Merge tag 'for-6.14-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux

Pull btrfs updates from David Sterba: "User visible changes, features: - rebuilding of the free space tree at mount time is done in more transactions, fix potential hangs when the transaction thread is blocked due to large amount of block groups - more read IO balancing strategies (experimental config), add two new ways how to select a device for read if the profiles allow that (all RAID1*), the current default selects the device by pid which is good on average but less performant for single reader workloads - select preferred device for all reads (namely for testing) - round-robin, balance reads across devices relevant for the requested IO range - add encoded write ioctl support to io_uring (read was added in 6.12), basis for writing send stream using that instead of syscalls, non-blocking mode is not yet implemented - support FS_IOC_READ_VERITY_METADATA, applications can use the metadata to do their own verification - pass inode's i_write_hint to bios, for parity with other filesystems, ioctls F_GET_RW_HINT/F_SET_RW_HINT Core: - in zoned mode: allow to directly reclaim a block group by simply resetting it, then it can be reused and another block group does not need to be allocated - super block validation now also does more comprehensive sys array validation, adding it to the points where superblock is validated (post-read, pre-write) - subpage mode fixes: - fix double accounting of blocks due to some races - improved or fixed error handling in a few cases (compression, delalloc) - raid stripe tree: - fix various cases with extent range splitting or deleting - implement hole punching to extent range - reduce number of stripe tree lookups during bio submission - more self-tests - updated self-tests (delayed refs) - error handling improvements - cleanups, refactoring - remove rest of backref caching infrastructure from relocation, not needed anymore - error message updates - remove unnecessary calls when extent buffer was marked dirty - unused parameter removal - code moved to new files Other code changes: add rb_find_add_cached() to the rb-tree API" * tag 'for-6.14-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux: (127 commits) btrfs: selftests: add a selftest for deleting two out of three extents btrfs: selftests: add test for punching a hole into 3 RAID stripe-extents btrfs: selftests: add selftest for punching holes into the RAID stripe extents btrfs: selftests: test RAID stripe-tree deletion spanning two items btrfs: selftests: don't split RAID extents in half btrfs: selftests: check for correct return value of failed lookup btrfs: don't use btrfs_set_item_key_safe on RAID stripe-extents btrfs: implement hole punching for RAID stripe extents btrfs: fix deletion of a range spanning parts two RAID stripe extents btrfs: fix tail delete of RAID stripe-extents btrfs: fix front delete range calculation for RAID stripe extents btrfs: assert RAID stripe-extent length is always greater than 0 btrfs: don't try to delete RAID stripe-extents if we don't need to btrfs: selftests: correct RAID stripe-tree feature flag setting btrfs: add io_uring interface for encoded writes btrfs: remove the unused locked_folio parameter from btrfs_cleanup_ordered_extents() btrfs: add extra error messages for delalloc range related errors btrfs: subpage: dump the involved bitmap when ASSERT() failed btrfs: subpage: fix the bitmap dump of the locked flags btrfs: do proper folio cleanup when run_delalloc_nocow() failed ...
2026-03-06 15:25:10 -08:00 · 2025-01-20 13:09:30 -08:00
parent 1851bccf60 9d0c23db26
commit 0eb4aaa230
63 changed files with 3751 additions and 1343 deletions
--- a/fs/btrfs/Makefile
+++ b/fs/btrfs/Makefile
@@ -44,4 +44,4 @@ btrfs-$(CONFIG_BTRFS_FS_RUN_SANITY_TESTS) += tests/free-space-tests.o \
 	tests/extent-buffer-tests.o tests/btrfs-tests.o \
 	tests/extent-io-tests.o tests/inode-tests.o tests/qgroup-tests.o \
 	tests/free-space-tree-tests.o tests/extent-map-tests.o \
-	tests/raid-stripe-tree-tests.o
+	tests/raid-stripe-tree-tests.o tests/delayed-refs-tests.o
--- a/fs/btrfs/async-thread.c
+++ b/fs/btrfs/async-thread.c
@@ -18,7 +18,7 @@ enum {
 };

 #define NO_THRESHOLD (-1)
-#define DFT_THRESHOLD (32)
+#define DEFAULT_THRESHOLD (32)

 struct btrfs_workqueue {
 	struct workqueue_struct *normal_wq;
@@ -94,9 +94,9 @@ struct btrfs_workqueue *btrfs_alloc_workqueue(struct btrfs_fs_info *fs_info,

 	ret->limit_active = limit_active;
 	if (thresh == 0)
-		thresh = DFT_THRESHOLD;
+		thresh = DEFAULT_THRESHOLD;
 	/* For low threshold, disabling threshold is a better choice */
-	if (thresh < DFT_THRESHOLD) {
+	if (thresh < DEFAULT_THRESHOLD) {
 		ret->current_active = limit_active;
 		ret->thresh = NO_THRESHOLD;
 	} else {
--- a/fs/btrfs/backref.c
+++ b/fs/btrfs/backref.c
@@ -250,6 +250,21 @@ static int prelim_ref_compare(const struct prelim_ref *ref1,
 	return 0;
 }

+static int prelim_ref_rb_add_cmp(const struct rb_node *new,
+				 const struct rb_node *exist)
+{
+	const struct prelim_ref *ref_new =
+		rb_entry(new, struct prelim_ref, rbnode);
+	const struct prelim_ref *ref_exist =
+		rb_entry(exist, struct prelim_ref, rbnode);
+
+	/*
+	 * prelim_ref_compare() expects the first parameter as the existing one,
+	 * different from the rb_find_add_cached() order.
+	 */
+	return prelim_ref_compare(ref_exist, ref_new);
+}
+
 static void update_share_count(struct share_check *sc, int oldcount,
 			       int newcount, const struct prelim_ref *newref)
 {
@@ -278,55 +293,39 @@ static void prelim_ref_insert(const struct btrfs_fs_info *fs_info,
 			      struct share_check *sc)
 {
 	struct rb_root_cached *root;
-	struct rb_node **p;
-	struct rb_node *parent = NULL;
-	struct prelim_ref *ref;
-	int result;
-	bool leftmost = true;
+	struct rb_node *exist;

 	root = &preftree->root;
-	p = &root->rb_root.rb_node;
+	exist = rb_find_add_cached(&newref->rbnode, root, prelim_ref_rb_add_cmp);
+	if (exist) {
+		struct prelim_ref *ref = rb_entry(exist, struct prelim_ref, rbnode);
+		/* Identical refs, merge them and free @newref */
+		struct extent_inode_elem *eie = ref->inode_list;

-	while (*p) {
-		parent = *p;
-		ref = rb_entry(parent, struct prelim_ref, rbnode);
-		result = prelim_ref_compare(ref, newref);
-		if (result < 0) {
-			p = &(*p)->rb_left;
-		} else if (result > 0) {
-			p = &(*p)->rb_right;
-			leftmost = false;
-		} else {
-			/* Identical refs, merge them and free @newref */
-			struct extent_inode_elem *eie = ref->inode_list;
+		while (eie && eie->next)
+			eie = eie->next;

-			while (eie && eie->next)
-				eie = eie->next;
-
-			if (!eie)
-				ref->inode_list = newref->inode_list;
-			else
-				eie->next = newref->inode_list;
-			trace_btrfs_prelim_ref_merge(fs_info, ref, newref,
-						     preftree->count);
-			/*
-			 * A delayed ref can have newref->count < 0.
-			 * The ref->count is updated to follow any
-			 * BTRFS_[ADD|DROP]_DELAYED_REF actions.
-			 */
-			update_share_count(sc, ref->count,
-					   ref->count + newref->count, newref);
-			ref->count += newref->count;
-			free_pref(newref);
-			return;
-		}
+		if (!eie)
+			ref->inode_list = newref->inode_list;
+		else
+			eie->next = newref->inode_list;
+		trace_btrfs_prelim_ref_merge(fs_info, ref, newref,
+							preftree->count);
+		/*
+		 * A delayed ref can have newref->count < 0.
+		 * The ref->count is updated to follow any
+		 * BTRFS_[ADD|DROP]_DELAYED_REF actions.
+		 */
+		update_share_count(sc, ref->count,
+					ref->count + newref->count, newref);
+		ref->count += newref->count;
+		free_pref(newref);
+		return;
 	}

 	update_share_count(sc, 0, newref->count, newref);
 	preftree->count++;
 	trace_btrfs_prelim_ref_insert(fs_info, newref, NULL, preftree->count);
-	rb_link_node(&newref->rbnode, parent, p);
-	rb_insert_color_cached(&newref->rbnode, root, leftmost);
 }

 /*
@@ -3022,9 +3021,6 @@ void btrfs_backref_init_cache(struct btrfs_fs_info *fs_info,
 	cache->rb_root = RB_ROOT;
 	for (i = 0; i < BTRFS_MAX_LEVEL; i++)
 		INIT_LIST_HEAD(&cache->pending[i]);
-	INIT_LIST_HEAD(&cache->changed);
-	INIT_LIST_HEAD(&cache->detached);
-	INIT_LIST_HEAD(&cache->leaves);
 	INIT_LIST_HEAD(&cache->pending_edge);
 	INIT_LIST_HEAD(&cache->useless_node);
 	cache->fs_info = fs_info;
@@ -3132,29 +3128,17 @@ void btrfs_backref_drop_node(struct btrfs_backref_cache *tree,
 void btrfs_backref_cleanup_node(struct btrfs_backref_cache *cache,
 				struct btrfs_backref_node *node)
 {
-	struct btrfs_backref_node *upper;
 	struct btrfs_backref_edge *edge;

 	if (!node)
 		return;

-	BUG_ON(!node->lowest && !node->detached);
 	while (!list_empty(&node->upper)) {
 		edge = list_entry(node->upper.next, struct btrfs_backref_edge,
 				  list[LOWER]);
-		upper = edge->node[UPPER];
 		list_del(&edge->list[LOWER]);
 		list_del(&edge->list[UPPER]);
 		btrfs_backref_free_edge(cache, edge);
-
-		/*
-		 * Add the node to leaf node list if no other child block
-		 * cached.
-		 */
-		if (list_empty(&upper->lower)) {
-			list_add_tail(&upper->lower, &cache->leaves);
-			upper->lowest = 1;
-		}
 	}

 	btrfs_backref_drop_node(cache, node);
@@ -3166,33 +3150,13 @@ void btrfs_backref_cleanup_node(struct btrfs_backref_cache *cache,
 void btrfs_backref_release_cache(struct btrfs_backref_cache *cache)
 {
 	struct btrfs_backref_node *node;
-	int i;

-	while (!list_empty(&cache->detached)) {
-		node = list_entry(cache->detached.next,
-				  struct btrfs_backref_node, list);
+	while ((node = rb_entry_safe(rb_first(&cache->rb_root),
+				     struct btrfs_backref_node, rb_node)))
 		btrfs_backref_cleanup_node(cache, node);
-	}

-	while (!list_empty(&cache->leaves)) {
-		node = list_entry(cache->leaves.next,
-				  struct btrfs_backref_node, lower);
-		btrfs_backref_cleanup_node(cache, node);
-	}
-
-	for (i = 0; i < BTRFS_MAX_LEVEL; i++) {
-		while (!list_empty(&cache->pending[i])) {
-			node = list_first_entry(&cache->pending[i],
-						struct btrfs_backref_node,
-						list);
-			btrfs_backref_cleanup_node(cache, node);
-		}
-	}
 	ASSERT(list_empty(&cache->pending_edge));
 	ASSERT(list_empty(&cache->useless_node));
-	ASSERT(list_empty(&cache->changed));
-	ASSERT(list_empty(&cache->detached));
-	ASSERT(RB_EMPTY_ROOT(&cache->rb_root));
 	ASSERT(!cache->nr_nodes);
 	ASSERT(!cache->nr_edges);
 }
@@ -3316,8 +3280,12 @@ static int handle_indirect_tree_backref(struct btrfs_trans_handle *trans,
 	root = btrfs_get_fs_root(fs_info, ref_key->offset, false);
 	if (IS_ERR(root))
 		return PTR_ERR(root);
-	if (!test_bit(BTRFS_ROOT_SHAREABLE, &root->state))
-		cur->cowonly = 1;
+
+	/* We shouldn't be using backref cache for non-shareable roots. */
+	if (unlikely(!test_bit(BTRFS_ROOT_SHAREABLE, &root->state))) {
+		btrfs_put_root(root);
+		return -EUCLEAN;
+	}

 	if (btrfs_root_level(&root->root_item) == cur->level) {
 		/* Tree root */
@@ -3403,8 +3371,15 @@ static int handle_indirect_tree_backref(struct btrfs_trans_handle *trans,
 				goto out;
 			}
 			upper->owner = btrfs_header_owner(eb);
-			if (!test_bit(BTRFS_ROOT_SHAREABLE, &root->state))
-				upper->cowonly = 1;
+
+			/* We shouldn't be using backref cache for non shareable roots. */
+			if (unlikely(!test_bit(BTRFS_ROOT_SHAREABLE, &root->state))) {
+				btrfs_put_root(root);
+				btrfs_backref_free_edge(cache, edge);
+				btrfs_backref_free_node(cache, upper);
+				ret = -EUCLEAN;
+				goto out;
+			}

 			/*
 			 * If we know the block isn't shared we can avoid
@@ -3595,15 +3570,9 @@ int btrfs_backref_finish_upper_links(struct btrfs_backref_cache *cache,

 	ASSERT(start->checked);

-	/* Insert this node to cache if it's not COW-only */
-	if (!start->cowonly) {
-		rb_node = rb_simple_insert(&cache->rb_root, start->bytenr,
-					   &start->rb_node);
-		if (rb_node)
-			btrfs_backref_panic(cache->fs_info, start->bytenr,
-					    -EEXIST);
-		list_add_tail(&start->lower, &cache->leaves);
-	}
+	rb_node = rb_simple_insert(&cache->rb_root, start->bytenr, &start->rb_node);
+	if (rb_node)
+		btrfs_backref_panic(cache->fs_info, start->bytenr, -EEXIST);

 	/*
 	 * Use breadth first search to iterate all related edges.
@@ -3642,11 +3611,6 @@ int btrfs_backref_finish_upper_links(struct btrfs_backref_cache *cache,
 		 * parents have already been linked.
 		 */
 		if (!RB_EMPTY_NODE(&upper->rb_node)) {
-			if (upper->lowest) {
-				list_del_init(&upper->lower);
-				upper->lowest = 0;
-			}
-
 			list_add_tail(&edge->list[UPPER], &upper->lower);
 			continue;
 		}
@@ -3657,23 +3621,13 @@ int btrfs_backref_finish_upper_links(struct btrfs_backref_cache *cache,
 			return -EUCLEAN;
 		}

-		/* Sanity check, COW-only node has non-COW-only parent */
-		if (start->cowonly != upper->cowonly) {
-			ASSERT(0);
+		rb_node = rb_simple_insert(&cache->rb_root, upper->bytenr,
+					   &upper->rb_node);
+		if (unlikely(rb_node)) {
+			btrfs_backref_panic(cache->fs_info, upper->bytenr, -EEXIST);
 			return -EUCLEAN;
 		}

-		/* Only cache non-COW-only (subvolume trees) tree blocks */
-		if (!upper->cowonly) {
-			rb_node = rb_simple_insert(&cache->rb_root, upper->bytenr,
-						   &upper->rb_node);
-			if (rb_node) {
-				btrfs_backref_panic(cache->fs_info,
-						upper->bytenr, -EEXIST);
-				return -EUCLEAN;
-			}
-		}
-
 		list_add_tail(&edge->list[UPPER], &upper->lower);

 		/*
--- a/fs/btrfs/backref.h
+++ b/fs/btrfs/backref.h
@@ -318,6 +318,12 @@ struct btrfs_backref_node {
 		u64 bytenr;
 	}; /* Use rb_simple_node for search/insert */

+	/*
+	 * This is a sanity check, whenever we COW a block we will update
+	 * new_bytenr with it's current location, and we will check this in
+	 * various places to validate that the cache makes sense, it shouldn't
+	 * be used for anything else.
+	 */
 	u64 new_bytenr;
 	/* Objectid of tree block owner, can be not uptodate */
 	u64 owner;
@@ -335,10 +341,6 @@ struct btrfs_backref_node {
 	struct extent_buffer *eb;
 	/* Level of the tree block */
 	unsigned int level:8;
-	/* Is the block in a non-shareable tree */
-	unsigned int cowonly:1;
-	/* 1 if no child node is in the cache */
-	unsigned int lowest:1;
 	/* Is the extent buffer locked */
 	unsigned int locked:1;
 	/* Has the block been processed */
@@ -391,12 +393,6 @@ struct btrfs_backref_cache {
 	 * level blocks may not reflect the new location
 	 */
 	struct list_head pending[BTRFS_MAX_LEVEL];
-	/* List of backref nodes with no child node */
-	struct list_head leaves;
-	/* List of blocks that have been COWed in current transaction */
-	struct list_head changed;
-	/* List of detached backref node. */
-	struct list_head detached;

 	u64 last_trans;

--- a/fs/btrfs/bio.c
+++ b/fs/btrfs/bio.c
@@ -453,6 +453,14 @@ static void btrfs_submit_dev_bio(struct btrfs_device *dev, struct bio *bio)
 		(unsigned long)dev->bdev->bd_dev, btrfs_dev_name(dev),
 		dev->devid, bio->bi_iter.bi_size);

+	/*
+	 * Track reads if tracking is enabled; ignore I/O operations before the
+	 * filesystem is fully initialized.
+	 */
+	if (dev->fs_devices->collect_fs_stats && bio_op(bio) == REQ_OP_READ && dev->fs_info)
+		percpu_counter_add(&dev->fs_info->stats_read_blocks,
+				   bio->bi_iter.bi_size >> dev->fs_info->sectorsize_bits);
+
 	if (bio->bi_opf & REQ_BTRFS_CGROUP_PUNT)
 		blkcg_punt_bio_submit(bio);
 	else
@@ -725,8 +733,7 @@ static bool btrfs_submit_chunk(struct btrfs_bio *bbio, int mirror_num)
 			bio->bi_opf |= REQ_OP_ZONE_APPEND;
 		}

-		if (is_data_bbio(bbio) && bioc &&
-		    btrfs_need_stripe_tree_update(bioc->fs_info, bioc->map_type)) {
+		if (is_data_bbio(bbio) && bioc && bioc->use_rst) {
 			/*
 			 * No locking for the list update, as we only add to
 			 * the list in the I/O submission path, and list
--- a/fs/btrfs/block-group.c
+++ b/fs/btrfs/block-group.c
@@ -173,43 +173,41 @@ void btrfs_put_block_group(struct btrfs_block_group *cache)
 	}
 }

+static int btrfs_bg_start_cmp(const struct rb_node *new,
+			      const struct rb_node *exist)
+{
+	const struct btrfs_block_group *new_bg =
+		rb_entry(new, struct btrfs_block_group, cache_node);
+	const struct btrfs_block_group *exist_bg =
+		rb_entry(exist, struct btrfs_block_group, cache_node);
+
+	if (new_bg->start < exist_bg->start)
+		return -1;
+	if (new_bg->start > exist_bg->start)
+		return 1;
+	return 0;
+}
+
 /*
 * This adds the block group to the fs_info rb tree for the block group cache
 */
 static int btrfs_add_block_group_cache(struct btrfs_fs_info *info,
 				       struct btrfs_block_group *block_group)
 {
-	struct rb_node **p;
-	struct rb_node *parent = NULL;
-	struct btrfs_block_group *cache;
-	bool leftmost = true;
+	struct rb_node *exist;
+	int ret = 0;

 	ASSERT(block_group->length != 0);

 	write_lock(&info->block_group_cache_lock);
-	p = &info->block_group_cache_tree.rb_root.rb_node;
-
-	while (*p) {
-		parent = *p;
-		cache = rb_entry(parent, struct btrfs_block_group, cache_node);
-		if (block_group->start < cache->start) {
-			p = &(*p)->rb_left;
-		} else if (block_group->start > cache->start) {
-			p = &(*p)->rb_right;
-			leftmost = false;
-		} else {
-			write_unlock(&info->block_group_cache_lock);
-			return -EEXIST;
-		}
-	}
-
-	rb_link_node(&block_group->cache_node, parent, p);
-	rb_insert_color_cached(&block_group->cache_node,
-			       &info->block_group_cache_tree, leftmost);

+	exist = rb_find_add_cached(&block_group->cache_node,
+			&info->block_group_cache_tree, btrfs_bg_start_cmp);
+	if (exist)
+		ret = -EEXIST;
 	write_unlock(&info->block_group_cache_lock);

-	return 0;
+	return ret;
 }

 /*
@@ -1223,7 +1221,7 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans,
 	block_group->space_info->total_bytes -= block_group->length;
 	block_group->space_info->bytes_readonly -=
 		(block_group->length - block_group->zone_unusable);
-	btrfs_space_info_update_bytes_zone_unusable(fs_info, block_group->space_info,
+	btrfs_space_info_update_bytes_zone_unusable(block_group->space_info,
 						    -block_group->zone_unusable);
 	block_group->space_info->disk_total -= block_group->length * factor;

@@ -1396,8 +1394,7 @@ static int inc_block_group_ro(struct btrfs_block_group *cache, int force)
 		if (btrfs_is_zoned(cache->fs_info)) {
 			/* Migrate zone_unusable bytes to readonly */
 			sinfo->bytes_readonly += cache->zone_unusable;
-			btrfs_space_info_update_bytes_zone_unusable(cache->fs_info, sinfo,
-								    -cache->zone_unusable);
+			btrfs_space_info_update_bytes_zone_unusable(sinfo, -cache->zone_unusable);
 			cache->zone_unusable = 0;
 		}
 		cache->ro++;
@@ -1645,8 +1642,7 @@ void btrfs_delete_unused_bgs(struct btrfs_fs_info *fs_info)
 		spin_lock(&space_info->lock);
 		spin_lock(&block_group->lock);

-		btrfs_space_info_update_bytes_pinned(fs_info, space_info,
-						     -block_group->pinned);
+		btrfs_space_info_update_bytes_pinned(space_info, -block_group->pinned);
 		space_info->bytes_readonly += block_group->pinned;
 		block_group->pinned = 0;

@@ -2672,7 +2668,6 @@ static int insert_dev_extent(struct btrfs_trans_handle *trans,
 	btrfs_set_dev_extent_chunk_offset(leaf, extent, chunk_offset);

 	btrfs_set_dev_extent_length(leaf, extent, num_bytes);
-	btrfs_mark_buffer_dirty(trans, leaf);
 out:
 	btrfs_free_path(path);
 	return ret;
@@ -3060,8 +3055,7 @@ void btrfs_dec_block_group_ro(struct btrfs_block_group *cache)
 				(cache->alloc_offset - cache->used - cache->pinned -
 				 cache->reserved) +
 				(cache->length - cache->zone_capacity);
-			btrfs_space_info_update_bytes_zone_unusable(cache->fs_info, sinfo,
-								    cache->zone_unusable);
+			btrfs_space_info_update_bytes_zone_unusable(sinfo, cache->zone_unusable);
 			sinfo->bytes_readonly -= cache->zone_unusable;
 		}
 		num_bytes = cache->length - cache->reserved -
@@ -3123,7 +3117,6 @@ static int update_block_group_item(struct btrfs_trans_handle *trans,
 						   cache->global_root_id);
 	btrfs_set_stack_block_group_flags(&bgi, cache->flags);
 	write_extent_buffer(leaf, &bgi, bi, sizeof(bgi));
-	btrfs_mark_buffer_dirty(trans, leaf);
 fail:
 	btrfs_release_path(path);
 	/*
@@ -3699,7 +3692,7 @@ int btrfs_update_block_group(struct btrfs_trans_handle *trans,
 		old_val -= num_bytes;
 		cache->used = old_val;
 		cache->pinned += num_bytes;
-		btrfs_space_info_update_bytes_pinned(info, space_info, num_bytes);
+		btrfs_space_info_update_bytes_pinned(space_info, num_bytes);
 		space_info->bytes_used -= num_bytes;
 		space_info->disk_used -= num_bytes * factor;
 		if (READ_ONCE(space_info->periodic_reclaim))
@@ -3781,8 +3774,7 @@ int btrfs_add_reserved_bytes(struct btrfs_block_group *cache,
 	space_info->bytes_reserved += num_bytes;
 	trace_btrfs_space_reservation(cache->fs_info, "space_info",
 				      space_info->flags, num_bytes, 1);
-	btrfs_space_info_update_bytes_may_use(cache->fs_info,
-					      space_info, -ram_bytes);
+	btrfs_space_info_update_bytes_may_use(space_info, -ram_bytes);
 	if (delalloc)
 		cache->delalloc_bytes += num_bytes;

--- a/fs/btrfs/block-rsv.c
+++ b/fs/btrfs/block-rsv.c
@@ -150,9 +150,7 @@ static u64 block_rsv_release_bytes(struct btrfs_fs_info *fs_info,
 			spin_unlock(&dest->lock);
 		}
 		if (num_bytes)
-			btrfs_space_info_free_bytes_may_use(fs_info,
-							    space_info,
-							    num_bytes);
+			btrfs_space_info_free_bytes_may_use(space_info, num_bytes);
 	}
 	if (qgroup_to_release_ret)
 		*qgroup_to_release_ret = qgroup_to_release;
@@ -383,13 +381,11 @@ void btrfs_update_global_block_rsv(struct btrfs_fs_info *fs_info)

 	if (block_rsv->reserved < block_rsv->size) {
 		num_bytes = block_rsv->size - block_rsv->reserved;
-		btrfs_space_info_update_bytes_may_use(fs_info, sinfo,
-						      num_bytes);
+		btrfs_space_info_update_bytes_may_use(sinfo, num_bytes);
 		block_rsv->reserved = block_rsv->size;
 	} else if (block_rsv->reserved > block_rsv->size) {
 		num_bytes = block_rsv->reserved - block_rsv->size;
-		btrfs_space_info_update_bytes_may_use(fs_info, sinfo,
-						      -num_bytes);
+		btrfs_space_info_update_bytes_may_use(sinfo, -num_bytes);
 		block_rsv->reserved = block_rsv->size;
 		btrfs_try_granting_tickets(fs_info, sinfo);
 	}
--- a/fs/btrfs/btrfs_inode.h
+++ b/fs/btrfs/btrfs_inode.h
@@ -526,7 +526,7 @@ bool btrfs_data_csum_ok(struct btrfs_bio *bbio, struct btrfs_device *dev,
 			u32 bio_offset, struct bio_vec *bv);
 noinline int can_nocow_extent(struct inode *inode, u64 offset, u64 *len,
 			      struct btrfs_file_extent *file_extent,
-			      bool nowait, bool strict);
+			      bool nowait);

 void btrfs_del_delalloc_inode(struct btrfs_inode *inode);
 struct inode *btrfs_lookup_dentry(struct inode *dir, struct dentry *dentry);
--- a/fs/btrfs/ctree.c
+++ b/fs/btrfs/ctree.c
@@ -37,19 +37,6 @@ static int push_node_left(struct btrfs_trans_handle *trans,
 static int balance_node_right(struct btrfs_trans_handle *trans,
 			      struct extent_buffer *dst_buf,
 			      struct extent_buffer *src_buf);
-
-static const struct btrfs_csums {
-	u16		size;
-	const char	name[10];
-	const char	driver[12];
-} btrfs_csums[] = {
-	[BTRFS_CSUM_TYPE_CRC32] = { .size = 4, .name = "crc32c" },
-	[BTRFS_CSUM_TYPE_XXHASH] = { .size = 8, .name = "xxhash64" },
-	[BTRFS_CSUM_TYPE_SHA256] = { .size = 32, .name = "sha256" },
-	[BTRFS_CSUM_TYPE_BLAKE2] = { .size = 32, .name = "blake2b",
-				     .driver = "blake2b-256" },
-};
-
 /*
 * The leaf data grows from end-to-front in the node.  this returns the address
 * of the start of the last item, which is the stop of the leaf data stack.
@@ -148,44 +135,6 @@ static inline void copy_leaf_items(const struct extent_buffer *dst,
 			      nr_items * sizeof(struct btrfs_item));
 }

-/* This exists for btrfs-progs usages. */
-u16 btrfs_csum_type_size(u16 type)
-{
-	return btrfs_csums[type].size;
-}
-
-int btrfs_super_csum_size(const struct btrfs_super_block *s)
-{
-	u16 t = btrfs_super_csum_type(s);
-	/*
-	 * csum type is validated at mount time
-	 */
-	return btrfs_csum_type_size(t);
-}
-
-const char *btrfs_super_csum_name(u16 csum_type)
-{
-	/* csum type is validated at mount time */
-	return btrfs_csums[csum_type].name;
-}
-
-/*
- * Return driver name if defined, otherwise the name that's also a valid driver
- * name
- */
-const char *btrfs_super_csum_driver(u16 csum_type)
-{
-	/* csum type is validated at mount time */
-	return btrfs_csums[csum_type].driver[0] ?
-		btrfs_csums[csum_type].driver :
-		btrfs_csums[csum_type].name;
-}
-
-size_t __attribute_const__ btrfs_get_num_csums(void)
-{
-	return ARRAY_SIZE(btrfs_csums);
-}
-
 struct btrfs_path *btrfs_alloc_path(void)
 {
 	might_sleep();
@@ -225,22 +174,6 @@ noinline void btrfs_release_path(struct btrfs_path *p)
 	}
 }

-/*
- * We want the transaction abort to print stack trace only for errors where the
- * cause could be a bug, eg. due to ENOSPC, and not for common errors that are
- * caused by external factors.
- */
-bool __cold abort_should_print_stack(int error)
-{
-	switch (error) {
-	case -EIO:
-	case -EROFS:
-	case -ENOMEM:
-		return false;
-	}
-	return true;
-}
-
 /*
 * safely gets a reference on the root node of a tree.  A lock
 * is not taken, so a concurrent writer may put a different node
@@ -3900,6 +3833,7 @@ static noinline int setup_leaf_for_split(struct btrfs_trans_handle *trans,
 	btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);

 	BUG_ON(key.type != BTRFS_EXTENT_DATA_KEY &&
+	       key.type != BTRFS_RAID_STRIPE_KEY &&
 	       key.type != BTRFS_EXTENT_CSUM_KEY);

 	if (btrfs_leaf_free_space(leaf) >= ins_len)
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -7,7 +7,6 @@
 #define BTRFS_CTREE_H

 #include "linux/cleanup.h"
-#include <linux/pagemap.h>
 #include <linux/spinlock.h>
 #include <linux/rbtree.h>
 #include <linux/mutex.h>
@@ -506,20 +505,6 @@ static inline u32 BTRFS_MAX_XATTR_SIZE(const struct btrfs_fs_info *info)
 	return BTRFS_MAX_ITEM_SIZE(info) - sizeof(struct btrfs_dir_item);
 }

-#define BTRFS_BYTES_TO_BLKS(fs_info, bytes) \
-				((bytes) >> (fs_info)->sectorsize_bits)
-
-static inline gfp_t btrfs_alloc_write_mask(struct address_space *mapping)
-{
-	return mapping_gfp_constraint(mapping, ~__GFP_FS);
-}
-
-void btrfs_error_unpin_extent_range(struct btrfs_fs_info *fs_info, u64 start, u64 end);
-int btrfs_discard_extent(struct btrfs_fs_info *fs_info, u64 bytenr,
-			 u64 num_bytes, u64 *actual_bytes);
-int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range);
-
-/* ctree.c */
 int __init btrfs_ctree_init(void);
 void __cold btrfs_ctree_exit(void);

@@ -756,18 +741,4 @@ static inline bool btrfs_is_data_reloc_root(const struct btrfs_root *root)
 	return root->root_key.objectid == BTRFS_DATA_RELOC_TREE_OBJECTID;
 }

-u16 btrfs_csum_type_size(u16 type);
-int btrfs_super_csum_size(const struct btrfs_super_block *s);
-const char *btrfs_super_csum_name(u16 csum_type);
-const char *btrfs_super_csum_driver(u16 csum_type);
-size_t __attribute_const__ btrfs_get_num_csums(void);
-
-/*
- * We use folio flag owner_2 to indicate there is an ordered extent with
- * unfinished IO.
- */
-#define folio_test_ordered(folio)	folio_test_owner_2(folio)
-#define folio_set_ordered(folio)	folio_set_owner_2(folio)
-#define folio_clear_ordered(folio)	folio_clear_owner_2(folio)
-
 #endif
--- a/fs/btrfs/delalloc-space.c
+++ b/fs/btrfs/delalloc-space.c
@@ -176,7 +176,7 @@ void btrfs_free_reserved_data_space_noquota(struct btrfs_fs_info *fs_info,
 	ASSERT(IS_ALIGNED(len, fs_info->sectorsize));

 	data_sinfo = fs_info->data_sinfo;
-	btrfs_space_info_free_bytes_may_use(fs_info, data_sinfo, len);
+	btrfs_space_info_free_bytes_may_use(data_sinfo, len);
 }

 /*
--- a/fs/btrfs/delayed-inode.c
+++ b/fs/btrfs/delayed-inode.c
@@ -366,40 +366,35 @@ static struct btrfs_delayed_item *__btrfs_lookup_delayed_item(
 	return NULL;
 }

+static int btrfs_delayed_item_cmp(const struct rb_node *new,
+				  const struct rb_node *exist)
+{
+	const struct btrfs_delayed_item *new_item =
+		rb_entry(new, struct btrfs_delayed_item, rb_node);
+	const struct btrfs_delayed_item *exist_item =
+		rb_entry(exist, struct btrfs_delayed_item, rb_node);
+
+	if (new_item->index < exist_item->index)
+		return -1;
+	if (new_item->index > exist_item->index)
+		return 1;
+	return 0;
+}
+
 static int __btrfs_add_delayed_item(struct btrfs_delayed_node *delayed_node,
 				    struct btrfs_delayed_item *ins)
 {
-	struct rb_node **p, *node;
-	struct rb_node *parent_node = NULL;
 	struct rb_root_cached *root;
-	struct btrfs_delayed_item *item;
-	bool leftmost = true;
+	struct rb_node *exist;

 	if (ins->type == BTRFS_DELAYED_INSERTION_ITEM)
 		root = &delayed_node->ins_root;
 	else
 		root = &delayed_node->del_root;

-	p = &root->rb_root.rb_node;
-	node = &ins->rb_node;
-
-	while (*p) {
-		parent_node = *p;
-		item = rb_entry(parent_node, struct btrfs_delayed_item,
-				 rb_node);
-
-		if (item->index < ins->index) {
-			p = &(*p)->rb_right;
-			leftmost = false;
-		} else if (item->index > ins->index) {
-			p = &(*p)->rb_left;
-		} else {
-			return -EEXIST;
-		}
-	}
-
-	rb_link_node(node, parent_node, p);
-	rb_insert_color_cached(node, root, leftmost);
+	exist = rb_find_add_cached(&ins->rb_node, root, btrfs_delayed_item_cmp);
+	if (exist)
+		return -EEXIST;

 	if (ins->type == BTRFS_DELAYED_INSERTION_ITEM &&
 	    ins->index >= delayed_node->index_cnt)
@@ -1038,7 +1033,6 @@ static int __btrfs_update_delayed_inode(struct btrfs_trans_handle *trans,
 				    struct btrfs_inode_item);
 	write_extent_buffer(leaf, &node->inode_item, (unsigned long)inode_item,
 			    sizeof(struct btrfs_inode_item));
-	btrfs_mark_buffer_dirty(trans, leaf);

 	if (!test_bit(BTRFS_DELAYED_NODE_DEL_IREF, &node->flags))
 		goto out;
@@ -1561,8 +1555,7 @@ release_node:
 	return ret;
 }

-static int btrfs_delete_delayed_insertion_item(struct btrfs_fs_info *fs_info,
-					       struct btrfs_delayed_node *node,
+static int btrfs_delete_delayed_insertion_item(struct btrfs_delayed_node *node,
 					       u64 index)
 {
 	struct btrfs_delayed_item *item;
@@ -1620,7 +1613,7 @@ int btrfs_delete_delayed_dir_index(struct btrfs_trans_handle *trans,
 	if (IS_ERR(node))
 		return PTR_ERR(node);

-	ret = btrfs_delete_delayed_insertion_item(trans->fs_info, node, index);
+	ret = btrfs_delete_delayed_insertion_item(node, index);
 	if (!ret)
 		goto end;

--- a/fs/btrfs/delayed-ref.c
+++ b/fs/btrfs/delayed-ref.c
@@ -93,6 +93,9 @@ void btrfs_update_delayed_refs_rsv(struct btrfs_trans_handle *trans)
 	u64 num_bytes;
 	u64 reserved_bytes;

+	if (btrfs_is_testing(fs_info))
+		return;
+
 	num_bytes = btrfs_calc_delayed_ref_bytes(fs_info, trans->delayed_ref_updates);
 	num_bytes += btrfs_calc_delayed_ref_csum_bytes(fs_info,
 						       trans->delayed_ref_csum_deletions);
@@ -254,7 +257,7 @@ int btrfs_delayed_refs_rsv_refill(struct btrfs_fs_info *fs_info,
 	spin_unlock(&block_rsv->lock);

 	if (to_free > 0)
-		btrfs_space_info_free_bytes_may_use(fs_info, space_info, to_free);
+		btrfs_space_info_free_bytes_may_use(space_info, to_free);

 	if (refilled_bytes > 0)
 		trace_btrfs_space_reservation(fs_info, "delayed_refs_rsv", 0,
@@ -265,8 +268,8 @@ int btrfs_delayed_refs_rsv_refill(struct btrfs_fs_info *fs_info,
 /*
 * compare two delayed data backrefs with same bytenr and type
 */
-static int comp_data_refs(struct btrfs_delayed_ref_node *ref1,
-			  struct btrfs_delayed_ref_node *ref2)
+static int comp_data_refs(const struct btrfs_delayed_ref_node *ref1,
+			  const struct btrfs_delayed_ref_node *ref2)
 {
 	if (ref1->data_ref.objectid < ref2->data_ref.objectid)
 		return -1;
@@ -279,8 +282,8 @@ static int comp_data_refs(struct btrfs_delayed_ref_node *ref1,
 	return 0;
 }

-static int comp_refs(struct btrfs_delayed_ref_node *ref1,
-		     struct btrfs_delayed_ref_node *ref2,
+static int comp_refs(const struct btrfs_delayed_ref_node *ref1,
+		     const struct btrfs_delayed_ref_node *ref2,
 		     bool check_seq)
 {
 	int ret = 0;
@@ -314,34 +317,25 @@ static int comp_refs(struct btrfs_delayed_ref_node *ref1,
 	return 0;
 }

+static int cmp_refs_node(const struct rb_node *new, const struct rb_node *exist)
+{
+	const struct btrfs_delayed_ref_node *new_node =
+		rb_entry(new, struct btrfs_delayed_ref_node, ref_node);
+	const struct btrfs_delayed_ref_node *exist_node =
+		rb_entry(exist, struct btrfs_delayed_ref_node, ref_node);
+
+	return comp_refs(new_node, exist_node, true);
+}
+
 static struct btrfs_delayed_ref_node* tree_insert(struct rb_root_cached *root,
 		struct btrfs_delayed_ref_node *ins)
 {
-	struct rb_node **p = &root->rb_root.rb_node;
 	struct rb_node *node = &ins->ref_node;
-	struct rb_node *parent_node = NULL;
-	struct btrfs_delayed_ref_node *entry;
-	bool leftmost = true;
+	struct rb_node *exist;

-	while (*p) {
-		int comp;
-
-		parent_node = *p;
-		entry = rb_entry(parent_node, struct btrfs_delayed_ref_node,
-				 ref_node);
-		comp = comp_refs(ins, entry, true);
-		if (comp < 0) {
-			p = &(*p)->rb_left;
-		} else if (comp > 0) {
-			p = &(*p)->rb_right;
-			leftmost = false;
-		} else {
-			return entry;
-		}
-	}
-
-	rb_link_node(node, parent_node, p);
-	rb_insert_color_cached(node, root, leftmost);
+	exist = rb_find_add_cached(node, root, cmp_refs_node);
+	if (exist)
+		return rb_entry(exist, struct btrfs_delayed_ref_node, ref_node);
 	return NULL;
 }

@@ -555,6 +549,32 @@ void btrfs_delete_ref_head(const struct btrfs_fs_info *fs_info,
 		delayed_refs->num_heads_ready--;
 }

+struct btrfs_delayed_ref_node *btrfs_select_delayed_ref(struct btrfs_delayed_ref_head *head)
+{
+	struct btrfs_delayed_ref_node *ref;
+
+	lockdep_assert_held(&head->mutex);
+	lockdep_assert_held(&head->lock);
+
+	if (RB_EMPTY_ROOT(&head->ref_tree.rb_root))
+		return NULL;
+
+	/*
+	 * Select a delayed ref of type BTRFS_ADD_DELAYED_REF first.
+	 * This is to prevent a ref count from going down to zero, which deletes
+	 * the extent item from the extent tree, when there still are references
+	 * to add, which would fail because they would not find the extent item.
+	 */
+	if (!list_empty(&head->ref_add_list))
+		return list_first_entry(&head->ref_add_list,
+					struct btrfs_delayed_ref_node, add_list);
+
+	ref = rb_entry(rb_first_cached(&head->ref_tree),
+		       struct btrfs_delayed_ref_node, ref_node);
+	ASSERT(list_empty(&ref->add_list));
+	return ref;
+}
+
 /*
 * Helper to insert the ref_node to the tail or merge with tail.
 *
@@ -1234,6 +1254,7 @@ void btrfs_destroy_delayed_refs(struct btrfs_transaction *trans)
 {
 	struct btrfs_delayed_ref_root *delayed_refs = &trans->delayed_refs;
 	struct btrfs_fs_info *fs_info = trans->fs_info;
+	bool testing = btrfs_is_testing(fs_info);

 	spin_lock(&delayed_refs->lock);
 	while (true) {
@@ -1263,7 +1284,7 @@ void btrfs_destroy_delayed_refs(struct btrfs_transaction *trans)
 		spin_unlock(&delayed_refs->lock);
 		mutex_unlock(&head->mutex);

-		if (pin_bytes) {
+		if (!testing && pin_bytes) {
 			struct btrfs_block_group *bg;

 			bg = btrfs_lookup_block_group(fs_info, head->bytenr);
@@ -1281,8 +1302,7 @@ void btrfs_destroy_delayed_refs(struct btrfs_transaction *trans)
 				spin_lock(&bg->space_info->lock);
 				spin_lock(&bg->lock);
 				bg->pinned += head->num_bytes;
-				btrfs_space_info_update_bytes_pinned(fs_info,
-								     bg->space_info,
+				btrfs_space_info_update_bytes_pinned(bg->space_info,
 								     head->num_bytes);
 				bg->reserved -= head->num_bytes;
 				bg->space_info->bytes_reserved -= head->num_bytes;
@@ -1295,12 +1315,15 @@ void btrfs_destroy_delayed_refs(struct btrfs_transaction *trans)
 			btrfs_error_unpin_extent_range(fs_info, head->bytenr,
 				head->bytenr + head->num_bytes - 1);
 		}
-		btrfs_cleanup_ref_head_accounting(fs_info, delayed_refs, head);
+		if (!testing)
+			btrfs_cleanup_ref_head_accounting(fs_info, delayed_refs, head);
 		btrfs_put_delayed_ref_head(head);
 		cond_resched();
 		spin_lock(&delayed_refs->lock);
 	}
-	btrfs_qgroup_destroy_extent_records(trans);
+
+	if (!testing)
+		btrfs_qgroup_destroy_extent_records(trans);

 	spin_unlock(&delayed_refs->lock);
 }
--- a/fs/btrfs/delayed-ref.h
+++ b/fs/btrfs/delayed-ref.h
@@ -402,6 +402,7 @@ struct btrfs_delayed_ref_head *btrfs_select_ref_head(
 		struct btrfs_delayed_ref_root *delayed_refs);
 void btrfs_unselect_ref_head(struct btrfs_delayed_ref_root *delayed_refs,
 			     struct btrfs_delayed_ref_head *head);
+struct btrfs_delayed_ref_node *btrfs_select_delayed_ref(struct btrfs_delayed_ref_head *head);

 int btrfs_check_delayed_seq(struct btrfs_fs_info *fs_info, u64 seq);

--- a/fs/btrfs/dev-replace.c
+++ b/fs/btrfs/dev-replace.c
@@ -440,9 +440,6 @@ int btrfs_run_dev_replace(struct btrfs_trans_handle *trans)
 		dev_replace->cursor_right);
 	dev_replace->item_needs_writeback = 0;
 	up_write(&dev_replace->rwsem);
-
-	btrfs_mark_buffer_dirty(trans, eb);
-
 out:
 	btrfs_free_path(path);

--- a/fs/btrfs/dir-item.c
+++ b/fs/btrfs/dir-item.c
@@ -92,7 +92,6 @@ int btrfs_insert_xattr_item(struct btrfs_trans_handle *trans,

 	write_extent_buffer(leaf, name, name_ptr, name_len);
 	write_extent_buffer(leaf, data, data_ptr, data_len);
-	btrfs_mark_buffer_dirty(trans, path->nodes[0]);

 	return ret;
 }
@@ -152,7 +151,6 @@ int btrfs_insert_dir_item(struct btrfs_trans_handle *trans,
 	name_ptr = (unsigned long)(dir_item + 1);

 	write_extent_buffer(leaf, name->name, name_ptr, name->len);
-	btrfs_mark_buffer_dirty(trans, leaf);

 second_insert:
 	/* FIXME, use some real flag for selecting the extra index */
--- a/fs/btrfs/direct-io.c
+++ b/fs/btrfs/direct-io.c
@@ -248,8 +248,7 @@ static int btrfs_get_blocks_direct_write(struct extent_map **map,
 		len = min(len, em->len - (start - em->start));
 		block_start = extent_map_block_start(em) + (start - em->start);

-		if (can_nocow_extent(inode, start, &len,
-				     &file_extent, false, false) == 1) {
+		if (can_nocow_extent(inode, start, &len, &file_extent, false) == 1) {
 			bg = btrfs_inc_nocow_writers(fs_info, block_start);
 			if (bg)
 				can_nocow = true;
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -226,7 +226,7 @@ int btrfs_read_extent_buffer(struct extent_buffer *eb,

 	while (1) {
 		clear_bit(EXTENT_BUFFER_CORRUPT, &eb->bflags);
-		ret = read_extent_buffer_pages(eb, WAIT_COMPLETE, mirror_num, check);
+		ret = read_extent_buffer_pages(eb, mirror_num, check);
 		if (!ret)
 			break;

@@ -1258,6 +1258,7 @@ void btrfs_free_fs_info(struct btrfs_fs_info *fs_info)
 {
 	struct percpu_counter *em_counter = &fs_info->evictable_extent_maps;

+	percpu_counter_destroy(&fs_info->stats_read_blocks);
 	percpu_counter_destroy(&fs_info->dirty_metadata_bytes);
 	percpu_counter_destroy(&fs_info->delalloc_bytes);
 	percpu_counter_destroy(&fs_info->ordered_bytes);
@@ -2327,6 +2328,71 @@ out:
 	return ret;
 }

+static int validate_sys_chunk_array(const struct btrfs_fs_info *fs_info,
+				    const struct btrfs_super_block *sb)
+{
+	unsigned int cur = 0; /* Offset inside the sys chunk array */
+	/*
+	 * At sb read time, fs_info is not fully initialized. Thus we have
+	 * to use super block sectorsize, which should have been validated.
+	 */
+	const u32 sectorsize = btrfs_super_sectorsize(sb);
+	u32 sys_array_size = btrfs_super_sys_array_size(sb);
+
+	if (sys_array_size > BTRFS_SYSTEM_CHUNK_ARRAY_SIZE) {
+		btrfs_err(fs_info, "system chunk array too big %u > %u",
+			  sys_array_size, BTRFS_SYSTEM_CHUNK_ARRAY_SIZE);
+		return -EUCLEAN;
+	}
+
+	while (cur < sys_array_size) {
+		struct btrfs_disk_key *disk_key;
+		struct btrfs_chunk *chunk;
+		struct btrfs_key key;
+		u64 type;
+		u16 num_stripes;
+		u32 len;
+		int ret;
+
+		disk_key = (struct btrfs_disk_key *)(sb->sys_chunk_array + cur);
+		len = sizeof(*disk_key);
+
+		if (cur + len > sys_array_size)
+			goto short_read;
+		cur += len;
+
+		btrfs_disk_key_to_cpu(&key, disk_key);
+		if (key.type != BTRFS_CHUNK_ITEM_KEY) {
+			btrfs_err(fs_info,
+			    "unexpected item type %u in sys_array at offset %u",
+				  key.type, cur);
+			return -EUCLEAN;
+		}
+		chunk = (struct btrfs_chunk *)(sb->sys_chunk_array + cur);
+		num_stripes = btrfs_stack_chunk_num_stripes(chunk);
+		if (cur + btrfs_chunk_item_size(num_stripes) > sys_array_size)
+			goto short_read;
+		type = btrfs_stack_chunk_type(chunk);
+		if (!(type & BTRFS_BLOCK_GROUP_SYSTEM)) {
+			btrfs_err(fs_info,
+			"invalid chunk type %llu in sys_array at offset %u",
+				  type, cur);
+			return -EUCLEAN;
+		}
+		ret = btrfs_check_chunk_valid(fs_info, NULL, chunk, key.offset,
+					      sectorsize);
+		if (ret < 0)
+			return ret;
+		cur += btrfs_chunk_item_size(num_stripes);
+	}
+	return 0;
+short_read:
+	btrfs_err(fs_info,
+	"super block sys chunk array short read, cur=%u sys_array_size=%u",
+		  cur, sys_array_size);
+	return -EUCLEAN;
+}
+
 /*
 * Real super block validation
 * NOTE: super csum type and incompat features will not be checked here.
@@ -2495,6 +2561,8 @@ int btrfs_validate_super(const struct btrfs_fs_info *fs_info,
 		ret = -EINVAL;
 	}

+	ret = validate_sys_chunk_array(fs_info, sb);
+
 	/*
 	 * Obvious sys_chunk_array corruptions, it must hold at least one key
 	 * and one chunk
@@ -2856,6 +2924,10 @@ static int init_mount_fs_info(struct btrfs_fs_info *fs_info, struct super_block
 	if (ret)
 		return ret;

+	ret = percpu_counter_init(&fs_info->stats_read_blocks, 0, GFP_KERNEL);
+	if (ret)
+		return ret;
+
 	fs_info->dirty_metadata_batch = PAGE_SIZE *
 					(1 + ilog2(nr_cpu_ids));

@@ -3321,6 +3393,7 @@ int __cold open_ctree(struct super_block *sb, struct btrfs_fs_devices *fs_device
 	fs_info->sectors_per_page = (PAGE_SIZE >> fs_info->sectorsize_bits);
 	fs_info->csums_per_leaf = BTRFS_MAX_ITEM_SIZE(fs_info) / fs_info->csum_size;
 	fs_info->stripesize = stripesize;
+	fs_info->fs_devices->fs_info = fs_info;

 	/*
 	 * Handle the space caching options appropriately now that we have the
--- a/fs/btrfs/disk-io.h
+++ b/fs/btrfs/disk-io.h
@@ -96,9 +96,6 @@ struct btrfs_root *btrfs_alloc_dummy_root(struct btrfs_fs_info *fs_info);
 /*
 * This function is used to grab the root, and avoid it is freed when we
 * access it. But it doesn't ensure that the tree is not dropped.
- *
- * If you want to ensure the whole tree is safe, you should use
- * 	fs_info->subvol_srcu
 */
 static inline struct btrfs_root *btrfs_grab_root(struct btrfs_root *root)
 {
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -570,7 +570,6 @@ static noinline int insert_extent_data_ref(struct btrfs_trans_handle *trans,
 			btrfs_set_extent_data_ref_count(leaf, ref, num_refs);
 		}
 	}
-	btrfs_mark_buffer_dirty(trans, leaf);
 	ret = 0;
 fail:
 	btrfs_release_path(path);
@@ -618,7 +617,6 @@ static noinline int remove_extent_data_ref(struct btrfs_trans_handle *trans,
 			btrfs_set_extent_data_ref_count(leaf, ref1, num_refs);
 		else if (key.type == BTRFS_SHARED_DATA_REF_KEY)
 			btrfs_set_shared_data_ref_count(leaf, ref2, num_refs);
-		btrfs_mark_buffer_dirty(trans, leaf);
 	}
 	return ret;
 }
@@ -1050,7 +1048,6 @@ void setup_inline_extent_backref(struct btrfs_trans_handle *trans,
 	} else {
 		btrfs_set_extent_inline_ref_offset(leaf, iref, root_objectid);
 	}
-	btrfs_mark_buffer_dirty(trans, leaf);
 }

 static int lookup_extent_backref(struct btrfs_trans_handle *trans,
@@ -1195,7 +1192,6 @@ static noinline_for_stack int update_inline_extent_backref(
 		item_size -= size;
 		btrfs_truncate_item(trans, path, item_size, 1);
 	}
-	btrfs_mark_buffer_dirty(trans, leaf);
 	return 0;
 }

@@ -1260,12 +1256,12 @@ static int btrfs_issue_discard(struct block_device *bdev, u64 start, u64 len,
 {
 	int j, ret = 0;
 	u64 bytes_left, end;
-	u64 aligned_start = ALIGN(start, 1 << SECTOR_SHIFT);
+	u64 aligned_start = ALIGN(start, SECTOR_SIZE);

 	/* Adjust the range to be aligned to 512B sectors if necessary. */
 	if (start != aligned_start) {
 		len -= aligned_start - start;
-		len = round_down(len, 1 << SECTOR_SHIFT);
+		len = round_down(len, SECTOR_SIZE);
 		start = aligned_start;
 	}

@@ -1527,7 +1523,6 @@ static int __btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
 	if (extent_op)
 		__run_delayed_extent_op(extent_op, leaf, item);

-	btrfs_mark_buffer_dirty(trans, leaf);
 	btrfs_release_path(path);

 	/* now insert the actual backref */
@@ -1711,8 +1706,6 @@ again:

 	ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
 	__run_delayed_extent_op(extent_op, leaf, ei);
-
-	btrfs_mark_buffer_dirty(trans, leaf);
 out:
 	btrfs_free_path(path);
 	return ret;
@@ -1803,30 +1796,6 @@ static int run_one_delayed_ref(struct btrfs_trans_handle *trans,
 	return ret;
 }

-static inline struct btrfs_delayed_ref_node *
-select_delayed_ref(struct btrfs_delayed_ref_head *head)
-{
-	struct btrfs_delayed_ref_node *ref;
-
-	if (RB_EMPTY_ROOT(&head->ref_tree.rb_root))
-		return NULL;
-
-	/*
-	 * Select a delayed ref of type BTRFS_ADD_DELAYED_REF first.
-	 * This is to prevent a ref count from going down to zero, which deletes
-	 * the extent item from the extent tree, when there still are references
-	 * to add, which would fail because they would not find the extent item.
-	 */
-	if (!list_empty(&head->ref_add_list))
-		return list_first_entry(&head->ref_add_list,
-				struct btrfs_delayed_ref_node, add_list);
-
-	ref = rb_entry(rb_first_cached(&head->ref_tree),
-		       struct btrfs_delayed_ref_node, ref_node);
-	ASSERT(list_empty(&ref->add_list));
-	return ref;
-}
-
 static struct btrfs_delayed_extent_op *cleanup_extent_op(
 				struct btrfs_delayed_ref_head *head)
 {
@@ -1959,7 +1928,7 @@ static int btrfs_run_delayed_refs_for_head(struct btrfs_trans_handle *trans,
 	lockdep_assert_held(&locked_ref->mutex);
 	lockdep_assert_held(&locked_ref->lock);

-	while ((ref = select_delayed_ref(locked_ref))) {
+	while ((ref = btrfs_select_delayed_ref(locked_ref))) {
 		if (ref->seq &&
 		    btrfs_check_delayed_seq(fs_info, ref->seq)) {
 			spin_unlock(&locked_ref->lock);
@@ -2230,10 +2199,11 @@ int btrfs_set_disk_extent_flags(struct btrfs_trans_handle *trans,
 	return ret;
 }

-static noinline int check_delayed_ref(struct btrfs_root *root,
+static noinline int check_delayed_ref(struct btrfs_inode *inode,
 				      struct btrfs_path *path,
-				      u64 objectid, u64 offset, u64 bytenr)
+				      u64 offset, u64 bytenr)
 {
+	struct btrfs_root *root = inode->root;
 	struct btrfs_delayed_ref_head *head;
 	struct btrfs_delayed_ref_node *ref;
 	struct btrfs_delayed_ref_root *delayed_refs;
@@ -2307,7 +2277,7 @@ static noinline int check_delayed_ref(struct btrfs_root *root,
 		 * then we have a cross reference.
 		 */
 		if (ref->ref_root != btrfs_root_id(root) ||
-		    ref_owner != objectid || ref_offset != offset) {
+		    ref_owner != btrfs_ino(inode) || ref_offset != offset) {
 			ret = 1;
 			break;
 		}
@@ -2318,11 +2288,53 @@ static noinline int check_delayed_ref(struct btrfs_root *root,
 	return ret;
 }

-static noinline int check_committed_ref(struct btrfs_root *root,
+/*
+ * Check if there are references for a data extent other than the one belonging
+ * to the given inode and offset.
+ *
+ * @inode:     The only inode we expect to find associated with the data extent.
+ * @path:      A path to use for searching the extent tree.
+ * @offset:    The only offset we expect to find associated with the data extent.
+ * @bytenr:    The logical address of the data extent.
+ *
+ * When the extent does not have any other references other than the one we
+ * expect to find, we always return a value of 0 with the path having a locked
+ * leaf that contains the extent's extent item - this is necessary to ensure
+ * we don't race with a task running delayed references, and our caller must
+ * have such a path when calling check_delayed_ref() - it must lock a delayed
+ * ref head while holding the leaf locked. In case the extent item is not found
+ * in the extent tree, we return -ENOENT with the path having the leaf (locked)
+ * where the extent item should be, in order to prevent races with another task
+ * running delayed references, so that we don't miss any reference when calling
+ * check_delayed_ref().
+ *
+ * Note: this may return false positives, and this is because we want to be
+ *       quick here as we're called in write paths (when flushing delalloc and
+ *       in the direct IO write path). For example we can have an extent with
+ *       a single reference but that reference is not inlined, or we may have
+ *       many references in the extent tree but we also have delayed references
+ *       that cancel all the reference except the one for our inode and offset,
+ *       but it would be expensive to do such checks and complex due to all
+ *       locking to avoid races between the checks and flushing delayed refs,
+ *       plus non-inline references may be located on leaves other than the one
+ *       that contains the extent item in the extent tree. The important thing
+ *       here is to not return false negatives and that the false positives are
+ *       not very common.
+ *
+ * Returns: 0 if there are no cross references and with the path having a locked
+ *          leaf from the extent tree that contains the extent's extent item.
+ *
+ *          1 if there are cross references (false positives can happen).
+ *
+ *          < 0 in case of an error. In case of -ENOENT the leaf in the extent
+ *          tree where the extent item should be located at is read locked and
+ *          accessible in the given path.
+ */
+static noinline int check_committed_ref(struct btrfs_inode *inode,
 					struct btrfs_path *path,
-					u64 objectid, u64 offset, u64 bytenr,
-					bool strict)
+					u64 offset, u64 bytenr)
 {
+	struct btrfs_root *root = inode->root;
 	struct btrfs_fs_info *fs_info = root->fs_info;
 	struct btrfs_root *extent_root = btrfs_extent_root(fs_info, bytenr);
 	struct extent_buffer *leaf;
@@ -2341,35 +2353,32 @@ static noinline int check_committed_ref(struct btrfs_root *root,

 	ret = btrfs_search_slot(NULL, extent_root, &key, path, 0, 0);
 	if (ret < 0)
-		goto out;
+		return ret;
 	if (ret == 0) {
 		/*
 		 * Key with offset -1 found, there would have to exist an extent
 		 * item with such offset, but this is out of the valid range.
 		 */
-		ret = -EUCLEAN;
-		goto out;
+		return -EUCLEAN;
 	}

-	ret = -ENOENT;
 	if (path->slots[0] == 0)
-		goto out;
+		return -ENOENT;

 	path->slots[0]--;
 	leaf = path->nodes[0];
 	btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);

 	if (key.objectid != bytenr || key.type != BTRFS_EXTENT_ITEM_KEY)
-		goto out;
+		return -ENOENT;

-	ret = 1;
 	item_size = btrfs_item_size(leaf, path->slots[0]);
 	ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
 	expected_size = sizeof(*ei) + btrfs_extent_inline_ref_size(BTRFS_EXTENT_DATA_REF_KEY);

 	/* No inline refs; we need to bail before checking for owner ref. */
 	if (item_size == sizeof(*ei))
-		goto out;
+		return 1;

 	/* Check for an owner ref; skip over it to the real inline refs. */
 	iref = (struct btrfs_extent_inline_ref *)(ei + 1);
@@ -2377,56 +2386,69 @@ static noinline int check_committed_ref(struct btrfs_root *root,
 	if (btrfs_fs_incompat(fs_info, SIMPLE_QUOTA) && type == BTRFS_EXTENT_OWNER_REF_KEY) {
 		expected_size += btrfs_extent_inline_ref_size(BTRFS_EXTENT_OWNER_REF_KEY);
 		iref = (struct btrfs_extent_inline_ref *)(iref + 1);
+		type = btrfs_get_extent_inline_ref_type(leaf, iref, BTRFS_REF_TYPE_DATA);
 	}

 	/* If extent item has more than 1 inline ref then it's shared */
 	if (item_size != expected_size)
-		goto out;
-
-	/*
-	 * If extent created before last snapshot => it's shared unless the
-	 * snapshot has been deleted. Use the heuristic if strict is false.
-	 */
-	if (!strict &&
-	    (btrfs_extent_generation(leaf, ei) <=
-	     btrfs_root_last_snapshot(&root->root_item)))
-		goto out;
+		return 1;

 	/* If this extent has SHARED_DATA_REF then it's shared */
-	type = btrfs_get_extent_inline_ref_type(leaf, iref, BTRFS_REF_TYPE_DATA);
 	if (type != BTRFS_EXTENT_DATA_REF_KEY)
-		goto out;
+		return 1;

 	ref = (struct btrfs_extent_data_ref *)(&iref->offset);
 	if (btrfs_extent_refs(leaf, ei) !=
 	    btrfs_extent_data_ref_count(leaf, ref) ||
 	    btrfs_extent_data_ref_root(leaf, ref) != btrfs_root_id(root) ||
-	    btrfs_extent_data_ref_objectid(leaf, ref) != objectid ||
+	    btrfs_extent_data_ref_objectid(leaf, ref) != btrfs_ino(inode) ||
 	    btrfs_extent_data_ref_offset(leaf, ref) != offset)
-		goto out;
+		return 1;

-	ret = 0;
-out:
-	return ret;
+	return 0;
 }

-int btrfs_cross_ref_exist(struct btrfs_root *root, u64 objectid, u64 offset,
-			  u64 bytenr, bool strict, struct btrfs_path *path)
+int btrfs_cross_ref_exist(struct btrfs_inode *inode, u64 offset,
+			  u64 bytenr, struct btrfs_path *path)
 {
 	int ret;

 	do {
-		ret = check_committed_ref(root, path, objectid,
-					  offset, bytenr, strict);
+		ret = check_committed_ref(inode, path, offset, bytenr);
 		if (ret && ret != -ENOENT)
 			goto out;

-		ret = check_delayed_ref(root, path, objectid, offset, bytenr);
+		/*
+		 * The path must have a locked leaf from the extent tree where
+		 * the extent item for our extent is located, in case it exists,
+		 * or where it should be located in case it doesn't exist yet
+		 * because it's new and its delayed ref was not yet flushed.
+		 * We need to lock the delayed ref head at check_delayed_ref(),
+		 * if one exists, while holding the leaf locked in order to not
+		 * race with delayed ref flushing, missing references and
+		 * incorrectly reporting that the extent is not shared.
+		 */
+		if (IS_ENABLED(CONFIG_BTRFS_ASSERT)) {
+			struct extent_buffer *leaf = path->nodes[0];
+
+			ASSERT(leaf != NULL);
+			btrfs_assert_tree_read_locked(leaf);
+
+			if (ret != -ENOENT) {
+				struct btrfs_key key;
+
+				btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
+				ASSERT(key.objectid == bytenr);
+				ASSERT(key.type == BTRFS_EXTENT_ITEM_KEY);
+			}
+		}
+
+		ret = check_delayed_ref(inode, path, offset, bytenr);
 	} while (ret == -EAGAIN && !path->nowait);

 out:
 	btrfs_release_path(path);
-	if (btrfs_is_data_reloc_root(root))
+	if (btrfs_is_data_reloc_root(inode->root))
 		WARN_ON(ret > 0);
 	return ret;
 }
@@ -2571,13 +2593,10 @@ static int pin_down_extent(struct btrfs_trans_handle *trans,
 			   struct btrfs_block_group *cache,
 			   u64 bytenr, u64 num_bytes, int reserved)
 {
-	struct btrfs_fs_info *fs_info = cache->fs_info;
-
 	spin_lock(&cache->space_info->lock);
 	spin_lock(&cache->lock);
 	cache->pinned += num_bytes;
-	btrfs_space_info_update_bytes_pinned(fs_info, cache->space_info,
-					     num_bytes);
+	btrfs_space_info_update_bytes_pinned(cache->space_info, num_bytes);
 	if (reserved) {
 		cache->reserved -= num_bytes;
 		cache->space_info->bytes_reserved -= num_bytes;
@@ -2724,15 +2743,15 @@ static int unpin_extent_range(struct btrfs_fs_info *fs_info,
 {
 	struct btrfs_block_group *cache = NULL;
 	struct btrfs_space_info *space_info;
-	struct btrfs_block_rsv *global_rsv = &fs_info->global_block_rsv;
 	struct btrfs_free_cluster *cluster = NULL;
-	u64 len;
 	u64 total_unpinned = 0;
 	u64 empty_cluster = 0;
 	bool readonly;
 	int ret = 0;

 	while (start <= end) {
+		u64 len;
+
 		readonly = false;
 		if (!cache ||
 		    start >= cache->start + cache->length) {
@@ -2778,37 +2797,19 @@ static int unpin_extent_range(struct btrfs_fs_info *fs_info,
 		spin_lock(&space_info->lock);
 		spin_lock(&cache->lock);
 		cache->pinned -= len;
-		btrfs_space_info_update_bytes_pinned(fs_info, space_info, -len);
+		btrfs_space_info_update_bytes_pinned(space_info, -len);
 		space_info->max_extent_size = 0;
 		if (cache->ro) {
 			space_info->bytes_readonly += len;
 			readonly = true;
 		} else if (btrfs_is_zoned(fs_info)) {
 			/* Need reset before reusing in a zoned block group */
-			btrfs_space_info_update_bytes_zone_unusable(fs_info, space_info,
-								    len);
+			btrfs_space_info_update_bytes_zone_unusable(space_info, len);
 			readonly = true;
 		}
 		spin_unlock(&cache->lock);
-		if (!readonly && return_free_space &&
-		    global_rsv->space_info == space_info) {
-			spin_lock(&global_rsv->lock);
-			if (!global_rsv->full) {
-				u64 to_add = min(len, global_rsv->size -
-						      global_rsv->reserved);
-
-				global_rsv->reserved += to_add;
-				btrfs_space_info_update_bytes_may_use(fs_info,
-						space_info, to_add);
-				if (global_rsv->reserved >= global_rsv->size)
-					global_rsv->full = 1;
-				len -= to_add;
-			}
-			spin_unlock(&global_rsv->lock);
-		}
-		/* Add to any tickets we may have */
-		if (!readonly && return_free_space && len)
-			btrfs_try_granting_tickets(fs_info, space_info);
+		if (!readonly && return_free_space)
+			btrfs_return_free_space(space_info, len);
 		spin_unlock(&space_info->lock);
 	}

@@ -3259,7 +3260,6 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 			}
 		} else {
 			btrfs_set_extent_refs(leaf, ei, refs);
-			btrfs_mark_buffer_dirty(trans, leaf);
 		}
 		if (found_extent) {
 			ret = remove_extent_backref(trans, extent_root, path,
@@ -4827,7 +4827,6 @@ static int alloc_reserved_file_extent(struct btrfs_trans_handle *trans,
 		btrfs_set_extent_data_ref_count(leaf, ref, ref_mod);
 	}

-	btrfs_mark_buffer_dirty(trans, path->nodes[0]);
 	btrfs_free_path(path);

 	return alloc_reserved_extent(trans, ins->objectid, ins->offset);
@@ -4902,7 +4901,6 @@ static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans,
 		btrfs_set_extent_inline_ref_offset(leaf, iref, node->ref_root);
 	}

-	btrfs_mark_buffer_dirty(trans, leaf);
 	btrfs_free_path(path);

 	return alloc_reserved_extent(trans, node->bytenr, fs_info->nodesize);
--- a/Show More
+++ b/Show More