1 // SPDX-License-Identifier: GPL-2.0
2 /*
3 * Data Access Monitor
4 *
5 * Author: SeongJae Park <sjpark@amazon.de>
6 */
7
8 #define pr_fmt(fmt) "damon: " fmt
9
10 #include <linux/damon.h>
11 #include <linux/delay.h>
12 #include <linux/kthread.h>
13 #include <linux/mm.h>
14 #include <linux/slab.h>
15 #include <linux/string.h>
16
17 #define CREATE_TRACE_POINTS
18 #include <trace/events/damon.h>
19
20 #ifdef CONFIG_DAMON_KUNIT_TEST
21 #undef DAMON_MIN_REGION
22 #define DAMON_MIN_REGION 1
23 #endif
24
25 static DEFINE_MUTEX(damon_lock);
26 static int nr_running_ctxs;
27
28 /*
29 * Construct a damon_region struct
30 *
31 * Returns the pointer to the new struct if success, or NULL otherwise
32 */
damon_new_region(unsigned long start,unsigned long end)33 struct damon_region *damon_new_region(unsigned long start, unsigned long end)
34 {
35 struct damon_region *region;
36
37 region = kmalloc(sizeof(*region), GFP_KERNEL);
38 if (!region)
39 return NULL;
40
41 region->ar.start = start;
42 region->ar.end = end;
43 region->nr_accesses = 0;
44 INIT_LIST_HEAD(®ion->list);
45
46 region->age = 0;
47 region->last_nr_accesses = 0;
48
49 return region;
50 }
51
damon_add_region(struct damon_region * r,struct damon_target * t)52 void damon_add_region(struct damon_region *r, struct damon_target *t)
53 {
54 list_add_tail(&r->list, &t->regions_list);
55 t->nr_regions++;
56 }
57
damon_del_region(struct damon_region * r,struct damon_target * t)58 static void damon_del_region(struct damon_region *r, struct damon_target *t)
59 {
60 list_del(&r->list);
61 t->nr_regions--;
62 }
63
damon_free_region(struct damon_region * r)64 static void damon_free_region(struct damon_region *r)
65 {
66 kfree(r);
67 }
68
damon_destroy_region(struct damon_region * r,struct damon_target * t)69 void damon_destroy_region(struct damon_region *r, struct damon_target *t)
70 {
71 damon_del_region(r, t);
72 damon_free_region(r);
73 }
74
damon_new_scheme(unsigned long min_sz_region,unsigned long max_sz_region,unsigned int min_nr_accesses,unsigned int max_nr_accesses,unsigned int min_age_region,unsigned int max_age_region,enum damos_action action,struct damos_quota * quota,struct damos_watermarks * wmarks)75 struct damos *damon_new_scheme(
76 unsigned long min_sz_region, unsigned long max_sz_region,
77 unsigned int min_nr_accesses, unsigned int max_nr_accesses,
78 unsigned int min_age_region, unsigned int max_age_region,
79 enum damos_action action, struct damos_quota *quota,
80 struct damos_watermarks *wmarks)
81 {
82 struct damos *scheme;
83
84 scheme = kmalloc(sizeof(*scheme), GFP_KERNEL);
85 if (!scheme)
86 return NULL;
87 scheme->min_sz_region = min_sz_region;
88 scheme->max_sz_region = max_sz_region;
89 scheme->min_nr_accesses = min_nr_accesses;
90 scheme->max_nr_accesses = max_nr_accesses;
91 scheme->min_age_region = min_age_region;
92 scheme->max_age_region = max_age_region;
93 scheme->action = action;
94 scheme->stat = (struct damos_stat){};
95 INIT_LIST_HEAD(&scheme->list);
96
97 scheme->quota.ms = quota->ms;
98 scheme->quota.sz = quota->sz;
99 scheme->quota.reset_interval = quota->reset_interval;
100 scheme->quota.weight_sz = quota->weight_sz;
101 scheme->quota.weight_nr_accesses = quota->weight_nr_accesses;
102 scheme->quota.weight_age = quota->weight_age;
103 scheme->quota.total_charged_sz = 0;
104 scheme->quota.total_charged_ns = 0;
105 scheme->quota.esz = 0;
106 scheme->quota.charged_sz = 0;
107 scheme->quota.charged_from = 0;
108 scheme->quota.charge_target_from = NULL;
109 scheme->quota.charge_addr_from = 0;
110
111 scheme->wmarks.metric = wmarks->metric;
112 scheme->wmarks.interval = wmarks->interval;
113 scheme->wmarks.high = wmarks->high;
114 scheme->wmarks.mid = wmarks->mid;
115 scheme->wmarks.low = wmarks->low;
116 scheme->wmarks.activated = true;
117
118 return scheme;
119 }
120
damon_add_scheme(struct damon_ctx * ctx,struct damos * s)121 void damon_add_scheme(struct damon_ctx *ctx, struct damos *s)
122 {
123 list_add_tail(&s->list, &ctx->schemes);
124 }
125
damon_del_scheme(struct damos * s)126 static void damon_del_scheme(struct damos *s)
127 {
128 list_del(&s->list);
129 }
130
damon_free_scheme(struct damos * s)131 static void damon_free_scheme(struct damos *s)
132 {
133 kfree(s);
134 }
135
damon_destroy_scheme(struct damos * s)136 void damon_destroy_scheme(struct damos *s)
137 {
138 damon_del_scheme(s);
139 damon_free_scheme(s);
140 }
141
142 /*
143 * Construct a damon_target struct
144 *
145 * Returns the pointer to the new struct if success, or NULL otherwise
146 */
damon_new_target(unsigned long id)147 struct damon_target *damon_new_target(unsigned long id)
148 {
149 struct damon_target *t;
150
151 t = kmalloc(sizeof(*t), GFP_KERNEL);
152 if (!t)
153 return NULL;
154
155 t->id = id;
156 t->nr_regions = 0;
157 INIT_LIST_HEAD(&t->regions_list);
158
159 return t;
160 }
161
damon_add_target(struct damon_ctx * ctx,struct damon_target * t)162 void damon_add_target(struct damon_ctx *ctx, struct damon_target *t)
163 {
164 list_add_tail(&t->list, &ctx->adaptive_targets);
165 }
166
damon_targets_empty(struct damon_ctx * ctx)167 bool damon_targets_empty(struct damon_ctx *ctx)
168 {
169 return list_empty(&ctx->adaptive_targets);
170 }
171
damon_del_target(struct damon_target * t)172 static void damon_del_target(struct damon_target *t)
173 {
174 list_del(&t->list);
175 }
176
damon_free_target(struct damon_target * t)177 void damon_free_target(struct damon_target *t)
178 {
179 struct damon_region *r, *next;
180
181 damon_for_each_region_safe(r, next, t)
182 damon_free_region(r);
183 kfree(t);
184 }
185
damon_destroy_target(struct damon_target * t)186 void damon_destroy_target(struct damon_target *t)
187 {
188 damon_del_target(t);
189 damon_free_target(t);
190 }
191
damon_nr_regions(struct damon_target * t)192 unsigned int damon_nr_regions(struct damon_target *t)
193 {
194 return t->nr_regions;
195 }
196
damon_new_ctx(void)197 struct damon_ctx *damon_new_ctx(void)
198 {
199 struct damon_ctx *ctx;
200
201 ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
202 if (!ctx)
203 return NULL;
204
205 ctx->sample_interval = 5 * 1000;
206 ctx->aggr_interval = 100 * 1000;
207 ctx->primitive_update_interval = 60 * 1000 * 1000;
208
209 ktime_get_coarse_ts64(&ctx->last_aggregation);
210 ctx->last_primitive_update = ctx->last_aggregation;
211
212 mutex_init(&ctx->kdamond_lock);
213
214 ctx->min_nr_regions = 10;
215 ctx->max_nr_regions = 1000;
216
217 INIT_LIST_HEAD(&ctx->adaptive_targets);
218 INIT_LIST_HEAD(&ctx->schemes);
219
220 return ctx;
221 }
222
damon_destroy_targets(struct damon_ctx * ctx)223 static void damon_destroy_targets(struct damon_ctx *ctx)
224 {
225 struct damon_target *t, *next_t;
226
227 if (ctx->primitive.cleanup) {
228 ctx->primitive.cleanup(ctx);
229 return;
230 }
231
232 damon_for_each_target_safe(t, next_t, ctx)
233 damon_destroy_target(t);
234 }
235
damon_destroy_ctx(struct damon_ctx * ctx)236 void damon_destroy_ctx(struct damon_ctx *ctx)
237 {
238 struct damos *s, *next_s;
239
240 damon_destroy_targets(ctx);
241
242 damon_for_each_scheme_safe(s, next_s, ctx)
243 damon_destroy_scheme(s);
244
245 kfree(ctx);
246 }
247
248 /**
249 * damon_set_targets() - Set monitoring targets.
250 * @ctx: monitoring context
251 * @ids: array of target ids
252 * @nr_ids: number of entries in @ids
253 *
254 * This function should not be called while the kdamond is running.
255 *
256 * Return: 0 on success, negative error code otherwise.
257 */
damon_set_targets(struct damon_ctx * ctx,unsigned long * ids,ssize_t nr_ids)258 int damon_set_targets(struct damon_ctx *ctx,
259 unsigned long *ids, ssize_t nr_ids)
260 {
261 ssize_t i;
262 struct damon_target *t, *next;
263
264 damon_destroy_targets(ctx);
265
266 for (i = 0; i < nr_ids; i++) {
267 t = damon_new_target(ids[i]);
268 if (!t) {
269 /* The caller should do cleanup of the ids itself */
270 damon_for_each_target_safe(t, next, ctx)
271 damon_destroy_target(t);
272 return -ENOMEM;
273 }
274 damon_add_target(ctx, t);
275 }
276
277 return 0;
278 }
279
280 /**
281 * damon_set_attrs() - Set attributes for the monitoring.
282 * @ctx: monitoring context
283 * @sample_int: time interval between samplings
284 * @aggr_int: time interval between aggregations
285 * @primitive_upd_int: time interval between monitoring primitive updates
286 * @min_nr_reg: minimal number of regions
287 * @max_nr_reg: maximum number of regions
288 *
289 * This function should not be called while the kdamond is running.
290 * Every time interval is in micro-seconds.
291 *
292 * Return: 0 on success, negative error code otherwise.
293 */
damon_set_attrs(struct damon_ctx * ctx,unsigned long sample_int,unsigned long aggr_int,unsigned long primitive_upd_int,unsigned long min_nr_reg,unsigned long max_nr_reg)294 int damon_set_attrs(struct damon_ctx *ctx, unsigned long sample_int,
295 unsigned long aggr_int, unsigned long primitive_upd_int,
296 unsigned long min_nr_reg, unsigned long max_nr_reg)
297 {
298 if (min_nr_reg < 3)
299 return -EINVAL;
300 if (min_nr_reg > max_nr_reg)
301 return -EINVAL;
302
303 ctx->sample_interval = sample_int;
304 ctx->aggr_interval = aggr_int;
305 ctx->primitive_update_interval = primitive_upd_int;
306 ctx->min_nr_regions = min_nr_reg;
307 ctx->max_nr_regions = max_nr_reg;
308
309 return 0;
310 }
311
312 /**
313 * damon_set_schemes() - Set data access monitoring based operation schemes.
314 * @ctx: monitoring context
315 * @schemes: array of the schemes
316 * @nr_schemes: number of entries in @schemes
317 *
318 * This function should not be called while the kdamond of the context is
319 * running.
320 *
321 * Return: 0 if success, or negative error code otherwise.
322 */
damon_set_schemes(struct damon_ctx * ctx,struct damos ** schemes,ssize_t nr_schemes)323 int damon_set_schemes(struct damon_ctx *ctx, struct damos **schemes,
324 ssize_t nr_schemes)
325 {
326 struct damos *s, *next;
327 ssize_t i;
328
329 damon_for_each_scheme_safe(s, next, ctx)
330 damon_destroy_scheme(s);
331 for (i = 0; i < nr_schemes; i++)
332 damon_add_scheme(ctx, schemes[i]);
333 return 0;
334 }
335
336 /**
337 * damon_nr_running_ctxs() - Return number of currently running contexts.
338 */
damon_nr_running_ctxs(void)339 int damon_nr_running_ctxs(void)
340 {
341 int nr_ctxs;
342
343 mutex_lock(&damon_lock);
344 nr_ctxs = nr_running_ctxs;
345 mutex_unlock(&damon_lock);
346
347 return nr_ctxs;
348 }
349
350 /* Returns the size upper limit for each monitoring region */
damon_region_sz_limit(struct damon_ctx * ctx)351 static unsigned long damon_region_sz_limit(struct damon_ctx *ctx)
352 {
353 struct damon_target *t;
354 struct damon_region *r;
355 unsigned long sz = 0;
356
357 damon_for_each_target(t, ctx) {
358 damon_for_each_region(r, t)
359 sz += r->ar.end - r->ar.start;
360 }
361
362 if (ctx->min_nr_regions)
363 sz /= ctx->min_nr_regions;
364 if (sz < DAMON_MIN_REGION)
365 sz = DAMON_MIN_REGION;
366
367 return sz;
368 }
369
370 static int kdamond_fn(void *data);
371
372 /*
373 * __damon_start() - Starts monitoring with given context.
374 * @ctx: monitoring context
375 *
376 * This function should be called while damon_lock is hold.
377 *
378 * Return: 0 on success, negative error code otherwise.
379 */
__damon_start(struct damon_ctx * ctx)380 static int __damon_start(struct damon_ctx *ctx)
381 {
382 int err = -EBUSY;
383
384 mutex_lock(&ctx->kdamond_lock);
385 if (!ctx->kdamond) {
386 err = 0;
387 ctx->kdamond = kthread_run(kdamond_fn, ctx, "kdamond.%d",
388 nr_running_ctxs);
389 if (IS_ERR(ctx->kdamond)) {
390 err = PTR_ERR(ctx->kdamond);
391 ctx->kdamond = NULL;
392 }
393 }
394 mutex_unlock(&ctx->kdamond_lock);
395
396 return err;
397 }
398
399 /**
400 * damon_start() - Starts the monitorings for a given group of contexts.
401 * @ctxs: an array of the pointers for contexts to start monitoring
402 * @nr_ctxs: size of @ctxs
403 *
404 * This function starts a group of monitoring threads for a group of monitoring
405 * contexts. One thread per each context is created and run in parallel. The
406 * caller should handle synchronization between the threads by itself. If a
407 * group of threads that created by other 'damon_start()' call is currently
408 * running, this function does nothing but returns -EBUSY.
409 *
410 * Return: 0 on success, negative error code otherwise.
411 */
damon_start(struct damon_ctx ** ctxs,int nr_ctxs)412 int damon_start(struct damon_ctx **ctxs, int nr_ctxs)
413 {
414 int i;
415 int err = 0;
416
417 mutex_lock(&damon_lock);
418 if (nr_running_ctxs) {
419 mutex_unlock(&damon_lock);
420 return -EBUSY;
421 }
422
423 for (i = 0; i < nr_ctxs; i++) {
424 err = __damon_start(ctxs[i]);
425 if (err)
426 break;
427 nr_running_ctxs++;
428 }
429 mutex_unlock(&damon_lock);
430
431 return err;
432 }
433
kdamond_usleep(unsigned long usecs)434 static void kdamond_usleep(unsigned long usecs)
435 {
436 /* See Documentation/timers/timers-howto.rst for the thresholds */
437 if (usecs > 20 * USEC_PER_MSEC)
438 schedule_timeout_idle(usecs_to_jiffies(usecs));
439 else
440 usleep_idle_range(usecs, usecs + 1);
441 }
442
443 /*
444 * __damon_stop() - Stops monitoring of given context.
445 * @ctx: monitoring context
446 *
447 * Return: 0 on success, negative error code otherwise.
448 */
__damon_stop(struct damon_ctx * ctx)449 static int __damon_stop(struct damon_ctx *ctx)
450 {
451 struct task_struct *tsk;
452
453 mutex_lock(&ctx->kdamond_lock);
454 tsk = ctx->kdamond;
455 if (tsk) {
456 get_task_struct(tsk);
457 mutex_unlock(&ctx->kdamond_lock);
458 kthread_stop(tsk);
459 put_task_struct(tsk);
460 return 0;
461 }
462 mutex_unlock(&ctx->kdamond_lock);
463
464 return -EPERM;
465 }
466
467 /**
468 * damon_stop() - Stops the monitorings for a given group of contexts.
469 * @ctxs: an array of the pointers for contexts to stop monitoring
470 * @nr_ctxs: size of @ctxs
471 *
472 * Return: 0 on success, negative error code otherwise.
473 */
damon_stop(struct damon_ctx ** ctxs,int nr_ctxs)474 int damon_stop(struct damon_ctx **ctxs, int nr_ctxs)
475 {
476 int i, err = 0;
477
478 for (i = 0; i < nr_ctxs; i++) {
479 /* nr_running_ctxs is decremented in kdamond_fn */
480 err = __damon_stop(ctxs[i]);
481 if (err)
482 return err;
483 }
484
485 return err;
486 }
487
488 /*
489 * damon_check_reset_time_interval() - Check if a time interval is elapsed.
490 * @baseline: the time to check whether the interval has elapsed since
491 * @interval: the time interval (microseconds)
492 *
493 * See whether the given time interval has passed since the given baseline
494 * time. If so, it also updates the baseline to current time for next check.
495 *
496 * Return: true if the time interval has passed, or false otherwise.
497 */
damon_check_reset_time_interval(struct timespec64 * baseline,unsigned long interval)498 static bool damon_check_reset_time_interval(struct timespec64 *baseline,
499 unsigned long interval)
500 {
501 struct timespec64 now;
502
503 ktime_get_coarse_ts64(&now);
504 if ((timespec64_to_ns(&now) - timespec64_to_ns(baseline)) <
505 interval * 1000)
506 return false;
507 *baseline = now;
508 return true;
509 }
510
511 /*
512 * Check whether it is time to flush the aggregated information
513 */
kdamond_aggregate_interval_passed(struct damon_ctx * ctx)514 static bool kdamond_aggregate_interval_passed(struct damon_ctx *ctx)
515 {
516 return damon_check_reset_time_interval(&ctx->last_aggregation,
517 ctx->aggr_interval);
518 }
519
520 /*
521 * Reset the aggregated monitoring results ('nr_accesses' of each region).
522 */
kdamond_reset_aggregated(struct damon_ctx * c)523 static void kdamond_reset_aggregated(struct damon_ctx *c)
524 {
525 struct damon_target *t;
526 unsigned int ti = 0; /* target's index */
527
528 damon_for_each_target(t, c) {
529 struct damon_region *r;
530
531 damon_for_each_region(r, t) {
532 trace_damon_aggregated(t, ti, r, damon_nr_regions(t));
533 r->last_nr_accesses = r->nr_accesses;
534 r->nr_accesses = 0;
535 }
536 ti++;
537 }
538 }
539
540 static void damon_split_region_at(struct damon_ctx *ctx,
541 struct damon_target *t, struct damon_region *r,
542 unsigned long sz_r);
543
__damos_valid_target(struct damon_region * r,struct damos * s)544 static bool __damos_valid_target(struct damon_region *r, struct damos *s)
545 {
546 unsigned long sz;
547
548 sz = r->ar.end - r->ar.start;
549 return s->min_sz_region <= sz && sz <= s->max_sz_region &&
550 s->min_nr_accesses <= r->nr_accesses &&
551 r->nr_accesses <= s->max_nr_accesses &&
552 s->min_age_region <= r->age && r->age <= s->max_age_region;
553 }
554
damos_valid_target(struct damon_ctx * c,struct damon_target * t,struct damon_region * r,struct damos * s)555 static bool damos_valid_target(struct damon_ctx *c, struct damon_target *t,
556 struct damon_region *r, struct damos *s)
557 {
558 bool ret = __damos_valid_target(r, s);
559
560 if (!ret || !s->quota.esz || !c->primitive.get_scheme_score)
561 return ret;
562
563 return c->primitive.get_scheme_score(c, t, r, s) >= s->quota.min_score;
564 }
565
damon_do_apply_schemes(struct damon_ctx * c,struct damon_target * t,struct damon_region * r)566 static void damon_do_apply_schemes(struct damon_ctx *c,
567 struct damon_target *t,
568 struct damon_region *r)
569 {
570 struct damos *s;
571
572 damon_for_each_scheme(s, c) {
573 struct damos_quota *quota = &s->quota;
574 unsigned long sz = r->ar.end - r->ar.start;
575 struct timespec64 begin, end;
576 unsigned long sz_applied = 0;
577
578 if (!s->wmarks.activated)
579 continue;
580
581 /* Check the quota */
582 if (quota->esz && quota->charged_sz >= quota->esz)
583 continue;
584
585 /* Skip previously charged regions */
586 if (quota->charge_target_from) {
587 if (t != quota->charge_target_from)
588 continue;
589 if (r == damon_last_region(t)) {
590 quota->charge_target_from = NULL;
591 quota->charge_addr_from = 0;
592 continue;
593 }
594 if (quota->charge_addr_from &&
595 r->ar.end <= quota->charge_addr_from)
596 continue;
597
598 if (quota->charge_addr_from && r->ar.start <
599 quota->charge_addr_from) {
600 sz = ALIGN_DOWN(quota->charge_addr_from -
601 r->ar.start, DAMON_MIN_REGION);
602 if (!sz) {
603 if (r->ar.end - r->ar.start <=
604 DAMON_MIN_REGION)
605 continue;
606 sz = DAMON_MIN_REGION;
607 }
608 damon_split_region_at(c, t, r, sz);
609 r = damon_next_region(r);
610 sz = r->ar.end - r->ar.start;
611 }
612 quota->charge_target_from = NULL;
613 quota->charge_addr_from = 0;
614 }
615
616 if (!damos_valid_target(c, t, r, s))
617 continue;
618
619 /* Apply the scheme */
620 if (c->primitive.apply_scheme) {
621 if (quota->esz &&
622 quota->charged_sz + sz > quota->esz) {
623 sz = ALIGN_DOWN(quota->esz - quota->charged_sz,
624 DAMON_MIN_REGION);
625 if (!sz)
626 goto update_stat;
627 damon_split_region_at(c, t, r, sz);
628 }
629 ktime_get_coarse_ts64(&begin);
630 sz_applied = c->primitive.apply_scheme(c, t, r, s);
631 ktime_get_coarse_ts64(&end);
632 quota->total_charged_ns += timespec64_to_ns(&end) -
633 timespec64_to_ns(&begin);
634 quota->charged_sz += sz;
635 if (quota->esz && quota->charged_sz >= quota->esz) {
636 quota->charge_target_from = t;
637 quota->charge_addr_from = r->ar.end + 1;
638 }
639 }
640 if (s->action != DAMOS_STAT)
641 r->age = 0;
642
643 update_stat:
644 s->stat.nr_tried++;
645 s->stat.sz_tried += sz;
646 if (sz_applied)
647 s->stat.nr_applied++;
648 s->stat.sz_applied += sz_applied;
649 }
650 }
651
652 /* Shouldn't be called if quota->ms and quota->sz are zero */
damos_set_effective_quota(struct damos_quota * quota)653 static void damos_set_effective_quota(struct damos_quota *quota)
654 {
655 unsigned long throughput;
656 unsigned long esz;
657
658 if (!quota->ms) {
659 quota->esz = quota->sz;
660 return;
661 }
662
663 if (quota->total_charged_ns)
664 throughput = quota->total_charged_sz * 1000000 /
665 quota->total_charged_ns;
666 else
667 throughput = PAGE_SIZE * 1024;
668 esz = throughput * quota->ms;
669
670 if (quota->sz && quota->sz < esz)
671 esz = quota->sz;
672 quota->esz = esz;
673 }
674
kdamond_apply_schemes(struct damon_ctx * c)675 static void kdamond_apply_schemes(struct damon_ctx *c)
676 {
677 struct damon_target *t;
678 struct damon_region *r, *next_r;
679 struct damos *s;
680
681 damon_for_each_scheme(s, c) {
682 struct damos_quota *quota = &s->quota;
683 unsigned long cumulated_sz;
684 unsigned int score, max_score = 0;
685
686 if (!s->wmarks.activated)
687 continue;
688
689 if (!quota->ms && !quota->sz)
690 continue;
691
692 /* New charge window starts */
693 if (time_after_eq(jiffies, quota->charged_from +
694 msecs_to_jiffies(
695 quota->reset_interval))) {
696 if (quota->esz && quota->charged_sz >= quota->esz)
697 s->stat.qt_exceeds++;
698 quota->total_charged_sz += quota->charged_sz;
699 quota->charged_from = jiffies;
700 quota->charged_sz = 0;
701 damos_set_effective_quota(quota);
702 }
703
704 if (!c->primitive.get_scheme_score)
705 continue;
706
707 /* Fill up the score histogram */
708 memset(quota->histogram, 0, sizeof(quota->histogram));
709 damon_for_each_target(t, c) {
710 damon_for_each_region(r, t) {
711 if (!__damos_valid_target(r, s))
712 continue;
713 score = c->primitive.get_scheme_score(
714 c, t, r, s);
715 quota->histogram[score] +=
716 r->ar.end - r->ar.start;
717 if (score > max_score)
718 max_score = score;
719 }
720 }
721
722 /* Set the min score limit */
723 for (cumulated_sz = 0, score = max_score; ; score--) {
724 cumulated_sz += quota->histogram[score];
725 if (cumulated_sz >= quota->esz || !score)
726 break;
727 }
728 quota->min_score = score;
729 }
730
731 damon_for_each_target(t, c) {
732 damon_for_each_region_safe(r, next_r, t)
733 damon_do_apply_schemes(c, t, r);
734 }
735 }
736
sz_damon_region(struct damon_region * r)737 static inline unsigned long sz_damon_region(struct damon_region *r)
738 {
739 return r->ar.end - r->ar.start;
740 }
741
742 /*
743 * Merge two adjacent regions into one region
744 */
damon_merge_two_regions(struct damon_target * t,struct damon_region * l,struct damon_region * r)745 static void damon_merge_two_regions(struct damon_target *t,
746 struct damon_region *l, struct damon_region *r)
747 {
748 unsigned long sz_l = sz_damon_region(l), sz_r = sz_damon_region(r);
749
750 l->nr_accesses = (l->nr_accesses * sz_l + r->nr_accesses * sz_r) /
751 (sz_l + sz_r);
752 l->age = (l->age * sz_l + r->age * sz_r) / (sz_l + sz_r);
753 l->ar.end = r->ar.end;
754 damon_destroy_region(r, t);
755 }
756
757 /*
758 * Merge adjacent regions having similar access frequencies
759 *
760 * t target affected by this merge operation
761 * thres '->nr_accesses' diff threshold for the merge
762 * sz_limit size upper limit of each region
763 */
damon_merge_regions_of(struct damon_target * t,unsigned int thres,unsigned long sz_limit)764 static void damon_merge_regions_of(struct damon_target *t, unsigned int thres,
765 unsigned long sz_limit)
766 {
767 struct damon_region *r, *prev = NULL, *next;
768
769 damon_for_each_region_safe(r, next, t) {
770 if (abs(r->nr_accesses - r->last_nr_accesses) > thres)
771 r->age = 0;
772 else
773 r->age++;
774
775 if (prev && prev->ar.end == r->ar.start &&
776 abs(prev->nr_accesses - r->nr_accesses) <= thres &&
777 sz_damon_region(prev) + sz_damon_region(r) <= sz_limit)
778 damon_merge_two_regions(t, prev, r);
779 else
780 prev = r;
781 }
782 }
783
784 /*
785 * Merge adjacent regions having similar access frequencies
786 *
787 * threshold '->nr_accesses' diff threshold for the merge
788 * sz_limit size upper limit of each region
789 *
790 * This function merges monitoring target regions which are adjacent and their
791 * access frequencies are similar. This is for minimizing the monitoring
792 * overhead under the dynamically changeable access pattern. If a merge was
793 * unnecessarily made, later 'kdamond_split_regions()' will revert it.
794 */
kdamond_merge_regions(struct damon_ctx * c,unsigned int threshold,unsigned long sz_limit)795 static void kdamond_merge_regions(struct damon_ctx *c, unsigned int threshold,
796 unsigned long sz_limit)
797 {
798 struct damon_target *t;
799
800 damon_for_each_target(t, c)
801 damon_merge_regions_of(t, threshold, sz_limit);
802 }
803
804 /*
805 * Split a region in two
806 *
807 * r the region to be split
808 * sz_r size of the first sub-region that will be made
809 */
damon_split_region_at(struct damon_ctx * ctx,struct damon_target * t,struct damon_region * r,unsigned long sz_r)810 static void damon_split_region_at(struct damon_ctx *ctx,
811 struct damon_target *t, struct damon_region *r,
812 unsigned long sz_r)
813 {
814 struct damon_region *new;
815
816 new = damon_new_region(r->ar.start + sz_r, r->ar.end);
817 if (!new)
818 return;
819
820 r->ar.end = new->ar.start;
821
822 new->age = r->age;
823 new->last_nr_accesses = r->last_nr_accesses;
824
825 damon_insert_region(new, r, damon_next_region(r), t);
826 }
827
828 /* Split every region in the given target into 'nr_subs' regions */
damon_split_regions_of(struct damon_ctx * ctx,struct damon_target * t,int nr_subs)829 static void damon_split_regions_of(struct damon_ctx *ctx,
830 struct damon_target *t, int nr_subs)
831 {
832 struct damon_region *r, *next;
833 unsigned long sz_region, sz_sub = 0;
834 int i;
835
836 damon_for_each_region_safe(r, next, t) {
837 sz_region = r->ar.end - r->ar.start;
838
839 for (i = 0; i < nr_subs - 1 &&
840 sz_region > 2 * DAMON_MIN_REGION; i++) {
841 /*
842 * Randomly select size of left sub-region to be at
843 * least 10 percent and at most 90% of original region
844 */
845 sz_sub = ALIGN_DOWN(damon_rand(1, 10) *
846 sz_region / 10, DAMON_MIN_REGION);
847 /* Do not allow blank region */
848 if (sz_sub == 0 || sz_sub >= sz_region)
849 continue;
850
851 damon_split_region_at(ctx, t, r, sz_sub);
852 sz_region = sz_sub;
853 }
854 }
855 }
856
857 /*
858 * Split every target region into randomly-sized small regions
859 *
860 * This function splits every target region into random-sized small regions if
861 * current total number of the regions is equal or smaller than half of the
862 * user-specified maximum number of regions. This is for maximizing the
863 * monitoring accuracy under the dynamically changeable access patterns. If a
864 * split was unnecessarily made, later 'kdamond_merge_regions()' will revert
865 * it.
866 */
kdamond_split_regions(struct damon_ctx * ctx)867 static void kdamond_split_regions(struct damon_ctx *ctx)
868 {
869 struct damon_target *t;
870 unsigned int nr_regions = 0;
871 static unsigned int last_nr_regions;
872 int nr_subregions = 2;
873
874 damon_for_each_target(t, ctx)
875 nr_regions += damon_nr_regions(t);
876
877 if (nr_regions > ctx->max_nr_regions / 2)
878 return;
879
880 /* Maybe the middle of the region has different access frequency */
881 if (last_nr_regions == nr_regions &&
882 nr_regions < ctx->max_nr_regions / 3)
883 nr_subregions = 3;
884
885 damon_for_each_target(t, ctx)
886 damon_split_regions_of(ctx, t, nr_subregions);
887
888 last_nr_regions = nr_regions;
889 }
890
891 /*
892 * Check whether it is time to check and apply the target monitoring regions
893 *
894 * Returns true if it is.
895 */
kdamond_need_update_primitive(struct damon_ctx * ctx)896 static bool kdamond_need_update_primitive(struct damon_ctx *ctx)
897 {
898 return damon_check_reset_time_interval(&ctx->last_primitive_update,
899 ctx->primitive_update_interval);
900 }
901
902 /*
903 * Check whether current monitoring should be stopped
904 *
905 * The monitoring is stopped when either the user requested to stop, or all
906 * monitoring targets are invalid.
907 *
908 * Returns true if need to stop current monitoring.
909 */
kdamond_need_stop(struct damon_ctx * ctx)910 static bool kdamond_need_stop(struct damon_ctx *ctx)
911 {
912 struct damon_target *t;
913
914 if (kthread_should_stop())
915 return true;
916
917 if (!ctx->primitive.target_valid)
918 return false;
919
920 damon_for_each_target(t, ctx) {
921 if (ctx->primitive.target_valid(t))
922 return false;
923 }
924
925 return true;
926 }
927
damos_wmark_metric_value(enum damos_wmark_metric metric)928 static unsigned long damos_wmark_metric_value(enum damos_wmark_metric metric)
929 {
930 struct sysinfo i;
931
932 switch (metric) {
933 case DAMOS_WMARK_FREE_MEM_RATE:
934 si_meminfo(&i);
935 return i.freeram * 1000 / i.totalram;
936 default:
937 break;
938 }
939 return -EINVAL;
940 }
941
942 /*
943 * Returns zero if the scheme is active. Else, returns time to wait for next
944 * watermark check in micro-seconds.
945 */
damos_wmark_wait_us(struct damos * scheme)946 static unsigned long damos_wmark_wait_us(struct damos *scheme)
947 {
948 unsigned long metric;
949
950 if (scheme->wmarks.metric == DAMOS_WMARK_NONE)
951 return 0;
952
953 metric = damos_wmark_metric_value(scheme->wmarks.metric);
954 /* higher than high watermark or lower than low watermark */
955 if (metric > scheme->wmarks.high || scheme->wmarks.low > metric) {
956 if (scheme->wmarks.activated)
957 pr_debug("deactivate a scheme (%d) for %s wmark\n",
958 scheme->action,
959 metric > scheme->wmarks.high ?
960 "high" : "low");
961 scheme->wmarks.activated = false;
962 return scheme->wmarks.interval;
963 }
964
965 /* inactive and higher than middle watermark */
966 if ((scheme->wmarks.high >= metric && metric >= scheme->wmarks.mid) &&
967 !scheme->wmarks.activated)
968 return scheme->wmarks.interval;
969
970 if (!scheme->wmarks.activated)
971 pr_debug("activate a scheme (%d)\n", scheme->action);
972 scheme->wmarks.activated = true;
973 return 0;
974 }
975
976 /* Returns negative error code if it's not activated but should return */
kdamond_wait_activation(struct damon_ctx * ctx)977 static int kdamond_wait_activation(struct damon_ctx *ctx)
978 {
979 struct damos *s;
980 unsigned long wait_time;
981 unsigned long min_wait_time = 0;
982
983 while (!kdamond_need_stop(ctx)) {
984 damon_for_each_scheme(s, ctx) {
985 wait_time = damos_wmark_wait_us(s);
986 if (!min_wait_time || wait_time < min_wait_time)
987 min_wait_time = wait_time;
988 }
989 if (!min_wait_time)
990 return 0;
991
992 kdamond_usleep(min_wait_time);
993 }
994 return -EBUSY;
995 }
996
997 /*
998 * The monitoring daemon that runs as a kernel thread
999 */
kdamond_fn(void * data)1000 static int kdamond_fn(void *data)
1001 {
1002 struct damon_ctx *ctx = (struct damon_ctx *)data;
1003 struct damon_target *t;
1004 struct damon_region *r, *next;
1005 unsigned int max_nr_accesses = 0;
1006 unsigned long sz_limit = 0;
1007 bool done = false;
1008
1009 pr_debug("kdamond (%d) starts\n", current->pid);
1010
1011 if (ctx->primitive.init)
1012 ctx->primitive.init(ctx);
1013 if (ctx->callback.before_start && ctx->callback.before_start(ctx))
1014 done = true;
1015
1016 sz_limit = damon_region_sz_limit(ctx);
1017
1018 while (!kdamond_need_stop(ctx) && !done) {
1019 if (kdamond_wait_activation(ctx))
1020 continue;
1021
1022 if (ctx->primitive.prepare_access_checks)
1023 ctx->primitive.prepare_access_checks(ctx);
1024 if (ctx->callback.after_sampling &&
1025 ctx->callback.after_sampling(ctx))
1026 done = true;
1027
1028 kdamond_usleep(ctx->sample_interval);
1029
1030 if (ctx->primitive.check_accesses)
1031 max_nr_accesses = ctx->primitive.check_accesses(ctx);
1032
1033 if (kdamond_aggregate_interval_passed(ctx)) {
1034 kdamond_merge_regions(ctx,
1035 max_nr_accesses / 10,
1036 sz_limit);
1037 if (ctx->callback.after_aggregation &&
1038 ctx->callback.after_aggregation(ctx))
1039 done = true;
1040 kdamond_apply_schemes(ctx);
1041 kdamond_reset_aggregated(ctx);
1042 kdamond_split_regions(ctx);
1043 if (ctx->primitive.reset_aggregated)
1044 ctx->primitive.reset_aggregated(ctx);
1045 }
1046
1047 if (kdamond_need_update_primitive(ctx)) {
1048 if (ctx->primitive.update)
1049 ctx->primitive.update(ctx);
1050 sz_limit = damon_region_sz_limit(ctx);
1051 }
1052 }
1053 damon_for_each_target(t, ctx) {
1054 damon_for_each_region_safe(r, next, t)
1055 damon_destroy_region(r, t);
1056 }
1057
1058 if (ctx->callback.before_terminate)
1059 ctx->callback.before_terminate(ctx);
1060 if (ctx->primitive.cleanup)
1061 ctx->primitive.cleanup(ctx);
1062
1063 pr_debug("kdamond (%d) finishes\n", current->pid);
1064 mutex_lock(&ctx->kdamond_lock);
1065 ctx->kdamond = NULL;
1066 mutex_unlock(&ctx->kdamond_lock);
1067
1068 mutex_lock(&damon_lock);
1069 nr_running_ctxs--;
1070 mutex_unlock(&damon_lock);
1071
1072 return 0;
1073 }
1074
1075 #include "core-test.h"
1076