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
|
/*
* QEMU throttling infrastructure
*
* Copyright (C) Nodalink, EURL. 2013-2014
* Copyright (C) Igalia, S.L. 2015
*
* Authors:
* Benoît Canet <benoit.canet@nodalink.com>
* Alberto Garcia <berto@igalia.com>
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License as
* published by the Free Software Foundation; either version 2 or
* (at your option) version 3 of the License.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, see <http://www.gnu.org/licenses/>.
*/
#include "qemu/osdep.h"
#include "qapi/error.h"
#include "qemu/throttle.h"
#include "qemu/timer.h"
#include "block/aio.h"
/* This function make a bucket leak
*
* @bkt: the bucket to make leak
* @delta_ns: the time delta
*/
void throttle_leak_bucket(LeakyBucket *bkt, int64_t delta_ns)
{
double leak;
/* compute how much to leak */
leak = (bkt->avg * (double) delta_ns) / NANOSECONDS_PER_SECOND;
/* make the bucket leak */
bkt->level = MAX(bkt->level - leak, 0);
/* if we allow bursts for more than one second we also need to
* keep track of bkt->burst_level so the bkt->max goal per second
* is attained */
if (bkt->burst_length > 1) {
leak = (bkt->max * (double) delta_ns) / NANOSECONDS_PER_SECOND;
bkt->burst_level = MAX(bkt->burst_level - leak, 0);
}
}
/* Calculate the time delta since last leak and make proportionals leaks
*
* @now: the current timestamp in ns
*/
static void throttle_do_leak(ThrottleState *ts, int64_t now)
{
/* compute the time elapsed since the last leak */
int64_t delta_ns = now - ts->previous_leak;
int i;
ts->previous_leak = now;
if (delta_ns <= 0) {
return;
}
/* make each bucket leak */
for (i = 0; i < BUCKETS_COUNT; i++) {
throttle_leak_bucket(&ts->cfg.buckets[i], delta_ns);
}
}
/* do the real job of computing the time to wait
*
* @limit: the throttling limit
* @extra: the number of operation to delay
* @ret: the time to wait in ns
*/
static int64_t throttle_do_compute_wait(double limit, double extra)
{
double wait = extra * NANOSECONDS_PER_SECOND;
wait /= limit;
return wait;
}
/* This function compute the wait time in ns that a leaky bucket should trigger
*
* @bkt: the leaky bucket we operate on
* @ret: the resulting wait time in ns or 0 if the operation can go through
*/
int64_t throttle_compute_wait(LeakyBucket *bkt)
{
double extra; /* the number of extra units blocking the io */
double bucket_size; /* I/O before throttling to bkt->avg */
double burst_bucket_size; /* Before throttling to bkt->max */
if (!bkt->avg) {
return 0;
}
if (!bkt->max) {
/* If bkt->max is 0 we still want to allow short bursts of I/O
* from the guest, otherwise every other request will be throttled
* and performance will suffer considerably. */
bucket_size = (double) bkt->avg / 10;
burst_bucket_size = 0;
} else {
/* If we have a burst limit then we have to wait until all I/O
* at burst rate has finished before throttling to bkt->avg */
bucket_size = bkt->max * bkt->burst_length;
burst_bucket_size = (double) bkt->max / 10;
}
/* If the main bucket is full then we have to wait */
extra = bkt->level - bucket_size;
if (extra > 0) {
return throttle_do_compute_wait(bkt->avg, extra);
}
/* If the main bucket is not full yet we still have to check the
* burst bucket in order to enforce the burst limit */
if (bkt->burst_length > 1) {
assert(bkt->max > 0); /* see throttle_is_valid() */
extra = bkt->burst_level - burst_bucket_size;
if (extra > 0) {
return throttle_do_compute_wait(bkt->max, extra);
}
}
return 0;
}
/* This function compute the time that must be waited while this IO
*
* @is_write: true if the current IO is a write, false if it's a read
* @ret: time to wait
*/
static int64_t throttle_compute_wait_for(ThrottleState *ts,
bool is_write)
{
BucketType to_check[2][4] = { {THROTTLE_BPS_TOTAL,
THROTTLE_OPS_TOTAL,
THROTTLE_BPS_READ,
THROTTLE_OPS_READ},
{THROTTLE_BPS_TOTAL,
THROTTLE_OPS_TOTAL,
THROTTLE_BPS_WRITE,
THROTTLE_OPS_WRITE}, };
int64_t wait, max_wait = 0;
int i;
for (i = 0; i < 4; i++) {
BucketType index = to_check[is_write][i];
wait = throttle_compute_wait(&ts->cfg.buckets[index]);
if (wait > max_wait) {
max_wait = wait;
}
}
return max_wait;
}
/* compute the timer for this type of operation
*
* @is_write: the type of operation
* @now: the current clock timestamp
* @next_timestamp: the resulting timer
* @ret: true if a timer must be set
*/
static bool throttle_compute_timer(ThrottleState *ts,
bool is_write,
int64_t now,
int64_t *next_timestamp)
{
int64_t wait;
/* leak proportionally to the time elapsed */
throttle_do_leak(ts, now);
/* compute the wait time if any */
wait = throttle_compute_wait_for(ts, is_write);
/* if the code must wait compute when the next timer should fire */
if (wait) {
*next_timestamp = now + wait;
return true;
}
/* else no need to wait at all */
*next_timestamp = now;
return false;
}
/* Add timers to event loop */
void throttle_timers_attach_aio_context(ThrottleTimers *tt,
AioContext *new_context)
{
tt->timers[0] = aio_timer_new(new_context, tt->clock_type, SCALE_NS,
tt->read_timer_cb, tt->timer_opaque);
tt->timers[1] = aio_timer_new(new_context, tt->clock_type, SCALE_NS,
tt->write_timer_cb, tt->timer_opaque);
}
/*
* Initialize the ThrottleConfig structure to a valid state
* @cfg: the config to initialize
*/
void throttle_config_init(ThrottleConfig *cfg)
{
unsigned i;
memset(cfg, 0, sizeof(*cfg));
for (i = 0; i < BUCKETS_COUNT; i++) {
cfg->buckets[i].burst_length = 1;
}
}
/* To be called first on the ThrottleState */
void throttle_init(ThrottleState *ts)
{
memset(ts, 0, sizeof(ThrottleState));
throttle_config_init(&ts->cfg);
}
/* To be called first on the ThrottleTimers */
void throttle_timers_init(ThrottleTimers *tt,
AioContext *aio_context,
QEMUClockType clock_type,
QEMUTimerCB *read_timer_cb,
QEMUTimerCB *write_timer_cb,
void *timer_opaque)
{
memset(tt, 0, sizeof(ThrottleTimers));
tt->clock_type = clock_type;
tt->read_timer_cb = read_timer_cb;
tt->write_timer_cb = write_timer_cb;
tt->timer_opaque = timer_opaque;
throttle_timers_attach_aio_context(tt, aio_context);
}
/* destroy a timer */
static void throttle_timer_destroy(QEMUTimer **timer)
{
assert(*timer != NULL);
timer_free(*timer);
*timer = NULL;
}
/* Remove timers from event loop */
void throttle_timers_detach_aio_context(ThrottleTimers *tt)
{
int i;
for (i = 0; i < 2; i++) {
throttle_timer_destroy(&tt->timers[i]);
}
}
/* To be called last on the ThrottleTimers */
void throttle_timers_destroy(ThrottleTimers *tt)
{
throttle_timers_detach_aio_context(tt);
}
/* is any throttling timer configured */
bool throttle_timers_are_initialized(ThrottleTimers *tt)
{
if (tt->timers[0]) {
return true;
}
return false;
}
/* Does any throttling must be done
*
* @cfg: the throttling configuration to inspect
* @ret: true if throttling must be done else false
*/
bool throttle_enabled(ThrottleConfig *cfg)
{
int i;
for (i = 0; i < BUCKETS_COUNT; i++) {
if (cfg->buckets[i].avg > 0) {
return true;
}
}
return false;
}
/* check if a throttling configuration is valid
* @cfg: the throttling configuration to inspect
* @ret: true if valid else false
* @errp: error object
*/
bool throttle_is_valid(ThrottleConfig *cfg, Error **errp)
{
int i;
bool bps_flag, ops_flag;
bool bps_max_flag, ops_max_flag;
bps_flag = cfg->buckets[THROTTLE_BPS_TOTAL].avg &&
(cfg->buckets[THROTTLE_BPS_READ].avg ||
cfg->buckets[THROTTLE_BPS_WRITE].avg);
ops_flag = cfg->buckets[THROTTLE_OPS_TOTAL].avg &&
(cfg->buckets[THROTTLE_OPS_READ].avg ||
cfg->buckets[THROTTLE_OPS_WRITE].avg);
bps_max_flag = cfg->buckets[THROTTLE_BPS_TOTAL].max &&
(cfg->buckets[THROTTLE_BPS_READ].max ||
cfg->buckets[THROTTLE_BPS_WRITE].max);
ops_max_flag = cfg->buckets[THROTTLE_OPS_TOTAL].max &&
(cfg->buckets[THROTTLE_OPS_READ].max ||
cfg->buckets[THROTTLE_OPS_WRITE].max);
if (bps_flag || ops_flag || bps_max_flag || ops_max_flag) {
error_setg(errp, "bps/iops/max total values and read/write values"
" cannot be used at the same time");
return false;
}
if (cfg->op_size &&
!cfg->buckets[THROTTLE_OPS_TOTAL].avg &&
!cfg->buckets[THROTTLE_OPS_READ].avg &&
!cfg->buckets[THROTTLE_OPS_WRITE].avg) {
error_setg(errp, "iops size requires an iops value to be set");
return false;
}
for (i = 0; i < BUCKETS_COUNT; i++) {
LeakyBucket *bkt = &cfg->buckets[i];
if (bkt->avg > THROTTLE_VALUE_MAX || bkt->max > THROTTLE_VALUE_MAX) {
error_setg(errp, "bps/iops/max values must be within [0, %lld]",
THROTTLE_VALUE_MAX);
return false;
}
if (!bkt->burst_length) {
error_setg(errp, "the burst length cannot be 0");
return false;
}
if (bkt->burst_length > 1 && !bkt->max) {
error_setg(errp, "burst length set without burst rate");
return false;
}
if (bkt->max && bkt->burst_length > THROTTLE_VALUE_MAX / bkt->max) {
error_setg(errp, "burst length too high for this burst rate");
return false;
}
if (bkt->max && !bkt->avg) {
error_setg(errp, "bps_max/iops_max require corresponding"
" bps/iops values");
return false;
}
if (bkt->max && bkt->max < bkt->avg) {
error_setg(errp, "bps_max/iops_max cannot be lower than bps/iops");
return false;
}
}
return true;
}
/* Used to configure the throttle
*
* @ts: the throttle state we are working on
* @clock_type: the group's clock_type
* @cfg: the config to set
*/
void throttle_config(ThrottleState *ts,
QEMUClockType clock_type,
ThrottleConfig *cfg)
{
int i;
ts->cfg = *cfg;
/* Zero bucket level */
for (i = 0; i < BUCKETS_COUNT; i++) {
ts->cfg.buckets[i].level = 0;
ts->cfg.buckets[i].burst_level = 0;
}
ts->previous_leak = qemu_clock_get_ns(clock_type);
}
/* used to get config
*
* @ts: the throttle state we are working on
* @cfg: the config to write
*/
void throttle_get_config(ThrottleState *ts, ThrottleConfig *cfg)
{
*cfg = ts->cfg;
}
/* Schedule the read or write timer if needed
*
* NOTE: this function is not unit tested due to it's usage of timer_mod
*
* @tt: the timers structure
* @is_write: the type of operation (read/write)
* @ret: true if the timer has been scheduled else false
*/
bool throttle_schedule_timer(ThrottleState *ts,
ThrottleTimers *tt,
bool is_write)
{
int64_t now = qemu_clock_get_ns(tt->clock_type);
int64_t next_timestamp;
bool must_wait;
must_wait = throttle_compute_timer(ts,
is_write,
now,
&next_timestamp);
/* request not throttled */
if (!must_wait) {
return false;
}
/* request throttled and timer pending -> do nothing */
if (timer_pending(tt->timers[is_write])) {
return true;
}
/* request throttled and timer not pending -> arm timer */
timer_mod(tt->timers[is_write], next_timestamp);
return true;
}
/* do the accounting for this operation
*
* @is_write: the type of operation (read/write)
* @size: the size of the operation
*/
void throttle_account(ThrottleState *ts, bool is_write, uint64_t size)
{
const BucketType bucket_types_size[2][2] = {
{ THROTTLE_BPS_TOTAL, THROTTLE_BPS_READ },
{ THROTTLE_BPS_TOTAL, THROTTLE_BPS_WRITE }
};
const BucketType bucket_types_units[2][2] = {
{ THROTTLE_OPS_TOTAL, THROTTLE_OPS_READ },
{ THROTTLE_OPS_TOTAL, THROTTLE_OPS_WRITE }
};
double units = 1.0;
unsigned i;
/* if cfg.op_size is defined and smaller than size we compute unit count */
if (ts->cfg.op_size && size > ts->cfg.op_size) {
units = (double) size / ts->cfg.op_size;
}
for (i = 0; i < 2; i++) {
LeakyBucket *bkt;
bkt = &ts->cfg.buckets[bucket_types_size[is_write][i]];
bkt->level += size;
if (bkt->burst_length > 1) {
bkt->burst_level += size;
}
bkt = &ts->cfg.buckets[bucket_types_units[is_write][i]];
bkt->level += units;
if (bkt->burst_length > 1) {
bkt->burst_level += units;
}
}
}
/* return a ThrottleConfig based on the options in a ThrottleLimits
*
* @arg: the ThrottleLimits object to read from
* @cfg: the ThrottleConfig to edit
* @errp: error object
*/
void throttle_limits_to_config(ThrottleLimits *arg, ThrottleConfig *cfg,
Error **errp)
{
if (arg->has_bps_total) {
cfg->buckets[THROTTLE_BPS_TOTAL].avg = arg->bps_total;
}
if (arg->has_bps_read) {
cfg->buckets[THROTTLE_BPS_READ].avg = arg->bps_read;
}
if (arg->has_bps_write) {
cfg->buckets[THROTTLE_BPS_WRITE].avg = arg->bps_write;
}
if (arg->has_iops_total) {
cfg->buckets[THROTTLE_OPS_TOTAL].avg = arg->iops_total;
}
if (arg->has_iops_read) {
cfg->buckets[THROTTLE_OPS_READ].avg = arg->iops_read;
}
if (arg->has_iops_write) {
cfg->buckets[THROTTLE_OPS_WRITE].avg = arg->iops_write;
}
if (arg->has_bps_total_max) {
cfg->buckets[THROTTLE_BPS_TOTAL].max = arg->bps_total_max;
}
if (arg->has_bps_read_max) {
cfg->buckets[THROTTLE_BPS_READ].max = arg->bps_read_max;
}
if (arg->has_bps_write_max) {
cfg->buckets[THROTTLE_BPS_WRITE].max = arg->bps_write_max;
}
if (arg->has_iops_total_max) {
cfg->buckets[THROTTLE_OPS_TOTAL].max = arg->iops_total_max;
}
if (arg->has_iops_read_max) {
cfg->buckets[THROTTLE_OPS_READ].max = arg->iops_read_max;
}
if (arg->has_iops_write_max) {
cfg->buckets[THROTTLE_OPS_WRITE].max = arg->iops_write_max;
}
if (arg->has_bps_total_max_length) {
if (arg->bps_total_max_length > UINT_MAX) {
error_setg(errp, "bps-total-max-length value must be in"
" the range [0, %u]", UINT_MAX);
return;
}
cfg->buckets[THROTTLE_BPS_TOTAL].burst_length = arg->bps_total_max_length;
}
if (arg->has_bps_read_max_length) {
if (arg->bps_read_max_length > UINT_MAX) {
error_setg(errp, "bps-read-max-length value must be in"
" the range [0, %u]", UINT_MAX);
return;
}
cfg->buckets[THROTTLE_BPS_READ].burst_length = arg->bps_read_max_length;
}
if (arg->has_bps_write_max_length) {
if (arg->bps_write_max_length > UINT_MAX) {
error_setg(errp, "bps-write-max-length value must be in"
" the range [0, %u]", UINT_MAX);
return;
}
cfg->buckets[THROTTLE_BPS_WRITE].burst_length = arg->bps_write_max_length;
}
if (arg->has_iops_total_max_length) {
if (arg->iops_total_max_length > UINT_MAX) {
error_setg(errp, "iops-total-max-length value must be in"
" the range [0, %u]", UINT_MAX);
return;
}
cfg->buckets[THROTTLE_OPS_TOTAL].burst_length = arg->iops_total_max_length;
}
if (arg->has_iops_read_max_length) {
if (arg->iops_read_max_length > UINT_MAX) {
error_setg(errp, "iops-read-max-length value must be in"
" the range [0, %u]", UINT_MAX);
return;
}
cfg->buckets[THROTTLE_OPS_READ].burst_length = arg->iops_read_max_length;
}
if (arg->has_iops_write_max_length) {
if (arg->iops_write_max_length > UINT_MAX) {
error_setg(errp, "iops-write-max-length value must be in"
" the range [0, %u]", UINT_MAX);
return;
}
cfg->buckets[THROTTLE_OPS_WRITE].burst_length = arg->iops_write_max_length;
}
if (arg->has_iops_size) {
cfg->op_size = arg->iops_size;
}
throttle_is_valid(cfg, errp);
}
/* write the options of a ThrottleConfig to a ThrottleLimits
*
* @cfg: the ThrottleConfig to read from
* @var: the ThrottleLimits to write to
*/
void throttle_config_to_limits(ThrottleConfig *cfg, ThrottleLimits *var)
{
var->bps_total = cfg->buckets[THROTTLE_BPS_TOTAL].avg;
var->bps_read = cfg->buckets[THROTTLE_BPS_READ].avg;
var->bps_write = cfg->buckets[THROTTLE_BPS_WRITE].avg;
var->iops_total = cfg->buckets[THROTTLE_OPS_TOTAL].avg;
var->iops_read = cfg->buckets[THROTTLE_OPS_READ].avg;
var->iops_write = cfg->buckets[THROTTLE_OPS_WRITE].avg;
var->bps_total_max = cfg->buckets[THROTTLE_BPS_TOTAL].max;
var->bps_read_max = cfg->buckets[THROTTLE_BPS_READ].max;
var->bps_write_max = cfg->buckets[THROTTLE_BPS_WRITE].max;
var->iops_total_max = cfg->buckets[THROTTLE_OPS_TOTAL].max;
var->iops_read_max = cfg->buckets[THROTTLE_OPS_READ].max;
var->iops_write_max = cfg->buckets[THROTTLE_OPS_WRITE].max;
var->bps_total_max_length = cfg->buckets[THROTTLE_BPS_TOTAL].burst_length;
var->bps_read_max_length = cfg->buckets[THROTTLE_BPS_READ].burst_length;
var->bps_write_max_length = cfg->buckets[THROTTLE_BPS_WRITE].burst_length;
var->iops_total_max_length = cfg->buckets[THROTTLE_OPS_TOTAL].burst_length;
var->iops_read_max_length = cfg->buckets[THROTTLE_OPS_READ].burst_length;
var->iops_write_max_length = cfg->buckets[THROTTLE_OPS_WRITE].burst_length;
var->iops_size = cfg->op_size;
var->has_bps_total = true;
var->has_bps_read = true;
var->has_bps_write = true;
var->has_iops_total = true;
var->has_iops_read = true;
var->has_iops_write = true;
var->has_bps_total_max = true;
var->has_bps_read_max = true;
var->has_bps_write_max = true;
var->has_iops_total_max = true;
var->has_iops_read_max = true;
var->has_iops_write_max = true;
var->has_bps_read_max_length = true;
var->has_bps_total_max_length = true;
var->has_bps_write_max_length = true;
var->has_iops_total_max_length = true;
var->has_iops_read_max_length = true;
var->has_iops_write_max_length = true;
var->has_iops_size = true;
}
|