ip4_port[] = { "lb4-gre4-port" , NULL };
const static char * const lb_dpo_gre4_ip6_port[] = { "lb6-gre4-port" , NULL };
const static char* const * const lb_dpo_gre4_port_nodes[DPO_PROTO_NUM] =
    {
        [DPO_PROTO_IP4]  = lb_dpo_gre4_ip4_port,
        [DPO_PROTO_IP6]  = lb_dpo_gre4_ip6_port,
    };

const static char * const lb_dpo_gre6_ip4_port[] = { "lb4-gre6-port" , NULL };
const static char * const lb_dpo_gre6_ip6_port[] = { "lb6-gre6-port" , NULL };
const static char* const * const lb_dpo_gre6_port_nodes[DPO_PROTO_NUM] =
    {
        [DPO_PROTO_IP4]  = lb_dpo_gre6_ip4_port,
        [DPO_PROTO_IP6]  = lb_dpo_gre6_ip6_port,
    };

const static char * const lb_dpo_l3dsr_ip4[] = {"lb4-l3dsr" , NULL};
const static char* const * const lb_dpo_l3dsr_nodes[DPO_PROTO_NUM] =
    {
        [DPO_PROTO_IP4]  = lb_dpo_l3dsr_ip4,
    };

const static char * const lb_dpo_l3dsr_ip4_port[] = {"lb4-l3dsr-port" , NULL};
const static char* const * const lb_dpo_l3dsr_port_nodes[DPO_PROTO_NUM] =
    {
        [DPO_PROTO_IP4]  = lb_dpo_l3dsr_ip4_port,
    };

const static char * const lb_dpo_nat4_ip4_port[] = { "lb4-nat4-port" , NULL };
const static char* const * const lb_dpo_nat4_port_nodes[DPO_PROTO_NUM] =
    {
        [DPO_PROTO_IP4]  = lb_dpo_nat4_ip4_port,
    };

const static char * const lb_dpo_nat6_ip6_port[] = { "lb6-nat6-port" , NULL };
const static char* const * const lb_dpo_nat6_port_nodes[DPO_PROTO_NUM] =
    {
        [DPO_PROTO_IP6]  = lb_dpo_nat6_ip6_port,
    };

u32 lb_hash_time_now(vlib_main_t * vm)
{
  return (u32) (vlib_time_now(vm) + 10000);
}

u8 *format_lb_main (u8 * s, va_list * args)
{
  vlib_thread_main_t *tm = vlib_get_thread_main();
  lb_main_t *lbm = &lb_main;
  s = format(s, "lb_main");
  s = format(s, " ip4-src-address: %U \n", format_ip4_address, &lbm->ip4_src_address);
  s = format(s, " ip6-src-address: %U \n", format_ip6_address, &lbm->ip6_src_address);
  s = format(s, " #vips: %u\n", pool_elts(lbm->vips));
  s = format(s, " #ass: %u\n", pool_elts(lbm->ass) - 1);

  u32 thread_index;
  for(thread_index = 0; thread_index < tm->n_vlib_mains; thread_index++ ) {
    lb_hash_t *h = lbm->per_cpu[thread_index].sticky_ht;
    if (h) {
      s = format(s, "core %d\n", thread_index);
      s = format(s, "  timeout: %ds\n", h->timeout);
      s = format(s, "  usage: %d / %d\n", lb_hash_elts(h, lb_hash_time_now(vlib_get_main())),  lb_hash_size(h));
    }
  }

  return s;
}

static char *lb_vip_type_strings[] = {
    [LB_VIP_TYPE_IP6_GRE6] = "ip6-gre6",
    [LB_VIP_TYPE_IP6_GRE4] = "ip6-gre4",
    [LB_VIP_TYPE_IP4_GRE6] = "ip4-gre6",
    [LB_VIP_TYPE_IP4_GRE4] = "ip4-gre4",
    [LB_VIP_TYPE_IP4_L3DSR] = "ip4-l3dsr",
    [LB_VIP_TYPE_IP4_NAT4] = "ip4-nat4",
    [LB_VIP_TYPE_IP6_NAT6] = "ip6-nat6",
};

u8 *format_lb_vip_type (u8 * s, va_list * args)
{
  lb_vip_type_t vipt = va_arg (*args, lb_vip_type_t);
  u32 i;
  for (i=0; i<LB_VIP_N_TYPES; i++)
    if (vipt == i)
      return format(s, lb_vip_type_strings[i]);
  return format(s, "_WRONG_TYPE_");
}

uword unformat_lb_vip_type (unformat_input_t * input, va_list * args)
{
  lb_vip_type_t *vipt = va_arg (*args, lb_vip_type_t *);
  u32 i;
  for (i=0; i<LB_VIP_N_TYPES; i++)
    if (unformat(input, lb_vip_type_strings[i])) {
      *vipt = i;
      return 1;
    }
  return 0;
}

u8 *format_lb_vip (u8 * s, va_list * args)
{
  lb_vip_t *vip = va_arg (*args, lb_vip_t *);
  s = format(s, "%U %U new_size:%u #as:%u%s",
             format_lb_vip_type, vip->type,
             format_ip46_prefix, &vip->prefix, vip->plen, IP46_TYPE_ANY,
             vip->new_flow_table_mask + 1,
             pool_elts(vip->as_indexes),
             (vip->flags & LB_VIP_FLAGS_USED)?"":" removed");

  if (vip->port != 0)
    {
      s = format(s, "  protocol:%u port:%u ", vip->protocol, vip->port);
    }

  if (vip->type == LB_VIP_TYPE_IP4_L3DSR)
    {
      s = format(s, "  dscp:%u", vip->encap_args.dscp);
    }
  else if ((vip->type == LB_VIP_TYPE_IP4_NAT4)
          || (vip->type == LB_VIP_TYPE_IP6_NAT6))
    {
      s = format (s, " type:%s port:%u target_port:%u",
         (vip->encap_args.srv_type == LB_SRV_TYPE_CLUSTERIP)?"clusterip":
             "nodeport",
         ntohs(vip->port), ntohs(vip->encap_args.target_port));
    }

  return s;
}

u8 *format_lb_as (u8 * s, va_list * args)
{
  lb_as_t *as = va_arg (*args, lb_as_t *);
  return format(s, "%U %s", format_ip46_address,
                &as->address, IP46_TYPE_ANY,
                (as->flags & LB_AS_FLAGS_USED)?"used":"removed");
}

u8 *format_lb_vip_detailed (u8 * s, va_list * args)
{
  lb_main_t *lbm = &lb_main;
  lb_vip_t *vip = va_arg (*args, lb_vip_t *);
  u32 indent = format_get_indent (s);

  s = format(s, "%U %U [%lu] %U%s\n"
                   "%U  new_size:%u\n",
                  format_white_space, indent,
                  format_lb_vip_type, vip->type,
                  vip - lbm->vips,
                  format_ip46_prefix, &vip->prefix, (u32) vip->plen, IP46_TYPE_ANY,
                  (vip->flags & LB_VIP_FLAGS_USED)?"":" removed",
                  format_white_space, indent,
                  vip->new_flow_table_mask + 1);

  if (vip->port != 0)
    {
      s = format(s, "%U  protocol:%u port:%u\n",
                 format_white_space, indent,
                 vip->protocol, vip->port);
    }

  if (vip->type == LB_VIP_TYPE_IP4_L3DSR)
    {
      s = format(s, "%U  dscp:%u\n",
                    format_white_space, indent,
                    vip->encap_args.dscp);
    }
  else if ((vip->type == LB_VIP_TYPE_IP4_NAT4)
          || (vip->type == LB_VIP_TYPE_IP6_NAT6))
    {
      s = format (s, "%U  type:%s port:%u target_port:%u",
         format_white_space, indent,
         (vip->encap_args.srv_type == LB_SRV_TYPE_CLUSTERIP)?"clusterip":
             "nodeport",
         ntohs(vip->port), ntohs(vip->encap_args.target_port));
    }

  //Print counters
  s = format(s, "%U  counters:\n",
             format_white_space, indent);
  u32 i;
  for (i=0; i<LB_N_VIP_COUNTERS; i++)
    s = format(s, "%U    %s: %Lu\n",
               format_white_space, indent,
               lbm->vip_counters[i].name,
               vlib_get_simple_counter(&lbm->vip_counters[i], vip - lbm->vips));


  s = format(s, "%U  #as:%u\n",
             format_white_space, indent,
             pool_elts(vip->as_indexes));

  //Let's count the buckets for each AS
  u32 *count = 0;
  vec_validate(count, pool_len(lbm->ass)); //Possibly big alloc for not much...
  lb_new_flow_entry_t *nfe;
  vec_foreach(nfe, vip->new_flow_table)
    count[nfe->as_index]++;

  lb_as_t *as;
  u32 *as_index;
  pool_foreach(as_index, vip->as_indexes, {
      as = &lbm->ass[*as_index];
      s = format(s, "%U    %U %u buckets   %Lu flows  dpo:%u %s\n",
                   format_white_space, indent,
                   format_ip46_address, &as->address, IP46_TYPE_ANY,
                   count[as - lbm->ass],
                   vlib_refcount_get(&lbm->as_refcount, as - lbm->ass),
                   as->dpo.dpoi_index,
                   (as->flags & LB_AS_FLAGS_USED)?"used":" removed");
  });

  vec_free(count);
  return s;
}

typedef struct {
  u32 as_index;
  u32 last;
  u32 skip;
} lb_pseudorand_t;

static int lb_pseudorand_compare(void *a, void *b)
{
  lb_as_t *asa, *asb;
  lb_main_t *lbm = &lb_main;
  asa = &lbm->ass[((lb_pseudorand_t *)a)->as_index];
  asb = &lbm->ass[((lb_pseudorand_t *)b)->as_index];
  return memcmp(&asa->address, &asb->address, sizeof(asb->address));
}

static void lb_vip_garbage_collection(lb_vip_t *vip)
{
  lb_main_t *lbm = &lb_main;
  lb_snat4_key_t m_key4;
  clib_bihash_kv_8_8_t kv4, value4;
  lb_snat6_key_t m_key6;
  clib_bihash_kv_24_8_t kv6, value6;
  lb_snat_mapping_t *m = 0;
  ASSERT (lbm->writer_lock[0]);

  u32 now = (u32) vlib_time_now(vlib_get_main());
  if (!clib_u32_loop_gt(now, vip->last_garbage_collection + LB_GARBAGE_RUN))
    return;

  vip->last_garbage_collection = now;
  lb_as_t *as;
  u32 *as_index;
  pool_foreach(as_index, vip->as_indexes, {
      as = &lbm->ass[*as_index];
      if (!(as->flags & LB_AS_FLAGS_USED) && //Not used
          clib_u32_loop_gt(now, as->last_used + LB_CONCURRENCY_TIMEOUT) &&
          (vlib_refcount_get(&lbm->as_refcount, as - lbm->ass) == 0))
        { //Not referenced

          if (lb_vip_is_nat4_port(vip)) {
              m_key4.addr = as->address.ip4;
              m_key4.port = vip->encap_args.target_port;
              m_key4.protocol = 0;
              m_key4.fib_index = 0;

              kv4.key = m_key4.as_u64;
              if(!clib_bihash_search_8_8(&lbm->mapping_by_as4, &kv4, &value4))
                m = pool_elt_at_index (lbm->snat_mappings, value4.value);
              ASSERT (m);

              kv4.value = m - lbm->snat_mappings;
              clib_bihash_add_del_8_8(&lbm->mapping_by_as4, &kv4, 0);
              pool_put (lbm->snat_mappings, m);
          } else if (lb_vip_is_nat6_port(vip)) {
              m_key6.addr.as_u64[0] = as->address.ip6.as_u64[0];
              m_key6.addr.as_u64[1] = as->address.ip6.as_u64[1];
              m_key6.port = vip->encap_args.target_port;
              m_key6.protocol = 0;
              m_key6.fib_index = 0;

              kv6.key[0] = m_key6.as_u64[0];
              kv6.key[1] = m_key6.as_u64[1];
              kv6.key[2] = m_key6.as_u64[2];

              if (!clib_bihash_search_24_8 (&lbm->mapping_by_as6, &kv6, &value6))
                m = pool_elt_at_index (lbm->snat_mappings, value6.value);
              ASSERT (m);

              kv6.value = m - lbm->snat_mappings;
              clib_bihash_add_del_24_8(&lbm->mapping_by_as6, &kv6, 0);
              pool_put (lbm->snat_mappings, m);
          }
          fib_entry_child_remove(as->next_hop_fib_entry_index,
                                as->next_hop_child_index);
          fib_table_entry_delete_index(as->next_hop_fib_entry_index,
                                       FIB_SOURCE_RR);
          as->next_hop_fib_entry_index = FIB_NODE_INDEX_INVALID;

          pool_put(vip->as_indexes, as_index);
          pool_put(lbm->ass, as);
        }
  });
}

void lb_garbage_collection()
{
  lb_main_t *lbm = &lb_main;
  lb_get_writer_lock();
  lb_vip_t *vip;
  u32 *to_be_removed_vips = 0, *i;
  pool_foreach(vip, lbm->vips, {
      lb_vip_garbage_collection(vip);

      if (!(vip->flags & LB_VIP_FLAGS_USED) &&
          (pool_elts(vip->as_indexes) == 0)) {
        vec_add1(to_be_removed_vips, vip - lbm->vips);
      }
  });

  vec_foreach(i, to_be_removed_vips) {
    vip = &lbm->vips[*i];
    pool_put(lbm->vips, vip);
    pool_free(vip->as_indexes);
  }

  vec_free(to_be_removed_vips);
  lb_put_writer_lock();
}

static void lb_vip_update_new_flow_table(lb_vip_t *vip)
{
  lb_main_t *lbm = &lb_main;
  lb_new_flow_entry_t *old_table;
  u32 i, *as_index;
  lb_new_flow_entry_t *new_flow_table = 0;
  lb_as_t *as;
  lb_pseudorand_t *pr, *sort_arr = 0;
  u32 count;

  ASSERT (lbm->writer_lock[0]); //We must have the lock

  //Check if some AS is configured or not
  i = 0;
  pool_foreach(as_index, vip->as_indexes, {
      as = &lbm->ass[*as_index];
      if (as->flags & LB_AS_FLAGS_USED) { //Not used anymore
        i = 1;
        goto out; //Not sure 'break' works in this macro-loop
      }
  });

out:
  if (i == 0) {
    //Only the default. i.e. no AS
    vec_validate(new_flow_table, vip->new_flow_table_mask);
    for (i=0; i<vec_len(new_flow_table); i++)
      new_flow_table[i].as_index = 0;

    goto finished;
  }

  //First, let's sort the ASs
  sort_arr = 0;
  vec_alloc(sort_arr, pool_elts(vip->as_indexes));

  i = 0;
  pool_foreach(as_index, vip->as_indexes, {
      as = &lbm->ass[*as_index];
      if (!(as->flags & LB_AS_FLAGS_USED)) //Not used anymore
        continue;

      sort_arr[i].as_index = as - lbm->ass;
      i++;
  });
  _vec_len(sort_arr) = i;

  vec_sort_with_function(sort_arr, lb_pseudorand_compare);

  //Now let's pseudo-randomly generate permutations
  vec_foreach(pr, sort_arr) {
    lb_as_t *as = &lbm->ass[pr->as_index];

    u64 seed = clib_xxhash(as->address.as_u64[0] ^
                           as->address.as_u64[1]);
    /* We have 2^n buckets.
     * skip must be prime with 2^n.
     * So skip must be odd.
     * MagLev actually state that M should be prime,
     * but this has a big computation cost (% operation).
     * Using 2^n is more better (& operation).
     */
    pr->skip = ((seed & 0xffffffff) | 1) & vip->new_flow_table_mask;
    pr->last = (seed >> 32) & vip->new_flow_table_mask;
  }

  //Let's create a new flow table
  vec_validate(new_flow_table, vip->new_flow_table_mask);
  for (i=0; i<vec_len(new_flow_table); i++)
    new_flow_table[i].as_index = ~0;

  u32 done = 0;
  while (1) {
    vec_foreach(pr, sort_arr) {
      while (1) {
        u32 last = pr->last;
        pr->last = (pr->last + pr->skip) & vip->new_flow_table_mask;
        if (new_flow_table[last].as_index == ~0) {
          new_flow_table[last].as_index = pr->as_index;
          break;
        }
      }
      done++;
      if (done == vec_len(new_flow_table))
        goto finished;
    }
  }

  vec_free(sort_arr);

finished:

//Count number of changed entries
  count = 0;
  for (i=0; i<vec_len(new_flow_table); i++)
    if (vip->new_flow_table == 0 ||
        new_flow_table[i].as_index != vip->new_flow_table[i].as_index)
      count++;

  old_table = vip->new_flow_table;
  vip->new_flow_table = new_flow_table;
  vec_free(old_table);
}

int lb_conf(ip4_address_t *ip4_address, ip6_address_t *ip6_address,
           u32 per_cpu_sticky_buckets, u32 flow_timeout)
{
  lb_main_t *lbm = &lb_main;

  if (!is_pow2(per_cpu_sticky_buckets))
    return VNET_API_ERROR_INVALID_MEMORY_SIZE;

  lb_get_writer_lock(); //Not exactly necessary but just a reminder that it exists for my future self
  lbm->ip4_src_address = *ip4_address;
  lbm->ip6_src_address = *ip6_address;
  lbm->per_cpu_sticky_buckets = per_cpu_sticky_buckets;
  lbm->flow_timeout = flow_timeout;
  lb_put_writer_lock();
  return 0;
}



static
int lb_vip_port_find_index(ip46_address_t *prefix, u8 plen,
                           u8 protocol, u16 port,
                           lb_lkp_type_t lkp_type,
                           u32 *vip_index)
{
  lb_main_t *lbm = &lb_main;
  lb_vip_t *vip;
  ASSERT (lbm->writer_lock[0]); //This must be called with the lock owned
  ip46_prefix_normalize(prefix, plen);
  pool_foreach(vip, lbm->vips, {
      if ((vip->flags & LB_AS_FLAGS_USED) &&
          vip->plen == plen &&
          vip->prefix.as_u64[0] == prefix->as_u64[0] &&
          vip->prefix.as_u64[1] == prefix->as_u64[1])
        {
          if((lkp_type == LB_LKP_SAME_IP_PORT &&
               vip->protocol == protocol &&
               vip->port == port) ||
             (lkp_type == LB_LKP_ALL_PORT_IP &&
               vip->port == 0) ||
             (lkp_type == LB_LKP_DIFF_IP_PORT &&
                (vip->protocol != protocol ||
                vip->port != port) ) )
            {
              *vip_index = vip - lbm->vips;
              return 0;
            }
        }
  });
  return VNET_API_ERROR_NO_SUCH_ENTRY;
}

static
int lb_vip_port_find_index_with_lock(ip46_address_t *prefix, u8 plen,
                                     u8 protocol, u16 port, u32 *vip_index)
{
  return lb_vip_port_find_index(prefix, plen, protocol, port,
                                LB_LKP_SAME_IP_PORT, vip_index);
}

static
int lb_vip_port_find_all_port_vip(ip46_address_t *prefix, u8 plen,
                                  u32 *vip_index)
{
  return lb_vip_port_find_index(prefix, plen, ~0, 0,
                                LB_LKP_ALL_PORT_IP, vip_index);
}

/* Find out per-port-vip entry with different protocol and port */
static
int lb_vip_port_find_diff_port(ip46_address_t *prefix, u8 plen,
                               u8 protocol, u16 port, u32 *vip_index)
{
  return lb_vip_port_find_index(prefix, plen, protocol, port,
                                LB_LKP_DIFF_IP_PORT, vip_index);
}

int lb_vip_find_index(ip46_address_t *prefix, u8 plen, u8 protocol,
                      u16 port, u32 *vip_index)
{
  int ret;
  lb_get_writer_lock();
  ret = lb_vip_port_find_index_with_lock(prefix, plen,
                                         protocol, port, vip_index);
  lb_put_writer_lock();
  return ret;
}

static int lb_as_find_index_vip(lb_vip_t *vip, ip46_address_t *address, u32 *as_index)
{
  lb_main_t *lbm = &lb_main;
  ASSERT (lbm->writer_lock[0]); //This must be called with the lock owned
  lb_as_t *as;
  u32 *asi;
  pool_foreach(asi, vip->as_indexes, {
      as = &lbm->ass[*asi];
      if (as->vip_index == (vip - lbm->vips) &&
          as->address.as_u64[0] == address->as_u64[0] &&
          as->address.as_u64[1] == address->as_u64[1])
      {
        *as_index = as - lbm->ass;
        return 0;
      }
  });
  return -1;
}

int lb_vip_add_ass(u32 vip_index, ip46_address_t *addresses, u32 n)
{
  lb_main_t *lbm = &lb_main;
  lb_get_writer_lock();
  lb_vip_t *vip;
  if (!(vip = lb_vip_get_by_index(vip_index))) {
    lb_put_writer_lock();
    return VNET_API_ERROR_NO_SUCH_ENTRY;
  }

  ip46_type_t type = lb_encap_is_ip4(vip)?IP46_TYPE_IP4:IP46_TYPE_IP6;
  u32 *to_be_added = 0;
  u32 *to_be_updated = 0;
  u32 i;
  u32 *ip;
  lb_snat_mapping_t *m;

  //Sanity check
  while (n--) {

    if (!lb_as_find_index_vip(vip, &addresses[n], &i)) {
      if (lbm->ass[i].flags & LB_AS_FLAGS_USED) {
        vec_free(to_be_added);
        vec_free(to_be_updated);
        lb_put_writer_lock();
        return VNET_API_ERROR_VALUE_EXIST;
      }
      vec_add1(to_be_updated, i);
      goto next;
    }

    if (ip46_address_type(&addresses[n]) != type) {
      vec_free(to_be_added);
      vec_free(to_be_updated);
      lb_put_writer_lock();
      return VNET_API_ERROR_INVALID_ADDRESS_FAMILY;
    }

    if (n) {
      u32 n2 = n;
      while(n2--) //Check for duplicates
        if (addresses[n2].as_u64[0] == addresses[n].as_u64[0] &&
            addresses[n2].as_u64[1] == addresses[n].as_u64[1])
          goto next;
    }

    vec_add1(to_be_added, n);

next:
    continue;
  }

  //Update reused ASs
  vec_foreach(ip, to_be_updated) {
    lbm->ass[*ip].flags = LB_AS_FLAGS_USED;
  }
  vec_free(to_be_updated);

  //Create those who have to be created
  vec_foreach(ip, to_be_added) {
    lb_as_t *as;
    u32 *as_index;
    pool_get(lbm->ass, as);
    as->address = addresses[*ip];
    as->flags = LB_AS_FLAGS_USED;
    as->vip_index = vip_index;
    pool_get(vip->as_indexes, as_index);
    *as_index = as - lbm->ass;

    /*
     * become a child of the FIB entry
     * so we are informed when its forwarding changes
     */
    fib_prefix_t nh = {};
    if (lb_encap_is_ip4(vip)) {
        nh.fp_addr.ip4 = as->address.ip4;
        nh.fp_len = 32;
        nh.fp_proto = FIB_PROTOCOL_IP4;
    } else {
        nh.fp_addr.ip6 = as->address.ip6;
        nh.fp_len = 128;
        nh.fp_proto = FIB_PROTOCOL_IP6;
    }

    as->next_hop_fib_entry_index =
        fib_table_entry_special_add(0,
                                    &nh,
                                    FIB_SOURCE_RR,
                                    FIB_ENTRY_FLAG_NONE);
    as->next_hop_child_index =
        fib_entry_child_add(as->next_hop_fib_entry_index,
                            lbm->fib_node_type,
                            as - lbm->ass);

    lb_as_stack(as);

    if ( lb_vip_is_nat4_port(vip) || lb_vip_is_nat6_port(vip) )
      {
        /* Add SNAT static mapping */
        pool_get (lbm->snat_mappings, m);
        memset (m, 0, sizeof (*m));
        if (lb_vip_is_nat4_port(vip)) {
            lb_snat4_key_t m_key4;
            clib_bihash_kv_8_8_t kv4;
            m_key4.addr = as->address.ip4;
            m_key4.port = vip->encap_args.target_port;
            m_key4.protocol = 0;
            m_key4.fib_index = 0;

            if (vip->encap_args.srv_type == LB_SRV_TYPE_CLUSTERIP)
              {
                m->src_ip.ip4 = vip->prefix.ip4;
              }
            else if (vip->encap_args.srv_type == LB_SRV_TYPE_NODEPORT)
              {
                m->src_ip.ip4 = lbm->ip4_src_address;
              }
            m->src_ip_is_ipv6 = 0;
            m->as_ip.ip4 = as->address.ip4;
            m->as_ip_is_ipv6 = 0;
            m->src_port = vip->port;
            m->target_port = vip->encap_args.target_port;
            m->vrf_id = 0;
            m->fib_index = 0;

            kv4.key = m_key4.as_u64;
            kv4.value = m - lbm->snat_mappings;
            clib_bihash_add_del_8_8(&lbm->mapping_by_as4, &kv4, 1);
        } else {
            lb_snat6_key_t m_key6;
            clib_bihash_kv_24_8_t kv6;
            m_key6.addr.as_u64[0] = as->address.ip6.as_u64[0];
            m_key6.addr.as_u64[1] = as->address.ip6.as_u64[1];
            m_key6.port = vip->encap_args.target_port;
            m_key6.protocol = 0;
            m_key6.fib_index = 0;

            if (vip->encap_args.srv_type == LB_SRV_TYPE_CLUSTERIP)
              {
                m->src_ip.ip6.as_u64[0] = vip->prefix.ip6.as_u64[0];
                m->src_ip.ip6.as_u64[1] = vip->prefix.ip6.as_u64[1];
              }
            else if (vip->encap_args.srv_type == LB_SRV_TYPE_NODEPORT)
              {
                m->src_ip.ip6.as_u64[0] = lbm->ip6_src_address.as_u64[0];
                m->src_ip.ip6.as_u64[1] = lbm->ip6_src_address.as_u64[1];
              }
            m->src_ip_is_ipv6 = 1;
            m->as_ip.ip6.as_u64[0] = as->address.ip6.as_u64[0];
            m->as_ip.ip6.as_u64[1] = as->address.ip6.as_u64[1];
            m->as_ip_is_ipv6 = 1;
            m->src_port = vip->port;
            m->target_port = vip->encap_args.target_port;
            m->vrf_id = 0;
            m->fib_index = 0;

            kv6.key[0] = m_key6.as_u64[0];
            kv6.key[1] = m_key6.as_u64[1];
            kv6.key[2] = m_key6.as_u64[2];
            kv6.value = m - lbm->snat_mappings;
            clib_bihash_add_del_24_8(&lbm->mapping_by_as6, &kv6, 1);
        }
      }
  }
  vec_free(to_be_added);

  //Recompute flows
  lb_vip_update_new_flow_table(vip);

  //Garbage collection maybe
  lb_vip_garbage_collection(vip);

  lb_put_writer_lock();
  return 0;
}

int
lb_flush_vip_as (u32 vip_index, u32 as_index)
{
  u32 thread_index;
  vlib_thread_main_t *tm = vlib_get_thread_main();
  lb_main_t *lbm = &lb_main;

  for(thread_index = 0; thread_index < tm->n_vlib_mains; thread_index++ ) {
    lb_hash_t *h = lbm->per_cpu[thread_index].sticky_ht;
    if (h != NULL) {
        u32 i;
        lb_hash_bucket_t *b;

        lb_hash_foreach_entry(h, b, i) {
          if ((b->vip[i] == vip_index)
              || (b->value[i] == as_index))
            {
              vlib_refcount_add(&lbm->as_refcount, thread_index, b->value[i], -1);
              vlib_refcount_add(&lbm->as_refcount, thread_index, 0, 1);
              b->vip[i] = ~0;
              b->value[i] = ~0;
            }
        }
    }
  }

  return 0;
}

int lb_vip_del_ass_withlock(u32 vip_index, ip46_address_t *addresses, u32 n,
                            u32 *as_index)
{
  lb_main_t *lbm = &lb_main;
  u32 now = (u32) vlib_time_now(vlib_get_main());
  u32 *ip = 0;

  lb_vip_t *vip;
  if (!(vip = lb_vip_get_by_index(vip_index))) {
    return VNET_API_ERROR_NO_SUCH_ENTRY;
  }

  u32 *indexes = NULL;
  while (n--) {
    if (lb_as_find_index_vip(vip, &addresses[n], as_index)) {
      vec_free(indexes);
      return VNET_API_ERROR_NO_SUCH_ENTRY;
    }

    if (n) { //Check for duplicates
      u32 n2 = n - 1;
      while(n2--) {
        if (addresses[n2].as_u64[0] == addresses[n].as_u64[0] &&
            addresses[n2].as_u64[1] == addresses[n].as_u64[1])
          goto next;
      }
    }

    vec_add1(indexes, *as_index);
next:
  continue;
  }

  //Garbage collection maybe
  lb_vip_garbage_collection(vip);

  if (indexes != NULL) {
    vec_foreach(ip, indexes) {
      lbm->ass[*ip].flags &= ~LB_AS_FLAGS_USED;
      lbm->ass[*ip].last_used = now;
    }

    //Recompute flows
    lb_vip_update_new_flow_table(vip);
  }

  vec_free(indexes);
  return 0;
}

int lb_vip_del_ass(u32 vip_index, ip46_address_t *addresses, u32 n, u8 flush)
{
  u32 as_index = 0;
  lb_get_writer_lock();
  int ret = lb_vip_del_ass_withlock(vip_index, addresses, n, &as_index);
  lb_put_writer_lock();

  if(flush)
    {
      /* flush flow table per as of per-port-vip */
      ret = lb_flush_vip_as(vip_index, as_index);
    }

  return ret;
}

static int
lb_vip_prefix_index_alloc (lb_main_t *lbm)
{
  /*
   * Check for dynamically allocaetd instance number.
   */
  u32 bit;

  bit = clib_bitmap_first_clear (lbm->vip_prefix_indexes);

  lbm->vip_prefix_indexes = clib_bitmap_set(lbm->vip_prefix_indexes, bit, 1);

  return bit;
}

static int
lb_vip_prefix_index_free (lb_main_t *lbm, u32 instance)
{

  if (clib_bitmap_get (lbm->vip_prefix_indexes, instance) == 0)
    {
      return -1;
    }

  lbm->vip_prefix_indexes = clib_bitmap_set (lbm->vip_prefix_indexes,
                                             instance, 0);

  return 0;
}

/**
 * Add the VIP adjacency to the ip4 or ip6 fib
 */
static void lb_vip_add_adjacency(lb_main_t *lbm, lb_vip_t *vip,
                                 u32 *vip_prefix_index)
{
  dpo_proto_t proto = 0;
  dpo_type_t dpo_type = 0;
  u32 vip_idx = 0;

  if (vip->port != 0)
    {
      /* for per-port vip, if VIP adjacency has been added,
       * no need to add adjacency. */
      if (!lb_vip_port_find_diff_port(&(vip->prefix), vip->plen,
                                      vip->protocol, vip->port, &vip_idx))
        {
          return;
        }

      /* Allocate an index for per-port vip */
      *vip_prefix_index = lb_vip_prefix_index_alloc(lbm);
    }
  else
    {
      *vip_prefix_index = vip - lbm->vips;
    }

  dpo_id_t dpo = DPO_INVALID;
  fib_prefix_t pfx = {};
  if (lb_vip_is_ip4(vip->type)) {
      pfx.fp_addr.ip4 = vip->prefix.ip4;
      pfx.fp_len = vip->plen - 96;
      pfx.fp_proto = FIB_PROTOCOL_IP4;
      proto = DPO_PROTO_IP4;
  } else {
      pfx.fp_addr.ip6 = vip->prefix.ip6;
      pfx.fp_len = vip->plen;
      pfx.fp_proto = FIB_PROTOCOL_IP6;
      proto = DPO_PROTO_IP6;
  }

  if (lb_vip_is_gre4(vip))
    dpo_type = lbm->dpo_gre4_type;
  else if (lb_vip_is_gre6(vip))
    dpo_type = lbm->dpo_gre6_type;
  else if (lb_vip_is_gre4_port(vip))
    dpo_type = lbm->dpo_gre4_port_type;
  else if (lb_vip_is_gre6_port(vip))
    dpo_type = lbm->dpo_gre6_port_type;
  else if (lb_vip_is_l3dsr(vip))
    dpo_type = lbm->dpo_l3dsr_type;
  else if (lb_vip_is_l3dsr_port(vip))
    dpo_type = lbm->dpo_l3dsr_port_type;
  else if(lb_vip_is_nat4_port(vip))
    dpo_type = lbm->dpo_nat4_port_type;
  else if (lb_vip_is_nat6_port(vip))
    dpo_type = lbm->dpo_nat6_port_type;

  dpo_set(&dpo, dpo_type, proto, *vip_prefix_index);
  fib_table_entry_special_dpo_add(0,
                                  &pfx,
                                  FIB_SOURCE_PLUGIN_HI,
                                  FIB_ENTRY_FLAG_EXCLUSIVE,
                                  &dpo);
  dpo_reset(&dpo);
}

/**
 * Add the VIP filter entry
 */
static int lb_vip_add_port_filter(lb_main_t *lbm, lb_vip_t *vip,
                                  u32 vip_prefix_index, u32 vip_idx)
{
  vip_port_key_t key;
  clib_bihash_kv_8_8_t kv;

  key.vip_prefix_index = vip_prefix_index;
  key.protocol = vip->protocol;
  key.port = clib_host_to_net_u16(vip->port);
  key.rsv = 0;

  kv.key = key.as_u64;
  kv.value = vip_idx;
  clib_bihash_add_del_8_8(&lbm->vip_index_per_port, &kv, 1);

  return 0;
}

/**
 * Del the VIP filter entry
 */
static int lb_vip_del_port_filter(lb_main_t *lbm, lb_vip_t *vip)
{
  vip_port_key_t key;
  clib_bihash_kv_8_8_t kv, value;
  lb_vip_t *m = 0;

  key.vip_prefix_index = vip->vip_prefix_index;
  key.protocol = vip->protocol;
  key.port = clib_host_to_net_u16(vip->port);

  kv.key = key.as_u64;
  if(clib_bihash_search_8_8(&lbm->vip_index_per_port, &kv, &value) == 0)
    m = pool_elt_at_index (lbm->vips, value.value);
  ASSERT (m);

  kv.value = m - lbm->vips;
  clib_bihash_add_del_8_8(&lbm->vip_index_per_port, &kv, 0);

  return 0;
}

/**
 * Deletes the adjacency associated with the VIP
 */
static void lb_vip_del_adjacency(lb_main_t *lbm, lb_vip_t *vip)
{
  fib_prefix_t pfx = {};
  u32 vip_idx = 0;

  if (vip->port != 0)
    {
      /* If this vip adjacency is used by other per-port vip,
       * no need to del this adjacency. */
      if (!lb_vip_port_find_diff_port(&(vip->prefix), vip->plen,
                                      vip->protocol, vip->port, &vip_idx))
        {
          lb_put_writer_lock();
          return;
        }

      /* Return vip_prefix_index for per-port vip */
      lb_vip_prefix_index_free(lbm, vip->vip_prefix_index);

    }

  if (lb_vip_is_ip4(vip->type)) {
      pfx.fp_addr.ip4 = vip->prefix.ip4;
      pfx.fp_len = vip->plen - 96;
      pfx.fp_proto = FIB_PROTOCOL_IP4;
  } else {
      pfx.fp_addr.ip6 = vip->prefix.ip6;
      pfx.fp_len = vip->plen;
      pfx.fp_proto = FIB_PROTOCOL_IP6;
  }
  fib_table_entry_special_remove(0, &pfx, FIB_SOURCE_PLUGIN_HI);
}

int lb_vip_add(lb_vip_add_args_t args, u32 *vip_index)
{
  lb_main_t *lbm = &lb_main;
  vlib_main_t *vm = vlib_get_main();
  lb_vip_t *vip;
  lb_vip_type_t type = args.type;
  u32 vip_prefix_index = 0;

  lb_get_writer_lock();
  ip46_prefix_normalize(&(args.prefix), args.plen);

  if (!lb_vip_port_find_index_with_lock(&(args.prefix), args.plen,
                                         args.protocol, args.port,
                                         vip_index))
    {
      lb_put_writer_lock();
      return VNET_API_ERROR_VALUE_EXIST;
    }

  /* Make sure we can't add a per-port VIP entry
   * when there already is an all-port VIP for the same prefix. */
  if ((args.port != 0) &&
      !lb_vip_port_find_all_port_vip(&(args.prefix), args.plen, vip_index))
    {
      lb_put_writer_lock();
      return VNET_API_ERROR_VALUE_EXIST;
    }

  /* Make sure we can't add a all-port VIP entry
   * when there already is an per-port VIP for the same prefix. */
  if ((args.port == 0) &&
      !lb_vip_port_find_diff_port(&(args.prefix), args.plen,
                                  args.protocol, args.port, vip_index))
    {
      lb_put_writer_lock();
      return VNET_API_ERROR_VALUE_EXIST;
    }

  /* Make sure all VIP for a given prefix (using different ports) have the same type. */
  if ((args.port != 0) &&
      !lb_vip_port_find_diff_port(&(args.prefix), args.plen,
                                  args.protocol, args.port, vip_index)
      && (args.type != lbm->vips[*vip_index].type))
    {
      lb_put_writer_lock();
      return VNET_API_ERROR_INVALID_ARGUMENT;
    }

  if (!is_pow2(args.new_length)) {
    lb_put_writer_lock();
    return VNET_API_ERROR_INVALID_MEMORY_SIZE;
  }

  if (ip46_prefix_is_ip4(&(args.prefix), args.plen) &&
      !lb_vip_is_ip4(type)) {
    lb_put_writer_lock();
    return VNET_API_ERROR_INVALID_ADDRESS_FAMILY;
  }

  if ((!ip46_prefix_is_ip4(&(args.prefix), args.plen)) &&
      !lb_vip_is_ip6(type)) {
    lb_put_writer_lock();
    return VNET_API_ERROR_INVALID_ADDRESS_FAMILY;
  }

  if ((type == LB_VIP_TYPE_IP4_L3DSR) &&
      (args.encap_args.dscp >= 64) )
    {
      lb_put_writer_lock();
      return VNET_API_ERROR_VALUE_EXIST;
    }

  //Allocate
  pool_get(lbm->vips, vip);

  //Init
  memcpy (&(vip->prefix), &(args.prefix), sizeof(args.prefix));
  vip->plen = args.plen;
  if (args.port != 0)
    {
      vip->protocol = args.protocol;
      vip->port = args.port;
    }
  else
    {
      vip->protocol = (u8)~0;
      vip->port = 0;
    }
  vip->last_garbage_collection = (u32) vlib_time_now(vlib_get_main());
  vip->type = args.type;

  if (args.type == LB_VIP_TYPE_IP4_L3DSR) {
      vip->encap_args.dscp = args.encap_args.dscp;
    }
  else if ((args.type == LB_VIP_TYPE_IP4_NAT4)
           ||(args.type == LB_VIP_TYPE_IP6_NAT6)) {
      vip->encap_args.srv_type = args.encap_args.srv_type;
      vip->encap_args.target_port =
          clib_host_to_net_u16(args.encap_args.target_port);
    }

  vip->flags = LB_VIP_FLAGS_USED;
  vip->as_indexes = 0;

  //Validate counters
  u32 i;
  for (i = 0; i < LB_N_VIP_COUNTERS; i++) {
    vlib_validate_simple_counter(&lbm->vip_counters[i], vip - lbm->vips);
    vlib_zero_simple_counter(&lbm->vip_counters[i], vip - lbm->vips);
  }

  //Configure new flow table
  vip->new_flow_table_mask = args.new_length - 1;
  vip->new_flow_table = 0;

  //Update flow hash table
  lb_vip_update_new_flow_table(vip);

  //Create adjacency to direct traffic
  lb_vip_add_adjacency(lbm, vip, &vip_prefix_index);

  if ( (lb_vip_is_nat4_port(vip) || lb_vip_is_nat6_port(vip))
      && (args.encap_args.srv_type == LB_SRV_TYPE_NODEPORT) )
    {
      u32 key;
      uword * entry;

      //Create maping from nodeport to vip_index
      key = clib_host_to_net_u16(args.port);
      entry = hash_get_mem (lbm->vip_index_by_nodeport, &key);
      if (entry) {
        lb_put_writer_lock();
        return VNET_API_ERROR_VALUE_EXIST;
      }

      hash_set_mem (lbm->vip_index_by_nodeport, &key, vip - lbm->vips);

      /* receive packets destined to NodeIP:NodePort */
      udp_register_dst_port (vm, args.port, lb4_nodeport_node.index, 1);
      udp_register_dst_port (vm, args.port, lb6_nodeport_node.index, 0);
    }

  *vip_index = vip - lbm->vips;
  //Create per-port vip filtering table
  if (args.port != 0)
    {
      lb_vip_add_port_filter(lbm, vip, vip_prefix_index, *vip_index);
      vip->vip_prefix_index = vip_prefix_index;
    }

  lb_put_writer_lock();
  return 0;
}

int lb_vip_del(u32 vip_index)
{
  lb_main_t *lbm = &lb_main;
  lb_vip_t *vip;

  /* Does not remove default vip, i.e. vip_index = 0 */
  if (vip_index == 0)
    return 0;

  lb_get_writer_lock();
  if (!(vip = lb_vip_get_by_index(vip_index))) {
    lb_put_writer_lock();
    return VNET_API_ERROR_NO_SUCH_ENTRY;
  }

  //FIXME: This operation is actually not working
  //We will need to remove state before performing this.

  {
    //Remove all ASs
    ip46_address_t *ass = 0;
    lb_as_t *as;
    u32 *as_index;
    pool_foreach(as_index, vip->as_indexes, {
        as = &lbm->ass[*as_index];
        vec_add1(ass, as->address);
    });
    if (vec_len(ass))
      lb_vip_del_ass_withlock(vip_index, ass, vec_len(ass), as_index);
    vec_free(ass);
  }

  //Delete adjacency
  lb_vip_del_adjacency(lbm, vip);

  //Delete per-port vip filtering entry
  if (vip->port != 0)
    {
      lb_vip_del_port_filter(lbm, vip);
    }

  //Set the VIP as unused
  vip->flags &= ~LB_VIP_FLAGS_USED;

  lb_put_writer_lock();
  return 0;
}

/* *INDENT-OFF* */
VLIB_PLUGIN_REGISTER () = {
    .version = VPP_BUILD_VER,
    .description = "Load Balancer",
};
/* *INDENT-ON* */

u8 *format_lb_dpo (u8 * s, va_list * va)
{
  index_t index = va_arg (*va, index_t);
  CLIB_UNUSED(u32 indent) = va_arg (*va, u32);
  lb_main_t *lbm = &lb_main;
  lb_vip_t *vip = pool_elt_at_index (lbm->vips, index);
  return format (s, "%U", format_lb_vip, vip);
}

static void lb_dpo_lock (dpo_id_t *dpo) {}
static void lb_dpo_unlock (dpo_id_t *dpo) {}

static fib_node_t *
lb_fib_node_get_node (fib_node_index_t index)
{
  lb_main_t *lbm = &lb_main;
  lb_as_t *as = pool_elt_at_index (lbm->ass, index);
  return (&as->fib_node);
}

static void
lb_fib_node_last_lock_gone (fib_node_t *node)
{
}

static lb_as_t *
lb_as_from_fib_node (fib_node_t *node)
{
  return ((lb_as_t*)(((char*)node) -
      STRUCT_OFFSET_OF(lb_as_t, fib_node)));
}

static void
lb_as_stack (lb_as_t *as)
{
  lb_main_t *lbm = &lb_main;
  lb_vip_t *vip = &lbm->vips[as->vip_index];
  dpo_type_t dpo_type = 0;

  if (lb_vip_is_gre4(vip))
    dpo_type = lbm->dpo_gre4_type;
  else if (lb_vip_is_gre6(vip))
    dpo_type = lbm->dpo_gre6_type;
  else if (lb_vip_is_gre4_port(vip))
    dpo_type = lbm->dpo_gre4_port_type;
  else if (lb_vip_is_gre6_port(vip))
    dpo_type = lbm->dpo_gre6_port_type;
  else if (lb_vip_is_l3dsr(vip))
    dpo_type = lbm->dpo_l3dsr_type;
  else if (lb_vip_is_l3dsr_port(vip))
    dpo_type = lbm->dpo_l3dsr_port_type;
  else if(lb_vip_is_nat4_port(vip))
    dpo_type = lbm->dpo_nat4_port_type;
  else if (lb_vip_is_nat6_port(vip))
    dpo_type = lbm->dpo_nat6_port_type;

  dpo_stack(dpo_type,
            lb_vip_is_ip4(vip->type)?DPO_PROTO_IP4:DPO_PROTO_IP6,
            &as->dpo,
            fib_entry_contribute_ip_forwarding(
                as->next_hop_fib_entry_index));
}

static fib_node_back_walk_rc_t
lb_fib_node_back_walk_notify (fib_node_t *node,
                 fib_node_back_walk_ctx_t *ctx)
{
    lb_as_stack(lb_as_from_fib_node(node));
    return (FIB_NODE_BACK_WALK_CONTINUE);
}

int lb_nat4_interface_add_del (u32 sw_if_index, int is_del)
{
  if (is_del)
    {
      vnet_feature_enable_disable ("ip4-unicast", "lb-nat4-in2out",
                                   sw_if_index, 0, 0, 0);
    }
  else
    {
      vnet_feature_enable_disable ("ip4-unicast", "lb-nat4-in2out",
                                   sw_if_index, 1, 0, 0);
    }

  return 0;
}

int lb_nat6_interface_add_del (u32 sw_if_index, int is_del)
{
  if (is_del)
    {
      vnet_feature_enable_disable ("ip6-unicast", "lb-nat6-in2out",
                                   sw_if_index, 0, 0, 0);
    }
  else
    {
      vnet_feature_enable_disable ("ip6-unicast", "lb-nat6-in2out",
                                   sw_if_index, 1, 0, 0);
    }

  return 0;
}

clib_error_t *
lb_init (vlib_main_t * vm)
{
  vlib_thread_main_t *tm = vlib_get_thread_main ();
  lb_main_t *lbm = &lb_main;
  lbm->vnet_main = vnet_get_main ();
  lbm->vlib_main = vm;

  lb_vip_t *default_vip;
  lb_as_t *default_as;
  fib_node_vft_t lb_fib_node_vft = {
      .fnv_get = lb_fib_node_get_node,
      .fnv_last_lock = lb_fib_node_last_lock_gone,
      .fnv_back_walk = lb_fib_node_back_walk_notify,
  };
  dpo_vft_t lb_vft = {
      .dv_lock = lb_dpo_lock,
      .dv_unlock = lb_dpo_unlock,
      .dv_format = format_lb_dpo,
  };

  //Allocate and init default VIP.
  lbm->vips = 0;
  pool_get(lbm->vips, default_vip);
  default_vip->prefix.ip6.as_u64[0] = 0xffffffffffffffffL;
  default_vip->prefix.ip6.as_u64[1] = 0xffffffffffffffffL;
  default_vip->protocol = ~0;
  default_vip->port = 0;
  default_vip->flags = LB_VIP_FLAGS_USED;

  lbm->per_cpu = 0;
  vec_validate(lbm->per_cpu, tm->n_vlib_mains - 1);
  lbm->writer_lock = clib_mem_alloc_aligned (CLIB_CACHE_LINE_BYTES,  CLIB_CACHE_LINE_BYTES);
  lbm->writer_lock[0] = 0;
  lbm->per_cpu_sticky_buckets = LB_DEFAULT_PER_CPU_STICKY_BUCKETS;
  lbm->flow_timeout = LB_DEFAULT_FLOW_TIMEOUT;
  lbm->ip4_src_address.as_u32 = 0xffffffff;
  lbm->ip6_src_address.as_u64[0] = 0xffffffffffffffffL;
  lbm->ip6_src_address.as_u64[1] = 0xffffffffffffffffL;
  lbm->dpo_gre4_type = dpo_register_new_type(&lb_vft, lb_dpo_gre4_nodes);
  lbm->dpo_gre6_type = dpo_register_new_type(&lb_vft, lb_dpo_gre6_nodes);
  lbm->dpo_gre4_port_type = dpo_register_new_type(&lb_vft,
                                                  lb_dpo_gre4_port_nodes);
  lbm->dpo_gre6_port_type = dpo_register_new_type(&lb_vft,
                                                  lb_dpo_gre6_port_nodes);
  lbm->dpo_l3dsr_type = dpo_register_new_type(&lb_vft,
                                              lb_dpo_l3dsr_nodes);
  lbm->dpo_l3dsr_port_type = dpo_register_new_type(&lb_vft,
                                                   lb_dpo_l3dsr_port_nodes);
  lbm->dpo_nat4_port_type = dpo_register_new_type(&lb_vft,
                                                  lb_dpo_nat4_port_nodes);
  lbm->dpo_nat6_port_type = dpo_register_new_type(&lb_vft,
                                                  lb_dpo_nat6_port_nodes);
  lbm->fib_node_type = fib_node_register_new_type(&lb_fib_node_vft);

  //Init AS reference counters
  vlib_refcount_init(&lbm->as_refcount);

  //Allocate and init default AS.
  lbm->ass = 0;
  pool_get(lbm->ass, default_as);
  default_as->flags = 0;
  default_as->dpo.dpoi_next_node = LB_NEXT_DROP;
  default_as->vip_index = ~0;
  default_as->address.ip6.as_u64[0] = 0xffffffffffffffffL;
  default_as->address.ip6.as_u64[1] = 0xffffffffffffffffL;

  lbm->vip_index_by_nodeport
    = hash_create_mem (0, sizeof(u16), sizeof (uword));

  clib_bihash_init_8_8 (&lbm->vip_index_per_port,
                        "vip_index_per_port", LB_VIP_PER_PORT_BUCKETS,
                        LB_VIP_PER_PORT_MEMORY_SIZE);

  clib_bihash_init_8_8 (&lbm->mapping_by_as4,
                        "mapping_by_as4", LB_MAPPING_BUCKETS,
                        LB_MAPPING_MEMORY_SIZE);

  clib_bihash_init_24_8 (&lbm->mapping_by_as6,
                        "mapping_by_as6", LB_MAPPING_BUCKETS,
                        LB_MAPPING_MEMORY_SIZE);

#define _(a,b,c) lbm->vip_counters[c].name = b;
  lb_foreach_vip_counter
#undef _
  return NULL;
}

VLIB_INIT_FUNCTION (lb_init);