1 // SPDX-License-Identifier: GPL-2.0
5 #include "btree_cache.h"
7 #include "btree_iter.h"
8 #include "btree_locking.h"
14 #include <linux/prefetch.h>
15 #include <linux/sched/mm.h>
17 const char * const bch2_btree_node_flags[] = {
24 void bch2_recalc_btree_reserve(struct bch_fs *c)
26 unsigned i, reserve = 16;
28 if (!c->btree_roots_known[0].b)
31 for (i = 0; i < btree_id_nr_alive(c); i++) {
32 struct btree_root *r = bch2_btree_id_root(c, i);
35 reserve += min_t(unsigned, 1, r->b->c.level) * 8;
38 c->btree_cache.reserve = reserve;
41 static inline unsigned btree_cache_can_free(struct btree_cache *bc)
43 return max_t(int, 0, bc->used - bc->reserve);
46 static void btree_node_to_freedlist(struct btree_cache *bc, struct btree *b)
48 if (b->c.lock.readers)
49 list_move(&b->list, &bc->freed_pcpu);
51 list_move(&b->list, &bc->freed_nonpcpu);
54 static void btree_node_data_free(struct bch_fs *c, struct btree *b)
56 struct btree_cache *bc = &c->btree_cache;
58 EBUG_ON(btree_node_write_in_flight(b));
60 clear_btree_node_just_written(b);
62 kvpfree(b->data, btree_bytes(c));
67 munmap(b->aux_data, btree_aux_data_bytes(b));
73 btree_node_to_freedlist(bc, b);
76 static int bch2_btree_cache_cmp_fn(struct rhashtable_compare_arg *arg,
79 const struct btree *b = obj;
80 const u64 *v = arg->key;
82 return b->hash_val == *v ? 0 : 1;
85 static const struct rhashtable_params bch_btree_cache_params = {
86 .head_offset = offsetof(struct btree, hash),
87 .key_offset = offsetof(struct btree, hash_val),
88 .key_len = sizeof(u64),
89 .obj_cmpfn = bch2_btree_cache_cmp_fn,
92 static int btree_node_data_alloc(struct bch_fs *c, struct btree *b, gfp_t gfp)
94 BUG_ON(b->data || b->aux_data);
96 b->data = kvpmalloc(btree_bytes(c), gfp);
98 return -BCH_ERR_ENOMEM_btree_node_mem_alloc;
100 b->aux_data = kvmalloc(btree_aux_data_bytes(b), gfp);
102 b->aux_data = mmap(NULL, btree_aux_data_bytes(b),
103 PROT_READ|PROT_WRITE|PROT_EXEC,
104 MAP_PRIVATE|MAP_ANONYMOUS, 0, 0);
105 if (b->aux_data == MAP_FAILED)
109 kvpfree(b->data, btree_bytes(c));
111 return -BCH_ERR_ENOMEM_btree_node_mem_alloc;
117 static struct btree *__btree_node_mem_alloc(struct bch_fs *c, gfp_t gfp)
121 b = kzalloc(sizeof(struct btree), gfp);
125 bkey_btree_ptr_init(&b->key);
126 INIT_LIST_HEAD(&b->list);
127 INIT_LIST_HEAD(&b->write_blocked);
128 b->byte_order = ilog2(btree_bytes(c));
132 struct btree *__bch2_btree_node_mem_alloc(struct bch_fs *c)
134 struct btree_cache *bc = &c->btree_cache;
137 b = __btree_node_mem_alloc(c, GFP_KERNEL);
141 if (btree_node_data_alloc(c, b, GFP_KERNEL)) {
146 bch2_btree_lock_init(&b->c, 0);
149 list_add(&b->list, &bc->freeable);
153 /* Btree in memory cache - hash table */
155 void bch2_btree_node_hash_remove(struct btree_cache *bc, struct btree *b)
157 int ret = rhashtable_remove_fast(&bc->table, &b->hash, bch_btree_cache_params);
161 /* Cause future lookups for this node to fail: */
165 int __bch2_btree_node_hash_insert(struct btree_cache *bc, struct btree *b)
168 b->hash_val = btree_ptr_hash_val(&b->key);
170 return rhashtable_lookup_insert_fast(&bc->table, &b->hash,
171 bch_btree_cache_params);
174 int bch2_btree_node_hash_insert(struct btree_cache *bc, struct btree *b,
175 unsigned level, enum btree_id id)
182 mutex_lock(&bc->lock);
183 ret = __bch2_btree_node_hash_insert(bc, b);
185 list_add_tail(&b->list, &bc->live);
186 mutex_unlock(&bc->lock);
192 static inline struct btree *btree_cache_find(struct btree_cache *bc,
193 const struct bkey_i *k)
195 u64 v = btree_ptr_hash_val(k);
197 return rhashtable_lookup_fast(&bc->table, &v, bch_btree_cache_params);
201 * this version is for btree nodes that have already been freed (we're not
202 * reaping a real btree node)
204 static int __btree_node_reclaim(struct bch_fs *c, struct btree *b, bool flush)
206 struct btree_cache *bc = &c->btree_cache;
209 lockdep_assert_held(&bc->lock);
211 if (b->flags & ((1U << BTREE_NODE_dirty)|
212 (1U << BTREE_NODE_read_in_flight)|
213 (1U << BTREE_NODE_write_in_flight))) {
215 return -BCH_ERR_ENOMEM_btree_node_reclaim;
217 /* XXX: waiting on IO with btree cache lock held */
218 bch2_btree_node_wait_on_read(b);
219 bch2_btree_node_wait_on_write(b);
222 if (!six_trylock_intent(&b->c.lock))
223 return -BCH_ERR_ENOMEM_btree_node_reclaim;
225 if (!six_trylock_write(&b->c.lock))
226 goto out_unlock_intent;
228 /* recheck under lock */
229 if (b->flags & ((1U << BTREE_NODE_read_in_flight)|
230 (1U << BTREE_NODE_write_in_flight))) {
233 six_unlock_write(&b->c.lock);
234 six_unlock_intent(&b->c.lock);
238 if (btree_node_noevict(b) ||
239 btree_node_write_blocked(b) ||
240 btree_node_will_make_reachable(b))
243 if (btree_node_dirty(b)) {
247 * Using the underscore version because we don't want to compact
248 * bsets after the write, since this node is about to be evicted
249 * - unless btree verify mode is enabled, since it runs out of
250 * the post write cleanup:
252 if (bch2_verify_btree_ondisk)
253 bch2_btree_node_write(c, b, SIX_LOCK_intent,
254 BTREE_WRITE_cache_reclaim);
256 __bch2_btree_node_write(c, b,
257 BTREE_WRITE_cache_reclaim);
259 six_unlock_write(&b->c.lock);
260 six_unlock_intent(&b->c.lock);
264 if (b->hash_val && !ret)
265 trace_and_count(c, btree_cache_reap, c, b);
268 six_unlock_write(&b->c.lock);
270 six_unlock_intent(&b->c.lock);
271 ret = -BCH_ERR_ENOMEM_btree_node_reclaim;
275 static int btree_node_reclaim(struct bch_fs *c, struct btree *b)
277 return __btree_node_reclaim(c, b, false);
280 static int btree_node_write_and_reclaim(struct bch_fs *c, struct btree *b)
282 return __btree_node_reclaim(c, b, true);
285 static unsigned long bch2_btree_cache_scan(struct shrinker *shrink,
286 struct shrink_control *sc)
288 struct bch_fs *c = shrink->private_data;
289 struct btree_cache *bc = &c->btree_cache;
291 unsigned long nr = sc->nr_to_scan;
292 unsigned long can_free = 0;
293 unsigned long freed = 0;
294 unsigned long touched = 0;
296 unsigned long ret = SHRINK_STOP;
297 bool trigger_writes = atomic_read(&bc->dirty) + nr >=
300 if (bch2_btree_shrinker_disabled)
303 mutex_lock(&bc->lock);
304 flags = memalloc_nofs_save();
307 * It's _really_ critical that we don't free too many btree nodes - we
308 * have to always leave ourselves a reserve. The reserve is how we
309 * guarantee that allocating memory for a new btree node can always
310 * succeed, so that inserting keys into the btree can always succeed and
311 * IO can always make forward progress:
313 can_free = btree_cache_can_free(bc);
314 nr = min_t(unsigned long, nr, can_free);
317 list_for_each_entry_safe(b, t, &bc->freeable, list) {
319 * Leave a few nodes on the freeable list, so that a btree split
320 * won't have to hit the system allocator:
330 if (!btree_node_reclaim(c, b)) {
331 btree_node_data_free(c, b);
332 six_unlock_write(&b->c.lock);
333 six_unlock_intent(&b->c.lock);
338 list_for_each_entry_safe(b, t, &bc->live, list) {
341 if (btree_node_accessed(b)) {
342 clear_btree_node_accessed(b);
343 } else if (!btree_node_reclaim(c, b)) {
345 btree_node_data_free(c, b);
347 bch2_btree_node_hash_remove(bc, b);
348 six_unlock_write(&b->c.lock);
349 six_unlock_intent(&b->c.lock);
353 } else if (trigger_writes &&
354 btree_node_dirty(b) &&
355 !btree_node_will_make_reachable(b) &&
356 !btree_node_write_blocked(b) &&
357 six_trylock_read(&b->c.lock)) {
358 list_move(&bc->live, &b->list);
359 mutex_unlock(&bc->lock);
360 __bch2_btree_node_write(c, b, BTREE_WRITE_cache_reclaim);
361 six_unlock_read(&b->c.lock);
364 mutex_lock(&bc->lock);
372 if (&t->list != &bc->live)
373 list_move_tail(&bc->live, &t->list);
375 mutex_unlock(&bc->lock);
378 memalloc_nofs_restore(flags);
379 trace_and_count(c, btree_cache_scan, sc->nr_to_scan, can_free, ret);
383 static unsigned long bch2_btree_cache_count(struct shrinker *shrink,
384 struct shrink_control *sc)
386 struct bch_fs *c = shrink->private_data;
387 struct btree_cache *bc = &c->btree_cache;
389 if (bch2_btree_shrinker_disabled)
392 return btree_cache_can_free(bc);
395 void bch2_fs_btree_cache_exit(struct bch_fs *c)
397 struct btree_cache *bc = &c->btree_cache;
401 shrinker_free(bc->shrink);
403 /* vfree() can allocate memory: */
404 flags = memalloc_nofs_save();
405 mutex_lock(&bc->lock);
408 list_move(&c->verify_data->list, &bc->live);
410 kvpfree(c->verify_ondisk, btree_bytes(c));
412 for (i = 0; i < btree_id_nr_alive(c); i++) {
413 struct btree_root *r = bch2_btree_id_root(c, i);
416 list_add(&r->b->list, &bc->live);
419 list_splice(&bc->freeable, &bc->live);
421 while (!list_empty(&bc->live)) {
422 b = list_first_entry(&bc->live, struct btree, list);
424 BUG_ON(btree_node_read_in_flight(b) ||
425 btree_node_write_in_flight(b));
427 if (btree_node_dirty(b))
428 bch2_btree_complete_write(c, b, btree_current_write(b));
429 clear_btree_node_dirty_acct(c, b);
431 btree_node_data_free(c, b);
434 BUG_ON(atomic_read(&c->btree_cache.dirty));
436 list_splice(&bc->freed_pcpu, &bc->freed_nonpcpu);
438 while (!list_empty(&bc->freed_nonpcpu)) {
439 b = list_first_entry(&bc->freed_nonpcpu, struct btree, list);
441 six_lock_exit(&b->c.lock);
445 mutex_unlock(&bc->lock);
446 memalloc_nofs_restore(flags);
448 if (bc->table_init_done)
449 rhashtable_destroy(&bc->table);
452 int bch2_fs_btree_cache_init(struct bch_fs *c)
454 struct btree_cache *bc = &c->btree_cache;
455 struct shrinker *shrink;
459 ret = rhashtable_init(&bc->table, &bch_btree_cache_params);
463 bc->table_init_done = true;
465 bch2_recalc_btree_reserve(c);
467 for (i = 0; i < bc->reserve; i++)
468 if (!__bch2_btree_node_mem_alloc(c))
471 list_splice_init(&bc->live, &bc->freeable);
473 mutex_init(&c->verify_lock);
475 shrink = shrinker_alloc(0, "%s-btree_cache", c->name);
479 shrink->count_objects = bch2_btree_cache_count;
480 shrink->scan_objects = bch2_btree_cache_scan;
482 shrink->private_data = c;
483 shrinker_register(shrink);
487 return -BCH_ERR_ENOMEM_fs_btree_cache_init;
490 void bch2_fs_btree_cache_init_early(struct btree_cache *bc)
492 mutex_init(&bc->lock);
493 INIT_LIST_HEAD(&bc->live);
494 INIT_LIST_HEAD(&bc->freeable);
495 INIT_LIST_HEAD(&bc->freed_pcpu);
496 INIT_LIST_HEAD(&bc->freed_nonpcpu);
500 * We can only have one thread cannibalizing other cached btree nodes at a time,
501 * or we'll deadlock. We use an open coded mutex to ensure that, which a
502 * cannibalize_bucket() will take. This means every time we unlock the root of
503 * the btree, we need to release this lock if we have it held.
505 void bch2_btree_cache_cannibalize_unlock(struct bch_fs *c)
507 struct btree_cache *bc = &c->btree_cache;
509 if (bc->alloc_lock == current) {
510 trace_and_count(c, btree_cache_cannibalize_unlock, c);
511 bc->alloc_lock = NULL;
512 closure_wake_up(&bc->alloc_wait);
516 int bch2_btree_cache_cannibalize_lock(struct bch_fs *c, struct closure *cl)
518 struct btree_cache *bc = &c->btree_cache;
519 struct task_struct *old;
521 old = cmpxchg(&bc->alloc_lock, NULL, current);
522 if (old == NULL || old == current)
526 trace_and_count(c, btree_cache_cannibalize_lock_fail, c);
527 return -BCH_ERR_ENOMEM_btree_cache_cannibalize_lock;
530 closure_wait(&bc->alloc_wait, cl);
532 /* Try again, after adding ourselves to waitlist */
533 old = cmpxchg(&bc->alloc_lock, NULL, current);
534 if (old == NULL || old == current) {
536 closure_wake_up(&bc->alloc_wait);
540 trace_and_count(c, btree_cache_cannibalize_lock_fail, c);
541 return -BCH_ERR_btree_cache_cannibalize_lock_blocked;
544 trace_and_count(c, btree_cache_cannibalize_lock, c);
548 static struct btree *btree_node_cannibalize(struct bch_fs *c)
550 struct btree_cache *bc = &c->btree_cache;
553 list_for_each_entry_reverse(b, &bc->live, list)
554 if (!btree_node_reclaim(c, b))
558 list_for_each_entry_reverse(b, &bc->live, list)
559 if (!btree_node_write_and_reclaim(c, b))
563 * Rare case: all nodes were intent-locked.
566 WARN_ONCE(1, "btree cache cannibalize failed\n");
571 struct btree *bch2_btree_node_mem_alloc(struct btree_trans *trans, bool pcpu_read_locks)
573 struct bch_fs *c = trans->c;
574 struct btree_cache *bc = &c->btree_cache;
575 struct list_head *freed = pcpu_read_locks
577 : &bc->freed_nonpcpu;
578 struct btree *b, *b2;
579 u64 start_time = local_clock();
582 flags = memalloc_nofs_save();
583 mutex_lock(&bc->lock);
586 * We never free struct btree itself, just the memory that holds the on
587 * disk node. Check the freed list before allocating a new one:
589 list_for_each_entry(b, freed, list)
590 if (!btree_node_reclaim(c, b)) {
591 list_del_init(&b->list);
595 b = __btree_node_mem_alloc(c, GFP_NOWAIT|__GFP_NOWARN);
597 mutex_unlock(&bc->lock);
598 bch2_trans_unlock(trans);
599 b = __btree_node_mem_alloc(c, GFP_KERNEL);
602 mutex_lock(&bc->lock);
605 bch2_btree_lock_init(&b->c, pcpu_read_locks ? SIX_LOCK_INIT_PCPU : 0);
607 BUG_ON(!six_trylock_intent(&b->c.lock));
608 BUG_ON(!six_trylock_write(&b->c.lock));
612 * btree_free() doesn't free memory; it sticks the node on the end of
613 * the list. Check if there's any freed nodes there:
615 list_for_each_entry(b2, &bc->freeable, list)
616 if (!btree_node_reclaim(c, b2)) {
617 swap(b->data, b2->data);
618 swap(b->aux_data, b2->aux_data);
619 btree_node_to_freedlist(bc, b2);
620 six_unlock_write(&b2->c.lock);
621 six_unlock_intent(&b2->c.lock);
625 mutex_unlock(&bc->lock);
627 if (btree_node_data_alloc(c, b, GFP_NOWAIT|__GFP_NOWARN)) {
628 bch2_trans_unlock(trans);
629 if (btree_node_data_alloc(c, b, GFP_KERNEL|__GFP_NOWARN))
633 mutex_lock(&bc->lock);
636 mutex_unlock(&bc->lock);
638 BUG_ON(btree_node_hashed(b));
639 BUG_ON(btree_node_dirty(b));
640 BUG_ON(btree_node_write_in_flight(b));
647 b->whiteout_u64s = 0;
648 bch2_btree_keys_init(b);
649 set_btree_node_accessed(b);
651 bch2_time_stats_update(&c->times[BCH_TIME_btree_node_mem_alloc],
654 memalloc_nofs_restore(flags);
657 mutex_lock(&bc->lock);
659 /* Try to cannibalize another cached btree node: */
660 if (bc->alloc_lock == current) {
661 b2 = btree_node_cannibalize(c);
662 clear_btree_node_just_written(b2);
663 bch2_btree_node_hash_remove(bc, b2);
666 swap(b->data, b2->data);
667 swap(b->aux_data, b2->aux_data);
668 btree_node_to_freedlist(bc, b2);
669 six_unlock_write(&b2->c.lock);
670 six_unlock_intent(&b2->c.lock);
673 list_del_init(&b->list);
676 mutex_unlock(&bc->lock);
678 trace_and_count(c, btree_cache_cannibalize, c);
682 mutex_unlock(&bc->lock);
683 memalloc_nofs_restore(flags);
684 return ERR_PTR(-BCH_ERR_ENOMEM_btree_node_mem_alloc);
687 /* Slowpath, don't want it inlined into btree_iter_traverse() */
688 static noinline struct btree *bch2_btree_node_fill(struct btree_trans *trans,
689 struct btree_path *path,
690 const struct bkey_i *k,
691 enum btree_id btree_id,
693 enum six_lock_type lock_type,
696 struct bch_fs *c = trans->c;
697 struct btree_cache *bc = &c->btree_cache;
701 BUG_ON(level + 1 >= BTREE_MAX_DEPTH);
703 * Parent node must be locked, else we could read in a btree node that's
706 if (path && !bch2_btree_node_relock(trans, path, level + 1)) {
707 trace_and_count(c, trans_restart_relock_parent_for_fill, trans, _THIS_IP_, path);
708 return ERR_PTR(btree_trans_restart(trans, BCH_ERR_transaction_restart_fill_relock));
711 b = bch2_btree_node_mem_alloc(trans, level != 0);
713 if (bch2_err_matches(PTR_ERR_OR_ZERO(b), ENOMEM)) {
714 trans->memory_allocation_failure = true;
715 trace_and_count(c, trans_restart_memory_allocation_failure, trans, _THIS_IP_, path);
716 return ERR_PTR(btree_trans_restart(trans, BCH_ERR_transaction_restart_fill_mem_alloc_fail));
723 * Btree nodes read in from disk should not have the accessed bit set
724 * initially, so that linear scans don't thrash the cache:
726 clear_btree_node_accessed(b);
728 bkey_copy(&b->key, k);
729 if (bch2_btree_node_hash_insert(bc, b, level, btree_id)) {
730 /* raced with another fill: */
732 /* mark as unhashed... */
735 mutex_lock(&bc->lock);
736 list_add(&b->list, &bc->freeable);
737 mutex_unlock(&bc->lock);
739 six_unlock_write(&b->c.lock);
740 six_unlock_intent(&b->c.lock);
744 set_btree_node_read_in_flight(b);
746 six_unlock_write(&b->c.lock);
747 seq = six_lock_seq(&b->c.lock);
748 six_unlock_intent(&b->c.lock);
750 /* Unlock before doing IO: */
752 bch2_trans_unlock_noassert(trans);
754 bch2_btree_node_read(c, b, sync);
760 int ret = bch2_trans_relock(trans) ?:
761 bch2_btree_path_relock_intent(trans, path);
763 BUG_ON(!trans->restarted);
768 if (!six_relock_type(&b->c.lock, lock_type, seq)) {
770 trace_and_count(c, trans_restart_relock_after_fill, trans, _THIS_IP_, path);
771 return ERR_PTR(btree_trans_restart(trans, BCH_ERR_transaction_restart_relock_after_fill));
777 static noinline void btree_bad_header(struct bch_fs *c, struct btree *b)
779 struct printbuf buf = PRINTBUF;
781 if (c->curr_recovery_pass <= BCH_RECOVERY_PASS_check_allocations)
785 "btree node header doesn't match ptr\n"
786 "btree %s level %u\n"
788 bch2_btree_id_str(b->c.btree_id), b->c.level);
789 bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(&b->key));
791 prt_printf(&buf, "\nheader: btree %s level %llu\n"
793 bch2_btree_id_str(BTREE_NODE_ID(b->data)),
794 BTREE_NODE_LEVEL(b->data));
795 bch2_bpos_to_text(&buf, b->data->min_key);
797 prt_printf(&buf, "\nmax ");
798 bch2_bpos_to_text(&buf, b->data->max_key);
800 bch2_fs_inconsistent(c, "%s", buf.buf);
804 static inline void btree_check_header(struct bch_fs *c, struct btree *b)
806 if (b->c.btree_id != BTREE_NODE_ID(b->data) ||
807 b->c.level != BTREE_NODE_LEVEL(b->data) ||
808 !bpos_eq(b->data->max_key, b->key.k.p) ||
809 (b->key.k.type == KEY_TYPE_btree_ptr_v2 &&
810 !bpos_eq(b->data->min_key,
811 bkey_i_to_btree_ptr_v2(&b->key)->v.min_key)))
812 btree_bad_header(c, b);
815 static struct btree *__bch2_btree_node_get(struct btree_trans *trans, struct btree_path *path,
816 const struct bkey_i *k, unsigned level,
817 enum six_lock_type lock_type,
818 unsigned long trace_ip)
820 struct bch_fs *c = trans->c;
821 struct btree_cache *bc = &c->btree_cache;
824 bool need_relock = false;
827 EBUG_ON(level >= BTREE_MAX_DEPTH);
829 b = btree_cache_find(bc, k);
832 * We must have the parent locked to call bch2_btree_node_fill(),
833 * else we could read in a btree node from disk that's been
836 b = bch2_btree_node_fill(trans, path, k, path->btree_id,
837 level, lock_type, true);
840 /* We raced and found the btree node in the cache */
847 if (btree_node_read_locked(path, level + 1))
848 btree_node_unlock(trans, path, level + 1);
850 ret = btree_node_lock(trans, path, &b->c, level, lock_type, trace_ip);
851 if (bch2_err_matches(ret, BCH_ERR_transaction_restart))
856 if (unlikely(b->hash_val != btree_ptr_hash_val(k) ||
857 b->c.level != level ||
859 six_unlock_type(&b->c.lock, lock_type);
860 if (bch2_btree_node_relock(trans, path, level + 1))
863 trace_and_count(c, trans_restart_btree_node_reused, trans, trace_ip, path);
864 return ERR_PTR(btree_trans_restart(trans, BCH_ERR_transaction_restart_lock_node_reused));
867 /* avoid atomic set bit if it's not needed: */
868 if (!btree_node_accessed(b))
869 set_btree_node_accessed(b);
872 if (unlikely(btree_node_read_in_flight(b))) {
873 u32 seq = six_lock_seq(&b->c.lock);
875 six_unlock_type(&b->c.lock, lock_type);
876 bch2_trans_unlock(trans);
879 bch2_btree_node_wait_on_read(b);
882 * should_be_locked is not set on this path yet, so we need to
883 * relock it specifically:
885 if (!six_relock_type(&b->c.lock, lock_type, seq))
889 if (unlikely(need_relock)) {
890 ret = bch2_trans_relock(trans) ?:
891 bch2_btree_path_relock_intent(trans, path);
893 six_unlock_type(&b->c.lock, lock_type);
898 prefetch(b->aux_data);
900 for_each_bset(b, t) {
901 void *p = (u64 *) b->aux_data + t->aux_data_offset;
903 prefetch(p + L1_CACHE_BYTES * 0);
904 prefetch(p + L1_CACHE_BYTES * 1);
905 prefetch(p + L1_CACHE_BYTES * 2);
908 if (unlikely(btree_node_read_error(b))) {
909 six_unlock_type(&b->c.lock, lock_type);
910 return ERR_PTR(-EIO);
913 EBUG_ON(b->c.btree_id != path->btree_id);
914 EBUG_ON(BTREE_NODE_LEVEL(b->data) != level);
915 btree_check_header(c, b);
921 * bch2_btree_node_get - find a btree node in the cache and lock it, reading it
922 * in from disk if necessary.
924 * @trans: btree transaction object
925 * @path: btree_path being traversed
926 * @k: pointer to btree node (generally KEY_TYPE_btree_ptr_v2)
927 * @level: level of btree node being looked up (0 == leaf node)
928 * @lock_type: SIX_LOCK_read or SIX_LOCK_intent
929 * @trace_ip: ip of caller of btree iterator code (i.e. caller of bch2_btree_iter_peek())
931 * The btree node will have either a read or a write lock held, depending on
932 * the @write parameter.
934 * Returns: btree node or ERR_PTR()
936 struct btree *bch2_btree_node_get(struct btree_trans *trans, struct btree_path *path,
937 const struct bkey_i *k, unsigned level,
938 enum six_lock_type lock_type,
939 unsigned long trace_ip)
941 struct bch_fs *c = trans->c;
946 EBUG_ON(level >= BTREE_MAX_DEPTH);
948 b = btree_node_mem_ptr(k);
951 * Check b->hash_val _before_ calling btree_node_lock() - this might not
952 * be the node we want anymore, and trying to lock the wrong node could
953 * cause an unneccessary transaction restart:
955 if (unlikely(!c->opts.btree_node_mem_ptr_optimization ||
957 b->hash_val != btree_ptr_hash_val(k)))
958 return __bch2_btree_node_get(trans, path, k, level, lock_type, trace_ip);
960 if (btree_node_read_locked(path, level + 1))
961 btree_node_unlock(trans, path, level + 1);
963 ret = btree_node_lock(trans, path, &b->c, level, lock_type, trace_ip);
964 if (bch2_err_matches(ret, BCH_ERR_transaction_restart))
969 if (unlikely(b->hash_val != btree_ptr_hash_val(k) ||
970 b->c.level != level ||
972 six_unlock_type(&b->c.lock, lock_type);
973 if (bch2_btree_node_relock(trans, path, level + 1))
974 return __bch2_btree_node_get(trans, path, k, level, lock_type, trace_ip);
976 trace_and_count(c, trans_restart_btree_node_reused, trans, trace_ip, path);
977 return ERR_PTR(btree_trans_restart(trans, BCH_ERR_transaction_restart_lock_node_reused));
980 if (unlikely(btree_node_read_in_flight(b))) {
981 six_unlock_type(&b->c.lock, lock_type);
982 return __bch2_btree_node_get(trans, path, k, level, lock_type, trace_ip);
985 prefetch(b->aux_data);
987 for_each_bset(b, t) {
988 void *p = (u64 *) b->aux_data + t->aux_data_offset;
990 prefetch(p + L1_CACHE_BYTES * 0);
991 prefetch(p + L1_CACHE_BYTES * 1);
992 prefetch(p + L1_CACHE_BYTES * 2);
995 /* avoid atomic set bit if it's not needed: */
996 if (!btree_node_accessed(b))
997 set_btree_node_accessed(b);
999 if (unlikely(btree_node_read_error(b))) {
1000 six_unlock_type(&b->c.lock, lock_type);
1001 return ERR_PTR(-EIO);
1004 EBUG_ON(b->c.btree_id != path->btree_id);
1005 EBUG_ON(BTREE_NODE_LEVEL(b->data) != level);
1006 btree_check_header(c, b);
1011 struct btree *bch2_btree_node_get_noiter(struct btree_trans *trans,
1012 const struct bkey_i *k,
1013 enum btree_id btree_id,
1017 struct bch_fs *c = trans->c;
1018 struct btree_cache *bc = &c->btree_cache;
1020 struct bset_tree *t;
1023 EBUG_ON(level >= BTREE_MAX_DEPTH);
1025 if (c->opts.btree_node_mem_ptr_optimization) {
1026 b = btree_node_mem_ptr(k);
1031 b = btree_cache_find(bc, k);
1036 b = bch2_btree_node_fill(trans, NULL, k, btree_id,
1037 level, SIX_LOCK_read, true);
1039 /* We raced and found the btree node in the cache */
1044 !bch2_btree_cache_cannibalize_lock(c, NULL))
1051 ret = btree_node_lock_nopath(trans, &b->c, SIX_LOCK_read, _THIS_IP_);
1052 if (bch2_err_matches(ret, BCH_ERR_transaction_restart))
1053 return ERR_PTR(ret);
1057 if (unlikely(b->hash_val != btree_ptr_hash_val(k) ||
1058 b->c.btree_id != btree_id ||
1059 b->c.level != level)) {
1060 six_unlock_read(&b->c.lock);
1065 /* XXX: waiting on IO with btree locks held: */
1066 __bch2_btree_node_wait_on_read(b);
1068 prefetch(b->aux_data);
1070 for_each_bset(b, t) {
1071 void *p = (u64 *) b->aux_data + t->aux_data_offset;
1073 prefetch(p + L1_CACHE_BYTES * 0);
1074 prefetch(p + L1_CACHE_BYTES * 1);
1075 prefetch(p + L1_CACHE_BYTES * 2);
1078 /* avoid atomic set bit if it's not needed: */
1079 if (!btree_node_accessed(b))
1080 set_btree_node_accessed(b);
1082 if (unlikely(btree_node_read_error(b))) {
1083 six_unlock_read(&b->c.lock);
1088 EBUG_ON(b->c.btree_id != btree_id);
1089 EBUG_ON(BTREE_NODE_LEVEL(b->data) != level);
1090 btree_check_header(c, b);
1092 bch2_btree_cache_cannibalize_unlock(c);
1096 int bch2_btree_node_prefetch(struct btree_trans *trans,
1097 struct btree_path *path,
1098 const struct bkey_i *k,
1099 enum btree_id btree_id, unsigned level)
1101 struct bch_fs *c = trans->c;
1102 struct btree_cache *bc = &c->btree_cache;
1105 BUG_ON(trans && !btree_node_locked(path, level + 1));
1106 BUG_ON(level >= BTREE_MAX_DEPTH);
1108 b = btree_cache_find(bc, k);
1112 b = bch2_btree_node_fill(trans, path, k, btree_id,
1113 level, SIX_LOCK_read, false);
1114 return PTR_ERR_OR_ZERO(b);
1117 void bch2_btree_node_evict(struct btree_trans *trans, const struct bkey_i *k)
1119 struct bch_fs *c = trans->c;
1120 struct btree_cache *bc = &c->btree_cache;
1123 b = btree_cache_find(bc, k);
1127 /* not allowed to wait on io with btree locks held: */
1129 /* XXX we're called from btree_gc which will be holding other btree
1132 __bch2_btree_node_wait_on_read(b);
1133 __bch2_btree_node_wait_on_write(b);
1135 btree_node_lock_nopath_nofail(trans, &b->c, SIX_LOCK_intent);
1136 btree_node_lock_nopath_nofail(trans, &b->c, SIX_LOCK_write);
1138 if (btree_node_dirty(b)) {
1139 __bch2_btree_node_write(c, b, BTREE_WRITE_cache_reclaim);
1140 six_unlock_write(&b->c.lock);
1141 six_unlock_intent(&b->c.lock);
1145 BUG_ON(btree_node_dirty(b));
1147 mutex_lock(&bc->lock);
1148 btree_node_data_free(c, b);
1149 bch2_btree_node_hash_remove(bc, b);
1150 mutex_unlock(&bc->lock);
1152 six_unlock_write(&b->c.lock);
1153 six_unlock_intent(&b->c.lock);
1156 const char *bch2_btree_id_str(enum btree_id btree)
1158 return btree < BTREE_ID_NR ? __bch2_btree_ids[btree] : "(unknown)";
1161 void bch2_btree_pos_to_text(struct printbuf *out, struct bch_fs *c, const struct btree *b)
1163 prt_printf(out, "%s level %u/%u\n ",
1164 bch2_btree_id_str(b->c.btree_id),
1166 bch2_btree_id_root(c, b->c.btree_id)->level);
1167 bch2_bkey_val_to_text(out, c, bkey_i_to_s_c(&b->key));
1170 void bch2_btree_node_to_text(struct printbuf *out, struct bch_fs *c, const struct btree *b)
1172 struct bset_stats stats;
1174 memset(&stats, 0, sizeof(stats));
1176 bch2_btree_keys_stats(b, &stats);
1178 prt_printf(out, "l %u ", b->c.level);
1179 bch2_bpos_to_text(out, b->data->min_key);
1180 prt_printf(out, " - ");
1181 bch2_bpos_to_text(out, b->data->max_key);
1182 prt_printf(out, ":\n"
1184 bch2_val_to_text(out, c, bkey_i_to_s_c(&b->key));
1189 bch2_bkey_format_to_text(out, &b->format);
1192 " unpack fn len: %u\n"
1193 " bytes used %zu/%zu (%zu%% full)\n"
1194 " sib u64s: %u, %u (merge threshold %u)\n"
1195 " nr packed keys %u\n"
1196 " nr unpacked keys %u\n"
1198 " failed unpacked %zu\n",
1200 b->nr.live_u64s * sizeof(u64),
1201 btree_bytes(c) - sizeof(struct btree_node),
1202 b->nr.live_u64s * 100 / btree_max_u64s(c),
1205 c->btree_foreground_merge_threshold,
1207 b->nr.unpacked_keys,
1212 void bch2_btree_cache_to_text(struct printbuf *out, const struct bch_fs *c)
1214 prt_printf(out, "nr nodes:\t\t%u\n", c->btree_cache.used);
1215 prt_printf(out, "nr dirty:\t\t%u\n", atomic_read(&c->btree_cache.dirty));
1216 prt_printf(out, "cannibalize lock:\t%p\n", c->btree_cache.alloc_lock);