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
|
/*
* Copyright (c) 2016 Cisco and/or its affiliates.
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at:
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include <vnet/ip/lookup.h>
#include <vnet/dpo/replicate_dpo.h>
#include <vnet/dpo/drop_dpo.h>
#include <vnet/adj/adj.h>
#undef REP_DEBUG
#ifdef REP_DEBUG
#define REP_DBG(_rep, _fmt, _args...) \
{ \
u8* _tmp =NULL; \
clib_warning("rep:[%s]:" _fmt, \
replicate_format(replicate_get_index((_rep)), \
0, _tmp), \
##_args); \
vec_free(_tmp); \
}
#else
#define REP_DBG(_p, _fmt, _args...)
#endif
/**
* Pool of all DPOs. It's not static so the DP can have fast access
*/
replicate_t *replicate_pool;
/**
* The one instance of replicate main
*/
replicate_main_t replicate_main;
static inline index_t
replicate_get_index (const replicate_t *rep)
{
return (rep - replicate_pool);
}
static inline dpo_id_t*
replicate_get_buckets (replicate_t *rep)
{
if (REP_HAS_INLINE_BUCKETS(rep))
{
return (rep->rep_buckets_inline);
}
else
{
return (rep->rep_buckets);
}
}
static replicate_t *
replicate_alloc_i (void)
{
replicate_t *rep;
pool_get_aligned(replicate_pool, rep, CLIB_CACHE_LINE_BYTES);
memset(rep, 0, sizeof(*rep));
vlib_validate_combined_counter(&(replicate_main.repm_counters),
replicate_get_index(rep));
vlib_zero_combined_counter(&(replicate_main.repm_counters),
replicate_get_index(rep));
return (rep);
}
static u8*
replicate_format (index_t repi,
replicate_format_flags_t flags,
u32 indent,
u8 *s)
{
vlib_counter_t to;
replicate_t *rep;
dpo_id_t *buckets;
u32 i;
rep = replicate_get(repi);
vlib_get_combined_counter(&(replicate_main.repm_counters), repi, &to);
buckets = replicate_get_buckets(rep);
s = format(s, "%U: ", format_dpo_type, DPO_REPLICATE);
s = format(s, "[index:%d buckets:%d ", repi, rep->rep_n_buckets);
s = format(s, "to:[%Ld:%Ld]]", to.packets, to.bytes);
for (i = 0; i < rep->rep_n_buckets; i++)
{
s = format(s, "\n%U", format_white_space, indent+2);
s = format(s, "[%d]", i);
s = format(s, " %U", format_dpo_id, &buckets[i], indent+6);
}
return (s);
}
u8*
format_replicate (u8 * s, va_list * args)
{
index_t repi = va_arg(*args, index_t);
replicate_format_flags_t flags = va_arg(*args, replicate_format_flags_t);
return (replicate_format(repi, flags, 0, s));
}
static u8*
format_replicate_dpo (u8 * s, va_list * args)
{
index_t repi = va_arg(*args, index_t);
u32 indent = va_arg(*args, u32);
return (replicate_format(repi, REPLICATE_FORMAT_DETAIL, indent, s));
}
static replicate_t *
replicate_create_i (u32 num_buckets,
dpo_proto_t rep_proto)
{
replicate_t *rep;
rep = replicate_alloc_i();
rep->rep_n_buckets = num_buckets;
rep->rep_proto = rep_proto;
if (!REP_HAS_INLINE_BUCKETS(rep))
{
vec_validate_aligned(rep->rep_buckets,
rep->rep_n_buckets - 1,
CLIB_CACHE_LINE_BYTES);
}
REP_DBG(rep, "create");
return (rep);
}
index_t
replicate_create (u32 n_buckets,
dpo_proto_t rep_proto)
{
return (replicate_get_index(replicate_create_i(n_buckets, rep_proto)));
}
static inline void
replicate_set_bucket_i (replicate_t *rep,
u32 bucket,
dpo_id_t *buckets,
const dpo_id_t *next)
{
dpo_stack(DPO_REPLICATE, rep->rep_proto, &buckets[bucket], next);
}
void
replicate_set_bucket (index_t repi,
u32 bucket,
const dpo_id_t *next)
{
replicate_t *rep;
dpo_id_t *buckets;
rep = replicate_get(repi);
buckets = replicate_get_buckets(rep);
ASSERT(bucket < rep->rep_n_buckets);
replicate_set_bucket_i(rep, bucket, buckets, next);
}
int
replicate_is_drop (const dpo_id_t *dpo)
{
replicate_t *rep;
if (DPO_REPLICATE != dpo->dpoi_type)
return (0);
rep = replicate_get(dpo->dpoi_index);
if (1 == rep->rep_n_buckets)
{
return (dpo_is_drop(replicate_get_bucket_i(rep, 0)));
}
return (0);
}
const dpo_id_t *
replicate_get_bucket (index_t repi,
u32 bucket)
{
replicate_t *rep;
rep = replicate_get(repi);
return (replicate_get_bucket_i(rep, bucket));
}
static load_balance_path_t *
replicate_multipath_next_hop_fixup (load_balance_path_t *nhs,
dpo_proto_t drop_proto)
{
if (0 == vec_len(nhs))
{
load_balance_path_t *nh;
/*
* we need something for the replicate. so use the drop
*/
vec_add2(nhs, nh, 1);
nh->path_weight = 1;
dpo_copy(&nh->path_dpo, drop_dpo_get(drop_proto));
}
return (nhs);
}
/*
* Fill in adjacencies in block based on corresponding
* next hop adjacencies.
*/
static void
replicate_fill_buckets (replicate_t *rep,
load_balance_path_t *nhs,
dpo_id_t *buckets,
u32 n_buckets)
{
load_balance_path_t * nh;
u16 ii, bucket;
bucket = 0;
/*
* the next-hops have normalised weights. that means their sum is the number
* of buckets we need to fill.
*/
vec_foreach (nh, nhs)
{
for (ii = 0; ii < nh->path_weight; ii++)
{
ASSERT(bucket < n_buckets);
replicate_set_bucket_i(rep, bucket++, buckets, &nh->path_dpo);
}
}
}
static inline void
replicate_set_n_buckets (replicate_t *rep,
u32 n_buckets)
{
rep->rep_n_buckets = n_buckets;
}
void
replicate_multipath_update (const dpo_id_t *dpo,
load_balance_path_t * next_hops)
{
load_balance_path_t * nh, * nhs;
dpo_id_t *tmp_dpo;
u32 ii, n_buckets;
replicate_t *rep;
ASSERT(DPO_REPLICATE == dpo->dpoi_type);
rep = replicate_get(dpo->dpoi_index);
nhs = replicate_multipath_next_hop_fixup(next_hops,
rep->rep_proto);
n_buckets = vec_len(nhs);
if (0 == rep->rep_n_buckets)
{
/*
* first time initialisation. no packets inflight, so we can write
* at leisure.
*/
replicate_set_n_buckets(rep, n_buckets);
if (!REP_HAS_INLINE_BUCKETS(rep))
vec_validate_aligned(rep->rep_buckets,
rep->rep_n_buckets - 1,
CLIB_CACHE_LINE_BYTES);
replicate_fill_buckets(rep, nhs,
replicate_get_buckets(rep),
n_buckets);
}
else
{
/*
* This is a modification of an existing replicate.
* We need to ensure that packets in flight see a consistent state, that
* is the number of reported buckets the REP has
* is not more than it actually has. So if the
* number of buckets is increasing, we must update the bucket array first,
* then the reported number. vice-versa if the number of buckets goes down.
*/
if (n_buckets == rep->rep_n_buckets)
{
/*
* no change in the number of buckets. we can simply fill what
* is new over what is old.
*/
replicate_fill_buckets(rep, nhs,
replicate_get_buckets(rep),
n_buckets);
}
else if (n_buckets > rep->rep_n_buckets)
{
/*
* we have more buckets. the old replicate map (if there is one)
* will remain valid, i.e. mapping to indices within range, so we
* update it last.
*/
if (n_buckets > REP_NUM_INLINE_BUCKETS &&
rep->rep_n_buckets <= REP_NUM_INLINE_BUCKETS)
{
/*
* the new increased number of buckets is crossing the threshold
* from the inline storage to out-line. Alloc the outline buckets
* first, then fixup the number. then reset the inlines.
*/
ASSERT(NULL == rep->rep_buckets);
vec_validate_aligned(rep->rep_buckets,
n_buckets - 1,
CLIB_CACHE_LINE_BYTES);
replicate_fill_buckets(rep, nhs,
rep->rep_buckets,
n_buckets);
CLIB_MEMORY_BARRIER();
replicate_set_n_buckets(rep, n_buckets);
CLIB_MEMORY_BARRIER();
for (ii = 0; ii < REP_NUM_INLINE_BUCKETS; ii++)
{
dpo_reset(&rep->rep_buckets_inline[ii]);
}
}
else
{
if (n_buckets <= REP_NUM_INLINE_BUCKETS)
{
/*
* we are not crossing the threshold and it's still inline buckets.
* we can write the new on the old..
*/
replicate_fill_buckets(rep, nhs,
replicate_get_buckets(rep),
n_buckets);
CLIB_MEMORY_BARRIER();
replicate_set_n_buckets(rep, n_buckets);
}
else
{
/*
* we are not crossing the threshold. We need a new bucket array to
* hold the increased number of choices.
*/
dpo_id_t *new_buckets, *old_buckets, *tmp_dpo;
new_buckets = NULL;
old_buckets = replicate_get_buckets(rep);
vec_validate_aligned(new_buckets,
n_buckets - 1,
CLIB_CACHE_LINE_BYTES);
replicate_fill_buckets(rep, nhs, new_buckets, n_buckets);
CLIB_MEMORY_BARRIER();
rep->rep_buckets = new_buckets;
CLIB_MEMORY_BARRIER();
replicate_set_n_buckets(rep, n_buckets);
vec_foreach(tmp_dpo, old_buckets)
{
dpo_reset(tmp_dpo);
}
vec_free(old_buckets);
}
}
}
else
{
/*
* bucket size shrinkage.
*/
if (n_buckets <= REP_NUM_INLINE_BUCKETS &&
rep->rep_n_buckets > REP_NUM_INLINE_BUCKETS)
{
/*
* the new decreased number of buckets is crossing the threshold
* from out-line storage to inline:
* 1 - Fill the inline buckets,
* 2 - fixup the number (and this point the inline buckets are
* used).
* 3 - free the outline buckets
*/
replicate_fill_buckets(rep, nhs,
rep->rep_buckets_inline,
n_buckets);
CLIB_MEMORY_BARRIER();
replicate_set_n_buckets(rep, n_buckets);
CLIB_MEMORY_BARRIER();
vec_foreach(tmp_dpo, rep->rep_buckets)
{
dpo_reset(tmp_dpo);
}
vec_free(rep->rep_buckets);
}
else
{
/*
* not crossing the threshold.
* 1 - update the number to the smaller size
* 2 - write the new buckets
* 3 - reset those no longer used.
*/
dpo_id_t *buckets;
u32 old_n_buckets;
old_n_buckets = rep->rep_n_buckets;
buckets = replicate_get_buckets(rep);
replicate_set_n_buckets(rep, n_buckets);
CLIB_MEMORY_BARRIER();
replicate_fill_buckets(rep, nhs,
buckets,
n_buckets);
for (ii = n_buckets; ii < old_n_buckets; ii++)
{
dpo_reset(&buckets[ii]);
}
}
}
}
vec_foreach (nh, nhs)
{
dpo_reset(&nh->path_dpo);
}
vec_free(nhs);
}
static void
replicate_lock (dpo_id_t *dpo)
{
replicate_t *rep;
rep = replicate_get(dpo->dpoi_index);
rep->rep_locks++;
}
static void
replicate_destroy (replicate_t *rep)
{
dpo_id_t *buckets;
int i;
buckets = replicate_get_buckets(rep);
for (i = 0; i < rep->rep_n_buckets; i++)
{
dpo_reset(&buckets[i]);
}
REP_DBG(rep, "destroy");
if (!REP_HAS_INLINE_BUCKETS(rep))
{
vec_free(rep->rep_buckets);
}
pool_put(replicate_pool, rep);
}
static void
replicate_unlock (dpo_id_t *dpo)
{
replicate_t *rep;
rep = replicate_get(dpo->dpoi_index);
rep->rep_locks--;
if (0 == rep->rep_locks)
{
replicate_destroy(rep);
}
}
static void
replicate_mem_show (void)
{
fib_show_memory_usage("replicate",
pool_elts(replicate_pool),
pool_len(replicate_pool),
sizeof(replicate_t));
}
const static dpo_vft_t rep_vft = {
.dv_lock = replicate_lock,
.dv_unlock = replicate_unlock,
.dv_format = format_replicate_dpo,
.dv_mem_show = replicate_mem_show,
};
/**
* @brief The per-protocol VLIB graph nodes that are assigned to a replicate
* object.
*
* this means that these graph nodes are ones from which a replicate is the
* parent object in the DPO-graph.
*/
const static char* const replicate_ip4_nodes[] =
{
"ip4-replicate",
NULL,
};
const static char* const replicate_ip6_nodes[] =
{
"ip6-replicate",
NULL,
};
const static char* const replicate_mpls_nodes[] =
{
"mpls-replicate",
NULL,
};
const static char* const * const replicate_nodes[DPO_PROTO_NUM] =
{
[DPO_PROTO_IP4] = replicate_ip4_nodes,
[DPO_PROTO_IP6] = replicate_ip6_nodes,
[DPO_PROTO_MPLS] = replicate_mpls_nodes,
};
void
replicate_module_init (void)
{
dpo_register(DPO_REPLICATE, &rep_vft, replicate_nodes);
}
static clib_error_t *
replicate_show (vlib_main_t * vm,
unformat_input_t * input,
vlib_cli_command_t * cmd)
{
index_t repi = INDEX_INVALID;
while (unformat_check_input (input) != UNFORMAT_END_OF_INPUT)
{
if (unformat (input, "%d", &repi))
;
else
break;
}
if (INDEX_INVALID != repi)
{
vlib_cli_output (vm, "%U", format_replicate, repi,
REPLICATE_FORMAT_DETAIL);
}
else
{
replicate_t *rep;
pool_foreach(rep, replicate_pool,
({
vlib_cli_output (vm, "%U", format_replicate,
replicate_get_index(rep),
REPLICATE_FORMAT_NONE);
}));
}
return 0;
}
VLIB_CLI_COMMAND (replicate_show_command, static) = {
.path = "show replicate",
.short_help = "show replicate [<index>]",
.function = replicate_show,
};
typedef struct replicate_trace_t_
{
index_t rep_index;
index_t dpo_index;
dpo_type_t dpo_type;
} replicate_trace_t;
static uword
replicate_inline (vlib_main_t * vm,
vlib_node_runtime_t * node,
vlib_frame_t * frame)
{
vlib_combined_counter_main_t * cm = &replicate_main.repm_counters;
u32 n_left_from, * from, * to_next, next_index;
u32 cpu_index = os_get_cpu_number();
from = vlib_frame_vector_args (frame);
n_left_from = frame->n_vectors;
next_index = node->cached_next_index;
while (n_left_from > 0)
{
u32 n_left_to_next;
vlib_get_next_frame (vm, node, next_index,
to_next, n_left_to_next);
while (n_left_from > 0 && n_left_to_next > 0)
{
u32 next0, ci0, bi0, bucket, repi0;
const replicate_t *rep0;
vlib_buffer_t * b0, *c0;
const dpo_id_t *dpo0;
bi0 = from[0];
to_next[0] = bi0;
from += 1;
to_next += 1;
n_left_from -= 1;
n_left_to_next -= 1;
b0 = vlib_get_buffer (vm, bi0);
repi0 = vnet_buffer (b0)->ip.adj_index[VLIB_TX];
rep0 = replicate_get(repi0);
vlib_increment_combined_counter(
cm, cpu_index, repi0, 1,
vlib_buffer_length_in_chain(vm, b0));
/* ship the original to the first bucket */
dpo0 = replicate_get_bucket_i(rep0, 0);
next0 = dpo0->dpoi_next_node;
vnet_buffer (b0)->ip.adj_index[VLIB_TX] = dpo0->dpoi_index;
if (PREDICT_FALSE(b0->flags & VLIB_BUFFER_IS_TRACED))
{
replicate_trace_t *t = vlib_add_trace (vm, node, b0, sizeof (*t));
t->rep_index = repi0;
t->dpo_index = dpo0->dpoi_index;
t->dpo_type = dpo0->dpoi_type;
}
vlib_validate_buffer_enqueue_x1 (vm, node, next_index,
to_next, n_left_to_next,
bi0, next0);
/* ship copies to the rest of the buckets */
for (bucket = 1; bucket < rep0->rep_n_buckets; bucket++)
{
/* Make a copy */
c0 = vlib_buffer_copy(vm, b0);
ci0 = vlib_get_buffer_index(vm, c0);
to_next[0] = ci0;
to_next += 1;
n_left_to_next -= 1;
dpo0 = replicate_get_bucket_i(rep0, bucket);
next0 = dpo0->dpoi_next_node;
vnet_buffer (c0)->ip.adj_index[VLIB_TX] = dpo0->dpoi_index;
if (PREDICT_FALSE(b0->flags & VLIB_BUFFER_IS_TRACED))
{
replicate_trace_t *t = vlib_add_trace (vm, node, b0, sizeof (*t));
t->rep_index = repi0;
t->dpo_index = dpo0->dpoi_index;
t->dpo_type = dpo0->dpoi_type;
}
vlib_validate_buffer_enqueue_x1 (vm, node, next_index,
to_next, n_left_to_next,
ci0, next0);
}
}
vlib_put_next_frame (vm, node, next_index, n_left_to_next);
}
return frame->n_vectors;
}
static u8 *
format_replicate_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 *);
replicate_trace_t *t = va_arg (*args, replicate_trace_t *);
s = format (s, "replicate: %d via %U:%d",
t->rep_index,
format_dpo_type, t->dpo_type,
t->dpo_index);
return s;
}
static uword
ip4_replicate (vlib_main_t * vm,
vlib_node_runtime_t * node,
vlib_frame_t * frame)
{
return (replicate_inline (vm, node, frame));
}
/**
* @brief
*/
VLIB_REGISTER_NODE (ip4_replicate_node) = {
.function = ip4_replicate,
.name = "ip4-replicate",
.vector_size = sizeof (u32),
.format_trace = format_replicate_trace,
.n_next_nodes = 1,
.next_nodes = {
[0] = "error-drop",
},
};
static uword
ip6_replicate (vlib_main_t * vm,
vlib_node_runtime_t * node,
vlib_frame_t * frame)
{
return (replicate_inline (vm, node, frame));
}
/**
* @brief
*/
VLIB_REGISTER_NODE (ip6_replicate_node) = {
.function = ip6_replicate,
.name = "ip6-replicate",
.vector_size = sizeof (u32),
.format_trace = format_replicate_trace,
.n_next_nodes = 1,
.next_nodes = {
[0] = "error-drop",
},
};
|