summaryrefslogtreecommitdiffstats
path: root/src/vnet/ip/ip.api
blob: 4b418bea71fdce1b88f711f7643582d9fab37348 (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
/* Hey Emacs use -*- mode: C -*- */
/*
 * Copyright (c) 2018 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.
 */

/** \file

    This file defines vpp IP control-plane API messages which are generally
    called through a shared memory interface. 
*/

option version = "1.4.0";
import "vnet/ip/ip_types.api";
import "vnet/fib/fib_types.api";
import "vnet/ethernet/ethernet_types.api";

/** \brief Add / del table request
           A table can be added multiple times, but need be deleted only once.
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param is_ipv6 - V4 or V6 table
    @param table_id - table ID associated with the route
                      This table ID will apply to both the unicats
		      and mlticast FIBs
    @param name - A client provided name/tag for the table. If this is
                  not set by the client, then VPP will generate something
		  meaningfull.
*/
autoreply define ip_table_add_del
{
  u32 client_index;
  u32 context;
  u32 table_id;
  u8 is_ipv6;
  u8 is_add;
  u8 name[64];
};

/** \brief Dump IP fib table
    @param client_index - opaque cookie to identify the sender
*/
define ip_fib_dump
{
  u32 client_index;
  u32 context;
};

/** \brief IP FIB table response
    @param table_id - IP fib table id
    @address_length - mask length
    @address - ip4 prefix
    @param count - the number of fib_path in path
    @param path  - array of of fib_path structures
*/
manual_endian manual_print define ip_fib_details
{
  u32 context;
  u32 table_id;
  u8  table_name[64];
  u8  address_length;
  u8  address[4];
  u32 count;
  u32 stats_index;
  vl_api_fib_path_t path[count];
};

/** \brief Dump IP6 fib table
    @param client_index - opaque cookie to identify the sender
*/
define ip6_fib_dump
{
  u32 client_index;
  u32 context;
};

/** \brief IP6 FIB table entry response
    @param table_id - IP6 fib table id
    @param address_length - mask length
    @param address - ip6 prefix
    @param count - the number of fib_path in path
    @param path  - array of of fib_path structures
*/
manual_endian manual_print define ip6_fib_details
{
  u32 context;
  u32 table_id;
  u8  table_name[64];
  u8  address_length;
  u8  address[16];
  u32 count;
  u32 stats_index;
  vl_api_fib_path_t path[count];
};

/** \brief Dump IP neighboors
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param sw_if_index - the interface to dump neighboors, ~0 == all
    @param is_ipv6 - [1|0] to indicate if address family is ipv[6|4]
*/
define ip_neighbor_dump
{
    u32 client_index;
    u32 context;
    u32 sw_if_index;
    u8  is_ipv6;
};

/** \brief IP neighboors dump response
    @param context - sender context which was passed in the request
    @param sw_if_index - The interface used to reach the neighbor
    @param stats_index - An index in the stats segment that can be used to read
                         the counters for this neighbour.
    @param is_static - [1|0] to indicate if neighbor is statically configured
    @param is_ipv6 - [1|0] to indicate if address family is ipv[6|4]
*/
define ip_neighbor_details {
    u32 context;
    u32 sw_if_index;
    u32 stats_index;
    u8  is_static;
    u8  is_ipv6;
    u8  mac_address[6];
    u8  ip_address[16];
};

/** \brief IP neighbor add / del request
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param sw_if_index - interface used to reach neighbor
    @param is_add - 1 to add neighbor, 0 to delete
    @param is_ipv6 - 1 for IPv6 neighbor, 0 for IPv4
    @param is_ipv6 - 1 for IPv6 neighbor, 0 for IPv4
    @param is_static - A static neighbor Entry - there are not flushed
                       If the interface goes down.
    @param is_no_adj_fib - Do not create a corresponding entry in the FIB
                           table for the neighbor.
    @param mac_address - l2 address of the neighbor
    @param dst_address - ip4 or ip6 address of the neighbor
*/
define ip_neighbor_add_del
{
  u32 client_index;
  u32 context;
  u32 sw_if_index;
  /* 1 = add, 0 = delete */
  u8 is_add;
  u8 is_ipv6;
  u8 is_static;
  u8 is_no_adj_fib;
  u8 mac_address[6];
  u8 dst_address[16];
};

define ip_neighbor_add_del_reply
{
  u32 context;
  i32 retval;
  u32 stats_index;
};

/** \brief Set the ip flow hash config for a fib request
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param vrf_id - vrf/fib id
    @param is_ipv6 - if non-zero the fib is ip6, else ip4
    @param src - if non-zero include src in flow hash
    @param dst - if non-zero include dst in flow hash
    @param sport - if non-zero include sport in flow hash
    @param dport - if non-zero include dport in flow hash
    @param proto -if non-zero include proto in flow hash
    @param reverse - if non-zero include reverse in flow hash
    @param symmetric - if non-zero include symmetry in flow hash
*/
autoreply define set_ip_flow_hash
{
  u32 client_index;
  u32 context;
  u32 vrf_id;
  u8 is_ipv6;
  u8 src;
  u8 dst;
  u8 sport;
  u8 dport;
  u8 proto;
  u8 reverse;
  u8 symmetric;
};

/** \brief IPv6 router advertisement config request
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param suppress -
    @param managed -
    @param other -
    @param ll_option -
    @param send_unicast -
    @param cease -
    @param is_no -
    @param default_router -
    @param max_interval -
    @param min_interval -
    @param lifetime -
    @param initial_count -
    @param initial_interval -
*/
autoreply define sw_interface_ip6nd_ra_config
{
  u32 client_index;
  u32 context;
  u32 sw_if_index;
  u8 suppress;
  u8 managed;
  u8 other;
  u8 ll_option;
  u8 send_unicast;
  u8 cease;
  u8 is_no;
  u8 default_router;
  u32 max_interval;
  u32 min_interval;
  u32 lifetime;
  u32 initial_count;
  u32 initial_interval;
};

/** \brief IPv6 router advertisement prefix config request
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param sw_if_index - The interface the RA prefix information is for
    @param address[] - The prefix to advertise
    @param address_length - the prefix length
    @param use_default - Revert to default settings
    @param no_advertise - Do not advertise this prefix
    @param off_link - The prefix is off link (it is not configured on the interface)
                      Configures the L-flag, When set, indicates that this
		      prefix can be used for on-link determination.
    @param no_autoconfig - Setting for the A-flag. When
                           set indicates that this prefix can be used for
                          stateless address configuration.
    @param no_onlink - The prefix is not on link. Make sure this is consistent
                       with the off_link parameter else YMMV
    @param is_no - add/delete
    @param val_lifetime - The length of time in
                     seconds (relative to the time the packet is sent)
                     that the prefix is valid for the purpose of on-link
                     determination.  A value of all one bits
                     (0xffffffff) represents infinity
    @param pref_lifetime - The length of time in
                     seconds (relative to the time the packet is sent)
                     that addresses generated from the prefix via
                     stateless address autoconfiguration remain
                     preferred [ADDRCONF].  A value of all one bits
                     (0xffffffff) represents infinity.
*/
autoreply define sw_interface_ip6nd_ra_prefix
{
  u32 client_index;
  u32 context;
  u32 sw_if_index;
  u8 address[16];
  u8 address_length;
  u8 use_default;
  u8 no_advertise;
  u8 off_link;
  u8 no_autoconfig;
  u8 no_onlink;
  u8 is_no;
  u32 val_lifetime;
  u32 pref_lifetime;
};

/** \brief IPv6 ND proxy config
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param sw_if_index - The interface the host is on
    @param address - The address of the host for which to proxy for
    @param is_add - Adding or deleting
*/
autoreply define ip6nd_proxy_add_del
{
  u32 client_index;
  u32 context;
  u32 sw_if_index;
  u8 is_del;
  u8 address[16];
};

/** \brief IPv6 ND proxy details returned after request
    @param context - sender context, to match reply w/ request
    @param retval - return code for the request
*/
define ip6nd_proxy_details
{
  u32 context;
  u32 sw_if_index;
  u8 address[16];
};

/** \brief IPv6 ND proxy dump request
    @param context - sender context, to match reply w/ request
    @param retval - return code for the request
    @param sw_if_index - The interface the host is on
    @param address - The address of the host for which to proxy for
*/
define ip6nd_proxy_dump
{
  u32 client_index;
  u32 context;
};

/** \brief Start / stop sending router solicitation
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param irt - initial retransmission time
    @param mrt - maximum retransmission time
    @param mrc - maximum retransmission count
    @param mrd - maximum retransmission duration
    @param sw_if_index - software interface index of interface
                         for sending router solicitation
    @param stop - if non-zero then stop sending router solicitation,
                  otherwise start sending router solicitation
*/
autoreply define ip6nd_send_router_solicitation
{
  u32 client_index;
  u32 context;
  u32 irt;
  u32 mrt;
  u32 mrc;
  u32 mrd;
  u32 sw_if_index;
  u8 stop;
};

/** \brief IPv6 interface enable / disable request
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param sw_if_index - interface used to reach neighbor
    @param enable - if non-zero enable ip6 on interface, else disable
*/
autoreply define sw_interface_ip6_enable_disable
{
  u32 client_index;
  u32 context;
  u32 sw_if_index;
  u8 enable;			/* set to true if enable */
};

/** \brief Add / del route request
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param sw_if_index - software index of the new vlan's parent interface
    @param vrf_id - fib table /vrf associated with the route
    @param lookup_in_vrf - 
    @param classify_table_index - 
    @param is_add - 1 if adding the route, 0 if deleting
    @param is_drop - Drop the packet
    @param is_unreach - Drop the packet and rate limit send ICMP unreachable
    @param is_prohibit - Drop the packet and rate limit send ICMP prohibited
    @param is_ipv6 - 0 if an ip4 route, else ip6
    @param is_local - The route will result in packets sent to VPP IP stack
    @param is_udp_encap - The path describes a UDP-o-IP encapsulation.
    @param is_classify - 
    @param is_multipath - Set to 1 if this is a multipath route, else 0
    @param is_dvr - Does the route resolve via a DVR interface.
    @param is_source_lookup - The the path is a deaggregate path (i.e. a lookup
    	   		      in another table) is the lookup on the packet's
			      source address or destination.
    @param next_hop_weight - Weight for Unequal cost multi-path
    @param next_hop_preference - Path that are up that have the best preference are
                                 are used for forwarding. lower value is better.
    @param next_hop_id - Used when the path resolves via an object that has a unique
                         identifier.
    @param dst_address_length - 
    @param dst_address[16] - 
    @param next_hop_address[16] - 
    @param next_hop_n_out_labels - the number of labels in the label stack
    @param next_hop_out_label_stack - the next-hop output label stack, outer most first
    @param next_hop_via_label - The next-hop is a resolved via a local label
*/
define ip_add_del_route
{
  u32 client_index;
  u32 context;
  u32 next_hop_sw_if_index;
  u32 table_id;
  u32 classify_table_index;
  u32 next_hop_table_id;
  u32 next_hop_id;
  u8 is_add;
  u8 is_drop;
  u8 is_unreach;
  u8 is_prohibit;
  u8 is_ipv6;
  u8 is_local;
  u8 is_classify;
  u8 is_multipath;
  u8 is_resolve_host;
  u8 is_resolve_attached;
  u8 is_dvr;
  u8 is_source_lookup;
  u8 is_udp_encap;
  u8 next_hop_weight;
  u8 next_hop_preference;
  u8 next_hop_proto;
  u8 dst_address_length;
  u8 dst_address[16];
  u8 next_hop_address[16];
  u8 next_hop_n_out_labels;
  u32 next_hop_via_label;
  vl_api_fib_mpls_label_t next_hop_out_label_stack[next_hop_n_out_labels];
};

define ip_add_del_route_reply
{
  u32 context;
  i32 retval;
  u32 stats_index;
};

/** \brief Add / del route request

    Adds a route, consisting both of the MFIB entry to match packets
    (which may already exist) and a path to send those packets down.
    Routes can be entered repeatedly to add multiple paths.  Deletions are
    per-path.

    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param table_id - fib table /vrf associated with the route
    @param is_add - true if adding a route; false if deleting one
    @param is_ipv6 - true iff all the addresses are v6
    @param entry_flags - see fib_entry_flag_t
    @param itf_flags - see mfib_entry_flags_t
    @param next_hop_afi - see dpo_proto_t; the type of destination description
    @param src_address - the source of the packet
    @param grp_address - the group the packet is destined to
    @param nh_address - the nexthop to forward the packet to
    @param next_hop_sw_if_index - interface to emit packet on

    BIER AFIs use the BIER imposition ID.  v4 and v6 AFIs use either the
    interface or the nexthop address.

    Note that if the route is source-specific (S is supplied, not all 0s),
    the prefix match is treated as exact (prefixlen /32 or /128).

    FIXME not complete yet
*/
define ip_mroute_add_del
{
  u32 client_index;
  u32 context;
  u32 next_hop_sw_if_index;
  u32 table_id;
  u32 entry_flags;
  u32 itf_flags;
  u32 rpf_id;
  u32 bier_imp;
  u16 grp_address_length;
  u8 next_hop_afi;
  u8 is_add;
  u8 is_ipv6;
  u8 is_local;
  u8 grp_address[16];
  u8 src_address[16];
  u8 nh_address[16];
};

define ip_mroute_add_del_reply
{
  u32 context;
  i32 retval;
  u32 stats_index;
};

/** \brief Dump IP multicast fib table
    @param client_index - opaque cookie to identify the sender
*/
define ip_mfib_dump
{
  u32 client_index;
  u32 context;
};

/** \brief IP Multicast FIB table response
    @param table_id - IP fib table id
    @address_length - mask length
    @grp_address - Group address/prefix
    @src_address - Source address
    @param count - the number of fib_path in path
    @param path  - array of of fib_path structures
*/
manual_endian manual_print define ip_mfib_details
{
  u32 context;
  u32 table_id;
  u32 entry_flags;
  u32 rpf_id;
  u8  address_length;
  u8  grp_address[4];
  u8  src_address[4];
  u32 count;
  u32 stats_index;
  vl_api_fib_path_t path[count];
};

/** \brief Dump IP6 multicast fib table
    @param client_index - opaque cookie to identify the sender
*/
define ip6_mfib_dump
{
  u32 client_index;
  u32 context;
};

/** \brief IP6 Multicast FIB table response
    @param table_id - IP fib table id
    @address_length - mask length
    @grp_address - Group address/prefix
    @src_address - Source address
    @param count - the number of fib_path in path
    @param path  - array of of fib_path structures
*/
manual_endian manual_print define ip6_mfib_details
{
  u32 context;
  u32 table_id;
  u8  address_length;
  u8  grp_address[16];
  u8  src_address[16];
  u32 count;
  vl_api_fib_path_t path[count];
};

define ip_address_details
{
  u32 context;
  u8 ip[16];
  u8 prefix_length;
  u32 sw_if_index;
  u8 is_ipv6;
};

define ip_address_dump
{
  u32 client_index;
  u32 context;
  u32 sw_if_index;
  u8 is_ipv6;
};

/** \brief IP unnumbered configurations
    @param sw_if_index The interface that has unnumbered configuration
    @param ip_sw_if_index The IP interface that it is unnnumbered to
*/
define ip_unnumbered_details
{
  u32 context;
  u32 sw_if_index;
  u32 ip_sw_if_index;
};

/** \brief Dump IP unnumbered configurations
    @param sw_if_index ~0 for all interfaces, else the interface desired
*/
define ip_unnumbered_dump
{
  u32 client_index;
  u32 context;
  u32 sw_if_index;
};

define ip_details
{
  u32 context;
  u32 sw_if_index;
  u8 is_ipv6;
};

define ip_dump
{
  u32 client_index;
  u32 context;
  u8 is_ipv6;
};

define mfib_signal_dump
{
  u32 client_index;
  u32 context;
};

define mfib_signal_details
{
  u32 context;
  u32 sw_if_index;
  u32 table_id;
  u16 grp_address_len;
  u8 grp_address[16];
  u8 src_address[16];
  u16 ip_packet_len;
  u8 ip_packet_data[256];
};

/** \brief IP punt policer
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param is_add - 1 to add neighbor, 0 to delete
    @param is_ipv6 - 1 for IPv6 neighbor, 0 for IPv4
    @param policer_index - Index of policer to use
*/
autoreply define ip_punt_police
{
  u32 client_index;
  u32 context;
  u32 policer_index;
  u8 is_add;
  u8 is_ip6;
};

/** \brief Punt redirect type
    @param rx_sw_if_index - specify the original RX interface of traffic
                            that should be redirected. ~0 means any interface.
    @param tx_sw_if_index - the TX interface to which traffic shoulde be
                            redirected.
    @param nh - the next-hop to redirect the traffic to.
    @param is_ipv6 - 1 for IPv6 neighbor, 0 for IPv4
*/
typeonly define punt_redirect
{
  u32 rx_sw_if_index;
  u32 tx_sw_if_index;
  vl_api_address_t nh;
};

/** \brief IP punt redirect
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param punt - punt definition
    @param is_add - 1 to add neighbor, 0 to delete
*/
autoreply define ip_punt_redirect
{
  u32 client_index;
  u32 context;
  vl_api_punt_redirect_t punt;
  u8 is_add;
};

define ip_punt_redirect_dump
{
  u32 client_index;
  u32 context;
  u32 sw_if_index;
  u8 is_ipv6;
};

define ip_punt_redirect_details
{
  u32 context;
  vl_api_punt_redirect_t punt;
};

autoreply define ip_container_proxy_add_del
{
  u32 client_index;
  u32 context;
  u8 ip[16];
  u8 is_ip4;
  u8 plen;
  u32 sw_if_index;
  u8 is_add;
};

define ip_container_proxy_dump
{
  u32 client_index;
  u32 context;
};

define ip_container_proxy_details
{
  u32 context;
  u32 sw_if_index;
  vl_api_prefix_t prefix;
};

/** \brief Configure IP source and L4 port-range check
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param is_ip6 - 1 if source address type is IPv6
    @param is_add - 1 if add, 0 if delete
    @param mask_length - mask length for address entry
    @param address - array of address bytes
    @param number_of_ranges - length of low_port and high_port arrays (must match)
    @param low_ports[32] - up to 32 low end of port range entries (must have corresponding high_ports entry)
    @param high_ports[32] - up to 32 high end of port range entries (must have corresponding low_ports entry)
    @param vrf_id - fib table/vrf id to associate the source and port-range check with
    @note To specify a single port set low_port and high_port entry the same
*/
autoreply define ip_source_and_port_range_check_add_del
{
  u32 client_index;
  u32 context;
  u8 is_ipv6;
  u8 is_add;
  u8 mask_length;
  u8 address[16];
  u8 number_of_ranges;
  u16 low_ports[32];
  u16 high_ports[32];
  u32 vrf_id;
};

/** \brief Set interface source and L4 port-range request
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param interface_id - interface index
    @param tcp_vrf_id - VRF associated with source and TCP port-range check
    @param udp_vrf_id - VRF associated with source and TCP port-range check
*/
autoreply define ip_source_and_port_range_check_interface_add_del
{
  u32 client_index;
  u32 context;
  u8 is_add;
  u32 sw_if_index;
  u32 tcp_in_vrf_id;
  u32 tcp_out_vrf_id;
  u32 udp_in_vrf_id;
  u32 udp_out_vrf_id;
};

/** \brief Set interface source check request
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param is_add - add or del
    @param loose - strict or loose
    @param sw_if_index - interface index
*/
autoreply define ip_source_check_interface_add_del
{
  u32 client_index;
  u32 context;
  u8 is_add;
  u8 loose;
  u32 sw_if_index;
};

/** \brief Enable/disable periodic IP neighbor scan
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param mode - 0: disable, 1: IPv4, 2: IPv6, 3: both IPv4/v6
    @param scan_interval - neighbor scan interval in minutes, 0: default to 1
    @param max_proc_time - max processing time per run in usec, 0: default to 20
    @param max_update - max neighbor probe/delete per run, 0: default to 10
    @param scan_int_delay - delay in msec to resume scan if exceed max proc
                            time or update, 0: default to 1
    @param stale_threshold - threshold in minutes for neighbor deletion, 
                             0: default to 4*scan_interval
*/
autoreply define ip_scan_neighbor_enable_disable
{
  u32 client_index;
  u32 context;
  u8 mode;
  u8 scan_interval;
  u8 max_proc_time;
  u8 max_update;
  u8 scan_int_delay;
  u8 stale_threshold;
};

/** \brief IP probe neighbor address on an interface by sending an
           ARP request (for IP4) or ICMP6 Neighbor Solicitation (for IP6)
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param sw_if_index - interface index
    @param dst_address - target IP address to send IP addr resolution request
    @param is_ipv6 - [1|0] to indicate if address family is IPv[6|4]
*/
autoreply define ip_probe_neighbor
{
  u32 client_index;
  u32 context;
  u32 sw_if_index;
  u8 dst_address[16];
  u8 is_ipv6;
};

/** \brief Register for IP4 ARP resolution event on receing ARP reply or
           MAC/IP info from ARP requests in L2 BDs
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param enable_disable - 1 => register for events, 0 => cancel registration
    @param pid - sender's pid
    @param address - exact IP4 address of interested arp resolution event, or
                     0 to get MAC/IP info from ARP requests in BDs
*/
autoreply define want_ip4_arp_events
{
  u32 client_index;
  u32 context;
  u8 enable_disable;
  u32 pid;
  u32 address;
};

/** \brief Tell client about an IP4 ARP resolution event or
           MAC/IP info from ARP requests in L2 BDs
    @param client_index - opaque cookie to identify the sender
    @param address - the exact ip4 address of interest
    @param pid - client pid registered to receive notification
    @param sw_if_index - interface which received ARP packet
    @param new_mac - the new mac address 
    @param mac_ip - 0: ARP resolution event, 1: MAC/IP info from L2 BDs
*/
define ip4_arp_event
{
  u32 client_index;
  u32 address;
  u32 pid;
  u32 sw_if_index;
  u8 new_mac[6];
  u8 mac_ip;
};

service {
  rpc want_ip4_arp_events returns want_ip4_arp_events_reply
    events ip4_arp_event;
};

/** \brief Register for IP6 ND resolution event on recieving NA reply
           MAC/IP info from ICMP6 Neighbor Solicitation in L2 BDs
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param enable_disable - 1 => register for events, 0 => cancel registration
    @param pid - sender's pid
    @param address - the exact IP6 address of interested ND resolution event, or
                     0 to get MAC/IP info from ICMP6 NS in L2 BDs.
*/
autoreply define want_ip6_nd_events
{
  u32 client_index;
  u32 context;
  u8 enable_disable;
  u32 pid;
  u8 address[16];
};

/** \brief Tell client about an IP6 ND resolution or
           MAC/IP info from ICMP6 Neighbor Solicitation in L2 BDs.
    @param client_index - opaque cookie to identify the sender
    @param pid - client pid registered to receive notification
    @param sw_if_index - interface which received ARP packet
    @param address - the exact ip6 address of interest
    @param new_mac - the new mac address 
    @param mac_ip - 0: ND resolution event, 1: MAC/IP info from L2 BDs
*/
define ip6_nd_event
{
  u32 client_index;
  u32 pid;
  u32 sw_if_index;
  u8 address[16];
  u8 new_mac[6];
  u8 mac_ip;
};

service {
  rpc want_ip6_ra_events returns want_ip6_ra_events_reply
    events ip6_ra_event;
};

/** \brief Register for ip6 router advertisement events
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param enable_disable - 1 => register for events, 0 => cancel registration
    @param pid - sender's pid
*/
autoreply define want_ip6_ra_events
{
  u32 client_index;
  u32 context;
  u8 enable_disable;
  u32 pid;
};

/** \brief Struct representing RA prefix info
    @param dst_address - RA prefix info destination address
    @param dst_address_length - RA prefix info destination address length
    @param flags - RA prefix info flags
    @param valid_time - RA prefix info valid time
    @param preferred_time - RA prefix info preferred time
*/
typeonly define ip6_ra_prefix_info
{
  u8 dst_address[16];
  u8 dst_address_length;
  u8 flags;
  u32 valid_time;
  u32 preferred_time;
};

/** \brief Tell client about a router advertisement event
    @param client_index - opaque cookie to identify the sender
    @param pid - client pid registered to receive notification
    @param current_hop_limit - RA current hop limit
    @param flags - RA flags
    @param router_lifetime_in_sec - RA lifetime in seconds
    @param neighbor_reachable_time_in_msec - RA neighbor reachable time in msec
    @param time_in_msec_between_retransmitted_neighbor_solicitations -
               time in msec between retransmitted neighbor solicitations
    @param n_prefixes -
    @param prefixes -
*/
define ip6_ra_event
{
  u32 client_index;
  u32 pid;
  u32 sw_if_index;
  u8 router_address[16];
  u8 current_hop_limit;
  u8 flags;
  u16 router_lifetime_in_sec;
  u32 neighbor_reachable_time_in_msec;
  u32 time_in_msec_between_retransmitted_neighbor_solicitations;
  u32 n_prefixes;
  vl_api_ip6_ra_prefix_info_t prefixes[n_prefixes];
};

service {
  rpc want_ip6_nd_events returns want_ip6_nd_events_reply
    events ip6_nd_event;
};

/** \brief Proxy ARP configuration type
    @param vrf_id - VRF / Fib table ID
    @param low_address[4] - Low address of the Proxy ARP range
    @param hi_address[4] - High address of the Proxy ARP range
*/
typeonly define proxy_arp
{
  u32 vrf_id;
  u8 low_address[4];
  u8 hi_address[4];
};

/** \brief Proxy ARP add / del request
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param is_add - 1 if adding the Proxy ARP range, 0 if deleting
    @param proxy - Proxy configuration
*/
autoreply define proxy_arp_add_del
{
  u32 client_index;
  u32 context;
  u8 is_add;
  vl_api_proxy_arp_t proxy;
};

/** \brief Proxy ARP dump request
 */
define proxy_arp_dump
{
  u32 client_index;
  u32 context;
};

/** \brief Proxy ARP dump details reply
 * @param proxy - Same data as used to configure
 */
define proxy_arp_details
{
  u32 context;
  vl_api_proxy_arp_t proxy;
};

/** \brief Proxy ARP add / del interface request
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param sw_if_index - Which interface to enable / disable Proxy Arp on
    @param enable_disable - 1 to enable Proxy ARP on interface, 0 to disable
*/
autoreply define proxy_arp_intfc_enable_disable
{
  u32 client_index;
  u32 context;
  u32 sw_if_index;
  /* 1 = on, 0 = off */
  u8 enable_disable;
};

/** \brief Proxy ARP interface dump request
 */
define proxy_arp_intfc_dump
{
  u32 client_index;
  u32 context;
};

/** \brief Proxy ARP interface dump details reply
 * @param sw_if_index The interface on which ARP proxy is enabled.
 */
define proxy_arp_intfc_details
{
  u32 context;
  u32 sw_if_index;
};

/** \brief Reset fib table request
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param vrf_id - vrf/table id of the fib table to reset
    @param is_ipv6 - an ipv6 fib to reset if non-zero, else ipv4
*/
autoreply define reset_fib
{
  u32 client_index;
  u32 context;
  u32 vrf_id;
  u8 is_ipv6;
};

/** \brief Set max allowed ARP or ip6 neighbor entries request
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param is_ipv6 - neighbor limit if non-zero, else ARP limit
    @param arp_neighbor_limit - the new limit, defaults are ~ 50k
*/
autoreply define set_arp_neighbor_limit
{
  u32 client_index;
  u32 context;
  u8 is_ipv6;
  u32 arp_neighbor_limit;
};

/** \brief IOAM enable : Enable in-band OAM
    @param id - profile id
    @param seqno - To enable Seqno Processing
    @param analyse - Enabling analysis of iOAM at decap node 
    @param pow_enable - Proof of Work enabled or not flag
    @param trace_enable - iOAM Trace enabled or not flag
*/
autoreply define ioam_enable
{
  u32 client_index;
  u32 context;
  u16 id;
  u8 seqno;
  u8 analyse;
  u8 pot_enable;
  u8 trace_enable;
  u32 node_id;
};

/** \brief iOAM disable
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param index - MAP Domain index
*/
autoreply define ioam_disable
{
  u32 client_index;
  u32 context;
  u16 id;
};

autoreply define ip_reassembly_set
{
  u32 client_index;
  u32 context;
  u32 timeout_ms;
  u32 max_reassemblies;
  u32 expire_walk_interval_ms;
  u8 is_ip6;
};

define ip_reassembly_get
{
  u32 client_index;
  u32 context;
  u8 is_ip6;
};

define ip_reassembly_get_reply
{
  u32 context;
  i32 retval;
  u32 timeout_ms;
  u32 max_reassemblies;
  u32 expire_walk_interval_ms;
  u8 is_ip6;
};

/** \brief Enable/disable reassembly feature
    @param client_index - opaque cookie to identify the sender
    @param context - sender context, to match reply w/ request
    @param sw_if_index - interface to enable/disable feature on
    @param enable_ip4 - enable ip4 reassembly if non-zero, disable if 0
    @param enable_ip6 - enable ip6 reassembly if non-zero, disable if 0
*/
autoreply define ip_reassembly_enable_disable
{
  u32 client_index;
  u32 context;
  u32 sw_if_index;
  u8 enable_ip4;
  u8 enable_ip6;
};

/*
 * Local Variables:
 * eval: (c-set-style "gnu")
 * End:
 */
href='#n2262'>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
/*
 * Copyright (c) 2015 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.
 */
/*
 * ip/ip4_forward.c: IP v4 forwarding
 *
 * Copyright (c) 2008 Eliot Dresselhaus
 *
 * Permission is hereby granted, free of charge, to any person obtaining
 * a copy of this software and associated documentation files (the
 * "Software"), to deal in the Software without restriction, including
 * without limitation the rights to use, copy, modify, merge, publish,
 * distribute, sublicense, and/or sell copies of the Software, and to
 * permit persons to whom the Software is furnished to do so, subject to
 * the following conditions:
 *
 * The above copyright notice and this permission notice shall be
 * included in all copies or substantial portions of the Software.
 *
 *  THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
 *  EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
 *  MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
 *  NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
 *  LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
 *  OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
 *  WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 */

#include <vnet/vnet.h>
#include <vnet/ip/ip.h>
#include <vnet/ip/ip_frag.h>
#include <vnet/ethernet/ethernet.h>	/* for ethernet_header_t */
#include <vnet/ethernet/arp_packet.h>	/* for ethernet_arp_header_t */
#include <vnet/ppp/ppp.h>
#include <vnet/srp/srp.h>	/* for srp_hw_interface_class */
#include <vnet/api_errno.h>	/* for API error numbers */
#include <vnet/fib/fib_table.h>	/* for FIB table and entry creation */
#include <vnet/fib/fib_entry.h>	/* for FIB table and entry creation */
#include <vnet/fib/fib_urpf_list.h>	/* for FIB uRPF check */
#include <vnet/fib/ip4_fib.h>
#include <vnet/dpo/load_balance.h>
#include <vnet/dpo/load_balance_map.h>
#include <vnet/dpo/classify_dpo.h>
#include <vnet/mfib/mfib_table.h>	/* for mFIB table and entry creation */

#include <vnet/ip/ip4_forward.h>
#include <vnet/interface_output.h>

/** @brief IPv4 lookup node.
    @node ip4-lookup

    This is the main IPv4 lookup dispatch node.

    @param vm vlib_main_t corresponding to the current thread
    @param node vlib_node_runtime_t
    @param frame vlib_frame_t whose contents should be dispatched

    @par Graph mechanics: buffer metadata, next index usage

    @em Uses:
    - <code>vnet_buffer(b)->sw_if_index[VLIB_RX]</code>
        - Indicates the @c sw_if_index value of the interface that the
	  packet was received on.
    - <code>vnet_buffer(b)->sw_if_index[VLIB_TX]</code>
        - When the value is @c ~0 then the node performs a longest prefix
          match (LPM) for the packet destination address in the FIB attached
          to the receive interface.
        - Otherwise perform LPM for the packet destination address in the
          indicated FIB. In this case <code>[VLIB_TX]</code> is a FIB index
          value (0, 1, ...) and not a VRF id.

    @em Sets:
    - <code>vnet_buffer(b)->ip.adj_index[VLIB_TX]</code>
        - The lookup result adjacency index.

    <em>Next Index:</em>
    - Dispatches the packet to the node index found in
      ip_adjacency_t @c adj->lookup_next_index
      (where @c adj is the lookup result adjacency).
*/
VLIB_NODE_FN (ip4_lookup_node) (vlib_main_t * vm, vlib_node_runtime_t * node,
				vlib_frame_t * frame)
{
  return ip4_lookup_inline (vm, node, frame,
			    /* lookup_for_responses_to_locally_received_packets */
			    0);

}

static u8 *format_ip4_lookup_trace (u8 * s, va_list * args);

/* *INDENT-OFF* */
VLIB_REGISTER_NODE (ip4_lookup_node) =
{
  .name = "ip4-lookup",
  .vector_size = sizeof (u32),
  .format_trace = format_ip4_lookup_trace,
  .n_next_nodes = IP_LOOKUP_N_NEXT,
  .next_nodes = IP4_LOOKUP_NEXT_NODES,
};
/* *INDENT-ON* */

VLIB_NODE_FN (ip4_load_balance_node) (vlib_main_t * vm,
				      vlib_node_runtime_t * node,
				      vlib_frame_t * frame)
{
  vlib_combined_counter_main_t *cm = &load_balance_main.lbm_via_counters;
  u32 n_left, *from;
  u32 thread_index = vm->thread_index;
  vlib_buffer_t *bufs[VLIB_FRAME_SIZE], **b = bufs;
  u16 nexts[VLIB_FRAME_SIZE], *next;

  from = vlib_frame_vector_args (frame);
  n_left = frame->n_vectors;
  next = nexts;

  vlib_get_buffers (vm, from, bufs, n_left);

  while (n_left >= 4)
    {
      const load_balance_t *lb0, *lb1;
      const ip4_header_t *ip0, *ip1;
      u32 lbi0, hc0, lbi1, hc1;
      const dpo_id_t *dpo0, *dpo1;

      /* Prefetch next iteration. */
      {
	vlib_prefetch_buffer_header (b[2], LOAD);
	vlib_prefetch_buffer_header (b[3], LOAD);

	CLIB_PREFETCH (b[2]->data, sizeof (ip0[0]), LOAD);
	CLIB_PREFETCH (b[3]->data, sizeof (ip0[0]), LOAD);
      }

      ip0 = vlib_buffer_get_current (b[0]);
      ip1 = vlib_buffer_get_current (b[1]);
      lbi0 = vnet_buffer (b[0])->ip.adj_index[VLIB_TX];
      lbi1 = vnet_buffer (b[1])->ip.adj_index[VLIB_TX];

      lb0 = load_balance_get (lbi0);
      lb1 = load_balance_get (lbi1);

      /*
       * this node is for via FIBs we can re-use the hash value from the
       * to node if present.
       * We don't want to use the same hash value at each level in the recursion
       * graph as that would lead to polarisation
       */
      hc0 = hc1 = 0;

      if (PREDICT_FALSE (lb0->lb_n_buckets > 1))
	{
	  if (PREDICT_TRUE (vnet_buffer (b[0])->ip.flow_hash))
	    {
	      hc0 = vnet_buffer (b[0])->ip.flow_hash =
		vnet_buffer (b[0])->ip.flow_hash >> 1;
	    }
	  else
	    {
	      hc0 = vnet_buffer (b[0])->ip.flow_hash =
		ip4_compute_flow_hash (ip0, lb0->lb_hash_config);
	    }
	  dpo0 = load_balance_get_fwd_bucket
	    (lb0, (hc0 & (lb0->lb_n_buckets_minus_1)));
	}
      else
	{
	  dpo0 = load_balance_get_bucket_i (lb0, 0);
	}
      if (PREDICT_FALSE (lb1->lb_n_buckets > 1))
	{
	  if (PREDICT_TRUE (vnet_buffer (b[1])->ip.flow_hash))
	    {
	      hc1 = vnet_buffer (b[1])->ip.flow_hash =
		vnet_buffer (b[1])->ip.flow_hash >> 1;
	    }
	  else
	    {
	      hc1 = vnet_buffer (b[1])->ip.flow_hash =
		ip4_compute_flow_hash (ip1, lb1->lb_hash_config);
	    }
	  dpo1 = load_balance_get_fwd_bucket
	    (lb1, (hc1 & (lb1->lb_n_buckets_minus_1)));
	}
      else
	{
	  dpo1 = load_balance_get_bucket_i (lb1, 0);
	}

      next[0] = dpo0->dpoi_next_node;
      next[1] = dpo1->dpoi_next_node;

      vnet_buffer (b[0])->ip.adj_index[VLIB_TX] = dpo0->dpoi_index;
      vnet_buffer (b[1])->ip.adj_index[VLIB_TX] = dpo1->dpoi_index;

      vlib_increment_combined_counter
	(cm, thread_index, lbi0, 1, vlib_buffer_length_in_chain (vm, b[0]));
      vlib_increment_combined_counter
	(cm, thread_index, lbi1, 1, vlib_buffer_length_in_chain (vm, b[1]));

      b += 2;
      next += 2;
      n_left -= 2;
    }

  while (n_left > 0)
    {
      const load_balance_t *lb0;
      const ip4_header_t *ip0;
      const dpo_id_t *dpo0;
      u32 lbi0, hc0;

      ip0 = vlib_buffer_get_current (b[0]);
      lbi0 = vnet_buffer (b[0])->ip.adj_index[VLIB_TX];

      lb0 = load_balance_get (lbi0);

      hc0 = 0;
      if (PREDICT_FALSE (lb0->lb_n_buckets > 1))
	{
	  if (PREDICT_TRUE (vnet_buffer (b[0])->ip.flow_hash))
	    {
	      hc0 = vnet_buffer (b[0])->ip.flow_hash =
		vnet_buffer (b[0])->ip.flow_hash >> 1;
	    }
	  else
	    {
	      hc0 = vnet_buffer (b[0])->ip.flow_hash =
		ip4_compute_flow_hash (ip0, lb0->lb_hash_config);
	    }
	  dpo0 = load_balance_get_fwd_bucket
	    (lb0, (hc0 & (lb0->lb_n_buckets_minus_1)));
	}
      else
	{
	  dpo0 = load_balance_get_bucket_i (lb0, 0);
	}

      next[0] = dpo0->dpoi_next_node;
      vnet_buffer (b[0])->ip.adj_index[VLIB_TX] = dpo0->dpoi_index;

      vlib_increment_combined_counter
	(cm, thread_index, lbi0, 1, vlib_buffer_length_in_chain (vm, b[0]));

      b += 1;
      next += 1;
      n_left -= 1;
    }

  vlib_buffer_enqueue_to_next (vm, node, from, nexts, frame->n_vectors);
  if (node->flags & VLIB_NODE_FLAG_TRACE)
    ip4_forward_next_trace (vm, node, frame, VLIB_TX);

  return frame->n_vectors;
}

/* *INDENT-OFF* */
VLIB_REGISTER_NODE (ip4_load_balance_node) =
{
  .name = "ip4-load-balance",
  .vector_size = sizeof (u32),
  .sibling_of = "ip4-lookup",
  .format_trace = format_ip4_lookup_trace,
};
/* *INDENT-ON* */

#ifndef CLIB_MARCH_VARIANT
/* get first interface address */
ip4_address_t *
ip4_interface_first_address (ip4_main_t * im, u32 sw_if_index,
			     ip_interface_address_t ** result_ia)
{
  ip_lookup_main_t *lm = &im->lookup_main;
  ip_interface_address_t *ia = 0;
  ip4_address_t *result = 0;

  /* *INDENT-OFF* */
  foreach_ip_interface_address
    (lm, ia, sw_if_index,
     1 /* honor unnumbered */ ,
     ({
       ip4_address_t * a =
         ip_interface_address_get_address (lm, ia);
       result = a;
       break;
     }));
  /* *INDENT-OFF* */
  if (result_ia)
    *result_ia = result ? ia : 0;
  return result;
}

static void
ip4_add_subnet_bcast_route (u32 fib_index,
                            fib_prefix_t *pfx,
                            u32 sw_if_index)
{
  vnet_sw_interface_flags_t iflags;

  iflags = vnet_sw_interface_get_flags(vnet_get_main(), sw_if_index);

  fib_table_entry_special_remove(fib_index,
                                 pfx,
                                 FIB_SOURCE_INTERFACE);

  if (iflags & VNET_SW_INTERFACE_FLAG_DIRECTED_BCAST)
    {
      fib_table_entry_update_one_path (fib_index, pfx,
                                       FIB_SOURCE_INTERFACE,
                                       FIB_ENTRY_FLAG_NONE,
                                       DPO_PROTO_IP4,
                                       /* No next-hop address */
                                       &ADJ_BCAST_ADDR,
                                       sw_if_index,
                                       // invalid FIB index
                                       ~0,
                                       1,
                                       // no out-label stack
                                       NULL,
                                       FIB_ROUTE_PATH_FLAG_NONE);
    }
  else
    {
        fib_table_entry_special_add(fib_index,
                                    pfx,
                                    FIB_SOURCE_INTERFACE,
                                    (FIB_ENTRY_FLAG_DROP |
                                     FIB_ENTRY_FLAG_LOOSE_URPF_EXEMPT));
    }
}

static void
ip4_add_interface_routes (u32 sw_if_index,
			  ip4_main_t * im, u32 fib_index,
			  ip_interface_address_t * a)
{
  ip_lookup_main_t *lm = &im->lookup_main;
  ip4_address_t *address = ip_interface_address_get_address (lm, a);
  fib_prefix_t pfx = {
    .fp_len = a->address_length,
    .fp_proto = FIB_PROTOCOL_IP4,
    .fp_addr.ip4 = *address,
  };

  if (pfx.fp_len <= 30)
    {
      /* a /30 or shorter - add a glean for the network address */
      fib_table_entry_update_one_path (fib_index, &pfx,
                                       FIB_SOURCE_INTERFACE,
                                       (FIB_ENTRY_FLAG_CONNECTED |
                                        FIB_ENTRY_FLAG_ATTACHED),
                                       DPO_PROTO_IP4,
                                       /* No next-hop address */
                                       NULL,
                                       sw_if_index,
                                       // invalid FIB index
                                       ~0,
                                       1,
                                       // no out-label stack
                                       NULL,
                                       FIB_ROUTE_PATH_FLAG_NONE);

      /* Add the two broadcast addresses as drop */
      fib_prefix_t net_pfx = {
        .fp_len = 32,
        .fp_proto = FIB_PROTOCOL_IP4,
        .fp_addr.ip4.as_u32 = address->as_u32 & im->fib_masks[pfx.fp_len],
      };
      if (net_pfx.fp_addr.ip4.as_u32 != pfx.fp_addr.ip4.as_u32)
        fib_table_entry_special_add(fib_index,
                                    &net_pfx,
                                    FIB_SOURCE_INTERFACE,
                                    (FIB_ENTRY_FLAG_DROP |
                                     FIB_ENTRY_FLAG_LOOSE_URPF_EXEMPT));
      net_pfx.fp_addr.ip4.as_u32 |= ~im->fib_masks[pfx.fp_len];
      if (net_pfx.fp_addr.ip4.as_u32 != pfx.fp_addr.ip4.as_u32)
        ip4_add_subnet_bcast_route(fib_index, &net_pfx, sw_if_index);
    }
  else if (pfx.fp_len == 31)
    {
      u32 mask = clib_host_to_net_u32(1);
      fib_prefix_t net_pfx = pfx;

      net_pfx.fp_len = 32;
      net_pfx.fp_addr.ip4.as_u32 ^= mask;

      /* a /31 - add the other end as an attached host */
      fib_table_entry_update_one_path (fib_index, &net_pfx,
                                       FIB_SOURCE_INTERFACE,
                                       (FIB_ENTRY_FLAG_ATTACHED),
                                       DPO_PROTO_IP4,
                                       &net_pfx.fp_addr,
                                       sw_if_index,
                                       // invalid FIB index
                                       ~0,
                                       1,
                                       NULL,
                                       FIB_ROUTE_PATH_FLAG_NONE);
    }
  pfx.fp_len = 32;

  if (sw_if_index < vec_len (lm->classify_table_index_by_sw_if_index))
    {
      u32 classify_table_index =
	lm->classify_table_index_by_sw_if_index[sw_if_index];
      if (classify_table_index != (u32) ~ 0)
	{
	  dpo_id_t dpo = DPO_INVALID;

	  dpo_set (&dpo,
		   DPO_CLASSIFY,
		   DPO_PROTO_IP4,
		   classify_dpo_create (DPO_PROTO_IP4, classify_table_index));

	  fib_table_entry_special_dpo_add (fib_index,
					   &pfx,
					   FIB_SOURCE_CLASSIFY,
					   FIB_ENTRY_FLAG_NONE, &dpo);
	  dpo_reset (&dpo);
	}
    }

  fib_table_entry_update_one_path (fib_index, &pfx,
                                   FIB_SOURCE_INTERFACE,
                                   (FIB_ENTRY_FLAG_CONNECTED |
                                    FIB_ENTRY_FLAG_LOCAL),
                                   DPO_PROTO_IP4,
                                   &pfx.fp_addr,
                                   sw_if_index,
                                   // invalid FIB index
                                   ~0,
				   1, NULL,
				   FIB_ROUTE_PATH_FLAG_NONE);
}

static void
ip4_del_interface_routes (ip4_main_t * im,
			  u32 fib_index,
			  ip4_address_t * address, u32 address_length)
{
  fib_prefix_t pfx = {
    .fp_len = address_length,
    .fp_proto = FIB_PROTOCOL_IP4,
    .fp_addr.ip4 = *address,
  };

  if (pfx.fp_len <= 30)
    {
      fib_prefix_t net_pfx = {
        .fp_len = 32,
        .fp_proto = FIB_PROTOCOL_IP4,
        .fp_addr.ip4.as_u32 = address->as_u32 & im->fib_masks[pfx.fp_len],
      };
      if (net_pfx.fp_addr.ip4.as_u32 != pfx.fp_addr.ip4.as_u32)
        fib_table_entry_special_remove(fib_index,
                                       &net_pfx,
                                       FIB_SOURCE_INTERFACE);
      net_pfx.fp_addr.ip4.as_u32 |= ~im->fib_masks[pfx.fp_len];
      if (net_pfx.fp_addr.ip4.as_u32 != pfx.fp_addr.ip4.as_u32)
        fib_table_entry_special_remove(fib_index,
                                       &net_pfx,
                                       FIB_SOURCE_INTERFACE);
      fib_table_entry_delete (fib_index, &pfx, FIB_SOURCE_INTERFACE);
    }
    else if (pfx.fp_len == 31)
    {
      u32 mask = clib_host_to_net_u32(1);
      fib_prefix_t net_pfx = pfx;

      net_pfx.fp_len = 32;
      net_pfx.fp_addr.ip4.as_u32 ^= mask;

      fib_table_entry_delete (fib_index, &net_pfx, FIB_SOURCE_INTERFACE);
    }

  pfx.fp_len = 32;
  fib_table_entry_delete (fib_index, &pfx, FIB_SOURCE_INTERFACE);
}

void
ip4_sw_interface_enable_disable (u32 sw_if_index, u32 is_enable)
{
  ip4_main_t *im = &ip4_main;

  vec_validate_init_empty (im->ip_enabled_by_sw_if_index, sw_if_index, 0);

  /*
   * enable/disable only on the 1<->0 transition
   */
  if (is_enable)
    {
      if (1 != ++im->ip_enabled_by_sw_if_index[sw_if_index])
	return;
    }
  else
    {
      ASSERT (im->ip_enabled_by_sw_if_index[sw_if_index] > 0);
      if (0 != --im->ip_enabled_by_sw_if_index[sw_if_index])
	return;
    }
  vnet_feature_enable_disable ("ip4-unicast", "ip4-not-enabled", sw_if_index,
			       !is_enable, 0, 0);


  vnet_feature_enable_disable ("ip4-multicast", "ip4-not-enabled",
			       sw_if_index, !is_enable, 0, 0);
}

static clib_error_t *
ip4_add_del_interface_address_internal (vlib_main_t * vm,
					u32 sw_if_index,
					ip4_address_t * address,
					u32 address_length, u32 is_del)
{
  vnet_main_t *vnm = vnet_get_main ();
  ip4_main_t *im = &ip4_main;
  ip_lookup_main_t *lm = &im->lookup_main;
  clib_error_t *error = 0;
  u32 if_address_index, elts_before;
  ip4_address_fib_t ip4_af, *addr_fib = 0;

  /* local0 interface doesn't support IP addressing  */
  if (sw_if_index == 0)
    {
      return
       clib_error_create ("local0 interface doesn't support IP addressing");
    }

  vec_validate (im->fib_index_by_sw_if_index, sw_if_index);
  ip4_addr_fib_init (&ip4_af, address,
		     vec_elt (im->fib_index_by_sw_if_index, sw_if_index));
  vec_add1 (addr_fib, ip4_af);

  /*
   * there is no support for adj-fib handling in the presence of overlapping
   * subnets on interfaces. Easy fix - disallow overlapping subnets, like
   * most routers do.
   */
  /* *INDENT-OFF* */
  if (!is_del)
    {
      /* When adding an address check that it does not conflict
         with an existing address on any interface in this table. */
      ip_interface_address_t *ia;
      vnet_sw_interface_t *sif;

      pool_foreach(sif, vnm->interface_main.sw_interfaces,
      ({
          if (im->fib_index_by_sw_if_index[sw_if_index] ==
              im->fib_index_by_sw_if_index[sif->sw_if_index])
            {
              foreach_ip_interface_address
                (&im->lookup_main, ia, sif->sw_if_index,
                 0 /* honor unnumbered */ ,
                 ({
                   ip4_address_t * x =
                     ip_interface_address_get_address
                     (&im->lookup_main, ia);
                   if (ip4_destination_matches_route
                       (im, address, x, ia->address_length) ||
                       ip4_destination_matches_route (im,
                                                      x,
                                                      address,
                                                      address_length))
                     {
                       vnm->api_errno = VNET_API_ERROR_DUPLICATE_IF_ADDRESS;

                       return
                         clib_error_create
                         ("failed to add %U which conflicts with %U for interface %U",
                          format_ip4_address_and_length, address,
                          address_length,
                          format_ip4_address_and_length, x,
                          ia->address_length,
                          format_vnet_sw_if_index_name, vnm,
                          sif->sw_if_index);
                     }
                 }));
            }
      }));
    }
  /* *INDENT-ON* */

  elts_before = pool_elts (lm->if_address_pool);

  error = ip_interface_address_add_del
    (lm, sw_if_index, addr_fib, address_length, is_del, &if_address_index);
  if (error)
    goto done;

  ip4_sw_interface_enable_disable (sw_if_index, !is_del);

  if (is_del)
    ip4_del_interface_routes (im, ip4_af.fib_index, address, address_length);
  else
    ip4_add_interface_routes (sw_if_index,
			      im, ip4_af.fib_index,
			      pool_elt_at_index
			      (lm->if_address_pool, if_address_index));

  /* If pool did not grow/shrink: add duplicate address. */
  if (elts_before != pool_elts (lm->if_address_pool))
    {
      ip4_add_del_interface_address_callback_t *cb;
      vec_foreach (cb, im->add_del_interface_address_callbacks)
	cb->function (im, cb->function_opaque, sw_if_index,
		      address, address_length, if_address_index, is_del);
    }

done:
  vec_free (addr_fib);
  return error;
}

clib_error_t *
ip4_add_del_interface_address (vlib_main_t * vm,
			       u32 sw_if_index,
			       ip4_address_t * address,
			       u32 address_length, u32 is_del)
{
  return ip4_add_del_interface_address_internal
    (vm, sw_if_index, address, address_length, is_del);
}

void
ip4_directed_broadcast (u32 sw_if_index, u8 enable)
{
  ip_interface_address_t *ia;
  ip4_main_t *im;

  im = &ip4_main;

  /*
   * when directed broadcast is enabled, the subnet braodcast route will forward
   * packets using an adjacency with a broadcast MAC. otherwise it drops
   */
  /* *INDENT-OFF* */
  foreach_ip_interface_address(&im->lookup_main, ia,
                               sw_if_index, 0,
     ({
       if (ia->address_length <= 30)
         {
           ip4_address_t *ipa;

           ipa = ip_interface_address_get_address (&im->lookup_main, ia);

           fib_prefix_t pfx = {
             .fp_len = 32,
             .fp_proto = FIB_PROTOCOL_IP4,
             .fp_addr = {
               .ip4.as_u32 = (ipa->as_u32 | ~im->fib_masks[ia->address_length]),
             },
           };

           ip4_add_subnet_bcast_route
             (fib_table_get_index_for_sw_if_index(FIB_PROTOCOL_IP4,
                                                  sw_if_index),
              &pfx, sw_if_index);
         }
     }));
  /* *INDENT-ON* */
}
#endif

/* Built-in ip4 unicast rx feature path definition */
/* *INDENT-OFF* */
VNET_FEATURE_ARC_INIT (ip4_unicast, static) =
{
  .arc_name = "ip4-unicast",
  .start_nodes = VNET_FEATURES ("ip4-input", "ip4-input-no-checksum"),
  .last_in_arc = "ip4-lookup",
  .arc_index_ptr = &ip4_main.lookup_main.ucast_feature_arc_index,
};

VNET_FEATURE_INIT (ip4_flow_classify, static) =
{
  .arc_name = "ip4-unicast",
  .node_name = "ip4-flow-classify",
  .runs_before = VNET_FEATURES ("ip4-inacl"),
};

VNET_FEATURE_INIT (ip4_inacl, static) =
{
  .arc_name = "ip4-unicast",
  .node_name = "ip4-inacl",
  .runs_before = VNET_FEATURES ("ip4-source-check-via-rx"),
};

VNET_FEATURE_INIT (ip4_source_check_1, static) =
{
  .arc_name = "ip4-unicast",
  .node_name = "ip4-source-check-via-rx",
  .runs_before = VNET_FEATURES ("ip4-source-check-via-any"),
};

VNET_FEATURE_INIT (ip4_source_check_2, static) =
{
  .arc_name = "ip4-unicast",
  .node_name = "ip4-source-check-via-any",
  .runs_before = VNET_FEATURES ("ip4-policer-classify"),
};

VNET_FEATURE_INIT (ip4_source_and_port_range_check_rx, static) =
{
  .arc_name = "ip4-unicast",
  .node_name = "ip4-source-and-port-range-check-rx",
  .runs_before = VNET_FEATURES ("ip4-policer-classify"),
};

VNET_FEATURE_INIT (ip4_policer_classify, static) =
{
  .arc_name = "ip4-unicast",
  .node_name = "ip4-policer-classify",
  .runs_before = VNET_FEATURES ("ipsec4-input-feature"),
};

VNET_FEATURE_INIT (ip4_ipsec, static) =
{
  .arc_name = "ip4-unicast",
  .node_name = "ipsec4-input-feature",
  .runs_before = VNET_FEATURES ("vpath-input-ip4"),
};

VNET_FEATURE_INIT (ip4_vpath, static) =
{
  .arc_name = "ip4-unicast",
  .node_name = "vpath-input-ip4",
  .runs_before = VNET_FEATURES ("ip4-vxlan-bypass"),
};

VNET_FEATURE_INIT (ip4_vxlan_bypass, static) =
{
  .arc_name = "ip4-unicast",
  .node_name = "ip4-vxlan-bypass",
  .runs_before = VNET_FEATURES ("ip4-lookup"),
};

VNET_FEATURE_INIT (ip4_not_enabled, static) =
{
  .arc_name = "ip4-unicast",
  .node_name = "ip4-not-enabled",
  .runs_before = VNET_FEATURES ("ip4-lookup"),
};

VNET_FEATURE_INIT (ip4_lookup, static) =
{
  .arc_name = "ip4-unicast",
  .node_name = "ip4-lookup",
  .runs_before = 0,	/* not before any other features */
};

/* Built-in ip4 multicast rx feature path definition */
VNET_FEATURE_ARC_INIT (ip4_multicast, static) =
{
  .arc_name = "ip4-multicast",
  .start_nodes = VNET_FEATURES ("ip4-input", "ip4-input-no-checksum"),
  .last_in_arc = "ip4-mfib-forward-lookup",
  .arc_index_ptr = &ip4_main.lookup_main.mcast_feature_arc_index,
};

VNET_FEATURE_INIT (ip4_vpath_mc, static) =
{
  .arc_name = "ip4-multicast",
  .node_name = "vpath-input-ip4",
  .runs_before = VNET_FEATURES ("ip4-mfib-forward-lookup"),
};

VNET_FEATURE_INIT (ip4_mc_not_enabled, static) =
{
  .arc_name = "ip4-multicast",
  .node_name = "ip4-not-enabled",
  .runs_before = VNET_FEATURES ("ip4-mfib-forward-lookup"),
};

VNET_FEATURE_INIT (ip4_lookup_mc, static) =
{
  .arc_name = "ip4-multicast",
  .node_name = "ip4-mfib-forward-lookup",
  .runs_before = 0,	/* last feature */
};

/* Source and port-range check ip4 tx feature path definition */
VNET_FEATURE_ARC_INIT (ip4_output, static) =
{
  .arc_name = "ip4-output",
  .start_nodes = VNET_FEATURES ("ip4-rewrite", "ip4-midchain", "ip4-dvr-dpo"),
  .last_in_arc = "interface-output",
  .arc_index_ptr = &ip4_main.lookup_main.output_feature_arc_index,
};

VNET_FEATURE_INIT (ip4_source_and_port_range_check_tx, static) =
{
  .arc_name = "ip4-output",
  .node_name = "ip4-source-and-port-range-check-tx",
  .runs_before = VNET_FEATURES ("ip4-outacl"),
};

VNET_FEATURE_INIT (ip4_outacl, static) =
{
  .arc_name = "ip4-output",
  .node_name = "ip4-outacl",
  .runs_before = VNET_FEATURES ("ipsec4-output-feature"),
};

VNET_FEATURE_INIT (ip4_ipsec_output, static) =
{
  .arc_name = "ip4-output",
  .node_name = "ipsec4-output-feature",
  .runs_before = VNET_FEATURES ("interface-output"),
};

/* Built-in ip4 tx feature path definition */
VNET_FEATURE_INIT (ip4_interface_output, static) =
{
  .arc_name = "ip4-output",
  .node_name = "interface-output",
  .runs_before = 0,	/* not before any other features */
};
/* *INDENT-ON* */

static clib_error_t *
ip4_sw_interface_add_del (vnet_main_t * vnm, u32 sw_if_index, u32 is_add)
{
  ip4_main_t *im = &ip4_main;

  /* Fill in lookup tables with default table (0). */
  vec_validate (im->fib_index_by_sw_if_index, sw_if_index);
  vec_validate (im->mfib_index_by_sw_if_index, sw_if_index);

  if (!is_add)
    {
      ip4_main_t *im4 = &ip4_main;
      ip_lookup_main_t *lm4 = &im4->lookup_main;
      ip_interface_address_t *ia = 0;
      ip4_address_t *address;
      vlib_main_t *vm = vlib_get_main ();

      vnet_sw_interface_update_unnumbered (sw_if_index, ~0, 0);
      /* *INDENT-OFF* */
      foreach_ip_interface_address (lm4, ia, sw_if_index, 0,
      ({
        address = ip_interface_address_get_address (lm4, ia);
        ip4_add_del_interface_address(vm, sw_if_index, address, ia->address_length, 1);
      }));
      /* *INDENT-ON* */
    }

  vnet_feature_enable_disable ("ip4-unicast", "ip4-not-enabled", sw_if_index,
			       is_add, 0, 0);

  vnet_feature_enable_disable ("ip4-multicast", "ip4-not-enabled",
			       sw_if_index, is_add, 0, 0);

  return /* no error */ 0;
}

VNET_SW_INTERFACE_ADD_DEL_FUNCTION (ip4_sw_interface_add_del);

/* Global IP4 main. */
#ifndef CLIB_MARCH_VARIANT
ip4_main_t ip4_main;
#endif /* CLIB_MARCH_VARIANT */

static clib_error_t *
ip4_lookup_init (vlib_main_t * vm)
{
  ip4_main_t *im = &ip4_main;
  clib_error_t *error;
  uword i;

  if ((error = vlib_call_init_function (vm, vnet_feature_init)))
    return error;
  if ((error = vlib_call_init_function (vm, ip4_mtrie_module_init)))
    return (error);
  if ((error = vlib_call_init_function (vm, fib_module_init)))
    return error;
  if ((error = vlib_call_init_function (vm, mfib_module_init)))
    return error;

  for (i = 0; i < ARRAY_LEN (im->fib_masks); i++)
    {
      u32 m;

      if (i < 32)
	m = pow2_mask (i) << (32 - i);
      else
	m = ~0;
      im->fib_masks[i] = clib_host_to_net_u32 (m);
    }

  ip_lookup_init (&im->lookup_main, /* is_ip6 */ 0);

  /* Create FIB with index 0 and table id of 0. */
  fib_table_find_or_create_and_lock (FIB_PROTOCOL_IP4, 0,
				     FIB_SOURCE_DEFAULT_ROUTE);
  mfib_table_find_or_create_and_lock (FIB_PROTOCOL_IP4, 0,
				      MFIB_SOURCE_DEFAULT_ROUTE);

  {
    pg_node_t *pn;
    pn = pg_get_node (ip4_lookup_node.index);
    pn->unformat_edit = unformat_pg_ip4_header;
  }

  {
    ethernet_arp_header_t h;

    clib_memset (&h, 0, sizeof (h));

#define _16(f,v) h.f = clib_host_to_net_u16 (v);
#define _8(f,v) h.f = v;
    _16 (l2_type, ETHERNET_ARP_HARDWARE_TYPE_ethernet);
    _16 (l3_type, ETHERNET_TYPE_IP4);
    _8 (n_l2_address_bytes, 6);
    _8 (n_l3_address_bytes, 4);
    _16 (opcode, ETHERNET_ARP_OPCODE_request);
#undef _16
#undef _8

    vlib_packet_template_init (vm, &im->ip4_arp_request_packet_template,
			       /* data */ &h,
			       sizeof (h),
			       /* alloc chunk size */ 8,
			       "ip4 arp");
  }

  return error;
}

VLIB_INIT_FUNCTION (ip4_lookup_init);

typedef struct
{
  /* Adjacency taken. */
  u32 dpo_index;
  u32 flow_hash;
  u32 fib_index;

  /* Packet data, possibly *after* rewrite. */
  u8 packet_data[64 - 1 * sizeof (u32)];
}
ip4_forward_next_trace_t;

#ifndef CLIB_MARCH_VARIANT
u8 *
format_ip4_forward_next_trace (u8 * s, va_list * args)
{
  CLIB_UNUSED (vlib_main_t * vm) = va_arg (*args, vlib_main_t *);
  CLIB_UNUSED (vlib_node_t * node) = va_arg (*args, vlib_node_t *);
  ip4_forward_next_trace_t *t = va_arg (*args, ip4_forward_next_trace_t *);
  u32 indent = format_get_indent (s);
  s = format (s, "%U%U",
	      format_white_space, indent,
	      format_ip4_header, t->packet_data, sizeof (t->packet_data));
  return s;
}
#endif

static u8 *
format_ip4_lookup_trace (u8 * s, va_list * args)
{
  CLIB_UNUSED (vlib_main_t * vm) = va_arg (*args, vlib_main_t *);
  CLIB_UNUSED (vlib_node_t * node) = va_arg (*args, vlib_node_t *);
  ip4_forward_next_trace_t *t = va_arg (*args, ip4_forward_next_trace_t *);
  u32 indent = format_get_indent (s);

  s = format (s, "fib %d dpo-idx %d flow hash: 0x%08x",
	      t->fib_index, t->dpo_index, t->flow_hash);
  s = format (s, "\n%U%U",
	      format_white_space, indent,
	      format_ip4_header, t->packet_data, sizeof (t->packet_data));
  return s;
}

static u8 *
format_ip4_rewrite_trace (u8 * s, va_list * args)
{
  CLIB_UNUSED (vlib_main_t * vm) = va_arg (*args, vlib_main_t *);
  CLIB_UNUSED (vlib_node_t * node) = va_arg (*args, vlib_node_t *);
  ip4_forward_next_trace_t *t = va_arg (*args, ip4_forward_next_trace_t *);
  u32 indent = format_get_indent (s);

  s = format (s, "tx_sw_if_index %d dpo-idx %d : %U flow hash: 0x%08x",
	      t->fib_index, t->dpo_index, format_ip_adjacency,
	      t->dpo_index, FORMAT_IP_ADJACENCY_NONE, t->flow_hash);
  s = format (s, "\n%U%U",
	      format_white_space, indent,
	      format_ip_adjacency_packet_data,
	      t->dpo_index, t->packet_data, sizeof (t->packet_data));
  return s;
}

#ifndef CLIB_MARCH_VARIANT
/* Common trace function for all ip4-forward next nodes. */
void
ip4_forward_next_trace (vlib_main_t * vm,
			vlib_node_runtime_t * node,
			vlib_frame_t * frame, vlib_rx_or_tx_t which_adj_index)
{
  u32 *from, n_left;
  ip4_main_t *im = &ip4_main;

  n_left = frame->n_vectors;
  from = vlib_frame_vector_args (frame);

  while (n_left >= 4)
    {
      u32 bi0, bi1;
      vlib_buffer_t *b0, *b1;
      ip4_forward_next_trace_t *t0, *t1;

      /* Prefetch next iteration. */
      vlib_prefetch_buffer_with_index (vm, from[2], LOAD);
      vlib_prefetch_buffer_with_index (vm, from[3], LOAD);

      bi0 = from[0];
      bi1 = from[1];

      b0 = vlib_get_buffer (vm, bi0);
      b1 = vlib_get_buffer (vm, bi1);

      if (b0->flags & VLIB_BUFFER_IS_TRACED)
	{
	  t0 = vlib_add_trace (vm, node, b0, sizeof (t0[0]));
	  t0->dpo_index = vnet_buffer (b0)->ip.adj_index[which_adj_index];
	  t0->flow_hash = vnet_buffer (b0)->ip.flow_hash;
	  t0->fib_index =
	    (vnet_buffer (b0)->sw_if_index[VLIB_TX] !=
	     (u32) ~ 0) ? vnet_buffer (b0)->sw_if_index[VLIB_TX] :
	    vec_elt (im->fib_index_by_sw_if_index,
		     vnet_buffer (b0)->sw_if_index[VLIB_RX]);

	  clib_memcpy_fast (t0->packet_data,
			    vlib_buffer_get_current (b0),
			    sizeof (t0->packet_data));
	}
      if (b1->flags & VLIB_BUFFER_IS_TRACED)
	{
	  t1 = vlib_add_trace (vm, node, b1, sizeof (t1[0]));
	  t1->dpo_index = vnet_buffer (b1)->ip.adj_index[which_adj_index];
	  t1->flow_hash = vnet_buffer (b1)->ip.flow_hash;
	  t1->fib_index =
	    (vnet_buffer (b1)->sw_if_index[VLIB_TX] !=
	     (u32) ~ 0) ? vnet_buffer (b1)->sw_if_index[VLIB_TX] :
	    vec_elt (im->fib_index_by_sw_if_index,
		     vnet_buffer (b1)->sw_if_index[VLIB_RX]);
	  clib_memcpy_fast (t1->packet_data, vlib_buffer_get_current (b1),
			    sizeof (t1->packet_data));
	}
      from += 2;
      n_left -= 2;
    }

  while (n_left >= 1)
    {
      u32 bi0;
      vlib_buffer_t *b0;
      ip4_forward_next_trace_t *t0;

      bi0 = from[0];

      b0 = vlib_get_buffer (vm, bi0);

      if (b0->flags & VLIB_BUFFER_IS_TRACED)
	{
	  t0 = vlib_add_trace (vm, node, b0, sizeof (t0[0]));
	  t0->dpo_index = vnet_buffer (b0)->ip.adj_index[which_adj_index];
	  t0->flow_hash = vnet_buffer (b0)->ip.flow_hash;
	  t0->fib_index =
	    (vnet_buffer (b0)->sw_if_index[VLIB_TX] !=
	     (u32) ~ 0) ? vnet_buffer (b0)->sw_if_index[VLIB_TX] :
	    vec_elt (im->fib_index_by_sw_if_index,
		     vnet_buffer (b0)->sw_if_index[VLIB_RX]);
	  clib_memcpy_fast (t0->packet_data, vlib_buffer_get_current (b0),
			    sizeof (t0->packet_data));
	}
      from += 1;
      n_left -= 1;
    }
}

/* Compute TCP/UDP/ICMP4 checksum in software. */
u16
ip4_tcp_udp_compute_checksum (vlib_main_t * vm, vlib_buffer_t * p0,
			      ip4_header_t * ip0)
{
  ip_csum_t sum0;
  u32 ip_header_length, payload_length_host_byte_order;
  u32 n_this_buffer, n_bytes_left, n_ip_bytes_this_buffer;
  u16 sum16;
  void *data_this_buffer;

  /* Initialize checksum with ip header. */
  ip_header_length = ip4_header_bytes (ip0);
  payload_length_host_byte_order =
    clib_net_to_host_u16 (ip0->length) - ip_header_length;
  sum0 =
    clib_host_to_net_u32 (payload_length_host_byte_order +
			  (ip0->protocol << 16));

  if (BITS (uword) == 32)
    {
      sum0 =
	ip_csum_with_carry (sum0,
			    clib_mem_unaligned (&ip0->src_address, u32));
      sum0 =
	ip_csum_with_carry (sum0,
			    clib_mem_unaligned (&ip0->dst_address, u32));
    }
  else
    sum0 =
      ip_csum_with_carry (sum0, clib_mem_unaligned (&ip0->src_address, u64));

  n_bytes_left = n_this_buffer = payload_length_host_byte_order;
  data_this_buffer = (void *) ip0 + ip_header_length;
  n_ip_bytes_this_buffer =
    p0->current_length - (((u8 *) ip0 - p0->data) - p0->current_data);
  if (n_this_buffer + ip_header_length > n_ip_bytes_this_buffer)
    {
      n_this_buffer = n_ip_bytes_this_buffer > ip_header_length ?
	n_ip_bytes_this_buffer - ip_header_length : 0;
    }
  while (1)
    {
      sum0 = ip_incremental_checksum (sum0, data_this_buffer, n_this_buffer);
      n_bytes_left -= n_this_buffer;
      if (n_bytes_left == 0)
	break;

      ASSERT (p0->flags & VLIB_BUFFER_NEXT_PRESENT);
      p0 = vlib_get_buffer (vm, p0->next_buffer);
      data_this_buffer = vlib_buffer_get_current (p0);
      n_this_buffer = clib_min (p0->current_length, n_bytes_left);
    }

  sum16 = ~ip_csum_fold (sum0);

  return sum16;
}

u32
ip4_tcp_udp_validate_checksum (vlib_main_t * vm, vlib_buffer_t * p0)
{
  ip4_header_t *ip0 = vlib_buffer_get_current (p0);
  udp_header_t *udp0;
  u16 sum16;

  ASSERT (ip0->protocol == IP_PROTOCOL_TCP
	  || ip0->protocol == IP_PROTOCOL_UDP);

  udp0 = (void *) (ip0 + 1);
  if (ip0->protocol == IP_PROTOCOL_UDP && udp0->checksum == 0)
    {
      p0->flags |= (VNET_BUFFER_F_L4_CHECKSUM_COMPUTED
		    | VNET_BUFFER_F_L4_CHECKSUM_CORRECT);
      return p0->flags;
    }

  sum16 = ip4_tcp_udp_compute_checksum (vm, p0, ip0);

  p0->flags |= (VNET_BUFFER_F_L4_CHECKSUM_COMPUTED
		| ((sum16 == 0) << VNET_BUFFER_F_LOG2_L4_CHECKSUM_CORRECT));

  return p0->flags;
}
#endif

/* *INDENT-OFF* */
VNET_FEATURE_ARC_INIT (ip4_local) =
{
  .arc_name  = "ip4-local",
  .start_nodes = VNET_FEATURES ("ip4-local"),
  .last_in_arc = "ip4-local-end-of-arc",
};
/* *INDENT-ON* */

static inline void
ip4_local_l4_csum_validate (vlib_main_t * vm, vlib_buffer_t * p,
			    ip4_header_t * ip, u8 is_udp, u8 * error,
			    u8 * good_tcp_udp)
{
  u32 flags0;
  flags0 = ip4_tcp_udp_validate_checksum (vm, p);
  *good_tcp_udp = (flags0 & VNET_BUFFER_F_L4_CHECKSUM_CORRECT) != 0;
  if (is_udp)
    {
      udp_header_t *udp;
      u32 ip_len, udp_len;
      i32 len_diff;
      udp = ip4_next_header (ip);
      /* Verify UDP length. */
      ip_len = clib_net_to_host_u16 (ip->length);
      udp_len = clib_net_to_host_u16 (udp->length);

      len_diff = ip_len - udp_len;
      *good_tcp_udp &= len_diff >= 0;
      *error = len_diff < 0 ? IP4_ERROR_UDP_LENGTH : *error;
    }
}

#define ip4_local_csum_is_offloaded(_b)					\
    _b->flags & VNET_BUFFER_F_OFFLOAD_TCP_CKSUM				\
	|| _b->flags & VNET_BUFFER_F_OFFLOAD_UDP_CKSUM

#define ip4_local_need_csum_check(is_tcp_udp, _b) 			\
    (is_tcp_udp && !(_b->flags & VNET_BUFFER_F_L4_CHECKSUM_COMPUTED 	\
	|| ip4_local_csum_is_offloaded (_b)))

#define ip4_local_csum_is_valid(_b)					\
    (_b->flags & VNET_BUFFER_F_L4_CHECKSUM_CORRECT			\
	|| (ip4_local_csum_is_offloaded (_b))) != 0

static inline void
ip4_local_check_l4_csum (vlib_main_t * vm, vlib_buffer_t * b,
			 ip4_header_t * ih, u8 * error)
{
  u8 is_udp, is_tcp_udp, good_tcp_udp;

  is_udp = ih->protocol == IP_PROTOCOL_UDP;
  is_tcp_udp = is_udp || ih->protocol == IP_PROTOCOL_TCP;

  if (PREDICT_FALSE (ip4_local_need_csum_check (is_tcp_udp, b)))
    ip4_local_l4_csum_validate (vm, b, ih, is_udp, error, &good_tcp_udp);
  else
    good_tcp_udp = ip4_local_csum_is_valid (b);

  ASSERT (IP4_ERROR_TCP_CHECKSUM + 1 == IP4_ERROR_UDP_CHECKSUM);
  *error = (is_tcp_udp && !good_tcp_udp
	    ? IP4_ERROR_TCP_CHECKSUM + is_udp : *error);
}

static inline void
ip4_local_check_l4_csum_x2 (vlib_main_t * vm, vlib_buffer_t ** b,
			    ip4_header_t ** ih, u8 * error)
{
  u8 is_udp[2], is_tcp_udp[2], good_tcp_udp[2];

  is_udp[0] = ih[0]->protocol == IP_PROTOCOL_UDP;
  is_udp[1] = ih[1]->protocol == IP_PROTOCOL_UDP;

  is_tcp_udp[0] = is_udp[0] || ih[0]->protocol == IP_PROTOCOL_TCP;
  is_tcp_udp[1] = is_udp[1] || ih[1]->protocol == IP_PROTOCOL_TCP;

  good_tcp_udp[0] = ip4_local_csum_is_valid (b[0]);
  good_tcp_udp[1] = ip4_local_csum_is_valid (b[1]);

  if (PREDICT_FALSE (ip4_local_need_csum_check (is_tcp_udp[0], b[0])
		     || ip4_local_need_csum_check (is_tcp_udp[1], b[1])))
    {
      if (is_tcp_udp[0])
	ip4_local_l4_csum_validate (vm, b[0], ih[0], is_udp[0], &error[0],
				    &good_tcp_udp[0]);
      if (is_tcp_udp[1])
	ip4_local_l4_csum_validate (vm, b[1], ih[1], is_udp[1], &error[1],
				    &good_tcp_udp[1]);
    }

  error[0] = (is_tcp_udp[0] && !good_tcp_udp[0] ?
	      IP4_ERROR_TCP_CHECKSUM + is_udp[0] : error[0]);
  error[1] = (is_tcp_udp[1] && !good_tcp_udp[1] ?
	      IP4_ERROR_TCP_CHECKSUM + is_udp[1] : error[1]);
}

static inline void
ip4_local_set_next_and_error (vlib_node_runtime_t * error_node,
			      vlib_buffer_t * b, u16 * next, u8 error,
			      u8 head_of_feature_arc)
{
  u8 arc_index = vnet_feat_arc_ip4_local.feature_arc_index;
  u32 next_index;

  *next = error != IP4_ERROR_UNKNOWN_PROTOCOL ? IP_LOCAL_NEXT_DROP : *next;
  b->error = error ? error_node->errors[error] : 0;
  if (head_of_feature_arc)
    {
      next_index = *next;
      if (PREDICT_TRUE (error == (u8) IP4_ERROR_UNKNOWN_PROTOCOL))
	{
	  vnet_feature_arc_start (arc_index,
				  vnet_buffer (b)->sw_if_index[VLIB_RX],
				  &next_index, b);
	  *next = next_index;
	}
    }
}

typedef struct
{
  ip4_address_t src;
  u32 lbi;
  u8 error;
  u8 first;
} ip4_local_last_check_t;

static inline void
ip4_local_check_src (vlib_buffer_t * b, ip4_header_t * ip0,
		     ip4_local_last_check_t * last_check, u8 * error0)
{
  ip4_fib_mtrie_leaf_t leaf0;
  ip4_fib_mtrie_t *mtrie0;
  const dpo_id_t *dpo0;
  load_balance_t *lb0;
  u32 lbi0;

  vnet_buffer (b)->ip.fib_index =
    vnet_buffer (b)->sw_if_index[VLIB_TX] != ~0 ?
    vnet_buffer (b)->sw_if_index[VLIB_TX] : vnet_buffer (b)->ip.fib_index;

  if (PREDICT_FALSE (last_check->first ||
		     (last_check->src.as_u32 != ip0->src_address.as_u32)))
    {
      mtrie0 = &ip4_fib_get (vnet_buffer (b)->ip.fib_index)->mtrie;
      leaf0 = ip4_fib_mtrie_lookup_step_one (mtrie0, &ip0->src_address);
      leaf0 = ip4_fib_mtrie_lookup_step (mtrie0, leaf0, &ip0->src_address, 2);
      leaf0 = ip4_fib_mtrie_lookup_step (mtrie0, leaf0, &ip0->src_address, 3);
      lbi0 = ip4_fib_mtrie_leaf_get_adj_index (leaf0);

      vnet_buffer (b)->ip.adj_index[VLIB_TX] = lbi0;
      vnet_buffer (b)->ip.adj_index[VLIB_RX] = lbi0;

      lb0 = load_balance_get (lbi0);
      dpo0 = load_balance_get_bucket_i (lb0, 0);

      /*
       * Must have a route to source otherwise we drop the packet.
       * ip4 broadcasts are accepted, e.g. to make dhcp client work
       *
       * The checks are:
       *  - the source is a recieve => it's from us => bogus, do this
       *    first since it sets a different error code.
       *  - uRPF check for any route to source - accept if passes.
       *  - allow packets destined to the broadcast address from unknown sources
       */

      *error0 = ((*error0 == IP4_ERROR_UNKNOWN_PROTOCOL
		  && dpo0->dpoi_type == DPO_RECEIVE) ?
		 IP4_ERROR_SPOOFED_LOCAL_PACKETS : *error0);
      *error0 = ((*error0 == IP4_ERROR_UNKNOWN_PROTOCOL
		  && !fib_urpf_check_size (lb0->lb_urpf)
		  && ip0->dst_address.as_u32 != 0xFFFFFFFF) ?
		 IP4_ERROR_SRC_LOOKUP_MISS : *error0);

      last_check->src.as_u32 = ip0->src_address.as_u32;
      last_check->lbi = lbi0;
      last_check->error = *error0;
    }
  else
    {
      vnet_buffer (b)->ip.adj_index[VLIB_TX] = last_check->lbi;
      vnet_buffer (b)->ip.adj_index[VLIB_RX] = last_check->lbi;
      *error0 = last_check->error;
      last_check->first = 0;
    }
}

static inline void
ip4_local_check_src_x2 (vlib_buffer_t ** b, ip4_header_t ** ip,
			ip4_local_last_check_t * last_check, u8 * error)
{
  ip4_fib_mtrie_leaf_t leaf[2];
  ip4_fib_mtrie_t *mtrie[2];
  const dpo_id_t *dpo[2];
  load_balance_t *lb[2];
  u32 not_last_hit;
  u32 lbi[2];

  not_last_hit = last_check->first;
  not_last_hit |= ip[0]->src_address.as_u32 ^ last_check->src.as_u32;
  not_last_hit |= ip[1]->src_address.as_u32 ^ last_check->src.as_u32;

  vnet_buffer (b[0])->ip.fib_index =
    vnet_buffer (b[0])->sw_if_index[VLIB_TX] != ~0 ?
    vnet_buffer (b[0])->sw_if_index[VLIB_TX] :
    vnet_buffer (b[0])->ip.fib_index;

  vnet_buffer (b[1])->ip.fib_index =
    vnet_buffer (b[1])->sw_if_index[VLIB_TX] != ~0 ?
    vnet_buffer (b[1])->sw_if_index[VLIB_TX] :
    vnet_buffer (b[1])->ip.fib_index;

  if (PREDICT_FALSE (not_last_hit))
    {
      mtrie[0] = &ip4_fib_get (vnet_buffer (b[0])->ip.fib_index)->mtrie;
      mtrie[1] = &ip4_fib_get (vnet_buffer (b[1])->ip.fib_index)->mtrie;

      leaf[0] = ip4_fib_mtrie_lookup_step_one (mtrie[0], &ip[0]->src_address);
      leaf[1] = ip4_fib_mtrie_lookup_step_one (mtrie[1], &ip[1]->src_address);

      leaf[0] = ip4_fib_mtrie_lookup_step (mtrie[0], leaf[0],
					   &ip[0]->src_address, 2);
      leaf[1] = ip4_fib_mtrie_lookup_step (mtrie[1], leaf[1],
					   &ip[1]->src_address, 2);

      leaf[0] = ip4_fib_mtrie_lookup_step (mtrie[0], leaf[0],
					   &ip[0]->src_address, 3);
      leaf[1] = ip4_fib_mtrie_lookup_step (mtrie[1], leaf[1],
					   &ip[1]->src_address, 3);

      lbi[0] = ip4_fib_mtrie_leaf_get_adj_index (leaf[0]);
      lbi[1] = ip4_fib_mtrie_leaf_get_adj_index (leaf[1]);

      vnet_buffer (b[0])->ip.adj_index[VLIB_TX] = lbi[0];
      vnet_buffer (b[0])->ip.adj_index[VLIB_RX] = lbi[0];

      vnet_buffer (b[1])->ip.adj_index[VLIB_TX] = lbi[1];
      vnet_buffer (b[1])->ip.adj_index[VLIB_RX] = lbi[1];

      lb[0] = load_balance_get (lbi[0]);
      lb[1] = load_balance_get (lbi[1]);

      dpo[0] = load_balance_get_bucket_i (lb[0], 0);
      dpo[1] = load_balance_get_bucket_i (lb[1], 0);

      error[0] = ((error[0] == IP4_ERROR_UNKNOWN_PROTOCOL &&
		   dpo[0]->dpoi_type == DPO_RECEIVE) ?
		  IP4_ERROR_SPOOFED_LOCAL_PACKETS : error[0]);
      error[0] = ((error[0] == IP4_ERROR_UNKNOWN_PROTOCOL &&
		   !fib_urpf_check_size (lb[0]->lb_urpf) &&
		   ip[0]->dst_address.as_u32 != 0xFFFFFFFF)
		  ? IP4_ERROR_SRC_LOOKUP_MISS : error[0]);

      error[1] = ((error[1] == IP4_ERROR_UNKNOWN_PROTOCOL &&
		   dpo[1]->dpoi_type == DPO_RECEIVE) ?
		  IP4_ERROR_SPOOFED_LOCAL_PACKETS : error[1]);
      error[1] = ((error[1] == IP4_ERROR_UNKNOWN_PROTOCOL &&
		   !fib_urpf_check_size (lb[1]->lb_urpf) &&
		   ip[1]->dst_address.as_u32 != 0xFFFFFFFF)
		  ? IP4_ERROR_SRC_LOOKUP_MISS : error[1]);

      last_check->src.as_u32 = ip[1]->src_address.as_u32;
      last_check->lbi = lbi[1];
      last_check->error = error[1];
    }
  else
    {
      vnet_buffer (b[0])->ip.adj_index[VLIB_TX] = last_check->lbi;
      vnet_buffer (b[0])->ip.adj_index[VLIB_RX] = last_check->lbi;

      vnet_buffer (b[1])->ip.adj_index[VLIB_TX] = last_check->lbi;
      vnet_buffer (b[1])->ip.adj_index[VLIB_RX] = last_check->lbi;

      error[0] = last_check->error;
      error[1] = last_check->error;
      last_check->first = 0;
    }
}

enum ip_local_packet_type_e
{
  IP_LOCAL_PACKET_TYPE_L4,
  IP_LOCAL_PACKET_TYPE_NAT,
  IP_LOCAL_PACKET_TYPE_FRAG,
};

/**
 * Determine packet type and next node.
 *
 * The expectation is that all packets that are not L4 will skip
 * checksums and source checks.
 */
always_inline u8
ip4_local_classify (vlib_buffer_t * b, ip4_header_t * ip, u16 * next)
{
  ip_lookup_main_t *lm = &ip4_main.lookup_main;

  if (PREDICT_FALSE (ip4_is_fragment (ip)))
    {
      *next = IP_LOCAL_NEXT_REASSEMBLY;
      return IP_LOCAL_PACKET_TYPE_FRAG;
    }
  if (PREDICT_FALSE (b->flags & VNET_BUFFER_F_IS_NATED))
    {
      *next = lm->local_next_by_ip_protocol[ip->protocol];
      return IP_LOCAL_PACKET_TYPE_NAT;
    }

  *next = lm->local_next_by_ip_protocol[ip->protocol];
  return IP_LOCAL_PACKET_TYPE_L4;
}

static inline uword
ip4_local_inline (vlib_main_t * vm,
		  vlib_node_runtime_t * node,
		  vlib_frame_t * frame, int head_of_feature_arc)
{
  u32 *from, n_left_from;
  vlib_node_runtime_t *error_node =
    vlib_node_get_runtime (vm, ip4_input_node.index);
  u16 nexts[VLIB_FRAME_SIZE], *next;
  vlib_buffer_t *bufs[VLIB_FRAME_SIZE], **b;
  ip4_header_t *ip[2];
  u8 error[2], pt[2];

  ip4_local_last_check_t last_check = {
    /*
     * 0.0.0.0 can appear as the source address of an IP packet,
     * as can any other address, hence the need to use the 'first'
     * member to make sure the .lbi is initialised for the first
     * packet.
     */
    .src = {.as_u32 = 0},
    .lbi = ~0,
    .error = IP4_ERROR_UNKNOWN_PROTOCOL,
    .first = 1,
  };

  from = vlib_frame_vector_args (frame);
  n_left_from = frame->n_vectors;

  if (node->flags & VLIB_NODE_FLAG_TRACE)
    ip4_forward_next_trace (vm, node, frame, VLIB_TX);

  vlib_get_buffers (vm, from, bufs, n_left_from);
  b = bufs;
  next = nexts;

  while (n_left_from >= 6)
    {
      u8 not_batch = 0;

      /* Prefetch next iteration. */
      {
	vlib_prefetch_buffer_header (b[4], LOAD);
	vlib_prefetch_buffer_header (b[5], LOAD);

	CLIB_PREFETCH (b[4]->data, CLIB_CACHE_LINE_BYTES, LOAD);
	CLIB_PREFETCH (b[5]->data, CLIB_CACHE_LINE_BYTES, LOAD);
      }

      error[0] = error[1] = IP4_ERROR_UNKNOWN_PROTOCOL;

      ip[0] = vlib_buffer_get_current (b[0]);
      ip[1] = vlib_buffer_get_current (b[1]);

      vnet_buffer (b[0])->l3_hdr_offset = b[0]->current_data;
      vnet_buffer (b[1])->l3_hdr_offset = b[1]->current_data;

      pt[0] = ip4_local_classify (b[0], ip[0], &next[0]);
      pt[1] = ip4_local_classify (b[1], ip[1], &next[1]);

      not_batch = pt[0] ^ pt[1];

      if (head_of_feature_arc == 0 || (pt[0] && not_batch == 0))
	goto skip_checks;

      if (PREDICT_TRUE (not_batch == 0))
	{
	  ip4_local_check_l4_csum_x2 (vm, b, ip, error);
	  ip4_local_check_src_x2 (b, ip, &last_check, error);
	}
      else
	{
	  if (!pt[0])
	    {
	      ip4_local_check_l4_csum (vm, b[0], ip[0], &error[0]);
	      ip4_local_check_src (b[0], ip[0], &last_check, &error[0]);
	    }
	  if (!pt[1])
	    {
	      ip4_local_check_l4_csum (vm, b[1], ip[1], &error[1]);
	      ip4_local_check_src (b[1], ip[1], &last_check, &error[1]);
	    }
	}

    skip_checks:

      ip4_local_set_next_and_error (error_node, b[0], &next[0], error[0],
				    head_of_feature_arc);
      ip4_local_set_next_and_error (error_node, b[1], &next[1], error[1],
				    head_of_feature_arc);

      b += 2;
      next += 2;
      n_left_from -= 2;
    }

  while (n_left_from > 0)
    {
      error[0] = IP4_ERROR_UNKNOWN_PROTOCOL;

      ip[0] = vlib_buffer_get_current (b[0]);
      vnet_buffer (b[0])->l3_hdr_offset = b[0]->current_data;
      pt[0] = ip4_local_classify (b[0], ip[0], &next[0]);

      if (head_of_feature_arc == 0 || pt[0])
	goto skip_check;

      ip4_local_check_l4_csum (vm, b[0], ip[0], &error[0]);
      ip4_local_check_src (b[0], ip[0], &last_check, &error[0]);

    skip_check:

      ip4_local_set_next_and_error (error_node, b[0], &next[0], error[0],
				    head_of_feature_arc);

      b += 1;
      next += 1;
      n_left_from -= 1;
    }

  vlib_buffer_enqueue_to_next (vm, node, from, nexts, frame->n_vectors);
  return frame->n_vectors;
}

VLIB_NODE_FN (ip4_local_node) (vlib_main_t * vm, vlib_node_runtime_t * node,
			       vlib_frame_t * frame)
{
  return ip4_local_inline (vm, node, frame, 1 /* head of feature arc */ );
}

/* *INDENT-OFF* */
VLIB_REGISTER_NODE (ip4_local_node) =
{
  .name = "ip4-local",
  .vector_size = sizeof (u32),
  .format_trace = format_ip4_forward_next_trace,
  .n_next_nodes = IP_LOCAL_N_NEXT,
  .next_nodes =
  {
    [IP_LOCAL_NEXT_DROP] = "ip4-drop",
    [IP_LOCAL_NEXT_PUNT] = "ip4-punt",
    [IP_LOCAL_NEXT_UDP_LOOKUP] = "ip4-udp-lookup",
    [IP_LOCAL_NEXT_ICMP] = "ip4-icmp-input",
    [IP_LOCAL_NEXT_REASSEMBLY] = "ip4-reassembly",
  },
};
/* *INDENT-ON* */


VLIB_NODE_FN (ip4_local_end_of_arc_node) (vlib_main_t * vm,
					  vlib_node_runtime_t * node,
					  vlib_frame_t * frame)
{
  return ip4_local_inline (vm, node, frame, 0 /* head of feature arc */ );
}

/* *INDENT-OFF* */
VLIB_REGISTER_NODE (ip4_local_end_of_arc_node) = {
  .name = "ip4-local-end-of-arc",
  .vector_size = sizeof (u32),

  .format_trace = format_ip4_forward_next_trace,
  .sibling_of = "ip4-local",
};

VNET_FEATURE_INIT (ip4_local_end_of_arc, static) = {
  .arc_name = "ip4-local",
  .node_name = "ip4-local-end-of-arc",
  .runs_before = 0, /* not before any other features */
};
/* *INDENT-ON* */

#ifndef CLIB_MARCH_VARIANT
void
ip4_register_protocol (u32 protocol, u32 node_index)
{
  vlib_main_t *vm = vlib_get_main ();
  ip4_main_t *im = &ip4_main;
  ip_lookup_main_t *lm = &im->lookup_main;

  ASSERT (protocol < ARRAY_LEN (lm->local_next_by_ip_protocol));
  lm->local_next_by_ip_protocol[protocol] =
    vlib_node_add_next (vm, ip4_local_node.index, node_index);
}
#endif

static clib_error_t *
show_ip_local_command_fn (vlib_main_t * vm,
			  unformat_input_t * input, vlib_cli_command_t * cmd)
{
  ip4_main_t *im = &ip4_main;
  ip_lookup_main_t *lm = &im->lookup_main;
  int i;

  vlib_cli_output (vm, "Protocols handled by ip4_local");
  for (i = 0; i < ARRAY_LEN (lm->local_next_by_ip_protocol); i++)
    {
      if (lm->local_next_by_ip_protocol[i] != IP_LOCAL_NEXT_PUNT)
	{
	  u32 node_index = vlib_get_node (vm,
					  ip4_local_node.index)->
	    next_nodes[lm->local_next_by_ip_protocol[i]];
	  vlib_cli_output (vm, "%d: %U", i, format_vlib_node_name, vm,
			   node_index);
	}
    }
  return 0;
}



/*?
 * Display the set of protocols handled by the local IPv4 stack.
 *
 * @cliexpar
 * Example of how to display local protocol table:
 * @cliexstart{show ip local}
 * Protocols handled by ip4_local
 * 1
 * 17
 * 47
 * @cliexend
?*/
/* *INDENT-OFF* */
VLIB_CLI_COMMAND (show_ip_local, static) =
{
  .path = "show ip local",
  .function = show_ip_local_command_fn,
  .short_help = "show ip local",
};
/* *INDENT-ON* */

always_inline uword
ip4_arp_inline (vlib_main_t * vm,
		vlib_node_runtime_t * node,
		vlib_frame_t * frame, int is_glean)
{
  vnet_main_t *vnm = vnet_get_main ();
  ip4_main_t *im = &ip4_main;
  ip_lookup_main_t *lm = &im->lookup_main;
  u32 *from, *to_next_drop;
  uword n_left_from, n_left_to_next_drop, next_index;
  u32 thread_index = vm->thread_index;
  u64 seed;

  if (node->flags & VLIB_NODE_FLAG_TRACE)
    ip4_forward_next_trace (vm, node, frame, VLIB_TX);

  seed = throttle_seed (&im->arp_throttle, thread_index, vlib_time_now (vm));

  from = vlib_frame_vector_args (frame);
  n_left_from = frame->n_vectors;
  next_index = node->cached_next_index;
  if (next_index == IP4_ARP_NEXT_DROP)
    next_index = IP4_ARP_N_NEXT;	/* point to first interface */

  while (n_left_from > 0)
    {
      vlib_get_next_frame (vm, node, IP4_ARP_NEXT_DROP,
			   to_next_drop, n_left_to_next_drop);

      while (n_left_from > 0 && n_left_to_next_drop > 0)
	{
	  u32 pi0, bi0, adj_index0, sw_if_index0;
	  ip_adjacency_t *adj0;
	  vlib_buffer_t *p0, *b0;
	  ip4_address_t resolve0;
	  ethernet_arp_header_t *h0;
	  vnet_hw_interface_t *hw_if0;
	  u64 r0;

	  pi0 = from[0];
	  p0 = vlib_get_buffer (vm, pi0);

	  from += 1;
	  n_left_from -= 1;
	  to_next_drop[0] = pi0;
	  to_next_drop += 1;
	  n_left_to_next_drop -= 1;

	  adj_index0 = vnet_buffer (p0)->ip.adj_index[VLIB_TX];
	  adj0 = adj_get (adj_index0);

	  if (is_glean)
	    {
	      /* resolve the packet's destination */
	      ip4_header_t *ip0 = vlib_buffer_get_current (p0);
	      resolve0 = ip0->dst_address;
	    }
	  else
	    {
	      /* resolve the incomplete adj */
	      resolve0 = adj0->sub_type.nbr.next_hop.ip4;
	    }

	  /* combine the address and interface for the hash key */
	  sw_if_index0 = adj0->rewrite_header.sw_if_index;
	  r0 = (u64) resolve0.data_u32 << 32;
	  r0 |= sw_if_index0;

	  if (throttle_check (&im->arp_throttle, thread_index, r0, seed))
	    {
	      p0->error = node->errors[IP4_ARP_ERROR_THROTTLED];
	      continue;
	    }

	  /*
	   * the adj has been updated to a rewrite but the node the DPO that got
	   * us here hasn't - yet. no big deal. we'll drop while we wait.
	   */
	  if (IP_LOOKUP_NEXT_REWRITE == adj0->lookup_next_index)
	    {
	      p0->error = node->errors[IP4_ARP_ERROR_RESOLVED];
	      continue;
	    }

	  /*
	   * Can happen if the control-plane is programming tables
	   * with traffic flowing; at least that's today's lame excuse.
	   */
	  if ((is_glean && adj0->lookup_next_index != IP_LOOKUP_NEXT_GLEAN)
	      || (!is_glean && adj0->lookup_next_index != IP_LOOKUP_NEXT_ARP))
	    {
	      p0->error = node->errors[IP4_ARP_ERROR_NON_ARP_ADJ];
	      continue;
	    }
	  /* Send ARP request. */
	  h0 =
	    vlib_packet_template_get_packet (vm,
					     &im->ip4_arp_request_packet_template,
					     &bi0);
	  b0 = vlib_get_buffer (vm, bi0);

	  /* copy the persistent fields from the original */
	  clib_memcpy_fast (b0->opaque2, p0->opaque2, sizeof (p0->opaque2));

	  /* Seems we're out of buffers */
	  if (PREDICT_FALSE (!h0))
	    {
	      p0->error = node->errors[IP4_ARP_ERROR_NO_BUFFERS];
	      continue;
	    }

	  /* Add rewrite/encap string for ARP packet. */
	  vnet_rewrite_one_header (adj0[0], h0, sizeof (ethernet_header_t));

	  hw_if0 = vnet_get_sup_hw_interface (vnm, sw_if_index0);

	  /* Src ethernet address in ARP header. */
	  mac_address_from_bytes (&h0->ip4_over_ethernet[0].mac,
				  hw_if0->hw_address);
	  if (is_glean)
	    {
	      /* The interface's source address is stashed in the Glean Adj */
	      h0->ip4_over_ethernet[0].ip4 =
		adj0->sub_type.glean.receive_addr.ip4;
	    }
	  else
	    {
	      /* Src IP address in ARP header. */
	      if (ip4_src_address_for_packet (lm, sw_if_index0,
					      &h0->ip4_over_ethernet[0].ip4))
		{
		  /* No source address available */
		  p0->error = node->errors[IP4_ARP_ERROR_NO_SOURCE_ADDRESS];
		  vlib_buffer_free (vm, &bi0, 1);
		  continue;
		}
	    }
	  h0->ip4_over_ethernet[1].ip4 = resolve0;

	  p0->error = node->errors[IP4_ARP_ERROR_REQUEST_SENT];

	  vlib_buffer_copy_trace_flag (vm, p0, bi0);
	  VLIB_BUFFER_TRACE_TRAJECTORY_INIT (b0);
	  vnet_buffer (b0)->sw_if_index[VLIB_TX] = sw_if_index0;

	  vlib_buffer_advance (b0, -adj0->rewrite_header.data_bytes);

	  vlib_set_next_frame_buffer (vm, node,
				      adj0->rewrite_header.next_index, bi0);
	}

      vlib_put_next_frame (vm, node, IP4_ARP_NEXT_DROP, n_left_to_next_drop);
    }

  return frame->n_vectors;
}

VLIB_NODE_FN (ip4_arp_node) (vlib_main_t * vm, vlib_node_runtime_t * node,
			     vlib_frame_t * frame)
{
  return (ip4_arp_inline (vm, node, frame, 0));
}

VLIB_NODE_FN (ip4_glean_node) (vlib_main_t * vm, vlib_node_runtime_t * node,
			       vlib_frame_t * frame)
{
  return (ip4_arp_inline (vm, node, frame, 1));
}

static char *ip4_arp_error_strings[] = {
  [IP4_ARP_ERROR_THROTTLED] = "ARP requests throttled",
  [IP4_ARP_ERROR_RESOLVED] = "ARP requests resolved",
  [IP4_ARP_ERROR_NO_BUFFERS] = "ARP requests out of buffer",
  [IP4_ARP_ERROR_REQUEST_SENT] = "ARP requests sent",
  [IP4_ARP_ERROR_NON_ARP_ADJ] = "ARPs to non-ARP adjacencies",
  [IP4_ARP_ERROR_NO_SOURCE_ADDRESS] = "no source address for ARP request",
};

/* *INDENT-OFF* */
VLIB_REGISTER_NODE (ip4_arp_node) =
{
  .name = "ip4-arp",
  .vector_size = sizeof (u32),
  .format_trace = format_ip4_forward_next_trace,
  .n_errors = ARRAY_LEN (ip4_arp_error_strings),
  .error_strings = ip4_arp_error_strings,
  .n_next_nodes = IP4_ARP_N_NEXT,
  .next_nodes =
  {
    [IP4_ARP_NEXT_DROP] = "error-drop",
  },
};

VLIB_REGISTER_NODE (ip4_glean_node) =
{
  .name = "ip4-glean",
  .vector_size = sizeof (u32),
  .format_trace = format_ip4_forward_next_trace,
  .n_errors = ARRAY_LEN (ip4_arp_error_strings),
  .error_strings = ip4_arp_error_strings,
  .n_next_nodes = IP4_ARP_N_NEXT,
  .next_nodes = {
  [IP4_ARP_NEXT_DROP] = "error-drop",
  },
};
/* *INDENT-ON* */

#define foreach_notrace_ip4_arp_error           \
_(THROTTLED)                                    \
_(RESOLVED)                                     \
_(NO_BUFFERS)                                   \
_(REQUEST_SENT)                                 \
_(NON_ARP_ADJ)                                  \
_(NO_SOURCE_ADDRESS)

static clib_error_t *
arp_notrace_init (vlib_main_t * vm)
{
  vlib_node_runtime_t *rt = vlib_node_get_runtime (vm, ip4_arp_node.index);

  /* don't trace ARP request packets */
#define _(a)                                    \
    vnet_pcap_drop_trace_filter_add_del         \
        (rt->errors[IP4_ARP_ERROR_##a],         \
         1 /* is_add */);
  foreach_notrace_ip4_arp_error;
#undef _
  return 0;
}

VLIB_INIT_FUNCTION (arp_notrace_init);


#ifndef CLIB_MARCH_VARIANT
/* Send an ARP request to see if given destination is reachable on given interface. */
clib_error_t *
ip4_probe_neighbor (vlib_main_t * vm, ip4_address_t * dst, u32 sw_if_index,
		    u8 refresh)
{
  vnet_main_t *vnm = vnet_get_main ();
  ip4_main_t *im = &ip4_main;
  ethernet_arp_header_t *h;
  ip4_address_t *src;
  ip_interface_address_t *ia;
  ip_adjacency_t *adj;
  vnet_hw_interface_t *hi;
  vnet_sw_interface_t *si;
  vlib_buffer_t *b;
  adj_index_t ai;
  u32 bi = 0;
  u8 unicast_rewrite = 0;

  si = vnet_get_sw_interface (vnm, sw_if_index);

  if (!(si->flags & VNET_SW_INTERFACE_FLAG_ADMIN_UP))
    {
      return clib_error_return (0, "%U: interface %U down",
				format_ip4_address, dst,
				format_vnet_sw_if_index_name, vnm,
				sw_if_index);
    }

  src =
    ip4_interface_address_matching_destination (im, dst, sw_if_index, &ia);
  if (!src)
    {
      vnm->api_errno = VNET_API_ERROR_NO_MATCHING_INTERFACE;
      return clib_error_return
	(0,
	 "no matching interface address for destination %U (interface %U)",
	 format_ip4_address, dst, format_vnet_sw_if_index_name, vnm,
	 sw_if_index);
    }

  h = vlib_packet_template_get_packet (vm,
				       &im->ip4_arp_request_packet_template,
				       &bi);

  if (!h)
    return clib_error_return (0, "ARP request packet allocation failed");

  hi = vnet_get_sup_hw_interface (vnm, sw_if_index);
  if (PREDICT_FALSE (!hi->hw_address))
    {
      return clib_error_return (0, "%U: interface %U do not support ip probe",
				format_ip4_address, dst,
				format_vnet_sw_if_index_name, vnm,
				sw_if_index);
    }

  mac_address_from_bytes (&h->ip4_over_ethernet[0].mac, hi->hw_address);

  h->ip4_over_ethernet[0].ip4 = src[0];
  h->ip4_over_ethernet[1].ip4 = dst[0];

  b = vlib_get_buffer (vm, bi);
  vnet_buffer (b)->sw_if_index[VLIB_RX] =
    vnet_buffer (b)->sw_if_index[VLIB_TX] = sw_if_index;

  ip46_address_t nh = {
    .ip4 = *dst,
  };

  ai = adj_nbr_add_or_lock (FIB_PROTOCOL_IP4,
			    VNET_LINK_IP4, &nh, sw_if_index);
  adj = adj_get (ai);

  /* Peer has been previously resolved, retrieve glean adj instead */
  if (adj->lookup_next_index == IP_LOOKUP_NEXT_REWRITE)
    {
      if (refresh)
	unicast_rewrite = 1;
      else
	{
	  adj_unlock (ai);
	  ai = adj_glean_add_or_lock (FIB_PROTOCOL_IP4,
				      VNET_LINK_IP4, sw_if_index, &nh);
	  adj = adj_get (ai);
	}
    }

  /* Add encapsulation string for software interface (e.g. ethernet header). */
  vnet_rewrite_one_header (adj[0], h, sizeof (ethernet_header_t));
  if (unicast_rewrite)
    {
      u16 *etype = vlib_buffer_get_current (b) - 2;
      etype[0] = clib_host_to_net_u16 (ETHERNET_TYPE_ARP);
    }
  vlib_buffer_advance (b, -adj->rewrite_header.data_bytes);

  {
    vlib_frame_t *f = vlib_get_frame_to_node (vm, hi->output_node_index);
    u32 *to_next = vlib_frame_vector_args (f);
    to_next[0] = bi;
    f->n_vectors = 1;
    vlib_put_frame_to_node (vm, hi->output_node_index, f);
  }

  adj_unlock (ai);
  return /* no error */ 0;
}
#endif

typedef enum
{
  IP4_REWRITE_NEXT_DROP,
  IP4_REWRITE_NEXT_ICMP_ERROR,
  IP4_REWRITE_NEXT_FRAGMENT,
  IP4_REWRITE_N_NEXT		/* Last */
} ip4_rewrite_next_t;

/**
 * This bits of an IPv4 address to mask to construct a multicast
 * MAC address
 */
#if CLIB_ARCH_IS_BIG_ENDIAN
#define IP4_MCAST_ADDR_MASK 0x007fffff
#else
#define IP4_MCAST_ADDR_MASK 0xffff7f00
#endif

always_inline void
ip4_mtu_check (vlib_buffer_t * b, u16 packet_len,
	       u16 adj_packet_bytes, bool df, u16 * next, u32 * error)
{
  if (packet_len > adj_packet_bytes)
    {
      *error = IP4_ERROR_MTU_EXCEEDED;
      if (df)
	{
	  icmp4_error_set_vnet_buffer
	    (b, ICMP4_destination_unreachable,
	     ICMP4_destination_unreachable_fragmentation_needed_and_dont_fragment_set,
	     adj_packet_bytes);
	  *next = IP4_REWRITE_NEXT_ICMP_ERROR;
	}
      else
	{
	  /* IP fragmentation */
	  ip_frag_set_vnet_buffer (b, adj_packet_bytes,
				   IP4_FRAG_NEXT_IP4_REWRITE, 0);
	  *next = IP4_REWRITE_NEXT_FRAGMENT;
	}
    }
}

/* Decrement TTL & update checksum.
   Works either endian, so no need for byte swap. */
static_always_inline void
ip4_ttl_and_checksum_check (vlib_buffer_t * b, ip4_header_t * ip, u16 * next,
			    u32 * error)
{
  i32 ttl;
  u32 checksum;
  if (PREDICT_FALSE (b->flags & VNET_BUFFER_F_LOCALLY_ORIGINATED))
    {
      b->flags &= ~VNET_BUFFER_F_LOCALLY_ORIGINATED;
      return;
    }

  ttl = ip->ttl;

  /* Input node should have reject packets with ttl 0. */
  ASSERT (ip->ttl > 0);

  checksum = ip->checksum + clib_host_to_net_u16 (0x0100);
  checksum += checksum >= 0xffff;

  ip->checksum = checksum;
  ttl -= 1;
  ip->ttl = ttl;

  /*
   * If the ttl drops below 1 when forwarding, generate
   * an ICMP response.
   */
  if (PREDICT_FALSE (ttl <= 0))
    {
      *error = IP4_ERROR_TIME_EXPIRED;
      vnet_buffer (b)->sw_if_index[VLIB_TX] = (u32) ~ 0;
      icmp4_error_set_vnet_buffer (b, ICMP4_time_exceeded,
				   ICMP4_time_exceeded_ttl_exceeded_in_transit,
				   0);
      *next = IP4_REWRITE_NEXT_ICMP_ERROR;
    }

  /* Verify checksum. */
  ASSERT ((ip->checksum == ip4_header_checksum (ip)) ||
	  (b->flags & VNET_BUFFER_F_OFFLOAD_IP_CKSUM));
}


always_inline uword
ip4_rewrite_inline_with_gso (vlib_main_t * vm,
			     vlib_node_runtime_t * node,
			     vlib_frame_t * frame,
			     int do_counters, int is_midchain, int is_mcast,
			     int do_gso)
{
  ip_lookup_main_t *lm = &ip4_main.lookup_main;
  u32 *from = vlib_frame_vector_args (frame);
  vlib_buffer_t *bufs[VLIB_FRAME_SIZE], **b;
  u16 nexts[VLIB_FRAME_SIZE], *next;
  u32 n_left_from;
  vlib_node_runtime_t *error_node =
    vlib_node_get_runtime (vm, ip4_input_node.index);

  n_left_from = frame->n_vectors;
  u32 thread_index = vm->thread_index;

  vlib_get_buffers (vm, from, bufs, n_left_from);
  clib_memset_u16 (nexts, IP4_REWRITE_NEXT_DROP, n_left_from);

  if (n_left_from >= 6)
    {
      int i;
      for (i = 2; i < 6; i++)
	vlib_prefetch_buffer_header (bufs[i], LOAD);
    }

  next = nexts;
  b = bufs;
  while (n_left_from >= 8)
    {
      ip_adjacency_t *adj0, *adj1;
      ip4_header_t *ip0, *ip1;
      u32 rw_len0, error0, adj_index0;
      u32 rw_len1, error1, adj_index1;
      u32 tx_sw_if_index0, tx_sw_if_index1;
      u8 *p;

      vlib_prefetch_buffer_header (b[6], LOAD);
      vlib_prefetch_buffer_header (b[7], LOAD);

      adj_index0 = vnet_buffer (b[0])->ip.adj_index[VLIB_TX];
      adj_index1 = vnet_buffer (b[1])->ip.adj_index[VLIB_TX];

      /*
       * pre-fetch the per-adjacency counters
       */
      if (do_counters)
	{
	  vlib_prefetch_combined_counter (&adjacency_counters,
					  thread_index, adj_index0);
	  vlib_prefetch_combined_counter (&adjacency_counters,
					  thread_index, adj_index1);
	}

      ip0 = vlib_buffer_get_current (b[0]);
      ip1 = vlib_buffer_get_current (b[1]);

      error0 = error1 = IP4_ERROR_NONE;

      ip4_ttl_and_checksum_check (b[0], ip0, next + 0, &error0);
      ip4_ttl_and_checksum_check (b[1], ip1, next + 1, &error1);

      /* Rewrite packet header and updates lengths. */
      adj0 = adj_get (adj_index0);
      adj1 = adj_get (adj_index1);

      /* Worth pipelining. No guarantee that adj0,1 are hot... */
      rw_len0 = adj0[0].rewrite_header.data_bytes;
      rw_len1 = adj1[0].rewrite_header.data_bytes;
      vnet_buffer (b[0])->ip.save_rewrite_length = rw_len0;
      vnet_buffer (b[1])->ip.save_rewrite_length = rw_len1;

      p = vlib_buffer_get_current (b[2]);
      CLIB_PREFETCH (p - CLIB_CACHE_LINE_BYTES, CLIB_CACHE_LINE_BYTES, STORE);
      CLIB_PREFETCH (p, CLIB_CACHE_LINE_BYTES, LOAD);

      p = vlib_buffer_get_current (b[3]);
      CLIB_PREFETCH (p - CLIB_CACHE_LINE_BYTES, CLIB_CACHE_LINE_BYTES, STORE);
      CLIB_PREFETCH (p, CLIB_CACHE_LINE_BYTES, LOAD);

      /* Check MTU of outgoing interface. */
      u16 ip0_len = clib_net_to_host_u16 (ip0->length);
      u16 ip1_len = clib_net_to_host_u16 (ip1->length);

      if (do_gso && (b[0]->flags & VNET_BUFFER_F_GSO))
	ip0_len = gso_mtu_sz (b[0]);
      if (do_gso && (b[1]->flags & VNET_BUFFER_F_GSO))
	ip1_len = gso_mtu_sz (b[1]);

      ip4_mtu_check (b[0], ip0_len,
		     adj0[0].rewrite_header.max_l3_packet_bytes,
		     ip0->flags_and_fragment_offset &
		     clib_host_to_net_u16 (IP4_HEADER_FLAG_DONT_FRAGMENT),
		     next + 0, &error0);
      ip4_mtu_check (b[1], ip1_len,
		     adj1[0].rewrite_header.max_l3_packet_bytes,
		     ip1->flags_and_fragment_offset &
		     clib_host_to_net_u16 (IP4_HEADER_FLAG_DONT_FRAGMENT),
		     next + 1, &error1);

      if (is_mcast)
	{
	  error0 = ((adj0[0].rewrite_header.sw_if_index ==
		     vnet_buffer (b[0])->sw_if_index[VLIB_RX]) ?
		    IP4_ERROR_SAME_INTERFACE : error0);
	  error1 = ((adj1[0].rewrite_header.sw_if_index ==
		     vnet_buffer (b[1])->sw_if_index[VLIB_RX]) ?
		    IP4_ERROR_SAME_INTERFACE : error1);
	}

      /* Don't adjust the buffer for ttl issue; icmp-error node wants
       * to see the IP header */
      if (PREDICT_TRUE (error0 == IP4_ERROR_NONE))
	{
	  u32 next_index = adj0[0].rewrite_header.next_index;
	  b[0]->current_data -= rw_len0;
	  b[0]->current_length += rw_len0;
	  tx_sw_if_index0 = adj0[0].rewrite_header.sw_if_index;
	  vnet_buffer (b[0])->sw_if_index[VLIB_TX] = tx_sw_if_index0;

	  if (PREDICT_FALSE
	      (adj0[0].rewrite_header.flags & VNET_REWRITE_HAS_FEATURES))
	    vnet_feature_arc_start (lm->output_feature_arc_index,
				    tx_sw_if_index0, &next_index, b[0]);
	  next[0] = next_index;
	}
      else
	{
	  b[0]->error = error_node->errors[error0];
	}
      if (PREDICT_TRUE (error1 == IP4_ERROR_NONE))
	{
	  u32 next_index = adj1[0].rewrite_header.next_index;
	  b[1]->current_data -= rw_len1;
	  b[1]->current_length += rw_len1;

	  tx_sw_if_index1 = adj1[0].rewrite_header.sw_if_index;
	  vnet_buffer (b[1])->sw_if_index[VLIB_TX] = tx_sw_if_index1;

	  if (PREDICT_FALSE
	      (adj1[0].rewrite_header.flags & VNET_REWRITE_HAS_FEATURES))
	    vnet_feature_arc_start (lm->output_feature_arc_index,
				    tx_sw_if_index1, &next_index, b[1]);
	  next[1] = next_index;
	}
      else
	{
	  b[1]->error = error_node->errors[error1];
	}
      if (is_midchain)
	{
	  calc_checksums (vm, b[0]);
	  calc_checksums (vm, b[1]);
	}
      /* Guess we are only writing on simple Ethernet header. */
      vnet_rewrite_two_headers (adj0[0], adj1[0],
				ip0, ip1, sizeof (ethernet_header_t));

      /*
       * Bump the per-adjacency counters
       */
      if (do_counters)
	{
	  vlib_increment_combined_counter
	    (&adjacency_counters,
	     thread_index,
	     adj_index0, 1, vlib_buffer_length_in_chain (vm, b[0]) + rw_len0);

	  vlib_increment_combined_counter
	    (&adjacency_counters,
	     thread_index,
	     adj_index1, 1, vlib_buffer_length_in_chain (vm, b[1]) + rw_len1);
	}

      if (is_midchain)
	{
	  if (adj0->sub_type.midchain.fixup_func)
	    adj0->sub_type.midchain.fixup_func
	      (vm, adj0, b[0], adj0->sub_type.midchain.fixup_data);
	  if (adj1->sub_type.midchain.fixup_func)
	    adj1->sub_type.midchain.fixup_func
	      (vm, adj1, b[1], adj1->sub_type.midchain.fixup_data);
	}

      if (is_mcast)
	{
	  /*
	   * copy bytes from the IP address into the MAC rewrite
	   */
	  vnet_ip_mcast_fixup_header (IP4_MCAST_ADDR_MASK,
				      adj0->rewrite_header.dst_mcast_offset,
				      &ip0->dst_address.as_u32, (u8 *) ip0);
	  vnet_ip_mcast_fixup_header (IP4_MCAST_ADDR_MASK,
				      adj1->rewrite_header.dst_mcast_offset,
				      &ip1->dst_address.as_u32, (u8 *) ip1);
	}

      next += 2;
      b += 2;
      n_left_from -= 2;
    }

  while (n_left_from > 0)
    {
      ip_adjacency_t *adj0;
      ip4_header_t *ip0;
      u32 rw_len0, adj_index0, error0;
      u32 tx_sw_if_index0;

      adj_index0 = vnet_buffer (b[0])->ip.adj_index[VLIB_TX];

      adj0 = adj_get (adj_index0);

      if (do_counters)
	vlib_prefetch_combined_counter (&adjacency_counters,
					thread_index, adj_index0);

      ip0 = vlib_buffer_get_current (b[0]);

      error0 = IP4_ERROR_NONE;

      ip4_ttl_and_checksum_check (b[0], ip0, next + 0, &error0);


      /* Update packet buffer attributes/set output interface. */
      rw_len0 = adj0[0].rewrite_header.data_bytes;
      vnet_buffer (b[0])->ip.save_rewrite_length = rw_len0;

      /* Check MTU of outgoing interface. */
      u16 ip0_len = clib_net_to_host_u16 (ip0->length);
      if (do_gso && (b[0]->flags & VNET_BUFFER_F_GSO))
	ip0_len = gso_mtu_sz (b[0]);

      ip4_mtu_check (b[0], ip0_len,
		     adj0[0].rewrite_header.max_l3_packet_bytes,
		     ip0->flags_and_fragment_offset &
		     clib_host_to_net_u16 (IP4_HEADER_FLAG_DONT_FRAGMENT),
		     next + 0, &error0);

      if (is_mcast)
	{
	  error0 = ((adj0[0].rewrite_header.sw_if_index ==
		     vnet_buffer (b[0])->sw_if_index[VLIB_RX]) ?
		    IP4_ERROR_SAME_INTERFACE : error0);
	}

      /* Don't adjust the buffer for ttl issue; icmp-error node wants
       * to see the IP header */
      if (PREDICT_TRUE (error0 == IP4_ERROR_NONE))
	{
	  u32 next_index = adj0[0].rewrite_header.next_index;
	  b[0]->current_data -= rw_len0;
	  b[0]->current_length += rw_len0;
	  tx_sw_if_index0 = adj0[0].rewrite_header.sw_if_index;
	  vnet_buffer (b[0])->sw_if_index[VLIB_TX] = tx_sw_if_index0;

	  if (PREDICT_FALSE
	      (adj0[0].rewrite_header.flags & VNET_REWRITE_HAS_FEATURES))
	    vnet_feature_arc_start (lm->output_feature_arc_index,
				    tx_sw_if_index0, &next_index, b[0]);
	  next[0] = next_index;
	}
      else
	{
	  b[0]->error = error_node->errors[error0];
	}
      if (is_midchain)
	{
	  calc_checksums (vm, b[0]);
	}
      /* Guess we are only writing on simple Ethernet header. */
      vnet_rewrite_one_header (adj0[0], ip0, sizeof (ethernet_header_t));

      if (do_counters)
	vlib_increment_combined_counter
	  (&adjacency_counters,
	   thread_index, adj_index0, 1,
	   vlib_buffer_length_in_chain (vm, b[0]) + rw_len0);

      if (is_midchain)
	{
	  if (adj0->sub_type.midchain.fixup_func)
	    adj0->sub_type.midchain.fixup_func
	      (vm, adj0, b[0], adj0->sub_type.midchain.fixup_data);
	}

      if (is_mcast)
	{
	  /*
	   * copy bytes from the IP address into the MAC rewrite
	   */
	  vnet_ip_mcast_fixup_header (IP4_MCAST_ADDR_MASK,
				      adj0->rewrite_header.dst_mcast_offset,
				      &ip0->dst_address.as_u32, (u8 *) ip0);
	}

      next += 1;
      b += 1;
      n_left_from -= 1;
    }


  /* Need to do trace after rewrites to pick up new packet data. */
  if (node->flags & VLIB_NODE_FLAG_TRACE)
    ip4_forward_next_trace (vm, node, frame, VLIB_TX);

  vlib_buffer_enqueue_to_next (vm, node, from, nexts, frame->n_vectors);
  return frame->n_vectors;
}

always_inline uword
ip4_rewrite_inline (vlib_main_t * vm,
		    vlib_node_runtime_t * node,
		    vlib_frame_t * frame,
		    int do_counters, int is_midchain, int is_mcast)
{
  vnet_main_t *vnm = vnet_get_main ();
  if (PREDICT_FALSE (vnm->interface_main.gso_interface_count > 0))
    return ip4_rewrite_inline_with_gso (vm, node, frame, do_counters,
					is_midchain, is_mcast,
					1 /* do_gso */ );
  else
    return ip4_rewrite_inline_with_gso (vm, node, frame, do_counters,
					is_midchain, is_mcast,
					0 /* no do_gso */ );
}


/** @brief IPv4 rewrite node.
    @node ip4-rewrite

    This is the IPv4 transit-rewrite node: decrement TTL, fix the ipv4
    header checksum, fetch the ip adjacency, check the outbound mtu,
    apply the adjacency rewrite, and send pkts to the adjacency
    rewrite header's rewrite_next_index.

    @param vm vlib_main_t corresponding to the current thread
    @param node vlib_node_runtime_t
    @param frame vlib_frame_t whose contents should be dispatched

    @par Graph mechanics: buffer metadata, next index usage

    @em Uses:
    - <code>vnet_buffer(b)->ip.adj_index[VLIB_TX]</code>
        - the rewrite adjacency index
    - <code>adj->lookup_next_index</code>
        - Must be IP_LOOKUP_NEXT_REWRITE or IP_LOOKUP_NEXT_ARP, otherwise
          the packet will be dropped.
    - <code>adj->rewrite_header</code>
        - Rewrite string length, rewrite string, next_index

    @em Sets:
    - <code>b->current_data, b->current_length</code>
        - Updated net of applying the rewrite string

    <em>Next Indices:</em>
    - <code> adj->rewrite_header.next_index </code>
      or @c ip4-drop
*/

VLIB_NODE_FN (ip4_rewrite_node) (vlib_main_t * vm, vlib_node_runtime_t * node,
				 vlib_frame_t * frame)
{
  if (adj_are_counters_enabled ())
    return ip4_rewrite_inline (vm, node, frame, 1, 0, 0);
  else
    return ip4_rewrite_inline (vm, node, frame, 0, 0, 0);
}

VLIB_NODE_FN (ip4_rewrite_bcast_node) (vlib_main_t * vm,
				       vlib_node_runtime_t * node,
				       vlib_frame_t * frame)
{
  if (adj_are_counters_enabled ())
    return ip4_rewrite_inline (vm, node, frame, 1, 0, 0);
  else
    return ip4_rewrite_inline (vm, node, frame, 0, 0, 0);
}

VLIB_NODE_FN (ip4_midchain_node) (vlib_main_t * vm,
				  vlib_node_runtime_t * node,
				  vlib_frame_t * frame)
{
  if (adj_are_counters_enabled ())
    return ip4_rewrite_inline (vm, node, frame, 1, 1, 0);
  else
    return ip4_rewrite_inline (vm, node, frame, 0, 1, 0);
}

VLIB_NODE_FN (ip4_rewrite_mcast_node) (vlib_main_t * vm,
				       vlib_node_runtime_t * node,
				       vlib_frame_t * frame)
{
  if (adj_are_counters_enabled ())
    return ip4_rewrite_inline (vm, node, frame, 1, 0, 1);
  else
    return ip4_rewrite_inline (vm, node, frame, 0, 0, 1);
}

VLIB_NODE_FN (ip4_mcast_midchain_node) (vlib_main_t * vm,
					vlib_node_runtime_t * node,
					vlib_frame_t * frame)
{
  if (adj_are_counters_enabled ())
    return ip4_rewrite_inline (vm, node, frame, 1, 1, 1);
  else
    return ip4_rewrite_inline (vm, node, frame, 0, 1, 1);
}

/* *INDENT-OFF* */
VLIB_REGISTER_NODE (ip4_rewrite_node) = {
  .name = "ip4-rewrite",
  .vector_size = sizeof (u32),

  .format_trace = format_ip4_rewrite_trace,

  .n_next_nodes = IP4_REWRITE_N_NEXT,
  .next_nodes = {
    [IP4_REWRITE_NEXT_DROP] = "ip4-drop",
    [IP4_REWRITE_NEXT_ICMP_ERROR] = "ip4-icmp-error",
    [IP4_REWRITE_NEXT_FRAGMENT] = "ip4-frag",
  },
};

VLIB_REGISTER_NODE (ip4_rewrite_bcast_node) = {
  .name = "ip4-rewrite-bcast",
  .vector_size = sizeof (u32),

  .format_trace = format_ip4_rewrite_trace,
  .sibling_of = "ip4-rewrite",
};

VLIB_REGISTER_NODE (ip4_rewrite_mcast_node) = {
  .name = "ip4-rewrite-mcast",
  .vector_size = sizeof (u32),

  .format_trace = format_ip4_rewrite_trace,
  .sibling_of = "ip4-rewrite",
};

VLIB_REGISTER_NODE (ip4_mcast_midchain_node) = {
  .name = "ip4-mcast-midchain",
  .vector_size = sizeof (u32),

  .format_trace = format_ip4_rewrite_trace,
  .sibling_of = "ip4-rewrite",
};

VLIB_REGISTER_NODE (ip4_midchain_node) = {
  .name = "ip4-midchain",
  .vector_size = sizeof (u32),
  .format_trace = format_ip4_forward_next_trace,
  .sibling_of =  "ip4-rewrite",
};
/* *INDENT-ON */

static int
ip4_lookup_validate (ip4_address_t * a, u32 fib_index0)
{
  ip4_fib_mtrie_t *mtrie0;
  ip4_fib_mtrie_leaf_t leaf0;
  u32 lbi0;

  mtrie0 = &ip4_fib_get (fib_index0)->mtrie;

  leaf0 = ip4_fib_mtrie_lookup_step_one (mtrie0, a);
  leaf0 = ip4_fib_mtrie_lookup_step (mtrie0, leaf0, a, 2);
  leaf0 = ip4_fib_mtrie_lookup_step (mtrie0, leaf0, a, 3);

  lbi0 = ip4_fib_mtrie_leaf_get_adj_index (leaf0);

  return lbi0 == ip4_fib_table_lookup_lb (ip4_fib_get (fib_index0), a);
}

static clib_error_t *
test_lookup_command_fn (vlib_main_t * vm,
			unformat_input_t * input, vlib_cli_command_t * cmd)
{
  ip4_fib_t *fib;
  u32 table_id = 0;
  f64 count = 1;
  u32 n;
  int i;
  ip4_address_t ip4_base_address;
  u64 errors = 0;

  while (unformat_check_input (input) != UNFORMAT_END_OF_INPUT)
    {
      if (unformat (input, "table %d", &table_id))
	{
	  /* Make sure the entry exists. */
	  fib = ip4_fib_get (table_id);
	  if ((fib) && (fib->index != table_id))
	    return clib_error_return (0, "<fib-index> %d does not exist",
				      table_id);
	}
      else if (unformat (input, "count %f", &count))
	;

      else if (unformat (input, "%U",
			 unformat_ip4_address, &ip4_base_address))
	;
      else
	return clib_error_return (0, "unknown input `%U'",
				  format_unformat_error, input);
    }

  n = count;

  for (i = 0; i < n; i++)
    {
      if (!ip4_lookup_validate (&ip4_base_address, table_id))
	errors++;

      ip4_base_address.as_u32 =
	clib_host_to_net_u32 (1 +
			      clib_net_to_host_u32 (ip4_base_address.as_u32));
    }

  if (errors)
    vlib_cli_output (vm, "%llu errors out of %d lookups\n", errors, n);
  else
    vlib_cli_output (vm, "No errors in %d lookups\n", n);

  return 0;
}

/*?
 * Perform a lookup of an IPv4 Address (or range of addresses) in the
 * given FIB table to determine if there is a conflict with the
 * adjacency table. The fib-id can be determined by using the
 * '<em>show ip fib</em>' command. If fib-id is not entered, default value
 * of 0 is used.
 *
 * @todo This command uses fib-id, other commands use table-id (not
 * just a name, they are different indexes). Would like to change this
 * to table-id for consistency.
 *
 * @cliexpar
 * Example of how to run the test lookup command:
 * @cliexstart{test lookup 172.16.1.1 table 1 count 2}
 * No errors in 2 lookups
 * @cliexend
?*/
/* *INDENT-OFF* */
VLIB_CLI_COMMAND (lookup_test_command, static) =
{
  .path = "test lookup",
  .short_help = "test lookup <ipv4-addr> [table <fib-id>] [count <nn>]",
  .function = test_lookup_command_fn,
};
/* *INDENT-ON* */

#ifndef CLIB_MARCH_VARIANT
int
vnet_set_ip4_flow_hash (u32 table_id, u32 flow_hash_config)
{
  u32 fib_index;

  fib_index = fib_table_find (FIB_PROTOCOL_IP4, table_id);

  if (~0 == fib_index)
    return VNET_API_ERROR_NO_SUCH_FIB;

  fib_table_set_flow_hash_config (fib_index, FIB_PROTOCOL_IP4,
				  flow_hash_config);

  return 0;
}
#endif

static clib_error_t *
set_ip_flow_hash_command_fn (vlib_main_t * vm,
			     unformat_input_t * input,
			     vlib_cli_command_t * cmd)
{
  int matched = 0;
  u32 table_id = 0;
  u32 flow_hash_config = 0;
  int rv;

  while (unformat_check_input (input) != UNFORMAT_END_OF_INPUT)
    {
      if (unformat (input, "table %d", &table_id))
	matched = 1;
#define _(a,v) \
    else if (unformat (input, #a)) { flow_hash_config |= v; matched=1;}
      foreach_flow_hash_bit
#undef _
	else
	break;
    }

  if (matched == 0)
    return clib_error_return (0, "unknown input `%U'",
			      format_unformat_error, input);

  rv = vnet_set_ip4_flow_hash (table_id, flow_hash_config);
  switch (rv)
    {
    case 0:
      break;

    case VNET_API_ERROR_NO_SUCH_FIB:
      return clib_error_return (0, "no such FIB table %d", table_id);

    default:
      clib_warning ("BUG: illegal flow hash config 0x%x", flow_hash_config);
      break;
    }

  return 0;
}

/*?
 * Configure the set of IPv4 fields used by the flow hash.
 *
 * @cliexpar
 * Example of how to set the flow hash on a given table:
 * @cliexcmd{set ip flow-hash table 7 dst sport dport proto}
 * Example of display the configured flow hash:
 * @cliexstart{show ip fib}
 * ipv4-VRF:0, fib_index 0, flow hash: src dst sport dport proto
 * 0.0.0.0/0
 *   unicast-ip4-chain
 *   [@0]: dpo-load-balance: [index:0 buckets:1 uRPF:0 to:[0:0]]
 *     [0] [@0]: dpo-drop ip6
 * 0.0.0.0/32
 *   unicast-ip4-chain
 *   [@0]: dpo-load-balance: [index:1 buckets:1 uRPF:1 to:[0:0]]
 *     [0] [@0]: dpo-drop ip6
 * 224.0.0.0/8
 *   unicast-ip4-chain
 *   [@0]: dpo-load-balance: [index:3 buckets:1 uRPF:3 to:[0:0]]
 *     [0] [@0]: dpo-drop ip6
 * 6.0.1.2/32
 *   unicast-ip4-chain
 *   [@0]: dpo-load-balance: [index:30 buckets:1 uRPF:29 to:[0:0]]
 *     [0] [@3]: arp-ipv4: via 6.0.0.1 af_packet0
 * 7.0.0.1/32
 *   unicast-ip4-chain
 *   [@0]: dpo-load-balance: [index:31 buckets:4 uRPF:30 to:[0:0]]
 *     [0] [@3]: arp-ipv4: via 6.0.0.2 af_packet0
 *     [1] [@3]: arp-ipv4: via 6.0.0.2 af_packet0
 *     [2] [@3]: arp-ipv4: via 6.0.0.2 af_packet0
 *     [3] [@3]: arp-ipv4: via 6.0.0.1 af_packet0
 * 240.0.0.0/8
 *   unicast-ip4-chain
 *   [@0]: dpo-load-balance: [index:2 buckets:1 uRPF:2 to:[0:0]]
 *     [0] [@0]: dpo-drop ip6
 * 255.255.255.255/32
 *   unicast-ip4-chain
 *   [@0]: dpo-load-balance: [index:4 buckets:1 uRPF:4 to:[0:0]]
 *     [0] [@0]: dpo-drop ip6
 * ipv4-VRF:7, fib_index 1, flow hash: dst sport dport proto
 * 0.0.0.0/0
 *   unicast-ip4-chain
 *   [@0]: dpo-load-balance: [index:12 buckets:1 uRPF:11 to:[0:0]]
 *     [0] [@0]: dpo-drop ip6
 * 0.0.0.0/32
 *   unicast-ip4-chain
 *   [@0]: dpo-load-balance: [index:13 buckets:1 uRPF:12 to:[0:0]]
 *     [0] [@0]: dpo-drop ip6
 * 172.16.1.0/24
 *   unicast-ip4-chain
 *   [@0]: dpo-load-balance: [index:17 buckets:1 uRPF:16 to:[0:0]]
 *     [0] [@4]: ipv4-glean: af_packet0
 * 172.16.1.1/32
 *   unicast-ip4-chain
 *   [@0]: dpo-load-balance: [index:18 buckets:1 uRPF:17 to:[1:84]]
 *     [0] [@2]: dpo-receive: 172.16.1.1 on af_packet0
 * 172.16.1.2/32
 *   unicast-ip4-chain
 *   [@0]: dpo-load-balance: [index:21 buckets:1 uRPF:20 to:[0:0]]
 *     [0] [@5]: ipv4 via 172.16.1.2 af_packet0: IP4: 02:fe:9e:70:7a:2b -> 26:a5:f6:9c:3a:36
 * 172.16.2.0/24
 *   unicast-ip4-chain
 *   [@0]: dpo-load-balance: [index:19 buckets:1 uRPF:18 to:[0:0]]
 *     [0] [@4]: ipv4-glean: af_packet1
 * 172.16.2.1/32
 *   unicast-ip4-chain
 *   [@0]: dpo-load-balance: [index:20 buckets:1 uRPF:19 to:[0:0]]
 *     [0] [@2]: dpo-receive: 172.16.2.1 on af_packet1
 * 224.0.0.0/8
 *   unicast-ip4-chain
 *   [@0]: dpo-load-balance: [index:15 buckets:1 uRPF:14 to:[0:0]]
 *     [0] [@0]: dpo-drop ip6
 * 240.0.0.0/8
 *   unicast-ip4-chain
 *   [@0]: dpo-load-balance: [index:14 buckets:1 uRPF:13 to:[0:0]]
 *     [0] [@0]: dpo-drop ip6
 * 255.255.255.255/32
 *   unicast-ip4-chain
 *   [@0]: dpo-load-balance: [index:16 buckets:1 uRPF:15 to:[0:0]]
 *     [0] [@0]: dpo-drop ip6
 * @cliexend
?*/
/* *INDENT-OFF* */
VLIB_CLI_COMMAND (set_ip_flow_hash_command, static) =
{
  .path = "set ip flow-hash",
  .short_help =
  "set ip flow-hash table <table-id> [src] [dst] [sport] [dport] [proto] [reverse]",
  .function = set_ip_flow_hash_command_fn,
};
/* *INDENT-ON* */

#ifndef CLIB_MARCH_VARIANT
int
vnet_set_ip4_classify_intfc (vlib_main_t * vm, u32 sw_if_index,
			     u32 table_index)
{
  vnet_main_t *vnm = vnet_get_main ();
  vnet_interface_main_t *im = &vnm->interface_main;
  ip4_main_t *ipm = &ip4_main;
  ip_lookup_main_t *lm = &ipm->lookup_main;
  vnet_classify_main_t *cm = &vnet_classify_main;
  ip4_address_t *if_addr;

  if (pool_is_free_index (im->sw_interfaces, sw_if_index))
    return VNET_API_ERROR_NO_MATCHING_INTERFACE;

  if (table_index != ~0 && pool_is_free_index (cm->tables, table_index))
    return VNET_API_ERROR_NO_SUCH_ENTRY;

  vec_validate (lm->classify_table_index_by_sw_if_index, sw_if_index);
  lm->classify_table_index_by_sw_if_index[sw_if_index] = table_index;

  if_addr = ip4_interface_first_address (ipm, sw_if_index, NULL);

  if (NULL != if_addr)
    {
      fib_prefix_t pfx = {
	.fp_len = 32,
	.fp_proto = FIB_PROTOCOL_IP4,
	.fp_addr.ip4 = *if_addr,
      };
      u32 fib_index;

      fib_index = fib_table_get_index_for_sw_if_index (FIB_PROTOCOL_IP4,
						       sw_if_index);


      if (table_index != (u32) ~ 0)
	{
	  dpo_id_t dpo = DPO_INVALID;

	  dpo_set (&dpo,
		   DPO_CLASSIFY,
		   DPO_PROTO_IP4,
		   classify_dpo_create (DPO_PROTO_IP4, table_index));

	  fib_table_entry_special_dpo_add (fib_index,
					   &pfx,
					   FIB_SOURCE_CLASSIFY,
					   FIB_ENTRY_FLAG_NONE, &dpo);
	  dpo_reset (&dpo);
	}
      else
	{
	  fib_table_entry_special_remove (fib_index,
					  &pfx, FIB_SOURCE_CLASSIFY);
	}
    }

  return 0;
}
#endif

static clib_error_t *
set_ip_classify_command_fn (vlib_main_t * vm,
			    unformat_input_t * input,
			    vlib_cli_command_t * cmd)
{
  u32 table_index = ~0;
  int table_index_set = 0;
  u32 sw_if_index = ~0;
  int rv;

  while (unformat_check_input (input) != UNFORMAT_END_OF_INPUT)
    {
      if (unformat (input, "table-index %d", &table_index))
	table_index_set = 1;
      else if (unformat (input, "intfc %U", unformat_vnet_sw_interface,
			 vnet_get_main (), &sw_if_index))
	;
      else
	break;
    }

  if (table_index_set == 0)
    return clib_error_return (0, "classify table-index must be specified");

  if (sw_if_index == ~0)
    return clib_error_return (0, "interface / subif must be specified");

  rv = vnet_set_ip4_classify_intfc (vm, sw_if_index, table_index);

  switch (rv)
    {
    case 0:
      break;

    case VNET_API_ERROR_NO_MATCHING_INTERFACE:
      return clib_error_return (0, "No such interface");

    case VNET_API_ERROR_NO_SUCH_ENTRY:
      return clib_error_return (0, "No such classifier table");
    }
  return 0;
}

/*?
 * Assign a classification table to an interface. The classification
 * table is created using the '<em>classify table</em>' and '<em>classify session</em>'
 * commands. Once the table is create, use this command to filter packets
 * on an interface.
 *
 * @cliexpar
 * Example of how to assign a classification table to an interface:
 * @cliexcmd{set ip classify intfc GigabitEthernet2/0/0 table-index 1}
?*/
/* *INDENT-OFF* */
VLIB_CLI_COMMAND (set_ip_classify_command, static) =
{
    .path = "set ip classify",
    .short_help =
    "set ip classify intfc <interface> table-index <classify-idx>",
    .function = set_ip_classify_command_fn,
};
/* *INDENT-ON* */

static clib_error_t *
ip4_config (vlib_main_t * vm, unformat_input_t * input)
{
  ip4_main_t *im = &ip4_main;
  uword heapsize = 0;

  while (unformat_check_input (input) != UNFORMAT_END_OF_INPUT)
    {
      if (unformat (input, "heap-size %U", unformat_memory_size, &heapsize))
	;
      else
	return clib_error_return (0,
				  "invalid heap-size parameter `%U'",
				  format_unformat_error, input);
    }

  im->mtrie_heap_size = heapsize;

  return 0;
}

VLIB_EARLY_CONFIG_FUNCTION (ip4_config, "ip");

/*
 * fd.io coding-style-patch-verification: ON
 *
 * Local Variables:
 * eval: (c-set-style "gnu")
 * End:
 */