summaryrefslogtreecommitdiff
path: root/drivers/md/dm-stats.c
blob: db2d997a6c1815ce6f7f72fa0fed51572afae856 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
// SPDX-License-Identifier: GPL-2.0-only
#include <linux/errno.h>
#include <linux/numa.h>
#include <linux/slab.h>
#include <linux/rculist.h>
#include <linux/threads.h>
#include <linux/preempt.h>
#include <linux/irqflags.h>
#include <linux/vmalloc.h>
#include <linux/mm.h>
#include <linux/module.h>
#include <linux/device-mapper.h>

#include "dm-core.h"
#include "dm-stats.h"

#define DM_MSG_PREFIX "stats"

static int dm_stat_need_rcu_barrier;

/*
 * Using 64-bit values to avoid overflow (which is a
 * problem that block/genhd.c's IO accounting has).
 */
struct dm_stat_percpu {
	unsigned long long sectors[2];
	unsigned long long ios[2];
	unsigned long long merges[2];
	unsigned long long ticks[2];
	unsigned long long io_ticks[2];
	unsigned long long io_ticks_total;
	unsigned long long time_in_queue;
	unsigned long long *histogram;
};

struct dm_stat_shared {
	atomic_t in_flight[2];
	unsigned long long stamp;
	struct dm_stat_percpu tmp;
};

struct dm_stat {
	struct list_head list_entry;
	int id;
	unsigned int stat_flags;
	size_t n_entries;
	sector_t start;
	sector_t end;
	sector_t step;
	unsigned int n_histogram_entries;
	unsigned long long *histogram_boundaries;
	const char *program_id;
	const char *aux_data;
	struct rcu_head rcu_head;
	size_t shared_alloc_size;
	size_t percpu_alloc_size;
	size_t histogram_alloc_size;
	struct dm_stat_percpu *stat_percpu[NR_CPUS];
	struct dm_stat_shared stat_shared[];
};

#define STAT_PRECISE_TIMESTAMPS		1

struct dm_stats_last_position {
	sector_t last_sector;
	unsigned int last_rw;
};

/*
 * A typo on the command line could possibly make the kernel run out of memory
 * and crash. To prevent the crash we account all used memory. We fail if we
 * exhaust 1/4 of all memory or 1/2 of vmalloc space.
 */
#define DM_STATS_MEMORY_FACTOR		4
#define DM_STATS_VMALLOC_FACTOR		2

static DEFINE_SPINLOCK(shared_memory_lock);

static unsigned long shared_memory_amount;

static bool __check_shared_memory(size_t alloc_size)
{
	size_t a;

	a = shared_memory_amount + alloc_size;
	if (a < shared_memory_amount)
		return false;
	if (a >> PAGE_SHIFT > totalram_pages() / DM_STATS_MEMORY_FACTOR)
		return false;
#ifdef CONFIG_MMU
	if (a > (VMALLOC_END - VMALLOC_START) / DM_STATS_VMALLOC_FACTOR)
		return false;
#endif
	return true;
}

static bool check_shared_memory(size_t alloc_size)
{
	bool ret;

	spin_lock_irq(&shared_memory_lock);

	ret = __check_shared_memory(alloc_size);

	spin_unlock_irq(&shared_memory_lock);

	return ret;
}

static bool claim_shared_memory(size_t alloc_size)
{
	spin_lock_irq(&shared_memory_lock);

	if (!__check_shared_memory(alloc_size)) {
		spin_unlock_irq(&shared_memory_lock);
		return false;
	}

	shared_memory_amount += alloc_size;

	spin_unlock_irq(&shared_memory_lock);

	return true;
}

static void free_shared_memory(size_t alloc_size)
{
	unsigned long flags;

	spin_lock_irqsave(&shared_memory_lock, flags);

	if (WARN_ON_ONCE(shared_memory_amount < alloc_size)) {
		spin_unlock_irqrestore(&shared_memory_lock, flags);
		DMCRIT("Memory usage accounting bug.");
		return;
	}

	shared_memory_amount -= alloc_size;

	spin_unlock_irqrestore(&shared_memory_lock, flags);
}

static void *dm_kvzalloc(size_t alloc_size, int node)
{
	void *p;

	if (!claim_shared_memory(alloc_size))
		return NULL;

	p = kvzalloc_node(alloc_size, GFP_KERNEL | __GFP_NOMEMALLOC, node);
	if (p)
		return p;

	free_shared_memory(alloc_size);

	return NULL;
}

static void dm_kvfree(void *ptr, size_t alloc_size)
{
	if (!ptr)
		return;

	free_shared_memory(alloc_size);

	kvfree(ptr);
}

static void dm_stat_free(struct rcu_head *head)
{
	int cpu;
	struct dm_stat *s = container_of(head, struct dm_stat, rcu_head);

	kfree(s->histogram_boundaries);
	kfree(s->program_id);
	kfree(s->aux_data);
	for_each_possible_cpu(cpu) {
		dm_kvfree(s->stat_percpu[cpu][0].histogram, s->histogram_alloc_size);
		dm_kvfree(s->stat_percpu[cpu], s->percpu_alloc_size);
	}
	dm_kvfree(s->stat_shared[0].tmp.histogram, s->histogram_alloc_size);
	dm_kvfree(s, s->shared_alloc_size);
}

static int dm_stat_in_flight(struct dm_stat_shared *shared)
{
	return atomic_read(&shared->in_flight[READ]) +
	       atomic_read(&shared->in_flight[WRITE]);
}

int dm_stats_init(struct dm_stats *stats)
{
	int cpu;
	struct dm_stats_last_position *last;

	mutex_init(&stats->mutex);
	INIT_LIST_HEAD(&stats->list);
	stats->precise_timestamps = false;
	stats->last = alloc_percpu(struct dm_stats_last_position);
	if (!stats->last)
		return -ENOMEM;

	for_each_possible_cpu(cpu) {
		last = per_cpu_ptr(stats->last, cpu);
		last->last_sector = (sector_t)ULLONG_MAX;
		last->last_rw = UINT_MAX;
	}

	return 0;
}

void dm_stats_cleanup(struct dm_stats *stats)
{
	size_t ni;
	struct dm_stat *s;
	struct dm_stat_shared *shared;

	while (!list_empty(&stats->list)) {
		s = container_of(stats->list.next, struct dm_stat, list_entry);
		list_del(&s->list_entry);
		for (ni = 0; ni < s->n_entries; ni++) {
			shared = &s->stat_shared[ni];
			if (WARN_ON(dm_stat_in_flight(shared))) {
				DMCRIT("leaked in-flight counter at index %lu "
				       "(start %llu, end %llu, step %llu): reads %d, writes %d",
				       (unsigned long)ni,
				       (unsigned long long)s->start,
				       (unsigned long long)s->end,
				       (unsigned long long)s->step,
				       atomic_read(&shared->in_flight[READ]),
				       atomic_read(&shared->in_flight[WRITE]));
			}
			cond_resched();
		}
		dm_stat_free(&s->rcu_head);
	}
	free_percpu(stats->last);
	mutex_destroy(&stats->mutex);
}

static void dm_stats_recalc_precise_timestamps(struct dm_stats *stats)
{
	struct list_head *l;
	struct dm_stat *tmp_s;
	bool precise_timestamps = false;

	list_for_each(l, &stats->list) {
		tmp_s = container_of(l, struct dm_stat, list_entry);
		if (tmp_s->stat_flags & STAT_PRECISE_TIMESTAMPS) {
			precise_timestamps = true;
			break;
		}
	}
	stats->precise_timestamps = precise_timestamps;
}

static int dm_stats_create(struct dm_stats *stats, sector_t start, sector_t end,
			   sector_t step, unsigned int stat_flags,
			   unsigned int n_histogram_entries,
			   unsigned long long *histogram_boundaries,
			   const char *program_id, const char *aux_data,
			   void (*suspend_callback)(struct mapped_device *),
			   void (*resume_callback)(struct mapped_device *),
			   struct mapped_device *md)
{
	struct list_head *l;
	struct dm_stat *s, *tmp_s;
	sector_t n_entries;
	size_t ni;
	size_t shared_alloc_size;
	size_t percpu_alloc_size;
	size_t histogram_alloc_size;
	struct dm_stat_percpu *p;
	int cpu;
	int ret_id;
	int r;

	if (end < start || !step)
		return -EINVAL;

	n_entries = end - start;
	if (dm_sector_div64(n_entries, step))
		n_entries++;

	if (n_entries != (size_t)n_entries || !(size_t)(n_entries + 1))
		return -EOVERFLOW;

	shared_alloc_size = struct_size(s, stat_shared, n_entries);
	if ((shared_alloc_size - sizeof(struct dm_stat)) / sizeof(struct dm_stat_shared) != n_entries)
		return -EOVERFLOW;

	percpu_alloc_size = (size_t)n_entries * sizeof(struct dm_stat_percpu);
	if (percpu_alloc_size / sizeof(struct dm_stat_percpu) != n_entries)
		return -EOVERFLOW;

	histogram_alloc_size = (n_histogram_entries + 1) * (size_t)n_entries * sizeof(unsigned long long);
	if (histogram_alloc_size / (n_histogram_entries + 1) != (size_t)n_entries * sizeof(unsigned long long))
		return -EOVERFLOW;

	if (!check_shared_memory(shared_alloc_size + histogram_alloc_size +
				 num_possible_cpus() * (percpu_alloc_size + histogram_alloc_size)))
		return -ENOMEM;

	s = dm_kvzalloc(shared_alloc_size, NUMA_NO_NODE);
	if (!s)
		return -ENOMEM;

	s->stat_flags = stat_flags;
	s->n_entries = n_entries;
	s->start = start;
	s->end = end;
	s->step = step;
	s->shared_alloc_size = shared_alloc_size;
	s->percpu_alloc_size = percpu_alloc_size;
	s->histogram_alloc_size = histogram_alloc_size;

	s->n_histogram_entries = n_histogram_entries;
	s->histogram_boundaries = kmemdup(histogram_boundaries,
					  s->n_histogram_entries * sizeof(unsigned long long), GFP_KERNEL);
	if (!s->histogram_boundaries) {
		r = -ENOMEM;
		goto out;
	}

	s->program_id = kstrdup(program_id, GFP_KERNEL);
	if (!s->program_id) {
		r = -ENOMEM;
		goto out;
	}
	s->aux_data = kstrdup(aux_data, GFP_KERNEL);
	if (!s->aux_data) {
		r = -ENOMEM;
		goto out;
	}

	for (ni = 0; ni < n_entries; ni++) {
		atomic_set(&s->stat_shared[ni].in_flight[READ], 0);
		atomic_set(&s->stat_shared[ni].in_flight[WRITE], 0);
		cond_resched();
	}

	if (s->n_histogram_entries) {
		unsigned long long *hi;

		hi = dm_kvzalloc(s->histogram_alloc_size, NUMA_NO_NODE);
		if (!hi) {
			r = -ENOMEM;
			goto out;
		}
		for (ni = 0; ni < n_entries; ni++) {
			s->stat_shared[ni].tmp.histogram = hi;
			hi += s->n_histogram_entries + 1;
			cond_resched();
		}
	}

	for_each_possible_cpu(cpu) {
		p = dm_kvzalloc(percpu_alloc_size, cpu_to_node(cpu));
		if (!p) {
			r = -ENOMEM;
			goto out;
		}
		s->stat_percpu[cpu] = p;
		if (s->n_histogram_entries) {
			unsigned long long *hi;

			hi = dm_kvzalloc(s->histogram_alloc_size, cpu_to_node(cpu));
			if (!hi) {
				r = -ENOMEM;
				goto out;
			}
			for (ni = 0; ni < n_entries; ni++) {
				p[ni].histogram = hi;
				hi += s->n_histogram_entries + 1;
				cond_resched();
			}
		}
	}

	/*
	 * Suspend/resume to make sure there is no i/o in flight,
	 * so that newly created statistics will be exact.
	 *
	 * (note: we couldn't suspend earlier because we must not
	 * allocate memory while suspended)
	 */
	suspend_callback(md);

	mutex_lock(&stats->mutex);
	s->id = 0;
	list_for_each(l, &stats->list) {
		tmp_s = container_of(l, struct dm_stat, list_entry);
		if (WARN_ON(tmp_s->id < s->id)) {
			r = -EINVAL;
			goto out_unlock_resume;
		}
		if (tmp_s->id > s->id)
			break;
		if (unlikely(s->id == INT_MAX)) {
			r = -ENFILE;
			goto out_unlock_resume;
		}
		s->id++;
	}
	ret_id = s->id;
	list_add_tail_rcu(&s->list_entry, l);

	dm_stats_recalc_precise_timestamps(stats);

	if (!static_key_enabled(&stats_enabled.key))
		static_branch_enable(&stats_enabled);

	mutex_unlock(&stats->mutex);

	resume_callback(md);

	return ret_id;

out_unlock_resume:
	mutex_unlock(&stats->mutex);
	resume_callback(md);
out:
	dm_stat_free(&s->rcu_head);
	return r;
}

static struct dm_stat *__dm_stats_find(struct dm_stats *stats, int id)
{
	struct dm_stat *s;

	list_for_each_entry(s, &stats->list, list_entry) {
		if (s->id > id)
			break;
		if (s->id == id)
			return s;
	}

	return NULL;
}

static int dm_stats_delete(struct dm_stats *stats, int id)
{
	struct dm_stat *s;
	int cpu;

	mutex_lock(&stats->mutex);

	s = __dm_stats_find(stats, id);
	if (!s) {
		mutex_unlock(&stats->mutex);
		return -ENOENT;
	}

	list_del_rcu(&s->list_entry);

	dm_stats_recalc_precise_timestamps(stats);

	mutex_unlock(&stats->mutex);

	/*
	 * vfree can't be called from RCU callback
	 */
	for_each_possible_cpu(cpu)
		if (is_vmalloc_addr(s->stat_percpu) ||
		    is_vmalloc_addr(s->stat_percpu[cpu][0].histogram))
			goto do_sync_free;
	if (is_vmalloc_addr(s) ||
	    is_vmalloc_addr(s->stat_shared[0].tmp.histogram)) {
do_sync_free:
		synchronize_rcu_expedited();
		dm_stat_free(&s->rcu_head);
	} else {
		WRITE_ONCE(dm_stat_need_rcu_barrier, 1);
		call_rcu(&s->rcu_head, dm_stat_free);
	}
	return 0;
}

static int dm_stats_list(struct dm_stats *stats, const char *program,
			 char *result, unsigned int maxlen)
{
	struct dm_stat *s;
	sector_t len;
	unsigned int sz = 0;

	/*
	 * Output format:
	 *   <region_id>: <start_sector>+<length> <step> <program_id> <aux_data>
	 */

	mutex_lock(&stats->mutex);
	list_for_each_entry(s, &stats->list, list_entry) {
		if (!program || !strcmp(program, s->program_id)) {
			len = s->end - s->start;
			DMEMIT("%d: %llu+%llu %llu %s %s", s->id,
				(unsigned long long)s->start,
				(unsigned long long)len,
				(unsigned long long)s->step,
				s->program_id,
				s->aux_data);
			if (s->stat_flags & STAT_PRECISE_TIMESTAMPS)
				DMEMIT(" precise_timestamps");
			if (s->n_histogram_entries) {
				unsigned int i;

				DMEMIT(" histogram:");
				for (i = 0; i < s->n_histogram_entries; i++) {
					if (i)
						DMEMIT(",");
					DMEMIT("%llu", s->histogram_boundaries[i]);
				}
			}
			DMEMIT("\n");
		}
		cond_resched();
	}
	mutex_unlock(&stats->mutex);

	return 1;
}

static void dm_stat_round(struct dm_stat *s, struct dm_stat_shared *shared,
			  struct dm_stat_percpu *p)
{
	/*
	 * This is racy, but so is part_round_stats_single.
	 */
	unsigned long long now, difference;
	unsigned int in_flight_read, in_flight_write;

	if (likely(!(s->stat_flags & STAT_PRECISE_TIMESTAMPS)))
		now = jiffies;
	else
		now = ktime_to_ns(ktime_get());

	difference = now - shared->stamp;
	if (!difference)
		return;

	in_flight_read = (unsigned int)atomic_read(&shared->in_flight[READ]);
	in_flight_write = (unsigned int)atomic_read(&shared->in_flight[WRITE]);
	if (in_flight_read)
		p->io_ticks[READ] += difference;
	if (in_flight_write)
		p->io_ticks[WRITE] += difference;
	if (in_flight_read + in_flight_write) {
		p->io_ticks_total += difference;
		p->time_in_queue += (in_flight_read + in_flight_write) * difference;
	}
	shared->stamp = now;
}

static void dm_stat_for_entry(struct dm_stat *s, size_t entry,
			      int idx, sector_t len,
			      struct dm_stats_aux *stats_aux, bool end,
			      unsigned long duration_jiffies)
{
	struct dm_stat_shared *shared = &s->stat_shared[entry];
	struct dm_stat_percpu *p;

	/*
	 * For strict correctness we should use local_irq_save/restore
	 * instead of preempt_disable/enable.
	 *
	 * preempt_disable/enable is racy if the driver finishes bios
	 * from non-interrupt context as well as from interrupt context
	 * or from more different interrupts.
	 *
	 * On 64-bit architectures the race only results in not counting some
	 * events, so it is acceptable.  On 32-bit architectures the race could
	 * cause the counter going off by 2^32, so we need to do proper locking
	 * there.
	 *
	 * part_stat_lock()/part_stat_unlock() have this race too.
	 */
#if BITS_PER_LONG == 32
	unsigned long flags;

	local_irq_save(flags);
#else
	preempt_disable();
#endif
	p = &s->stat_percpu[smp_processor_id()][entry];

	if (!end) {
		dm_stat_round(s, shared, p);
		atomic_inc(&shared->in_flight[idx]);
	} else {
		unsigned long long duration;

		dm_stat_round(s, shared, p);
		atomic_dec(&shared->in_flight[idx]);
		p->sectors[idx] += len;
		p->ios[idx] += 1;
		p->merges[idx] += stats_aux->merged;
		if (!(s->stat_flags & STAT_PRECISE_TIMESTAMPS)) {
			p->ticks[idx] += duration_jiffies;
			duration = jiffies_to_msecs(duration_jiffies);
		} else {
			p->ticks[idx] += stats_aux->duration_ns;
			duration = stats_aux->duration_ns;
		}
		if (s->n_histogram_entries) {
			unsigned int lo = 0, hi = s->n_histogram_entries + 1;

			while (lo + 1 < hi) {
				unsigned int mid = (lo + hi) / 2;

				if (s->histogram_boundaries[mid - 1] > duration)
					hi = mid;
				else
					lo = mid;
			}
			p->histogram[lo]++;
		}
	}

#if BITS_PER_LONG == 32
	local_irq_restore(flags);
#else
	preempt_enable();
#endif
}

static void __dm_stat_bio(struct dm_stat *s, int bi_rw,
			  sector_t bi_sector, sector_t end_sector,
			  bool end, unsigned long duration_jiffies,
			  struct dm_stats_aux *stats_aux)
{
	sector_t rel_sector, offset, todo, fragment_len;
	size_t entry;

	if (end_sector <= s->start || bi_sector >= s->end)
		return;
	if (unlikely(bi_sector < s->start)) {
		rel_sector = 0;
		todo = end_sector - s->start;
	} else {
		rel_sector = bi_sector - s->start;
		todo = end_sector - bi_sector;
	}
	if (unlikely(end_sector > s->end))
		todo -= (end_sector - s->end);

	offset = dm_sector_div64(rel_sector, s->step);
	entry = rel_sector;
	do {
		if (WARN_ON_ONCE(entry >= s->n_entries)) {
			DMCRIT("Invalid area access in region id %d", s->id);
			return;
		}
		fragment_len = todo;
		if (fragment_len > s->step - offset)
			fragment_len = s->step - offset;
		dm_stat_for_entry(s, entry, bi_rw, fragment_len,
				  stats_aux, end, duration_jiffies);
		todo -= fragment_len;
		entry++;
		offset = 0;
	} while (unlikely(todo != 0));
}

void dm_stats_account_io(struct dm_stats *stats, unsigned long bi_rw,
			 sector_t bi_sector, unsigned int bi_sectors, bool end,
			 unsigned long start_time,
			 struct dm_stats_aux *stats_aux)
{
	struct dm_stat *s;
	sector_t end_sector;
	struct dm_stats_last_position *last;
	bool got_precise_time;
	unsigned long duration_jiffies = 0;

	if (unlikely(!bi_sectors))
		return;

	end_sector = bi_sector + bi_sectors;

	if (!end) {
		/*
		 * A race condition can at worst result in the merged flag being
		 * misrepresented, so we don't have to disable preemption here.
		 */
		last = raw_cpu_ptr(stats->last);
		stats_aux->merged =
			(bi_sector == (READ_ONCE(last->last_sector) &&
				       ((bi_rw == WRITE) ==
					(READ_ONCE(last->last_rw) == WRITE))
				       ));
		WRITE_ONCE(last->last_sector, end_sector);
		WRITE_ONCE(last->last_rw, bi_rw);
	} else
		duration_jiffies = jiffies - start_time;

	rcu_read_lock();

	got_precise_time = false;
	list_for_each_entry_rcu(s, &stats->list, list_entry) {
		if (s->stat_flags & STAT_PRECISE_TIMESTAMPS && !got_precise_time) {
			/* start (!end) duration_ns is set by DM core's alloc_io() */
			if (end)
				stats_aux->duration_ns = ktime_to_ns(ktime_get()) - stats_aux->duration_ns;
			got_precise_time = true;
		}
		__dm_stat_bio(s, bi_rw, bi_sector, end_sector, end, duration_jiffies, stats_aux);
	}

	rcu_read_unlock();
}

static void __dm_stat_init_temporary_percpu_totals(struct dm_stat_shared *shared,
						   struct dm_stat *s, size_t x)
{
	int cpu;
	struct dm_stat_percpu *p;

	local_irq_disable();
	p = &s->stat_percpu[smp_processor_id()][x];
	dm_stat_round(s, shared, p);
	local_irq_enable();

	shared->tmp.sectors[READ] = 0;
	shared->tmp.sectors[WRITE] = 0;
	shared->tmp.ios[READ] = 0;
	shared->tmp.ios[WRITE] = 0;
	shared->tmp.merges[READ] = 0;
	shared->tmp.merges[WRITE] = 0;
	shared->tmp.ticks[READ] = 0;
	shared->tmp.ticks[WRITE] = 0;
	shared->tmp.io_ticks[READ] = 0;
	shared->tmp.io_ticks[WRITE] = 0;
	shared->tmp.io_ticks_total = 0;
	shared->tmp.time_in_queue = 0;

	if (s->n_histogram_entries)
		memset(shared->tmp.histogram, 0, (s->n_histogram_entries + 1) * sizeof(unsigned long long));

	for_each_possible_cpu(cpu) {
		p = &s->stat_percpu[cpu][x];
		shared->tmp.sectors[READ] += READ_ONCE(p->sectors[READ]);
		shared->tmp.sectors[WRITE] += READ_ONCE(p->sectors[WRITE]);
		shared->tmp.ios[READ] += READ_ONCE(p->ios[READ]);
		shared->tmp.ios[WRITE] += READ_ONCE(p->ios[WRITE]);
		shared->tmp.merges[READ] += READ_ONCE(p->merges[READ]);
		shared->tmp.merges[WRITE] += READ_ONCE(p->merges[WRITE]);
		shared->tmp.ticks[READ] += READ_ONCE(p->ticks[READ]);
		shared->tmp.ticks[WRITE] += READ_ONCE(p->ticks[WRITE]);
		shared->tmp.io_ticks[READ] += READ_ONCE(p->io_ticks[READ]);
		shared->tmp.io_ticks[WRITE] += READ_ONCE(p->io_ticks[WRITE]);
		shared->tmp.io_ticks_total += READ_ONCE(p->io_ticks_total);
		shared->tmp.time_in_queue += READ_ONCE(p->time_in_queue);
		if (s->n_histogram_entries) {
			unsigned int i;

			for (i = 0; i < s->n_histogram_entries + 1; i++)
				shared->tmp.histogram[i] += READ_ONCE(p->histogram[i]);
		}
	}
}

static void __dm_stat_clear(struct dm_stat *s, size_t idx_start, size_t idx_end,
			    bool init_tmp_percpu_totals)
{
	size_t x;
	struct dm_stat_shared *shared;
	struct dm_stat_percpu *p;

	for (x = idx_start; x < idx_end; x++) {
		shared = &s->stat_shared[x];
		if (init_tmp_percpu_totals)
			__dm_stat_init_temporary_percpu_totals(shared, s, x);
		local_irq_disable();
		p = &s->stat_percpu[smp_processor_id()][x];
		p->sectors[READ] -= shared->tmp.sectors[READ];
		p->sectors[WRITE] -= shared->tmp.sectors[WRITE];
		p->ios[READ] -= shared->tmp.ios[READ];
		p->ios[WRITE] -= shared->tmp.ios[WRITE];
		p->merges[READ] -= shared->tmp.merges[READ];
		p->merges[WRITE] -= shared->tmp.merges[WRITE];
		p->ticks[READ] -= shared->tmp.ticks[READ];
		p->ticks[WRITE] -= shared->tmp.ticks[WRITE];
		p->io_ticks[READ] -= shared->tmp.io_ticks[READ];
		p->io_ticks[WRITE] -= shared->tmp.io_ticks[WRITE];
		p->io_ticks_total -= shared->tmp.io_ticks_total;
		p->time_in_queue -= shared->tmp.time_in_queue;
		local_irq_enable();
		if (s->n_histogram_entries) {
			unsigned int i;

			for (i = 0; i < s->n_histogram_entries + 1; i++) {
				local_irq_disable();
				p = &s->stat_percpu[smp_processor_id()][x];
				p->histogram[i] -= shared->tmp.histogram[i];
				local_irq_enable();
			}
		}
		cond_resched();
	}
}

static int dm_stats_clear(struct dm_stats *stats, int id)
{
	struct dm_stat *s;

	mutex_lock(&stats->mutex);

	s = __dm_stats_find(stats, id);
	if (!s) {
		mutex_unlock(&stats->mutex);
		return -ENOENT;
	}

	__dm_stat_clear(s, 0, s->n_entries, true);

	mutex_unlock(&stats->mutex);

	return 1;
}

/*
 * This is like jiffies_to_msec, but works for 64-bit values.
 */
static unsigned long long dm_jiffies_to_msec64(struct dm_stat *s, unsigned long long j)
{
	unsigned long long result;
	unsigned int mult;

	if (s->stat_flags & STAT_PRECISE_TIMESTAMPS)
		return j;

	result = 0;
	if (j)
		result = jiffies_to_msecs(j & 0x3fffff);
	if (j >= 1 << 22) {
		mult = jiffies_to_msecs(1 << 22);
		result += (unsigned long long)mult * (unsigned long long)jiffies_to_msecs((j >> 22) & 0x3fffff);
	}
	if (j >= 1ULL << 44)
		result += (unsigned long long)mult * (unsigned long long)mult * (unsigned long long)jiffies_to_msecs(j >> 44);

	return result;
}

static int dm_stats_print(struct dm_stats *stats, int id,
			  size_t idx_start, size_t idx_len,
			  bool clear, char *result, unsigned int maxlen)
{
	unsigned int sz = 0;
	struct dm_stat *s;
	size_t x;
	sector_t start, end, step;
	size_t idx_end;
	struct dm_stat_shared *shared;

	/*
	 * Output format:
	 *   <start_sector>+<length> counters
	 */

	mutex_lock(&stats->mutex);

	s = __dm_stats_find(stats, id);
	if (!s) {
		mutex_unlock(&stats->mutex);
		return -ENOENT;
	}

	idx_end = idx_start + idx_len;
	if (idx_end < idx_start ||
	    idx_end > s->n_entries)
		idx_end = s->n_entries;

	if (idx_start > idx_end)
		idx_start = idx_end;

	step = s->step;
	start = s->start + (step * idx_start);

	for (x = idx_start; x < idx_end; x++, start = end) {
		shared = &s->stat_shared[x];
		end = start + step;
		if (unlikely(end > s->end))
			end = s->end;

		__dm_stat_init_temporary_percpu_totals(shared, s, x);

		DMEMIT("%llu+%llu %llu %llu %llu %llu %llu %llu %llu %llu %d %llu %llu %llu %llu",
		       (unsigned long long)start,
		       (unsigned long long)step,
		       shared->tmp.ios[READ],
		       shared->tmp.merges[READ],
		       shared->tmp.sectors[READ],
		       dm_jiffies_to_msec64(s, shared->tmp.ticks[READ]),
		       shared->tmp.ios[WRITE],
		       shared->tmp.merges[WRITE],
		       shared->tmp.sectors[WRITE],
		       dm_jiffies_to_msec64(s, shared->tmp.ticks[WRITE]),
		       dm_stat_in_flight(shared),
		       dm_jiffies_to_msec64(s, shared->tmp.io_ticks_total),
		       dm_jiffies_to_msec64(s, shared->tmp.time_in_queue),
		       dm_jiffies_to_msec64(s, shared->tmp.io_ticks[READ]),
		       dm_jiffies_to_msec64(s, shared->tmp.io_ticks[WRITE]));
		if (s->n_histogram_entries) {
			unsigned int i;

			for (i = 0; i < s->n_histogram_entries + 1; i++)
				DMEMIT("%s%llu", !i ? " " : ":", shared->tmp.histogram[i]);
		}
		DMEMIT("\n");

		if (unlikely(sz + 1 >= maxlen))
			goto buffer_overflow;

		cond_resched();
	}

	if (clear)
		__dm_stat_clear(s, idx_start, idx_end, false);

buffer_overflow:
	mutex_unlock(&stats->mutex);

	return 1;
}

static int dm_stats_set_aux(struct dm_stats *stats, int id, const char *aux_data)
{
	struct dm_stat *s;
	const char *new_aux_data;

	mutex_lock(&stats->mutex);

	s = __dm_stats_find(stats, id);
	if (!s) {
		mutex_unlock(&stats->mutex);
		return -ENOENT;
	}

	new_aux_data = kstrdup(aux_data, GFP_KERNEL);
	if (!new_aux_data) {
		mutex_unlock(&stats->mutex);
		return -ENOMEM;
	}

	kfree(s->aux_data);
	s->aux_data = new_aux_data;

	mutex_unlock(&stats->mutex);

	return 0;
}

static int parse_histogram(const char *h, unsigned int *n_histogram_entries,
			   unsigned long long **histogram_boundaries)
{
	const char *q;
	unsigned int n;
	unsigned long long last;

	*n_histogram_entries = 1;
	for (q = h; *q; q++)
		if (*q == ',')
			(*n_histogram_entries)++;

	*histogram_boundaries = kmalloc_array(*n_histogram_entries,
					      sizeof(unsigned long long),
					      GFP_KERNEL);
	if (!*histogram_boundaries)
		return -ENOMEM;

	n = 0;
	last = 0;
	while (1) {
		unsigned long long hi;
		int s;
		char ch;

		s = sscanf(h, "%llu%c", &hi, &ch);
		if (!s || (s == 2 && ch != ','))
			return -EINVAL;
		if (hi <= last)
			return -EINVAL;
		last = hi;
		(*histogram_boundaries)[n] = hi;
		if (s == 1)
			return 0;
		h = strchr(h, ',') + 1;
		n++;
	}
}

static int message_stats_create(struct mapped_device *md,
				unsigned int argc, char **argv,
				char *result, unsigned int maxlen)
{
	int r;
	int id;
	char dummy;
	unsigned long long start, end, len, step;
	unsigned int divisor;
	const char *program_id, *aux_data;
	unsigned int stat_flags = 0;
	unsigned int n_histogram_entries = 0;
	unsigned long long *histogram_boundaries = NULL;
	struct dm_arg_set as, as_backup;
	const char *a;
	unsigned int feature_args;

	/*
	 * Input format:
	 *   <range> <step> [<extra_parameters> <parameters>] [<program_id> [<aux_data>]]
	 */

	if (argc < 3)
		goto ret_einval;

	as.argc = argc;
	as.argv = argv;
	dm_consume_args(&as, 1);

	a = dm_shift_arg(&as);
	if (!strcmp(a, "-")) {
		start = 0;
		len = dm_get_size(md);
		if (!len)
			len = 1;
	} else if (sscanf(a, "%llu+%llu%c", &start, &len, &dummy) != 2 ||
		   start != (sector_t)start || len != (sector_t)len)
		goto ret_einval;

	end = start + len;
	if (start >= end)
		goto ret_einval;

	a = dm_shift_arg(&as);
	if (sscanf(a, "/%u%c", &divisor, &dummy) == 1) {
		if (!divisor)
			return -EINVAL;
		step = end - start;
		if (do_div(step, divisor))
			step++;
		if (!step)
			step = 1;
	} else if (sscanf(a, "%llu%c", &step, &dummy) != 1 ||
		   step != (sector_t)step || !step)
		goto ret_einval;

	as_backup = as;
	a = dm_shift_arg(&as);
	if (a && sscanf(a, "%u%c", &feature_args, &dummy) == 1) {
		while (feature_args--) {
			a = dm_shift_arg(&as);
			if (!a)
				goto ret_einval;
			if (!strcasecmp(a, "precise_timestamps"))
				stat_flags |= STAT_PRECISE_TIMESTAMPS;
			else if (!strncasecmp(a, "histogram:", 10)) {
				if (n_histogram_entries)
					goto ret_einval;
				r = parse_histogram(a + 10, &n_histogram_entries, &histogram_boundaries);
				if (r)
					goto ret;
			} else
				goto ret_einval;
		}
	} else {
		as = as_backup;
	}

	program_id = "-";
	aux_data = "-";

	a = dm_shift_arg(&as);
	if (a)
		program_id = a;

	a = dm_shift_arg(&as);
	if (a)
		aux_data = a;

	if (as.argc)
		goto ret_einval;

	/*
	 * If a buffer overflow happens after we created the region,
	 * it's too late (the userspace would retry with a larger
	 * buffer, but the region id that caused the overflow is already
	 * leaked).  So we must detect buffer overflow in advance.
	 */
	snprintf(result, maxlen, "%d", INT_MAX);
	if (dm_message_test_buffer_overflow(result, maxlen)) {
		r = 1;
		goto ret;
	}

	id = dm_stats_create(dm_get_stats(md), start, end, step, stat_flags,
			     n_histogram_entries, histogram_boundaries, program_id, aux_data,
			     dm_internal_suspend_fast, dm_internal_resume_fast, md);
	if (id < 0) {
		r = id;
		goto ret;
	}

	snprintf(result, maxlen, "%d", id);

	r = 1;
	goto ret;

ret_einval:
	r = -EINVAL;
ret:
	kfree(histogram_boundaries);
	return r;
}

static int message_stats_delete(struct mapped_device *md,
				unsigned int argc, char **argv)
{
	int id;
	char dummy;

	if (argc != 2)
		return -EINVAL;

	if (sscanf(argv[1], "%d%c", &id, &dummy) != 1 || id < 0)
		return -EINVAL;

	return dm_stats_delete(dm_get_stats(md), id);
}

static int message_stats_clear(struct mapped_device *md,
			       unsigned int argc, char **argv)
{
	int id;
	char dummy;

	if (argc != 2)
		return -EINVAL;

	if (sscanf(argv[1], "%d%c", &id, &dummy) != 1 || id < 0)
		return -EINVAL;

	return dm_stats_clear(dm_get_stats(md), id);
}

static int message_stats_list(struct mapped_device *md,
			      unsigned int argc, char **argv,
			      char *result, unsigned int maxlen)
{
	int r;
	const char *program = NULL;

	if (argc < 1 || argc > 2)
		return -EINVAL;

	if (argc > 1) {
		program = kstrdup(argv[1], GFP_KERNEL);
		if (!program)
			return -ENOMEM;
	}

	r = dm_stats_list(dm_get_stats(md), program, result, maxlen);

	kfree(program);

	return r;
}

static int message_stats_print(struct mapped_device *md,
			       unsigned int argc, char **argv, bool clear,
			       char *result, unsigned int maxlen)
{
	int id;
	char dummy;
	unsigned long idx_start = 0, idx_len = ULONG_MAX;

	if (argc != 2 && argc != 4)
		return -EINVAL;

	if (sscanf(argv[1], "%d%c", &id, &dummy) != 1 || id < 0)
		return -EINVAL;

	if (argc > 3) {
		if (strcmp(argv[2], "-") &&
		    sscanf(argv[2], "%lu%c", &idx_start, &dummy) != 1)
			return -EINVAL;
		if (strcmp(argv[3], "-") &&
		    sscanf(argv[3], "%lu%c", &idx_len, &dummy) != 1)
			return -EINVAL;
	}

	return dm_stats_print(dm_get_stats(md), id, idx_start, idx_len, clear,
			      result, maxlen);
}

static int message_stats_set_aux(struct mapped_device *md,
				 unsigned int argc, char **argv)
{
	int id;
	char dummy;

	if (argc != 3)
		return -EINVAL;

	if (sscanf(argv[1], "%d%c", &id, &dummy) != 1 || id < 0)
		return -EINVAL;

	return dm_stats_set_aux(dm_get_stats(md), id, argv[2]);
}

int dm_stats_message(struct mapped_device *md, unsigned int argc, char **argv,
		     char *result, unsigned int maxlen)
{
	int r;

	/* All messages here must start with '@' */
	if (!strcasecmp(argv[0], "@stats_create"))
		r = message_stats_create(md, argc, argv, result, maxlen);
	else if (!strcasecmp(argv[0], "@stats_delete"))
		r = message_stats_delete(md, argc, argv);
	else if (!strcasecmp(argv[0], "@stats_clear"))
		r = message_stats_clear(md, argc, argv);
	else if (!strcasecmp(argv[0], "@stats_list"))
		r = message_stats_list(md, argc, argv, result, maxlen);
	else if (!strcasecmp(argv[0], "@stats_print"))
		r = message_stats_print(md, argc, argv, false, result, maxlen);
	else if (!strcasecmp(argv[0], "@stats_print_clear"))
		r = message_stats_print(md, argc, argv, true, result, maxlen);
	else if (!strcasecmp(argv[0], "@stats_set_aux"))
		r = message_stats_set_aux(md, argc, argv);
	else
		return 2; /* this wasn't a stats message */

	if (r == -EINVAL)
		DMCRIT("Invalid parameters for message %s", argv[0]);

	return r;
}

int __init dm_statistics_init(void)
{
	shared_memory_amount = 0;
	dm_stat_need_rcu_barrier = 0;
	return 0;
}

void dm_statistics_exit(void)
{
	if (dm_stat_need_rcu_barrier)
		rcu_barrier();
	if (WARN_ON(shared_memory_amount))
		DMCRIT("shared_memory_amount leaked: %lu", shared_memory_amount);
}

module_param_named(stats_current_allocated_bytes, shared_memory_amount, ulong, 0444);
MODULE_PARM_DESC(stats_current_allocated_bytes, "Memory currently used by statistics");