Loading...
1// SPDX-License-Identifier: GPL-2.0
2/*
3 * Primary bucket allocation code
4 *
5 * Copyright 2012 Google, Inc.
6 *
7 * Allocation in bcache is done in terms of buckets:
8 *
9 * Each bucket has associated an 8 bit gen; this gen corresponds to the gen in
10 * btree pointers - they must match for the pointer to be considered valid.
11 *
12 * Thus (assuming a bucket has no dirty data or metadata in it) we can reuse a
13 * bucket simply by incrementing its gen.
14 *
15 * The gens (along with the priorities; it's really the gens are important but
16 * the code is named as if it's the priorities) are written in an arbitrary list
17 * of buckets on disk, with a pointer to them in the journal header.
18 *
19 * When we invalidate a bucket, we have to write its new gen to disk and wait
20 * for that write to complete before we use it - otherwise after a crash we
21 * could have pointers that appeared to be good but pointed to data that had
22 * been overwritten.
23 *
24 * Since the gens and priorities are all stored contiguously on disk, we can
25 * batch this up: We fill up the free_inc list with freshly invalidated buckets,
26 * call prio_write(), and when prio_write() finishes we pull buckets off the
27 * free_inc list and optionally discard them.
28 *
29 * free_inc isn't the only freelist - if it was, we'd often to sleep while
30 * priorities and gens were being written before we could allocate. c->free is a
31 * smaller freelist, and buckets on that list are always ready to be used.
32 *
33 * If we've got discards enabled, that happens when a bucket moves from the
34 * free_inc list to the free list.
35 *
36 * There is another freelist, because sometimes we have buckets that we know
37 * have nothing pointing into them - these we can reuse without waiting for
38 * priorities to be rewritten. These come from freed btree nodes and buckets
39 * that garbage collection discovered no longer had valid keys pointing into
40 * them (because they were overwritten). That's the unused list - buckets on the
41 * unused list move to the free list, optionally being discarded in the process.
42 *
43 * It's also important to ensure that gens don't wrap around - with respect to
44 * either the oldest gen in the btree or the gen on disk. This is quite
45 * difficult to do in practice, but we explicitly guard against it anyways - if
46 * a bucket is in danger of wrapping around we simply skip invalidating it that
47 * time around, and we garbage collect or rewrite the priorities sooner than we
48 * would have otherwise.
49 *
50 * bch_bucket_alloc() allocates a single bucket from a specific cache.
51 *
52 * bch_bucket_alloc_set() allocates one bucket from different caches
53 * out of a cache set.
54 *
55 * free_some_buckets() drives all the processes described above. It's called
56 * from bch_bucket_alloc() and a few other places that need to make sure free
57 * buckets are ready.
58 *
59 * invalidate_buckets_(lru|fifo)() find buckets that are available to be
60 * invalidated, and then invalidate them and stick them on the free_inc list -
61 * in either lru or fifo order.
62 */
63
64#include "bcache.h"
65#include "btree.h"
66
67#include <linux/blkdev.h>
68#include <linux/kthread.h>
69#include <linux/random.h>
70#include <trace/events/bcache.h>
71
72#define MAX_OPEN_BUCKETS 128
73
74/* Bucket heap / gen */
75
76uint8_t bch_inc_gen(struct cache *ca, struct bucket *b)
77{
78 uint8_t ret = ++b->gen;
79
80 ca->set->need_gc = max(ca->set->need_gc, bucket_gc_gen(b));
81 WARN_ON_ONCE(ca->set->need_gc > BUCKET_GC_GEN_MAX);
82
83 return ret;
84}
85
86void bch_rescale_priorities(struct cache_set *c, int sectors)
87{
88 struct cache *ca;
89 struct bucket *b;
90 unsigned long next = c->nbuckets * c->cache->sb.bucket_size / 1024;
91 int r;
92
93 atomic_sub(sectors, &c->rescale);
94
95 do {
96 r = atomic_read(&c->rescale);
97
98 if (r >= 0)
99 return;
100 } while (atomic_cmpxchg(&c->rescale, r, r + next) != r);
101
102 mutex_lock(&c->bucket_lock);
103
104 c->min_prio = USHRT_MAX;
105
106 ca = c->cache;
107 for_each_bucket(b, ca)
108 if (b->prio &&
109 b->prio != BTREE_PRIO &&
110 !atomic_read(&b->pin)) {
111 b->prio--;
112 c->min_prio = min(c->min_prio, b->prio);
113 }
114
115 mutex_unlock(&c->bucket_lock);
116}
117
118/*
119 * Background allocation thread: scans for buckets to be invalidated,
120 * invalidates them, rewrites prios/gens (marking them as invalidated on disk),
121 * then optionally issues discard commands to the newly free buckets, then puts
122 * them on the various freelists.
123 */
124
125static inline bool can_inc_bucket_gen(struct bucket *b)
126{
127 return bucket_gc_gen(b) < BUCKET_GC_GEN_MAX;
128}
129
130bool bch_can_invalidate_bucket(struct cache *ca, struct bucket *b)
131{
132 BUG_ON(!ca->set->gc_mark_valid);
133
134 return (!GC_MARK(b) ||
135 GC_MARK(b) == GC_MARK_RECLAIMABLE) &&
136 !atomic_read(&b->pin) &&
137 can_inc_bucket_gen(b);
138}
139
140void __bch_invalidate_one_bucket(struct cache *ca, struct bucket *b)
141{
142 lockdep_assert_held(&ca->set->bucket_lock);
143 BUG_ON(GC_MARK(b) && GC_MARK(b) != GC_MARK_RECLAIMABLE);
144
145 if (GC_SECTORS_USED(b))
146 trace_bcache_invalidate(ca, b - ca->buckets);
147
148 bch_inc_gen(ca, b);
149 b->prio = INITIAL_PRIO;
150 atomic_inc(&b->pin);
151}
152
153static void bch_invalidate_one_bucket(struct cache *ca, struct bucket *b)
154{
155 __bch_invalidate_one_bucket(ca, b);
156
157 fifo_push(&ca->free_inc, b - ca->buckets);
158}
159
160/*
161 * Determines what order we're going to reuse buckets, smallest bucket_prio()
162 * first: we also take into account the number of sectors of live data in that
163 * bucket, and in order for that multiply to make sense we have to scale bucket
164 *
165 * Thus, we scale the bucket priorities so that the bucket with the smallest
166 * prio is worth 1/8th of what INITIAL_PRIO is worth.
167 */
168
169#define bucket_prio(b) \
170({ \
171 unsigned int min_prio = (INITIAL_PRIO - ca->set->min_prio) / 8; \
172 \
173 (b->prio - ca->set->min_prio + min_prio) * GC_SECTORS_USED(b); \
174})
175
176#define bucket_max_cmp(l, r) (bucket_prio(l) < bucket_prio(r))
177#define bucket_min_cmp(l, r) (bucket_prio(l) > bucket_prio(r))
178
179static void invalidate_buckets_lru(struct cache *ca)
180{
181 struct bucket *b;
182 ssize_t i;
183
184 ca->heap.used = 0;
185
186 for_each_bucket(b, ca) {
187 if (!bch_can_invalidate_bucket(ca, b))
188 continue;
189
190 if (!heap_full(&ca->heap))
191 heap_add(&ca->heap, b, bucket_max_cmp);
192 else if (bucket_max_cmp(b, heap_peek(&ca->heap))) {
193 ca->heap.data[0] = b;
194 heap_sift(&ca->heap, 0, bucket_max_cmp);
195 }
196 }
197
198 for (i = ca->heap.used / 2 - 1; i >= 0; --i)
199 heap_sift(&ca->heap, i, bucket_min_cmp);
200
201 while (!fifo_full(&ca->free_inc)) {
202 if (!heap_pop(&ca->heap, b, bucket_min_cmp)) {
203 /*
204 * We don't want to be calling invalidate_buckets()
205 * multiple times when it can't do anything
206 */
207 ca->invalidate_needs_gc = 1;
208 wake_up_gc(ca->set);
209 return;
210 }
211
212 bch_invalidate_one_bucket(ca, b);
213 }
214}
215
216static void invalidate_buckets_fifo(struct cache *ca)
217{
218 struct bucket *b;
219 size_t checked = 0;
220
221 while (!fifo_full(&ca->free_inc)) {
222 if (ca->fifo_last_bucket < ca->sb.first_bucket ||
223 ca->fifo_last_bucket >= ca->sb.nbuckets)
224 ca->fifo_last_bucket = ca->sb.first_bucket;
225
226 b = ca->buckets + ca->fifo_last_bucket++;
227
228 if (bch_can_invalidate_bucket(ca, b))
229 bch_invalidate_one_bucket(ca, b);
230
231 if (++checked >= ca->sb.nbuckets) {
232 ca->invalidate_needs_gc = 1;
233 wake_up_gc(ca->set);
234 return;
235 }
236 }
237}
238
239static void invalidate_buckets_random(struct cache *ca)
240{
241 struct bucket *b;
242 size_t checked = 0;
243
244 while (!fifo_full(&ca->free_inc)) {
245 size_t n;
246
247 get_random_bytes(&n, sizeof(n));
248
249 n %= (size_t) (ca->sb.nbuckets - ca->sb.first_bucket);
250 n += ca->sb.first_bucket;
251
252 b = ca->buckets + n;
253
254 if (bch_can_invalidate_bucket(ca, b))
255 bch_invalidate_one_bucket(ca, b);
256
257 if (++checked >= ca->sb.nbuckets / 2) {
258 ca->invalidate_needs_gc = 1;
259 wake_up_gc(ca->set);
260 return;
261 }
262 }
263}
264
265static void invalidate_buckets(struct cache *ca)
266{
267 BUG_ON(ca->invalidate_needs_gc);
268
269 switch (CACHE_REPLACEMENT(&ca->sb)) {
270 case CACHE_REPLACEMENT_LRU:
271 invalidate_buckets_lru(ca);
272 break;
273 case CACHE_REPLACEMENT_FIFO:
274 invalidate_buckets_fifo(ca);
275 break;
276 case CACHE_REPLACEMENT_RANDOM:
277 invalidate_buckets_random(ca);
278 break;
279 }
280}
281
282#define allocator_wait(ca, cond) \
283do { \
284 while (1) { \
285 set_current_state(TASK_INTERRUPTIBLE); \
286 if (cond) \
287 break; \
288 \
289 mutex_unlock(&(ca)->set->bucket_lock); \
290 if (kthread_should_stop() || \
291 test_bit(CACHE_SET_IO_DISABLE, &ca->set->flags)) { \
292 set_current_state(TASK_RUNNING); \
293 goto out; \
294 } \
295 \
296 schedule(); \
297 mutex_lock(&(ca)->set->bucket_lock); \
298 } \
299 __set_current_state(TASK_RUNNING); \
300} while (0)
301
302static int bch_allocator_push(struct cache *ca, long bucket)
303{
304 unsigned int i;
305
306 /* Prios/gens are actually the most important reserve */
307 if (fifo_push(&ca->free[RESERVE_PRIO], bucket))
308 return true;
309
310 for (i = 0; i < RESERVE_NR; i++)
311 if (fifo_push(&ca->free[i], bucket))
312 return true;
313
314 return false;
315}
316
317static int bch_allocator_thread(void *arg)
318{
319 struct cache *ca = arg;
320
321 mutex_lock(&ca->set->bucket_lock);
322
323 while (1) {
324 /*
325 * First, we pull buckets off of the unused and free_inc lists,
326 * possibly issue discards to them, then we add the bucket to
327 * the free list:
328 */
329 while (1) {
330 long bucket;
331
332 if (!fifo_pop(&ca->free_inc, bucket))
333 break;
334
335 if (ca->discard) {
336 mutex_unlock(&ca->set->bucket_lock);
337 blkdev_issue_discard(ca->bdev,
338 bucket_to_sector(ca->set, bucket),
339 ca->sb.bucket_size, GFP_KERNEL);
340 mutex_lock(&ca->set->bucket_lock);
341 }
342
343 allocator_wait(ca, bch_allocator_push(ca, bucket));
344 wake_up(&ca->set->btree_cache_wait);
345 wake_up(&ca->set->bucket_wait);
346 }
347
348 /*
349 * We've run out of free buckets, we need to find some buckets
350 * we can invalidate. First, invalidate them in memory and add
351 * them to the free_inc list:
352 */
353
354retry_invalidate:
355 allocator_wait(ca, ca->set->gc_mark_valid &&
356 !ca->invalidate_needs_gc);
357 invalidate_buckets(ca);
358
359 /*
360 * Now, we write their new gens to disk so we can start writing
361 * new stuff to them:
362 */
363 allocator_wait(ca, !atomic_read(&ca->set->prio_blocked));
364 if (CACHE_SYNC(&ca->sb)) {
365 /*
366 * This could deadlock if an allocation with a btree
367 * node locked ever blocked - having the btree node
368 * locked would block garbage collection, but here we're
369 * waiting on garbage collection before we invalidate
370 * and free anything.
371 *
372 * But this should be safe since the btree code always
373 * uses btree_check_reserve() before allocating now, and
374 * if it fails it blocks without btree nodes locked.
375 */
376 if (!fifo_full(&ca->free_inc))
377 goto retry_invalidate;
378
379 if (bch_prio_write(ca, false) < 0) {
380 ca->invalidate_needs_gc = 1;
381 wake_up_gc(ca->set);
382 }
383 }
384 }
385out:
386 wait_for_kthread_stop();
387 return 0;
388}
389
390/* Allocation */
391
392long bch_bucket_alloc(struct cache *ca, unsigned int reserve, bool wait)
393{
394 DEFINE_WAIT(w);
395 struct bucket *b;
396 long r;
397
398
399 /* No allocation if CACHE_SET_IO_DISABLE bit is set */
400 if (unlikely(test_bit(CACHE_SET_IO_DISABLE, &ca->set->flags)))
401 return -1;
402
403 /* fastpath */
404 if (fifo_pop(&ca->free[RESERVE_NONE], r) ||
405 fifo_pop(&ca->free[reserve], r))
406 goto out;
407
408 if (!wait) {
409 trace_bcache_alloc_fail(ca, reserve);
410 return -1;
411 }
412
413 do {
414 prepare_to_wait(&ca->set->bucket_wait, &w,
415 TASK_UNINTERRUPTIBLE);
416
417 mutex_unlock(&ca->set->bucket_lock);
418 schedule();
419 mutex_lock(&ca->set->bucket_lock);
420 } while (!fifo_pop(&ca->free[RESERVE_NONE], r) &&
421 !fifo_pop(&ca->free[reserve], r));
422
423 finish_wait(&ca->set->bucket_wait, &w);
424out:
425 if (ca->alloc_thread)
426 wake_up_process(ca->alloc_thread);
427
428 trace_bcache_alloc(ca, reserve);
429
430 if (expensive_debug_checks(ca->set)) {
431 size_t iter;
432 long i;
433 unsigned int j;
434
435 for (iter = 0; iter < prio_buckets(ca) * 2; iter++)
436 BUG_ON(ca->prio_buckets[iter] == (uint64_t) r);
437
438 for (j = 0; j < RESERVE_NR; j++)
439 fifo_for_each(i, &ca->free[j], iter)
440 BUG_ON(i == r);
441 fifo_for_each(i, &ca->free_inc, iter)
442 BUG_ON(i == r);
443 }
444
445 b = ca->buckets + r;
446
447 BUG_ON(atomic_read(&b->pin) != 1);
448
449 SET_GC_SECTORS_USED(b, ca->sb.bucket_size);
450
451 if (reserve <= RESERVE_PRIO) {
452 SET_GC_MARK(b, GC_MARK_METADATA);
453 SET_GC_MOVE(b, 0);
454 b->prio = BTREE_PRIO;
455 } else {
456 SET_GC_MARK(b, GC_MARK_RECLAIMABLE);
457 SET_GC_MOVE(b, 0);
458 b->prio = INITIAL_PRIO;
459 }
460
461 if (ca->set->avail_nbuckets > 0) {
462 ca->set->avail_nbuckets--;
463 bch_update_bucket_in_use(ca->set, &ca->set->gc_stats);
464 }
465
466 return r;
467}
468
469void __bch_bucket_free(struct cache *ca, struct bucket *b)
470{
471 SET_GC_MARK(b, 0);
472 SET_GC_SECTORS_USED(b, 0);
473
474 if (ca->set->avail_nbuckets < ca->set->nbuckets) {
475 ca->set->avail_nbuckets++;
476 bch_update_bucket_in_use(ca->set, &ca->set->gc_stats);
477 }
478}
479
480void bch_bucket_free(struct cache_set *c, struct bkey *k)
481{
482 unsigned int i;
483
484 for (i = 0; i < KEY_PTRS(k); i++)
485 __bch_bucket_free(c->cache, PTR_BUCKET(c, k, i));
486}
487
488int __bch_bucket_alloc_set(struct cache_set *c, unsigned int reserve,
489 struct bkey *k, bool wait)
490{
491 struct cache *ca;
492 long b;
493
494 /* No allocation if CACHE_SET_IO_DISABLE bit is set */
495 if (unlikely(test_bit(CACHE_SET_IO_DISABLE, &c->flags)))
496 return -1;
497
498 lockdep_assert_held(&c->bucket_lock);
499
500 bkey_init(k);
501
502 ca = c->cache;
503 b = bch_bucket_alloc(ca, reserve, wait);
504 if (b == -1)
505 goto err;
506
507 k->ptr[0] = MAKE_PTR(ca->buckets[b].gen,
508 bucket_to_sector(c, b),
509 ca->sb.nr_this_dev);
510
511 SET_KEY_PTRS(k, 1);
512
513 return 0;
514err:
515 bch_bucket_free(c, k);
516 bkey_put(c, k);
517 return -1;
518}
519
520int bch_bucket_alloc_set(struct cache_set *c, unsigned int reserve,
521 struct bkey *k, bool wait)
522{
523 int ret;
524
525 mutex_lock(&c->bucket_lock);
526 ret = __bch_bucket_alloc_set(c, reserve, k, wait);
527 mutex_unlock(&c->bucket_lock);
528 return ret;
529}
530
531/* Sector allocator */
532
533struct open_bucket {
534 struct list_head list;
535 unsigned int last_write_point;
536 unsigned int sectors_free;
537 BKEY_PADDED(key);
538};
539
540/*
541 * We keep multiple buckets open for writes, and try to segregate different
542 * write streams for better cache utilization: first we try to segregate flash
543 * only volume write streams from cached devices, secondly we look for a bucket
544 * where the last write to it was sequential with the current write, and
545 * failing that we look for a bucket that was last used by the same task.
546 *
547 * The ideas is if you've got multiple tasks pulling data into the cache at the
548 * same time, you'll get better cache utilization if you try to segregate their
549 * data and preserve locality.
550 *
551 * For example, dirty sectors of flash only volume is not reclaimable, if their
552 * dirty sectors mixed with dirty sectors of cached device, such buckets will
553 * be marked as dirty and won't be reclaimed, though the dirty data of cached
554 * device have been written back to backend device.
555 *
556 * And say you've starting Firefox at the same time you're copying a
557 * bunch of files. Firefox will likely end up being fairly hot and stay in the
558 * cache awhile, but the data you copied might not be; if you wrote all that
559 * data to the same buckets it'd get invalidated at the same time.
560 *
561 * Both of those tasks will be doing fairly random IO so we can't rely on
562 * detecting sequential IO to segregate their data, but going off of the task
563 * should be a sane heuristic.
564 */
565static struct open_bucket *pick_data_bucket(struct cache_set *c,
566 const struct bkey *search,
567 unsigned int write_point,
568 struct bkey *alloc)
569{
570 struct open_bucket *ret, *ret_task = NULL;
571
572 list_for_each_entry_reverse(ret, &c->data_buckets, list)
573 if (UUID_FLASH_ONLY(&c->uuids[KEY_INODE(&ret->key)]) !=
574 UUID_FLASH_ONLY(&c->uuids[KEY_INODE(search)]))
575 continue;
576 else if (!bkey_cmp(&ret->key, search))
577 goto found;
578 else if (ret->last_write_point == write_point)
579 ret_task = ret;
580
581 ret = ret_task ?: list_first_entry(&c->data_buckets,
582 struct open_bucket, list);
583found:
584 if (!ret->sectors_free && KEY_PTRS(alloc)) {
585 ret->sectors_free = c->cache->sb.bucket_size;
586 bkey_copy(&ret->key, alloc);
587 bkey_init(alloc);
588 }
589
590 if (!ret->sectors_free)
591 ret = NULL;
592
593 return ret;
594}
595
596/*
597 * Allocates some space in the cache to write to, and k to point to the newly
598 * allocated space, and updates KEY_SIZE(k) and KEY_OFFSET(k) (to point to the
599 * end of the newly allocated space).
600 *
601 * May allocate fewer sectors than @sectors, KEY_SIZE(k) indicates how many
602 * sectors were actually allocated.
603 *
604 * If s->writeback is true, will not fail.
605 */
606bool bch_alloc_sectors(struct cache_set *c,
607 struct bkey *k,
608 unsigned int sectors,
609 unsigned int write_point,
610 unsigned int write_prio,
611 bool wait)
612{
613 struct open_bucket *b;
614 BKEY_PADDED(key) alloc;
615 unsigned int i;
616
617 /*
618 * We might have to allocate a new bucket, which we can't do with a
619 * spinlock held. So if we have to allocate, we drop the lock, allocate
620 * and then retry. KEY_PTRS() indicates whether alloc points to
621 * allocated bucket(s).
622 */
623
624 bkey_init(&alloc.key);
625 spin_lock(&c->data_bucket_lock);
626
627 while (!(b = pick_data_bucket(c, k, write_point, &alloc.key))) {
628 unsigned int watermark = write_prio
629 ? RESERVE_MOVINGGC
630 : RESERVE_NONE;
631
632 spin_unlock(&c->data_bucket_lock);
633
634 if (bch_bucket_alloc_set(c, watermark, &alloc.key, wait))
635 return false;
636
637 spin_lock(&c->data_bucket_lock);
638 }
639
640 /*
641 * If we had to allocate, we might race and not need to allocate the
642 * second time we call pick_data_bucket(). If we allocated a bucket but
643 * didn't use it, drop the refcount bch_bucket_alloc_set() took:
644 */
645 if (KEY_PTRS(&alloc.key))
646 bkey_put(c, &alloc.key);
647
648 for (i = 0; i < KEY_PTRS(&b->key); i++)
649 EBUG_ON(ptr_stale(c, &b->key, i));
650
651 /* Set up the pointer to the space we're allocating: */
652
653 for (i = 0; i < KEY_PTRS(&b->key); i++)
654 k->ptr[i] = b->key.ptr[i];
655
656 sectors = min(sectors, b->sectors_free);
657
658 SET_KEY_OFFSET(k, KEY_OFFSET(k) + sectors);
659 SET_KEY_SIZE(k, sectors);
660 SET_KEY_PTRS(k, KEY_PTRS(&b->key));
661
662 /*
663 * Move b to the end of the lru, and keep track of what this bucket was
664 * last used for:
665 */
666 list_move_tail(&b->list, &c->data_buckets);
667 bkey_copy_key(&b->key, k);
668 b->last_write_point = write_point;
669
670 b->sectors_free -= sectors;
671
672 for (i = 0; i < KEY_PTRS(&b->key); i++) {
673 SET_PTR_OFFSET(&b->key, i, PTR_OFFSET(&b->key, i) + sectors);
674
675 atomic_long_add(sectors,
676 &c->cache->sectors_written);
677 }
678
679 if (b->sectors_free < c->cache->sb.block_size)
680 b->sectors_free = 0;
681
682 /*
683 * k takes refcounts on the buckets it points to until it's inserted
684 * into the btree, but if we're done with this bucket we just transfer
685 * get_data_bucket()'s refcount.
686 */
687 if (b->sectors_free)
688 for (i = 0; i < KEY_PTRS(&b->key); i++)
689 atomic_inc(&PTR_BUCKET(c, &b->key, i)->pin);
690
691 spin_unlock(&c->data_bucket_lock);
692 return true;
693}
694
695/* Init */
696
697void bch_open_buckets_free(struct cache_set *c)
698{
699 struct open_bucket *b;
700
701 while (!list_empty(&c->data_buckets)) {
702 b = list_first_entry(&c->data_buckets,
703 struct open_bucket, list);
704 list_del(&b->list);
705 kfree(b);
706 }
707}
708
709int bch_open_buckets_alloc(struct cache_set *c)
710{
711 int i;
712
713 spin_lock_init(&c->data_bucket_lock);
714
715 for (i = 0; i < MAX_OPEN_BUCKETS; i++) {
716 struct open_bucket *b = kzalloc(sizeof(*b), GFP_KERNEL);
717
718 if (!b)
719 return -ENOMEM;
720
721 list_add(&b->list, &c->data_buckets);
722 }
723
724 return 0;
725}
726
727int bch_cache_allocator_start(struct cache *ca)
728{
729 struct task_struct *k = kthread_run(bch_allocator_thread,
730 ca, "bcache_allocator");
731 if (IS_ERR(k))
732 return PTR_ERR(k);
733
734 ca->alloc_thread = k;
735 return 0;
736}
1/*
2 * Primary bucket allocation code
3 *
4 * Copyright 2012 Google, Inc.
5 *
6 * Allocation in bcache is done in terms of buckets:
7 *
8 * Each bucket has associated an 8 bit gen; this gen corresponds to the gen in
9 * btree pointers - they must match for the pointer to be considered valid.
10 *
11 * Thus (assuming a bucket has no dirty data or metadata in it) we can reuse a
12 * bucket simply by incrementing its gen.
13 *
14 * The gens (along with the priorities; it's really the gens are important but
15 * the code is named as if it's the priorities) are written in an arbitrary list
16 * of buckets on disk, with a pointer to them in the journal header.
17 *
18 * When we invalidate a bucket, we have to write its new gen to disk and wait
19 * for that write to complete before we use it - otherwise after a crash we
20 * could have pointers that appeared to be good but pointed to data that had
21 * been overwritten.
22 *
23 * Since the gens and priorities are all stored contiguously on disk, we can
24 * batch this up: We fill up the free_inc list with freshly invalidated buckets,
25 * call prio_write(), and when prio_write() finishes we pull buckets off the
26 * free_inc list and optionally discard them.
27 *
28 * free_inc isn't the only freelist - if it was, we'd often to sleep while
29 * priorities and gens were being written before we could allocate. c->free is a
30 * smaller freelist, and buckets on that list are always ready to be used.
31 *
32 * If we've got discards enabled, that happens when a bucket moves from the
33 * free_inc list to the free list.
34 *
35 * There is another freelist, because sometimes we have buckets that we know
36 * have nothing pointing into them - these we can reuse without waiting for
37 * priorities to be rewritten. These come from freed btree nodes and buckets
38 * that garbage collection discovered no longer had valid keys pointing into
39 * them (because they were overwritten). That's the unused list - buckets on the
40 * unused list move to the free list, optionally being discarded in the process.
41 *
42 * It's also important to ensure that gens don't wrap around - with respect to
43 * either the oldest gen in the btree or the gen on disk. This is quite
44 * difficult to do in practice, but we explicitly guard against it anyways - if
45 * a bucket is in danger of wrapping around we simply skip invalidating it that
46 * time around, and we garbage collect or rewrite the priorities sooner than we
47 * would have otherwise.
48 *
49 * bch_bucket_alloc() allocates a single bucket from a specific cache.
50 *
51 * bch_bucket_alloc_set() allocates one or more buckets from different caches
52 * out of a cache set.
53 *
54 * free_some_buckets() drives all the processes described above. It's called
55 * from bch_bucket_alloc() and a few other places that need to make sure free
56 * buckets are ready.
57 *
58 * invalidate_buckets_(lru|fifo)() find buckets that are available to be
59 * invalidated, and then invalidate them and stick them on the free_inc list -
60 * in either lru or fifo order.
61 */
62
63#include "bcache.h"
64#include "btree.h"
65
66#include <linux/blkdev.h>
67#include <linux/freezer.h>
68#include <linux/kthread.h>
69#include <linux/random.h>
70#include <trace/events/bcache.h>
71
72/* Bucket heap / gen */
73
74uint8_t bch_inc_gen(struct cache *ca, struct bucket *b)
75{
76 uint8_t ret = ++b->gen;
77
78 ca->set->need_gc = max(ca->set->need_gc, bucket_gc_gen(b));
79 WARN_ON_ONCE(ca->set->need_gc > BUCKET_GC_GEN_MAX);
80
81 return ret;
82}
83
84void bch_rescale_priorities(struct cache_set *c, int sectors)
85{
86 struct cache *ca;
87 struct bucket *b;
88 unsigned next = c->nbuckets * c->sb.bucket_size / 1024;
89 unsigned i;
90 int r;
91
92 atomic_sub(sectors, &c->rescale);
93
94 do {
95 r = atomic_read(&c->rescale);
96
97 if (r >= 0)
98 return;
99 } while (atomic_cmpxchg(&c->rescale, r, r + next) != r);
100
101 mutex_lock(&c->bucket_lock);
102
103 c->min_prio = USHRT_MAX;
104
105 for_each_cache(ca, c, i)
106 for_each_bucket(b, ca)
107 if (b->prio &&
108 b->prio != BTREE_PRIO &&
109 !atomic_read(&b->pin)) {
110 b->prio--;
111 c->min_prio = min(c->min_prio, b->prio);
112 }
113
114 mutex_unlock(&c->bucket_lock);
115}
116
117/*
118 * Background allocation thread: scans for buckets to be invalidated,
119 * invalidates them, rewrites prios/gens (marking them as invalidated on disk),
120 * then optionally issues discard commands to the newly free buckets, then puts
121 * them on the various freelists.
122 */
123
124static inline bool can_inc_bucket_gen(struct bucket *b)
125{
126 return bucket_gc_gen(b) < BUCKET_GC_GEN_MAX;
127}
128
129bool bch_can_invalidate_bucket(struct cache *ca, struct bucket *b)
130{
131 BUG_ON(!ca->set->gc_mark_valid);
132
133 return (!GC_MARK(b) ||
134 GC_MARK(b) == GC_MARK_RECLAIMABLE) &&
135 !atomic_read(&b->pin) &&
136 can_inc_bucket_gen(b);
137}
138
139void __bch_invalidate_one_bucket(struct cache *ca, struct bucket *b)
140{
141 lockdep_assert_held(&ca->set->bucket_lock);
142 BUG_ON(GC_MARK(b) && GC_MARK(b) != GC_MARK_RECLAIMABLE);
143
144 if (GC_SECTORS_USED(b))
145 trace_bcache_invalidate(ca, b - ca->buckets);
146
147 bch_inc_gen(ca, b);
148 b->prio = INITIAL_PRIO;
149 atomic_inc(&b->pin);
150}
151
152static void bch_invalidate_one_bucket(struct cache *ca, struct bucket *b)
153{
154 __bch_invalidate_one_bucket(ca, b);
155
156 fifo_push(&ca->free_inc, b - ca->buckets);
157}
158
159/*
160 * Determines what order we're going to reuse buckets, smallest bucket_prio()
161 * first: we also take into account the number of sectors of live data in that
162 * bucket, and in order for that multiply to make sense we have to scale bucket
163 *
164 * Thus, we scale the bucket priorities so that the bucket with the smallest
165 * prio is worth 1/8th of what INITIAL_PRIO is worth.
166 */
167
168#define bucket_prio(b) \
169({ \
170 unsigned min_prio = (INITIAL_PRIO - ca->set->min_prio) / 8; \
171 \
172 (b->prio - ca->set->min_prio + min_prio) * GC_SECTORS_USED(b); \
173})
174
175#define bucket_max_cmp(l, r) (bucket_prio(l) < bucket_prio(r))
176#define bucket_min_cmp(l, r) (bucket_prio(l) > bucket_prio(r))
177
178static void invalidate_buckets_lru(struct cache *ca)
179{
180 struct bucket *b;
181 ssize_t i;
182
183 ca->heap.used = 0;
184
185 for_each_bucket(b, ca) {
186 if (!bch_can_invalidate_bucket(ca, b))
187 continue;
188
189 if (!heap_full(&ca->heap))
190 heap_add(&ca->heap, b, bucket_max_cmp);
191 else if (bucket_max_cmp(b, heap_peek(&ca->heap))) {
192 ca->heap.data[0] = b;
193 heap_sift(&ca->heap, 0, bucket_max_cmp);
194 }
195 }
196
197 for (i = ca->heap.used / 2 - 1; i >= 0; --i)
198 heap_sift(&ca->heap, i, bucket_min_cmp);
199
200 while (!fifo_full(&ca->free_inc)) {
201 if (!heap_pop(&ca->heap, b, bucket_min_cmp)) {
202 /*
203 * We don't want to be calling invalidate_buckets()
204 * multiple times when it can't do anything
205 */
206 ca->invalidate_needs_gc = 1;
207 wake_up_gc(ca->set);
208 return;
209 }
210
211 bch_invalidate_one_bucket(ca, b);
212 }
213}
214
215static void invalidate_buckets_fifo(struct cache *ca)
216{
217 struct bucket *b;
218 size_t checked = 0;
219
220 while (!fifo_full(&ca->free_inc)) {
221 if (ca->fifo_last_bucket < ca->sb.first_bucket ||
222 ca->fifo_last_bucket >= ca->sb.nbuckets)
223 ca->fifo_last_bucket = ca->sb.first_bucket;
224
225 b = ca->buckets + ca->fifo_last_bucket++;
226
227 if (bch_can_invalidate_bucket(ca, b))
228 bch_invalidate_one_bucket(ca, b);
229
230 if (++checked >= ca->sb.nbuckets) {
231 ca->invalidate_needs_gc = 1;
232 wake_up_gc(ca->set);
233 return;
234 }
235 }
236}
237
238static void invalidate_buckets_random(struct cache *ca)
239{
240 struct bucket *b;
241 size_t checked = 0;
242
243 while (!fifo_full(&ca->free_inc)) {
244 size_t n;
245 get_random_bytes(&n, sizeof(n));
246
247 n %= (size_t) (ca->sb.nbuckets - ca->sb.first_bucket);
248 n += ca->sb.first_bucket;
249
250 b = ca->buckets + n;
251
252 if (bch_can_invalidate_bucket(ca, b))
253 bch_invalidate_one_bucket(ca, b);
254
255 if (++checked >= ca->sb.nbuckets / 2) {
256 ca->invalidate_needs_gc = 1;
257 wake_up_gc(ca->set);
258 return;
259 }
260 }
261}
262
263static void invalidate_buckets(struct cache *ca)
264{
265 BUG_ON(ca->invalidate_needs_gc);
266
267 switch (CACHE_REPLACEMENT(&ca->sb)) {
268 case CACHE_REPLACEMENT_LRU:
269 invalidate_buckets_lru(ca);
270 break;
271 case CACHE_REPLACEMENT_FIFO:
272 invalidate_buckets_fifo(ca);
273 break;
274 case CACHE_REPLACEMENT_RANDOM:
275 invalidate_buckets_random(ca);
276 break;
277 }
278}
279
280#define allocator_wait(ca, cond) \
281do { \
282 while (1) { \
283 set_current_state(TASK_INTERRUPTIBLE); \
284 if (cond) \
285 break; \
286 \
287 mutex_unlock(&(ca)->set->bucket_lock); \
288 if (kthread_should_stop()) \
289 return 0; \
290 \
291 try_to_freeze(); \
292 schedule(); \
293 mutex_lock(&(ca)->set->bucket_lock); \
294 } \
295 __set_current_state(TASK_RUNNING); \
296} while (0)
297
298static int bch_allocator_push(struct cache *ca, long bucket)
299{
300 unsigned i;
301
302 /* Prios/gens are actually the most important reserve */
303 if (fifo_push(&ca->free[RESERVE_PRIO], bucket))
304 return true;
305
306 for (i = 0; i < RESERVE_NR; i++)
307 if (fifo_push(&ca->free[i], bucket))
308 return true;
309
310 return false;
311}
312
313static int bch_allocator_thread(void *arg)
314{
315 struct cache *ca = arg;
316
317 mutex_lock(&ca->set->bucket_lock);
318
319 while (1) {
320 /*
321 * First, we pull buckets off of the unused and free_inc lists,
322 * possibly issue discards to them, then we add the bucket to
323 * the free list:
324 */
325 while (!fifo_empty(&ca->free_inc)) {
326 long bucket;
327
328 fifo_pop(&ca->free_inc, bucket);
329
330 if (ca->discard) {
331 mutex_unlock(&ca->set->bucket_lock);
332 blkdev_issue_discard(ca->bdev,
333 bucket_to_sector(ca->set, bucket),
334 ca->sb.bucket_size, GFP_KERNEL, 0);
335 mutex_lock(&ca->set->bucket_lock);
336 }
337
338 allocator_wait(ca, bch_allocator_push(ca, bucket));
339 wake_up(&ca->set->btree_cache_wait);
340 wake_up(&ca->set->bucket_wait);
341 }
342
343 /*
344 * We've run out of free buckets, we need to find some buckets
345 * we can invalidate. First, invalidate them in memory and add
346 * them to the free_inc list:
347 */
348
349retry_invalidate:
350 allocator_wait(ca, ca->set->gc_mark_valid &&
351 !ca->invalidate_needs_gc);
352 invalidate_buckets(ca);
353
354 /*
355 * Now, we write their new gens to disk so we can start writing
356 * new stuff to them:
357 */
358 allocator_wait(ca, !atomic_read(&ca->set->prio_blocked));
359 if (CACHE_SYNC(&ca->set->sb)) {
360 /*
361 * This could deadlock if an allocation with a btree
362 * node locked ever blocked - having the btree node
363 * locked would block garbage collection, but here we're
364 * waiting on garbage collection before we invalidate
365 * and free anything.
366 *
367 * But this should be safe since the btree code always
368 * uses btree_check_reserve() before allocating now, and
369 * if it fails it blocks without btree nodes locked.
370 */
371 if (!fifo_full(&ca->free_inc))
372 goto retry_invalidate;
373
374 bch_prio_write(ca);
375 }
376 }
377}
378
379/* Allocation */
380
381long bch_bucket_alloc(struct cache *ca, unsigned reserve, bool wait)
382{
383 DEFINE_WAIT(w);
384 struct bucket *b;
385 long r;
386
387 /* fastpath */
388 if (fifo_pop(&ca->free[RESERVE_NONE], r) ||
389 fifo_pop(&ca->free[reserve], r))
390 goto out;
391
392 if (!wait) {
393 trace_bcache_alloc_fail(ca, reserve);
394 return -1;
395 }
396
397 do {
398 prepare_to_wait(&ca->set->bucket_wait, &w,
399 TASK_UNINTERRUPTIBLE);
400
401 mutex_unlock(&ca->set->bucket_lock);
402 schedule();
403 mutex_lock(&ca->set->bucket_lock);
404 } while (!fifo_pop(&ca->free[RESERVE_NONE], r) &&
405 !fifo_pop(&ca->free[reserve], r));
406
407 finish_wait(&ca->set->bucket_wait, &w);
408out:
409 wake_up_process(ca->alloc_thread);
410
411 trace_bcache_alloc(ca, reserve);
412
413 if (expensive_debug_checks(ca->set)) {
414 size_t iter;
415 long i;
416 unsigned j;
417
418 for (iter = 0; iter < prio_buckets(ca) * 2; iter++)
419 BUG_ON(ca->prio_buckets[iter] == (uint64_t) r);
420
421 for (j = 0; j < RESERVE_NR; j++)
422 fifo_for_each(i, &ca->free[j], iter)
423 BUG_ON(i == r);
424 fifo_for_each(i, &ca->free_inc, iter)
425 BUG_ON(i == r);
426 }
427
428 b = ca->buckets + r;
429
430 BUG_ON(atomic_read(&b->pin) != 1);
431
432 SET_GC_SECTORS_USED(b, ca->sb.bucket_size);
433
434 if (reserve <= RESERVE_PRIO) {
435 SET_GC_MARK(b, GC_MARK_METADATA);
436 SET_GC_MOVE(b, 0);
437 b->prio = BTREE_PRIO;
438 } else {
439 SET_GC_MARK(b, GC_MARK_RECLAIMABLE);
440 SET_GC_MOVE(b, 0);
441 b->prio = INITIAL_PRIO;
442 }
443
444 return r;
445}
446
447void __bch_bucket_free(struct cache *ca, struct bucket *b)
448{
449 SET_GC_MARK(b, 0);
450 SET_GC_SECTORS_USED(b, 0);
451}
452
453void bch_bucket_free(struct cache_set *c, struct bkey *k)
454{
455 unsigned i;
456
457 for (i = 0; i < KEY_PTRS(k); i++)
458 __bch_bucket_free(PTR_CACHE(c, k, i),
459 PTR_BUCKET(c, k, i));
460}
461
462int __bch_bucket_alloc_set(struct cache_set *c, unsigned reserve,
463 struct bkey *k, int n, bool wait)
464{
465 int i;
466
467 lockdep_assert_held(&c->bucket_lock);
468 BUG_ON(!n || n > c->caches_loaded || n > 8);
469
470 bkey_init(k);
471
472 /* sort by free space/prio of oldest data in caches */
473
474 for (i = 0; i < n; i++) {
475 struct cache *ca = c->cache_by_alloc[i];
476 long b = bch_bucket_alloc(ca, reserve, wait);
477
478 if (b == -1)
479 goto err;
480
481 k->ptr[i] = PTR(ca->buckets[b].gen,
482 bucket_to_sector(c, b),
483 ca->sb.nr_this_dev);
484
485 SET_KEY_PTRS(k, i + 1);
486 }
487
488 return 0;
489err:
490 bch_bucket_free(c, k);
491 bkey_put(c, k);
492 return -1;
493}
494
495int bch_bucket_alloc_set(struct cache_set *c, unsigned reserve,
496 struct bkey *k, int n, bool wait)
497{
498 int ret;
499 mutex_lock(&c->bucket_lock);
500 ret = __bch_bucket_alloc_set(c, reserve, k, n, wait);
501 mutex_unlock(&c->bucket_lock);
502 return ret;
503}
504
505/* Sector allocator */
506
507struct open_bucket {
508 struct list_head list;
509 unsigned last_write_point;
510 unsigned sectors_free;
511 BKEY_PADDED(key);
512};
513
514/*
515 * We keep multiple buckets open for writes, and try to segregate different
516 * write streams for better cache utilization: first we look for a bucket where
517 * the last write to it was sequential with the current write, and failing that
518 * we look for a bucket that was last used by the same task.
519 *
520 * The ideas is if you've got multiple tasks pulling data into the cache at the
521 * same time, you'll get better cache utilization if you try to segregate their
522 * data and preserve locality.
523 *
524 * For example, say you've starting Firefox at the same time you're copying a
525 * bunch of files. Firefox will likely end up being fairly hot and stay in the
526 * cache awhile, but the data you copied might not be; if you wrote all that
527 * data to the same buckets it'd get invalidated at the same time.
528 *
529 * Both of those tasks will be doing fairly random IO so we can't rely on
530 * detecting sequential IO to segregate their data, but going off of the task
531 * should be a sane heuristic.
532 */
533static struct open_bucket *pick_data_bucket(struct cache_set *c,
534 const struct bkey *search,
535 unsigned write_point,
536 struct bkey *alloc)
537{
538 struct open_bucket *ret, *ret_task = NULL;
539
540 list_for_each_entry_reverse(ret, &c->data_buckets, list)
541 if (!bkey_cmp(&ret->key, search))
542 goto found;
543 else if (ret->last_write_point == write_point)
544 ret_task = ret;
545
546 ret = ret_task ?: list_first_entry(&c->data_buckets,
547 struct open_bucket, list);
548found:
549 if (!ret->sectors_free && KEY_PTRS(alloc)) {
550 ret->sectors_free = c->sb.bucket_size;
551 bkey_copy(&ret->key, alloc);
552 bkey_init(alloc);
553 }
554
555 if (!ret->sectors_free)
556 ret = NULL;
557
558 return ret;
559}
560
561/*
562 * Allocates some space in the cache to write to, and k to point to the newly
563 * allocated space, and updates KEY_SIZE(k) and KEY_OFFSET(k) (to point to the
564 * end of the newly allocated space).
565 *
566 * May allocate fewer sectors than @sectors, KEY_SIZE(k) indicates how many
567 * sectors were actually allocated.
568 *
569 * If s->writeback is true, will not fail.
570 */
571bool bch_alloc_sectors(struct cache_set *c, struct bkey *k, unsigned sectors,
572 unsigned write_point, unsigned write_prio, bool wait)
573{
574 struct open_bucket *b;
575 BKEY_PADDED(key) alloc;
576 unsigned i;
577
578 /*
579 * We might have to allocate a new bucket, which we can't do with a
580 * spinlock held. So if we have to allocate, we drop the lock, allocate
581 * and then retry. KEY_PTRS() indicates whether alloc points to
582 * allocated bucket(s).
583 */
584
585 bkey_init(&alloc.key);
586 spin_lock(&c->data_bucket_lock);
587
588 while (!(b = pick_data_bucket(c, k, write_point, &alloc.key))) {
589 unsigned watermark = write_prio
590 ? RESERVE_MOVINGGC
591 : RESERVE_NONE;
592
593 spin_unlock(&c->data_bucket_lock);
594
595 if (bch_bucket_alloc_set(c, watermark, &alloc.key, 1, wait))
596 return false;
597
598 spin_lock(&c->data_bucket_lock);
599 }
600
601 /*
602 * If we had to allocate, we might race and not need to allocate the
603 * second time we call find_data_bucket(). If we allocated a bucket but
604 * didn't use it, drop the refcount bch_bucket_alloc_set() took:
605 */
606 if (KEY_PTRS(&alloc.key))
607 bkey_put(c, &alloc.key);
608
609 for (i = 0; i < KEY_PTRS(&b->key); i++)
610 EBUG_ON(ptr_stale(c, &b->key, i));
611
612 /* Set up the pointer to the space we're allocating: */
613
614 for (i = 0; i < KEY_PTRS(&b->key); i++)
615 k->ptr[i] = b->key.ptr[i];
616
617 sectors = min(sectors, b->sectors_free);
618
619 SET_KEY_OFFSET(k, KEY_OFFSET(k) + sectors);
620 SET_KEY_SIZE(k, sectors);
621 SET_KEY_PTRS(k, KEY_PTRS(&b->key));
622
623 /*
624 * Move b to the end of the lru, and keep track of what this bucket was
625 * last used for:
626 */
627 list_move_tail(&b->list, &c->data_buckets);
628 bkey_copy_key(&b->key, k);
629 b->last_write_point = write_point;
630
631 b->sectors_free -= sectors;
632
633 for (i = 0; i < KEY_PTRS(&b->key); i++) {
634 SET_PTR_OFFSET(&b->key, i, PTR_OFFSET(&b->key, i) + sectors);
635
636 atomic_long_add(sectors,
637 &PTR_CACHE(c, &b->key, i)->sectors_written);
638 }
639
640 if (b->sectors_free < c->sb.block_size)
641 b->sectors_free = 0;
642
643 /*
644 * k takes refcounts on the buckets it points to until it's inserted
645 * into the btree, but if we're done with this bucket we just transfer
646 * get_data_bucket()'s refcount.
647 */
648 if (b->sectors_free)
649 for (i = 0; i < KEY_PTRS(&b->key); i++)
650 atomic_inc(&PTR_BUCKET(c, &b->key, i)->pin);
651
652 spin_unlock(&c->data_bucket_lock);
653 return true;
654}
655
656/* Init */
657
658void bch_open_buckets_free(struct cache_set *c)
659{
660 struct open_bucket *b;
661
662 while (!list_empty(&c->data_buckets)) {
663 b = list_first_entry(&c->data_buckets,
664 struct open_bucket, list);
665 list_del(&b->list);
666 kfree(b);
667 }
668}
669
670int bch_open_buckets_alloc(struct cache_set *c)
671{
672 int i;
673
674 spin_lock_init(&c->data_bucket_lock);
675
676 for (i = 0; i < 6; i++) {
677 struct open_bucket *b = kzalloc(sizeof(*b), GFP_KERNEL);
678 if (!b)
679 return -ENOMEM;
680
681 list_add(&b->list, &c->data_buckets);
682 }
683
684 return 0;
685}
686
687int bch_cache_allocator_start(struct cache *ca)
688{
689 struct task_struct *k = kthread_run(bch_allocator_thread,
690 ca, "bcache_allocator");
691 if (IS_ERR(k))
692 return PTR_ERR(k);
693
694 ca->alloc_thread = k;
695 return 0;
696}