Linux 4.19.133
[linux/fpc-iii.git] / drivers / md / persistent-data / dm-btree.c
blob8aae0624a2971e939fa79acd656f98035049460d
1 /*
2 * Copyright (C) 2011 Red Hat, Inc.
4 * This file is released under the GPL.
5 */
7 #include "dm-btree-internal.h"
8 #include "dm-space-map.h"
9 #include "dm-transaction-manager.h"
11 #include <linux/export.h>
12 #include <linux/device-mapper.h>
14 #define DM_MSG_PREFIX "btree"
16 /*----------------------------------------------------------------
17 * Array manipulation
18 *--------------------------------------------------------------*/
19 static void memcpy_disk(void *dest, const void *src, size_t len)
20 __dm_written_to_disk(src)
22 memcpy(dest, src, len);
23 __dm_unbless_for_disk(src);
26 static void array_insert(void *base, size_t elt_size, unsigned nr_elts,
27 unsigned index, void *elt)
28 __dm_written_to_disk(elt)
30 if (index < nr_elts)
31 memmove(base + (elt_size * (index + 1)),
32 base + (elt_size * index),
33 (nr_elts - index) * elt_size);
35 memcpy_disk(base + (elt_size * index), elt, elt_size);
38 /*----------------------------------------------------------------*/
40 /* makes the assumption that no two keys are the same. */
41 static int bsearch(struct btree_node *n, uint64_t key, int want_hi)
43 int lo = -1, hi = le32_to_cpu(n->header.nr_entries);
45 while (hi - lo > 1) {
46 int mid = lo + ((hi - lo) / 2);
47 uint64_t mid_key = le64_to_cpu(n->keys[mid]);
49 if (mid_key == key)
50 return mid;
52 if (mid_key < key)
53 lo = mid;
54 else
55 hi = mid;
58 return want_hi ? hi : lo;
61 int lower_bound(struct btree_node *n, uint64_t key)
63 return bsearch(n, key, 0);
66 static int upper_bound(struct btree_node *n, uint64_t key)
68 return bsearch(n, key, 1);
71 void inc_children(struct dm_transaction_manager *tm, struct btree_node *n,
72 struct dm_btree_value_type *vt)
74 unsigned i;
75 uint32_t nr_entries = le32_to_cpu(n->header.nr_entries);
77 if (le32_to_cpu(n->header.flags) & INTERNAL_NODE)
78 for (i = 0; i < nr_entries; i++)
79 dm_tm_inc(tm, value64(n, i));
80 else if (vt->inc)
81 for (i = 0; i < nr_entries; i++)
82 vt->inc(vt->context, value_ptr(n, i));
85 static int insert_at(size_t value_size, struct btree_node *node, unsigned index,
86 uint64_t key, void *value)
87 __dm_written_to_disk(value)
89 uint32_t nr_entries = le32_to_cpu(node->header.nr_entries);
90 __le64 key_le = cpu_to_le64(key);
92 if (index > nr_entries ||
93 index >= le32_to_cpu(node->header.max_entries)) {
94 DMERR("too many entries in btree node for insert");
95 __dm_unbless_for_disk(value);
96 return -ENOMEM;
99 __dm_bless_for_disk(&key_le);
101 array_insert(node->keys, sizeof(*node->keys), nr_entries, index, &key_le);
102 array_insert(value_base(node), value_size, nr_entries, index, value);
103 node->header.nr_entries = cpu_to_le32(nr_entries + 1);
105 return 0;
108 /*----------------------------------------------------------------*/
111 * We want 3n entries (for some n). This works more nicely for repeated
112 * insert remove loops than (2n + 1).
114 static uint32_t calc_max_entries(size_t value_size, size_t block_size)
116 uint32_t total, n;
117 size_t elt_size = sizeof(uint64_t) + value_size; /* key + value */
119 block_size -= sizeof(struct node_header);
120 total = block_size / elt_size;
121 n = total / 3; /* rounds down */
123 return 3 * n;
126 int dm_btree_empty(struct dm_btree_info *info, dm_block_t *root)
128 int r;
129 struct dm_block *b;
130 struct btree_node *n;
131 size_t block_size;
132 uint32_t max_entries;
134 r = new_block(info, &b);
135 if (r < 0)
136 return r;
138 block_size = dm_bm_block_size(dm_tm_get_bm(info->tm));
139 max_entries = calc_max_entries(info->value_type.size, block_size);
141 n = dm_block_data(b);
142 memset(n, 0, block_size);
143 n->header.flags = cpu_to_le32(LEAF_NODE);
144 n->header.nr_entries = cpu_to_le32(0);
145 n->header.max_entries = cpu_to_le32(max_entries);
146 n->header.value_size = cpu_to_le32(info->value_type.size);
148 *root = dm_block_location(b);
149 unlock_block(info, b);
151 return 0;
153 EXPORT_SYMBOL_GPL(dm_btree_empty);
155 /*----------------------------------------------------------------*/
158 * Deletion uses a recursive algorithm, since we have limited stack space
159 * we explicitly manage our own stack on the heap.
161 #define MAX_SPINE_DEPTH 64
162 struct frame {
163 struct dm_block *b;
164 struct btree_node *n;
165 unsigned level;
166 unsigned nr_children;
167 unsigned current_child;
170 struct del_stack {
171 struct dm_btree_info *info;
172 struct dm_transaction_manager *tm;
173 int top;
174 struct frame spine[MAX_SPINE_DEPTH];
177 static int top_frame(struct del_stack *s, struct frame **f)
179 if (s->top < 0) {
180 DMERR("btree deletion stack empty");
181 return -EINVAL;
184 *f = s->spine + s->top;
186 return 0;
189 static int unprocessed_frames(struct del_stack *s)
191 return s->top >= 0;
194 static void prefetch_children(struct del_stack *s, struct frame *f)
196 unsigned i;
197 struct dm_block_manager *bm = dm_tm_get_bm(s->tm);
199 for (i = 0; i < f->nr_children; i++)
200 dm_bm_prefetch(bm, value64(f->n, i));
203 static bool is_internal_level(struct dm_btree_info *info, struct frame *f)
205 return f->level < (info->levels - 1);
208 static int push_frame(struct del_stack *s, dm_block_t b, unsigned level)
210 int r;
211 uint32_t ref_count;
213 if (s->top >= MAX_SPINE_DEPTH - 1) {
214 DMERR("btree deletion stack out of memory");
215 return -ENOMEM;
218 r = dm_tm_ref(s->tm, b, &ref_count);
219 if (r)
220 return r;
222 if (ref_count > 1)
224 * This is a shared node, so we can just decrement it's
225 * reference counter and leave the children.
227 dm_tm_dec(s->tm, b);
229 else {
230 uint32_t flags;
231 struct frame *f = s->spine + ++s->top;
233 r = dm_tm_read_lock(s->tm, b, &btree_node_validator, &f->b);
234 if (r) {
235 s->top--;
236 return r;
239 f->n = dm_block_data(f->b);
240 f->level = level;
241 f->nr_children = le32_to_cpu(f->n->header.nr_entries);
242 f->current_child = 0;
244 flags = le32_to_cpu(f->n->header.flags);
245 if (flags & INTERNAL_NODE || is_internal_level(s->info, f))
246 prefetch_children(s, f);
249 return 0;
252 static void pop_frame(struct del_stack *s)
254 struct frame *f = s->spine + s->top--;
256 dm_tm_dec(s->tm, dm_block_location(f->b));
257 dm_tm_unlock(s->tm, f->b);
260 static void unlock_all_frames(struct del_stack *s)
262 struct frame *f;
264 while (unprocessed_frames(s)) {
265 f = s->spine + s->top--;
266 dm_tm_unlock(s->tm, f->b);
270 int dm_btree_del(struct dm_btree_info *info, dm_block_t root)
272 int r;
273 struct del_stack *s;
276 * dm_btree_del() is called via an ioctl, as such should be
277 * considered an FS op. We can't recurse back into the FS, so we
278 * allocate GFP_NOFS.
280 s = kmalloc(sizeof(*s), GFP_NOFS);
281 if (!s)
282 return -ENOMEM;
283 s->info = info;
284 s->tm = info->tm;
285 s->top = -1;
287 r = push_frame(s, root, 0);
288 if (r)
289 goto out;
291 while (unprocessed_frames(s)) {
292 uint32_t flags;
293 struct frame *f;
294 dm_block_t b;
296 r = top_frame(s, &f);
297 if (r)
298 goto out;
300 if (f->current_child >= f->nr_children) {
301 pop_frame(s);
302 continue;
305 flags = le32_to_cpu(f->n->header.flags);
306 if (flags & INTERNAL_NODE) {
307 b = value64(f->n, f->current_child);
308 f->current_child++;
309 r = push_frame(s, b, f->level);
310 if (r)
311 goto out;
313 } else if (is_internal_level(info, f)) {
314 b = value64(f->n, f->current_child);
315 f->current_child++;
316 r = push_frame(s, b, f->level + 1);
317 if (r)
318 goto out;
320 } else {
321 if (info->value_type.dec) {
322 unsigned i;
324 for (i = 0; i < f->nr_children; i++)
325 info->value_type.dec(info->value_type.context,
326 value_ptr(f->n, i));
328 pop_frame(s);
331 out:
332 if (r) {
333 /* cleanup all frames of del_stack */
334 unlock_all_frames(s);
336 kfree(s);
338 return r;
340 EXPORT_SYMBOL_GPL(dm_btree_del);
342 /*----------------------------------------------------------------*/
344 static int btree_lookup_raw(struct ro_spine *s, dm_block_t block, uint64_t key,
345 int (*search_fn)(struct btree_node *, uint64_t),
346 uint64_t *result_key, void *v, size_t value_size)
348 int i, r;
349 uint32_t flags, nr_entries;
351 do {
352 r = ro_step(s, block);
353 if (r < 0)
354 return r;
356 i = search_fn(ro_node(s), key);
358 flags = le32_to_cpu(ro_node(s)->header.flags);
359 nr_entries = le32_to_cpu(ro_node(s)->header.nr_entries);
360 if (i < 0 || i >= nr_entries)
361 return -ENODATA;
363 if (flags & INTERNAL_NODE)
364 block = value64(ro_node(s), i);
366 } while (!(flags & LEAF_NODE));
368 *result_key = le64_to_cpu(ro_node(s)->keys[i]);
369 memcpy(v, value_ptr(ro_node(s), i), value_size);
371 return 0;
374 int dm_btree_lookup(struct dm_btree_info *info, dm_block_t root,
375 uint64_t *keys, void *value_le)
377 unsigned level, last_level = info->levels - 1;
378 int r = -ENODATA;
379 uint64_t rkey;
380 __le64 internal_value_le;
381 struct ro_spine spine;
383 init_ro_spine(&spine, info);
384 for (level = 0; level < info->levels; level++) {
385 size_t size;
386 void *value_p;
388 if (level == last_level) {
389 value_p = value_le;
390 size = info->value_type.size;
392 } else {
393 value_p = &internal_value_le;
394 size = sizeof(uint64_t);
397 r = btree_lookup_raw(&spine, root, keys[level],
398 lower_bound, &rkey,
399 value_p, size);
401 if (!r) {
402 if (rkey != keys[level]) {
403 exit_ro_spine(&spine);
404 return -ENODATA;
406 } else {
407 exit_ro_spine(&spine);
408 return r;
411 root = le64_to_cpu(internal_value_le);
413 exit_ro_spine(&spine);
415 return r;
417 EXPORT_SYMBOL_GPL(dm_btree_lookup);
419 static int dm_btree_lookup_next_single(struct dm_btree_info *info, dm_block_t root,
420 uint64_t key, uint64_t *rkey, void *value_le)
422 int r, i;
423 uint32_t flags, nr_entries;
424 struct dm_block *node;
425 struct btree_node *n;
427 r = bn_read_lock(info, root, &node);
428 if (r)
429 return r;
431 n = dm_block_data(node);
432 flags = le32_to_cpu(n->header.flags);
433 nr_entries = le32_to_cpu(n->header.nr_entries);
435 if (flags & INTERNAL_NODE) {
436 i = lower_bound(n, key);
437 if (i < 0) {
439 * avoid early -ENODATA return when all entries are
440 * higher than the search @key.
442 i = 0;
444 if (i >= nr_entries) {
445 r = -ENODATA;
446 goto out;
449 r = dm_btree_lookup_next_single(info, value64(n, i), key, rkey, value_le);
450 if (r == -ENODATA && i < (nr_entries - 1)) {
451 i++;
452 r = dm_btree_lookup_next_single(info, value64(n, i), key, rkey, value_le);
455 } else {
456 i = upper_bound(n, key);
457 if (i < 0 || i >= nr_entries) {
458 r = -ENODATA;
459 goto out;
462 *rkey = le64_to_cpu(n->keys[i]);
463 memcpy(value_le, value_ptr(n, i), info->value_type.size);
465 out:
466 dm_tm_unlock(info->tm, node);
467 return r;
470 int dm_btree_lookup_next(struct dm_btree_info *info, dm_block_t root,
471 uint64_t *keys, uint64_t *rkey, void *value_le)
473 unsigned level;
474 int r = -ENODATA;
475 __le64 internal_value_le;
476 struct ro_spine spine;
478 init_ro_spine(&spine, info);
479 for (level = 0; level < info->levels - 1u; level++) {
480 r = btree_lookup_raw(&spine, root, keys[level],
481 lower_bound, rkey,
482 &internal_value_le, sizeof(uint64_t));
483 if (r)
484 goto out;
486 if (*rkey != keys[level]) {
487 r = -ENODATA;
488 goto out;
491 root = le64_to_cpu(internal_value_le);
494 r = dm_btree_lookup_next_single(info, root, keys[level], rkey, value_le);
495 out:
496 exit_ro_spine(&spine);
497 return r;
500 EXPORT_SYMBOL_GPL(dm_btree_lookup_next);
503 * Splits a node by creating a sibling node and shifting half the nodes
504 * contents across. Assumes there is a parent node, and it has room for
505 * another child.
507 * Before:
508 * +--------+
509 * | Parent |
510 * +--------+
513 * +----------+
514 * | A ++++++ |
515 * +----------+
518 * After:
519 * +--------+
520 * | Parent |
521 * +--------+
522 * | |
523 * v +------+
524 * +---------+ |
525 * | A* +++ | v
526 * +---------+ +-------+
527 * | B +++ |
528 * +-------+
530 * Where A* is a shadow of A.
532 static int btree_split_sibling(struct shadow_spine *s, unsigned parent_index,
533 uint64_t key)
535 int r;
536 size_t size;
537 unsigned nr_left, nr_right;
538 struct dm_block *left, *right, *parent;
539 struct btree_node *ln, *rn, *pn;
540 __le64 location;
542 left = shadow_current(s);
544 r = new_block(s->info, &right);
545 if (r < 0)
546 return r;
548 ln = dm_block_data(left);
549 rn = dm_block_data(right);
551 nr_left = le32_to_cpu(ln->header.nr_entries) / 2;
552 nr_right = le32_to_cpu(ln->header.nr_entries) - nr_left;
554 ln->header.nr_entries = cpu_to_le32(nr_left);
556 rn->header.flags = ln->header.flags;
557 rn->header.nr_entries = cpu_to_le32(nr_right);
558 rn->header.max_entries = ln->header.max_entries;
559 rn->header.value_size = ln->header.value_size;
560 memcpy(rn->keys, ln->keys + nr_left, nr_right * sizeof(rn->keys[0]));
562 size = le32_to_cpu(ln->header.flags) & INTERNAL_NODE ?
563 sizeof(uint64_t) : s->info->value_type.size;
564 memcpy(value_ptr(rn, 0), value_ptr(ln, nr_left),
565 size * nr_right);
568 * Patch up the parent
570 parent = shadow_parent(s);
572 pn = dm_block_data(parent);
573 location = cpu_to_le64(dm_block_location(left));
574 __dm_bless_for_disk(&location);
575 memcpy_disk(value_ptr(pn, parent_index),
576 &location, sizeof(__le64));
578 location = cpu_to_le64(dm_block_location(right));
579 __dm_bless_for_disk(&location);
581 r = insert_at(sizeof(__le64), pn, parent_index + 1,
582 le64_to_cpu(rn->keys[0]), &location);
583 if (r) {
584 unlock_block(s->info, right);
585 return r;
588 if (key < le64_to_cpu(rn->keys[0])) {
589 unlock_block(s->info, right);
590 s->nodes[1] = left;
591 } else {
592 unlock_block(s->info, left);
593 s->nodes[1] = right;
596 return 0;
600 * Splits a node by creating two new children beneath the given node.
602 * Before:
603 * +----------+
604 * | A ++++++ |
605 * +----------+
608 * After:
609 * +------------+
610 * | A (shadow) |
611 * +------------+
612 * | |
613 * +------+ +----+
614 * | |
615 * v v
616 * +-------+ +-------+
617 * | B +++ | | C +++ |
618 * +-------+ +-------+
620 static int btree_split_beneath(struct shadow_spine *s, uint64_t key)
622 int r;
623 size_t size;
624 unsigned nr_left, nr_right;
625 struct dm_block *left, *right, *new_parent;
626 struct btree_node *pn, *ln, *rn;
627 __le64 val;
629 new_parent = shadow_current(s);
631 pn = dm_block_data(new_parent);
632 size = le32_to_cpu(pn->header.flags) & INTERNAL_NODE ?
633 sizeof(__le64) : s->info->value_type.size;
635 /* create & init the left block */
636 r = new_block(s->info, &left);
637 if (r < 0)
638 return r;
640 ln = dm_block_data(left);
641 nr_left = le32_to_cpu(pn->header.nr_entries) / 2;
643 ln->header.flags = pn->header.flags;
644 ln->header.nr_entries = cpu_to_le32(nr_left);
645 ln->header.max_entries = pn->header.max_entries;
646 ln->header.value_size = pn->header.value_size;
647 memcpy(ln->keys, pn->keys, nr_left * sizeof(pn->keys[0]));
648 memcpy(value_ptr(ln, 0), value_ptr(pn, 0), nr_left * size);
650 /* create & init the right block */
651 r = new_block(s->info, &right);
652 if (r < 0) {
653 unlock_block(s->info, left);
654 return r;
657 rn = dm_block_data(right);
658 nr_right = le32_to_cpu(pn->header.nr_entries) - nr_left;
660 rn->header.flags = pn->header.flags;
661 rn->header.nr_entries = cpu_to_le32(nr_right);
662 rn->header.max_entries = pn->header.max_entries;
663 rn->header.value_size = pn->header.value_size;
664 memcpy(rn->keys, pn->keys + nr_left, nr_right * sizeof(pn->keys[0]));
665 memcpy(value_ptr(rn, 0), value_ptr(pn, nr_left),
666 nr_right * size);
668 /* new_parent should just point to l and r now */
669 pn->header.flags = cpu_to_le32(INTERNAL_NODE);
670 pn->header.nr_entries = cpu_to_le32(2);
671 pn->header.max_entries = cpu_to_le32(
672 calc_max_entries(sizeof(__le64),
673 dm_bm_block_size(
674 dm_tm_get_bm(s->info->tm))));
675 pn->header.value_size = cpu_to_le32(sizeof(__le64));
677 val = cpu_to_le64(dm_block_location(left));
678 __dm_bless_for_disk(&val);
679 pn->keys[0] = ln->keys[0];
680 memcpy_disk(value_ptr(pn, 0), &val, sizeof(__le64));
682 val = cpu_to_le64(dm_block_location(right));
683 __dm_bless_for_disk(&val);
684 pn->keys[1] = rn->keys[0];
685 memcpy_disk(value_ptr(pn, 1), &val, sizeof(__le64));
687 unlock_block(s->info, left);
688 unlock_block(s->info, right);
689 return 0;
692 static int btree_insert_raw(struct shadow_spine *s, dm_block_t root,
693 struct dm_btree_value_type *vt,
694 uint64_t key, unsigned *index)
696 int r, i = *index, top = 1;
697 struct btree_node *node;
699 for (;;) {
700 r = shadow_step(s, root, vt);
701 if (r < 0)
702 return r;
704 node = dm_block_data(shadow_current(s));
707 * We have to patch up the parent node, ugly, but I don't
708 * see a way to do this automatically as part of the spine
709 * op.
711 if (shadow_has_parent(s) && i >= 0) { /* FIXME: second clause unness. */
712 __le64 location = cpu_to_le64(dm_block_location(shadow_current(s)));
714 __dm_bless_for_disk(&location);
715 memcpy_disk(value_ptr(dm_block_data(shadow_parent(s)), i),
716 &location, sizeof(__le64));
719 node = dm_block_data(shadow_current(s));
721 if (node->header.nr_entries == node->header.max_entries) {
722 if (top)
723 r = btree_split_beneath(s, key);
724 else
725 r = btree_split_sibling(s, i, key);
727 if (r < 0)
728 return r;
731 node = dm_block_data(shadow_current(s));
733 i = lower_bound(node, key);
735 if (le32_to_cpu(node->header.flags) & LEAF_NODE)
736 break;
738 if (i < 0) {
739 /* change the bounds on the lowest key */
740 node->keys[0] = cpu_to_le64(key);
741 i = 0;
744 root = value64(node, i);
745 top = 0;
748 if (i < 0 || le64_to_cpu(node->keys[i]) != key)
749 i++;
751 *index = i;
752 return 0;
755 static bool need_insert(struct btree_node *node, uint64_t *keys,
756 unsigned level, unsigned index)
758 return ((index >= le32_to_cpu(node->header.nr_entries)) ||
759 (le64_to_cpu(node->keys[index]) != keys[level]));
762 static int insert(struct dm_btree_info *info, dm_block_t root,
763 uint64_t *keys, void *value, dm_block_t *new_root,
764 int *inserted)
765 __dm_written_to_disk(value)
767 int r;
768 unsigned level, index = -1, last_level = info->levels - 1;
769 dm_block_t block = root;
770 struct shadow_spine spine;
771 struct btree_node *n;
772 struct dm_btree_value_type le64_type;
774 init_le64_type(info->tm, &le64_type);
775 init_shadow_spine(&spine, info);
777 for (level = 0; level < (info->levels - 1); level++) {
778 r = btree_insert_raw(&spine, block, &le64_type, keys[level], &index);
779 if (r < 0)
780 goto bad;
782 n = dm_block_data(shadow_current(&spine));
784 if (need_insert(n, keys, level, index)) {
785 dm_block_t new_tree;
786 __le64 new_le;
788 r = dm_btree_empty(info, &new_tree);
789 if (r < 0)
790 goto bad;
792 new_le = cpu_to_le64(new_tree);
793 __dm_bless_for_disk(&new_le);
795 r = insert_at(sizeof(uint64_t), n, index,
796 keys[level], &new_le);
797 if (r)
798 goto bad;
801 if (level < last_level)
802 block = value64(n, index);
805 r = btree_insert_raw(&spine, block, &info->value_type,
806 keys[level], &index);
807 if (r < 0)
808 goto bad;
810 n = dm_block_data(shadow_current(&spine));
812 if (need_insert(n, keys, level, index)) {
813 if (inserted)
814 *inserted = 1;
816 r = insert_at(info->value_type.size, n, index,
817 keys[level], value);
818 if (r)
819 goto bad_unblessed;
820 } else {
821 if (inserted)
822 *inserted = 0;
824 if (info->value_type.dec &&
825 (!info->value_type.equal ||
826 !info->value_type.equal(
827 info->value_type.context,
828 value_ptr(n, index),
829 value))) {
830 info->value_type.dec(info->value_type.context,
831 value_ptr(n, index));
833 memcpy_disk(value_ptr(n, index),
834 value, info->value_type.size);
837 *new_root = shadow_root(&spine);
838 exit_shadow_spine(&spine);
840 return 0;
842 bad:
843 __dm_unbless_for_disk(value);
844 bad_unblessed:
845 exit_shadow_spine(&spine);
846 return r;
849 int dm_btree_insert(struct dm_btree_info *info, dm_block_t root,
850 uint64_t *keys, void *value, dm_block_t *new_root)
851 __dm_written_to_disk(value)
853 return insert(info, root, keys, value, new_root, NULL);
855 EXPORT_SYMBOL_GPL(dm_btree_insert);
857 int dm_btree_insert_notify(struct dm_btree_info *info, dm_block_t root,
858 uint64_t *keys, void *value, dm_block_t *new_root,
859 int *inserted)
860 __dm_written_to_disk(value)
862 return insert(info, root, keys, value, new_root, inserted);
864 EXPORT_SYMBOL_GPL(dm_btree_insert_notify);
866 /*----------------------------------------------------------------*/
868 static int find_key(struct ro_spine *s, dm_block_t block, bool find_highest,
869 uint64_t *result_key, dm_block_t *next_block)
871 int i, r;
872 uint32_t flags;
874 do {
875 r = ro_step(s, block);
876 if (r < 0)
877 return r;
879 flags = le32_to_cpu(ro_node(s)->header.flags);
880 i = le32_to_cpu(ro_node(s)->header.nr_entries);
881 if (!i)
882 return -ENODATA;
883 else
884 i--;
886 if (find_highest)
887 *result_key = le64_to_cpu(ro_node(s)->keys[i]);
888 else
889 *result_key = le64_to_cpu(ro_node(s)->keys[0]);
891 if (next_block || flags & INTERNAL_NODE) {
892 if (find_highest)
893 block = value64(ro_node(s), i);
894 else
895 block = value64(ro_node(s), 0);
898 } while (flags & INTERNAL_NODE);
900 if (next_block)
901 *next_block = block;
902 return 0;
905 static int dm_btree_find_key(struct dm_btree_info *info, dm_block_t root,
906 bool find_highest, uint64_t *result_keys)
908 int r = 0, count = 0, level;
909 struct ro_spine spine;
911 init_ro_spine(&spine, info);
912 for (level = 0; level < info->levels; level++) {
913 r = find_key(&spine, root, find_highest, result_keys + level,
914 level == info->levels - 1 ? NULL : &root);
915 if (r == -ENODATA) {
916 r = 0;
917 break;
919 } else if (r)
920 break;
922 count++;
924 exit_ro_spine(&spine);
926 return r ? r : count;
929 int dm_btree_find_highest_key(struct dm_btree_info *info, dm_block_t root,
930 uint64_t *result_keys)
932 return dm_btree_find_key(info, root, true, result_keys);
934 EXPORT_SYMBOL_GPL(dm_btree_find_highest_key);
936 int dm_btree_find_lowest_key(struct dm_btree_info *info, dm_block_t root,
937 uint64_t *result_keys)
939 return dm_btree_find_key(info, root, false, result_keys);
941 EXPORT_SYMBOL_GPL(dm_btree_find_lowest_key);
943 /*----------------------------------------------------------------*/
946 * FIXME: We shouldn't use a recursive algorithm when we have limited stack
947 * space. Also this only works for single level trees.
949 static int walk_node(struct dm_btree_info *info, dm_block_t block,
950 int (*fn)(void *context, uint64_t *keys, void *leaf),
951 void *context)
953 int r;
954 unsigned i, nr;
955 struct dm_block *node;
956 struct btree_node *n;
957 uint64_t keys;
959 r = bn_read_lock(info, block, &node);
960 if (r)
961 return r;
963 n = dm_block_data(node);
965 nr = le32_to_cpu(n->header.nr_entries);
966 for (i = 0; i < nr; i++) {
967 if (le32_to_cpu(n->header.flags) & INTERNAL_NODE) {
968 r = walk_node(info, value64(n, i), fn, context);
969 if (r)
970 goto out;
971 } else {
972 keys = le64_to_cpu(*key_ptr(n, i));
973 r = fn(context, &keys, value_ptr(n, i));
974 if (r)
975 goto out;
979 out:
980 dm_tm_unlock(info->tm, node);
981 return r;
984 int dm_btree_walk(struct dm_btree_info *info, dm_block_t root,
985 int (*fn)(void *context, uint64_t *keys, void *leaf),
986 void *context)
988 BUG_ON(info->levels > 1);
989 return walk_node(info, root, fn, context);
991 EXPORT_SYMBOL_GPL(dm_btree_walk);
993 /*----------------------------------------------------------------*/
995 static void prefetch_values(struct dm_btree_cursor *c)
997 unsigned i, nr;
998 __le64 value_le;
999 struct cursor_node *n = c->nodes + c->depth - 1;
1000 struct btree_node *bn = dm_block_data(n->b);
1001 struct dm_block_manager *bm = dm_tm_get_bm(c->info->tm);
1003 BUG_ON(c->info->value_type.size != sizeof(value_le));
1005 nr = le32_to_cpu(bn->header.nr_entries);
1006 for (i = 0; i < nr; i++) {
1007 memcpy(&value_le, value_ptr(bn, i), sizeof(value_le));
1008 dm_bm_prefetch(bm, le64_to_cpu(value_le));
1012 static bool leaf_node(struct dm_btree_cursor *c)
1014 struct cursor_node *n = c->nodes + c->depth - 1;
1015 struct btree_node *bn = dm_block_data(n->b);
1017 return le32_to_cpu(bn->header.flags) & LEAF_NODE;
1020 static int push_node(struct dm_btree_cursor *c, dm_block_t b)
1022 int r;
1023 struct cursor_node *n = c->nodes + c->depth;
1025 if (c->depth >= DM_BTREE_CURSOR_MAX_DEPTH - 1) {
1026 DMERR("couldn't push cursor node, stack depth too high");
1027 return -EINVAL;
1030 r = bn_read_lock(c->info, b, &n->b);
1031 if (r)
1032 return r;
1034 n->index = 0;
1035 c->depth++;
1037 if (c->prefetch_leaves || !leaf_node(c))
1038 prefetch_values(c);
1040 return 0;
1043 static void pop_node(struct dm_btree_cursor *c)
1045 c->depth--;
1046 unlock_block(c->info, c->nodes[c->depth].b);
1049 static int inc_or_backtrack(struct dm_btree_cursor *c)
1051 struct cursor_node *n;
1052 struct btree_node *bn;
1054 for (;;) {
1055 if (!c->depth)
1056 return -ENODATA;
1058 n = c->nodes + c->depth - 1;
1059 bn = dm_block_data(n->b);
1061 n->index++;
1062 if (n->index < le32_to_cpu(bn->header.nr_entries))
1063 break;
1065 pop_node(c);
1068 return 0;
1071 static int find_leaf(struct dm_btree_cursor *c)
1073 int r = 0;
1074 struct cursor_node *n;
1075 struct btree_node *bn;
1076 __le64 value_le;
1078 for (;;) {
1079 n = c->nodes + c->depth - 1;
1080 bn = dm_block_data(n->b);
1082 if (le32_to_cpu(bn->header.flags) & LEAF_NODE)
1083 break;
1085 memcpy(&value_le, value_ptr(bn, n->index), sizeof(value_le));
1086 r = push_node(c, le64_to_cpu(value_le));
1087 if (r) {
1088 DMERR("push_node failed");
1089 break;
1093 if (!r && (le32_to_cpu(bn->header.nr_entries) == 0))
1094 return -ENODATA;
1096 return r;
1099 int dm_btree_cursor_begin(struct dm_btree_info *info, dm_block_t root,
1100 bool prefetch_leaves, struct dm_btree_cursor *c)
1102 int r;
1104 c->info = info;
1105 c->root = root;
1106 c->depth = 0;
1107 c->prefetch_leaves = prefetch_leaves;
1109 r = push_node(c, root);
1110 if (r)
1111 return r;
1113 return find_leaf(c);
1115 EXPORT_SYMBOL_GPL(dm_btree_cursor_begin);
1117 void dm_btree_cursor_end(struct dm_btree_cursor *c)
1119 while (c->depth)
1120 pop_node(c);
1122 EXPORT_SYMBOL_GPL(dm_btree_cursor_end);
1124 int dm_btree_cursor_next(struct dm_btree_cursor *c)
1126 int r = inc_or_backtrack(c);
1127 if (!r) {
1128 r = find_leaf(c);
1129 if (r)
1130 DMERR("find_leaf failed");
1133 return r;
1135 EXPORT_SYMBOL_GPL(dm_btree_cursor_next);
1137 int dm_btree_cursor_skip(struct dm_btree_cursor *c, uint32_t count)
1139 int r = 0;
1141 while (count-- && !r)
1142 r = dm_btree_cursor_next(c);
1144 return r;
1146 EXPORT_SYMBOL_GPL(dm_btree_cursor_skip);
1148 int dm_btree_cursor_get_value(struct dm_btree_cursor *c, uint64_t *key, void *value_le)
1150 if (c->depth) {
1151 struct cursor_node *n = c->nodes + c->depth - 1;
1152 struct btree_node *bn = dm_block_data(n->b);
1154 if (le32_to_cpu(bn->header.flags) & INTERNAL_NODE)
1155 return -EINVAL;
1157 *key = le64_to_cpu(*key_ptr(bn, n->index));
1158 memcpy(value_le, value_ptr(bn, n->index), c->info->value_type.size);
1159 return 0;
1161 } else
1162 return -ENODATA;
1164 EXPORT_SYMBOL_GPL(dm_btree_cursor_get_value);