forked from mtuncer/memcached
-
Notifications
You must be signed in to change notification settings - Fork 0
/
items.c
1478 lines (1348 loc) · 50.4 KB
/
items.c
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
/* -*- Mode: C; tab-width: 4; c-basic-offset: 4; indent-tabs-mode: nil -*- */
#include "memcached.h"
#include <sys/stat.h>
#include <sys/socket.h>
#include <sys/signal.h>
#include <sys/resource.h>
#include <fcntl.h>
#include <netinet/in.h>
#include <errno.h>
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include <time.h>
#include <assert.h>
#include <unistd.h>
/* Forward Declarations */
static void item_link_q(item *it);
static void item_unlink_q(item *it);
#define HOT_LRU 0
#define WARM_LRU 64
#define COLD_LRU 128
#define NOEXP_LRU 192
static unsigned int lru_type_map[4] = {HOT_LRU, WARM_LRU, COLD_LRU, NOEXP_LRU};
#define CLEAR_LRU(id) (id & ~(3<<6))
#define LARGEST_ID POWER_LARGEST
typedef struct {
uint64_t evicted;
uint64_t evicted_nonzero;
uint64_t reclaimed;
uint64_t outofmemory;
uint64_t tailrepairs;
uint64_t expired_unfetched;
uint64_t evicted_unfetched;
uint64_t crawler_reclaimed;
uint64_t crawler_items_checked;
uint64_t lrutail_reflocked;
uint64_t moves_to_cold;
uint64_t moves_to_warm;
uint64_t moves_within_lru;
uint64_t direct_reclaims;
rel_time_t evicted_time;
} itemstats_t;
typedef struct {
uint64_t histo[60];
uint64_t ttl_hourplus;
uint64_t noexp;
uint64_t reclaimed;
uint64_t seen;
rel_time_t start_time;
rel_time_t end_time;
bool run_complete;
} crawlerstats_t;
static item *heads[LARGEST_ID];
static item *tails[LARGEST_ID];
static crawler crawlers[LARGEST_ID];
static itemstats_t itemstats[LARGEST_ID];
static unsigned int sizes[LARGEST_ID];
static crawlerstats_t crawlerstats[MAX_NUMBER_OF_SLAB_CLASSES];
static int crawler_count = 0;
static volatile int do_run_lru_crawler_thread = 0;
static volatile int do_run_lru_maintainer_thread = 0;
static int lru_crawler_initialized = 0;
static int lru_maintainer_initialized = 0;
static int lru_maintainer_check_clsid = 0;
static pthread_mutex_t lru_crawler_lock = PTHREAD_MUTEX_INITIALIZER;
static pthread_cond_t lru_crawler_cond = PTHREAD_COND_INITIALIZER;
static pthread_mutex_t lru_crawler_stats_lock = PTHREAD_MUTEX_INITIALIZER;
static pthread_mutex_t lru_maintainer_lock = PTHREAD_MUTEX_INITIALIZER;
static pthread_mutex_t cas_id_lock = PTHREAD_MUTEX_INITIALIZER;
void item_stats_reset(void) {
int i;
for (i = 0; i < LARGEST_ID; i++) {
pthread_mutex_lock(&lru_locks[i]);
memset(&itemstats[i], 0, sizeof(itemstats_t));
pthread_mutex_unlock(&lru_locks[i]);
}
}
static int lru_pull_tail(const int orig_id, const int cur_lru,
const unsigned int total_chunks, const bool do_evict, const uint32_t cur_hv);
static int lru_crawler_start(uint32_t id, uint32_t remaining);
/* Get the next CAS id for a new item. */
/* TODO: refactor some atomics for this. */
uint64_t get_cas_id(void) {
static uint64_t cas_id = 0;
pthread_mutex_lock(&cas_id_lock);
uint64_t next_id = ++cas_id;
pthread_mutex_unlock(&cas_id_lock);
return next_id;
}
static int is_flushed(item *it) {
rel_time_t oldest_live = settings.oldest_live;
uint64_t cas = ITEM_get_cas(it);
uint64_t oldest_cas = settings.oldest_cas;
if (oldest_live == 0 || oldest_live > current_time)
return 0;
if ((it->time <= oldest_live)
|| (oldest_cas != 0 && cas != 0 && cas < oldest_cas)) {
return 1;
}
return 0;
}
static unsigned int noexp_lru_size(int slabs_clsid) {
int id = CLEAR_LRU(slabs_clsid);
unsigned int ret;
pthread_mutex_lock(&lru_locks[id]);
ret = sizes[id];
pthread_mutex_unlock(&lru_locks[id]);
return ret;
}
/* Enable this for reference-count debugging. */
#if 0
# define DEBUG_REFCNT(it,op) \
fprintf(stderr, "item %x refcnt(%c) %d %c%c%c\n", \
it, op, it->refcount, \
(it->it_flags & ITEM_LINKED) ? 'L' : ' ', \
(it->it_flags & ITEM_SLABBED) ? 'S' : ' ')
#else
# define DEBUG_REFCNT(it,op) while(0)
#endif
/**
* Generates the variable-sized part of the header for an object.
*
* key - The key
* nkey - The length of the key
* flags - key flags
* nbytes - Number of bytes to hold value and addition CRLF terminator
* suffix - Buffer for the "VALUE" line suffix (flags, size).
* nsuffix - The length of the suffix is stored here.
*
* Returns the total size of the header.
*/
static size_t item_make_header(const uint8_t nkey, const int flags, const int nbytes,
char *suffix, uint8_t *nsuffix) {
/* suffix is defined at 40 chars elsewhere.. */
*nsuffix = (uint8_t) snprintf(suffix, 40, " %d %d\r\n", flags, nbytes - 2);
return sizeof(item) + nkey + *nsuffix + nbytes;
}
item *do_item_alloc(char *key, const size_t nkey, const int flags,
const rel_time_t exptime, const int nbytes,
const uint32_t cur_hv) {
int i;
uint8_t nsuffix;
item *it = NULL;
char suffix[40];
unsigned int total_chunks;
size_t ntotal = item_make_header(nkey + 1, flags, nbytes, suffix, &nsuffix);
if (settings.use_cas) {
ntotal += sizeof(uint64_t);
}
unsigned int id = slabs_clsid(ntotal);
if (id == 0)
return 0;
/* If no memory is available, attempt a direct LRU juggle/eviction */
/* This is a race in order to simplify lru_pull_tail; in cases where
* locked items are on the tail, you want them to fall out and cause
* occasional OOM's, rather than internally work around them.
* This also gives one fewer code path for slab alloc/free
*/
for (i = 0; i < 5; i++) {
/* Try to reclaim memory first */
if (!settings.lru_maintainer_thread) {
lru_pull_tail(id, COLD_LRU, 0, false, cur_hv);
}
it = slabs_alloc(ntotal, id, &total_chunks);
if (settings.expirezero_does_not_evict)
total_chunks -= noexp_lru_size(id);
if (it == NULL) {
if (settings.lru_maintainer_thread) {
lru_pull_tail(id, HOT_LRU, total_chunks, false, cur_hv);
lru_pull_tail(id, WARM_LRU, total_chunks, false, cur_hv);
lru_pull_tail(id, COLD_LRU, total_chunks, true, cur_hv);
} else {
lru_pull_tail(id, COLD_LRU, 0, true, cur_hv);
}
} else {
break;
}
}
if (i > 0) {
pthread_mutex_lock(&lru_locks[id]);
itemstats[id].direct_reclaims += i;
pthread_mutex_unlock(&lru_locks[id]);
}
if (it == NULL) {
pthread_mutex_lock(&lru_locks[id]);
itemstats[id].outofmemory++;
pthread_mutex_unlock(&lru_locks[id]);
return NULL;
}
assert(it->slabs_clsid == 0);
//assert(it != heads[id]);
/* Refcount is seeded to 1 by slabs_alloc() */
it->next = it->prev = it->h_next = 0;
/* Items are initially loaded into the HOT_LRU. This is '0' but I want at
* least a note here. Compiler (hopefully?) optimizes this out.
*/
if (settings.lru_maintainer_thread) {
if (exptime == 0 && settings.expirezero_does_not_evict) {
id |= NOEXP_LRU;
} else {
id |= HOT_LRU;
}
} else {
/* There is only COLD in compat-mode */
id |= COLD_LRU;
}
it->slabs_clsid = id;
DEBUG_REFCNT(it, '*');
it->it_flags = settings.use_cas ? ITEM_CAS : 0;
it->nkey = nkey;
it->nbytes = nbytes;
memcpy(ITEM_key(it), key, nkey);
it->exptime = exptime;
memcpy(ITEM_suffix(it), suffix, (size_t)nsuffix);
it->nsuffix = nsuffix;
return it;
}
void item_free(item *it) {
size_t ntotal = ITEM_ntotal(it);
unsigned int clsid;
assert((it->it_flags & ITEM_LINKED) == 0);
assert(it != heads[it->slabs_clsid]);
assert(it != tails[it->slabs_clsid]);
assert(it->refcount == 0);
/* so slab size changer can tell later if item is already free or not */
clsid = ITEM_clsid(it);
DEBUG_REFCNT(it, 'F');
slabs_free(it, ntotal, clsid);
}
/**
* Returns true if an item will fit in the cache (its size does not exceed
* the maximum for a cache entry.)
*/
bool item_size_ok(const size_t nkey, const int flags, const int nbytes) {
char prefix[40];
uint8_t nsuffix;
size_t ntotal = item_make_header(nkey + 1, flags, nbytes,
prefix, &nsuffix);
if (settings.use_cas) {
ntotal += sizeof(uint64_t);
}
return slabs_clsid(ntotal) != 0;
}
static void do_item_link_q(item *it) { /* item is the new head */
item **head, **tail;
assert((it->it_flags & ITEM_SLABBED) == 0);
head = &heads[it->slabs_clsid];
tail = &tails[it->slabs_clsid];
assert(it != *head);
assert((*head && *tail) || (*head == 0 && *tail == 0));
it->prev = 0;
it->next = *head;
if (it->next) it->next->prev = it;
*head = it;
if (*tail == 0) *tail = it;
sizes[it->slabs_clsid]++;
return;
}
static void item_link_q(item *it) {
pthread_mutex_lock(&lru_locks[it->slabs_clsid]);
do_item_link_q(it);
pthread_mutex_unlock(&lru_locks[it->slabs_clsid]);
}
static void do_item_unlink_q(item *it) {
item **head, **tail;
head = &heads[it->slabs_clsid];
tail = &tails[it->slabs_clsid];
if (*head == it) {
assert(it->prev == 0);
*head = it->next;
}
if (*tail == it) {
assert(it->next == 0);
*tail = it->prev;
}
assert(it->next != it);
assert(it->prev != it);
if (it->next) it->next->prev = it->prev;
if (it->prev) it->prev->next = it->next;
sizes[it->slabs_clsid]--;
return;
}
static void item_unlink_q(item *it) {
pthread_mutex_lock(&lru_locks[it->slabs_clsid]);
do_item_unlink_q(it);
pthread_mutex_unlock(&lru_locks[it->slabs_clsid]);
}
int do_item_link(item *it, const uint32_t hv) {
MEMCACHED_ITEM_LINK(ITEM_key(it), it->nkey, it->nbytes);
assert((it->it_flags & (ITEM_LINKED|ITEM_SLABBED)) == 0);
it->it_flags |= ITEM_LINKED;
it->time = current_time;
STATS_LOCK();
stats.curr_bytes += ITEM_ntotal(it);
stats.curr_items += 1;
stats.total_items += 1;
STATS_UNLOCK();
/* Allocate a new CAS ID on link. */
ITEM_set_cas(it, (settings.use_cas) ? get_cas_id() : 0);
assoc_insert(it, hv);
item_link_q(it);
refcount_incr(&it->refcount);
return 1;
}
void do_item_unlink(item *it, const uint32_t hv) {
MEMCACHED_ITEM_UNLINK(ITEM_key(it), it->nkey, it->nbytes);
if ((it->it_flags & ITEM_LINKED) != 0) {
it->it_flags &= ~ITEM_LINKED;
STATS_LOCK();
stats.curr_bytes -= ITEM_ntotal(it);
stats.curr_items -= 1;
STATS_UNLOCK();
assoc_delete(ITEM_key(it), it->nkey, hv);
item_unlink_q(it);
do_item_remove(it);
}
}
/* FIXME: Is it necessary to keep this copy/pasted code? */
void do_item_unlink_nolock(item *it, const uint32_t hv) {
MEMCACHED_ITEM_UNLINK(ITEM_key(it), it->nkey, it->nbytes);
if ((it->it_flags & ITEM_LINKED) != 0) {
it->it_flags &= ~ITEM_LINKED;
STATS_LOCK();
stats.curr_bytes -= ITEM_ntotal(it);
stats.curr_items -= 1;
STATS_UNLOCK();
assoc_delete(ITEM_key(it), it->nkey, hv);
do_item_unlink_q(it);
do_item_remove(it);
}
}
void do_item_remove(item *it) {
MEMCACHED_ITEM_REMOVE(ITEM_key(it), it->nkey, it->nbytes);
assert((it->it_flags & ITEM_SLABBED) == 0);
assert(it->refcount > 0);
if (refcount_decr(&it->refcount) == 0) {
item_free(it);
}
}
/* Copy/paste to avoid adding two extra branches for all common calls, since
* _nolock is only used in an uncommon case where we want to relink. */
void do_item_update_nolock(item *it) {
MEMCACHED_ITEM_UPDATE(ITEM_key(it), it->nkey, it->nbytes);
if (it->time < current_time - ITEM_UPDATE_INTERVAL) {
assert((it->it_flags & ITEM_SLABBED) == 0);
if ((it->it_flags & ITEM_LINKED) != 0) {
do_item_unlink_q(it);
it->time = current_time;
do_item_link_q(it);
}
}
}
/* Bump the last accessed time, or relink if we're in compat mode */
void do_item_update(item *it) {
MEMCACHED_ITEM_UPDATE(ITEM_key(it), it->nkey, it->nbytes);
if (it->time < current_time - ITEM_UPDATE_INTERVAL) {
assert((it->it_flags & ITEM_SLABBED) == 0);
if ((it->it_flags & ITEM_LINKED) != 0) {
it->time = current_time;
if (!settings.lru_maintainer_thread) {
item_unlink_q(it);
item_link_q(it);
}
}
}
}
int do_item_replace(item *it, item *new_it, const uint32_t hv) {
MEMCACHED_ITEM_REPLACE(ITEM_key(it), it->nkey, it->nbytes,
ITEM_key(new_it), new_it->nkey, new_it->nbytes);
assert((it->it_flags & ITEM_SLABBED) == 0);
do_item_unlink(it, hv);
return do_item_link(new_it, hv);
}
/*@null@*/
/* This is walking the line of violating lock order, but I think it's safe.
* If the LRU lock is held, an item in the LRU cannot be wiped and freed.
* The data could possibly be overwritten, but this is only accessing the
* headers.
* It may not be the best idea to leave it like this, but for now it's safe.
* FIXME: only dumps the hot LRU with the new LRU's.
*/
char *item_cachedump(const unsigned int slabs_clsid, const unsigned int limit, unsigned int *bytes) {
unsigned int memlimit = 2 * 1024 * 1024; /* 2MB max response size */
char *buffer;
unsigned int bufcurr;
item *it;
unsigned int len;
unsigned int shown = 0;
char key_temp[KEY_MAX_LENGTH + 1];
char temp[512];
unsigned int id = slabs_clsid;
if (!settings.lru_maintainer_thread)
id |= COLD_LRU;
pthread_mutex_lock(&lru_locks[id]);
it = heads[id];
buffer = malloc((size_t)memlimit);
if (buffer == 0) {
return NULL;
}
bufcurr = 0;
while (it != NULL && (limit == 0 || shown < limit)) {
assert(it->nkey <= KEY_MAX_LENGTH);
if (it->nbytes == 0 && it->nkey == 0) {
it = it->next;
continue;
}
/* Copy the key since it may not be null-terminated in the struct */
strncpy(key_temp, ITEM_key(it), it->nkey);
key_temp[it->nkey] = 0x00; /* terminate */
len = snprintf(temp, sizeof(temp), "ITEM %s [%d b; %lu s]\r\n",
key_temp, it->nbytes - 2,
(unsigned long)it->exptime + process_started);
if (bufcurr + len + 6 > memlimit) /* 6 is END\r\n\0 */
break;
memcpy(buffer + bufcurr, temp, len);
bufcurr += len;
shown++;
it = it->next;
}
memcpy(buffer + bufcurr, "END\r\n", 6);
bufcurr += 5;
*bytes = bufcurr;
pthread_mutex_unlock(&lru_locks[id]);
return buffer;
}
void item_stats_evictions(uint64_t *evicted) {
int n;
for (n = 0; n < MAX_NUMBER_OF_SLAB_CLASSES; n++) {
int i;
int x;
for (x = 0; x < 4; x++) {
i = n | lru_type_map[x];
pthread_mutex_lock(&lru_locks[i]);
evicted[n] += itemstats[i].evicted;
pthread_mutex_unlock(&lru_locks[i]);
}
}
}
void item_stats_totals(ADD_STAT add_stats, void *c) {
itemstats_t totals;
memset(&totals, 0, sizeof(itemstats_t));
int n;
for (n = 0; n < MAX_NUMBER_OF_SLAB_CLASSES; n++) {
int x;
int i;
for (x = 0; x < 4; x++) {
i = n | lru_type_map[x];
pthread_mutex_lock(&lru_locks[i]);
totals.expired_unfetched += itemstats[i].expired_unfetched;
totals.evicted_unfetched += itemstats[i].evicted_unfetched;
totals.evicted += itemstats[i].evicted;
totals.reclaimed += itemstats[i].reclaimed;
totals.crawler_reclaimed += itemstats[i].crawler_reclaimed;
totals.crawler_items_checked += itemstats[i].crawler_items_checked;
totals.lrutail_reflocked += itemstats[i].lrutail_reflocked;
totals.moves_to_cold += itemstats[i].moves_to_cold;
totals.moves_to_warm += itemstats[i].moves_to_warm;
totals.moves_within_lru += itemstats[i].moves_within_lru;
totals.direct_reclaims += itemstats[i].direct_reclaims;
pthread_mutex_unlock(&lru_locks[i]);
}
}
APPEND_STAT("expired_unfetched", "%llu",
(unsigned long long)totals.expired_unfetched);
APPEND_STAT("evicted_unfetched", "%llu",
(unsigned long long)totals.evicted_unfetched);
APPEND_STAT("evictions", "%llu",
(unsigned long long)totals.evicted);
APPEND_STAT("reclaimed", "%llu",
(unsigned long long)totals.reclaimed);
APPEND_STAT("crawler_reclaimed", "%llu",
(unsigned long long)totals.crawler_reclaimed);
APPEND_STAT("crawler_items_checked", "%llu",
(unsigned long long)totals.crawler_items_checked);
APPEND_STAT("lrutail_reflocked", "%llu",
(unsigned long long)totals.lrutail_reflocked);
if (settings.lru_maintainer_thread) {
APPEND_STAT("moves_to_cold", "%llu",
(unsigned long long)totals.moves_to_cold);
APPEND_STAT("moves_to_warm", "%llu",
(unsigned long long)totals.moves_to_warm);
APPEND_STAT("moves_within_lru", "%llu",
(unsigned long long)totals.moves_within_lru);
APPEND_STAT("direct_reclaims", "%llu",
(unsigned long long)totals.direct_reclaims);
}
}
void item_stats(ADD_STAT add_stats, void *c) {
itemstats_t totals;
int n;
for (n = 0; n < MAX_NUMBER_OF_SLAB_CLASSES; n++) {
memset(&totals, 0, sizeof(itemstats_t));
int x;
int i;
unsigned int size = 0;
unsigned int age = 0;
unsigned int lru_size_map[4];
const char *fmt = "items:%d:%s";
char key_str[STAT_KEY_LEN];
char val_str[STAT_VAL_LEN];
int klen = 0, vlen = 0;
for (x = 0; x < 4; x++) {
i = n | lru_type_map[x];
pthread_mutex_lock(&lru_locks[i]);
totals.evicted += itemstats[i].evicted;
totals.evicted_nonzero += itemstats[i].evicted_nonzero;
totals.outofmemory += itemstats[i].outofmemory;
totals.tailrepairs += itemstats[i].tailrepairs;
totals.reclaimed += itemstats[i].reclaimed;
totals.expired_unfetched += itemstats[i].expired_unfetched;
totals.evicted_unfetched += itemstats[i].evicted_unfetched;
totals.crawler_reclaimed += itemstats[i].crawler_reclaimed;
totals.crawler_items_checked += itemstats[i].crawler_items_checked;
totals.lrutail_reflocked += itemstats[i].lrutail_reflocked;
totals.moves_to_cold += itemstats[i].moves_to_cold;
totals.moves_to_warm += itemstats[i].moves_to_warm;
totals.moves_within_lru += itemstats[i].moves_within_lru;
totals.direct_reclaims += itemstats[i].direct_reclaims;
size += sizes[i];
lru_size_map[x] = sizes[i];
if (lru_type_map[x] == COLD_LRU && tails[i] != NULL)
age = current_time - tails[i]->time;
pthread_mutex_unlock(&lru_locks[i]);
}
if (size == 0)
continue;
APPEND_NUM_FMT_STAT(fmt, n, "number", "%u", size);
if (settings.lru_maintainer_thread) {
APPEND_NUM_FMT_STAT(fmt, n, "number_hot", "%u", lru_size_map[0]);
APPEND_NUM_FMT_STAT(fmt, n, "number_warm", "%u", lru_size_map[1]);
APPEND_NUM_FMT_STAT(fmt, n, "number_cold", "%u", lru_size_map[2]);
if (settings.expirezero_does_not_evict)
APPEND_NUM_FMT_STAT(fmt, n, "number_noexp", "%u", lru_size_map[3]);
}
APPEND_NUM_FMT_STAT(fmt, n, "age", "%u", age);
APPEND_NUM_FMT_STAT(fmt, n, "evicted",
"%llu", (unsigned long long)totals.evicted);
APPEND_NUM_FMT_STAT(fmt, n, "evicted_nonzero",
"%llu", (unsigned long long)totals.evicted_nonzero);
APPEND_NUM_FMT_STAT(fmt, n, "evicted_time",
"%u", totals.evicted_time);
APPEND_NUM_FMT_STAT(fmt, n, "outofmemory",
"%llu", (unsigned long long)totals.outofmemory);
APPEND_NUM_FMT_STAT(fmt, n, "tailrepairs",
"%llu", (unsigned long long)totals.tailrepairs);
APPEND_NUM_FMT_STAT(fmt, n, "reclaimed",
"%llu", (unsigned long long)totals.reclaimed);
APPEND_NUM_FMT_STAT(fmt, n, "expired_unfetched",
"%llu", (unsigned long long)totals.expired_unfetched);
APPEND_NUM_FMT_STAT(fmt, n, "evicted_unfetched",
"%llu", (unsigned long long)totals.evicted_unfetched);
APPEND_NUM_FMT_STAT(fmt, n, "crawler_reclaimed",
"%llu", (unsigned long long)totals.crawler_reclaimed);
APPEND_NUM_FMT_STAT(fmt, n, "crawler_items_checked",
"%llu", (unsigned long long)totals.crawler_items_checked);
APPEND_NUM_FMT_STAT(fmt, n, "lrutail_reflocked",
"%llu", (unsigned long long)totals.lrutail_reflocked);
if (settings.lru_maintainer_thread) {
APPEND_NUM_FMT_STAT(fmt, n, "moves_to_cold",
"%llu", (unsigned long long)totals.moves_to_cold);
APPEND_NUM_FMT_STAT(fmt, n, "moves_to_warm",
"%llu", (unsigned long long)totals.moves_to_warm);
APPEND_NUM_FMT_STAT(fmt, n, "moves_within_lru",
"%llu", (unsigned long long)totals.moves_within_lru);
APPEND_NUM_FMT_STAT(fmt, n, "direct_reclaims",
"%llu", (unsigned long long)totals.direct_reclaims);
}
}
/* getting here means both ascii and binary terminators fit */
add_stats(NULL, 0, NULL, 0, c);
}
/** dumps out a list of objects of each size, with granularity of 32 bytes */
/*@null@*/
/* Locks are correct based on a technicality. Holds LRU lock while doing the
* work, so items can't go invalid, and it's only looking at header sizes
* which don't change.
*/
void item_stats_sizes(ADD_STAT add_stats, void *c) {
/* max 1MB object, divided into 32 bytes size buckets */
const int num_buckets = 32768;
unsigned int *histogram = calloc(num_buckets, sizeof(int));
if (histogram != NULL) {
int i;
/* build the histogram */
for (i = 0; i < LARGEST_ID; i++) {
pthread_mutex_lock(&lru_locks[i]);
item *iter = heads[i];
while (iter) {
int ntotal = ITEM_ntotal(iter);
int bucket = ntotal / 32;
if ((ntotal % 32) != 0) bucket++;
if (bucket < num_buckets) histogram[bucket]++;
iter = iter->next;
}
pthread_mutex_unlock(&lru_locks[i]);
}
/* write the buffer */
for (i = 0; i < num_buckets; i++) {
if (histogram[i] != 0) {
char key[8];
snprintf(key, sizeof(key), "%d", i * 32);
APPEND_STAT(key, "%u", histogram[i]);
}
}
free(histogram);
}
add_stats(NULL, 0, NULL, 0, c);
}
/** wrapper around assoc_find which does the lazy expiration logic */
item *do_item_get(const char *key, const size_t nkey, const uint32_t hv) {
item *it = assoc_find(key, nkey, hv);
if (it != NULL) {
refcount_incr(&it->refcount);
/* Optimization for slab reassignment. prevents popular items from
* jamming in busy wait. Can only do this here to satisfy lock order
* of item_lock, slabs_lock. */
/* This was made unsafe by removal of the cache_lock:
* slab_rebalance_signal and slab_rebal.* are modified in a separate
* thread under slabs_lock. If slab_rebalance_signal = 1, slab_start =
* NULL (0), but slab_end is still equal to some value, this would end
* up unlinking every item fetched.
* This is either an acceptable loss, or if slab_rebalance_signal is
* true, slab_start/slab_end should be put behind the slabs_lock.
* Which would cause a huge potential slowdown.
* Could also use a specific lock for slab_rebal.* and
* slab_rebalance_signal (shorter lock?)
*/
/*if (slab_rebalance_signal &&
((void *)it >= slab_rebal.slab_start && (void *)it < slab_rebal.slab_end)) {
do_item_unlink(it, hv);
do_item_remove(it);
it = NULL;
}*/
}
int was_found = 0;
if (settings.verbose > 2) {
int ii;
if (it == NULL) {
fprintf(stderr, "> NOT FOUND ");
} else {
fprintf(stderr, "> FOUND KEY ");
was_found++;
}
for (ii = 0; ii < nkey; ++ii) {
fprintf(stderr, "%c", key[ii]);
}
}
if (it != NULL) {
if (is_flushed(it)) {
do_item_unlink(it, hv);
do_item_remove(it);
it = NULL;
if (was_found) {
fprintf(stderr, " -nuked by flush");
}
} else if (it->exptime != 0 && it->exptime <= current_time) {
do_item_unlink(it, hv);
do_item_remove(it);
it = NULL;
if (was_found) {
fprintf(stderr, " -nuked by expire");
}
} else {
it->it_flags |= ITEM_FETCHED|ITEM_ACTIVE;
DEBUG_REFCNT(it, '+');
}
}
if (settings.verbose > 2)
fprintf(stderr, "\n");
return it;
}
item *do_item_touch(const char *key, size_t nkey, uint32_t exptime,
const uint32_t hv) {
item *it = do_item_get(key, nkey, hv);
if (it != NULL) {
it->exptime = exptime;
}
return it;
}
/*** LRU MAINTENANCE THREAD ***/
/* Returns number of items remove, expired, or evicted.
* Callable from worker threads or the LRU maintainer thread */
static int lru_pull_tail(const int orig_id, const int cur_lru,
const unsigned int total_chunks, const bool do_evict, const uint32_t cur_hv) {
item *it = NULL;
int id = orig_id;
int removed = 0;
if (id == 0)
return 0;
int tries = 5;
item *search;
item *next_it;
void *hold_lock = NULL;
unsigned int move_to_lru = 0;
uint64_t limit;
id |= cur_lru;
pthread_mutex_lock(&lru_locks[id]);
search = tails[id];
/* We walk up *only* for locked items, and if bottom is expired. */
for (; tries > 0 && search != NULL; tries--, search=next_it) {
/* we might relink search mid-loop, so search->prev isn't reliable */
next_it = search->prev;
if (search->nbytes == 0 && search->nkey == 0 && search->it_flags == 1) {
/* We are a crawler, ignore it. */
tries++;
continue;
}
uint32_t hv = hash(ITEM_key(search), search->nkey);
/* Attempt to hash item lock the "search" item. If locked, no
* other callers can incr the refcount. Also skip ourselves. */
if (hv == cur_hv || (hold_lock = item_trylock(hv)) == NULL)
continue;
/* Now see if the item is refcount locked */
if (refcount_incr(&search->refcount) != 2) {
/* Note pathological case with ref'ed items in tail.
* Can still unlink the item, but it won't be reusable yet */
itemstats[id].lrutail_reflocked++;
/* In case of refcount leaks, enable for quick workaround. */
/* WARNING: This can cause terrible corruption */
if (settings.tail_repair_time &&
search->time + settings.tail_repair_time < current_time) {
itemstats[id].tailrepairs++;
search->refcount = 1;
/* This will call item_remove -> item_free since refcnt is 1 */
do_item_unlink_nolock(search, hv);
item_trylock_unlock(hold_lock);
continue;
}
}
/* Expired or flushed */
if ((search->exptime != 0 && search->exptime < current_time)
|| is_flushed(search)) {
itemstats[id].reclaimed++;
if ((search->it_flags & ITEM_FETCHED) == 0) {
itemstats[id].expired_unfetched++;
}
/* refcnt 2 -> 1 */
do_item_unlink_nolock(search, hv);
/* refcnt 1 -> 0 -> item_free */
do_item_remove(search);
item_trylock_unlock(hold_lock);
removed++;
/* If all we're finding are expired, can keep going */
continue;
}
/* If we're HOT_LRU or WARM_LRU and over size limit, send to COLD_LRU.
* If we're COLD_LRU, send to WARM_LRU unless we need to evict
*/
switch (cur_lru) {
case HOT_LRU:
limit = total_chunks * settings.hot_lru_pct / 100;
case WARM_LRU:
limit = total_chunks * settings.warm_lru_pct / 100;
if (sizes[id] > limit) {
itemstats[id].moves_to_cold++;
move_to_lru = COLD_LRU;
do_item_unlink_q(search);
it = search;
removed++;
break;
} else if ((search->it_flags & ITEM_ACTIVE) != 0) {
/* Only allow ACTIVE relinking if we're not too large. */
itemstats[id].moves_within_lru++;
search->it_flags &= ~ITEM_ACTIVE;
do_item_update_nolock(search);
do_item_remove(search);
item_trylock_unlock(hold_lock);
} else {
/* Don't want to move to COLD, not active, bail out */
it = search;
}
break;
case COLD_LRU:
it = search; /* No matter what, we're stopping */
if (do_evict) {
if (settings.evict_to_free == 0) {
/* Don't think we need a counter for this. It'll OOM. */
break;
}
itemstats[id].evicted++;
itemstats[id].evicted_time = current_time - search->time;
if (search->exptime != 0)
itemstats[id].evicted_nonzero++;
if ((search->it_flags & ITEM_FETCHED) == 0) {
itemstats[id].evicted_unfetched++;
}
do_item_unlink_nolock(search, hv);
removed++;
} else if ((search->it_flags & ITEM_ACTIVE) != 0
&& settings.lru_maintainer_thread) {
itemstats[id].moves_to_warm++;
search->it_flags &= ~ITEM_ACTIVE;
move_to_lru = WARM_LRU;
do_item_unlink_q(search);
removed++;
}
break;
}
if (it != NULL)
break;
}
pthread_mutex_unlock(&lru_locks[id]);
if (it != NULL) {
if (move_to_lru) {
it->slabs_clsid = ITEM_clsid(it);
it->slabs_clsid |= move_to_lru;
item_link_q(it);
}
do_item_remove(it);
item_trylock_unlock(hold_lock);
}
return removed;
}
/* Loop up to N times:
* If too many items are in HOT_LRU, push to COLD_LRU
* If too many items are in WARM_LRU, push to COLD_LRU
* If too many items are in COLD_LRU, poke COLD_LRU tail
* 1000 loops with 1ms min sleep gives us under 1m items shifted/sec. The
* locks can't handle much more than that. Leaving a TODO for how to
* autoadjust in the future.
*/
static int lru_maintainer_juggle(const int slabs_clsid) {
int i;
int did_moves = 0;
bool mem_limit_reached = false;
unsigned int total_chunks = 0;
/* TODO: if free_chunks below high watermark, increase aggressiveness */
slabs_available_chunks(slabs_clsid, &mem_limit_reached, &total_chunks);
if (settings.expirezero_does_not_evict)
total_chunks -= noexp_lru_size(slabs_clsid);
/* Juggle HOT/WARM up to N times */
for (i = 0; i < 1000; i++) {
int do_more = 0;
if (lru_pull_tail(slabs_clsid, HOT_LRU, total_chunks, false, 0) ||
lru_pull_tail(slabs_clsid, WARM_LRU, total_chunks, false, 0)) {
do_more++;
}
do_more += lru_pull_tail(slabs_clsid, COLD_LRU, total_chunks, false, 0);
if (do_more == 0)
break;
did_moves++;
}
return did_moves;
}
/* Will crawl all slab classes a minimum of once per hour */
#define MAX_MAINTCRAWL_WAIT 60 * 60
/* Hoping user input will improve this function. This is all a wild guess.
* Operation: Kicks crawler for each slab id. Crawlers take some statistics as
* to items with nonzero expirations. It then buckets how many items will
* expire per minute for the next hour.
* This function checks the results of a run, and if it things more than 1% of
* expirable objects are ready to go, kick the crawler again to reap.
* It will also kick the crawler once per minute regardless, waiting a minute
* longer for each time it has no work to do, up to an hour wait time.
* The latter is to avoid newly started daemons from waiting too long before
* retrying a crawl.
*/
static void lru_maintainer_crawler_check(void) {
int i;
static rel_time_t last_crawls[MAX_NUMBER_OF_SLAB_CLASSES];
static rel_time_t next_crawl_wait[MAX_NUMBER_OF_SLAB_CLASSES];
for (i = POWER_SMALLEST; i < MAX_NUMBER_OF_SLAB_CLASSES; i++) {
crawlerstats_t *s = &crawlerstats[i];
/* We've not successfully kicked off a crawl yet. */
if (last_crawls[i] == 0) {
if (lru_crawler_start(i, 0) > 0) {
last_crawls[i] = current_time;
}
}
pthread_mutex_lock(&lru_crawler_stats_lock);
if (s->run_complete) {
int x;
/* Should we crawl again? */
uint64_t possible_reclaims = s->seen - s->noexp;
uint64_t available_reclaims = 0;
/* Need to think we can free at least 1% of the items before
* crawling. */
/* FIXME: Configurable? */
uint64_t low_watermark = (s->seen / 100) + 1;
rel_time_t since_run = current_time - s->end_time;
/* Don't bother if the payoff is too low. */
if (settings.verbose > 1)
fprintf(stderr, "maint crawler: low_watermark: %llu, possible_reclaims: %llu, since_run: %u\n",
(unsigned long long)low_watermark, (unsigned long long)possible_reclaims,
(unsigned int)since_run);
for (x = 0; x < 60; x++) {
if (since_run < (x * 60) + 60)
break;
available_reclaims += s->histo[x];
}
if (available_reclaims > low_watermark) {
last_crawls[i] = 0;
if (next_crawl_wait[i] > 60)
next_crawl_wait[i] -= 60;
} else if (since_run > 5 && since_run > next_crawl_wait[i]) {
last_crawls[i] = 0;
if (next_crawl_wait[i] < MAX_MAINTCRAWL_WAIT)
next_crawl_wait[i] += 60;
}
if (settings.verbose > 1)
fprintf(stderr, "maint crawler: available reclaims: %llu, next_crawl: %u\n", (unsigned long long)available_reclaims, next_crawl_wait[i]);
}
pthread_mutex_unlock(&lru_crawler_stats_lock);
}
}
static pthread_t lru_maintainer_tid;
#define MAX_LRU_MAINTAINER_SLEEP 1000000
#define MIN_LRU_MAINTAINER_SLEEP 1000
static void *lru_maintainer_thread(void *arg) {
int i;
useconds_t to_sleep = MIN_LRU_MAINTAINER_SLEEP;
rel_time_t last_crawler_check = 0;
pthread_mutex_lock(&lru_maintainer_lock);