extent_map.c

   1// SPDX-License-Identifier: GPL-2.0
   2
   3#include <linux/err.h>
   4#include <linux/slab.h>
   5#include <linux/spinlock.h>
   6#include "messages.h"
   7#include "ctree.h"
   8#include "extent_map.h"
   9#include "compression.h"
  10#include "btrfs_inode.h"
  11#include "disk-io.h"
  12
  13
  14static struct kmem_cache *extent_map_cache;
  15
  16int __init btrfs_extent_map_init(void)
  17{
  18	extent_map_cache = kmem_cache_create("btrfs_extent_map",
  19					     sizeof(struct extent_map), 0, 0, NULL);
  20	if (!extent_map_cache)
  21		return -ENOMEM;
  22	return 0;
  23}
  24
  25void __cold btrfs_extent_map_exit(void)
  26{
  27	kmem_cache_destroy(extent_map_cache);
  28}
  29
  30/*
  31 * Initialize the extent tree @tree.  Should be called for each new inode or
  32 * other user of the extent_map interface.
  33 */
  34void btrfs_extent_map_tree_init(struct extent_map_tree *tree)
  35{
  36	tree->root = RB_ROOT;
  37	INIT_LIST_HEAD(&tree->modified_extents);
  38	rwlock_init(&tree->lock);
  39}
  40
  41/*
  42 * Allocate a new extent_map structure.  The new structure is returned with a
  43 * reference count of one and needs to be freed using free_extent_map()
  44 */
  45struct extent_map *btrfs_alloc_extent_map(void)
  46{
  47	struct extent_map *em;
  48	em = kmem_cache_zalloc(extent_map_cache, GFP_NOFS);
  49	if (!em)
  50		return NULL;
  51	RB_CLEAR_NODE(&em->rb_node);
  52	refcount_set(&em->refs, 1);
  53	INIT_LIST_HEAD(&em->list);
  54	return em;
  55}
  56
  57/*
  58 * Drop the reference out on @em by one and free the structure if the reference
  59 * count hits zero.
  60 */
  61void btrfs_free_extent_map(struct extent_map *em)
  62{
  63	if (!em)
  64		return;
  65	if (refcount_dec_and_test(&em->refs)) {
  66		WARN_ON(btrfs_extent_map_in_tree(em));
  67		WARN_ON(!list_empty(&em->list));
  68		kmem_cache_free(extent_map_cache, em);
  69	}
  70}
  71
  72/* Do the math around the end of an extent, handling wrapping. */
  73static u64 range_end(u64 start, u64 len)
  74{
  75	if (start + len < start)
  76		return (u64)-1;
  77	return start + len;
  78}
  79
  80static void remove_em(struct btrfs_inode *inode, struct extent_map *em)
  81{
  82	struct btrfs_fs_info *fs_info = inode->root->fs_info;
  83
  84	rb_erase(&em->rb_node, &inode->extent_tree.root);
  85	RB_CLEAR_NODE(&em->rb_node);
  86
  87	if (!btrfs_is_testing(fs_info) && btrfs_is_fstree(btrfs_root_id(inode->root)))
  88		percpu_counter_dec(&fs_info->evictable_extent_maps);
  89}
  90
  91static int tree_insert(struct rb_root *root, struct extent_map *em)
  92{
  93	struct rb_node **p = &root->rb_node;
  94	struct rb_node *parent = NULL;
  95	struct extent_map *entry = NULL;
  96	struct rb_node *orig_parent = NULL;
  97	u64 end = range_end(em->start, em->len);
  98
  99	while (*p) {
 100		parent = *p;
 101		entry = rb_entry(parent, struct extent_map, rb_node);
 102
 103		if (em->start < entry->start)
 104			p = &(*p)->rb_left;
 105		else if (em->start >= btrfs_extent_map_end(entry))
 106			p = &(*p)->rb_right;
 107		else
 108			return -EEXIST;
 109	}
 110
 111	orig_parent = parent;
 112	while (parent && em->start >= btrfs_extent_map_end(entry)) {
 113		parent = rb_next(parent);
 114		entry = rb_entry(parent, struct extent_map, rb_node);
 115	}
 116	if (parent)
 117		if (end > entry->start && em->start < btrfs_extent_map_end(entry))
 118			return -EEXIST;
 119
 120	parent = orig_parent;
 121	entry = rb_entry(parent, struct extent_map, rb_node);
 122	while (parent && em->start < entry->start) {
 123		parent = rb_prev(parent);
 124		entry = rb_entry(parent, struct extent_map, rb_node);
 125	}
 126	if (parent)
 127		if (end > entry->start && em->start < btrfs_extent_map_end(entry))
 128			return -EEXIST;
 129
 130	rb_link_node(&em->rb_node, orig_parent, p);
 131	rb_insert_color(&em->rb_node, root);
 132	return 0;
 133}
 134
 135/*
 136 * Search through the tree for an extent_map with a given offset.  If it can't
 137 * be found, try to find some neighboring extents
 138 */
 139static struct rb_node *tree_search(struct rb_root *root, u64 offset,
 140				   struct rb_node **prev_or_next_ret)
 141{
 142	struct rb_node *n = root->rb_node;
 143	struct rb_node *prev = NULL;
 144	struct rb_node *orig_prev = NULL;
 145	struct extent_map *entry;
 146	struct extent_map *prev_entry = NULL;
 147
 148	ASSERT(prev_or_next_ret);
 149
 150	while (n) {
 151		entry = rb_entry(n, struct extent_map, rb_node);
 152		prev = n;
 153		prev_entry = entry;
 154
 155		if (offset < entry->start)
 156			n = n->rb_left;
 157		else if (offset >= btrfs_extent_map_end(entry))
 158			n = n->rb_right;
 159		else
 160			return n;
 161	}
 162
 163	orig_prev = prev;
 164	while (prev && offset >= btrfs_extent_map_end(prev_entry)) {
 165		prev = rb_next(prev);
 166		prev_entry = rb_entry(prev, struct extent_map, rb_node);
 167	}
 168
 169	/*
 170	 * Previous extent map found, return as in this case the caller does not
 171	 * care about the next one.
 172	 */
 173	if (prev) {
 174		*prev_or_next_ret = prev;
 175		return NULL;
 176	}
 177
 178	prev = orig_prev;
 179	prev_entry = rb_entry(prev, struct extent_map, rb_node);
 180	while (prev && offset < prev_entry->start) {
 181		prev = rb_prev(prev);
 182		prev_entry = rb_entry(prev, struct extent_map, rb_node);
 183	}
 184	*prev_or_next_ret = prev;
 185
 186	return NULL;
 187}
 188
 189static inline u64 extent_map_block_len(const struct extent_map *em)
 190{
 191	if (btrfs_extent_map_is_compressed(em))
 192		return em->disk_num_bytes;
 193	return em->len;
 194}
 195
 196static inline u64 extent_map_block_end(const struct extent_map *em)
 197{
 198	const u64 block_start = btrfs_extent_map_block_start(em);
 199	const u64 block_end = block_start + extent_map_block_len(em);
 200
 201	if (block_end < block_start)
 202		return (u64)-1;
 203
 204	return block_end;
 205}
 206
 207static bool can_merge_extent_map(const struct extent_map *em)
 208{
 209	if (em->flags & EXTENT_FLAG_PINNED)
 210		return false;
 211
 212	/* Don't merge compressed extents, we need to know their actual size. */
 213	if (btrfs_extent_map_is_compressed(em))
 214		return false;
 215
 216	if (em->flags & EXTENT_FLAG_LOGGING)
 217		return false;
 218
 219	/*
 220	 * We don't want to merge stuff that hasn't been written to the log yet
 221	 * since it may not reflect exactly what is on disk, and that would be
 222	 * bad.
 223	 */
 224	if (!list_empty(&em->list))
 225		return false;
 226
 227	return true;
 228}
 229
 230/* Check to see if two extent_map structs are adjacent and safe to merge. */
 231static bool mergeable_maps(const struct extent_map *prev, const struct extent_map *next)
 232{
 233	if (btrfs_extent_map_end(prev) != next->start)
 234		return false;
 235
 236	/*
 237	 * The merged flag is not an on-disk flag, it just indicates we had the
 238	 * extent maps of 2 (or more) adjacent extents merged, so factor it out.
 239	 */
 240	if ((prev->flags & ~EXTENT_FLAG_MERGED) !=
 241	    (next->flags & ~EXTENT_FLAG_MERGED))
 242		return false;
 243
 244	if (next->disk_bytenr < EXTENT_MAP_LAST_BYTE - 1)
 245		return btrfs_extent_map_block_start(next) == extent_map_block_end(prev);
 246
 247	/* HOLES and INLINE extents. */
 248	return next->disk_bytenr == prev->disk_bytenr;
 249}
 250
 251/*
 252 * Handle the on-disk data extents merge for @prev and @next.
 253 *
 254 * @prev:    left extent to merge
 255 * @next:    right extent to merge
 256 * @merged:  the extent we will not discard after the merge; updated with new values
 257 *
 258 * After this, one of the two extents is the new merged extent and the other is
 259 * removed from the tree and likely freed. Note that @merged is one of @prev/@next
 260 * so there is const/non-const aliasing occurring here.
 261 *
 262 * Only touches disk_bytenr/disk_num_bytes/offset/ram_bytes.
 263 * For now only uncompressed regular extent can be merged.
 264 */
 265static void merge_ondisk_extents(const struct extent_map *prev, const struct extent_map *next,
 266				 struct extent_map *merged)
 267{
 268	u64 new_disk_bytenr;
 269	u64 new_disk_num_bytes;
 270	u64 new_offset;
 271
 272	/* @prev and @next should not be compressed. */
 273	ASSERT(!btrfs_extent_map_is_compressed(prev));
 274	ASSERT(!btrfs_extent_map_is_compressed(next));
 275
 276	/*
 277	 * There are two different cases where @prev and @next can be merged.
 278	 *
 279	 * 1) They are referring to the same data extent:
 280	 *
 281	 * |<----- data extent A ----->|
 282	 *    |<- prev ->|<- next ->|
 283	 *
 284	 * 2) They are referring to different data extents but still adjacent:
 285	 *
 286	 * |<-- data extent A -->|<-- data extent B -->|
 287	 *            |<- prev ->|<- next ->|
 288	 *
 289	 * The calculation here always merges the data extents first, then updates
 290	 * @offset using the new data extents.
 291	 *
 292	 * For case 1), the merged data extent would be the same.
 293	 * For case 2), we just merge the two data extents into one.
 294	 */
 295	new_disk_bytenr = min(prev->disk_bytenr, next->disk_bytenr);
 296	new_disk_num_bytes = max(prev->disk_bytenr + prev->disk_num_bytes,
 297				 next->disk_bytenr + next->disk_num_bytes) -
 298			     new_disk_bytenr;
 299	new_offset = prev->disk_bytenr + prev->offset - new_disk_bytenr;
 300
 301	merged->disk_bytenr = new_disk_bytenr;
 302	merged->disk_num_bytes = new_disk_num_bytes;
 303	merged->ram_bytes = new_disk_num_bytes;
 304	merged->offset = new_offset;
 305}
 306
 307static void dump_extent_map(struct btrfs_fs_info *fs_info, const char *prefix,
 308			    struct extent_map *em)
 309{
 310	if (!IS_ENABLED(CONFIG_BTRFS_DEBUG))
 311		return;
 312	btrfs_crit(fs_info,
 313"%s, start=%llu len=%llu disk_bytenr=%llu disk_num_bytes=%llu ram_bytes=%llu offset=%llu flags=0x%x",
 314		prefix, em->start, em->len, em->disk_bytenr, em->disk_num_bytes,
 315		em->ram_bytes, em->offset, em->flags);
 316	ASSERT(0);
 317}
 318
 319/* Internal sanity checks for btrfs debug builds. */
 320static void validate_extent_map(struct btrfs_fs_info *fs_info, struct extent_map *em)
 321{
 322	if (!IS_ENABLED(CONFIG_BTRFS_DEBUG))
 323		return;
 324	if (em->disk_bytenr < EXTENT_MAP_LAST_BYTE) {
 325		if (em->disk_num_bytes == 0)
 326			dump_extent_map(fs_info, "zero disk_num_bytes", em);
 327		if (em->offset + em->len > em->ram_bytes)
 328			dump_extent_map(fs_info, "ram_bytes too small", em);
 329		if (em->offset + em->len > em->disk_num_bytes &&
 330		    !btrfs_extent_map_is_compressed(em))
 331			dump_extent_map(fs_info, "disk_num_bytes too small", em);
 332		if (!btrfs_extent_map_is_compressed(em) &&
 333		    em->ram_bytes != em->disk_num_bytes)
 334			dump_extent_map(fs_info,
 335		"ram_bytes mismatch with disk_num_bytes for non-compressed em",
 336					em);
 337	} else if (em->offset) {
 338		dump_extent_map(fs_info, "non-zero offset for hole/inline", em);
 339	}
 340}
 341
 342static void try_merge_map(struct btrfs_inode *inode, struct extent_map *em)
 343{
 344	struct btrfs_fs_info *fs_info = inode->root->fs_info;
 345	struct extent_map *merge = NULL;
 346	struct rb_node *rb;
 347
 348	/*
 349	 * We can't modify an extent map that is in the tree and that is being
 350	 * used by another task, as it can cause that other task to see it in
 351	 * inconsistent state during the merging. We always have 1 reference for
 352	 * the tree and 1 for this task (which is unpinning the extent map or
 353	 * clearing the logging flag), so anything > 2 means it's being used by
 354	 * other tasks too.
 355	 */
 356	if (refcount_read(&em->refs) > 2)
 357		return;
 358
 359	if (!can_merge_extent_map(em))
 360		return;
 361
 362	if (em->start != 0) {
 363		rb = rb_prev(&em->rb_node);
 364		merge = rb_entry_safe(rb, struct extent_map, rb_node);
 365
 366		if (rb && can_merge_extent_map(merge) && mergeable_maps(merge, em)) {
 367			em->start = merge->start;
 368			em->len += merge->len;
 369			em->generation = max(em->generation, merge->generation);
 370
 371			if (em->disk_bytenr < EXTENT_MAP_LAST_BYTE)
 372				merge_ondisk_extents(merge, em, em);
 373			em->flags |= EXTENT_FLAG_MERGED;
 374
 375			validate_extent_map(fs_info, em);
 376			remove_em(inode, merge);
 377			btrfs_free_extent_map(merge);
 378		}
 379	}
 380
 381	rb = rb_next(&em->rb_node);
 382	merge = rb_entry_safe(rb, struct extent_map, rb_node);
 383
 384	if (rb && can_merge_extent_map(merge) && mergeable_maps(em, merge)) {
 385		em->len += merge->len;
 386		if (em->disk_bytenr < EXTENT_MAP_LAST_BYTE)
 387			merge_ondisk_extents(em, merge, em);
 388		validate_extent_map(fs_info, em);
 389		em->generation = max(em->generation, merge->generation);
 390		em->flags |= EXTENT_FLAG_MERGED;
 391		remove_em(inode, merge);
 392		btrfs_free_extent_map(merge);
 393	}
 394}
 395
 396/*
 397 * Unpin an extent from the cache.
 398 *
 399 * @inode:	the inode from which we are unpinning an extent range
 400 * @start:	logical offset in the file
 401 * @len:	length of the extent
 402 * @gen:	generation that this extent has been modified in
 403 *
 404 * Called after an extent has been written to disk properly.  Set the generation
 405 * to the generation that actually added the file item to the inode so we know
 406 * we need to sync this extent when we call fsync().
 407 *
 408 * Returns: 0	     on success
 409 * 	    -ENOENT  when the extent is not found in the tree
 410 * 	    -EUCLEAN if the found extent does not match the expected start
 411 */
 412int btrfs_unpin_extent_cache(struct btrfs_inode *inode, u64 start, u64 len, u64 gen)
 413{
 414	struct btrfs_fs_info *fs_info = inode->root->fs_info;
 415	struct extent_map_tree *tree = &inode->extent_tree;
 416	int ret = 0;
 417	struct extent_map *em;
 418
 419	write_lock(&tree->lock);
 420	em = btrfs_lookup_extent_mapping(tree, start, len);
 421
 422	if (WARN_ON(!em)) {
 423		btrfs_warn(fs_info,
 424"no extent map found for inode %llu (root %lld) when unpinning extent range [%llu, %llu), generation %llu",
 425			   btrfs_ino(inode), btrfs_root_id(inode->root),
 426			   start, start + len, gen);
 427		ret = -ENOENT;
 428		goto out;
 429	}
 430
 431	if (WARN_ON(em->start != start)) {
 432		btrfs_warn(fs_info,
 433"found extent map for inode %llu (root %lld) with unexpected start offset %llu when unpinning extent range [%llu, %llu), generation %llu",
 434			   btrfs_ino(inode), btrfs_root_id(inode->root),
 435			   em->start, start, start + len, gen);
 436		ret = -EUCLEAN;
 437		goto out;
 438	}
 439
 440	em->generation = gen;
 441	em->flags &= ~EXTENT_FLAG_PINNED;
 442
 443	try_merge_map(inode, em);
 444
 445out:
 446	write_unlock(&tree->lock);
 447	btrfs_free_extent_map(em);
 448	return ret;
 449
 450}
 451
 452void btrfs_clear_em_logging(struct btrfs_inode *inode, struct extent_map *em)
 453{
 454	lockdep_assert_held_write(&inode->extent_tree.lock);
 455
 456	em->flags &= ~EXTENT_FLAG_LOGGING;
 457	if (btrfs_extent_map_in_tree(em))
 458		try_merge_map(inode, em);
 459}
 460
 461static inline void setup_extent_mapping(struct btrfs_inode *inode,
 462					struct extent_map *em,
 463					bool modified)
 464{
 465	refcount_inc(&em->refs);
 466
 467	ASSERT(list_empty(&em->list));
 468
 469	if (modified)
 470		list_add(&em->list, &inode->extent_tree.modified_extents);
 471	else
 472		try_merge_map(inode, em);
 473}
 474
 475/*
 476 * Add a new extent map to an inode's extent map tree.
 477 *
 478 * @inode:	the target inode
 479 * @em:		map to insert
 480 * @modified:	indicate whether the given @em should be added to the
 481 *	        modified list, which indicates the extent needs to be logged
 482 *
 483 * Insert @em into the @inode's extent map tree or perform a simple
 484 * forward/backward merge with existing mappings.  The extent_map struct passed
 485 * in will be inserted into the tree directly, with an additional reference
 486 * taken, or a reference dropped if the merge attempt was successful.
 487 */
 488static int add_extent_mapping(struct btrfs_inode *inode,
 489			      struct extent_map *em, bool modified)
 490{
 491	struct extent_map_tree *tree = &inode->extent_tree;
 492	struct btrfs_root *root = inode->root;
 493	struct btrfs_fs_info *fs_info = root->fs_info;
 494	int ret;
 495
 496	lockdep_assert_held_write(&tree->lock);
 497
 498	validate_extent_map(fs_info, em);
 499	ret = tree_insert(&tree->root, em);
 500	if (ret)
 501		return ret;
 502
 503	setup_extent_mapping(inode, em, modified);
 504
 505	if (!btrfs_is_testing(fs_info) && btrfs_is_fstree(btrfs_root_id(root)))
 506		percpu_counter_inc(&fs_info->evictable_extent_maps);
 507
 508	return 0;
 509}
 510
 511static struct extent_map *lookup_extent_mapping(struct extent_map_tree *tree,
 512						u64 start, u64 len, bool strict)
 513{
 514	struct extent_map *em;
 515	struct rb_node *rb_node;
 516	struct rb_node *prev_or_next = NULL;
 517	u64 end = range_end(start, len);
 518
 519	rb_node = tree_search(&tree->root, start, &prev_or_next);
 520	if (!rb_node) {
 521		if (prev_or_next)
 522			rb_node = prev_or_next;
 523		else
 524			return NULL;
 525	}
 526
 527	em = rb_entry(rb_node, struct extent_map, rb_node);
 528
 529	if (strict && !(end > em->start && start < btrfs_extent_map_end(em)))
 530		return NULL;
 531
 532	refcount_inc(&em->refs);
 533	return em;
 534}
 535
 536/*
 537 * Lookup extent_map that intersects @start + @len range.
 538 *
 539 * @tree:	tree to lookup in
 540 * @start:	byte offset to start the search
 541 * @len:	length of the lookup range
 542 *
 543 * Find and return the first extent_map struct in @tree that intersects the
 544 * [start, len] range.  There may be additional objects in the tree that
 545 * intersect, so check the object returned carefully to make sure that no
 546 * additional lookups are needed.
 547 */
 548struct extent_map *btrfs_lookup_extent_mapping(struct extent_map_tree *tree,
 549					       u64 start, u64 len)
 550{
 551	return lookup_extent_mapping(tree, start, len, true);
 552}
 553
 554/*
 555 * Find a nearby extent map intersecting @start + @len (not an exact search).
 556 *
 557 * @tree:	tree to lookup in
 558 * @start:	byte offset to start the search
 559 * @len:	length of the lookup range
 560 *
 561 * Find and return the first extent_map struct in @tree that intersects the
 562 * [start, len] range.
 563 *
 564 * If one can't be found, any nearby extent may be returned
 565 */
 566struct extent_map *btrfs_search_extent_mapping(struct extent_map_tree *tree,
 567					       u64 start, u64 len)
 568{
 569	return lookup_extent_mapping(tree, start, len, false);
 570}
 571
 572/*
 573 * Remove an extent_map from its inode's extent tree.
 574 *
 575 * @inode:	the inode the extent map belongs to
 576 * @em:		extent map being removed
 577 *
 578 * Remove @em from the extent tree of @inode.  No reference counts are dropped,
 579 * and no checks are done to see if the range is in use.
 580 */
 581void btrfs_remove_extent_mapping(struct btrfs_inode *inode, struct extent_map *em)
 582{
 583	struct extent_map_tree *tree = &inode->extent_tree;
 584
 585	lockdep_assert_held_write(&tree->lock);
 586
 587	WARN_ON(em->flags & EXTENT_FLAG_PINNED);
 588	if (!(em->flags & EXTENT_FLAG_LOGGING))
 589		list_del_init(&em->list);
 590
 591	remove_em(inode, em);
 592}
 593
 594static void replace_extent_mapping(struct btrfs_inode *inode,
 595				   struct extent_map *cur,
 596				   struct extent_map *new,
 597				   bool modified)
 598{
 599	struct btrfs_fs_info *fs_info = inode->root->fs_info;
 600	struct extent_map_tree *tree = &inode->extent_tree;
 601
 602	lockdep_assert_held_write(&tree->lock);
 603
 604	validate_extent_map(fs_info, new);
 605
 606	WARN_ON(cur->flags & EXTENT_FLAG_PINNED);
 607	ASSERT(btrfs_extent_map_in_tree(cur));
 608	if (!(cur->flags & EXTENT_FLAG_LOGGING))
 609		list_del_init(&cur->list);
 610	rb_replace_node(&cur->rb_node, &new->rb_node, &tree->root);
 611	RB_CLEAR_NODE(&cur->rb_node);
 612
 613	setup_extent_mapping(inode, new, modified);
 614}
 615
 616static struct extent_map *next_extent_map(const struct extent_map *em)
 617{
 618	struct rb_node *next;
 619
 620	next = rb_next(&em->rb_node);
 621	if (!next)
 622		return NULL;
 623	return container_of(next, struct extent_map, rb_node);
 624}
 625
 626static struct extent_map *prev_extent_map(struct extent_map *em)
 627{
 628	struct rb_node *prev;
 629
 630	prev = rb_prev(&em->rb_node);
 631	if (!prev)
 632		return NULL;
 633	return container_of(prev, struct extent_map, rb_node);
 634}
 635
 636/*
 637 * Helper for btrfs_get_extent.  Given an existing extent in the tree,
 638 * the existing extent is the nearest extent to map_start,
 639 * and an extent that you want to insert, deal with overlap and insert
 640 * the best fitted new extent into the tree.
 641 */
 642static noinline int merge_extent_mapping(struct btrfs_inode *inode,
 643					 struct extent_map *existing,
 644					 struct extent_map *em,
 645					 u64 map_start)
 646{
 647	struct extent_map *prev;
 648	struct extent_map *next;
 649	u64 start;
 650	u64 end;
 651	u64 start_diff;
 652
 653	if (map_start < em->start || map_start >= btrfs_extent_map_end(em))
 654		return -EINVAL;
 655
 656	if (existing->start > map_start) {
 657		next = existing;
 658		prev = prev_extent_map(next);
 659	} else {
 660		prev = existing;
 661		next = next_extent_map(prev);
 662	}
 663
 664	start = prev ? btrfs_extent_map_end(prev) : em->start;
 665	start = max_t(u64, start, em->start);
 666	end = next ? next->start : btrfs_extent_map_end(em);
 667	end = min_t(u64, end, btrfs_extent_map_end(em));
 668	start_diff = start - em->start;
 669	em->start = start;
 670	em->len = end - start;
 671	if (em->disk_bytenr < EXTENT_MAP_LAST_BYTE)
 672		em->offset += start_diff;
 673	return add_extent_mapping(inode, em, false);
 674}
 675
 676/*
 677 * Add extent mapping into an inode's extent map tree.
 678 *
 679 * @inode:    target inode
 680 * @em_in:    extent we are inserting
 681 * @start:    start of the logical range btrfs_get_extent() is requesting
 682 * @len:      length of the logical range btrfs_get_extent() is requesting
 683 *
 684 * Note that @em_in's range may be different from [start, start+len),
 685 * but they must be overlapped.
 686 *
 687 * Insert @em_in into the inode's extent map tree. In case there is an
 688 * overlapping range, handle the -EEXIST by either:
 689 * a) Returning the existing extent in @em_in if @start is within the
 690 *    existing em.
 691 * b) Merge the existing extent with @em_in passed in.
 692 *
 693 * Return 0 on success, otherwise -EEXIST.
 694 *
 695 */
 696int btrfs_add_extent_mapping(struct btrfs_inode *inode,
 697			     struct extent_map **em_in, u64 start, u64 len)
 698{
 699	int ret;
 700	struct extent_map *em = *em_in;
 701	struct btrfs_fs_info *fs_info = inode->root->fs_info;
 702
 703	/*
 704	 * Tree-checker should have rejected any inline extent with non-zero
 705	 * file offset. Here just do a sanity check.
 706	 */
 707	if (em->disk_bytenr == EXTENT_MAP_INLINE)
 708		ASSERT(em->start == 0);
 709
 710	ret = add_extent_mapping(inode, em, false);
 711	/* it is possible that someone inserted the extent into the tree
 712	 * while we had the lock dropped.  It is also possible that
 713	 * an overlapping map exists in the tree
 714	 */
 715	if (ret == -EEXIST) {
 716		struct extent_map *existing;
 717
 718		existing = btrfs_search_extent_mapping(&inode->extent_tree, start, len);
 719
 720		trace_btrfs_handle_em_exist(fs_info, existing, em, start, len);
 721
 722		/*
 723		 * existing will always be non-NULL, since there must be
 724		 * extent causing the -EEXIST.
 725		 */
 726		if (start >= existing->start &&
 727		    start < btrfs_extent_map_end(existing)) {
 728			btrfs_free_extent_map(em);
 729			*em_in = existing;
 730			ret = 0;
 731		} else {
 732			u64 orig_start = em->start;
 733			u64 orig_len = em->len;
 734
 735			/*
 736			 * The existing extent map is the one nearest to
 737			 * the [start, start + len) range which overlaps
 738			 */
 739			ret = merge_extent_mapping(inode, existing, em, start);
 740			if (WARN_ON(ret)) {
 741				btrfs_free_extent_map(em);
 742				*em_in = NULL;
 743				btrfs_warn(fs_info,
 744"extent map merge error existing [%llu, %llu) with em [%llu, %llu) start %llu",
 745					   existing->start, btrfs_extent_map_end(existing),
 746					   orig_start, orig_start + orig_len, start);
 747			}
 748			btrfs_free_extent_map(existing);
 749		}
 750	}
 751
 752	ASSERT(ret == 0 || ret == -EEXIST);
 753	return ret;
 754}
 755
 756/*
 757 * Drop all extent maps from a tree in the fastest possible way, rescheduling
 758 * if needed. This avoids searching the tree, from the root down to the first
 759 * extent map, before each deletion.
 760 */
 761static void drop_all_extent_maps_fast(struct btrfs_inode *inode)
 762{
 763	struct extent_map_tree *tree = &inode->extent_tree;
 764	struct rb_node *node;
 765
 766	write_lock(&tree->lock);
 767	node = rb_first(&tree->root);
 768	while (node) {
 769		struct extent_map *em;
 770		struct rb_node *next = rb_next(node);
 771
 772		em = rb_entry(node, struct extent_map, rb_node);
 773		em->flags &= ~(EXTENT_FLAG_PINNED | EXTENT_FLAG_LOGGING);
 774		btrfs_remove_extent_mapping(inode, em);
 775		btrfs_free_extent_map(em);
 776
 777		if (cond_resched_rwlock_write(&tree->lock))
 778			node = rb_first(&tree->root);
 779		else
 780			node = next;
 781	}
 782	write_unlock(&tree->lock);
 783}
 784
 785/*
 786 * Drop all extent maps in a given range.
 787 *
 788 * @inode:       The target inode.
 789 * @start:       Start offset of the range.
 790 * @end:         End offset of the range (inclusive value).
 791 * @skip_pinned: Indicate if pinned extent maps should be ignored or not.
 792 *
 793 * This drops all the extent maps that intersect the given range [@start, @end].
 794 * Extent maps that partially overlap the range and extend behind or beyond it,
 795 * are split.
 796 * The caller should have locked an appropriate file range in the inode's io
 797 * tree before calling this function.
 798 */
 799void btrfs_drop_extent_map_range(struct btrfs_inode *inode, u64 start, u64 end,
 800				 bool skip_pinned)
 801{
 802	struct extent_map *split;
 803	struct extent_map *split2;
 804	struct extent_map *em;
 805	struct extent_map_tree *em_tree = &inode->extent_tree;
 806	u64 len = end - start + 1;
 807
 808	WARN_ON(end < start);
 809	if (end == (u64)-1) {
 810		if (start == 0 && !skip_pinned) {
 811			drop_all_extent_maps_fast(inode);
 812			return;
 813		}
 814		len = (u64)-1;
 815	} else {
 816		/* Make end offset exclusive for use in the loop below. */
 817		end++;
 818	}
 819
 820	/*
 821	 * It's ok if we fail to allocate the extent maps, see the comment near
 822	 * the bottom of the loop below. We only need two spare extent maps in
 823	 * the worst case, where the first extent map that intersects our range
 824	 * starts before the range and the last extent map that intersects our
 825	 * range ends after our range (and they might be the same extent map),
 826	 * because we need to split those two extent maps at the boundaries.
 827	 */
 828	split = btrfs_alloc_extent_map();
 829	split2 = btrfs_alloc_extent_map();
 830
 831	write_lock(&em_tree->lock);
 832	em = btrfs_lookup_extent_mapping(em_tree, start, len);
 833
 834	while (em) {
 835		/* extent_map_end() returns exclusive value (last byte + 1). */
 836		const u64 em_end = btrfs_extent_map_end(em);
 837		struct extent_map *next_em = NULL;
 838		u64 gen;
 839		unsigned long flags;
 840		bool modified;
 841
 842		if (em_end < end) {
 843			next_em = next_extent_map(em);
 844			if (next_em) {
 845				if (next_em->start < end)
 846					refcount_inc(&next_em->refs);
 847				else
 848					next_em = NULL;
 849			}
 850		}
 851
 852		if (skip_pinned && (em->flags & EXTENT_FLAG_PINNED)) {
 853			start = em_end;
 854			goto next;
 855		}
 856
 857		flags = em->flags;
 858		/*
 859		 * In case we split the extent map, we want to preserve the
 860		 * EXTENT_FLAG_LOGGING flag on our extent map, but we don't want
 861		 * it on the new extent maps.
 862		 */
 863		em->flags &= ~(EXTENT_FLAG_PINNED | EXTENT_FLAG_LOGGING);
 864		modified = !list_empty(&em->list);
 865
 866		/*
 867		 * The extent map does not cross our target range, so no need to
 868		 * split it, we can remove it directly.
 869		 */
 870		if (em->start >= start && em_end <= end)
 871			goto remove_em;
 872
 873		gen = em->generation;
 874
 875		if (em->start < start) {
 876			if (!split) {
 877				split = split2;
 878				split2 = NULL;
 879				if (!split)
 880					goto remove_em;
 881			}
 882			split->start = em->start;
 883			split->len = start - em->start;
 884
 885			if (em->disk_bytenr < EXTENT_MAP_LAST_BYTE) {
 886				split->disk_bytenr = em->disk_bytenr;
 887				split->disk_num_bytes = em->disk_num_bytes;
 888				split->offset = em->offset;
 889				split->ram_bytes = em->ram_bytes;
 890			} else {
 891				split->disk_bytenr = em->disk_bytenr;
 892				split->disk_num_bytes = 0;
 893				split->offset = 0;
 894				split->ram_bytes = split->len;
 895			}
 896
 897			split->generation = gen;
 898			split->flags = flags;
 899			replace_extent_mapping(inode, em, split, modified);
 900			btrfs_free_extent_map(split);
 901			split = split2;
 902			split2 = NULL;
 903		}
 904		if (em_end > end) {
 905			if (!split) {
 906				split = split2;
 907				split2 = NULL;
 908				if (!split)
 909					goto remove_em;
 910			}
 911			split->start = end;
 912			split->len = em_end - end;
 913			split->disk_bytenr = em->disk_bytenr;
 914			split->flags = flags;
 915			split->generation = gen;
 916
 917			if (em->disk_bytenr < EXTENT_MAP_LAST_BYTE) {
 918				split->disk_num_bytes = em->disk_num_bytes;
 919				split->offset = em->offset + end - em->start;
 920				split->ram_bytes = em->ram_bytes;
 921			} else {
 922				split->disk_num_bytes = 0;
 923				split->offset = 0;
 924				split->ram_bytes = split->len;
 925			}
 926
 927			if (btrfs_extent_map_in_tree(em)) {
 928				replace_extent_mapping(inode, em, split, modified);
 929			} else {
 930				int ret;
 931
 932				ret = add_extent_mapping(inode, split, modified);
 933				/* Logic error, shouldn't happen. */
 934				ASSERT(ret == 0);
 935				if (WARN_ON(ret != 0) && modified)
 936					btrfs_set_inode_full_sync(inode);
 937			}
 938			btrfs_free_extent_map(split);
 939			split = NULL;
 940		}
 941remove_em:
 942		if (btrfs_extent_map_in_tree(em)) {
 943			/*
 944			 * If the extent map is still in the tree it means that
 945			 * either of the following is true:
 946			 *
 947			 * 1) It fits entirely in our range (doesn't end beyond
 948			 *    it or starts before it);
 949			 *
 950			 * 2) It starts before our range and/or ends after our
 951			 *    range, and we were not able to allocate the extent
 952			 *    maps for split operations, @split and @split2.
 953			 *
 954			 * If we are at case 2) then we just remove the entire
 955			 * extent map - this is fine since if anyone needs it to
 956			 * access the subranges outside our range, will just
 957			 * load it again from the subvolume tree's file extent
 958			 * item. However if the extent map was in the list of
 959			 * modified extents, then we must mark the inode for a
 960			 * full fsync, otherwise a fast fsync will miss this
 961			 * extent if it's new and needs to be logged.
 962			 */
 963			if ((em->start < start || em_end > end) && modified) {
 964				ASSERT(!split);
 965				btrfs_set_inode_full_sync(inode);
 966			}
 967			btrfs_remove_extent_mapping(inode, em);
 968		}
 969
 970		/*
 971		 * Once for the tree reference (we replaced or removed the
 972		 * extent map from the tree).
 973		 */
 974		btrfs_free_extent_map(em);
 975next:
 976		/* Once for us (for our lookup reference). */
 977		btrfs_free_extent_map(em);
 978
 979		em = next_em;
 980	}
 981
 982	write_unlock(&em_tree->lock);
 983
 984	btrfs_free_extent_map(split);
 985	btrfs_free_extent_map(split2);
 986}
 987
 988/*
 989 * Replace a range in the inode's extent map tree with a new extent map.
 990 *
 991 * @inode:      The target inode.
 992 * @new_em:     The new extent map to add to the inode's extent map tree.
 993 * @modified:   Indicate if the new extent map should be added to the list of
 994 *              modified extents (for fast fsync tracking).
 995 *
 996 * Drops all the extent maps in the inode's extent map tree that intersect the
 997 * range of the new extent map and adds the new extent map to the tree.
 998 * The caller should have locked an appropriate file range in the inode's io
 999 * tree before calling this function.
1000 */
1001int btrfs_replace_extent_map_range(struct btrfs_inode *inode,
1002				   struct extent_map *new_em,
1003				   bool modified)
1004{
1005	const u64 end = new_em->start + new_em->len - 1;
1006	struct extent_map_tree *tree = &inode->extent_tree;
1007	int ret;
1008
1009	ASSERT(!btrfs_extent_map_in_tree(new_em));
1010
1011	/*
1012	 * The caller has locked an appropriate file range in the inode's io
1013	 * tree, but getting -EEXIST when adding the new extent map can still
1014	 * happen in case there are extents that partially cover the range, and
1015	 * this is due to two tasks operating on different parts of the extent.
1016	 * See commit 18e83ac75bfe67 ("Btrfs: fix unexpected EEXIST from
1017	 * btrfs_get_extent") for an example and details.
1018	 */
1019	do {
1020		btrfs_drop_extent_map_range(inode, new_em->start, end, false);
1021		write_lock(&tree->lock);
1022		ret = add_extent_mapping(inode, new_em, modified);
1023		write_unlock(&tree->lock);
1024	} while (ret == -EEXIST);
1025
1026	return ret;
1027}
1028
1029/*
1030 * Split off the first pre bytes from the extent_map at [start, start + len],
1031 * and set the block_start for it to new_logical.
1032 *
1033 * This function is used when an ordered_extent needs to be split.
1034 */
1035int btrfs_split_extent_map(struct btrfs_inode *inode, u64 start, u64 len, u64 pre,
1036			   u64 new_logical)
1037{
1038	struct extent_map_tree *em_tree = &inode->extent_tree;
1039	struct extent_map *em;
1040	struct extent_map *split_pre = NULL;
1041	struct extent_map *split_mid = NULL;
1042	int ret = 0;
1043	unsigned long flags;
1044
1045	ASSERT(pre != 0);
1046	ASSERT(pre < len);
1047
1048	split_pre = btrfs_alloc_extent_map();
1049	if (!split_pre)
1050		return -ENOMEM;
1051	split_mid = btrfs_alloc_extent_map();
1052	if (!split_mid) {
1053		ret = -ENOMEM;
1054		goto out_free_pre;
1055	}
1056
1057	btrfs_lock_extent(&inode->io_tree, start, start + len - 1, NULL);
1058	write_lock(&em_tree->lock);
1059	em = btrfs_lookup_extent_mapping(em_tree, start, len);
1060	if (unlikely(!em)) {
1061		ret = -EIO;
1062		goto out_unlock;
1063	}
1064
1065	ASSERT(em->len == len);
1066	ASSERT(!btrfs_extent_map_is_compressed(em));
1067	ASSERT(em->disk_bytenr < EXTENT_MAP_LAST_BYTE);
1068	ASSERT(em->flags & EXTENT_FLAG_PINNED);
1069	ASSERT(!(em->flags & EXTENT_FLAG_LOGGING));
1070	ASSERT(!list_empty(&em->list));
1071
1072	flags = em->flags;
1073	em->flags &= ~EXTENT_FLAG_PINNED;
1074
1075	/* First, replace the em with a new extent_map starting from * em->start */
1076	split_pre->start = em->start;
1077	split_pre->len = pre;
1078	split_pre->disk_bytenr = new_logical;
1079	split_pre->disk_num_bytes = split_pre->len;
1080	split_pre->offset = 0;
1081	split_pre->ram_bytes = split_pre->len;
1082	split_pre->flags = flags;
1083	split_pre->generation = em->generation;
1084
1085	replace_extent_mapping(inode, em, split_pre, true);
1086
1087	/*
1088	 * Now we only have an extent_map at:
1089	 *     [em->start, em->start + pre]
1090	 */
1091
1092	/* Insert the middle extent_map. */
1093	split_mid->start = em->start + pre;
1094	split_mid->len = em->len - pre;
1095	split_mid->disk_bytenr = btrfs_extent_map_block_start(em) + pre;
1096	split_mid->disk_num_bytes = split_mid->len;
1097	split_mid->offset = 0;
1098	split_mid->ram_bytes = split_mid->len;
1099	split_mid->flags = flags;
1100	split_mid->generation = em->generation;
1101	add_extent_mapping(inode, split_mid, true);
1102
1103	/* Once for us */
1104	btrfs_free_extent_map(em);
1105	/* Once for the tree */
1106	btrfs_free_extent_map(em);
1107
1108out_unlock:
1109	write_unlock(&em_tree->lock);
1110	btrfs_unlock_extent(&inode->io_tree, start, start + len - 1, NULL);
1111	btrfs_free_extent_map(split_mid);
1112out_free_pre:
1113	btrfs_free_extent_map(split_pre);
1114	return ret;
1115}
1116
1117struct btrfs_em_shrink_ctx {
1118	long nr_to_scan;
1119	long scanned;
1120};
1121
1122static long btrfs_scan_inode(struct btrfs_inode *inode, struct btrfs_em_shrink_ctx *ctx)
1123{
1124	struct btrfs_fs_info *fs_info = inode->root->fs_info;
1125	const u64 cur_fs_gen = btrfs_get_fs_generation(fs_info);
1126	struct extent_map_tree *tree = &inode->extent_tree;
1127	long nr_dropped = 0;
1128	struct rb_node *node;
1129
1130	lockdep_assert_held_write(&tree->lock);
1131
1132	/*
1133	 * Take the mmap lock so that we serialize with the inode logging phase
1134	 * of fsync because we may need to set the full sync flag on the inode,
1135	 * in case we have to remove extent maps in the tree's list of modified
1136	 * extents. If we set the full sync flag in the inode while an fsync is
1137	 * in progress, we may risk missing new extents because before the flag
1138	 * is set, fsync decides to only wait for writeback to complete and then
1139	 * during inode logging it sees the flag set and uses the subvolume tree
1140	 * to find new extents, which may not be there yet because ordered
1141	 * extents haven't completed yet.
1142	 *
1143	 * We also do a try lock because we don't want to block for too long and
1144	 * we are holding the extent map tree's lock in write mode.
1145	 */
1146	if (!down_read_trylock(&inode->i_mmap_lock))
1147		return 0;
1148
1149	node = rb_first(&tree->root);
1150	while (node) {
1151		struct rb_node *next = rb_next(node);
1152		struct extent_map *em;
1153
1154		em = rb_entry(node, struct extent_map, rb_node);
1155		ctx->scanned++;
1156
1157		if (em->flags & EXTENT_FLAG_PINNED)
1158			goto next;
1159
1160		/*
1161		 * If the inode is in the list of modified extents (new) and its
1162		 * generation is the same (or is greater than) the current fs
1163		 * generation, it means it was not yet persisted so we have to
1164		 * set the full sync flag so that the next fsync will not miss
1165		 * it.
1166		 */
1167		if (!list_empty(&em->list) && em->generation >= cur_fs_gen)
1168			btrfs_set_inode_full_sync(inode);
1169
1170		btrfs_remove_extent_mapping(inode, em);
1171		trace_btrfs_extent_map_shrinker_remove_em(inode, em);
1172		/* Drop the reference for the tree. */
1173		btrfs_free_extent_map(em);
1174		nr_dropped++;
1175next:
1176		if (ctx->scanned >= ctx->nr_to_scan)
1177			break;
1178
1179		/*
1180		 * Stop if we need to reschedule or there's contention on the
1181		 * lock. This is to avoid slowing other tasks trying to take the
1182		 * lock.
1183		 */
1184		if (need_resched() || rwlock_needbreak(&tree->lock) ||
1185		    btrfs_fs_closing(fs_info))
1186			break;
1187		node = next;
1188	}
1189	up_read(&inode->i_mmap_lock);
1190
1191	return nr_dropped;
1192}
1193
1194static struct btrfs_inode *find_first_inode_to_shrink(struct btrfs_root *root,
1195						      u64 min_ino)
1196{
1197	struct btrfs_inode *inode;
1198	unsigned long from = min_ino;
1199
1200	xa_lock(&root->inodes);
1201	while (true) {
1202		struct extent_map_tree *tree;
1203
1204		inode = xa_find(&root->inodes, &from, ULONG_MAX, XA_PRESENT);
1205		if (!inode)
1206			break;
1207
1208		tree = &inode->extent_tree;
1209
1210		/*
1211		 * We want to be fast so if the lock is busy we don't want to
1212		 * spend time waiting for it (some task is about to do IO for
1213		 * the inode).
1214		 */
1215		if (!write_trylock(&tree->lock))
1216			goto next;
1217
1218		/*
1219		 * Skip inode if it doesn't have loaded extent maps, so we avoid
1220		 * getting a reference and doing an iput later. This includes
1221		 * cases like files that were opened for things like stat(2), or
1222		 * files with all extent maps previously released through the
1223		 * release folio callback (btrfs_release_folio()) or released in
1224		 * a previous run, or directories which never have extent maps.
1225		 */
1226		if (RB_EMPTY_ROOT(&tree->root)) {
1227			write_unlock(&tree->lock);
1228			goto next;
1229		}
1230
1231		if (igrab(&inode->vfs_inode))
1232			break;
1233
1234		write_unlock(&tree->lock);
1235next:
1236		from = btrfs_ino(inode) + 1;
1237		cond_resched_lock(&root->inodes.xa_lock);
1238	}
1239	xa_unlock(&root->inodes);
1240
1241	return inode;
1242}
1243
1244static long btrfs_scan_root(struct btrfs_root *root, struct btrfs_em_shrink_ctx *ctx)
1245{
1246	struct btrfs_fs_info *fs_info = root->fs_info;
1247	struct btrfs_inode *inode;
1248	long nr_dropped = 0;
1249	u64 min_ino = fs_info->em_shrinker_last_ino + 1;
1250
1251	inode = find_first_inode_to_shrink(root, min_ino);
1252	while (inode) {
1253		nr_dropped += btrfs_scan_inode(inode, ctx);
1254		write_unlock(&inode->extent_tree.lock);
1255
1256		min_ino = btrfs_ino(inode) + 1;
1257		fs_info->em_shrinker_last_ino = btrfs_ino(inode);
1258		iput(&inode->vfs_inode);
1259
1260		if (ctx->scanned >= ctx->nr_to_scan || btrfs_fs_closing(fs_info))
1261			break;
1262
1263		cond_resched();
1264
1265		inode = find_first_inode_to_shrink(root, min_ino);
1266	}
1267
1268	if (inode) {
1269		/*
1270		 * There are still inodes in this root or we happened to process
1271		 * the last one and reached the scan limit. In either case set
1272		 * the current root to this one, so we'll resume from the next
1273		 * inode if there is one or we will find out this was the last
1274		 * one and move to the next root.
1275		 */
1276		fs_info->em_shrinker_last_root = btrfs_root_id(root);
1277	} else {
1278		/*
1279		 * No more inodes in this root, set extent_map_shrinker_last_ino to 0 so
1280		 * that when processing the next root we start from its first inode.
1281		 */
1282		fs_info->em_shrinker_last_ino = 0;
1283		fs_info->em_shrinker_last_root = btrfs_root_id(root) + 1;
1284	}
1285
1286	return nr_dropped;
1287}
1288
1289static void btrfs_extent_map_shrinker_worker(struct work_struct *work)
1290{
1291	struct btrfs_fs_info *fs_info;
1292	struct btrfs_em_shrink_ctx ctx;
1293	u64 start_root_id;
1294	u64 next_root_id;
1295	bool cycled = false;
1296	long nr_dropped = 0;
1297
1298	fs_info = container_of(work, struct btrfs_fs_info, em_shrinker_work);
1299
1300	ctx.scanned = 0;
1301	ctx.nr_to_scan = atomic64_read(&fs_info->em_shrinker_nr_to_scan);
1302
1303	start_root_id = fs_info->em_shrinker_last_root;
1304	next_root_id = fs_info->em_shrinker_last_root;
1305
1306	if (trace_btrfs_extent_map_shrinker_scan_enter_enabled()) {
1307		s64 nr = percpu_counter_sum_positive(&fs_info->evictable_extent_maps);
1308
1309		trace_btrfs_extent_map_shrinker_scan_enter(fs_info, nr);
1310	}
1311
1312	while (ctx.scanned < ctx.nr_to_scan && !btrfs_fs_closing(fs_info)) {
1313		struct btrfs_root *root;
1314		unsigned long count;
1315
1316		cond_resched();
1317
1318		spin_lock(&fs_info->fs_roots_radix_lock);
1319		count = radix_tree_gang_lookup(&fs_info->fs_roots_radix,
1320					       (void **)&root,
1321					       (unsigned long)next_root_id, 1);
1322		if (count == 0) {
1323			spin_unlock(&fs_info->fs_roots_radix_lock);
1324			if (start_root_id > 0 && !cycled) {
1325				next_root_id = 0;
1326				fs_info->em_shrinker_last_root = 0;
1327				fs_info->em_shrinker_last_ino = 0;
1328				cycled = true;
1329				continue;
1330			}
1331			break;
1332		}
1333		next_root_id = btrfs_root_id(root) + 1;
1334		root = btrfs_grab_root(root);
1335		spin_unlock(&fs_info->fs_roots_radix_lock);
1336
1337		if (!root)
1338			continue;
1339
1340		if (btrfs_is_fstree(btrfs_root_id(root)))
1341			nr_dropped += btrfs_scan_root(root, &ctx);
1342
1343		btrfs_put_root(root);
1344	}
1345
1346	if (trace_btrfs_extent_map_shrinker_scan_exit_enabled()) {
1347		s64 nr = percpu_counter_sum_positive(&fs_info->evictable_extent_maps);
1348
1349		trace_btrfs_extent_map_shrinker_scan_exit(fs_info, nr_dropped, nr);
1350	}
1351
1352	atomic64_set(&fs_info->em_shrinker_nr_to_scan, 0);
1353}
1354
1355void btrfs_free_extent_maps(struct btrfs_fs_info *fs_info, long nr_to_scan)
1356{
1357	/*
1358	 * Do nothing if the shrinker is already running. In case of high memory
1359	 * pressure we can have a lot of tasks calling us and all passing the
1360	 * same nr_to_scan value, but in reality we may need only to free
1361	 * nr_to_scan extent maps (or less). In case we need to free more than
1362	 * that, we will be called again by the fs shrinker, so no worries about
1363	 * not doing enough work to reclaim memory from extent maps.
1364	 * We can also be repeatedly called with the same nr_to_scan value
1365	 * simply because the shrinker runs asynchronously and multiple calls
1366	 * to this function are made before the shrinker does enough progress.
1367	 *
1368	 * That's why we set the atomic counter to nr_to_scan only if its
1369	 * current value is zero, instead of incrementing the counter by
1370	 * nr_to_scan.
1371	 */
1372	if (atomic64_cmpxchg(&fs_info->em_shrinker_nr_to_scan, 0, nr_to_scan) != 0)
1373		return;
1374
1375	queue_work(system_dfl_wq, &fs_info->em_shrinker_work);
1376}
1377
1378void btrfs_init_extent_map_shrinker_work(struct btrfs_fs_info *fs_info)
1379{
1380	atomic64_set(&fs_info->em_shrinker_nr_to_scan, 0);
1381	INIT_WORK(&fs_info->em_shrinker_work, btrfs_extent_map_shrinker_worker);
1382}