Merge tag 'for-6.15-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux

Pull btrfs updates from David Sterba:
"User visible changes:

- fall back to buffered write if direct io is done on a file that
requires checksums
- this avoids a problem with checksum mismatch errors, observed
e.g. on virtual images when writes to pages under writeback
cause the checksum mismatch reports
- this may lead to some performance degradation but currently the
recommended setup for VM images is to use the NOCOW file
attribute that also disables checksums

- fast/realtime zstd levels -15 to -1
- supported by mount options (compress=zstd:-5) and defrag ioctl
- improved speed, reduced compression ratio, check the commit for
sample measurements

- defrag ioctl extended to accept negative compression levels

- subpage mode
- remove warning when subpage mode is used, the feature is now
reasonably complete and tested
- in debug mode allow to create 2K b-tree nodes to allow testing
subpage on x86_64 with 4K pages too

Performance improvements:

- in send, better file path caching improves runtime (on sample load
by -30%)

- on s390x with hardware zlib support prepare the input buffer in a
better way to get the best results from the acceleration

- minor speed improvement in encoded read, avoid memory allocation in
synchronous mode

Core:

- enable stable writes on inodes, replacing manually waiting for
writeback and allowing to skip that on inodes without checksums

- add last checks and warnings for out-of-band dirty writes to pages,
requiring a fixup ("fixup worker"), this should not be necessary
since 5.8 where get_user_page() and pin_user_pages*() prevent this
- long history behind that, we'll be happy to remove the whole
infrastructure in the near future

- more folio API conversions and preparations for large folio support

- subpage cleanups and refactoring, split handling of data and
metadata to allow future support for large folios

- readpage works as block-by-block, no change for normal mode, this
is preparation for future subpage updates

- block group refcount fixes and hardening

- delayed iput fixes

- in zoned mode, fix zone activation on filesystem with missing
devices

Cleanups:

- inode parameter cleanups

- path auto-freeing updates

- code flow simplifications in send

- redundant parameter cleanups"

* tag 'for-6.15-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux: (164 commits)
btrfs: zoned: fix zone finishing with missing devices
btrfs: zoned: fix zone activation with missing devices
btrfs: remove end_no_trans label from btrfs_log_inode_parent()
btrfs: simplify condition for logging new dentries at btrfs_log_inode_parent()
btrfs: remove redundant else statement from btrfs_log_inode_parent()
btrfs: use memcmp_extent_buffer() at replay_one_extent()
btrfs: update outdated comment for overwrite_item()
btrfs: use variables to store extent buffer and slot at overwrite_item()
btrfs: avoid unnecessary memory allocation and copy at overwrite_item()
btrfs: don't clobber ret in btrfs_validate_super()
btrfs: prepare btrfs_page_mkwrite() for large folios
btrfs: prepare extent_io.c for future large folio support
btrfs: prepare btrfs_launcher_folio() for large folios support
btrfs: replace PAGE_SIZE with folio_size for subpage.[ch]
btrfs: add a size parameter to btrfs_alloc_subpage()
btrfs: subpage: make btrfs_is_subpage() check against a folio
btrfs: add extra warning if delayed iput is added when it's not allowed
btrfs: avoid redundant path slot assignment in btrfs_search_forward()
btrfs: remove unnecessary btrfs_key local variable in btrfs_search_forward()
btrfs: simplify the return value handling in search_ioctl()
...

fs/btrfs/accessors.h

···
 #include <linux/string.h>
 #include <linux/mm.h>
 #include <uapi/linux/btrfs_tree.h>
+#include "extent_io.h"
 
 struct extent_buffer;
 

fs/btrfs/acl.h

···
 #ifndef BTRFS_ACL_H
 #define BTRFS_ACL_H
 
+#include <linux/types.h>
+
 struct posix_acl;
 struct inode;
 struct btrfs_trans_handle;

fs/btrfs/async-thread.c

···
 {
 	int new_current_active;
 	long pending;
-	int need_change = 0;
+	bool need_change = false;
 
 	if (wq->thresh == NO_THRESHOLD)
 		return;
···
 		new_current_active--;
 	new_current_active = clamp_val(new_current_active, 1, wq->limit_active);
 	if (new_current_active != wq->current_active)  {
-		need_change = 1;
+		need_change = true;
 		wq->current_active = new_current_active;
 	}
 out:
 	spin_unlock(&wq->thres_lock);
 
-	if (need_change) {
+	if (need_change)
 		workqueue_set_max_active(wq->normal_wq, wq->current_active);
-	}
 }
 
 static void run_ordered_work(struct btrfs_workqueue *wq,
···
 	struct btrfs_work *work = container_of(normal_work, struct btrfs_work,
 					       normal_work);
 	struct btrfs_workqueue *wq = work->wq;
-	int need_order = 0;
+	bool need_order = false;
 
 	/*
 	 * We should not touch things inside work in the following cases:
···
 	 * So we save the needed things here.
 	 */
 	if (work->ordered_func)
-		need_order = 1;
+		need_order = true;
 
 	trace_btrfs_work_sched(work);
 	thresh_exec_hook(wq);

fs/btrfs/backref.c

···
 		ASSERT(ctx->roots == NULL);
 
 	key.objectid = ctx->bytenr;
-	key.offset = (u64)-1;
 	if (btrfs_fs_incompat(ctx->fs_info, SKINNY_METADATA))
 		key.type = BTRFS_METADATA_ITEM_KEY;
 	else
 		key.type = BTRFS_EXTENT_ITEM_KEY;
+	key.offset = (u64)-1;
 
 	path = btrfs_alloc_path();
 	if (!path)
···
 	struct btrfs_extent_item *ei;
 	struct btrfs_key key;
 
+	key.objectid = logical;
 	if (btrfs_fs_incompat(fs_info, SKINNY_METADATA))
 		key.type = BTRFS_METADATA_ITEM_KEY;
 	else
 		key.type = BTRFS_EXTENT_ITEM_KEY;
-	key.objectid = logical;
 	key.offset = (u64)-1;
 
 	ret = btrfs_search_slot(NULL, extent_root, &key, path, 0, 0);

fs/btrfs/bio.c

···
 	return bbio;
 }
 
-/* Free a bio that was never submitted to the underlying device. */
-static void btrfs_cleanup_bio(struct btrfs_bio *bbio)
-{
-	if (bbio_has_ordered_extent(bbio))
-		btrfs_put_ordered_extent(bbio->ordered);
-	bio_put(&bbio->bio);
-}
-
-static void __btrfs_bio_end_io(struct btrfs_bio *bbio)
-{
-	if (bbio_has_ordered_extent(bbio)) {
-		struct btrfs_ordered_extent *ordered = bbio->ordered;
-
-		bbio->end_io(bbio);
-		btrfs_put_ordered_extent(ordered);
-	} else {
-		bbio->end_io(bbio);
-	}
-}
-
 void btrfs_bio_end_io(struct btrfs_bio *bbio, blk_status_t status)
 {
 	bbio->bio.bi_status = status;
 	if (bbio->bio.bi_pool == &btrfs_clone_bioset) {
 		struct btrfs_bio *orig_bbio = bbio->private;
 
-		btrfs_cleanup_bio(bbio);
+		/* Free bio that was never submitted to the underlying device. */
+		if (bbio_has_ordered_extent(bbio))
+			btrfs_put_ordered_extent(bbio->ordered);
+		bio_put(&bbio->bio);
+
 		bbio = orig_bbio;
 	}
 
···
 		/* Load split bio's error which might be set above. */
 		if (status == BLK_STS_OK)
 			bbio->bio.bi_status = READ_ONCE(bbio->status);
-		__btrfs_bio_end_io(bbio);
+
+		if (bbio_has_ordered_extent(bbio)) {
+			struct btrfs_ordered_extent *ordered = bbio->ordered;
+
+			bbio->end_io(bbio);
+			btrfs_put_ordered_extent(ordered);
+		} else {
+			bbio->end_io(bbio);
+		}
 	}
 }
 
···
 
 	/* If an error occurred we just want to clean up the bio and move on. */
 	if (bio->bi_status) {
-		btrfs_bio_end_io(async->bbio, async->bbio->bio.bi_status);
+		btrfs_bio_end_io(async->bbio, bio->bi_status);
 		return;
 	}
 

fs/btrfs/block-group.c

···
 /*
  * 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)
+static int btrfs_add_block_group_cache(struct btrfs_block_group *block_group)
 {
+	struct btrfs_fs_info *fs_info = block_group->fs_info;
 	struct rb_node *exist;
 	int ret = 0;
 
 	ASSERT(block_group->length != 0);
 
-	write_lock(&info->block_group_cache_lock);
+	write_lock(&fs_info->block_group_cache_lock);
 
 	exist = rb_find_add_cached(&block_group->cache_node,
-			&info->block_group_cache_tree, btrfs_bg_start_cmp);
+			&fs_info->block_group_cache_tree, btrfs_bg_start_cmp);
 	if (exist)
 		ret = -EEXIST;
-	write_unlock(&info->block_group_cache_lock);
+	write_unlock(&fs_info->block_group_cache_lock);
 
 	return ret;
 }
···
 	struct btrfs_root *extent_root;
 	u64 search_offset;
 	u64 search_end = block_group->start + block_group->length;
-	struct btrfs_path *path;
+	BTRFS_PATH_AUTO_FREE(path);
 	struct btrfs_key search_key;
 	int ret = 0;
 
···
 
 	lockdep_assert_held(&caching_ctl->mutex);
 	lockdep_assert_held_read(&fs_info->commit_root_sem);
-	btrfs_free_path(path);
 	return ret;
 }
 
···
 	path->reada = READA_FORWARD;
 
 	key.objectid = last;
-	key.offset = 0;
 	key.type = BTRFS_EXTENT_ITEM_KEY;
+	key.offset = 0;
 
 next:
 	ret = btrfs_search_slot(NULL, extent_root, &key, path, 0, 0);
···
 
 		if (key.objectid < last) {
 			key.objectid = last;
-			key.offset = 0;
 			key.type = BTRFS_EXTENT_ITEM_KEY;
+			key.offset = 0;
 			btrfs_release_path(path);
 			goto next;
 		}
···
 }
 
 /*
+ * Link the block_group to a list via bg_list.
+ *
+ * @bg:       The block_group to link to the list.
+ * @list:     The list to link it to.
+ *
+ * Use this rather than list_add_tail() directly to ensure proper respect
+ * to locking and refcounting.
+ *
+ * Returns: true if the bg was linked with a refcount bump and false otherwise.
+ */
+static bool btrfs_link_bg_list(struct btrfs_block_group *bg, struct list_head *list)
+{
+	struct btrfs_fs_info *fs_info = bg->fs_info;
+	bool added = false;
+
+	spin_lock(&fs_info->unused_bgs_lock);
+	if (list_empty(&bg->bg_list)) {
+		btrfs_get_block_group(bg);
+		list_add_tail(&bg->bg_list, list);
+		added = true;
+	}
+	spin_unlock(&fs_info->unused_bgs_lock);
+	return added;
+}
+
+/*
  * Process the unused_bgs list and remove any that don't have any allocated
  * space inside of them.
  */
···
 			 * drop under the "next" label for the
 			 * fs_info->unused_bgs list.
 			 */
-			btrfs_get_block_group(block_group);
-			list_add_tail(&block_group->bg_list, &retry_list);
+			btrfs_link_bg_list(block_group, &retry_list);
 
 			trace_btrfs_skip_unused_block_group(block_group);
 			spin_unlock(&block_group->lock);
···
 	list_sort(NULL, &fs_info->reclaim_bgs, reclaim_bgs_cmp);
 	while (!list_empty(&fs_info->reclaim_bgs)) {
 		u64 zone_unusable;
-		u64 reclaimed;
+		u64 used;
+		u64 reserved;
 		int ret = 0;
 
 		bg = list_first_entry(&fs_info->reclaim_bgs,
···
 			up_write(&space_info->groups_sem);
 			goto next;
 		}
+
+		/*
+		 * Cache the zone_unusable value before turning the block group
+		 * to read only. As soon as the block group is read only it's
+		 * zone_unusable value gets moved to the block group's read-only
+		 * bytes and isn't available for calculations anymore. We also
+		 * cache it before unlocking the block group, to prevent races
+		 * (reports from KCSAN and such tools) with tasks updating it.
+		 */
+		zone_unusable = bg->zone_unusable;
+
 		spin_unlock(&bg->lock);
 		spin_unlock(&space_info->lock);
 
···
 			goto next;
 		}
 
-		/*
-		 * Cache the zone_unusable value before turning the block group
-		 * to read only. As soon as the blog group is read only it's
-		 * zone_unusable value gets moved to the block group's read-only
-		 * bytes and isn't available for calculations anymore.
-		 */
-		zone_unusable = bg->zone_unusable;
 		ret = inc_block_group_ro(bg, 0);
 		up_write(&space_info->groups_sem);
 		if (ret < 0)
 			goto next;
 
+		/*
+		 * The amount of bytes reclaimed corresponds to the sum of the
+		 * "used" and "reserved" counters. We have set the block group
+		 * to RO above, which prevents reservations from happening but
+		 * we may have existing reservations for which allocation has
+		 * not yet been done - btrfs_update_block_group() was not yet
+		 * called, which is where we will transfer a reserved extent's
+		 * size from the "reserved" counter to the "used" counter - this
+		 * happens when running delayed references. When we relocate the
+		 * chunk below, relocation first flushes dellaloc, waits for
+		 * ordered extent completion (which is where we create delayed
+		 * references for data extents) and commits the current
+		 * transaction (which runs delayed references), and only after
+		 * it does the actual work to move extents out of the block
+		 * group. So the reported amount of reclaimed bytes is
+		 * effectively the sum of the 'used' and 'reserved' counters.
+		 */
+		spin_lock(&bg->lock);
+		used = bg->used;
+		reserved = bg->reserved;
+		spin_unlock(&bg->lock);
+
 		btrfs_info(fs_info,
-			"reclaiming chunk %llu with %llu%% used %llu%% unusable",
+	"reclaiming chunk %llu with %llu%% used %llu%% reserved %llu%% unusable",
 				bg->start,
-				div64_u64(bg->used * 100, bg->length),
+				div64_u64(used * 100, bg->length),
+				div64_u64(reserved * 100, bg->length),
 				div64_u64(zone_unusable * 100, bg->length));
 		trace_btrfs_reclaim_block_group(bg);
-		reclaimed = bg->used;
 		ret = btrfs_relocate_chunk(fs_info, bg->start);
 		if (ret) {
 			btrfs_dec_block_group_ro(bg);
 			btrfs_err(fs_info, "error relocating chunk %llu",
 				  bg->start);
-			reclaimed = 0;
+			used = 0;
+			reserved = 0;
 			spin_lock(&space_info->lock);
 			space_info->reclaim_errors++;
 			if (READ_ONCE(space_info->periodic_reclaim))
···
 		}
 		spin_lock(&space_info->lock);
 		space_info->reclaim_count++;
-		space_info->reclaim_bytes += reclaimed;
+		space_info->reclaim_bytes += used;
+		space_info->reclaim_bytes += reserved;
 		spin_unlock(&space_info->lock);
 
 next:
-		if (ret && !READ_ONCE(space_info->periodic_reclaim)) {
-			/* Refcount held by the reclaim_bgs list after splice. */
-			spin_lock(&fs_info->unused_bgs_lock);
-			/*
-			 * This block group might be added to the unused list
-			 * during the above process. Move it back to the
-			 * reclaim list otherwise.
-			 */
-			if (list_empty(&bg->bg_list)) {
-				btrfs_get_block_group(bg);
-				list_add_tail(&bg->bg_list, &retry_list);
-			}
-			spin_unlock(&fs_info->unused_bgs_lock);
-		}
+		if (ret && !READ_ONCE(space_info->periodic_reclaim))
+			btrfs_link_bg_list(bg, &retry_list);
 		btrfs_put_block_group(bg);
 
 		mutex_unlock(&fs_info->reclaim_bgs_lock);
···
 {
 	struct btrfs_fs_info *fs_info = bg->fs_info;
 
-	spin_lock(&fs_info->unused_bgs_lock);
-	if (list_empty(&bg->bg_list)) {
-		btrfs_get_block_group(bg);
+	if (btrfs_link_bg_list(bg, &fs_info->reclaim_bgs))
 		trace_btrfs_add_reclaim_block_group(bg);
-		list_add_tail(&bg->bg_list, &fs_info->reclaim_bgs);
-	}
-	spin_unlock(&fs_info->unused_bgs_lock);
 }
 
 static int read_bg_from_eb(struct btrfs_fs_info *fs_info, const struct btrfs_key *key,
···
 			goto error;
 	}
 
-	ret = btrfs_add_block_group_cache(info, cache);
+	ret = btrfs_add_block_group_cache(cache);
 	if (ret) {
 		btrfs_remove_free_space_cache(cache);
 		goto error;
···
 		bg->cached = BTRFS_CACHE_FINISHED;
 		bg->used = map->chunk_len;
 		bg->flags = map->type;
-		ret = btrfs_add_block_group_cache(fs_info, bg);
+		ret = btrfs_add_block_group_cache(bg);
 		/*
 		 * We may have some valid block group cache added already, in
 		 * that case we skip to the next one.
···
 		return fill_dummy_bgs(info);
 
 	key.objectid = 0;
-	key.offset = 0;
 	key.type = BTRFS_BLOCK_GROUP_ITEM_KEY;
+	key.offset = 0;
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
···
 {
 	struct btrfs_fs_info *fs_info = device->fs_info;
 	struct btrfs_root *root = fs_info->dev_root;
-	struct btrfs_path *path;
+	BTRFS_PATH_AUTO_FREE(path);
 	struct btrfs_dev_extent *extent;
 	struct extent_buffer *leaf;
 	struct btrfs_key key;
···
 	key.offset = start;
 	ret = btrfs_insert_empty_item(trans, root, path, &key, sizeof(*extent));
 	if (ret)
-		goto out;
+		return ret;
 
 	leaf = path->nodes[0];
 	extent = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_dev_extent);
···
 	btrfs_set_dev_extent_chunk_objectid(leaf, extent,
 					    BTRFS_FIRST_CHUNK_TREE_OBJECTID);
 	btrfs_set_dev_extent_chunk_offset(leaf, extent, chunk_offset);
-
 	btrfs_set_dev_extent_length(leaf, extent, num_bytes);
-out:
-	btrfs_free_path(path);
+
 	return ret;
 }
 
···
 		/* Already aborted the transaction if it failed. */
 next:
 		btrfs_dec_delayed_refs_rsv_bg_inserts(fs_info);
+
+		spin_lock(&fs_info->unused_bgs_lock);
 		list_del_init(&block_group->bg_list);
 		clear_bit(BLOCK_GROUP_FLAG_NEW, &block_group->runtime_flags);
+		btrfs_put_block_group(block_group);
+		spin_unlock(&fs_info->unused_bgs_lock);
 
 		/*
 		 * If the block group is still unused, add it to the list of
···
 	cache->space_info = btrfs_find_space_info(fs_info, cache->flags);
 	ASSERT(cache->space_info);
 
-	ret = btrfs_add_block_group_cache(fs_info, cache);
+	ret = btrfs_add_block_group_cache(cache);
 	if (ret) {
 		btrfs_remove_free_space_cache(cache);
 		btrfs_put_block_group(cache);
···
 	}
 #endif
 
-	list_add_tail(&cache->bg_list, &trans->new_bgs);
+	btrfs_link_bg_list(cache, &trans->new_bgs);
 	btrfs_inc_delayed_refs_rsv_bg_inserts(fs_info);
 
 	set_avail_alloc_bits(fs_info, type);
···
 	struct btrfs_fs_info *fs_info = trans->fs_info;
 	struct btrfs_block_group *cache, *tmp;
 	struct btrfs_transaction *cur_trans = trans->transaction;
-	struct btrfs_path *path;
+	BTRFS_PATH_AUTO_FREE(path);
 
 	if (list_empty(&cur_trans->dirty_bgs) ||
 	    !btrfs_test_opt(fs_info, SPACE_CACHE))
···
 			cache_save_setup(cache, trans, path);
 	}
 
-	btrfs_free_path(path);
 	return 0;
 }
 
···
 	struct btrfs_transaction *cur_trans = trans->transaction;
 	int ret = 0;
 	int should_put;
-	struct btrfs_path *path = NULL;
+	BTRFS_PATH_AUTO_FREE(path);
 	LIST_HEAD(dirty);
 	struct list_head *io = &cur_trans->io_bgs;
 	int loops = 0;
···
 		btrfs_cleanup_dirty_bgs(cur_trans, fs_info);
 	}
 
-	btrfs_free_path(path);
 	return ret;
 }
 
···
 	struct btrfs_transaction *cur_trans = trans->transaction;
 	int ret = 0;
 	int should_put;
-	struct btrfs_path *path;
+	BTRFS_PATH_AUTO_FREE(path);
 	struct list_head *io = &cur_trans->io_bgs;
 
 	path = btrfs_alloc_path();
···
 		btrfs_put_block_group(cache);
 	}
 
-	btrfs_free_path(path);
 	return ret;
 }
 

fs/btrfs/btrfs_inode.h

···
 	 * different from prop_compress and takes precedence if set.
 	 */
 	u8 defrag_compress;
+	s8 defrag_compress_level;
 
 	/*
 	 * Lock for counters and all fields used to determine if the inode is in
···
 	lockdep_assert_held(&inode->vfs_inode.i_rwsem);
 }
 
+static inline void btrfs_update_inode_mapping_flags(struct btrfs_inode *inode)
+{
+	if (inode->flags & BTRFS_INODE_NODATASUM)
+		mapping_clear_stable_writes(inode->vfs_inode.i_mapping);
+	else
+		mapping_set_stable_writes(inode->vfs_inode.i_mapping);
+}
+
 /* Array of bytes with variable length, hexadecimal format 0x1234 */
 #define CSUM_FMT				"0x%*phN"
 #define CSUM_FMT_VALUE(size, bytes)		size, bytes
···
 			    u32 pgoff, u8 *csum, const u8 * const csum_expected);
 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,
+noinline int can_nocow_extent(struct btrfs_inode *inode, u64 offset, u64 *len,
 			      struct btrfs_file_extent *file_extent,
 			      bool nowait);
 
···
 int btrfs_drop_inode(struct inode *inode);
 int __init btrfs_init_cachep(void);
 void __cold btrfs_destroy_cachep(void);
-struct inode *btrfs_iget_path(u64 ino, struct btrfs_root *root,
-			      struct btrfs_path *path);
-struct inode *btrfs_iget(u64 ino, struct btrfs_root *root);
+struct btrfs_inode *btrfs_iget_path(u64 ino, struct btrfs_root *root,
+				    struct btrfs_path *path);
+struct btrfs_inode *btrfs_iget(u64 ino, struct btrfs_root *root);
 struct extent_map *btrfs_get_extent(struct btrfs_inode *inode,
 				    struct folio *folio, u64 start, u64 len);
 int btrfs_update_inode(struct btrfs_trans_handle *trans,

fs/btrfs/compression.c

···
 	&btrfs_zstd_compress,
 };
 
-static struct list_head *alloc_workspace(int type, unsigned int level)
+static struct list_head *alloc_workspace(int type, int level)
 {
 	switch (type) {
 	case BTRFS_COMPRESS_NONE: return alloc_heuristic_ws();
···
  * Preallocation makes a forward progress guarantees and we do not return
  * errors.
  */
-struct list_head *btrfs_get_workspace(int type, unsigned int level)
+struct list_head *btrfs_get_workspace(int type, int level)
 {
 	struct workspace_manager *wsm;
 	struct list_head *workspace;
···
  * Adjust @level according to the limits of the compression algorithm or
  * fallback to default
  */
-static unsigned int btrfs_compress_set_level(int type, unsigned level)
+static int btrfs_compress_set_level(unsigned int type, int level)
 {
 	const struct btrfs_compress_op *ops = btrfs_compress_op[type];
 
 	if (level == 0)
 		level = ops->default_level;
 	else
-		level = min(level, ops->max_level);
+		level = min(max(level, ops->min_level), ops->max_level);
 
 	return level;
+}
+
+/*
+ * Check whether the @level is within the valid range for the given type.
+ */
+bool btrfs_compress_level_valid(unsigned int type, int level)
+{
+	const struct btrfs_compress_op *ops = btrfs_compress_op[type];
+
+	return ops->min_level <= level && level <= ops->max_level;
 }
 
 /* Wrapper around find_get_page(), with extra error message. */
···
  * @total_out is an in/out parameter, must be set to the input length and will
  * be also used to return the total number of compressed bytes
  */
-int btrfs_compress_folios(unsigned int type_level, struct address_space *mapping,
+int btrfs_compress_folios(unsigned int type, int level, struct address_space *mapping,
 			 u64 start, struct folio **folios, unsigned long *out_folios,
 			 unsigned long *total_in, unsigned long *total_out)
 {
-	int type = btrfs_compress_type(type_level);
-	int level = btrfs_compress_level(type_level);
 	const unsigned long orig_len = *total_out;
 	struct list_head *workspace;
 	int ret;
···
 
 /*
  * Convert the compression suffix (eg. after "zlib" starting with ":") to
- * level, unrecognized string will set the default level
+ * level, unrecognized string will set the default level. Negative level
+ * numbers are allowed.
  */
-unsigned int btrfs_compress_str2level(unsigned int type, const char *str)
+int btrfs_compress_str2level(unsigned int type, const char *str)
 {
-	unsigned int level = 0;
+	int level = 0;
 	int ret;
 
 	if (!type)
 		return 0;
 
 	if (str[0] == ':') {
-		ret = kstrtouint(str + 1, 10, &level);
+		ret = kstrtoint(str + 1, 10, &level);
 		if (ret)
 			level = 0;
 	}

fs/btrfs/compression.h

···
 	struct btrfs_bio bbio;
 };
 
-static inline unsigned int btrfs_compress_type(unsigned int type_level)
-{
-	return (type_level & 0xF);
-}
-
-static inline unsigned int btrfs_compress_level(unsigned int type_level)
-{
-	return ((type_level & 0xF0) >> 4);
-}
-
 /* @range_end must be exclusive. */
 static inline u32 btrfs_calc_input_length(u64 range_end, u64 cur)
 {
···
 int __init btrfs_init_compress(void);
 void __cold btrfs_exit_compress(void);
 
-int btrfs_compress_folios(unsigned int type_level, struct address_space *mapping,
+bool btrfs_compress_level_valid(unsigned int type, int level);
+int btrfs_compress_folios(unsigned int type, int level, struct address_space *mapping,
 			  u64 start, struct folio **folios, unsigned long *out_folios,
 			 unsigned long *total_in, unsigned long *total_out);
 int btrfs_decompress(int type, const u8 *data_in, struct folio *dest_folio,
···
 				   bool writeback);
 void btrfs_submit_compressed_read(struct btrfs_bio *bbio);
 
-unsigned int btrfs_compress_str2level(unsigned int type, const char *str);
+int btrfs_compress_str2level(unsigned int type, const char *str);
 
 struct folio *btrfs_alloc_compr_folio(void);
 void btrfs_free_compr_folio(struct folio *folio);
···
 	wait_queue_head_t ws_wait;
 };
 
-struct list_head *btrfs_get_workspace(int type, unsigned int level);
+struct list_head *btrfs_get_workspace(int type, int level);
 void btrfs_put_workspace(int type, struct list_head *ws);
 
 struct btrfs_compress_op {
 	struct workspace_manager *workspace_manager;
 	/* Maximum level supported by the compression algorithm */
-	unsigned int max_level;
-	unsigned int default_level;
+	int min_level;
+	int max_level;
+	int default_level;
 };
 
 /* The heuristic workspaces are managed via the 0th workspace manager */
···
 		size_t destlen);
 void zstd_init_workspace_manager(void);
 void zstd_cleanup_workspace_manager(void);
-struct list_head *zstd_alloc_workspace(unsigned int level);
+struct list_head *zstd_alloc_workspace(int level);
 void zstd_free_workspace(struct list_head *ws);
-struct list_head *zstd_get_workspace(unsigned int level);
+struct list_head *zstd_get_workspace(int level);
 void zstd_put_workspace(struct list_head *ws);
 
 #endif

fs/btrfs/ctree.c

···
 		      u32 data_size)
 {
 	int ret = 0;
-	struct btrfs_path *path;
+	BTRFS_PATH_AUTO_FREE(path);
 	struct extent_buffer *leaf;
 	unsigned long ptr;
 
···
 		write_extent_buffer(leaf, data, ptr, data_size);
 		btrfs_mark_buffer_dirty(trans, leaf);
 	}
-	btrfs_free_path(path);
 	return ret;
 }
 
···
 			 u64 min_trans)
 {
 	struct extent_buffer *cur;
-	struct btrfs_key found_key;
 	int slot;
 	int sret;
 	u32 nritems;
···
 				goto find_next_key;
 			ret = 0;
 			path->slots[level] = slot;
-			btrfs_item_key_to_cpu(cur, &found_key, slot);
+			/* Save our key for returning back. */
+			btrfs_item_key_to_cpu(cur, min_key, slot);
 			goto out;
 		}
 		if (sret && slot > 0)
···
 		 * we didn't find a candidate key in this node, walk forward
 		 * and find another one
 		 */
+		path->slots[level] = slot;
 		if (slot >= nritems) {
-			path->slots[level] = slot;
 			sret = btrfs_find_next_key(root, path, min_key, level,
 						  min_trans);
 			if (sret == 0) {
···
 				goto out;
 			}
 		}
-		/* save our key for returning back */
-		btrfs_node_key_to_cpu(cur, &found_key, slot);
-		path->slots[level] = slot;
 		if (level == path->lowest_level) {
 			ret = 0;
+			/* Save our key for returning back. */
+			btrfs_node_key_to_cpu(cur, min_key, slot);
 			goto out;
 		}
 		cur = btrfs_read_node_slot(cur, slot);
···
 	}
 out:
 	path->keep_locks = keep_locks;
-	if (ret == 0) {
+	if (ret == 0)
 		btrfs_unlock_up_safe(path, path->lowest_level + 1);
-		memcpy(min_key, &found_key, sizeof(found_key));
-	}
 	return ret;
 }
 

fs/btrfs/ctree.h

···
 #ifndef BTRFS_CTREE_H
 #define BTRFS_CTREE_H
 
-#include "linux/cleanup.h"
+#include <linux/cleanup.h>
 #include <linux/spinlock.h>
 #include <linux/rbtree.h>
 #include <linux/mutex.h>

fs/btrfs/defrag.c

···
 				  struct file_ra_state *ra)
 {
 	struct btrfs_root *inode_root;
-	struct inode *inode;
+	struct btrfs_inode *inode;
 	struct btrfs_ioctl_defrag_range_args range;
 	int ret = 0;
 	u64 cur = 0;
···
 		goto cleanup;
 	}
 
-	if (cur >= i_size_read(inode)) {
-		iput(inode);
+	if (cur >= i_size_read(&inode->vfs_inode)) {
+		iput(&inode->vfs_inode);
 		goto cleanup;
 	}
 
 	/* Do a chunk of defrag */
-	clear_bit(BTRFS_INODE_IN_DEFRAG, &BTRFS_I(inode)->runtime_flags);
+	clear_bit(BTRFS_INODE_IN_DEFRAG, &inode->runtime_flags);
 	memset(&range, 0, sizeof(range));
 	range.len = (u64)-1;
 	range.start = cur;
 	range.extent_thresh = defrag->extent_thresh;
-	file_ra_state_init(ra, inode->i_mapping);
+	file_ra_state_init(ra, inode->vfs_inode.i_mapping);
 
 	sb_start_write(fs_info->sb);
 	ret = btrfs_defrag_file(inode, ra, &range, defrag->transid,
-				       BTRFS_DEFRAG_BATCH);
+				BTRFS_DEFRAG_BATCH);
 	sb_end_write(fs_info->sb);
-	iput(inode);
+	iput(&inode->vfs_inode);
 
 	if (ret < 0)
 		goto cleanup;
···
  * (Mostly for autodefrag, which sets @max_to_defrag thus we may exit early without
  *  defragging all the range).
  */
-int btrfs_defrag_file(struct inode *inode, struct file_ra_state *ra,
+int btrfs_defrag_file(struct btrfs_inode *inode, struct file_ra_state *ra,
 		      struct btrfs_ioctl_defrag_range_args *range,
 		      u64 newer_than, unsigned long max_to_defrag)
 {
-	struct btrfs_fs_info *fs_info = inode_to_fs_info(inode);
+	struct btrfs_fs_info *fs_info = inode->root->fs_info;
 	unsigned long sectors_defragged = 0;
-	u64 isize = i_size_read(inode);
+	u64 isize = i_size_read(&inode->vfs_inode);
 	u64 cur;
 	u64 last_byte;
 	bool do_compress = (range->flags & BTRFS_DEFRAG_RANGE_COMPRESS);
 	int compress_type = BTRFS_COMPRESS_ZLIB;
+	int compress_level = 0;
 	int ret = 0;
 	u32 extent_thresh = range->extent_thresh;
 	pgoff_t start_index;
···
 		return -EINVAL;
 
 	if (do_compress) {
-		if (range->compress_type >= BTRFS_NR_COMPRESS_TYPES)
-			return -EINVAL;
-		if (range->compress_type)
-			compress_type = range->compress_type;
+		if (range->flags & BTRFS_DEFRAG_RANGE_COMPRESS_LEVEL) {
+			if (range->compress.type >= BTRFS_NR_COMPRESS_TYPES)
+				return -EINVAL;
+			if (range->compress.type) {
+				compress_type  = range->compress.type;
+				compress_level = range->compress.level;
+				if (!btrfs_compress_level_valid(compress_type, compress_level))
+					return -EINVAL;
+			}
+		} else {
+			if (range->compress_type >= BTRFS_NR_COMPRESS_TYPES)
+				return -EINVAL;
+			if (range->compress_type)
+				compress_type = range->compress_type;
+		}
 	}
 
 	if (extent_thresh == 0)
···
 	 * defrag range can be written sequentially.
 	 */
 	start_index = cur >> PAGE_SHIFT;
-	if (start_index < inode->i_mapping->writeback_index)
-		inode->i_mapping->writeback_index = start_index;
+	if (start_index < inode->vfs_inode.i_mapping->writeback_index)
+		inode->vfs_inode.i_mapping->writeback_index = start_index;
 
 	while (cur < last_byte) {
 		const unsigned long prev_sectors_defragged = sectors_defragged;
···
 			       (SZ_256K >> PAGE_SHIFT)) << PAGE_SHIFT) - 1;
 		cluster_end = min(cluster_end, last_byte);
 
-		btrfs_inode_lock(BTRFS_I(inode), 0);
-		if (IS_SWAPFILE(inode)) {
+		btrfs_inode_lock(inode, 0);
+		if (IS_SWAPFILE(&inode->vfs_inode)) {
 			ret = -ETXTBSY;
-			btrfs_inode_unlock(BTRFS_I(inode), 0);
+			btrfs_inode_unlock(inode, 0);
 			break;
 		}
-		if (!(inode->i_sb->s_flags & SB_ACTIVE)) {
-			btrfs_inode_unlock(BTRFS_I(inode), 0);
+		if (!(inode->vfs_inode.i_sb->s_flags & SB_ACTIVE)) {
+			btrfs_inode_unlock(inode, 0);
 			break;
 		}
-		if (do_compress)
-			BTRFS_I(inode)->defrag_compress = compress_type;
-		ret = defrag_one_cluster(BTRFS_I(inode), ra, cur,
+		if (do_compress) {
+			inode->defrag_compress = compress_type;
+			inode->defrag_compress_level = compress_level;
+		}
+		ret = defrag_one_cluster(inode, ra, cur,
 				cluster_end + 1 - cur, extent_thresh,
 				newer_than, do_compress, &sectors_defragged,
 				max_to_defrag, &last_scanned);
 
 		if (sectors_defragged > prev_sectors_defragged)
-			balance_dirty_pages_ratelimited(inode->i_mapping);
+			balance_dirty_pages_ratelimited(inode->vfs_inode.i_mapping);
 
-		btrfs_inode_unlock(BTRFS_I(inode), 0);
+		btrfs_inode_unlock(inode, 0);
 		if (ret < 0)
 			break;
 		cur = max(cluster_end + 1, last_scanned);
···
 		 * need to be written back immediately.
 		 */
 		if (range->flags & BTRFS_DEFRAG_RANGE_START_IO) {
-			filemap_flush(inode->i_mapping);
+			filemap_flush(inode->vfs_inode.i_mapping);
 			if (test_bit(BTRFS_INODE_HAS_ASYNC_EXTENT,
-				     &BTRFS_I(inode)->runtime_flags))
-				filemap_flush(inode->i_mapping);
+				     &inode->runtime_flags))
+				filemap_flush(inode->vfs_inode.i_mapping);
 		}
 		if (range->compress_type == BTRFS_COMPRESS_LZO)
 			btrfs_set_fs_incompat(fs_info, COMPRESS_LZO);
···
 		ret = sectors_defragged;
 	}
 	if (do_compress) {
-		btrfs_inode_lock(BTRFS_I(inode), 0);
-		BTRFS_I(inode)->defrag_compress = BTRFS_COMPRESS_NONE;
-		btrfs_inode_unlock(BTRFS_I(inode), 0);
+		btrfs_inode_lock(inode, 0);
+		inode->defrag_compress = BTRFS_COMPRESS_NONE;
+		btrfs_inode_unlock(inode, 0);
 	}
 	return ret;
 }

fs/btrfs/defrag.h

···
 #include <linux/types.h>
 #include <linux/compiler_types.h>
 
-struct inode;
 struct file_ra_state;
+struct btrfs_inode;
 struct btrfs_fs_info;
 struct btrfs_root;
 struct btrfs_trans_handle;
 struct btrfs_ioctl_defrag_range_args;
 
-int btrfs_defrag_file(struct inode *inode, struct file_ra_state *ra,
+int btrfs_defrag_file(struct btrfs_inode *inode, struct file_ra_state *ra,
 		      struct btrfs_ioctl_defrag_range_args *range,
 		      u64 newer_than, unsigned long max_to_defrag);
 int __init btrfs_auto_defrag_init(void);

fs/btrfs/delayed-inode.c

···
 				     struct btrfs_inode *inode)
 {
 	struct btrfs_delayed_node *delayed_node = btrfs_get_delayed_node(inode);
-	struct btrfs_path *path;
+	BTRFS_PATH_AUTO_FREE(path);
 	struct btrfs_block_rsv *block_rsv;
 	int ret;
 
···
 	ret = __btrfs_commit_inode_delayed_items(trans, path, delayed_node);
 
 	btrfs_release_delayed_node(delayed_node);
-	btrfs_free_path(path);
 	trans->block_rsv = block_rsv;
 
 	return ret;
···
 
 static void fill_stack_inode_item(struct btrfs_trans_handle *trans,
 				  struct btrfs_inode_item *inode_item,
-				  struct inode *inode)
+				  struct btrfs_inode *inode)
 {
+	struct inode *vfs_inode = &inode->vfs_inode;
 	u64 flags;
 
-	btrfs_set_stack_inode_uid(inode_item, i_uid_read(inode));
-	btrfs_set_stack_inode_gid(inode_item, i_gid_read(inode));
-	btrfs_set_stack_inode_size(inode_item, BTRFS_I(inode)->disk_i_size);
-	btrfs_set_stack_inode_mode(inode_item, inode->i_mode);
-	btrfs_set_stack_inode_nlink(inode_item, inode->i_nlink);
-	btrfs_set_stack_inode_nbytes(inode_item, inode_get_bytes(inode));
-	btrfs_set_stack_inode_generation(inode_item,
-					 BTRFS_I(inode)->generation);
+	btrfs_set_stack_inode_uid(inode_item, i_uid_read(vfs_inode));
+	btrfs_set_stack_inode_gid(inode_item, i_gid_read(vfs_inode));
+	btrfs_set_stack_inode_size(inode_item, inode->disk_i_size);
+	btrfs_set_stack_inode_mode(inode_item, vfs_inode->i_mode);
+	btrfs_set_stack_inode_nlink(inode_item, vfs_inode->i_nlink);
+	btrfs_set_stack_inode_nbytes(inode_item, inode_get_bytes(vfs_inode));
+	btrfs_set_stack_inode_generation(inode_item, inode->generation);
 	btrfs_set_stack_inode_sequence(inode_item,
-				       inode_peek_iversion(inode));
+				       inode_peek_iversion(vfs_inode));
 	btrfs_set_stack_inode_transid(inode_item, trans->transid);
-	btrfs_set_stack_inode_rdev(inode_item, inode->i_rdev);
-	flags = btrfs_inode_combine_flags(BTRFS_I(inode)->flags,
-					  BTRFS_I(inode)->ro_flags);
+	btrfs_set_stack_inode_rdev(inode_item, vfs_inode->i_rdev);
+	flags = btrfs_inode_combine_flags(inode->flags, inode->ro_flags);
 	btrfs_set_stack_inode_flags(inode_item, flags);
 	btrfs_set_stack_inode_block_group(inode_item, 0);
 
 	btrfs_set_stack_timespec_sec(&inode_item->atime,
-				     inode_get_atime_sec(inode));
+				     inode_get_atime_sec(vfs_inode));
 	btrfs_set_stack_timespec_nsec(&inode_item->atime,
-				      inode_get_atime_nsec(inode));
+				      inode_get_atime_nsec(vfs_inode));
 
 	btrfs_set_stack_timespec_sec(&inode_item->mtime,
-				     inode_get_mtime_sec(inode));
+				     inode_get_mtime_sec(vfs_inode));
 	btrfs_set_stack_timespec_nsec(&inode_item->mtime,
-				      inode_get_mtime_nsec(inode));
+				      inode_get_mtime_nsec(vfs_inode));
 
 	btrfs_set_stack_timespec_sec(&inode_item->ctime,
-				     inode_get_ctime_sec(inode));
+				     inode_get_ctime_sec(vfs_inode));
 	btrfs_set_stack_timespec_nsec(&inode_item->ctime,
-				      inode_get_ctime_nsec(inode));
+				      inode_get_ctime_nsec(vfs_inode));
 
-	btrfs_set_stack_timespec_sec(&inode_item->otime, BTRFS_I(inode)->i_otime_sec);
-	btrfs_set_stack_timespec_nsec(&inode_item->otime, BTRFS_I(inode)->i_otime_nsec);
+	btrfs_set_stack_timespec_sec(&inode_item->otime, inode->i_otime_sec);
+	btrfs_set_stack_timespec_nsec(&inode_item->otime, inode->i_otime_nsec);
 }
 
-int btrfs_fill_inode(struct inode *inode, u32 *rdev)
+int btrfs_fill_inode(struct btrfs_inode *inode, u32 *rdev)
 {
-	struct btrfs_fs_info *fs_info = BTRFS_I(inode)->root->fs_info;
+	struct btrfs_fs_info *fs_info = inode->root->fs_info;
 	struct btrfs_delayed_node *delayed_node;
 	struct btrfs_inode_item *inode_item;
+	struct inode *vfs_inode = &inode->vfs_inode;
 
-	delayed_node = btrfs_get_delayed_node(BTRFS_I(inode));
+	delayed_node = btrfs_get_delayed_node(inode);
 	if (!delayed_node)
 		return -ENOENT;
 
···
 
 	inode_item = &delayed_node->inode_item;
 
-	i_uid_write(inode, btrfs_stack_inode_uid(inode_item));
-	i_gid_write(inode, btrfs_stack_inode_gid(inode_item));
-	btrfs_i_size_write(BTRFS_I(inode), btrfs_stack_inode_size(inode_item));
-	btrfs_inode_set_file_extent_range(BTRFS_I(inode), 0,
-			round_up(i_size_read(inode), fs_info->sectorsize));
-	inode->i_mode = btrfs_stack_inode_mode(inode_item);
-	set_nlink(inode, btrfs_stack_inode_nlink(inode_item));
-	inode_set_bytes(inode, btrfs_stack_inode_nbytes(inode_item));
-	BTRFS_I(inode)->generation = btrfs_stack_inode_generation(inode_item);
-        BTRFS_I(inode)->last_trans = btrfs_stack_inode_transid(inode_item);
+	i_uid_write(vfs_inode, btrfs_stack_inode_uid(inode_item));
+	i_gid_write(vfs_inode, btrfs_stack_inode_gid(inode_item));
+	btrfs_i_size_write(inode, btrfs_stack_inode_size(inode_item));
+	btrfs_inode_set_file_extent_range(inode, 0,
+			round_up(i_size_read(vfs_inode), fs_info->sectorsize));
+	vfs_inode->i_mode = btrfs_stack_inode_mode(inode_item);
+	set_nlink(vfs_inode, btrfs_stack_inode_nlink(inode_item));
+	inode_set_bytes(vfs_inode, btrfs_stack_inode_nbytes(inode_item));
+	inode->generation = btrfs_stack_inode_generation(inode_item);
+	inode->last_trans = btrfs_stack_inode_transid(inode_item);
 
-	inode_set_iversion_queried(inode,
-				   btrfs_stack_inode_sequence(inode_item));
-	inode->i_rdev = 0;
+	inode_set_iversion_queried(vfs_inode, btrfs_stack_inode_sequence(inode_item));
+	vfs_inode->i_rdev = 0;
 	*rdev = btrfs_stack_inode_rdev(inode_item);
 	btrfs_inode_split_flags(btrfs_stack_inode_flags(inode_item),
-				&BTRFS_I(inode)->flags, &BTRFS_I(inode)->ro_flags);
+				&inode->flags, &inode->ro_flags);
 
-	inode_set_atime(inode, btrfs_stack_timespec_sec(&inode_item->atime),
+	inode_set_atime(vfs_inode, btrfs_stack_timespec_sec(&inode_item->atime),
 			btrfs_stack_timespec_nsec(&inode_item->atime));
 
-	inode_set_mtime(inode, btrfs_stack_timespec_sec(&inode_item->mtime),
+	inode_set_mtime(vfs_inode, btrfs_stack_timespec_sec(&inode_item->mtime),
 			btrfs_stack_timespec_nsec(&inode_item->mtime));
 
-	inode_set_ctime(inode, btrfs_stack_timespec_sec(&inode_item->ctime),
+	inode_set_ctime(vfs_inode, btrfs_stack_timespec_sec(&inode_item->ctime),
 			btrfs_stack_timespec_nsec(&inode_item->ctime));
 
-	BTRFS_I(inode)->i_otime_sec = btrfs_stack_timespec_sec(&inode_item->otime);
-	BTRFS_I(inode)->i_otime_nsec = btrfs_stack_timespec_nsec(&inode_item->otime);
+	inode->i_otime_sec = btrfs_stack_timespec_sec(&inode_item->otime);
+	inode->i_otime_nsec = btrfs_stack_timespec_nsec(&inode_item->otime);
 
-	inode->i_generation = BTRFS_I(inode)->generation;
-	if (S_ISDIR(inode->i_mode))
-		BTRFS_I(inode)->index_cnt = (u64)-1;
+	vfs_inode->i_generation = inode->generation;
+	if (S_ISDIR(vfs_inode->i_mode))
+		inode->index_cnt = (u64)-1;
 
 	mutex_unlock(&delayed_node->mutex);
 	btrfs_release_delayed_node(delayed_node);
···
 
 	mutex_lock(&delayed_node->mutex);
 	if (test_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &delayed_node->flags)) {
-		fill_stack_inode_item(trans, &delayed_node->inode_item,
-				      &inode->vfs_inode);
+		fill_stack_inode_item(trans, &delayed_node->inode_item, inode);
 		goto release_node;
 	}
 
···
 	if (ret)
 		goto release_node;
 
-	fill_stack_inode_item(trans, &delayed_node->inode_item, &inode->vfs_inode);
+	fill_stack_inode_item(trans, &delayed_node->inode_item, inode);
 	set_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &delayed_node->flags);
 	delayed_node->count++;
 	atomic_inc(&root->fs_info->delayed_root->items);

fs/btrfs/delayed-inode.h

···
 
 int btrfs_delayed_update_inode(struct btrfs_trans_handle *trans,
 			       struct btrfs_inode *inode);
-int btrfs_fill_inode(struct inode *inode, u32 *rdev);
+int btrfs_fill_inode(struct btrfs_inode *inode, u32 *rdev);
 int btrfs_delayed_delete_inode_ref(struct btrfs_inode *inode);
 
 /* Used for drop dead root */

fs/btrfs/delayed-ref.h

···
 #include <linux/spinlock.h>
 #include <linux/slab.h>
 #include <uapi/linux/btrfs_tree.h>
+#include "fs.h"
+#include "messages.h"
 
 struct btrfs_trans_handle;
 struct btrfs_fs_info;

fs/btrfs/dev-replace.c

···
 	struct extent_buffer *eb;
 	int slot;
 	int ret = 0;
-	struct btrfs_path *path = NULL;
+	BTRFS_PATH_AUTO_FREE(path);
 	int item_size;
 	struct btrfs_dev_replace_item *ptr;
 	u64 src_devid;
···
 		return 0;
 
 	path = btrfs_alloc_path();
-	if (!path) {
-		ret = -ENOMEM;
-		goto out;
-	}
+	if (!path)
+		return -ENOMEM;
 
 	key.objectid = 0;
 	key.type = BTRFS_DEV_REPLACE_KEY;
···
 		if (btrfs_find_device(fs_info->fs_devices, &args)) {
 			btrfs_err(fs_info,
 			"found replace target device without a valid replace item");
-			ret = -EUCLEAN;
-			goto out;
+			return -EUCLEAN;
 		}
-		ret = 0;
 		dev_replace->replace_state =
 			BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED;
 		dev_replace->cont_reading_from_srcdev_mode =
···
 		dev_replace->tgtdev = NULL;
 		dev_replace->is_valid = 0;
 		dev_replace->item_needs_writeback = 0;
-		goto out;
+		return 0;
 	}
 	slot = path->slots[0];
 	eb = path->nodes[0];
···
 		break;
 	}
 
-out:
-	btrfs_free_path(path);
 	return ret;
 }
 
···
 	struct btrfs_fs_info *fs_info = trans->fs_info;
 	int ret;
 	struct btrfs_root *dev_root = fs_info->dev_root;
-	struct btrfs_path *path;
+	BTRFS_PATH_AUTO_FREE(path);
 	struct btrfs_key key;
 	struct extent_buffer *eb;
 	struct btrfs_dev_replace_item *ptr;
···
 	key.offset = 0;
 
 	path = btrfs_alloc_path();
-	if (!path) {
-		ret = -ENOMEM;
-		goto out;
-	}
+	if (!path)
+		return -ENOMEM;
+
 	ret = btrfs_search_slot(trans, dev_root, &key, path, -1, 1);
 	if (ret < 0) {
 		btrfs_warn(fs_info,
 			   "error %d while searching for dev_replace item!",
 			   ret);
-		goto out;
+		return ret;
 	}
 
 	if (ret == 0 &&
···
 			btrfs_warn(fs_info,
 				   "delete too small dev_replace item failed %d!",
 				   ret);
-			goto out;
+			return ret;
 		}
 		ret = 1;
 	}
···
 		if (ret < 0) {
 			btrfs_warn(fs_info,
 				   "insert dev_replace item failed %d!", ret);
-			goto out;
+			return ret;
 		}
 	}
 
···
 		dev_replace->cursor_right);
 	dev_replace->item_needs_writeback = 0;
 	up_write(&dev_replace->rwsem);
-out:
-	btrfs_free_path(path);
 
 	return ret;
 }

fs/btrfs/dir-item.c

···
 	int data_size;
 	struct extent_buffer *leaf;
 	int slot;
-	struct btrfs_path *path;
+	BTRFS_PATH_AUTO_FREE(path);
 
 	path = btrfs_alloc_path();
 	if (!path)
···
 	if (IS_ERR(di)) {
 		ret = PTR_ERR(di);
 		/* Nothing found, we're safe */
-		if (ret == -ENOENT) {
-			ret = 0;
-			goto out;
-		}
+		if (ret == -ENOENT)
+			return 0;
 
 		if (ret < 0)
-			goto out;
+			return ret;
 	}
 
 	/* we found an item, look for our name in the item */
 	if (di) {
 		/* our exact name was found */
-		ret = -EEXIST;
-		goto out;
+		return -EEXIST;
 	}
 
 	/* See if there is room in the item to insert this name. */
···
 	slot = path->slots[0];
 	if (data_size + btrfs_item_size(leaf, slot) +
 	    sizeof(struct btrfs_item) > BTRFS_LEAF_DATA_SIZE(root->fs_info)) {
-		ret = -EOVERFLOW;
-	} else {
-		/* plenty of insertion room */
-		ret = 0;
+		return -EOVERFLOW;
 	}
-out:
-	btrfs_free_path(path);
-	return ret;
+
+	/* Plenty of insertion room. */
+	return 0;
 }
 
 /*

fs/btrfs/dir-item.h

···
 struct btrfs_fs_info;
 struct btrfs_key;
 struct btrfs_path;
+struct btrfs_inode;
 struct btrfs_root;
 struct btrfs_trans_handle;
 

fs/btrfs/direct-io.c

···
 		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) == 1) {
+		if (can_nocow_extent(BTRFS_I(inode), start, &len, &file_extent,
+				     false) == 1) {
 			bg = btrfs_inc_nocow_writers(fs_info, block_start);
 			if (bg)
 				can_nocow = true;
···
 	}
 
 	if (check_direct_IO(fs_info, from, pos)) {
+		btrfs_inode_unlock(BTRFS_I(inode), ilock_flags);
+		goto buffered;
+	}
+	/*
+	 * We can't control the folios being passed in, applications can write
+	 * to them while a direct IO write is in progress.  This means the
+	 * content might change after we calculated the data checksum.
+	 * Therefore we can end up storing a checksum that doesn't match the
+	 * persisted data.
+	 *
+	 * To be extra safe and avoid false data checksum mismatch, if the
+	 * inode requires data checksum, just fallback to buffered IO.
+	 * For buffered IO we have full control of page cache and can ensure
+	 * no one is modifying the content during writeback.
+	 */
+	if (!(BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM)) {
 		btrfs_inode_unlock(BTRFS_I(inode), ilock_flags);
 		goto buffered;
 	}

fs/btrfs/direct-io.h

···
 
 #include <linux/types.h>
 
+struct kiocb;
+
 int __init btrfs_init_dio(void);
 void __cold btrfs_destroy_dio(void);
 

fs/btrfs/discard.c

···
 	block_group->discard_eligible_time = 0;
 	queued = !list_empty(&block_group->discard_list);
 	list_del_init(&block_group->discard_list);
-	/*
-	 * If the block group is currently running in the discard workfn, we
-	 * don't want to deref it, since it's still being used by the workfn.
-	 * The workfn will notice this case and deref the block group when it is
-	 * finished.
-	 */
-	if (queued && !running)
+	if (queued)
 		btrfs_put_block_group(block_group);
 
 	spin_unlock(&discard_ctl->lock);
···
 			block_group->discard_cursor = block_group->start;
 			block_group->discard_state = BTRFS_DISCARD_EXTENTS;
 		}
-		discard_ctl->block_group = block_group;
 	}
 	if (block_group) {
+		btrfs_get_block_group(block_group);
+		discard_ctl->block_group = block_group;
 		*discard_state = block_group->discard_state;
 		*discard_index = block_group->discard_index;
 	}
···
 
 	block_group = peek_discard_list(discard_ctl, &discard_state,
 					&discard_index, now);
-	if (!block_group || !btrfs_run_discard_work(discard_ctl))
+	if (!block_group)
 		return;
+	if (!btrfs_run_discard_work(discard_ctl)) {
+		spin_lock(&discard_ctl->lock);
+		btrfs_put_block_group(block_group);
+		discard_ctl->block_group = NULL;
+		spin_unlock(&discard_ctl->lock);
+		return;
+	}
 	if (now < block_group->discard_eligible_time) {
+		spin_lock(&discard_ctl->lock);
+		btrfs_put_block_group(block_group);
+		discard_ctl->block_group = NULL;
+		spin_unlock(&discard_ctl->lock);
 		btrfs_discard_schedule_work(discard_ctl, false);
 		return;
 	}
···
 	spin_lock(&discard_ctl->lock);
 	discard_ctl->prev_discard = trimmed;
 	discard_ctl->prev_discard_time = now;
-	/*
-	 * If the block group was removed from the discard list while it was
-	 * running in this workfn, then we didn't deref it, since this function
-	 * still owned that reference. But we set the discard_ctl->block_group
-	 * back to NULL, so we can use that condition to know that now we need
-	 * to deref the block_group.
-	 */
-	if (discard_ctl->block_group == NULL)
-		btrfs_put_block_group(block_group);
+	btrfs_put_block_group(block_group);
 	discard_ctl->block_group = NULL;
 	__btrfs_discard_schedule_work(discard_ctl, now, false);
 	spin_unlock(&discard_ctl->lock);

fs/btrfs/discard.h

···
 #ifndef BTRFS_DISCARD_H
 #define BTRFS_DISCARD_H
 
+#include <linux/types.h>
 #include <linux/sizes.h>
 
 struct btrfs_fs_info;

fs/btrfs/disk-io.c

···
 				      int mirror_num)
 {
 	struct btrfs_fs_info *fs_info = eb->fs_info;
-	int num_folios = num_extent_folios(eb);
 	int ret = 0;
 
 	if (sb_rdonly(fs_info->sb))
 		return -EROFS;
 
-	for (int i = 0; i < num_folios; i++) {
+	for (int i = 0; i < num_extent_folios(eb); i++) {
 		struct folio *folio = eb->folios[i];
 		u64 start = max_t(u64, eb->start, folio_pos(folio));
 		u64 end = min_t(u64, eb->start + eb->len,
···
 
 	if (WARN_ON_ONCE(found_start != eb->start))
 		return BLK_STS_IOERR;
-	if (WARN_ON(!btrfs_folio_test_uptodate(fs_info, eb->folios[0],
-					       eb->start, eb->len)))
+	if (WARN_ON(!btrfs_meta_folio_test_uptodate(eb->folios[0], eb)))
 		return BLK_STS_IOERR;
 
 	ASSERT(memcmp_extent_buffer(eb, fs_info->fs_devices->metadata_uuid,
···
 					const struct btrfs_key *key)
 {
 	struct btrfs_root *root;
-	struct btrfs_path *path;
+	BTRFS_PATH_AUTO_FREE(path);
 
 	path = btrfs_alloc_path();
 	if (!path)
 		return ERR_PTR(-ENOMEM);
 	root = read_tree_root_path(tree_root, path, key);
-	btrfs_free_path(path);
 
 	return root;
 }
 
 /*
- * Initialize subvolume root in-memory structure
+ * Initialize subvolume root in-memory structure.
  *
  * @anon_dev:	anonymous device to attach to the root, if zero, allocate new
+ *
+ * In case of failure the caller is responsible to call btrfs_free_fs_root()
  */
 static int btrfs_init_fs_root(struct btrfs_root *root, dev_t anon_dev)
 {
···
 		if (!anon_dev) {
 			ret = get_anon_bdev(&root->anon_dev);
 			if (ret)
-				goto fail;
+				return ret;
 		} else {
 			root->anon_dev = anon_dev;
 		}
···
 	ret = btrfs_init_root_free_objectid(root);
 	if (ret) {
 		mutex_unlock(&root->objectid_mutex);
-		goto fail;
+		return ret;
 	}
 
 	ASSERT(root->free_objectid <= BTRFS_LAST_FREE_OBJECTID);
···
 	mutex_unlock(&root->objectid_mutex);
 
 	return 0;
-fail:
-	/* The caller is responsible to call btrfs_free_fs_root */
-	return ret;
 }
 
 static struct btrfs_root *btrfs_lookup_fs_root(struct btrfs_fs_info *fs_info,
···
 
 static int load_global_roots(struct btrfs_root *tree_root)
 {
-	struct btrfs_path *path;
-	int ret = 0;
+	BTRFS_PATH_AUTO_FREE(path);
+	int ret;
 
 	path = btrfs_alloc_path();
 	if (!path)
···
 	ret = load_global_roots_objectid(tree_root, path,
 					 BTRFS_EXTENT_TREE_OBJECTID, "extent");
 	if (ret)
-		goto out;
+		return ret;
 	ret = load_global_roots_objectid(tree_root, path,
 					 BTRFS_CSUM_TREE_OBJECTID, "csum");
 	if (ret)
-		goto out;
+		return ret;
 	if (!btrfs_fs_compat_ro(tree_root->fs_info, FREE_SPACE_TREE))
-		goto out;
+		return ret;
 	ret = load_global_roots_objectid(tree_root, path,
 					 BTRFS_FREE_SPACE_TREE_OBJECTID,
 					 "free space");
-out:
-	btrfs_free_path(path);
+
 	return ret;
 }
 
···
 	 * Check sectorsize and nodesize first, other check will need it.
 	 * Check all possible sectorsize(4K, 8K, 16K, 32K, 64K) here.
 	 */
-	if (!is_power_of_2(sectorsize) || sectorsize < 4096 ||
+	if (!is_power_of_2(sectorsize) || sectorsize < BTRFS_MIN_BLOCKSIZE ||
 	    sectorsize > BTRFS_MAX_METADATA_BLOCKSIZE) {
 		btrfs_err(fs_info, "invalid sectorsize %llu", sectorsize);
 		ret = -EINVAL;
 	}
 
 	/*
-	 * We only support at most two sectorsizes: 4K and PAGE_SIZE.
+	 * We only support at most 3 sectorsizes: 4K, PAGE_SIZE, MIN_BLOCKSIZE.
+	 *
+	 * For 4K page sized systems with non-debug builds, all 3 matches (4K).
+	 * For 4K page sized systems with debug builds, there are two block sizes
+	 * supported. (4K and 2K)
 	 *
 	 * We can support 16K sectorsize with 64K page size without problem,
 	 * but such sectorsize/pagesize combination doesn't make much sense.
 	 * 4K will be our future standard, PAGE_SIZE is supported from the very
 	 * beginning.
 	 */
-	if (sectorsize > PAGE_SIZE || (sectorsize != SZ_4K && sectorsize != PAGE_SIZE)) {
+	if (sectorsize > PAGE_SIZE || (sectorsize != SZ_4K &&
+				       sectorsize != PAGE_SIZE &&
+				       sectorsize != BTRFS_MIN_BLOCKSIZE)) {
 		btrfs_err(fs_info,
 			"sectorsize %llu not yet supported for page size %lu",
 			sectorsize, PAGE_SIZE);
···
 			  btrfs_super_bytenr(sb), BTRFS_SUPER_INFO_OFFSET);
 		ret = -EINVAL;
 	}
+
+	if (ret)
+		return ret;
 
 	ret = validate_sys_chunk_array(fs_info, sb);
 
···
 	fs_info->nodesize = nodesize;
 	fs_info->sectorsize = sectorsize;
 	fs_info->sectorsize_bits = ilog2(sectorsize);
-	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;
···
 	 * to something non-standard make sure we truncate it to sectorsize.
 	 */
 	fs_info->max_inline = min_t(u64, fs_info->max_inline, fs_info->sectorsize);
-
-	if (sectorsize < PAGE_SIZE)
-		btrfs_warn(fs_info,
-		"read-write for sector size %u with page size %lu is experimental",
-			   sectorsize, PAGE_SIZE);
 
 	ret = btrfs_init_workqueues(fs_info);
 	if (ret)
···
 	btrfs_cleanup_defrag_inodes(fs_info);
 
 	/*
+	 * Handle the error fs first, as it will flush and wait for all ordered
+	 * extents.  This will generate delayed iputs, thus we want to handle
+	 * it first.
+	 */
+	if (unlikely(BTRFS_FS_ERROR(fs_info)))
+		btrfs_error_commit_super(fs_info);
+
+	/*
 	 * Wait for any fixup workers to complete.
 	 * If we don't wait for them here and they are still running by the time
 	 * we call kthread_stop() against the cleaner kthread further below, we
···
 	 * when we call kthread_stop().
 	 */
 	btrfs_flush_workqueue(fs_info->delalloc_workers);
+
+	/*
+	 * We can have ordered extents getting their last reference dropped from
+	 * the fs_info->workers queue because for async writes for data bios we
+	 * queue a work for that queue, at btrfs_wq_submit_bio(), that runs
+	 * run_one_async_done() which calls btrfs_bio_end_io() in case the bio
+	 * has an error, and that later function can do the final
+	 * btrfs_put_ordered_extent() on the ordered extent attached to the bio,
+	 * which adds a delayed iput for the inode. So we must flush the queue
+	 * so that we don't have delayed iputs after committing the current
+	 * transaction below and stopping the cleaner and transaction kthreads.
+	 */
+	btrfs_flush_workqueue(fs_info->workers);
+
+	/*
+	 * When finishing a compressed write bio we schedule a work queue item
+	 * to finish an ordered extent - btrfs_finish_compressed_write_work()
+	 * calls btrfs_finish_ordered_extent() which in turns does a call to
+	 * btrfs_queue_ordered_fn(), and that queues the ordered extent
+	 * completion either in the endio_write_workers work queue or in the
+	 * fs_info->endio_freespace_worker work queue. We flush those queues
+	 * below, so before we flush them we must flush this queue for the
+	 * workers of compressed writes.
+	 */
+	flush_workqueue(fs_info->compressed_write_workers);
 
 	/*
 	 * After we parked the cleaner kthread, ordered extents may have
···
 	/* Ordered extents for free space inodes. */
 	btrfs_flush_workqueue(fs_info->endio_freespace_worker);
 	btrfs_run_delayed_iputs(fs_info);
+	/* There should be no more workload to generate new delayed iputs. */
+	set_bit(BTRFS_FS_STATE_NO_DELAYED_IPUT, &fs_info->fs_state);
 
 	cancel_work_sync(&fs_info->async_reclaim_work);
 	cancel_work_sync(&fs_info->async_data_reclaim_work);
···
 		if (ret)
 			btrfs_err(fs_info, "commit super ret %d", ret);
 	}
-
-	if (BTRFS_FS_ERROR(fs_info))
-		btrfs_error_commit_super(fs_info);
 
 	kthread_stop(fs_info->transaction_kthread);
 	kthread_stop(fs_info->cleaner_kthread);
···
 {
 	/* cleanup FS via transaction */
 	btrfs_cleanup_transaction(fs_info);
-
-	mutex_lock(&fs_info->cleaner_mutex);
-	btrfs_run_delayed_iputs(fs_info);
-	mutex_unlock(&fs_info->cleaner_mutex);
 
 	down_write(&fs_info->cleanup_work_sem);
 	up_write(&fs_info->cleanup_work_sem);
···
 
 int btrfs_init_root_free_objectid(struct btrfs_root *root)
 {
-	struct btrfs_path *path;
+	BTRFS_PATH_AUTO_FREE(path);
 	int ret;
 	struct extent_buffer *l;
 	struct btrfs_key search_key;
···
 	search_key.offset = (u64)-1;
 	ret = btrfs_search_slot(NULL, root, &search_key, path, 0, 0);
 	if (ret < 0)
-		goto error;
+		return ret;
 	if (ret == 0) {
 		/*
 		 * Key with offset -1 found, there would have to exist a root
 		 * with such id, but this is out of valid range.
 		 */
-		ret = -EUCLEAN;
-		goto error;
+		return -EUCLEAN;
 	}
 	if (path->slots[0] > 0) {
 		slot = path->slots[0] - 1;
···
 	} else {
 		root->free_objectid = BTRFS_FIRST_FREE_OBJECTID;
 	}
-	ret = 0;
-error:
-	btrfs_free_path(path);
-	return ret;
+
+	return 0;
 }
 
 int btrfs_get_free_objectid(struct btrfs_root *root, u64 *objectid)

fs/btrfs/export.c

···
 {
 	struct btrfs_fs_info *fs_info = btrfs_sb(sb);
 	struct btrfs_root *root;
-	struct inode *inode;
+	struct btrfs_inode *inode;
 
 	if (objectid < BTRFS_FIRST_FREE_OBJECTID)
 		return ERR_PTR(-ESTALE);
···
 	if (IS_ERR(inode))
 		return ERR_CAST(inode);
 
-	if (generation != 0 && generation != inode->i_generation) {
-		iput(inode);
+	if (generation != 0 && generation != inode->vfs_inode.i_generation) {
+		iput(&inode->vfs_inode);
 		return ERR_PTR(-ESTALE);
 	}
 
-	return d_obtain_alias(inode);
+	return d_obtain_alias(&inode->vfs_inode);
 }
 
 static struct dentry *btrfs_fh_to_parent(struct super_block *sb, struct fid *fh,
···
 
 struct dentry *btrfs_get_parent(struct dentry *child)
 {
-	struct inode *dir = d_inode(child);
-	struct btrfs_fs_info *fs_info = btrfs_sb(dir->i_sb);
-	struct btrfs_root *root = BTRFS_I(dir)->root;
+	struct btrfs_inode *dir = BTRFS_I(d_inode(child));
+	struct btrfs_inode *inode;
+	struct btrfs_root *root = dir->root;
+	struct btrfs_fs_info *fs_info = root->fs_info;
 	struct btrfs_path *path;
 	struct extent_buffer *leaf;
 	struct btrfs_root_ref *ref;
···
 	if (!path)
 		return ERR_PTR(-ENOMEM);
 
-	if (btrfs_ino(BTRFS_I(dir)) == BTRFS_FIRST_FREE_OBJECTID) {
+	if (btrfs_ino(dir) == BTRFS_FIRST_FREE_OBJECTID) {
 		key.objectid = btrfs_root_id(root);
 		key.type = BTRFS_ROOT_BACKREF_KEY;
 		key.offset = (u64)-1;
 		root = fs_info->tree_root;
 	} else {
-		key.objectid = btrfs_ino(BTRFS_I(dir));
+		key.objectid = btrfs_ino(dir);
 		key.type = BTRFS_INODE_REF_KEY;
 		key.offset = (u64)-1;
 	}
···
 					found_key.offset, 0);
 	}
 
-	return d_obtain_alias(btrfs_iget(key.objectid, root));
+	inode = btrfs_iget(key.objectid, root);
+	if (IS_ERR(inode))
+		return ERR_CAST(inode);
+
+	return d_obtain_alias(&inode->vfs_inode);
 fail:
 	btrfs_free_path(path);
 	return ERR_PTR(ret);
···
 static int btrfs_get_name(struct dentry *parent, char *name,
 			  struct dentry *child)
 {
-	struct inode *inode = d_inode(child);
-	struct inode *dir = d_inode(parent);
-	struct btrfs_fs_info *fs_info = inode_to_fs_info(inode);
-	struct btrfs_path *path;
-	struct btrfs_root *root = BTRFS_I(dir)->root;
+	struct btrfs_inode *inode = BTRFS_I(d_inode(child));
+	struct btrfs_inode *dir = BTRFS_I(d_inode(parent));
+	struct btrfs_root *root = dir->root;
+	struct btrfs_fs_info *fs_info = root->fs_info;
+	BTRFS_PATH_AUTO_FREE(path);
 	struct btrfs_inode_ref *iref;
 	struct btrfs_root_ref *rref;
 	struct extent_buffer *leaf;
···
 	int ret;
 	u64 ino;
 
-	if (!S_ISDIR(dir->i_mode))
+	if (!S_ISDIR(dir->vfs_inode.i_mode))
 		return -EINVAL;
 
-	ino = btrfs_ino(BTRFS_I(inode));
+	ino = btrfs_ino(inode);
 
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
 
 	if (ino == BTRFS_FIRST_FREE_OBJECTID) {
-		key.objectid = btrfs_root_id(BTRFS_I(inode)->root);
+		key.objectid = btrfs_root_id(inode->root);
 		key.type = BTRFS_ROOT_BACKREF_KEY;
 		key.offset = (u64)-1;
 		root = fs_info->tree_root;
 	} else {
 		key.objectid = ino;
-		key.offset = btrfs_ino(BTRFS_I(dir));
 		key.type = BTRFS_INODE_REF_KEY;
+		key.offset = btrfs_ino(dir);
 	}
 
 	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
 	if (ret < 0) {
-		btrfs_free_path(path);
 		return ret;
 	} else if (ret > 0) {
-		if (ino == BTRFS_FIRST_FREE_OBJECTID) {
+		if (ino == BTRFS_FIRST_FREE_OBJECTID)
 			path->slots[0]--;
-		} else {
-			btrfs_free_path(path);
+		else
 			return -ENOENT;
-		}
 	}
 	leaf = path->nodes[0];
 
···
 	}
 
 	read_extent_buffer(leaf, name, name_ptr, name_len);
-	btrfs_free_path(path);
 
 	/*
 	 * have to add the null termination to make sure that reconnect_path

fs/btrfs/extent-io-tree.c

···
 	return tree_search_for_insert(tree, offset, NULL, NULL);
 }
 
-static void extent_io_tree_panic(const struct extent_io_tree *tree,
-				 const struct extent_state *state,
-				 const char *opname,
-				 int err)
+static void __cold extent_io_tree_panic(const struct extent_io_tree *tree,
+					const struct extent_state *state,
+					const char *opname,
+					int err)
 {
 	btrfs_panic(extent_io_tree_to_fs_info(tree), err,
 		    "extent io tree error on %s state start %llu end %llu",

fs/btrfs/extent-tree.c

···
 int btrfs_lookup_data_extent(struct btrfs_fs_info *fs_info, u64 start, u64 len)
 {
 	struct btrfs_root *root = btrfs_extent_root(fs_info, start);
-	int ret;
 	struct btrfs_key key;
-	struct btrfs_path *path;
+	BTRFS_PATH_AUTO_FREE(path);
 
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
 
 	key.objectid = start;
-	key.offset = len;
 	key.type = BTRFS_EXTENT_ITEM_KEY;
-	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
-	btrfs_free_path(path);
-	return ret;
+	key.offset = len;
+	return btrfs_search_slot(NULL, root, &key, path, 0, 0);
 }
 
 /*
···
 	struct btrfs_root *extent_root;
 	struct btrfs_delayed_ref_head *head;
 	struct btrfs_delayed_ref_root *delayed_refs;
-	struct btrfs_path *path;
+	BTRFS_PATH_AUTO_FREE(path);
 	struct btrfs_key key;
 	u64 num_refs;
 	u64 extent_flags;
···
 
 search_again:
 	key.objectid = bytenr;
-	key.offset = offset;
 	if (metadata)
 		key.type = BTRFS_METADATA_ITEM_KEY;
 	else
 		key.type = BTRFS_EXTENT_ITEM_KEY;
+	key.offset = offset;
 
 	extent_root = btrfs_extent_root(fs_info, bytenr);
 	ret = btrfs_search_slot(NULL, extent_root, &key, path, 0, 0);
 	if (ret < 0)
-		goto out_free;
+		return ret;
 
 	if (ret > 0 && key.type == BTRFS_METADATA_ITEM_KEY) {
 		if (path->slots[0]) {
···
 			"unexpected extent item size, has %u expect >= %zu",
 				  item_size, sizeof(*ei));
 			btrfs_abort_transaction(trans, ret);
-			goto out_free;
+			return ret;
 		}
 
 		ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
···
 		"unexpected zero reference count for extent item (%llu %u %llu)",
 				  key.objectid, key.type, key.offset);
 			btrfs_abort_transaction(trans, ret);
-			goto out_free;
+			return ret;
 		}
 		extent_flags = btrfs_extent_flags(leaf, ei);
 		owner = btrfs_get_extent_owner_root(fs_info, leaf, path->slots[0]);
···
 		*flags = extent_flags;
 	if (owning_root)
 		*owning_root = owner;
-out_free:
-	btrfs_free_path(path);
+
 	return ret;
 }
 
···
 				  struct btrfs_delayed_ref_node *node,
 				  struct btrfs_delayed_extent_op *extent_op)
 {
-	struct btrfs_path *path;
+	BTRFS_PATH_AUTO_FREE(path);
 	struct extent_buffer *leaf;
 	struct btrfs_extent_item *item;
 	struct btrfs_key key;
···
 					   node->parent, node->ref_root, owner,
 					   offset, refs_to_add, extent_op);
 	if ((ret < 0 && ret != -EAGAIN) || !ret)
-		goto out;
+		return ret;
 
 	/*
 	 * Ok we had -EAGAIN which means we didn't have space to insert and
···
 
 	if (ret)
 		btrfs_abort_transaction(trans, ret);
-out:
-	btrfs_free_path(path);
+
 	return ret;
 }
 
···
 	struct btrfs_fs_info *fs_info = trans->fs_info;
 	struct btrfs_root *root;
 	struct btrfs_key key;
-	struct btrfs_path *path;
+	BTRFS_PATH_AUTO_FREE(path);
 	struct btrfs_extent_item *ei;
 	struct extent_buffer *leaf;
 	u32 item_size;
···
 again:
 	ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
 	if (ret < 0) {
-		goto out;
+		return ret;
 	} else if (ret > 0) {
 		if (metadata) {
 			if (path->slots[0] > 0) {
···
 				metadata = 0;
 
 				key.objectid = head->bytenr;
-				key.offset = head->num_bytes;
 				key.type = BTRFS_EXTENT_ITEM_KEY;
+				key.offset = head->num_bytes;
 				goto again;
 			}
 		} else {
···
 			btrfs_err(fs_info,
 		  "missing extent item for extent %llu num_bytes %llu level %d",
 				  head->bytenr, head->num_bytes, head->level);
-			goto out;
+			return ret;
 		}
 	}
 
···
 			  "unexpected extent item size, has %u expect >= %zu",
 			  item_size, sizeof(*ei));
 		btrfs_abort_transaction(trans, ret);
-		goto out;
+		return ret;
 	}
 
 	ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
 	__run_delayed_extent_op(extent_op, leaf, ei);
-out:
-	btrfs_free_path(path);
+
 	return ret;
 }
 
···
 	int ret;
 
 	key.objectid = bytenr;
-	key.offset = (u64)-1;
 	key.type = BTRFS_EXTENT_ITEM_KEY;
+	key.offset = (u64)-1;
 
 	ret = btrfs_search_slot(NULL, extent_root, &key, path, 0, 0);
 	if (ret < 0)
···
 						   block_group->length,
 						   &trimmed);
 
+		/*
+		 * Not strictly necessary to lock, as the block_group should be
+		 * read-only from btrfs_delete_unused_bgs().
+		 */
+		ASSERT(block_group->ro);
+		spin_lock(&fs_info->unused_bgs_lock);
 		list_del_init(&block_group->bg_list);
+		spin_unlock(&fs_info->unused_bgs_lock);
+
 		btrfs_unfreeze_block_group(block_group);
 		btrfs_put_block_group(block_group);
 
···
 {
 	struct btrfs_delayed_ref_root *delayed_refs;
 	struct btrfs_delayed_ref_head *head;
-	struct btrfs_path *path;
+	BTRFS_PATH_AUTO_FREE(path);
 	struct btrfs_extent_inline_ref *iref;
 	int ret;
 	bool exists = false;
···
 		 * If we get 0 then we found our reference, return 1, else
 		 * return the error if it's not -ENOENT;
 		 */
-		btrfs_free_path(path);
 		return (ret < 0 ) ? ret : 1;
 	}
 
···
 	mutex_unlock(&head->mutex);
 out:
 	spin_unlock(&delayed_refs->lock);
-	btrfs_free_path(path);
 	return exists ? 1 : 0;
 }
 
···
 			struct extent_buffer *parent)
 {
 	struct btrfs_fs_info *fs_info = root->fs_info;
-	struct btrfs_path *path;
+	BTRFS_PATH_AUTO_FREE(path);
 	struct walk_control *wc;
 	int level;
 	int parent_level;
···
 		return -ENOMEM;
 
 	wc = kzalloc(sizeof(*wc), GFP_NOFS);
-	if (!wc) {
-		btrfs_free_path(path);
+	if (!wc)
 		return -ENOMEM;
-	}
 
 	btrfs_assert_tree_write_locked(parent);
 	parent_level = btrfs_header_level(parent);
···
 	}
 
 	kfree(wc);
-	btrfs_free_path(path);
 	return ret;
 }
 

fs/btrfs/extent-tree.h

···
 #define BTRFS_EXTENT_TREE_H
 
 #include <linux/types.h>
-#include "misc.h"
 #include "block-group.h"
 #include "locking.h"
 

fs/btrfs/extent_io.c

···
 	struct btrfs_fs_info *fs_info = folio_to_fs_info(folio);
 
 	ASSERT(folio_pos(folio) <= start &&
-	       start + len <= folio_pos(folio) + PAGE_SIZE);
+	       start + len <= folio_pos(folio) + folio_size(folio));
 
 	if (uptodate && btrfs_verify_folio(folio, start, len))
 		btrfs_folio_set_uptodate(fs_info, folio, start, len);
 	else
 		btrfs_folio_clear_uptodate(fs_info, folio, start, len);
 
-	if (!btrfs_is_subpage(fs_info, folio->mapping))
+	if (!btrfs_is_subpage(fs_info, folio))
 		folio_unlock(folio);
 	else
 		btrfs_folio_end_lock(fs_info, folio, start, len);
···
 static void begin_folio_read(struct btrfs_fs_info *fs_info, struct folio *folio)
 {
 	ASSERT(folio_test_locked(folio));
-	if (!btrfs_is_subpage(fs_info, folio->mapping))
+	if (!btrfs_is_subpage(fs_info, folio))
 		return;
 
 	ASSERT(folio_test_private(folio));
-	btrfs_folio_set_lock(fs_info, folio, folio_pos(folio), PAGE_SIZE);
+	btrfs_folio_set_lock(fs_info, folio, folio_pos(folio), folio_size(folio));
 }
 
 /*
···
 {
 	struct btrfs_inode *inode = folio_to_inode(folio);
 
-	ASSERT(pg_offset + size <= PAGE_SIZE);
+	ASSERT(pg_offset + size <= folio_size(folio));
 	ASSERT(bio_ctrl->end_io_func);
 
 	if (bio_ctrl->bbio &&
···
 	if (folio->mapping)
 		lockdep_assert_held(&folio->mapping->i_private_lock);
 
-	if (fs_info->nodesize >= PAGE_SIZE) {
+	if (!btrfs_meta_is_subpage(fs_info)) {
 		if (!folio_test_private(folio))
 			folio_attach_private(folio, eb);
 		else
···
 
 	fs_info = folio_to_fs_info(folio);
 
-	if (btrfs_is_subpage(fs_info, folio->mapping))
+	if (btrfs_is_subpage(fs_info, folio))
 		return btrfs_attach_subpage(fs_info, folio, BTRFS_SUBPAGE_DATA);
 
 	folio_attach_private(folio, (void *)EXTENT_FOLIO_PRIVATE);
···
 		return;
 
 	fs_info = folio_to_fs_info(folio);
-	if (btrfs_is_subpage(fs_info, folio->mapping))
-		return btrfs_detach_subpage(fs_info, folio);
+	if (btrfs_is_subpage(fs_info, folio))
+		return btrfs_detach_subpage(fs_info, folio, BTRFS_SUBPAGE_DATA);
 
 	folio_detach_private(folio);
 }
···
 	struct inode *inode = folio->mapping->host;
 	struct btrfs_fs_info *fs_info = inode_to_fs_info(inode);
 	u64 start = folio_pos(folio);
-	const u64 end = start + PAGE_SIZE - 1;
-	u64 cur = start;
+	const u64 end = start + folio_size(folio) - 1;
 	u64 extent_offset;
 	u64 last_byte = i_size_read(inode);
-	u64 block_start;
 	struct extent_map *em;
 	int ret = 0;
-	size_t pg_offset = 0;
-	size_t iosize;
-	size_t blocksize = fs_info->sectorsize;
+	const size_t blocksize = fs_info->sectorsize;
 
 	ret = set_folio_extent_mapped(folio);
 	if (ret < 0) {
···
 	if (folio_contains(folio, last_byte >> PAGE_SHIFT)) {
 		size_t zero_offset = offset_in_folio(folio, last_byte);
 
-		if (zero_offset) {
-			iosize = folio_size(folio) - zero_offset;
-			folio_zero_range(folio, zero_offset, iosize);
-		}
+		if (zero_offset)
+			folio_zero_range(folio, zero_offset,
+					 folio_size(folio) - zero_offset);
 	}
 	bio_ctrl->end_io_func = end_bbio_data_read;
 	begin_folio_read(fs_info, folio);
-	while (cur <= end) {
+	for (u64 cur = start; cur <= end; cur += blocksize) {
 		enum btrfs_compression_type compress_type = BTRFS_COMPRESS_NONE;
+		unsigned long pg_offset = offset_in_folio(folio, cur);
 		bool force_bio_submit = false;
 		u64 disk_bytenr;
+		u64 block_start;
 
 		ASSERT(IS_ALIGNED(cur, fs_info->sectorsize));
 		if (cur >= last_byte) {
-			iosize = folio_size(folio) - pg_offset;
-			folio_zero_range(folio, pg_offset, iosize);
-			end_folio_read(folio, true, cur, iosize);
+			folio_zero_range(folio, pg_offset, end - cur + 1);
+			end_folio_read(folio, true, cur, end - cur + 1);
 			break;
+		}
+		if (btrfs_folio_test_uptodate(fs_info, folio, cur, blocksize)) {
+			end_folio_read(folio, true, cur, blocksize);
+			continue;
 		}
 		em = get_extent_map(BTRFS_I(inode), folio, cur, end - cur + 1, em_cached);
 		if (IS_ERR(em)) {
···
 
 		compress_type = extent_map_compression(em);
 
-		iosize = min(extent_map_end(em) - cur, end - cur + 1);
-		iosize = ALIGN(iosize, blocksize);
 		if (compress_type != BTRFS_COMPRESS_NONE)
 			disk_bytenr = em->disk_bytenr;
 		else
 			disk_bytenr = extent_map_block_start(em) + extent_offset;
-		block_start = extent_map_block_start(em);
+
 		if (em->flags & EXTENT_FLAG_PREALLOC)
 			block_start = EXTENT_MAP_HOLE;
+		else
+			block_start = extent_map_block_start(em);
 
 		/*
 		 * If we have a file range that points to a compressed extent
···
 
 		/* we've found a hole, just zero and go on */
 		if (block_start == EXTENT_MAP_HOLE) {
-			folio_zero_range(folio, pg_offset, iosize);
-
-			end_folio_read(folio, true, cur, iosize);
-			cur = cur + iosize;
-			pg_offset += iosize;
+			folio_zero_range(folio, pg_offset, blocksize);
+			end_folio_read(folio, true, cur, blocksize);
 			continue;
 		}
 		/* the get_extent function already copied into the folio */
 		if (block_start == EXTENT_MAP_INLINE) {
-			end_folio_read(folio, true, cur, iosize);
-			cur = cur + iosize;
-			pg_offset += iosize;
+			end_folio_read(folio, true, cur, blocksize);
 			continue;
 		}
 
···
 
 		if (force_bio_submit)
 			submit_one_bio(bio_ctrl);
-		submit_extent_folio(bio_ctrl, disk_bytenr, folio, iosize,
+		submit_extent_folio(bio_ctrl, disk_bytenr, folio, blocksize,
 				    pg_offset);
-		cur = cur + iosize;
-		pg_offset += iosize;
+	}
+	return 0;
+}
+
+/*
+ * Check if we can skip waiting the @ordered extent covering the block at @fileoff.
+ *
+ * @fileoff:	Both input and output.
+ *		Input as the file offset where the check should start at.
+ *		Output as where the next check should start at,
+ *		if the function returns true.
+ *
+ * Return true if we can skip to @fileoff. The caller needs to check the new
+ * @fileoff value to make sure it covers the full range, before skipping the
+ * full OE.
+ *
+ * Return false if we must wait for the ordered extent.
+ */
+static bool can_skip_one_ordered_range(struct btrfs_inode *inode,
+				       struct btrfs_ordered_extent *ordered,
+				       u64 *fileoff)
+{
+	const struct btrfs_fs_info *fs_info = inode->root->fs_info;
+	struct folio *folio;
+	const u32 blocksize = fs_info->sectorsize;
+	u64 cur = *fileoff;
+	bool ret;
+
+	folio = filemap_get_folio(inode->vfs_inode.i_mapping, cur >> PAGE_SHIFT);
+
+	/*
+	 * We should have locked the folio(s) for range [start, end], thus
+	 * there must be a folio and it must be locked.
+	 */
+	ASSERT(!IS_ERR(folio));
+	ASSERT(folio_test_locked(folio));
+
+	/*
+	 * There are several cases for the folio and OE combination:
+	 *
+	 * 1) Folio has no private flag
+	 *    The OE has all its IO done but not yet finished, and folio got
+	 *    invalidated.
+	 *
+	 * Have we have to wait for the OE to finish, as it may contain the
+	 * to-be-inserted data checksum.
+	 * Without the data checksum inserted into the csum tree, read will
+	 * just fail with missing csum.
+	 */
+	if (!folio_test_private(folio)) {
+		ret = false;
+		goto out;
 	}
 
-	return 0;
+	/*
+	 * 2) The first block is DIRTY.
+	 *
+	 * This means the OE is created by some other folios whose file pos is
+	 * before this one. And since we are holding the folio lock, the writeback
+	 * of this folio cannot start.
+	 *
+	 * We must skip the whole OE, because it will never start until we
+	 * finished our folio read and unlocked the folio.
+	 */
+	if (btrfs_folio_test_dirty(fs_info, folio, cur, blocksize)) {
+		u64 range_len = min(folio_pos(folio) + folio_size(folio),
+				    ordered->file_offset + ordered->num_bytes) - cur;
+
+		ret = true;
+		/*
+		 * At least inside the folio, all the remaining blocks should
+		 * also be dirty.
+		 */
+		ASSERT(btrfs_folio_test_dirty(fs_info, folio, cur, range_len));
+		*fileoff = ordered->file_offset + ordered->num_bytes;
+		goto out;
+	}
+
+	/*
+	 * 3) The first block is uptodate.
+	 *
+	 * At least the first block can be skipped, but we are still not fully
+	 * sure. E.g. if the OE has some other folios in the range that cannot
+	 * be skipped.
+	 * So we return true and update @next_ret to the OE/folio boundary.
+	 */
+	if (btrfs_folio_test_uptodate(fs_info, folio, cur, blocksize)) {
+		u64 range_len = min(folio_pos(folio) + folio_size(folio),
+				    ordered->file_offset + ordered->num_bytes) - cur;
+
+		/*
+		 * The whole range to the OE end or folio boundary should also
+		 * be uptodate.
+		 */
+		ASSERT(btrfs_folio_test_uptodate(fs_info, folio, cur, range_len));
+		ret = true;
+		*fileoff = cur + range_len;
+		goto out;
+	}
+
+	/*
+	 * 4) The first block is not uptodate.
+	 *
+	 * This means the folio is invalidated after the writeback was finished,
+	 * but by some other operations (e.g. block aligned buffered write) the
+	 * folio is inserted into filemap.
+	 * Very much the same as case 1).
+	 */
+	ret = false;
+out:
+	folio_put(folio);
+	return ret;
+}
+
+static bool can_skip_ordered_extent(struct btrfs_inode *inode,
+				    struct btrfs_ordered_extent *ordered,
+				    u64 start, u64 end)
+{
+	const u64 range_end = min(end, ordered->file_offset + ordered->num_bytes - 1);
+	u64 cur = max(start, ordered->file_offset);
+
+	while (cur < range_end) {
+		bool can_skip;
+
+		can_skip = can_skip_one_ordered_range(inode, ordered, &cur);
+		if (!can_skip)
+			return false;
+	}
+	return true;
+}
+
+/*
+ * Locking helper to make sure we get a stable view of extent maps for the
+ * involved range.
+ *
+ * This is for folio read paths (read and readahead), thus the involved range
+ * should have all the folios locked.
+ */
+static void lock_extents_for_read(struct btrfs_inode *inode, u64 start, u64 end,
+				  struct extent_state **cached_state)
+{
+	u64 cur_pos;
+
+	/* Caller must provide a valid @cached_state. */
+	ASSERT(cached_state);
+
+	/* The range must at least be page aligned, as all read paths are folio based. */
+	ASSERT(IS_ALIGNED(start, PAGE_SIZE));
+	ASSERT(IS_ALIGNED(end + 1, PAGE_SIZE));
+
+again:
+	lock_extent(&inode->io_tree, start, end, cached_state);
+	cur_pos = start;
+	while (cur_pos < end) {
+		struct btrfs_ordered_extent *ordered;
+
+		ordered = btrfs_lookup_ordered_range(inode, cur_pos,
+						     end - cur_pos + 1);
+		/*
+		 * No ordered extents in the range, and we hold the extent lock,
+		 * no one can modify the extent maps in the range, we're safe to return.
+		 */
+		if (!ordered)
+			break;
+
+		/* Check if we can skip waiting for the whole OE. */
+		if (can_skip_ordered_extent(inode, ordered, start, end)) {
+			cur_pos = min(ordered->file_offset + ordered->num_bytes,
+				      end + 1);
+			btrfs_put_ordered_extent(ordered);
+			continue;
+		}
+
+		/* Now wait for the OE to finish. */
+		unlock_extent(&inode->io_tree, start, end, cached_state);
+		btrfs_start_ordered_extent_nowriteback(ordered, start, end + 1 - start);
+		btrfs_put_ordered_extent(ordered);
+		/* We have unlocked the whole range, restart from the beginning. */
+		goto again;
+	}
 }
 
 int btrfs_read_folio(struct file *file, struct folio *folio)
···
 	struct extent_map *em_cached = NULL;
 	int ret;
 
-	btrfs_lock_and_flush_ordered_range(inode, start, end, &cached_state);
+	lock_extents_for_read(inode, start, end, &cached_state);
 	ret = btrfs_do_readpage(folio, &em_cached, &bio_ctrl, NULL);
 	unlock_extent(&inode->io_tree, start, end, &cached_state);
 
···
 	unsigned int start_bit;
 	unsigned int nbits;
 
-	ASSERT(start >= folio_start && start + len <= folio_start + PAGE_SIZE);
+	ASSERT(start >= folio_start && start + len <= folio_start + folio_size(folio));
 	start_bit = (start - folio_start) >> fs_info->sectorsize_bits;
 	nbits = len >> fs_info->sectorsize_bits;
 	ASSERT(bitmap_test_range_all_zero(delalloc_bitmap, start_bit, nbits));
···
 {
 	struct btrfs_fs_info *fs_info = folio_to_fs_info(folio);
 	const u64 folio_start = folio_pos(folio);
-	const unsigned int bitmap_size = fs_info->sectors_per_page;
+	const unsigned int bitmap_size = btrfs_blocks_per_folio(fs_info, folio);
 	unsigned int start_bit;
 	unsigned int first_zero;
 	unsigned int first_set;
 
-	ASSERT(start >= folio_start && start < folio_start + PAGE_SIZE);
+	ASSERT(start >= folio_start && start < folio_start + folio_size(folio));
 
 	start_bit = (start - folio_start) >> fs_info->sectorsize_bits;
 	first_set = find_next_bit(delalloc_bitmap, bitmap_size, start_bit);
···
 {
 	struct btrfs_fs_info *fs_info = inode_to_fs_info(&inode->vfs_inode);
 	struct writeback_control *wbc = bio_ctrl->wbc;
-	const bool is_subpage = btrfs_is_subpage(fs_info, folio->mapping);
+	const bool is_subpage = btrfs_is_subpage(fs_info, folio);
 	const u64 page_start = folio_pos(folio);
 	const u64 page_end = page_start + folio_size(folio) - 1;
+	const unsigned int blocks_per_folio = btrfs_blocks_per_folio(fs_info, folio);
 	unsigned long delalloc_bitmap = 0;
 	/*
 	 * Save the last found delalloc end. As the delalloc end can go beyond
···
 	int bit;
 
 	/* Save the dirty bitmap as our submission bitmap will be a subset of it. */
-	if (btrfs_is_subpage(fs_info, inode->vfs_inode.i_mapping)) {
-		ASSERT(fs_info->sectors_per_page > 1);
+	if (btrfs_is_subpage(fs_info, folio)) {
+		ASSERT(blocks_per_folio > 1);
 		btrfs_get_subpage_dirty_bitmap(fs_info, folio, &bio_ctrl->submit_bitmap);
 	} else {
 		bio_ctrl->submit_bitmap = 1;
 	}
 
-	for_each_set_bit(bit, &bio_ctrl->submit_bitmap, fs_info->sectors_per_page) {
+	for_each_set_bit(bit, &bio_ctrl->submit_bitmap, blocks_per_folio) {
 		u64 start = page_start + (bit << fs_info->sectorsize_bits);
 
 		btrfs_folio_set_lock(fs_info, folio, start, fs_info->sectorsize);
···
 					     btrfs_root_id(inode->root),
 					     btrfs_ino(inode),
 					     folio_pos(folio),
-					     fs_info->sectors_per_page,
+					     blocks_per_folio,
 					     &bio_ctrl->submit_bitmap,
 					     found_start, found_len, ret);
 		} else {
···
 		unsigned int bitmap_size = min(
 				(last_finished_delalloc_end - page_start) >>
 				fs_info->sectorsize_bits,
-				fs_info->sectors_per_page);
+				blocks_per_folio);
 
 		for_each_set_bit(bit, &bio_ctrl->submit_bitmap, bitmap_size)
 			btrfs_mark_ordered_io_finished(inode, folio,
···
 		delalloc_end = page_end;
 	/*
 	 * delalloc_end is already one less than the total length, so
-	 * we don't subtract one from PAGE_SIZE
+	 * we don't subtract one from PAGE_SIZE.
 	 */
 	delalloc_to_write +=
 		DIV_ROUND_UP(delalloc_end + 1 - page_start, PAGE_SIZE);
···
 	 * If all ranges are submitted asynchronously, we just need to account
 	 * for them here.
 	 */
-	if (bitmap_empty(&bio_ctrl->submit_bitmap, fs_info->sectors_per_page)) {
+	if (bitmap_empty(&bio_ctrl->submit_bitmap, blocks_per_folio)) {
 		wbc->nr_to_write -= delalloc_to_write;
 		return 1;
 	}
···
 	bool submitted_io = false;
 	bool error = false;
 	const u64 folio_start = folio_pos(folio);
+	const unsigned int blocks_per_folio = btrfs_blocks_per_folio(fs_info, folio);
 	u64 cur;
 	int bit;
 	int ret = 0;
···
 	       start + len <= folio_start + folio_size(folio));
 
 	ret = btrfs_writepage_cow_fixup(folio);
-	if (ret) {
+	if (ret == -EAGAIN) {
 		/* Fixup worker will requeue */
 		folio_redirty_for_writepage(bio_ctrl->wbc, folio);
 		folio_unlock(folio);
 		return 1;
 	}
+	if (ret < 0)
+		return ret;
 
 	for (cur = start; cur < start + len; cur += fs_info->sectorsize)
 		set_bit((cur - folio_start) >> fs_info->sectorsize_bits, &range_bitmap);
 	bitmap_and(&bio_ctrl->submit_bitmap, &bio_ctrl->submit_bitmap, &range_bitmap,
-		   fs_info->sectors_per_page);
+		   blocks_per_folio);
 
 	bio_ctrl->end_io_func = end_bbio_data_write;
 
-	for_each_set_bit(bit, &bio_ctrl->submit_bitmap, fs_info->sectors_per_page) {
+	for_each_set_bit(bit, &bio_ctrl->submit_bitmap, blocks_per_folio) {
 		cur = folio_pos(folio) + (bit << fs_info->sectorsize_bits);
 
 		if (cur >= i_size) {
···
 	size_t pg_offset;
 	loff_t i_size = i_size_read(&inode->vfs_inode);
 	unsigned long end_index = i_size >> PAGE_SHIFT;
+	const unsigned int blocks_per_folio = btrfs_blocks_per_folio(fs_info, folio);
 
 	trace_extent_writepage(folio, &inode->vfs_inode, bio_ctrl->wbc);
 
···
 	 * The proper bitmap can only be initialized until writepage_delalloc().
 	 */
 	bio_ctrl->submit_bitmap = (unsigned long)-1;
+
+	/*
+	 * If the page is dirty but without private set, it's marked dirty
+	 * without informing the fs.
+	 * Nowadays that is a bug, since the introduction of
+	 * pin_user_pages*().
+	 *
+	 * So here we check if the page has private set to rule out such
+	 * case.
+	 * But we also have a long history of relying on the COW fixup,
+	 * so here we only enable this check for experimental builds until
+	 * we're sure it's safe.
+	 */
+	if (IS_ENABLED(CONFIG_BTRFS_EXPERIMENTAL) &&
+	    unlikely(!folio_test_private(folio))) {
+		WARN_ON(IS_ENABLED(CONFIG_BTRFS_DEBUG));
+		btrfs_err_rl(fs_info,
+	"root %lld ino %llu folio %llu is marked dirty without notifying the fs",
+			     inode->root->root_key.objectid,
+			     btrfs_ino(inode), folio_pos(folio));
+		ret = -EUCLEAN;
+		goto done;
+	}
+
 	ret = set_folio_extent_mapped(folio);
 	if (ret < 0)
 		goto done;
···
 		goto done;
 
 	ret = extent_writepage_io(inode, folio, folio_pos(folio),
-				  PAGE_SIZE, bio_ctrl, i_size);
+				  folio_size(folio), bio_ctrl, i_size);
 	if (ret == 1)
 		return 0;
 	if (ret < 0)
 		btrfs_err_rl(fs_info,
 "failed to submit blocks, root=%lld inode=%llu folio=%llu submit_bitmap=%*pbl: %d",
 			     btrfs_root_id(inode->root), btrfs_ino(inode),
-			     folio_pos(folio), fs_info->sectors_per_page,
+			     folio_pos(folio), blocks_per_folio,
 			     &bio_ctrl->submit_bitmap, ret);
 
 	bio_ctrl->wbc->nr_to_write--;
···
 static void end_bbio_meta_write(struct btrfs_bio *bbio)
 {
 	struct extent_buffer *eb = bbio->private;
-	struct btrfs_fs_info *fs_info = eb->fs_info;
 	struct folio_iter fi;
-	u32 bio_offset = 0;
 
 	if (bbio->bio.bi_status != BLK_STS_OK)
 		set_btree_ioerr(eb);
 
 	bio_for_each_folio_all(fi, &bbio->bio) {
-		u64 start = eb->start + bio_offset;
-		struct folio *folio = fi.folio;
-		u32 len = fi.length;
-
-		btrfs_folio_clear_writeback(fs_info, folio, start, len);
-		bio_offset += len;
+		btrfs_meta_folio_clear_writeback(fi.folio, eb);
 	}
 
 	clear_bit(EXTENT_BUFFER_WRITEBACK, &eb->bflags);
···
 	wbc_init_bio(wbc, &bbio->bio);
 	bbio->inode = BTRFS_I(eb->fs_info->btree_inode);
 	bbio->file_offset = eb->start;
-	if (fs_info->nodesize < PAGE_SIZE) {
-		struct folio *folio = eb->folios[0];
-		bool ret;
+	for (int i = 0; i < num_extent_folios(eb); i++) {
+		struct folio *folio = eb->folios[i];
+		u64 range_start = max_t(u64, eb->start, folio_pos(folio));
+		u32 range_len = min_t(u64, folio_pos(folio) + folio_size(folio),
+				      eb->start + eb->len) - range_start;
 
 		folio_lock(folio);
-		btrfs_subpage_set_writeback(fs_info, folio, eb->start, eb->len);
-		if (btrfs_subpage_clear_and_test_dirty(fs_info, folio, eb->start,
-						       eb->len)) {
-			folio_clear_dirty_for_io(folio);
-			wbc->nr_to_write--;
-		}
-		ret = bio_add_folio(&bbio->bio, folio, eb->len,
-				    eb->start - folio_pos(folio));
-		ASSERT(ret);
-		wbc_account_cgroup_owner(wbc, folio, eb->len);
-		folio_unlock(folio);
-	} else {
-		int num_folios = num_extent_folios(eb);
-
-		for (int i = 0; i < num_folios; i++) {
-			struct folio *folio = eb->folios[i];
-			bool ret;
-
-			folio_lock(folio);
-			folio_clear_dirty_for_io(folio);
-			folio_start_writeback(folio);
-			ret = bio_add_folio(&bbio->bio, folio, eb->folio_size, 0);
-			ASSERT(ret);
-			wbc_account_cgroup_owner(wbc, folio, eb->folio_size);
+		btrfs_meta_folio_clear_dirty(folio, eb);
+		btrfs_meta_folio_set_writeback(folio, eb);
+		if (!folio_test_dirty(folio))
 			wbc->nr_to_write -= folio_nr_pages(folio);
-			folio_unlock(folio);
-		}
+		bio_add_folio_nofail(&bbio->bio, folio, range_len,
+				     offset_in_folio(folio, range_start));
+		wbc_account_cgroup_owner(wbc, folio, range_len);
+		folio_unlock(folio);
 	}
 	btrfs_submit_bbio(bbio, 0);
 }
···
 	u64 folio_start = folio_pos(folio);
 	int bit_start = 0;
 	int sectors_per_node = fs_info->nodesize >> fs_info->sectorsize_bits;
+	const unsigned int blocks_per_folio = btrfs_blocks_per_folio(fs_info, folio);
 
 	/* Lock and write each dirty extent buffers in the range */
-	while (bit_start < fs_info->sectors_per_page) {
+	while (bit_start < blocks_per_folio) {
 		struct btrfs_subpage *subpage = folio_get_private(folio);
 		struct extent_buffer *eb;
 		unsigned long flags;
···
 			break;
 		}
 		spin_lock_irqsave(&subpage->lock, flags);
-		if (!test_bit(bit_start + btrfs_bitmap_nr_dirty * fs_info->sectors_per_page,
+		if (!test_bit(bit_start + btrfs_bitmap_nr_dirty * blocks_per_folio,
 			      subpage->bitmaps)) {
 			spin_unlock_irqrestore(&subpage->lock, flags);
 			spin_unlock(&folio->mapping->i_private_lock);
···
 	if (!folio_test_private(folio))
 		return 0;
 
-	if (folio_to_fs_info(folio)->nodesize < PAGE_SIZE)
+	if (btrfs_meta_is_subpage(folio_to_fs_info(folio)))
 		return submit_eb_subpage(folio, wbc);
 
 	spin_lock(&mapping->i_private_lock);
···
 			done_index = folio_next_index(folio);
 			/*
 			 * At this point we hold neither the i_pages lock nor
-			 * the page lock: the page may be truncated or
-			 * invalidated (changing page->mapping to NULL),
-			 * or even swizzled back from swapper_space to
-			 * tmpfs file mapping
+			 * the folio lock: the folio may be truncated or
+			 * invalidated (changing folio->mapping to NULL).
 			 */
 			if (!folio_trylock(folio)) {
 				submit_write_bio(bio_ctrl, 0);
···
 			 * regular submission.
 			 */
 			if (wbc->sync_mode != WB_SYNC_NONE ||
-			    btrfs_is_subpage(inode_to_fs_info(inode), mapping)) {
+			    btrfs_is_subpage(inode_to_fs_info(inode), folio)) {
 				if (folio_test_writeback(folio))
 					submit_write_bio(bio_ctrl, 0);
 				folio_wait_writeback(folio);
···
 	ASSERT(IS_ALIGNED(start, sectorsize) && IS_ALIGNED(end + 1, sectorsize));
 
 	while (cur <= end) {
-		u64 cur_end = min(round_down(cur, PAGE_SIZE) + PAGE_SIZE - 1, end);
-		u32 cur_len = cur_end + 1 - cur;
+		u64 cur_end;
+		u32 cur_len;
 		struct folio *folio;
 
 		folio = filemap_get_folio(mapping, cur >> PAGE_SHIFT);
···
 		 * code is just in case, but shouldn't actually be run.
 		 */
 		if (IS_ERR(folio)) {
+			cur_end = min(round_down(cur, PAGE_SIZE) + PAGE_SIZE - 1, end);
+			cur_len = cur_end + 1 - cur;
 			btrfs_mark_ordered_io_finished(BTRFS_I(inode), NULL,
 						       cur, cur_len, false);
 			mapping_set_error(mapping, PTR_ERR(folio));
-			cur = cur_end + 1;
+			cur = cur_end;
 			continue;
 		}
+
+		cur_end = min_t(u64, folio_pos(folio) + folio_size(folio) - 1, end);
+		cur_len = cur_end + 1 - cur;
 
 		ASSERT(folio_test_locked(folio));
 		if (pages_dirty && folio != locked_folio)
···
 	struct extent_map *em_cached = NULL;
 	u64 prev_em_start = (u64)-1;
 
-	btrfs_lock_and_flush_ordered_range(inode, start, end, &cached_state);
+	lock_extents_for_read(inode, start, end, &cached_state);
 
 	while ((folio = readahead_folio(rac)) != NULL)
 		btrfs_do_readpage(folio, &em_cached, &bio_ctrl, &prev_em_start);
···
 				     struct folio *folio)
 {
 	u64 start = folio_pos(folio);
-	u64 end = start + PAGE_SIZE - 1;
+	u64 end = start + folio_size(folio) - 1;
 	bool ret;
 
 	if (test_range_bit_exists(tree, start, end, EXTENT_LOCKED)) {
···
 bool try_release_extent_mapping(struct folio *folio, gfp_t mask)
 {
 	u64 start = folio_pos(folio);
-	u64 end = start + PAGE_SIZE - 1;
+	u64 end = start + folio_size(folio) - 1;
 	struct btrfs_inode *inode = folio_to_inode(folio);
 	struct extent_io_tree *io_tree = &inode->io_tree;
 
···
 		return;
 	}
 
-	if (fs_info->nodesize >= PAGE_SIZE) {
+	if (!btrfs_meta_is_subpage(fs_info)) {
 		/*
 		 * We do this since we'll remove the pages after we've
 		 * removed the eb from the radix tree, so we could race
···
 	 * attached to one dummy eb, no sharing.
 	 */
 	if (!mapped) {
-		btrfs_detach_subpage(fs_info, folio);
+		btrfs_detach_subpage(fs_info, folio, BTRFS_SUBPAGE_METADATA);
 		return;
 	}
 
···
 	 * page range and no unfinished IO.
 	 */
 	if (!folio_range_has_eb(folio))
-		btrfs_detach_subpage(fs_info, folio);
+		btrfs_detach_subpage(fs_info, folio, BTRFS_SUBPAGE_METADATA);
 
 	spin_unlock(&folio->mapping->i_private_lock);
 }
···
 	kmem_cache_free(extent_buffer_cache, eb);
 }
 
-static struct extent_buffer *
-__alloc_extent_buffer(struct btrfs_fs_info *fs_info, u64 start,
-		      unsigned long len)
+static struct extent_buffer *__alloc_extent_buffer(struct btrfs_fs_info *fs_info,
+						   u64 start)
 {
 	struct extent_buffer *eb = NULL;
 
 	eb = kmem_cache_zalloc(extent_buffer_cache, GFP_NOFS|__GFP_NOFAIL);
 	eb->start = start;
-	eb->len = len;
+	eb->len = fs_info->nodesize;
 	eb->fs_info = fs_info;
 	init_rwsem(&eb->lock);
 
···
 	spin_lock_init(&eb->refs_lock);
 	atomic_set(&eb->refs, 1);
 
-	ASSERT(len <= BTRFS_MAX_METADATA_BLOCKSIZE);
+	ASSERT(eb->len <= BTRFS_MAX_METADATA_BLOCKSIZE);
 
 	return eb;
 }
···
 struct extent_buffer *btrfs_clone_extent_buffer(const struct extent_buffer *src)
 {
 	struct extent_buffer *new;
-	int num_folios = num_extent_folios(src);
 	int ret;
 
-	new = __alloc_extent_buffer(src->fs_info, src->start, src->len);
+	new = __alloc_extent_buffer(src->fs_info, src->start);
 	if (new == NULL)
 		return NULL;
 
···
 		return NULL;
 	}
 
-	for (int i = 0; i < num_folios; i++) {
+	for (int i = 0; i < num_extent_folios(src); i++) {
 		struct folio *folio = new->folios[i];
 
 		ret = attach_extent_buffer_folio(new, folio, NULL);
···
 	return new;
 }
 
-struct extent_buffer *__alloc_dummy_extent_buffer(struct btrfs_fs_info *fs_info,
-						  u64 start, unsigned long len)
+struct extent_buffer *alloc_dummy_extent_buffer(struct btrfs_fs_info *fs_info,
+						u64 start)
 {
 	struct extent_buffer *eb;
-	int num_folios = 0;
 	int ret;
 
-	eb = __alloc_extent_buffer(fs_info, start, len);
+	eb = __alloc_extent_buffer(fs_info, start);
 	if (!eb)
 		return NULL;
 
 	ret = alloc_eb_folio_array(eb, false);
 	if (ret)
-		goto err;
+		goto out;
 
-	num_folios = num_extent_folios(eb);
-	for (int i = 0; i < num_folios; i++) {
+	for (int i = 0; i < num_extent_folios(eb); i++) {
 		ret = attach_extent_buffer_folio(eb, eb->folios[i], NULL);
 		if (ret < 0)
-			goto err;
+			goto out_detach;
 	}
 
 	set_extent_buffer_uptodate(eb);
···
 	set_bit(EXTENT_BUFFER_UNMAPPED, &eb->bflags);
 
 	return eb;
-err:
-	for (int i = 0; i < num_folios; i++) {
+
+out_detach:
+	for (int i = 0; i < num_extent_folios(eb); i++) {
 		if (eb->folios[i]) {
 			detach_extent_buffer_folio(eb, eb->folios[i]);
 			folio_put(eb->folios[i]);
 		}
 	}
+out:
 	kmem_cache_free(extent_buffer_cache, eb);
 	return NULL;
-}
-
-struct extent_buffer *alloc_dummy_extent_buffer(struct btrfs_fs_info *fs_info,
-						u64 start)
-{
-	return __alloc_dummy_extent_buffer(fs_info, start, fs_info->nodesize);
 }
 
 static void check_buffer_tree_ref(struct extent_buffer *eb)
···
 
 static void mark_extent_buffer_accessed(struct extent_buffer *eb)
 {
-	int num_folios= num_extent_folios(eb);
-
 	check_buffer_tree_ref(eb);
 
-	for (int i = 0; i < num_folios; i++)
+	for (int i = 0; i < num_extent_folios(eb); i++)
 		folio_mark_accessed(eb->folios[i]);
 }
 
···
 	return eb;
 }
 
-#ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS
 struct extent_buffer *alloc_test_extent_buffer(struct btrfs_fs_info *fs_info,
 					u64 start)
 {
+#ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS
 	struct extent_buffer *eb, *exists = NULL;
 	int ret;
 
···
 free_eb:
 	btrfs_release_extent_buffer(eb);
 	return exists;
-}
+#else
+	/* Stub to avoid linker error when compiled with optimizations turned off. */
+	return NULL;
 #endif
+}
 
 static struct extent_buffer *grab_extent_buffer(struct btrfs_fs_info *fs_info,
 						struct folio *folio)
···
 	 * don't try to insert two ebs for the same bytenr.  So here we always
 	 * return NULL and just continue.
 	 */
-	if (fs_info->nodesize < PAGE_SIZE)
+	if (btrfs_meta_is_subpage(fs_info))
 		return NULL;
 
 	/* Page not yet attached to an extent buffer */
···
 
 finish:
 	spin_lock(&mapping->i_private_lock);
-	if (existing_folio && fs_info->nodesize < PAGE_SIZE) {
+	if (existing_folio && btrfs_meta_is_subpage(fs_info)) {
 		/* We're going to reuse the existing page, can drop our folio now. */
 		__free_page(folio_page(eb->folios[i], 0));
 		eb->folios[i] = existing_folio;
···
 struct extent_buffer *alloc_extent_buffer(struct btrfs_fs_info *fs_info,
 					  u64 start, u64 owner_root, int level)
 {
-	unsigned long len = fs_info->nodesize;
-	int num_folios;
 	int attached = 0;
 	struct extent_buffer *eb;
 	struct extent_buffer *existing_eb = NULL;
···
 	if (eb)
 		return eb;
 
-	eb = __alloc_extent_buffer(fs_info, start, len);
+	eb = __alloc_extent_buffer(fs_info, start);
 	if (!eb)
 		return ERR_PTR(-ENOMEM);
 
···
 	 * The memory will be freed by attach_extent_buffer_page() or freed
 	 * manually if we exit earlier.
 	 */
-	if (fs_info->nodesize < PAGE_SIZE) {
-		prealloc = btrfs_alloc_subpage(fs_info, BTRFS_SUBPAGE_METADATA);
+	if (btrfs_meta_is_subpage(fs_info)) {
+		prealloc = btrfs_alloc_subpage(fs_info, PAGE_SIZE, BTRFS_SUBPAGE_METADATA);
 		if (IS_ERR(prealloc)) {
 			ret = PTR_ERR(prealloc);
 			goto out;
···
 		goto out;
 	}
 
-	num_folios = num_extent_folios(eb);
 	/* Attach all pages to the filemap. */
-	for (int i = 0; i < num_folios; i++) {
+	for (int i = 0; i < num_extent_folios(eb); i++) {
 		struct folio *folio;
 
 		ret = attach_eb_folio_to_filemap(eb, i, prealloc, &existing_eb);
···
 		 * and free the allocated page.
 		 */
 		folio = eb->folios[i];
-		WARN_ON(btrfs_folio_test_dirty(fs_info, folio, eb->start, eb->len));
+		WARN_ON(btrfs_meta_folio_test_dirty(folio, eb));
 
 		/*
 		 * Check if the current page is physically contiguous with previous eb
···
 		if (i && folio_page(eb->folios[i - 1], 0) + 1 != folio_page(folio, 0))
 			page_contig = false;
 
-		if (!btrfs_folio_test_uptodate(fs_info, folio, eb->start, eb->len))
+		if (!btrfs_meta_folio_test_uptodate(folio, eb))
 			uptodate = 0;
 
 		/*
···
 	 * btree_release_folio will correctly detect that a page belongs to a
 	 * live buffer and won't free them prematurely.
 	 */
-	for (int i = 0; i < num_folios; i++)
+	for (int i = 0; i < num_extent_folios(eb); i++)
 		folio_unlock(eb->folios[i]);
 	return eb;
 
···
 	}
 	/*
 	 * Now all pages of that extent buffer is unmapped, set UNMAPPED flag,
-	 * so it can be cleaned up without utilizing page->mapping.
+	 * so it can be cleaned up without utilizing folio->mapping.
 	 */
 	set_bit(EXTENT_BUFFER_UNMAPPED, &eb->bflags);
 
···
 	release_extent_buffer(eb);
 }
 
-static void btree_clear_folio_dirty(struct folio *folio)
+static void btree_clear_folio_dirty_tag(struct folio *folio)
 {
-	ASSERT(folio_test_dirty(folio));
+	ASSERT(!folio_test_dirty(folio));
 	ASSERT(folio_test_locked(folio));
-	folio_clear_dirty_for_io(folio);
 	xa_lock_irq(&folio->mapping->i_pages);
 	if (!folio_test_dirty(folio))
 		__xa_clear_mark(&folio->mapping->i_pages,
···
 	xa_unlock_irq(&folio->mapping->i_pages);
 }
 
-static void clear_subpage_extent_buffer_dirty(const struct extent_buffer *eb)
-{
-	struct btrfs_fs_info *fs_info = eb->fs_info;
-	struct folio *folio = eb->folios[0];
-	bool last;
-
-	/* btree_clear_folio_dirty() needs page locked. */
-	folio_lock(folio);
-	last = btrfs_subpage_clear_and_test_dirty(fs_info, folio, eb->start, eb->len);
-	if (last)
-		btree_clear_folio_dirty(folio);
-	folio_unlock(folio);
-	WARN_ON(atomic_read(&eb->refs) == 0);
-}
-
 void btrfs_clear_buffer_dirty(struct btrfs_trans_handle *trans,
 			      struct extent_buffer *eb)
 {
 	struct btrfs_fs_info *fs_info = eb->fs_info;
-	int num_folios;
 
 	btrfs_assert_tree_write_locked(eb);
 
···
 	percpu_counter_add_batch(&fs_info->dirty_metadata_bytes, -eb->len,
 				 fs_info->dirty_metadata_batch);
 
-	if (eb->fs_info->nodesize < PAGE_SIZE)
-		return clear_subpage_extent_buffer_dirty(eb);
-
-	num_folios = num_extent_folios(eb);
-	for (int i = 0; i < num_folios; i++) {
+	for (int i = 0; i < num_extent_folios(eb); i++) {
 		struct folio *folio = eb->folios[i];
+		bool last;
 
 		if (!folio_test_dirty(folio))
 			continue;
 		folio_lock(folio);
-		btree_clear_folio_dirty(folio);
+		last = btrfs_meta_folio_clear_and_test_dirty(folio, eb);
+		if (last)
+			btree_clear_folio_dirty_tag(folio);
 		folio_unlock(folio);
 	}
 	WARN_ON(atomic_read(&eb->refs) == 0);
···
 
 void set_extent_buffer_dirty(struct extent_buffer *eb)
 {
-	int num_folios;
 	bool was_dirty;
 
 	check_buffer_tree_ref(eb);
 
 	was_dirty = test_and_set_bit(EXTENT_BUFFER_DIRTY, &eb->bflags);
 
-	num_folios = num_extent_folios(eb);
 	WARN_ON(atomic_read(&eb->refs) == 0);
 	WARN_ON(!test_bit(EXTENT_BUFFER_TREE_REF, &eb->bflags));
 	WARN_ON(test_bit(EXTENT_BUFFER_ZONED_ZEROOUT, &eb->bflags));
 
 	if (!was_dirty) {
-		bool subpage = eb->fs_info->nodesize < PAGE_SIZE;
+		bool subpage = btrfs_meta_is_subpage(eb->fs_info);
 
 		/*
 		 * For subpage case, we can have other extent buffers in the
-		 * same page, and in clear_subpage_extent_buffer_dirty() we
+		 * same page, and in clear_extent_buffer_dirty() we
 		 * have to clear page dirty without subpage lock held.
 		 * This can cause race where our page gets dirty cleared after
 		 * we just set it.
 		 *
-		 * Thankfully, clear_subpage_extent_buffer_dirty() has locked
+		 * Thankfully, clear_extent_buffer_dirty() has locked
 		 * its page for other reasons, we can use page lock to prevent
 		 * the above race.
 		 */
 		if (subpage)
 			folio_lock(eb->folios[0]);
-		for (int i = 0; i < num_folios; i++)
-			btrfs_folio_set_dirty(eb->fs_info, eb->folios[i],
-					      eb->start, eb->len);
+		for (int i = 0; i < num_extent_folios(eb); i++)
+			btrfs_meta_folio_set_dirty(eb->folios[i], eb);
 		if (subpage)
 			folio_unlock(eb->folios[0]);
 		percpu_counter_add_batch(&eb->fs_info->dirty_metadata_bytes,
···
 					 eb->fs_info->dirty_metadata_batch);
 	}
 #ifdef CONFIG_BTRFS_DEBUG
-	for (int i = 0; i < num_folios; i++)
+	for (int i = 0; i < num_extent_folios(eb); i++)
 		ASSERT(folio_test_dirty(eb->folios[i]));
 #endif
 }
 
 void clear_extent_buffer_uptodate(struct extent_buffer *eb)
 {
-	struct btrfs_fs_info *fs_info = eb->fs_info;
-	int num_folios = num_extent_folios(eb);
 
 	clear_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags);
-	for (int i = 0; i < num_folios; i++) {
+	for (int i = 0; i < num_extent_folios(eb); i++) {
 		struct folio *folio = eb->folios[i];
 
 		if (!folio)
 			continue;
 
-		/*
-		 * This is special handling for metadata subpage, as regular
-		 * btrfs_is_subpage() can not handle cloned/dummy metadata.
-		 */
-		if (fs_info->nodesize >= PAGE_SIZE)
-			folio_clear_uptodate(folio);
-		else
-			btrfs_subpage_clear_uptodate(fs_info, folio,
-						     eb->start, eb->len);
+		btrfs_meta_folio_clear_uptodate(folio, eb);
 	}
 }
 
 void set_extent_buffer_uptodate(struct extent_buffer *eb)
 {
-	struct btrfs_fs_info *fs_info = eb->fs_info;
-	int num_folios = num_extent_folios(eb);
 
 	set_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags);
-	for (int i = 0; i < num_folios; i++) {
-		struct folio *folio = eb->folios[i];
-
-		/*
-		 * This is special handling for metadata subpage, as regular
-		 * btrfs_is_subpage() can not handle cloned/dummy metadata.
-		 */
-		if (fs_info->nodesize >= PAGE_SIZE)
-			folio_mark_uptodate(folio);
-		else
-			btrfs_subpage_set_uptodate(fs_info, folio,
-						   eb->start, eb->len);
-	}
+	for (int i = 0; i < num_extent_folios(eb); i++)
+		btrfs_meta_folio_set_uptodate(eb->folios[i], eb);
 }
 
 static void clear_extent_buffer_reading(struct extent_buffer *eb)
···
 static void end_bbio_meta_read(struct btrfs_bio *bbio)
 {
 	struct extent_buffer *eb = bbio->private;
-	struct btrfs_fs_info *fs_info = eb->fs_info;
 	bool uptodate = !bbio->bio.bi_status;
-	struct folio_iter fi;
-	u32 bio_offset = 0;
 
 	/*
 	 * If the extent buffer is marked UPTODATE before the read operation
···
 		set_bit(EXTENT_BUFFER_READ_ERR, &eb->bflags);
 	}
 
-	bio_for_each_folio_all(fi, &bbio->bio) {
-		struct folio *folio = fi.folio;
-		u64 start = eb->start + bio_offset;
-		u32 len = fi.length;
-
-		if (uptodate)
-			btrfs_folio_set_uptodate(fs_info, folio, start, len);
-		else
-			btrfs_folio_clear_uptodate(fs_info, folio, start, len);
-
-		bio_offset += len;
-	}
-
 	clear_extent_buffer_reading(eb);
 	free_extent_buffer(eb);
 
···
 				    const struct btrfs_tree_parent_check *check)
 {
 	struct btrfs_bio *bbio;
-	bool ret;
 
 	if (test_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags))
 		return 0;
···
 	bbio->inode = BTRFS_I(eb->fs_info->btree_inode);
 	bbio->file_offset = eb->start;
 	memcpy(&bbio->parent_check, check, sizeof(*check));
-	if (eb->fs_info->nodesize < PAGE_SIZE) {
-		ret = bio_add_folio(&bbio->bio, eb->folios[0], eb->len,
-				    eb->start - folio_pos(eb->folios[0]));
-		ASSERT(ret);
-	} else {
-		int num_folios = num_extent_folios(eb);
+	for (int i = 0; i < num_extent_folios(eb); i++) {
+		struct folio *folio = eb->folios[i];
+		u64 range_start = max_t(u64, eb->start, folio_pos(folio));
+		u32 range_len = min_t(u64, folio_pos(folio) + folio_size(folio),
+				      eb->start + eb->len) - range_start;
 
-		for (int i = 0; i < num_folios; i++) {
-			struct folio *folio = eb->folios[i];
-
-			ret = bio_add_folio(&bbio->bio, folio, eb->folio_size, 0);
-			ASSERT(ret);
-		}
+		bio_add_folio_nofail(&bbio->bio, folio, range_len,
+				     offset_in_folio(folio, range_start));
 	}
 	btrfs_submit_bbio(bbio, mirror_num);
 	return 0;
···
 	if (test_bit(EXTENT_BUFFER_WRITE_ERR, &eb->bflags))
 		return;
 
-	if (fs_info->nodesize < PAGE_SIZE) {
+	if (btrfs_meta_is_subpage(fs_info)) {
 		folio = eb->folios[0];
 		ASSERT(i == 0);
 		if (WARN_ON(!btrfs_subpage_test_uptodate(fs_info, folio,
···
 {
 	struct extent_buffer *eb;
 
-	if (folio_to_fs_info(folio)->nodesize < PAGE_SIZE)
+	if (btrfs_meta_is_subpage(folio_to_fs_info(folio)))
 		return try_release_subpage_extent_buffer(folio);
 
 	/*

fs/btrfs/extent_io.h

···
 
 struct extent_buffer *alloc_extent_buffer(struct btrfs_fs_info *fs_info,
 					  u64 start, u64 owner_root, int level);
-struct extent_buffer *__alloc_dummy_extent_buffer(struct btrfs_fs_info *fs_info,
-						  u64 start, unsigned long len);
 struct extent_buffer *alloc_dummy_extent_buffer(struct btrfs_fs_info *fs_info,
 						u64 start);
 struct extent_buffer *btrfs_clone_extent_buffer(const struct extent_buffer *src);
···
 				u64 bytenr, u64 owner_root, u64 gen, int level);
 void btrfs_readahead_node_child(struct extent_buffer *node, int slot);
 
-static inline int num_extent_pages(const struct extent_buffer *eb)
+/* Note: this can be used in for loops without caching the value in a variable. */
+static inline int __pure num_extent_pages(const struct extent_buffer *eb)
 {
 	/*
 	 * For sectorsize == PAGE_SIZE case, since nodesize is always aligned to
···
  * As we can have either one large folio covering the whole eb
  * (either nodesize <= PAGE_SIZE, or high order folio), or multiple
  * single-paged folios.
+ *
+ * Note: this can be used in for loops without caching the value in a variable.
  */
-static inline int num_extent_folios(const struct extent_buffer *eb)
+static inline int __pure num_extent_folios(const struct extent_buffer *eb)
 {
 	if (folio_order(eb->folios[0]))
 		return 1;

fs/btrfs/file-item.c

···
 	int ret = 0;
 	struct btrfs_file_extent_item *item;
 	struct btrfs_key file_key;
-	struct btrfs_path *path;
+	BTRFS_PATH_AUTO_FREE(path);
 	struct extent_buffer *leaf;
 
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
+
 	file_key.objectid = objectid;
-	file_key.offset = pos;
 	file_key.type = BTRFS_EXTENT_DATA_KEY;
+	file_key.offset = pos;
 
 	ret = btrfs_insert_empty_item(trans, root, path, &file_key,
 				      sizeof(*item));
 	if (ret < 0)
-		goto out;
+		return ret;
 	leaf = path->nodes[0];
 	item = btrfs_item_ptr(leaf, path->slots[0],
 			      struct btrfs_file_extent_item);
···
 	btrfs_set_file_extent_compression(leaf, item, 0);
 	btrfs_set_file_extent_encryption(leaf, item, 0);
 	btrfs_set_file_extent_other_encoding(leaf, item, 0);
-out:
-	btrfs_free_path(path);
+
 	return ret;
 }
 
···
 	int csums_in_item;
 
 	file_key.objectid = BTRFS_EXTENT_CSUM_OBJECTID;
-	file_key.offset = bytenr;
 	file_key.type = BTRFS_EXTENT_CSUM_KEY;
+	file_key.offset = bytenr;
 	ret = btrfs_search_slot(trans, root, &file_key, path, 0, cow);
 	if (ret < 0)
 		goto fail;
···
 	int cow = mod != 0;
 
 	file_key.objectid = objectid;
-	file_key.offset = offset;
 	file_key.type = BTRFS_EXTENT_DATA_KEY;
+	file_key.offset = offset;
 
 	return btrfs_search_slot(trans, root, &file_key, path, ins_len, cow);
 }
···
 	struct btrfs_inode *inode = bbio->inode;
 	struct btrfs_fs_info *fs_info = inode->root->fs_info;
 	struct bio *bio = &bbio->bio;
-	struct btrfs_path *path;
+	BTRFS_PATH_AUTO_FREE(path);
 	const u32 sectorsize = fs_info->sectorsize;
 	const u32 csum_size = fs_info->csum_size;
 	u32 orig_len = bio->bi_iter.bi_size;
···
 
 	if (nblocks * csum_size > BTRFS_BIO_INLINE_CSUM_SIZE) {
 		bbio->csum = kmalloc_array(nblocks, csum_size, GFP_NOFS);
-		if (!bbio->csum) {
-			btrfs_free_path(path);
+		if (!bbio->csum)
 			return BLK_STS_RESOURCE;
-		}
 	} else {
 		bbio->csum = bbio->csum_inline;
 	}
···
 		bio_offset += count * sectorsize;
 	}
 
-	btrfs_free_path(path);
 	return ret;
 }
 
···
 	path->nowait = nowait;
 
 	key.objectid = BTRFS_EXTENT_CSUM_OBJECTID;
-	key.offset = start;
 	key.type = BTRFS_EXTENT_CSUM_KEY;
+	key.offset = start;
 
 	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
 	if (ret < 0)
···
 		    struct btrfs_root *root, u64 bytenr, u64 len)
 {
 	struct btrfs_fs_info *fs_info = trans->fs_info;
-	struct btrfs_path *path;
+	BTRFS_PATH_AUTO_FREE(path);
 	struct btrfs_key key;
 	u64 end_byte = bytenr + len;
 	u64 csum_end;
···
 
 	while (1) {
 		key.objectid = BTRFS_EXTENT_CSUM_OBJECTID;
-		key.offset = end_byte - 1;
 		key.type = BTRFS_EXTENT_CSUM_KEY;
+		key.offset = end_byte - 1;
 
 		ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
 		if (ret > 0) {
···
 		}
 		btrfs_release_path(path);
 	}
-	btrfs_free_path(path);
 	return ret;
 }
 
···
 	found_next = 0;
 	bytenr = sums->logical + total_bytes;
 	file_key.objectid = BTRFS_EXTENT_CSUM_OBJECTID;
-	file_key.offset = bytenr;
 	file_key.type = BTRFS_EXTENT_CSUM_KEY;
+	file_key.offset = bytenr;
 
 	item = btrfs_lookup_csum(trans, root, path, bytenr, 1);
 	if (!IS_ERR(item)) {

fs/btrfs/file-item.h

···
 #ifndef BTRFS_FILE_ITEM_H
 #define BTRFS_FILE_ITEM_H
 
+#include <linux/blk_types.h>
 #include <linux/list.h>
 #include <uapi/linux/btrfs_tree.h>
+#include "ctree.h"
 #include "accessors.h"
 
 struct extent_map;

fs/btrfs/file.c

···
 {
 	u64 clamp_start = max_t(u64, pos, folio_pos(folio));
 	u64 clamp_end = min_t(u64, pos + len, folio_pos(folio) + folio_size(folio));
+	const u32 blocksize = inode_to_fs_info(inode)->sectorsize;
 	int ret = 0;
 
 	if (folio_test_uptodate(folio))
 		return 0;
 
 	if (!force_uptodate &&
-	    IS_ALIGNED(clamp_start, PAGE_SIZE) &&
-	    IS_ALIGNED(clamp_end, PAGE_SIZE))
+	    IS_ALIGNED(clamp_start, blocksize) &&
+	    IS_ALIGNED(clamp_end, blocksize))
 		return 0;
 
 	ret = btrfs_read_folio(NULL, folio);
···
 			ret = PTR_ERR(folio);
 		return ret;
 	}
-	folio_wait_writeback(folio);
 	/* Only support page sized folio yet. */
 	ASSERT(folio_order(folio) == 0);
 	ret = set_folio_extent_mapped(folio);
···
 		btrfs_lock_and_flush_ordered_range(inode, lockstart, lockend,
 						   &cached_state);
 	}
-	ret = can_nocow_extent(&inode->vfs_inode, lockstart, &num_bytes,
-			       NULL, nowait);
+	ret = can_nocow_extent(inode, lockstart, &num_bytes, NULL, nowait);
 	if (ret <= 0)
 		btrfs_drew_write_unlock(&root->snapshot_lock);
 	else
···
 	struct extent_changeset *data_reserved = NULL;
 	unsigned long zero_start;
 	loff_t size;
+	size_t fsize = folio_size(folio);
 	vm_fault_t ret;
 	int ret2;
 	int reserved = 0;
···
 
 	ASSERT(folio_order(folio) == 0);
 
-	reserved_space = PAGE_SIZE;
+	reserved_space = fsize;
 
 	sb_start_pagefault(inode->i_sb);
 	page_start = folio_pos(folio);
···
 	 * We can't set the delalloc bits if there are pending ordered
 	 * extents.  Drop our locks and wait for them to finish.
 	 */
-	ordered = btrfs_lookup_ordered_range(BTRFS_I(inode), page_start, PAGE_SIZE);
+	ordered = btrfs_lookup_ordered_range(BTRFS_I(inode), page_start, fsize);
 	if (ordered) {
 		unlock_extent(io_tree, page_start, page_end, &cached_state);
 		folio_unlock(folio);
···
 
 	if (folio->index == ((size - 1) >> PAGE_SHIFT)) {
 		reserved_space = round_up(size - page_start, fs_info->sectorsize);
-		if (reserved_space < PAGE_SIZE) {
+		if (reserved_space < fsize) {
 			end = page_start + reserved_space - 1;
 			btrfs_delalloc_release_space(BTRFS_I(inode),
 					data_reserved, page_start,
-					PAGE_SIZE - reserved_space, true);
+					fsize - reserved_space, true);
 		}
 	}
 
···
 	if (page_start + folio_size(folio) > size)
 		zero_start = offset_in_folio(folio, size);
 	else
-		zero_start = PAGE_SIZE;
+		zero_start = fsize;
 
-	if (zero_start != PAGE_SIZE)
+	if (zero_start != fsize)
 		folio_zero_range(folio, zero_start, folio_size(folio) - zero_start);
 
-	btrfs_folio_clear_checked(fs_info, folio, page_start, PAGE_SIZE);
+	btrfs_folio_clear_checked(fs_info, folio, page_start, fsize);
 	btrfs_folio_set_dirty(fs_info, folio, page_start, end + 1 - page_start);
 	btrfs_folio_set_uptodate(fs_info, folio, page_start, end + 1 - page_start);
 
···
 	unlock_extent(io_tree, page_start, page_end, &cached_state);
 	up_read(&BTRFS_I(inode)->i_mmap_lock);
 
-	btrfs_delalloc_release_extents(BTRFS_I(inode), PAGE_SIZE);
+	btrfs_delalloc_release_extents(BTRFS_I(inode), fsize);
 	sb_end_pagefault(inode->i_sb);
 	extent_changeset_free(data_reserved);
 	return VM_FAULT_LOCKED;
···
 	folio_unlock(folio);
 	up_read(&BTRFS_I(inode)->i_mmap_lock);
 out:
-	btrfs_delalloc_release_extents(BTRFS_I(inode), PAGE_SIZE);
+	btrfs_delalloc_release_extents(BTRFS_I(inode), fsize);
 	btrfs_delalloc_release_space(BTRFS_I(inode), data_reserved, page_start,
 				     reserved_space, (ret != 0));
 out_noreserve:

fs/btrfs/file.h

···
 struct extent_state;
 struct kiocb;
 struct iov_iter;
+struct inode;
+struct folio;
 struct page;
 struct btrfs_ioctl_encoded_io_args;
 struct btrfs_drop_extents_args;

fs/btrfs/free-space-cache.c

···
 	struct btrfs_disk_key disk_key;
 	struct btrfs_free_space_header *header;
 	struct extent_buffer *leaf;
-	struct inode *inode = NULL;
+	struct btrfs_inode *inode;
 	unsigned nofs_flag;
 	int ret;
 
 	key.objectid = BTRFS_FREE_SPACE_OBJECTID;
-	key.offset = offset;
 	key.type = 0;
+	key.offset = offset;
 
 	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
 	if (ret < 0)
···
 	btrfs_release_path(path);
 	memalloc_nofs_restore(nofs_flag);
 	if (IS_ERR(inode))
-		return inode;
+		return ERR_CAST(inode);
 
-	mapping_set_gfp_mask(inode->i_mapping,
-			mapping_gfp_constraint(inode->i_mapping,
+	mapping_set_gfp_mask(inode->vfs_inode.i_mapping,
+			mapping_gfp_constraint(inode->vfs_inode.i_mapping,
 			~(__GFP_FS | __GFP_HIGHMEM)));
 
-	return inode;
+	return &inode->vfs_inode;
 }
 
 struct inode *lookup_free_space_inode(struct btrfs_block_group *block_group,
···
 	btrfs_release_path(path);
 
 	key.objectid = BTRFS_FREE_SPACE_OBJECTID;
-	key.offset = offset;
 	key.type = 0;
+	key.offset = offset;
 	ret = btrfs_insert_empty_item(trans, root, path, &key,
 				      sizeof(struct btrfs_free_space_header));
 	if (ret < 0) {
···
 				  struct inode *inode,
 				  struct btrfs_block_group *block_group)
 {
-	struct btrfs_path *path;
+	BTRFS_PATH_AUTO_FREE(path);
 	struct btrfs_key key;
 	int ret = 0;
 
···
 	if (IS_ERR(inode)) {
 		if (PTR_ERR(inode) != -ENOENT)
 			ret = PTR_ERR(inode);
-		goto out;
+		return ret;
 	}
 	ret = btrfs_orphan_add(trans, BTRFS_I(inode));
 	if (ret) {
 		btrfs_add_delayed_iput(BTRFS_I(inode));
-		goto out;
+		return ret;
 	}
 	clear_nlink(inode);
 	/* One for the block groups ref */
···
 	if (ret) {
 		if (ret > 0)
 			ret = 0;
-		goto out;
+		return ret;
 	}
-	ret = btrfs_del_item(trans, trans->fs_info->tree_root, path);
-out:
-	btrfs_free_path(path);
-	return ret;
+	return btrfs_del_item(trans, trans->fs_info->tree_root, path);
 }
 
 int btrfs_truncate_free_space_cache(struct btrfs_trans_handle *trans,
···
 
 static int io_ctl_prepare_pages(struct btrfs_io_ctl *io_ctl, bool uptodate)
 {
-	struct page *page;
+	struct folio *folio;
 	struct inode *inode = io_ctl->inode;
 	gfp_t mask = btrfs_alloc_write_mask(inode->i_mapping);
 	int i;
···
 	for (i = 0; i < io_ctl->num_pages; i++) {
 		int ret;
 
-		page = find_or_create_page(inode->i_mapping, i, mask);
-		if (!page) {
+		folio = __filemap_get_folio(inode->i_mapping, i,
+					    FGP_LOCK | FGP_ACCESSED | FGP_CREAT,
+					    mask);
+		if (IS_ERR(folio)) {
 			io_ctl_drop_pages(io_ctl);
 			return -ENOMEM;
 		}
 
-		ret = set_folio_extent_mapped(page_folio(page));
+		ret = set_folio_extent_mapped(folio);
 		if (ret < 0) {
-			unlock_page(page);
-			put_page(page);
+			folio_unlock(folio);
+			folio_put(folio);
 			io_ctl_drop_pages(io_ctl);
 			return ret;
 		}
 
-		io_ctl->pages[i] = page;
-		if (uptodate && !PageUptodate(page)) {
-			btrfs_read_folio(NULL, page_folio(page));
-			lock_page(page);
-			if (page->mapping != inode->i_mapping) {
+		io_ctl->pages[i] = &folio->page;
+		if (uptodate && !folio_test_uptodate(folio)) {
+			btrfs_read_folio(NULL, folio);
+			folio_lock(folio);
+			if (folio->mapping != inode->i_mapping) {
 				btrfs_err(BTRFS_I(inode)->root->fs_info,
 					  "free space cache page truncated");
 				io_ctl_drop_pages(io_ctl);
 				return -EIO;
 			}
-			if (!PageUptodate(page)) {
+			if (!folio_test_uptodate(folio)) {
 				btrfs_err(BTRFS_I(inode)->root->fs_info,
 					   "error reading free space cache");
 				io_ctl_drop_pages(io_ctl);
···
 		return 0;
 
 	key.objectid = BTRFS_FREE_SPACE_OBJECTID;
-	key.offset = offset;
 	key.type = 0;
+	key.offset = offset;
 
 	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
 	if (ret < 0)
···
 	int ret;
 
 	key.objectid = BTRFS_FREE_SPACE_OBJECTID;
-	key.offset = offset;
 	key.type = 0;
+	key.offset = offset;
 
 	ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
 	if (ret < 0) {

fs/btrfs/free-space-tree.c

···
 				    struct btrfs_block_group *block_group)
 {
 	struct btrfs_root *extent_root;
-	struct btrfs_path *path, *path2;
+	BTRFS_PATH_AUTO_FREE(path);
+	BTRFS_PATH_AUTO_FREE(path2);
 	struct btrfs_key key;
 	u64 start, end;
 	int ret;
···
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
-	path->reada = READA_FORWARD;
 
 	path2 = btrfs_alloc_path();
-	if (!path2) {
-		btrfs_free_path(path);
+	if (!path2)
 		return -ENOMEM;
-	}
+
+	path->reada = READA_FORWARD;
 
 	ret = add_new_free_space_info(trans, block_group, path2);
 	if (ret)
-		goto out;
+		return ret;
 
 	mutex_lock(&block_group->free_space_lock);
 
···
 	ret = 0;
 out_locked:
 	mutex_unlock(&block_group->free_space_lock);
-out:
-	btrfs_free_path(path2);
-	btrfs_free_path(path);
+
 	return ret;
 }
 
···
 static int clear_free_space_tree(struct btrfs_trans_handle *trans,
 				 struct btrfs_root *root)
 {
-	struct btrfs_path *path;
+	BTRFS_PATH_AUTO_FREE(path);
 	struct btrfs_key key;
 	int nr;
 	int ret;
···
 	while (1) {
 		ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
 		if (ret < 0)
-			goto out;
+			return ret;
 
 		nr = btrfs_header_nritems(path->nodes[0]);
 		if (!nr)
···
 		path->slots[0] = 0;
 		ret = btrfs_del_items(trans, root, path, 0, nr);
 		if (ret)
-			goto out;
+			return ret;
 
 		btrfs_release_path(path);
 	}
 
-	ret = 0;
-out:
-	btrfs_free_path(path);
-	return ret;
+	return 0;
 }
 
 int btrfs_delete_free_space_tree(struct btrfs_fs_info *fs_info)
···
 {
 	struct btrfs_block_group *block_group;
 	struct btrfs_free_space_info *info;
-	struct btrfs_path *path;
+	BTRFS_PATH_AUTO_FREE(path);
 	u32 extent_count, flags;
-	int ret;
 
 	block_group = caching_ctl->block_group;
 
···
 	path->reada = READA_FORWARD;
 
 	info = search_free_space_info(NULL, block_group, path, 0);
-	if (IS_ERR(info)) {
-		ret = PTR_ERR(info);
-		goto out;
-	}
+	if (IS_ERR(info))
+		return PTR_ERR(info);
+
 	extent_count = btrfs_free_space_extent_count(path->nodes[0], info);
 	flags = btrfs_free_space_flags(path->nodes[0], info);
 
···
 	 * there.
 	 */
 	if (flags & BTRFS_FREE_SPACE_USING_BITMAPS)
-		ret = load_free_space_bitmaps(caching_ctl, path, extent_count);
+		return load_free_space_bitmaps(caching_ctl, path, extent_count);
 	else
-		ret = load_free_space_extents(caching_ctl, path, extent_count);
-
-out:
-	btrfs_free_path(path);
-	return ret;
+		return load_free_space_extents(caching_ctl, path, extent_count);
 }

fs/btrfs/fs.c

···
 // SPDX-License-Identifier: GPL-2.0
 
 #include "messages.h"
-#include "ctree.h"
 #include "fs.h"
 #include "accessors.h"
 #include "volumes.h"

fs/btrfs/fs.h

···
 struct btrfs_stripe_hash_table;
 struct btrfs_space_info;
 
+/*
+ * Minimum data and metadata block size.
+ *
+ * Normally it's 4K, but for testing subpage block size on 4K page systems, we
+ * allow DEBUG builds to accept 2K page size.
+ */
+#ifdef CONFIG_BTRFS_DEBUG
+#define BTRFS_MIN_BLOCKSIZE	(SZ_2K)
+#else
+#define BTRFS_MIN_BLOCKSIZE	(SZ_4K)
+#endif
+
 #define BTRFS_MAX_EXTENT_SIZE SZ_128M
 
 #define BTRFS_OLDEST_GENERATION	0ULL
···
 
 	/* Indicates there was an error cleaning up a log tree. */
 	BTRFS_FS_STATE_LOG_CLEANUP_ERROR,
+
+	/* No more delayed iput can be queued. */
+	BTRFS_FS_STATE_NO_DELAYED_IPUT,
 
 	BTRFS_FS_STATE_COUNT
 };
···
 	u64 last_trans_log_full_commit;
 	unsigned long long mount_opt;
 
-	unsigned long compress_type:4;
-	unsigned int compress_level;
+	int compress_type;
+	int compress_level;
 	u32 commit_interval;
 	/*
 	 * It is a suggestive number, the read side is safe even it gets a
···
 	 * running.
 	 */
 	refcount_t scrub_workers_refcnt;
-	u32 sectors_per_page;
 	struct workqueue_struct *scrub_workers;
 
 	struct btrfs_discard_ctl discard_ctl;
···
 #endif
 
 	return div_u64(size + fs_info->max_extent_size - 1, fs_info->max_extent_size);
+}
+
+static inline unsigned int btrfs_blocks_per_folio(const struct btrfs_fs_info *fs_info,
+						  const struct folio *folio)
+{
+	return folio_size(folio) >> fs_info->sectorsize_bits;
 }
 
 bool btrfs_exclop_start(struct btrfs_fs_info *fs_info,

fs/btrfs/inode-item.c

···
 	int del_len = name->len + sizeof(*ref);
 
 	key.objectid = inode_objectid;
-	key.offset = ref_objectid;
 	key.type = BTRFS_INODE_REF_KEY;
+	key.offset = ref_objectid;
 
 	path = btrfs_alloc_path();
 	if (!path)
···
 	int ins_len = name->len + sizeof(*ref);
 
 	key.objectid = inode_objectid;
-	key.offset = ref_objectid;
 	key.type = BTRFS_INODE_REF_KEY;
+	key.offset = ref_objectid;
 
 	path = btrfs_alloc_path();
 	if (!path)
···
 	path->reada = READA_BACK;
 
 	key.objectid = control->ino;
-	key.offset = (u64)-1;
 	key.type = (u8)-1;
+	key.offset = (u64)-1;
 
 search_again:
 	/*

fs/btrfs/inode.c

···
 		size_t datasize;
 
 		key.objectid = btrfs_ino(inode);
-		key.offset = 0;
 		key.type = BTRFS_EXTENT_DATA_KEY;
+		key.offset = 0;
 
 		datasize = btrfs_file_extent_calc_inline_size(cur_size);
 		ret = btrfs_insert_empty_item(trans, root, path, &key,
···
 	if (offset != 0)
 		return false;
 
-	/*
-	 * Due to the page size limit, for subpage we can only trigger the
-	 * writeback for the dirty sectors of page, that means data writeback
-	 * is doing more writeback than what we want.
-	 *
-	 * This is especially unexpected for some call sites like fallocate,
-	 * where we only increase i_size after everything is done.
-	 * This means we can trigger inline extent even if we didn't want to.
-	 * So here we skip inline extent creation completely.
-	 */
-	if (fs_info->sectorsize != PAGE_SIZE)
-		return false;
-
 	/* Inline extents are limited to sectorsize. */
 	if (size > fs_info->sectorsize)
+		return false;
+
+	/* We do not allow a non-compressed extent to be as large as block size. */
+	if (data_len >= fs_info->sectorsize)
 		return false;
 
 	/* We cannot exceed the maximum inline data size. */
···
 	 * And at reserve time, it's always aligned to page size, so
 	 * just free one page here.
 	 */
-	btrfs_qgroup_free_data(inode, NULL, 0, PAGE_SIZE, NULL);
+	btrfs_qgroup_free_data(inode, NULL, 0, fs_info->sectorsize, NULL);
 	btrfs_free_path(path);
 	btrfs_end_transaction(trans);
 	return ret;
···
 		btrfs_add_inode_defrag(inode, small_write);
 }
 
-static int extent_range_clear_dirty_for_io(struct inode *inode, u64 start, u64 end)
+static int extent_range_clear_dirty_for_io(struct btrfs_inode *inode, u64 start, u64 end)
 {
 	unsigned long end_index = end >> PAGE_SHIFT;
 	struct folio *folio;
···
 
 	for (unsigned long index = start >> PAGE_SHIFT;
 	     index <= end_index; index++) {
-		folio = filemap_get_folio(inode->i_mapping, index);
+		folio = filemap_get_folio(inode->vfs_inode.i_mapping, index);
 		if (IS_ERR(folio)) {
 			if (!ret)
 				ret = PTR_ERR(folio);
 			continue;
 		}
-		btrfs_folio_clamp_clear_dirty(inode_to_fs_info(inode), folio, start,
+		btrfs_folio_clamp_clear_dirty(inode->root->fs_info, folio, start,
 					      end + 1 - start);
 		folio_put(folio);
 	}
···
 	unsigned int poff;
 	int i;
 	int compress_type = fs_info->compress_type;
+	int compress_level = fs_info->compress_level;
 
 	inode_should_defrag(inode, start, end, end - start + 1, SZ_16K);
 
···
 	 * Otherwise applications with the file mmap'd can wander in and change
 	 * the page contents while we are compressing them.
 	 */
-	ret = extent_range_clear_dirty_for_io(&inode->vfs_inode, start, end);
+	ret = extent_range_clear_dirty_for_io(inode, start, end);
 
 	/*
 	 * All the folios should have been locked thus no failure.
···
 		goto cleanup_and_bail_uncompressed;
 	}
 
-	if (inode->defrag_compress)
+	if (inode->defrag_compress) {
 		compress_type = inode->defrag_compress;
-	else if (inode->prop_compress)
+		compress_level = inode->defrag_compress_level;
+	} else if (inode->prop_compress) {
 		compress_type = inode->prop_compress;
+	}
 
 	/* Compression level is applied here. */
-	ret = btrfs_compress_folios(compress_type | (fs_info->compress_level << 4),
+	ret = btrfs_compress_folios(compress_type, compress_level,
 				    mapping, start, folios, &nr_folios, &total_in,
 				    &total_compressed);
 	if (ret)
···
 			       &wbc, false);
 	wbc_detach_inode(&wbc);
 	if (ret < 0) {
-		btrfs_cleanup_ordered_extents(inode, start, end - start + 1);
 		if (locked_folio)
 			btrfs_folio_end_lock(inode->root->fs_info, locked_folio,
 					     start, async_extent->ram_size);
···
  * - Else all pages except for @locked_folio are unlocked.
  *
  * When a failure happens in the second or later iteration of the
- * while-loop, the ordered extents created in previous iterations are kept
- * intact. So, the caller must clean them up by calling
- * btrfs_cleanup_ordered_extents(). See btrfs_run_delalloc_range() for
- * example.
+ * while-loop, the ordered extents created in previous iterations are cleaned up.
  */
 static noinline int cow_file_range(struct btrfs_inode *inode,
 				   struct folio *locked_folio, u64 start,
···
 
 	/*
 	 * For the range (1). We have already instantiated the ordered extents
-	 * for this region. They are cleaned up by
-	 * btrfs_cleanup_ordered_extents() in e.g,
-	 * btrfs_run_delalloc_range().
+	 * for this region, thus we need to cleanup those ordered extents.
 	 * EXTENT_DELALLOC_NEW | EXTENT_DEFRAG | EXTENT_CLEAR_META_RESV
-	 * are also handled by the cleanup function.
+	 * are also handled by the ordered extents cleanup.
 	 *
 	 * So here we only clear EXTENT_LOCKED and EXTENT_DELALLOC flag, and
 	 * finish the writeback of the involved folios, which will be never submitted.
···
 
 		if (!locked_folio)
 			mapping_set_error(inode->vfs_inode.i_mapping, ret);
+
+		btrfs_cleanup_ordered_extents(inode, orig_start, start - orig_start);
 		extent_clear_unlock_delalloc(inode, orig_start, start - 1,
 					     locked_folio, NULL, clear_bits, page_ops);
 	}
···
 	mapping_set_error(mapping, error);
 }
 
+static int nocow_one_range(struct btrfs_inode *inode, struct folio *locked_folio,
+			   struct extent_state **cached,
+			   struct can_nocow_file_extent_args *nocow_args,
+			   u64 file_pos, bool is_prealloc)
+{
+	struct btrfs_ordered_extent *ordered;
+	u64 len = nocow_args->file_extent.num_bytes;
+	u64 end = file_pos + len - 1;
+	int ret = 0;
+
+	lock_extent(&inode->io_tree, file_pos, end, cached);
+
+	if (is_prealloc) {
+		struct extent_map *em;
+
+		em = btrfs_create_io_em(inode, file_pos, &nocow_args->file_extent,
+					BTRFS_ORDERED_PREALLOC);
+		if (IS_ERR(em)) {
+			unlock_extent(&inode->io_tree, file_pos, end, cached);
+			return PTR_ERR(em);
+		}
+		free_extent_map(em);
+	}
+
+	ordered = btrfs_alloc_ordered_extent(inode, file_pos, &nocow_args->file_extent,
+					     is_prealloc
+					     ? (1 << BTRFS_ORDERED_PREALLOC)
+					     : (1 << BTRFS_ORDERED_NOCOW));
+	if (IS_ERR(ordered)) {
+		if (is_prealloc)
+			btrfs_drop_extent_map_range(inode, file_pos, end, false);
+		unlock_extent(&inode->io_tree, file_pos, end, cached);
+		return PTR_ERR(ordered);
+	}
+
+	if (btrfs_is_data_reloc_root(inode->root))
+		/*
+		 * Errors are handled later, as we must prevent
+		 * extent_clear_unlock_delalloc() in error handler from freeing
+		 * metadata of the created ordered extent.
+		 */
+		ret = btrfs_reloc_clone_csums(ordered);
+	btrfs_put_ordered_extent(ordered);
+
+	extent_clear_unlock_delalloc(inode, file_pos, end, locked_folio, cached,
+				     EXTENT_LOCKED | EXTENT_DELALLOC |
+				     EXTENT_CLEAR_DATA_RESV,
+				     PAGE_UNLOCK | PAGE_SET_ORDERED);
+	/*
+	 * On error, we need to cleanup the ordered extents we created.
+	 *
+	 * We do not clear the folio Dirty flags because they are set and
+	 * cleaered by the caller.
+	 */
+	if (ret < 0)
+		btrfs_cleanup_ordered_extents(inode, file_pos, end);
+	return ret;
+}
+
 /*
  * when nowcow writeback call back.  This checks for snapshots or COW copies
  * of the extents that exist in the file, and COWs the file as required.
···
 
 	while (cur_offset <= end) {
 		struct btrfs_block_group *nocow_bg = NULL;
-		struct btrfs_ordered_extent *ordered;
 		struct btrfs_key found_key;
 		struct btrfs_file_extent_item *fi;
 		struct extent_buffer *leaf;
 		struct extent_state *cached_state = NULL;
 		u64 extent_end;
-		u64 nocow_end;
 		int extent_type;
-		bool is_prealloc;
 
 		ret = btrfs_lookup_file_extent(NULL, root, path, ino,
 					       cur_offset, 0);
···
 		if (cow_start != (u64)-1) {
 			ret = fallback_to_cow(inode, locked_folio, cow_start,
 					      found_key.offset - 1);
-			cow_start = (u64)-1;
 			if (ret) {
 				cow_end = found_key.offset - 1;
 				btrfs_dec_nocow_writers(nocow_bg);
 				goto error;
 			}
+			cow_start = (u64)-1;
 		}
 
-		nocow_end = cur_offset + nocow_args.file_extent.num_bytes - 1;
-		lock_extent(&inode->io_tree, cur_offset, nocow_end, &cached_state);
-
-		is_prealloc = extent_type == BTRFS_FILE_EXTENT_PREALLOC;
-		if (is_prealloc) {
-			struct extent_map *em;
-
-			em = btrfs_create_io_em(inode, cur_offset,
-						&nocow_args.file_extent,
-						BTRFS_ORDERED_PREALLOC);
-			if (IS_ERR(em)) {
-				unlock_extent(&inode->io_tree, cur_offset,
-					      nocow_end, &cached_state);
-				btrfs_dec_nocow_writers(nocow_bg);
-				ret = PTR_ERR(em);
-				goto error;
-			}
-			free_extent_map(em);
-		}
-
-		ordered = btrfs_alloc_ordered_extent(inode, cur_offset,
-				&nocow_args.file_extent,
-				is_prealloc
-				? (1 << BTRFS_ORDERED_PREALLOC)
-				: (1 << BTRFS_ORDERED_NOCOW));
+		ret = nocow_one_range(inode, locked_folio, &cached_state,
+				      &nocow_args, cur_offset,
+				      extent_type == BTRFS_FILE_EXTENT_PREALLOC);
 		btrfs_dec_nocow_writers(nocow_bg);
-		if (IS_ERR(ordered)) {
-			if (is_prealloc) {
-				btrfs_drop_extent_map_range(inode, cur_offset,
-							    nocow_end, false);
-			}
-			unlock_extent(&inode->io_tree, cur_offset,
-				      nocow_end, &cached_state);
-			ret = PTR_ERR(ordered);
+		if (ret < 0)
 			goto error;
-		}
-
-		if (btrfs_is_data_reloc_root(root))
-			/*
-			 * Error handled later, as we must prevent
-			 * extent_clear_unlock_delalloc() in error handler
-			 * from freeing metadata of created ordered extent.
-			 */
-			ret = btrfs_reloc_clone_csums(ordered);
-		btrfs_put_ordered_extent(ordered);
-
-		extent_clear_unlock_delalloc(inode, cur_offset, nocow_end,
-					     locked_folio, &cached_state,
-					     EXTENT_LOCKED | EXTENT_DELALLOC |
-					     EXTENT_CLEAR_DATA_RESV,
-					     PAGE_UNLOCK | PAGE_SET_ORDERED);
-
 		cur_offset = extent_end;
-
-		/*
-		 * btrfs_reloc_clone_csums() error, now we're OK to call error
-		 * handler, as metadata for created ordered extent will only
-		 * be freed by btrfs_finish_ordered_io().
-		 */
-		if (ret)
-			goto error;
 	}
 	btrfs_release_path(path);
 
···
 
 	if (cow_start != (u64)-1) {
 		ret = fallback_to_cow(inode, locked_folio, cow_start, end);
-		cow_start = (u64)-1;
 		if (ret) {
 			cow_end = end;
 			goto error;
 		}
+		cow_start = (u64)-1;
 	}
 
 	btrfs_free_path(path);
···
 	 *    start         cur_offset             end
 	 *    |/////////////|                      |
 	 *
+	 *    In this case, cow_start should be (u64)-1.
+	 *
 	 *    For range [start, cur_offset) the folios are already unlocked (except
 	 *    @locked_folio), EXTENT_DELALLOC already removed.
-	 *    Only need to clear the dirty flag as they will never be submitted.
-	 *    Ordered extent and extent maps are handled by
-	 *    btrfs_mark_ordered_io_finished() inside run_delalloc_range().
+	 *    Need to clear the dirty flags and finish the ordered extents.
 	 *
-	 * 2) Failed with error from fallback_to_cow()
-	 *    start         cur_offset  cow_end    end
+	 * 2) Failed with error before calling fallback_to_cow()
+	 *
+	 *    start         cow_start              end
+	 *    |/////////////|                      |
+	 *
+	 *    In this case, only @cow_start is set, @cur_offset is between
+	 *    [cow_start, end)
+	 *
+	 *    It's mostly the same as case 1), just replace @cur_offset with
+	 *    @cow_start.
+	 *
+	 * 3) Failed with error from fallback_to_cow()
+	 *
+	 *    start         cow_start   cow_end    end
 	 *    |/////////////|-----------|          |
 	 *
-	 *    For range [start, cur_offset) it's the same as case 1).
-	 *    But for range [cur_offset, cow_end), the folios have dirty flag
-	 *    cleared and unlocked, EXTENT_DEALLLOC cleared by cow_file_range().
+	 *    In this case, both @cow_start and @cow_end is set.
 	 *
-	 *    Thus we should not call extent_clear_unlock_delalloc() on range
-	 *    [cur_offset, cow_end), as the folios are already unlocked.
+	 *    For range [start, cow_start) it's the same as case 1).
+	 *    But for range [cow_start, cow_end), all the cleanup is handled by
+	 *    cow_file_range(), we should not touch anything in that range.
 	 *
-	 * So clear the folio dirty flags for [start, cur_offset) first.
+	 * So for all above cases, if @cow_start is set, cleanup ordered extents
+	 * for range [start, @cow_start), other wise cleanup range [start, @cur_offset).
 	 */
-	if (cur_offset > start)
+	if (cow_start != (u64)-1)
+		cur_offset = cow_start;
+
+	if (cur_offset > start) {
+		btrfs_cleanup_ordered_extents(inode, start, cur_offset - start);
 		cleanup_dirty_folios(inode, locked_folio, start, cur_offset - 1, ret);
+	}
 
 	/*
 	 * If an error happened while a COW region is outstanding, cur_offset
···
 
 	if (should_nocow(inode, start, end)) {
 		ret = run_delalloc_nocow(inode, locked_folio, start, end);
-		goto out;
+		return ret;
 	}
 
 	if (btrfs_inode_can_compress(inode) &&
···
 	else
 		ret = cow_file_range(inode, locked_folio, start, end, NULL,
 				     false, false);
-
-out:
-	if (ret < 0)
-		btrfs_cleanup_ordered_extents(inode, start, end - start + 1);
 	return ret;
 }
 
···
 		return 0;
 
 	/*
+	 * For experimental build, we error out instead of EAGAIN.
+	 *
+	 * We should not hit such out-of-band dirty folios anymore.
+	 */
+	if (IS_ENABLED(CONFIG_BTRFS_EXPERIMENTAL)) {
+		WARN_ON(IS_ENABLED(CONFIG_BTRFS_DEBUG));
+		btrfs_err_rl(fs_info,
+	"root %lld ino %llu folio %llu is marked dirty without notifying the fs",
+			     BTRFS_I(inode)->root->root_key.objectid,
+			     btrfs_ino(BTRFS_I(inode)),
+			     folio_pos(folio));
+		return -EUCLEAN;
+	}
+
+	/*
 	 * folio_checked is set below when we create a fixup worker for this
 	 * folio, don't try to create another one if we're already
 	 * folio_test_checked.
···
 	 * We are already holding a reference to this inode from
 	 * write_cache_pages.  We need to hold it because the space reservation
 	 * takes place outside of the folio lock, and we can't trust
-	 * page->mapping outside of the folio lock.
+	 * folio->mapping outside of the folio lock.
 	 */
 	ihold(inode);
 	btrfs_folio_set_checked(fs_info, folio, folio_pos(folio), folio_size(folio));
···
 
 	if (!drop_args.extent_inserted) {
 		ins.objectid = btrfs_ino(inode);
-		ins.offset = file_pos;
 		ins.type = BTRFS_EXTENT_DATA_KEY;
+		ins.offset = file_pos;
 
 		ret = btrfs_insert_empty_item(trans, root, path, &ins,
 					      sizeof(*stack_fi));
···
 		btrfs_update_inode_bytes(inode, num_bytes, drop_args.bytes_found);
 
 	ins.objectid = disk_bytenr;
-	ins.offset = disk_num_bytes;
 	ins.type = BTRFS_EXTENT_ITEM_KEY;
+	ins.offset = disk_num_bytes;
 
 	ret = btrfs_inode_set_file_extent_range(inode, file_pos, ram_bytes);
 	if (ret)
···
 	if (atomic_add_unless(&inode->vfs_inode.i_count, -1, 1))
 		return;
 
+	WARN_ON_ONCE(test_bit(BTRFS_FS_STATE_NO_DELAYED_IPUT, &fs_info->fs_state));
 	atomic_inc(&fs_info->nr_delayed_iputs);
 	/*
 	 * Need to be irq safe here because we can be called from either an irq
···
 	struct extent_buffer *leaf;
 	struct btrfs_key key, found_key;
 	struct btrfs_trans_handle *trans;
-	struct inode *inode;
 	u64 last_objectid = 0;
 	int ret = 0, nr_unlink = 0;
 
···
 	key.offset = (u64)-1;
 
 	while (1) {
+		struct btrfs_inode *inode;
+
 		ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
 		if (ret < 0)
 			goto out;
···
 		 * deleted but wasn't. The inode number may have been reused,
 		 * but either way, we can delete the orphan item.
 		 */
-		if (!inode || inode->i_nlink) {
+		if (!inode || inode->vfs_inode.i_nlink) {
 			if (inode) {
-				ret = btrfs_drop_verity_items(BTRFS_I(inode));
-				iput(inode);
+				ret = btrfs_drop_verity_items(inode);
+				iput(&inode->vfs_inode);
 				inode = NULL;
 				if (ret)
 					goto out;
···
 		nr_unlink++;
 
 		/* this will do delete_inode and everything for us */
-		iput(inode);
+		iput(&inode->vfs_inode);
 	}
 	/* release the path since we're done with it */
 	btrfs_release_path(path);
···
  *
  * On failure clean up the inode.
  */
-static int btrfs_read_locked_inode(struct inode *inode, struct btrfs_path *path)
+static int btrfs_read_locked_inode(struct btrfs_inode *inode, struct btrfs_path *path)
 {
-	struct btrfs_fs_info *fs_info = inode_to_fs_info(inode);
+	struct btrfs_root *root = inode->root;
+	struct btrfs_fs_info *fs_info = root->fs_info;
 	struct extent_buffer *leaf;
 	struct btrfs_inode_item *inode_item;
-	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct inode *vfs_inode = &inode->vfs_inode;
 	struct btrfs_key location;
 	unsigned long ptr;
 	int maybe_acls;
···
 	bool filled = false;
 	int first_xattr_slot;
 
-	ret = btrfs_init_file_extent_tree(BTRFS_I(inode));
+	ret = btrfs_init_file_extent_tree(inode);
 	if (ret)
 		goto out;
 
···
 
 	ASSERT(path);
 
-	btrfs_get_inode_key(BTRFS_I(inode), &location);
+	btrfs_get_inode_key(inode, &location);
 
 	ret = btrfs_lookup_inode(NULL, root, path, &location, 0);
 	if (ret) {
···
 
 	inode_item = btrfs_item_ptr(leaf, path->slots[0],
 				    struct btrfs_inode_item);
-	inode->i_mode = btrfs_inode_mode(leaf, inode_item);
-	set_nlink(inode, btrfs_inode_nlink(leaf, inode_item));
-	i_uid_write(inode, btrfs_inode_uid(leaf, inode_item));
-	i_gid_write(inode, btrfs_inode_gid(leaf, inode_item));
-	btrfs_i_size_write(BTRFS_I(inode), btrfs_inode_size(leaf, inode_item));
-	btrfs_inode_set_file_extent_range(BTRFS_I(inode), 0,
-			round_up(i_size_read(inode), fs_info->sectorsize));
+	vfs_inode->i_mode = btrfs_inode_mode(leaf, inode_item);
+	set_nlink(vfs_inode, btrfs_inode_nlink(leaf, inode_item));
+	i_uid_write(vfs_inode, btrfs_inode_uid(leaf, inode_item));
+	i_gid_write(vfs_inode, btrfs_inode_gid(leaf, inode_item));
+	btrfs_i_size_write(inode, btrfs_inode_size(leaf, inode_item));
+	btrfs_inode_set_file_extent_range(inode, 0,
+			round_up(i_size_read(vfs_inode), fs_info->sectorsize));
 
-	inode_set_atime(inode, btrfs_timespec_sec(leaf, &inode_item->atime),
+	inode_set_atime(vfs_inode, btrfs_timespec_sec(leaf, &inode_item->atime),
 			btrfs_timespec_nsec(leaf, &inode_item->atime));
 
-	inode_set_mtime(inode, btrfs_timespec_sec(leaf, &inode_item->mtime),
+	inode_set_mtime(vfs_inode, btrfs_timespec_sec(leaf, &inode_item->mtime),
 			btrfs_timespec_nsec(leaf, &inode_item->mtime));
 
-	inode_set_ctime(inode, btrfs_timespec_sec(leaf, &inode_item->ctime),
+	inode_set_ctime(vfs_inode, btrfs_timespec_sec(leaf, &inode_item->ctime),
 			btrfs_timespec_nsec(leaf, &inode_item->ctime));
 
-	BTRFS_I(inode)->i_otime_sec = btrfs_timespec_sec(leaf, &inode_item->otime);
-	BTRFS_I(inode)->i_otime_nsec = btrfs_timespec_nsec(leaf, &inode_item->otime);
+	inode->i_otime_sec = btrfs_timespec_sec(leaf, &inode_item->otime);
+	inode->i_otime_nsec = btrfs_timespec_nsec(leaf, &inode_item->otime);
 
-	inode_set_bytes(inode, btrfs_inode_nbytes(leaf, inode_item));
-	BTRFS_I(inode)->generation = btrfs_inode_generation(leaf, inode_item);
-	BTRFS_I(inode)->last_trans = btrfs_inode_transid(leaf, inode_item);
+	inode_set_bytes(vfs_inode, btrfs_inode_nbytes(leaf, inode_item));
+	inode->generation = btrfs_inode_generation(leaf, inode_item);
+	inode->last_trans = btrfs_inode_transid(leaf, inode_item);
 
-	inode_set_iversion_queried(inode,
-				   btrfs_inode_sequence(leaf, inode_item));
-	inode->i_generation = BTRFS_I(inode)->generation;
-	inode->i_rdev = 0;
+	inode_set_iversion_queried(vfs_inode, btrfs_inode_sequence(leaf, inode_item));
+	vfs_inode->i_generation = inode->generation;
+	vfs_inode->i_rdev = 0;
 	rdev = btrfs_inode_rdev(leaf, inode_item);
 
-	if (S_ISDIR(inode->i_mode))
-		BTRFS_I(inode)->index_cnt = (u64)-1;
+	if (S_ISDIR(vfs_inode->i_mode))
+		inode->index_cnt = (u64)-1;
 
 	btrfs_inode_split_flags(btrfs_inode_flags(leaf, inode_item),
-				&BTRFS_I(inode)->flags, &BTRFS_I(inode)->ro_flags);
+				&inode->flags, &inode->ro_flags);
+	btrfs_update_inode_mapping_flags(inode);
 
 cache_index:
 	/*
···
 	 * This is required for both inode re-read from disk and delayed inode
 	 * in the delayed_nodes xarray.
 	 */
-	if (BTRFS_I(inode)->last_trans == btrfs_get_fs_generation(fs_info))
-		set_bit(BTRFS_INODE_NEEDS_FULL_SYNC,
-			&BTRFS_I(inode)->runtime_flags);
+	if (inode->last_trans == btrfs_get_fs_generation(fs_info))
+		set_bit(BTRFS_INODE_NEEDS_FULL_SYNC, &inode->runtime_flags);
 
 	/*
 	 * We don't persist the id of the transaction where an unlink operation
···
 	 * transaction commits on fsync if our inode is a directory, or if our
 	 * inode is not a directory, logging its parent unnecessarily.
 	 */
-	BTRFS_I(inode)->last_unlink_trans = BTRFS_I(inode)->last_trans;
+	inode->last_unlink_trans = inode->last_trans;
 
 	/*
 	 * Same logic as for last_unlink_trans. We don't persist the generation
···
 	 * operation, so after eviction and reloading the inode we must be
 	 * pessimistic and assume the last transaction that modified the inode.
 	 */
-	BTRFS_I(inode)->last_reflink_trans = BTRFS_I(inode)->last_trans;
+	inode->last_reflink_trans = inode->last_trans;
 
 	path->slots[0]++;
-	if (inode->i_nlink != 1 ||
+	if (vfs_inode->i_nlink != 1 ||
 	    path->slots[0] >= btrfs_header_nritems(leaf))
 		goto cache_acl;
 
 	btrfs_item_key_to_cpu(leaf, &location, path->slots[0]);
-	if (location.objectid != btrfs_ino(BTRFS_I(inode)))
+	if (location.objectid != btrfs_ino(inode))
 		goto cache_acl;
 
 	ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
···
 		struct btrfs_inode_ref *ref;
 
 		ref = (struct btrfs_inode_ref *)ptr;
-		BTRFS_I(inode)->dir_index = btrfs_inode_ref_index(leaf, ref);
+		inode->dir_index = btrfs_inode_ref_index(leaf, ref);
 	} else if (location.type == BTRFS_INODE_EXTREF_KEY) {
 		struct btrfs_inode_extref *extref;
 
 		extref = (struct btrfs_inode_extref *)ptr;
-		BTRFS_I(inode)->dir_index = btrfs_inode_extref_index(leaf,
-								     extref);
+		inode->dir_index = btrfs_inode_extref_index(leaf, extref);
 	}
 cache_acl:
 	/*
···
 	 * any xattrs or acls
 	 */
 	maybe_acls = acls_after_inode_item(leaf, path->slots[0],
-			btrfs_ino(BTRFS_I(inode)), &first_xattr_slot);
+					   btrfs_ino(inode), &first_xattr_slot);
 	if (first_xattr_slot != -1) {
 		path->slots[0] = first_xattr_slot;
 		ret = btrfs_load_inode_props(inode, path);
 		if (ret)
 			btrfs_err(fs_info,
 				  "error loading props for ino %llu (root %llu): %d",
-				  btrfs_ino(BTRFS_I(inode)),
-				  btrfs_root_id(root), ret);
+				  btrfs_ino(inode), btrfs_root_id(root), ret);
 	}
 
 	if (!maybe_acls)
-		cache_no_acl(inode);
+		cache_no_acl(vfs_inode);
 
-	switch (inode->i_mode & S_IFMT) {
+	switch (vfs_inode->i_mode & S_IFMT) {
 	case S_IFREG:
-		inode->i_mapping->a_ops = &btrfs_aops;
-		inode->i_fop = &btrfs_file_operations;
-		inode->i_op = &btrfs_file_inode_operations;
+		vfs_inode->i_mapping->a_ops = &btrfs_aops;
+		vfs_inode->i_fop = &btrfs_file_operations;
+		vfs_inode->i_op = &btrfs_file_inode_operations;
 		break;
 	case S_IFDIR:
-		inode->i_fop = &btrfs_dir_file_operations;
-		inode->i_op = &btrfs_dir_inode_operations;
+		vfs_inode->i_fop = &btrfs_dir_file_operations;
+		vfs_inode->i_op = &btrfs_dir_inode_operations;
 		break;
 	case S_IFLNK:
-		inode->i_op = &btrfs_symlink_inode_operations;
-		inode_nohighmem(inode);
-		inode->i_mapping->a_ops = &btrfs_aops;
+		vfs_inode->i_op = &btrfs_symlink_inode_operations;
+		inode_nohighmem(vfs_inode);
+		vfs_inode->i_mapping->a_ops = &btrfs_aops;
 		break;
 	default:
-		inode->i_op = &btrfs_special_inode_operations;
-		init_special_inode(inode, inode->i_mode, rdev);
+		vfs_inode->i_op = &btrfs_special_inode_operations;
+		init_special_inode(vfs_inode, vfs_inode->i_mode, rdev);
 		break;
 	}
 
 	btrfs_sync_inode_flags_to_i_flags(inode);
 
-	ret = btrfs_add_inode_to_root(BTRFS_I(inode), true);
+	ret = btrfs_add_inode_to_root(inode, true);
 	if (ret)
 		goto out;
 
 	return 0;
 out:
-	iget_failed(inode);
+	iget_failed(vfs_inode);
 	return ret;
 }
 
···
 		args->root == BTRFS_I(inode)->root;
 }
 
-static struct inode *btrfs_iget_locked(u64 ino, struct btrfs_root *root)
+static struct btrfs_inode *btrfs_iget_locked(u64 ino, struct btrfs_root *root)
 {
 	struct inode *inode;
 	struct btrfs_iget_args args;
···
 	inode = iget5_locked_rcu(root->fs_info->sb, hashval, btrfs_find_actor,
 			     btrfs_init_locked_inode,
 			     (void *)&args);
-	return inode;
+	if (!inode)
+		return NULL;
+	return BTRFS_I(inode);
 }
 
 /*
  * Get an inode object given its inode number and corresponding root.  Path is
  * preallocated to prevent recursing back to iget through allocator.
  */
-struct inode *btrfs_iget_path(u64 ino, struct btrfs_root *root,
-			      struct btrfs_path *path)
+struct btrfs_inode *btrfs_iget_path(u64 ino, struct btrfs_root *root,
+				    struct btrfs_path *path)
 {
-	struct inode *inode;
+	struct btrfs_inode *inode;
 	int ret;
 
 	inode = btrfs_iget_locked(ino, root);
 	if (!inode)
 		return ERR_PTR(-ENOMEM);
 
-	if (!(inode->i_state & I_NEW))
+	if (!(inode->vfs_inode.i_state & I_NEW))
 		return inode;
 
 	ret = btrfs_read_locked_inode(inode, path);
 	if (ret)
 		return ERR_PTR(ret);
 
-	unlock_new_inode(inode);
+	unlock_new_inode(&inode->vfs_inode);
 	return inode;
 }
 
 /*
  * Get an inode object given its inode number and corresponding root.
  */
-struct inode *btrfs_iget(u64 ino, struct btrfs_root *root)
+struct btrfs_inode *btrfs_iget(u64 ino, struct btrfs_root *root)
 {
-	struct inode *inode;
+	struct btrfs_inode *inode;
 	struct btrfs_path *path;
 	int ret;
 
···
 	if (!inode)
 		return ERR_PTR(-ENOMEM);
 
-	if (!(inode->i_state & I_NEW))
+	if (!(inode->vfs_inode.i_state & I_NEW))
 		return inode;
 
 	path = btrfs_alloc_path();
···
 	if (ret)
 		return ERR_PTR(ret);
 
-	unlock_new_inode(inode);
+	unlock_new_inode(&inode->vfs_inode);
 	return inode;
 }
 
-static struct inode *new_simple_dir(struct inode *dir,
-				    struct btrfs_key *key,
-				    struct btrfs_root *root)
+static struct btrfs_inode *new_simple_dir(struct inode *dir,
+					  struct btrfs_key *key,
+					  struct btrfs_root *root)
 {
 	struct timespec64 ts;
-	struct inode *inode = new_inode(dir->i_sb);
+	struct inode *vfs_inode;
+	struct btrfs_inode *inode;
 
-	if (!inode)
+	vfs_inode = new_inode(dir->i_sb);
+	if (!vfs_inode)
 		return ERR_PTR(-ENOMEM);
 
-	BTRFS_I(inode)->root = btrfs_grab_root(root);
-	BTRFS_I(inode)->ref_root_id = key->objectid;
-	set_bit(BTRFS_INODE_ROOT_STUB, &BTRFS_I(inode)->runtime_flags);
-	set_bit(BTRFS_INODE_DUMMY, &BTRFS_I(inode)->runtime_flags);
+	inode = BTRFS_I(vfs_inode);
+	inode->root = btrfs_grab_root(root);
+	inode->ref_root_id = key->objectid;
+	set_bit(BTRFS_INODE_ROOT_STUB, &inode->runtime_flags);
+	set_bit(BTRFS_INODE_DUMMY, &inode->runtime_flags);
 
-	btrfs_set_inode_number(BTRFS_I(inode), BTRFS_EMPTY_SUBVOL_DIR_OBJECTID);
+	btrfs_set_inode_number(inode, BTRFS_EMPTY_SUBVOL_DIR_OBJECTID);
 	/*
 	 * We only need lookup, the rest is read-only and there's no inode
 	 * associated with the dentry
 	 */
-	inode->i_op = &simple_dir_inode_operations;
-	inode->i_opflags &= ~IOP_XATTR;
-	inode->i_fop = &simple_dir_operations;
-	inode->i_mode = S_IFDIR | S_IRUGO | S_IWUSR | S_IXUGO;
+	vfs_inode->i_op = &simple_dir_inode_operations;
+	vfs_inode->i_opflags &= ~IOP_XATTR;
+	vfs_inode->i_fop = &simple_dir_operations;
+	vfs_inode->i_mode = S_IFDIR | S_IRUGO | S_IWUSR | S_IXUGO;
 
-	ts = inode_set_ctime_current(inode);
-	inode_set_mtime_to_ts(inode, ts);
-	inode_set_atime_to_ts(inode, inode_get_atime(dir));
-	BTRFS_I(inode)->i_otime_sec = ts.tv_sec;
-	BTRFS_I(inode)->i_otime_nsec = ts.tv_nsec;
+	ts = inode_set_ctime_current(vfs_inode);
+	inode_set_mtime_to_ts(vfs_inode, ts);
+	inode_set_atime_to_ts(vfs_inode, inode_get_atime(dir));
+	inode->i_otime_sec = ts.tv_sec;
+	inode->i_otime_nsec = ts.tv_nsec;
 
-	inode->i_uid = dir->i_uid;
-	inode->i_gid = dir->i_gid;
+	vfs_inode->i_uid = dir->i_uid;
+	vfs_inode->i_gid = dir->i_gid;
 
 	return inode;
 }
···
 static_assert(BTRFS_FT_SOCK == FT_SOCK);
 static_assert(BTRFS_FT_SYMLINK == FT_SYMLINK);
 
-static inline u8 btrfs_inode_type(struct inode *inode)
+static inline u8 btrfs_inode_type(const struct btrfs_inode *inode)
 {
-	return fs_umode_to_ftype(inode->i_mode);
+	return fs_umode_to_ftype(inode->vfs_inode.i_mode);
 }
 
 struct inode *btrfs_lookup_dentry(struct inode *dir, struct dentry *dentry)
 {
 	struct btrfs_fs_info *fs_info = inode_to_fs_info(dir);
-	struct inode *inode;
+	struct btrfs_inode *inode;
 	struct btrfs_root *root = BTRFS_I(dir)->root;
 	struct btrfs_root *sub_root = root;
 	struct btrfs_key location = { 0 };
···
 	if (location.type == BTRFS_INODE_ITEM_KEY) {
 		inode = btrfs_iget(location.objectid, root);
 		if (IS_ERR(inode))
-			return inode;
+			return ERR_CAST(inode);
 
 		/* Do extra check against inode mode with di_type */
 		if (btrfs_inode_type(inode) != di_type) {
 			btrfs_crit(fs_info,
 "inode mode mismatch with dir: inode mode=0%o btrfs type=%u dir type=%u",
-				  inode->i_mode, btrfs_inode_type(inode),
+				  inode->vfs_inode.i_mode, btrfs_inode_type(inode),
 				  di_type);
-			iput(inode);
+			iput(&inode->vfs_inode);
 			return ERR_PTR(-EUCLEAN);
 		}
-		return inode;
+		return &inode->vfs_inode;
 	}
 
 	ret = fixup_tree_root_location(fs_info, BTRFS_I(dir), dentry,
···
 		btrfs_put_root(sub_root);
 
 		if (IS_ERR(inode))
-			return inode;
+			return ERR_CAST(inode);
 
 		down_read(&fs_info->cleanup_work_sem);
-		if (!sb_rdonly(inode->i_sb))
+		if (!sb_rdonly(inode->vfs_inode.i_sb))
 			ret = btrfs_orphan_cleanup(sub_root);
 		up_read(&fs_info->cleanup_work_sem);
 		if (ret) {
-			iput(inode);
+			iput(&inode->vfs_inode);
 			inode = ERR_PTR(ret);
 		}
 	}
 
-	return inode;
+	if (IS_ERR(inode))
+		return ERR_CAST(inode);
+
+	return &inode->vfs_inode;
 }
 
 static int btrfs_dentry_delete(const struct dentry *dentry)
···
 			inode->flags |= BTRFS_INODE_NODATASUM;
 	}
 
-	btrfs_sync_inode_flags_to_i_flags(&inode->vfs_inode);
+	btrfs_sync_inode_flags_to_i_flags(inode);
 }
 
 int btrfs_create_new_inode(struct btrfs_trans_handle *trans,
···
 		if (btrfs_test_opt(fs_info, NODATACOW))
 			BTRFS_I(inode)->flags |= BTRFS_INODE_NODATACOW |
 				BTRFS_INODE_NODATASUM;
+		btrfs_update_inode_mapping_flags(BTRFS_I(inode));
 	}
 
 	ret = btrfs_insert_inode_locked(inode);
···
 	path = NULL;
 
 	if (args->subvol) {
-		struct inode *parent;
+		struct btrfs_inode *parent;
 
 		/*
 		 * Subvolumes inherit properties from their parent subvolume,
···
 		if (IS_ERR(parent)) {
 			ret = PTR_ERR(parent);
 		} else {
-			ret = btrfs_inode_inherit_props(trans, inode, parent);
-			iput(parent);
+			ret = btrfs_inode_inherit_props(trans, BTRFS_I(inode),
+							parent);
+			iput(&parent->vfs_inode);
 		}
 	} else {
-		ret = btrfs_inode_inherit_props(trans, inode, dir);
+		ret = btrfs_inode_inherit_props(trans, BTRFS_I(inode),
+						BTRFS_I(dir));
 	}
 	if (ret) {
 		btrfs_err(fs_info,
···
 		return ret;
 
 	ret = btrfs_insert_dir_item(trans, name, parent_inode, &key,
-				    btrfs_inode_type(&inode->vfs_inode), index);
+				    btrfs_inode_type(inode), index);
 	if (ret == -EEXIST || ret == -EOVERFLOW)
 		goto fail_dir_item;
 	else if (ret) {
···
 {
 	int ret;
 	struct extent_buffer *leaf = path->nodes[0];
+	const u32 blocksize = leaf->fs_info->sectorsize;
 	char *tmp;
 	size_t max_size;
 	unsigned long inline_size;
···
 
 	read_extent_buffer(leaf, tmp, ptr, inline_size);
 
-	max_size = min_t(unsigned long, PAGE_SIZE, max_size);
+	max_size = min_t(unsigned long, blocksize, max_size);
 	ret = btrfs_decompress(compress_type, tmp, folio, 0, inline_size,
 			       max_size);
 
···
 	 * cover that region here.
 	 */
 
-	if (max_size < PAGE_SIZE)
-		folio_zero_range(folio, max_size, PAGE_SIZE - max_size);
+	if (max_size < blocksize)
+		folio_zero_range(folio, max_size, blocksize - max_size);
 	kfree(tmp);
 	return ret;
 }
 
 static int read_inline_extent(struct btrfs_path *path, struct folio *folio)
 {
+	const u32 blocksize = path->nodes[0]->fs_info->sectorsize;
 	struct btrfs_file_extent_item *fi;
 	void *kaddr;
 	size_t copy_size;
···
 	if (btrfs_file_extent_compression(path->nodes[0], fi) != BTRFS_COMPRESS_NONE)
 		return uncompress_inline(path, folio, fi);
 
-	copy_size = min_t(u64, PAGE_SIZE,
+	copy_size = min_t(u64, blocksize,
 			  btrfs_file_extent_ram_bytes(path->nodes[0], fi));
 	kaddr = kmap_local_folio(folio, 0);
 	read_extent_buffer(path->nodes[0], kaddr,
 			   btrfs_file_extent_inline_start(fi), copy_size);
 	kunmap_local(kaddr);
-	if (copy_size < PAGE_SIZE)
-		folio_zero_range(folio, copy_size, PAGE_SIZE - copy_size);
+	if (copy_size < blocksize)
+		folio_zero_range(folio, copy_size, blocksize - copy_size);
 	return 0;
 }
 
···
  * NOTE: This only checks the file extents, caller is responsible to wait for
  *	 any ordered extents.
  */
-noinline int can_nocow_extent(struct inode *inode, u64 offset, u64 *len,
+noinline int can_nocow_extent(struct btrfs_inode *inode, u64 offset, u64 *len,
 			      struct btrfs_file_extent *file_extent,
 			      bool nowait)
 {
-	struct btrfs_fs_info *fs_info = inode_to_fs_info(inode);
+	struct btrfs_root *root = inode->root;
+	struct btrfs_fs_info *fs_info = root->fs_info;
 	struct can_nocow_file_extent_args nocow_args = { 0 };
 	struct btrfs_path *path;
 	int ret;
 	struct extent_buffer *leaf;
-	struct btrfs_root *root = BTRFS_I(inode)->root;
-	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
+	struct extent_io_tree *io_tree = &inode->io_tree;
 	struct btrfs_file_extent_item *fi;
 	struct btrfs_key key;
 	int found_type;
···
 		return -ENOMEM;
 	path->nowait = nowait;
 
-	ret = btrfs_lookup_file_extent(NULL, root, path,
-			btrfs_ino(BTRFS_I(inode)), offset, 0);
+	ret = btrfs_lookup_file_extent(NULL, root, path, btrfs_ino(inode),
+				       offset, 0);
 	if (ret < 0)
 		goto out;
 
···
 	ret = 0;
 	leaf = path->nodes[0];
 	btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
-	if (key.objectid != btrfs_ino(BTRFS_I(inode)) ||
+	if (key.objectid != btrfs_ino(inode) ||
 	    key.type != BTRFS_EXTENT_DATA_KEY) {
 		/* not our file or wrong item type, must cow */
 		goto out;
···
 	nocow_args.end = offset + *len - 1;
 	nocow_args.free_path = true;
 
-	ret = can_nocow_file_extent(path, &key, BTRFS_I(inode), &nocow_args);
+	ret = can_nocow_file_extent(path, &key, inode, &nocow_args);
 	/* can_nocow_file_extent() has freed the path. */
 	path = NULL;
 
···
 				  nocow_args.file_extent.offset))
 		goto out;
 
-	if (!(BTRFS_I(inode)->flags & BTRFS_INODE_NODATACOW) &&
+	if (!(inode->flags & BTRFS_INODE_NODATACOW) &&
 	    found_type == BTRFS_FILE_EXTENT_PREALLOC) {
 		u64 range_end;
 
···
 	struct btrfs_fs_info *fs_info = folio_to_fs_info(folio);
 	struct btrfs_subpage *subpage;
 
-	if (!btrfs_is_subpage(fs_info, folio->mapping))
+	if (!btrfs_is_subpage(fs_info, folio))
 		return;
 
 	ASSERT(folio_test_private(folio) && folio_get_private(folio));
···
 static int btrfs_launder_folio(struct folio *folio)
 {
 	return btrfs_qgroup_free_data(folio_to_inode(folio), NULL, folio_pos(folio),
-				      PAGE_SIZE, NULL);
+				      folio_size(folio), NULL);
 }
 
 static bool __btrfs_release_folio(struct folio *folio, gfp_t gfp_flags)
···
 				 struct writeback_control *wbc, bool snapshot,
 				 bool in_reclaim_context)
 {
-	struct btrfs_inode *binode;
-	struct inode *inode;
 	struct btrfs_delalloc_work *work, *next;
 	LIST_HEAD(works);
 	LIST_HEAD(splice);
···
 	spin_lock(&root->delalloc_lock);
 	list_splice_init(&root->delalloc_inodes, &splice);
 	while (!list_empty(&splice)) {
-		binode = list_entry(splice.next, struct btrfs_inode,
-				    delalloc_inodes);
+		struct btrfs_inode *inode;
+		struct inode *tmp_inode;
 
-		list_move_tail(&binode->delalloc_inodes,
-			       &root->delalloc_inodes);
+		inode = list_entry(splice.next, struct btrfs_inode, delalloc_inodes);
+
+		list_move_tail(&inode->delalloc_inodes, &root->delalloc_inodes);
 
 		if (in_reclaim_context &&
-		    test_bit(BTRFS_INODE_NO_DELALLOC_FLUSH, &binode->runtime_flags))
+		    test_bit(BTRFS_INODE_NO_DELALLOC_FLUSH, &inode->runtime_flags))
 			continue;
 
-		inode = igrab(&binode->vfs_inode);
-		if (!inode) {
+		tmp_inode = igrab(&inode->vfs_inode);
+		if (!tmp_inode) {
 			cond_resched_lock(&root->delalloc_lock);
 			continue;
 		}
 		spin_unlock(&root->delalloc_lock);
 
 		if (snapshot)
-			set_bit(BTRFS_INODE_SNAPSHOT_FLUSH,
-				&binode->runtime_flags);
+			set_bit(BTRFS_INODE_SNAPSHOT_FLUSH, &inode->runtime_flags);
 		if (full_flush) {
-			work = btrfs_alloc_delalloc_work(inode);
+			work = btrfs_alloc_delalloc_work(&inode->vfs_inode);
 			if (!work) {
-				iput(inode);
+				iput(&inode->vfs_inode);
 				ret = -ENOMEM;
 				goto out;
 			}
···
 			btrfs_queue_work(root->fs_info->flush_workers,
 					 &work->work);
 		} else {
-			ret = filemap_fdatawrite_wbc(inode->i_mapping, wbc);
-			btrfs_add_delayed_iput(BTRFS_I(inode));
+			ret = filemap_fdatawrite_wbc(inode->vfs_inode.i_mapping, wbc);
+			btrfs_add_delayed_iput(inode);
 			if (ret || wbc->nr_to_write <= 0)
 				goto out;
 		}
···
 	struct extent_buffer *leaf;
 
 	name_len = strlen(symname);
-	if (name_len > BTRFS_MAX_INLINE_DATA_SIZE(fs_info))
+	/*
+	 * Symlinks utilize uncompressed inline extent data, which should not
+	 * reach block size.
+	 */
+	if (name_len > BTRFS_MAX_INLINE_DATA_SIZE(fs_info) ||
+	    name_len >= fs_info->sectorsize)
 		return -ENAMETOOLONG;
 
 	inode = new_inode(dir->i_sb);
···
 		goto out;
 	}
 	key.objectid = btrfs_ino(BTRFS_I(inode));
-	key.offset = 0;
 	key.type = BTRFS_EXTENT_DATA_KEY;
+	key.offset = 0;
 	datasize = btrfs_file_extent_calc_inline_size(name_len);
 	err = btrfs_insert_empty_item(trans, root, path, &key,
 				      datasize);
···
 }
 
 struct btrfs_encoded_read_private {
-	struct completion done;
+	struct completion *sync_reads;
 	void *uring_ctx;
 	refcount_t pending_refs;
 	blk_status_t status;
···
 
 	if (bbio->bio.bi_status) {
 		/*
-		 * The memory barrier implied by the atomic_dec_return() here
-		 * pairs with the memory barrier implied by the
-		 * atomic_dec_return() or io_wait_event() in
-		 * btrfs_encoded_read_regular_fill_pages() to ensure that this
-		 * write is observed before the load of status in
+		 * The memory barrier implied by the refcount_dec_and_test() here
+		 * pairs with the memory barrier implied by the refcount_dec_and_test()
+		 * in btrfs_encoded_read_regular_fill_pages() to ensure that
+		 * this write is observed before the load of status in
 		 * btrfs_encoded_read_regular_fill_pages().
 		 */
 		WRITE_ONCE(priv->status, bbio->bio.bi_status);
···
 			btrfs_uring_read_extent_endio(priv->uring_ctx, err);
 			kfree(priv);
 		} else {
-			complete(&priv->done);
+			complete(priv->sync_reads);
 		}
 	}
 	bio_put(&bbio->bio);
···
 					  struct page **pages, void *uring_ctx)
 {
 	struct btrfs_fs_info *fs_info = inode->root->fs_info;
-	struct btrfs_encoded_read_private *priv;
+	struct btrfs_encoded_read_private *priv, sync_priv;
+	struct completion sync_reads;
 	unsigned long i = 0;
 	struct btrfs_bio *bbio;
 	int ret;
 
-	priv = kmalloc(sizeof(struct btrfs_encoded_read_private), GFP_NOFS);
-	if (!priv)
-		return -ENOMEM;
+	/*
+	 * Fast path for synchronous reads which completes in this call, io_uring
+	 * needs longer time span.
+	 */
+	if (uring_ctx) {
+		priv = kmalloc(sizeof(struct btrfs_encoded_read_private), GFP_NOFS);
+		if (!priv)
+			return -ENOMEM;
+	} else {
+		priv = &sync_priv;
+		init_completion(&sync_reads);
+		priv->sync_reads = &sync_reads;
+	}
 
-	init_completion(&priv->done);
 	refcount_set(&priv->pending_refs, 1);
 	priv->status = 0;
 	priv->uring_ctx = uring_ctx;
···
 		return -EIOCBQUEUED;
 	} else {
 		if (!refcount_dec_and_test(&priv->pending_refs))
-			wait_for_completion_io(&priv->done);
+			wait_for_completion_io(&sync_reads);
 		/* See btrfs_encoded_read_endio() for ordering. */
-		ret = blk_status_to_errno(READ_ONCE(priv->status));
-		kfree(priv);
-		return ret;
+		return blk_status_to_errno(READ_ONCE(priv->status));
 	}
 }
 

fs/btrfs/ioctl.c

···
 #endif
 
 /* Mask out flags that are inappropriate for the given type of inode. */
-static unsigned int btrfs_mask_fsflags_for_type(struct inode *inode,
-		unsigned int flags)
+static unsigned int btrfs_mask_fsflags_for_type(const struct inode *inode,
+						unsigned int flags)
 {
 	if (S_ISDIR(inode->i_mode))
 		return flags;
···
  * Export internal inode flags to the format expected by the FS_IOC_GETFLAGS
  * ioctl.
  */
-static unsigned int btrfs_inode_flags_to_fsflags(struct btrfs_inode *binode)
+static unsigned int btrfs_inode_flags_to_fsflags(const struct btrfs_inode *inode)
 {
 	unsigned int iflags = 0;
-	u32 flags = binode->flags;
-	u32 ro_flags = binode->ro_flags;
+	u32 flags = inode->flags;
+	u32 ro_flags = inode->ro_flags;
 
 	if (flags & BTRFS_INODE_SYNC)
 		iflags |= FS_SYNC_FL;
···
 /*
  * Update inode->i_flags based on the btrfs internal flags.
  */
-void btrfs_sync_inode_flags_to_i_flags(struct inode *inode)
+void btrfs_sync_inode_flags_to_i_flags(struct btrfs_inode *inode)
 {
-	struct btrfs_inode *binode = BTRFS_I(inode);
 	unsigned int new_fl = 0;
 
-	if (binode->flags & BTRFS_INODE_SYNC)
+	if (inode->flags & BTRFS_INODE_SYNC)
 		new_fl |= S_SYNC;
-	if (binode->flags & BTRFS_INODE_IMMUTABLE)
+	if (inode->flags & BTRFS_INODE_IMMUTABLE)
 		new_fl |= S_IMMUTABLE;
-	if (binode->flags & BTRFS_INODE_APPEND)
+	if (inode->flags & BTRFS_INODE_APPEND)
 		new_fl |= S_APPEND;
-	if (binode->flags & BTRFS_INODE_NOATIME)
+	if (inode->flags & BTRFS_INODE_NOATIME)
 		new_fl |= S_NOATIME;
-	if (binode->flags & BTRFS_INODE_DIRSYNC)
+	if (inode->flags & BTRFS_INODE_DIRSYNC)
 		new_fl |= S_DIRSYNC;
-	if (binode->ro_flags & BTRFS_INODE_RO_VERITY)
+	if (inode->ro_flags & BTRFS_INODE_RO_VERITY)
 		new_fl |= S_VERITY;
 
-	set_mask_bits(&inode->i_flags,
+	set_mask_bits(&inode->vfs_inode.i_flags,
 		      S_SYNC | S_APPEND | S_IMMUTABLE | S_NOATIME | S_DIRSYNC |
 		      S_VERITY, new_fl);
 }
···
 	return 0;
 }
 
-static int check_fsflags_compatible(struct btrfs_fs_info *fs_info,
+static int check_fsflags_compatible(const struct btrfs_fs_info *fs_info,
 				    unsigned int flags)
 {
 	if (btrfs_is_zoned(fs_info) && (flags & FS_NOCOW_FL))
···
  */
 int btrfs_fileattr_get(struct dentry *dentry, struct fileattr *fa)
 {
-	struct btrfs_inode *binode = BTRFS_I(d_inode(dentry));
+	const struct btrfs_inode *inode = BTRFS_I(d_inode(dentry));
 
-	fileattr_fill_flags(fa, btrfs_inode_flags_to_fsflags(binode));
+	fileattr_fill_flags(fa, btrfs_inode_flags_to_fsflags(inode));
 	return 0;
 }
 
 int btrfs_fileattr_set(struct mnt_idmap *idmap,
 		       struct dentry *dentry, struct fileattr *fa)
 {
-	struct inode *inode = d_inode(dentry);
-	struct btrfs_fs_info *fs_info = inode_to_fs_info(inode);
-	struct btrfs_inode *binode = BTRFS_I(inode);
-	struct btrfs_root *root = binode->root;
+	struct btrfs_inode *inode = BTRFS_I(d_inode(dentry));
+	struct btrfs_root *root = inode->root;
+	struct btrfs_fs_info *fs_info = root->fs_info;
 	struct btrfs_trans_handle *trans;
 	unsigned int fsflags, old_fsflags;
 	int ret;
 	const char *comp = NULL;
-	u32 binode_flags;
+	u32 inode_flags;
 
 	if (btrfs_root_readonly(root))
 		return -EROFS;
···
 	if (fileattr_has_fsx(fa))
 		return -EOPNOTSUPP;
 
-	fsflags = btrfs_mask_fsflags_for_type(inode, fa->flags);
-	old_fsflags = btrfs_inode_flags_to_fsflags(binode);
+	fsflags = btrfs_mask_fsflags_for_type(&inode->vfs_inode, fa->flags);
+	old_fsflags = btrfs_inode_flags_to_fsflags(inode);
 	ret = check_fsflags(old_fsflags, fsflags);
 	if (ret)
 		return ret;
···
 	if (ret)
 		return ret;
 
-	binode_flags = binode->flags;
+	inode_flags = inode->flags;
 	if (fsflags & FS_SYNC_FL)
-		binode_flags |= BTRFS_INODE_SYNC;
+		inode_flags |= BTRFS_INODE_SYNC;
 	else
-		binode_flags &= ~BTRFS_INODE_SYNC;
+		inode_flags &= ~BTRFS_INODE_SYNC;
 	if (fsflags & FS_IMMUTABLE_FL)
-		binode_flags |= BTRFS_INODE_IMMUTABLE;
+		inode_flags |= BTRFS_INODE_IMMUTABLE;
 	else
-		binode_flags &= ~BTRFS_INODE_IMMUTABLE;
+		inode_flags &= ~BTRFS_INODE_IMMUTABLE;
 	if (fsflags & FS_APPEND_FL)
-		binode_flags |= BTRFS_INODE_APPEND;
+		inode_flags |= BTRFS_INODE_APPEND;
 	else
-		binode_flags &= ~BTRFS_INODE_APPEND;
+		inode_flags &= ~BTRFS_INODE_APPEND;
 	if (fsflags & FS_NODUMP_FL)
-		binode_flags |= BTRFS_INODE_NODUMP;
+		inode_flags |= BTRFS_INODE_NODUMP;
 	else
-		binode_flags &= ~BTRFS_INODE_NODUMP;
+		inode_flags &= ~BTRFS_INODE_NODUMP;
 	if (fsflags & FS_NOATIME_FL)
-		binode_flags |= BTRFS_INODE_NOATIME;
+		inode_flags |= BTRFS_INODE_NOATIME;
 	else
-		binode_flags &= ~BTRFS_INODE_NOATIME;
+		inode_flags &= ~BTRFS_INODE_NOATIME;
 
 	/* If coming from FS_IOC_FSSETXATTR then skip unconverted flags */
 	if (!fa->flags_valid) {
···
 	}
 
 	if (fsflags & FS_DIRSYNC_FL)
-		binode_flags |= BTRFS_INODE_DIRSYNC;
+		inode_flags |= BTRFS_INODE_DIRSYNC;
 	else
-		binode_flags &= ~BTRFS_INODE_DIRSYNC;
+		inode_flags &= ~BTRFS_INODE_DIRSYNC;
 	if (fsflags & FS_NOCOW_FL) {
-		if (S_ISREG(inode->i_mode)) {
+		if (S_ISREG(inode->vfs_inode.i_mode)) {
 			/*
 			 * It's safe to turn csums off here, no extents exist.
 			 * Otherwise we want the flag to reflect the real COW
 			 * status of the file and will not set it.
 			 */
-			if (inode->i_size == 0)
-				binode_flags |= BTRFS_INODE_NODATACOW |
-						BTRFS_INODE_NODATASUM;
+			if (inode->vfs_inode.i_size == 0)
+				inode_flags |= BTRFS_INODE_NODATACOW |
+					       BTRFS_INODE_NODATASUM;
 		} else {
-			binode_flags |= BTRFS_INODE_NODATACOW;
+			inode_flags |= BTRFS_INODE_NODATACOW;
 		}
 	} else {
 		/*
 		 * Revert back under same assumptions as above
 		 */
-		if (S_ISREG(inode->i_mode)) {
-			if (inode->i_size == 0)
-				binode_flags &= ~(BTRFS_INODE_NODATACOW |
-						  BTRFS_INODE_NODATASUM);
+		if (S_ISREG(inode->vfs_inode.i_mode)) {
+			if (inode->vfs_inode.i_size == 0)
+				inode_flags &= ~(BTRFS_INODE_NODATACOW |
+						 BTRFS_INODE_NODATASUM);
 		} else {
-			binode_flags &= ~BTRFS_INODE_NODATACOW;
+			inode_flags &= ~BTRFS_INODE_NODATACOW;
 		}
 	}
 
···
 	 * things smaller.
 	 */
 	if (fsflags & FS_NOCOMP_FL) {
-		binode_flags &= ~BTRFS_INODE_COMPRESS;
-		binode_flags |= BTRFS_INODE_NOCOMPRESS;
+		inode_flags &= ~BTRFS_INODE_COMPRESS;
+		inode_flags |= BTRFS_INODE_NOCOMPRESS;
 	} else if (fsflags & FS_COMPR_FL) {
 
-		if (IS_SWAPFILE(inode))
+		if (IS_SWAPFILE(&inode->vfs_inode))
 			return -ETXTBSY;
 
-		binode_flags |= BTRFS_INODE_COMPRESS;
-		binode_flags &= ~BTRFS_INODE_NOCOMPRESS;
+		inode_flags |= BTRFS_INODE_COMPRESS;
+		inode_flags &= ~BTRFS_INODE_NOCOMPRESS;
 
 		comp = btrfs_compress_type2str(fs_info->compress_type);
 		if (!comp || comp[0] == 0)
 			comp = btrfs_compress_type2str(BTRFS_COMPRESS_ZLIB);
 	} else {
-		binode_flags &= ~(BTRFS_INODE_COMPRESS | BTRFS_INODE_NOCOMPRESS);
+		inode_flags &= ~(BTRFS_INODE_COMPRESS | BTRFS_INODE_NOCOMPRESS);
 	}
 
 	/*
···
 		return PTR_ERR(trans);
 
 	if (comp) {
-		ret = btrfs_set_prop(trans, BTRFS_I(inode), "btrfs.compression",
+		ret = btrfs_set_prop(trans, inode, "btrfs.compression",
 				     comp, strlen(comp), 0);
 		if (ret) {
 			btrfs_abort_transaction(trans, ret);
 			goto out_end_trans;
 		}
 	} else {
-		ret = btrfs_set_prop(trans, BTRFS_I(inode), "btrfs.compression",
-				     NULL, 0, 0);
+		ret = btrfs_set_prop(trans, inode, "btrfs.compression", NULL, 0, 0);
 		if (ret && ret != -ENODATA) {
 			btrfs_abort_transaction(trans, ret);
 			goto out_end_trans;
···
 	}
 
 update_flags:
-	binode->flags = binode_flags;
+	inode->flags = inode_flags;
+	btrfs_update_inode_mapping_flags(inode);
 	btrfs_sync_inode_flags_to_i_flags(inode);
-	inode_inc_iversion(inode);
-	inode_set_ctime_current(inode);
-	ret = btrfs_update_inode(trans, BTRFS_I(inode));
+	inode_inc_iversion(&inode->vfs_inode);
+	inode_set_ctime_current(&inode->vfs_inode);
+	ret = btrfs_update_inode(trans, inode);
 
  out_end_trans:
 	btrfs_end_transaction(trans);
 	return ret;
 }
 
-static int btrfs_ioctl_getversion(struct inode *inode, int __user *arg)
+static int btrfs_ioctl_getversion(const struct inode *inode, int __user *arg)
 {
 	return put_user(inode->i_generation, arg);
 }
···
  * Calculate the number of transaction items to reserve for creating a subvolume
  * or snapshot, not including the inode, directory entries, or parent directory.
  */
-static unsigned int create_subvol_num_items(struct btrfs_qgroup_inherit *inherit)
+static unsigned int create_subvol_num_items(const struct btrfs_qgroup_inherit *inherit)
 {
 	/*
 	 * 1 to add root block
···
 	btrfs_set_root_dirid(root_item, BTRFS_FIRST_FREE_OBJECTID);
 
 	key.objectid = objectid;
-	key.offset = 0;
 	key.type = BTRFS_ROOT_ITEM_KEY;
+	key.offset = 0;
 	ret = btrfs_insert_root(trans, fs_info->tree_root, &key,
 				root_item);
 	if (ret) {
···
 
 /* copy of may_create in fs/namei.c() */
 static inline int btrfs_may_create(struct mnt_idmap *idmap,
-				   struct inode *dir, struct dentry *child)
+				   struct inode *dir, const struct dentry *child)
 {
 	if (d_really_is_positive(child))
 		return -EEXIST;
···
 					void __user *arg)
 {
 	BTRFS_DEV_LOOKUP_ARGS(args);
-	struct inode *inode = file_inode(file);
-	struct btrfs_fs_info *fs_info = inode_to_fs_info(inode);
+	struct btrfs_root *root = BTRFS_I(file_inode(file))->root;
+	struct btrfs_fs_info *fs_info = root->fs_info;
 	u64 new_size;
 	u64 old_size;
 	u64 devid = 1;
-	struct btrfs_root *root = BTRFS_I(inode)->root;
 	struct btrfs_ioctl_vol_args *vol_args;
-	struct btrfs_trans_handle *trans;
 	struct btrfs_device *device = NULL;
 	char *sizestr;
-	char *retptr;
 	char *devstr = NULL;
 	int ret = 0;
 	int mod = 0;
···
 	if (!strcmp(sizestr, "max"))
 		new_size = bdev_nr_bytes(device->bdev);
 	else {
+		char *retptr;
+
 		if (sizestr[0] == '-') {
 			mod = -1;
 			sizestr++;
···
 	new_size = round_down(new_size, fs_info->sectorsize);
 
 	if (new_size > old_size) {
+		struct btrfs_trans_handle *trans;
+
 		trans = btrfs_start_transaction(root, 0);
 		if (IS_ERR(trans)) {
 			ret = PTR_ERR(trans);
···
 	return ret;
 }
 
-static noinline int btrfs_ioctl_subvol_getflags(struct inode *inode,
+static noinline int btrfs_ioctl_subvol_getflags(struct btrfs_inode *inode,
 						void __user *arg)
 {
-	struct btrfs_fs_info *fs_info = inode_to_fs_info(inode);
-	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_root *root = inode->root;
+	struct btrfs_fs_info *fs_info = root->fs_info;
 	int ret = 0;
 	u64 flags = 0;
 
-	if (btrfs_ino(BTRFS_I(inode)) != BTRFS_FIRST_FREE_OBJECTID)
+	if (btrfs_ino(inode) != BTRFS_FIRST_FREE_OBJECTID)
 		return -EINVAL;
 
 	down_read(&fs_info->subvol_sem);
···
 	return ret;
 }
 
-static noinline int key_in_sk(struct btrfs_key *key,
-			      struct btrfs_ioctl_search_key *sk)
+static noinline int key_in_sk(const struct btrfs_key *key,
+			      const struct btrfs_ioctl_search_key *sk)
 {
 	struct btrfs_key test;
 	int ret;
···
 
 static noinline int copy_to_sk(struct btrfs_path *path,
 			       struct btrfs_key *key,
-			       struct btrfs_ioctl_search_key *sk,
+			       const struct btrfs_ioctl_search_key *sk,
 			       u64 *buf_size,
 			       char __user *ubuf,
 			       unsigned long *sk_offset,
···
 		}
 
 		sh.objectid = key->objectid;
-		sh.offset = key->offset;
 		sh.type = key->type;
+		sh.offset = key->offset;
 		sh.len = item_len;
 		sh.transid = found_transid;
 
···
 	return ret;
 }
 
-static noinline int search_ioctl(struct inode *inode,
+static noinline int search_ioctl(struct btrfs_root *root,
 				 struct btrfs_ioctl_search_key *sk,
 				 u64 *buf_size,
 				 char __user *ubuf)
 {
-	struct btrfs_fs_info *info = inode_to_fs_info(inode);
-	struct btrfs_root *root;
+	struct btrfs_fs_info *info = root->fs_info;
 	struct btrfs_key key;
 	struct btrfs_path *path;
 	int ret;
···
 		return -ENOMEM;
 
 	if (sk->tree_id == 0) {
-		/* search the root of the inode that was passed */
-		root = btrfs_grab_root(BTRFS_I(inode)->root);
+		/* Search the root that we got passed. */
+		root = btrfs_grab_root(root);
 	} else {
+		/* Look up the root from the arguments. */
 		root = btrfs_get_fs_root(info, sk->tree_id, true);
 		if (IS_ERR(root)) {
 			btrfs_free_path(path);
···
 	key.offset = sk->min_offset;
 
 	while (1) {
-		ret = -EFAULT;
 		/*
 		 * Ensure that the whole user buffer is faulted in at sub-page
 		 * granularity, otherwise the loop may live-lock.
 		 */
-		if (fault_in_subpage_writeable(ubuf + sk_offset,
-					       *buf_size - sk_offset))
+		if (fault_in_subpage_writeable(ubuf + sk_offset, *buf_size - sk_offset)) {
+			ret = -EFAULT;
 			break;
+		}
 
 		ret = btrfs_search_forward(root, &key, path, sk->min_transid);
-		if (ret != 0) {
-			if (ret > 0)
-				ret = 0;
-			goto err;
-		}
+		if (ret)
+			break;
+
 		ret = copy_to_sk(path, &key, sk, buf_size, ubuf,
 				 &sk_offset, &num_found);
 		btrfs_release_path(path);
···
 			break;
 
 	}
+	/* Normalize return values from btrfs_search_forward() and copy_to_sk(). */
 	if (ret > 0)
 		ret = 0;
-err:
+
 	sk->nr_items = num_found;
 	btrfs_put_root(root);
 	btrfs_free_path(path);
 	return ret;
 }
 
-static noinline int btrfs_ioctl_tree_search(struct inode *inode,
+static noinline int btrfs_ioctl_tree_search(struct btrfs_root *root,
 					    void __user *argp)
 {
 	struct btrfs_ioctl_search_args __user *uargs = argp;
···
 
 	buf_size = sizeof(uargs->buf);
 
-	ret = search_ioctl(inode, &sk, &buf_size, uargs->buf);
+	ret = search_ioctl(root, &sk, &buf_size, uargs->buf);
 
 	/*
 	 * In the origin implementation an overflow is handled by returning a
···
 	return ret;
 }
 
-static noinline int btrfs_ioctl_tree_search_v2(struct inode *inode,
+static noinline int btrfs_ioctl_tree_search_v2(struct btrfs_root *root,
 					       void __user *argp)
 {
 	struct btrfs_ioctl_search_args_v2 __user *uarg = argp;
···
 	if (buf_size > buf_limit)
 		buf_size = buf_limit;
 
-	ret = search_ioctl(inode, &args.key, &buf_size,
+	ret = search_ioctl(root, &args.key, &buf_size,
 			   (char __user *)(&uarg->buf[0]));
 	if (ret == 0 && copy_to_user(&uarg->key, &args.key, sizeof(args.key)))
 		ret = -EFAULT;
···
 	struct btrfs_path *path;
 	struct btrfs_key key, key2;
 	struct extent_buffer *leaf;
-	struct inode *temp_inode;
 	char *ptr;
 	int slot;
 	int len;
···
 		key.type = BTRFS_INODE_REF_KEY;
 		key.offset = (u64)-1;
 		while (1) {
+			struct btrfs_inode *temp_inode;
+
 			ret = btrfs_search_backwards(root, &key, path);
 			if (ret < 0)
 				goto out_put;
···
 				ret = PTR_ERR(temp_inode);
 				goto out_put;
 			}
-			ret = inode_permission(idmap, temp_inode,
+			ret = inode_permission(idmap, &temp_inode->vfs_inode,
 					       MAY_READ | MAY_EXEC);
-			iput(temp_inode);
+			iput(&temp_inode->vfs_inode);
 			if (ret) {
 				ret = -EACCES;
 				goto out_put;
···
 			/* the rest are all set to zero by kzalloc */
 			range.len = (u64)-1;
 		}
-		ret = btrfs_defrag_file(file_inode(file), &file->f_ra,
+		ret = btrfs_defrag_file(BTRFS_I(file_inode(file)), &file->f_ra,
 					&range, BTRFS_OLDEST_GENERATION, 0);
 		if (ret > 0)
 			ret = 0;
···
 	return ret;
 }
 
-static long btrfs_ioctl_fs_info(struct btrfs_fs_info *fs_info,
+static long btrfs_ioctl_fs_info(const struct btrfs_fs_info *fs_info,
 				void __user *arg)
 {
 	struct btrfs_ioctl_fs_info_args *fi_args;
···
 	return ret;
 }
 
-static long btrfs_ioctl_dev_info(struct btrfs_fs_info *fs_info,
+static long btrfs_ioctl_dev_info(const struct btrfs_fs_info *fs_info,
 				 void __user *arg)
 {
 	BTRFS_DEV_LOOKUP_ARGS(args);
···
 	return 0;
 }
 
-static int check_feature_bits(struct btrfs_fs_info *fs_info,
+static int check_feature_bits(const struct btrfs_fs_info *fs_info,
 			      enum btrfs_feature_set set,
 			      u64 change_mask, u64 flags, u64 supported_flags,
 			      u64 safe_set, u64 safe_clear)
···
 	return ret;
 }
 
-static int _btrfs_ioctl_send(struct btrfs_inode *inode, void __user *argp, bool compat)
+static int _btrfs_ioctl_send(struct btrfs_root *root, void __user *argp, bool compat)
 {
 	struct btrfs_ioctl_send_args *arg;
 	int ret;
···
 		if (IS_ERR(arg))
 			return PTR_ERR(arg);
 	}
-	ret = btrfs_ioctl_send(inode, arg);
+	ret = btrfs_ioctl_send(root, arg);
 	kfree(arg);
 	return ret;
 }
···
 	case BTRFS_IOC_SNAP_DESTROY_V2:
 		return btrfs_ioctl_snap_destroy(file, argp, true);
 	case BTRFS_IOC_SUBVOL_GETFLAGS:
-		return btrfs_ioctl_subvol_getflags(inode, argp);
+		return btrfs_ioctl_subvol_getflags(BTRFS_I(inode), argp);
 	case BTRFS_IOC_SUBVOL_SETFLAGS:
 		return btrfs_ioctl_subvol_setflags(file, argp);
 	case BTRFS_IOC_DEFAULT_SUBVOL:
···
 	case BTRFS_IOC_DEV_INFO:
 		return btrfs_ioctl_dev_info(fs_info, argp);
 	case BTRFS_IOC_TREE_SEARCH:
-		return btrfs_ioctl_tree_search(inode, argp);
+		return btrfs_ioctl_tree_search(root, argp);
 	case BTRFS_IOC_TREE_SEARCH_V2:
-		return btrfs_ioctl_tree_search_v2(inode, argp);
+		return btrfs_ioctl_tree_search_v2(root, argp);
 	case BTRFS_IOC_INO_LOOKUP:
 		return btrfs_ioctl_ino_lookup(root, argp);
 	case BTRFS_IOC_INO_PATHS:
···
 		return btrfs_ioctl_set_received_subvol_32(file, argp);
 #endif
 	case BTRFS_IOC_SEND:
-		return _btrfs_ioctl_send(BTRFS_I(inode), argp, false);
+		return _btrfs_ioctl_send(root, argp, false);
 #if defined(CONFIG_64BIT) && defined(CONFIG_COMPAT)
 	case BTRFS_IOC_SEND_32:
-		return _btrfs_ioctl_send(BTRFS_I(inode), argp, true);
+		return _btrfs_ioctl_send(root, argp, true);
 #endif
 	case BTRFS_IOC_GET_DEV_STATS:
 		return btrfs_ioctl_get_dev_stats(fs_info, argp);

fs/btrfs/ioctl.h

···
 struct dentry;
 struct mnt_idmap;
 struct fileattr;
+struct io_uring_cmd;
+struct btrfs_inode;
 struct btrfs_fs_info;
 struct btrfs_ioctl_balance_args;
 
···
 int btrfs_fileattr_set(struct mnt_idmap *idmap,
 		       struct dentry *dentry, struct fileattr *fa);
 int btrfs_ioctl_get_supported_features(void __user *arg);
-void btrfs_sync_inode_flags_to_i_flags(struct inode *inode);
+void btrfs_sync_inode_flags_to_i_flags(struct btrfs_inode *inode);
 void btrfs_update_ioctl_balance_args(struct btrfs_fs_info *fs_info,
 				     struct btrfs_ioctl_balance_args *bargs);
 int btrfs_uring_cmd(struct io_uring_cmd *cmd, unsigned int issue_flags);

fs/btrfs/locking.c

···
 #include <linux/page-flags.h>
 #include <asm/bug.h>
 #include <trace/events/btrfs.h>
-#include "misc.h"
 #include "ctree.h"
 #include "extent_io.h"
 #include "locking.h"

fs/btrfs/ordered-data.c

···
 /*
  * Start IO and wait for a given ordered extent to finish.
  *
- * Wait on page writeback for all the pages in the extent and the IO completion
- * code to insert metadata into the btree corresponding to the extent.
+ * Wait on page writeback for all the pages in the extent but not in
+ * [@nowriteback_start, @nowriteback_start + @nowriteback_len) and the
+ * IO completion code to insert metadata into the btree corresponding to the extent.
  */
-void btrfs_start_ordered_extent(struct btrfs_ordered_extent *entry)
+void btrfs_start_ordered_extent_nowriteback(struct btrfs_ordered_extent *entry,
+					    u64 nowriteback_start, u32 nowriteback_len)
 {
 	u64 start = entry->file_offset;
 	u64 end = start + entry->num_bytes - 1;
···
 	 * start IO on any dirty ones so the wait doesn't stall waiting
 	 * for the flusher thread to find them
 	 */
-	if (!test_bit(BTRFS_ORDERED_DIRECT, &entry->flags))
-		filemap_fdatawrite_range(inode->vfs_inode.i_mapping, start, end);
+	if (!test_bit(BTRFS_ORDERED_DIRECT, &entry->flags)) {
+		if (!nowriteback_len) {
+			filemap_fdatawrite_range(inode->vfs_inode.i_mapping, start, end);
+		} else {
+			if (start < nowriteback_start)
+				filemap_fdatawrite_range(inode->vfs_inode.i_mapping, start,
+							 nowriteback_start - 1);
+			if (nowriteback_start + nowriteback_len < end)
+				filemap_fdatawrite_range(inode->vfs_inode.i_mapping,
+							 nowriteback_start + nowriteback_len,
+							 end);
+		}
+	}
 
 	if (!freespace_inode)
 		btrfs_might_wait_for_event(inode->root->fs_info, btrfs_ordered_extent);

fs/btrfs/ordered-data.h

···
 struct inode;
 struct page;
 struct extent_state;
+struct btrfs_block_group;
 struct btrfs_inode;
 struct btrfs_root;
 struct btrfs_fs_info;
···
 			   struct btrfs_ordered_sum *sum);
 struct btrfs_ordered_extent *btrfs_lookup_ordered_extent(struct btrfs_inode *inode,
 							 u64 file_offset);
-void btrfs_start_ordered_extent(struct btrfs_ordered_extent *entry);
+void btrfs_start_ordered_extent_nowriteback(struct btrfs_ordered_extent *entry,
+				u64 nowriteback_start, u32 nowriteback_len);
+static inline void btrfs_start_ordered_extent(struct btrfs_ordered_extent *entry)
+{
+	return btrfs_start_ordered_extent_nowriteback(entry, 0, 0);
+}
+
 int btrfs_wait_ordered_range(struct btrfs_inode *inode, u64 start, u64 len);
 struct btrfs_ordered_extent *
 btrfs_lookup_first_ordered_extent(struct btrfs_inode *inode, u64 file_offset);

fs/btrfs/print-tree.h

···
 #ifndef BTRFS_PRINT_TREE_H
 #define BTRFS_PRINT_TREE_H
 
+#include <linux/types.h>
+
 /* Buffer size to contain tree name and possibly additional data (offset) */
 #define BTRFS_ROOT_NAME_BUF_LEN				48
 

fs/btrfs/props.c

···
 	const char *xattr_name;
 	int (*validate)(const struct btrfs_inode *inode, const char *value,
 			size_t len);
-	int (*apply)(struct inode *inode, const char *value, size_t len);
-	const char *(*extract)(const struct inode *inode);
+	int (*apply)(struct btrfs_inode *inode, const char *value, size_t len);
+	const char *(*extract)(const struct btrfs_inode *inode);
 	bool (*ignore)(const struct btrfs_inode *inode);
 	int inheritable;
 };
···
 		if (ret)
 			return ret;
 
-		ret = handler->apply(&inode->vfs_inode, NULL, 0);
+		ret = handler->apply(inode, NULL, 0);
 		ASSERT(ret == 0);
 
 		return ret;
···
 			     value_len, flags);
 	if (ret)
 		return ret;
-	ret = handler->apply(&inode->vfs_inode, value, value_len);
+	ret = handler->apply(inode, value, value_len);
 	if (ret) {
 		btrfs_setxattr(trans, &inode->vfs_inode, handler->xattr_name, NULL,
 			       0, flags);
···
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	int ret;
 
-	ret = handler->apply(inode, value, len);
+	ret = handler->apply(BTRFS_I(inode), value, len);
 	if (unlikely(ret))
 		btrfs_warn(root->fs_info,
 			   "error applying prop %s to ino %llu (root %llu): %d",
···
 		set_bit(BTRFS_INODE_HAS_PROPS, &BTRFS_I(inode)->runtime_flags);
 }
 
-int btrfs_load_inode_props(struct inode *inode, struct btrfs_path *path)
+int btrfs_load_inode_props(struct btrfs_inode *inode, struct btrfs_path *path)
 {
-	struct btrfs_root *root = BTRFS_I(inode)->root;
-	u64 ino = btrfs_ino(BTRFS_I(inode));
+	struct btrfs_root *root = inode->root;
+	u64 ino = btrfs_ino(inode);
 
-	return iterate_object_props(root, path, ino, inode_prop_iterator, inode);
+	return iterate_object_props(root, path, ino, inode_prop_iterator,
+				    &inode->vfs_inode);
 }
 
 static int prop_compression_validate(const struct btrfs_inode *inode,
···
 	return -EINVAL;
 }
 
-static int prop_compression_apply(struct inode *inode, const char *value,
+static int prop_compression_apply(struct btrfs_inode *inode, const char *value,
 				  size_t len)
 {
-	struct btrfs_fs_info *fs_info = inode_to_fs_info(inode);
+	struct btrfs_fs_info *fs_info = inode->root->fs_info;
 	int type;
 
 	/* Reset to defaults */
 	if (len == 0) {
-		BTRFS_I(inode)->flags &= ~BTRFS_INODE_COMPRESS;
-		BTRFS_I(inode)->flags &= ~BTRFS_INODE_NOCOMPRESS;
-		BTRFS_I(inode)->prop_compress = BTRFS_COMPRESS_NONE;
+		inode->flags &= ~BTRFS_INODE_COMPRESS;
+		inode->flags &= ~BTRFS_INODE_NOCOMPRESS;
+		inode->prop_compress = BTRFS_COMPRESS_NONE;
 		return 0;
 	}
 
 	/* Set NOCOMPRESS flag */
 	if ((len == 2 && strncmp("no", value, 2) == 0) ||
 	    (len == 4 && strncmp("none", value, 4) == 0)) {
-		BTRFS_I(inode)->flags |= BTRFS_INODE_NOCOMPRESS;
-		BTRFS_I(inode)->flags &= ~BTRFS_INODE_COMPRESS;
-		BTRFS_I(inode)->prop_compress = BTRFS_COMPRESS_NONE;
+		inode->flags |= BTRFS_INODE_NOCOMPRESS;
+		inode->flags &= ~BTRFS_INODE_COMPRESS;
+		inode->prop_compress = BTRFS_COMPRESS_NONE;
 
 		return 0;
 	}
···
 		return -EINVAL;
 	}
 
-	BTRFS_I(inode)->flags &= ~BTRFS_INODE_NOCOMPRESS;
-	BTRFS_I(inode)->flags |= BTRFS_INODE_COMPRESS;
-	BTRFS_I(inode)->prop_compress = type;
+	inode->flags &= ~BTRFS_INODE_NOCOMPRESS;
+	inode->flags |= BTRFS_INODE_COMPRESS;
+	inode->prop_compress = type;
 
 	return 0;
 }
···
 	return false;
 }
 
-static const char *prop_compression_extract(const struct inode *inode)
+static const char *prop_compression_extract(const struct btrfs_inode *inode)
 {
-	switch (BTRFS_I(inode)->prop_compress) {
+	switch (inode->prop_compress) {
 	case BTRFS_COMPRESS_ZLIB:
 	case BTRFS_COMPRESS_LZO:
 	case BTRFS_COMPRESS_ZSTD:
-		return btrfs_compress_type2str(BTRFS_I(inode)->prop_compress);
+		return btrfs_compress_type2str(inode->prop_compress);
 	default:
 		break;
 	}
···
 };
 
 int btrfs_inode_inherit_props(struct btrfs_trans_handle *trans,
-			      struct inode *inode, const struct inode *parent)
+			      struct btrfs_inode *inode,
+			      const struct btrfs_inode *parent)
 {
-	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_root *root = inode->root;
 	struct btrfs_fs_info *fs_info = root->fs_info;
 	int ret;
 	int i;
 	bool need_reserve = false;
 
-	if (!test_bit(BTRFS_INODE_HAS_PROPS,
-		      &BTRFS_I(parent)->runtime_flags))
+	if (!test_bit(BTRFS_INODE_HAS_PROPS, &parent->runtime_flags))
 		return 0;
 
 	for (i = 0; i < ARRAY_SIZE(prop_handlers); i++) {
···
 		if (!h->inheritable)
 			continue;
 
-		if (h->ignore(BTRFS_I(inode)))
+		if (h->ignore(inode))
 			continue;
 
 		value = h->extract(parent);
···
 		 * This is not strictly necessary as the property should be
 		 * valid, but in case it isn't, don't propagate it further.
 		 */
-		ret = h->validate(BTRFS_I(inode), value, strlen(value));
+		ret = h->validate(inode, value, strlen(value));
 		if (ret)
 			continue;
 
···
 				return ret;
 		}
 
-		ret = btrfs_setxattr(trans, inode, h->xattr_name, value,
+		ret = btrfs_setxattr(trans, &inode->vfs_inode, h->xattr_name, value,
 				     strlen(value), 0);
 		if (!ret) {
 			ret = h->apply(inode, value, strlen(value));
 			if (ret)
-				btrfs_setxattr(trans, inode, h->xattr_name,
+				btrfs_setxattr(trans, &inode->vfs_inode, h->xattr_name,
 					       NULL, 0, 0);
 			else
-				set_bit(BTRFS_INODE_HAS_PROPS,
-					&BTRFS_I(inode)->runtime_flags);
+				set_bit(BTRFS_INODE_HAS_PROPS, &inode->runtime_flags);
 		}
 
 		if (need_reserve) {

fs/btrfs/props.h

···
 #ifndef BTRFS_PROPS_H
 #define BTRFS_PROPS_H
 
+#include <linux/types.h>
 #include <linux/compiler_types.h>
 
-struct inode;
 struct btrfs_inode;
 struct btrfs_path;
 struct btrfs_trans_handle;
···
 			const char *value, size_t value_len);
 bool btrfs_ignore_prop(const struct btrfs_inode *inode, const char *name);
 
-int btrfs_load_inode_props(struct inode *inode, struct btrfs_path *path);
+int btrfs_load_inode_props(struct btrfs_inode *inode, struct btrfs_path *path);
 
 int btrfs_inode_inherit_props(struct btrfs_trans_handle *trans,
-			      struct inode *inode,
-			      const struct inode *dir);
+			      struct btrfs_inode *inode,
+			      const struct btrfs_inode *dir);
 
 #endif

fs/btrfs/qgroup.c

···
 		return -ENOMEM;
 
 	key.objectid = 0;
-	key.offset = 0;
 	key.type = 0;
+	key.offset = 0;
 
 	while (1) {
 		ret = btrfs_search_slot(trans, root, &key, path, -1, 1);

fs/btrfs/qgroup.h

···
 struct btrfs_trans_handle;
 struct btrfs_delayed_ref_root;
 struct btrfs_inode;
+struct btrfs_transaction;
+struct btrfs_block_group;
+struct btrfs_qgroup_swapped_blocks;
 
 /*
  * Btrfs qgroup overview

fs/btrfs/raid-stripe-tree.h

···
 #include <linux/types.h>
 #include <uapi/linux/btrfs_tree.h>
 #include "fs.h"
+#include "accessors.h"
 
 #define BTRFS_RST_SUPP_BLOCK_GROUP_MASK    (BTRFS_BLOCK_GROUP_DUP |		\
 					    BTRFS_BLOCK_GROUP_RAID1_MASK |	\

fs/btrfs/reflink.c

···
  * the source inode to destination inode when possible. When not possible we
  * copy the inline extent's data into the respective page of the inode.
  */
-static int clone_copy_inline_extent(struct inode *dst,
+static int clone_copy_inline_extent(struct btrfs_inode *inode,
 				    struct btrfs_path *path,
 				    struct btrfs_key *new_key,
 				    const u64 drop_start,
···
 				    char *inline_data,
 				    struct btrfs_trans_handle **trans_out)
 {
-	struct btrfs_fs_info *fs_info = inode_to_fs_info(dst);
-	struct btrfs_root *root = BTRFS_I(dst)->root;
+	struct btrfs_root *root = inode->root;
+	struct btrfs_fs_info *fs_info = root->fs_info;
 	const u64 aligned_end = ALIGN(new_key->offset + datal,
 				      fs_info->sectorsize);
 	struct btrfs_trans_handle *trans = NULL;
···
 	struct btrfs_key key;
 
 	if (new_key->offset > 0) {
-		ret = copy_inline_to_page(BTRFS_I(dst), new_key->offset,
+		ret = copy_inline_to_page(inode, new_key->offset,
 					  inline_data, size, datal, comp_type);
 		goto out;
 	}
 
-	key.objectid = btrfs_ino(BTRFS_I(dst));
+	key.objectid = btrfs_ino(inode);
 	key.type = BTRFS_EXTENT_DATA_KEY;
 	key.offset = 0;
 	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
···
 				goto copy_inline_extent;
 		}
 		btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
-		if (key.objectid == btrfs_ino(BTRFS_I(dst)) &&
+		if (key.objectid == btrfs_ino(inode) &&
 		    key.type == BTRFS_EXTENT_DATA_KEY) {
 			/*
 			 * There's an implicit hole at file offset 0, copy the
···
 			ASSERT(key.offset > 0);
 			goto copy_to_page;
 		}
-	} else if (i_size_read(dst) <= datal) {
+	} else if (i_size_read(&inode->vfs_inode) <= datal) {
 		struct btrfs_file_extent_item *ei;
 
 		ei = btrfs_item_ptr(path->nodes[0], path->slots[0],
···
 	 * We have no extent items, or we have an extent at offset 0 which may
 	 * or may not be inlined. All these cases are dealt the same way.
 	 */
-	if (i_size_read(dst) > datal) {
+	if (i_size_read(&inode->vfs_inode) > datal) {
 		/*
 		 * At the destination offset 0 we have either a hole, a regular
 		 * extent or an inline extent larger then the one we want to
···
 	drop_args.start = drop_start;
 	drop_args.end = aligned_end;
 	drop_args.drop_cache = true;
-	ret = btrfs_drop_extents(trans, root, BTRFS_I(dst), &drop_args);
+	ret = btrfs_drop_extents(trans, root, inode, &drop_args);
 	if (ret)
 		goto out;
 	ret = btrfs_insert_empty_item(trans, root, path, new_key, size);
···
 			    btrfs_item_ptr_offset(path->nodes[0],
 						  path->slots[0]),
 			    size);
-	btrfs_update_inode_bytes(BTRFS_I(dst), datal, drop_args.bytes_found);
-	btrfs_set_inode_full_sync(BTRFS_I(dst));
-	ret = btrfs_inode_set_file_extent_range(BTRFS_I(dst), 0, aligned_end);
+	btrfs_update_inode_bytes(inode, datal, drop_args.bytes_found);
+	btrfs_set_inode_full_sync(inode);
+	ret = btrfs_inode_set_file_extent_range(inode, 0, aligned_end);
 out:
 	if (!ret && !trans) {
 		/*
···
 	 */
 	btrfs_release_path(path);
 
-	ret = copy_inline_to_page(BTRFS_I(dst), new_key->offset,
+	ret = copy_inline_to_page(inode, new_key->offset,
 				  inline_data, size, datal, comp_type);
 	goto out;
 }
···
 				goto out;
 			}
 
-			ret = clone_copy_inline_extent(inode, path, &new_key,
+			ret = clone_copy_inline_extent(BTRFS_I(inode), path, &new_key,
 						       drop_start, datal, size,
 						       comp, buf, &trans);
 			if (ret)
···
 	return ret;
 }
 
-static void btrfs_double_mmap_lock(struct inode *inode1, struct inode *inode2)
+static void btrfs_double_mmap_lock(struct btrfs_inode *inode1, struct btrfs_inode *inode2)
 {
 	if (inode1 < inode2)
 		swap(inode1, inode2);
-	down_write(&BTRFS_I(inode1)->i_mmap_lock);
-	down_write_nested(&BTRFS_I(inode2)->i_mmap_lock, SINGLE_DEPTH_NESTING);
+	down_write(&inode1->i_mmap_lock);
+	down_write_nested(&inode2->i_mmap_lock, SINGLE_DEPTH_NESTING);
 }
 
-static void btrfs_double_mmap_unlock(struct inode *inode1, struct inode *inode2)
+static void btrfs_double_mmap_unlock(struct btrfs_inode *inode1, struct btrfs_inode *inode2)
 {
-	up_write(&BTRFS_I(inode1)->i_mmap_lock);
-	up_write(&BTRFS_I(inode2)->i_mmap_lock);
+	up_write(&inode1->i_mmap_lock);
+	up_write(&inode2->i_mmap_lock);
 }
 
-static int btrfs_extent_same_range(struct inode *src, u64 loff, u64 len,
-				   struct inode *dst, u64 dst_loff)
+static int btrfs_extent_same_range(struct btrfs_inode *src, u64 loff, u64 len,
+				   struct btrfs_inode *dst, u64 dst_loff)
 {
 	const u64 end = dst_loff + len - 1;
 	struct extent_state *cached_state = NULL;
-	struct btrfs_fs_info *fs_info = BTRFS_I(src)->root->fs_info;
+	struct btrfs_fs_info *fs_info = src->root->fs_info;
 	const u64 bs = fs_info->sectorsize;
 	int ret;
 
···
 	 * because we have already locked the inode's i_mmap_lock in exclusive
 	 * mode.
 	 */
-	lock_extent(&BTRFS_I(dst)->io_tree, dst_loff, end, &cached_state);
-	ret = btrfs_clone(src, dst, loff, len, ALIGN(len, bs), dst_loff, 1);
-	unlock_extent(&BTRFS_I(dst)->io_tree, dst_loff, end, &cached_state);
+	lock_extent(&dst->io_tree, dst_loff, end, &cached_state);
+	ret = btrfs_clone(&src->vfs_inode, &dst->vfs_inode, loff, len,
+			  ALIGN(len, bs), dst_loff, 1);
+	unlock_extent(&dst->io_tree, dst_loff, end, &cached_state);
 
 	btrfs_btree_balance_dirty(fs_info);
 
···
 	chunk_count = div_u64(olen, BTRFS_MAX_DEDUPE_LEN);
 
 	for (i = 0; i < chunk_count; i++) {
-		ret = btrfs_extent_same_range(src, loff, BTRFS_MAX_DEDUPE_LEN,
-					      dst, dst_loff);
+		ret = btrfs_extent_same_range(BTRFS_I(src), loff, BTRFS_MAX_DEDUPE_LEN,
+					      BTRFS_I(dst), dst_loff);
 		if (ret)
 			goto out;
 
···
 	}
 
 	if (tail_len > 0)
-		ret = btrfs_extent_same_range(src, loff, tail_len, dst, dst_loff);
+		ret = btrfs_extent_same_range(BTRFS_I(src), loff, tail_len,
+					      BTRFS_I(dst), dst_loff);
 out:
 	spin_lock(&root_dst->root_item_lock);
 	root_dst->dedupe_in_progress--;
···
 				       struct file *file_out, loff_t pos_out,
 				       loff_t *len, unsigned int remap_flags)
 {
-	struct inode *inode_in = file_inode(file_in);
-	struct inode *inode_out = file_inode(file_out);
-	u64 bs = BTRFS_I(inode_out)->root->fs_info->sectorsize;
+	struct btrfs_inode *inode_in = BTRFS_I(file_inode(file_in));
+	struct btrfs_inode *inode_out = BTRFS_I(file_inode(file_out));
+	u64 bs = inode_out->root->fs_info->sectorsize;
 	u64 wb_len;
 	int ret;
 
 	if (!(remap_flags & REMAP_FILE_DEDUP)) {
-		struct btrfs_root *root_out = BTRFS_I(inode_out)->root;
+		struct btrfs_root *root_out = inode_out->root;
 
 		if (btrfs_root_readonly(root_out))
 			return -EROFS;
 
-		ASSERT(inode_in->i_sb == inode_out->i_sb);
+		ASSERT(inode_in->vfs_inode.i_sb == inode_out->vfs_inode.i_sb);
 	}
 
 	/* Don't make the dst file partly checksummed */
-	if ((BTRFS_I(inode_in)->flags & BTRFS_INODE_NODATASUM) !=
-	    (BTRFS_I(inode_out)->flags & BTRFS_INODE_NODATASUM)) {
+	if ((inode_in->flags & BTRFS_INODE_NODATASUM) !=
+	    (inode_out->flags & BTRFS_INODE_NODATASUM)) {
 		return -EINVAL;
 	}
 
···
 	 *    to complete so that new file extent items are in the fs tree.
 	 */
 	if (*len == 0 && !(remap_flags & REMAP_FILE_DEDUP))
-		wb_len = ALIGN(inode_in->i_size, bs) - ALIGN_DOWN(pos_in, bs);
+		wb_len = ALIGN(inode_in->vfs_inode.i_size, bs) - ALIGN_DOWN(pos_in, bs);
 	else
 		wb_len = ALIGN(*len, bs);
 
···
 	 * Also we don't need to check ASYNC_EXTENT, as async extent will be
 	 * CoWed anyway, not affecting nocow part.
 	 */
-	ret = filemap_flush(inode_in->i_mapping);
+	ret = filemap_flush(inode_in->vfs_inode.i_mapping);
 	if (ret < 0)
 		return ret;
 
-	ret = btrfs_wait_ordered_range(BTRFS_I(inode_in), ALIGN_DOWN(pos_in, bs),
-				       wb_len);
+	ret = btrfs_wait_ordered_range(inode_in, ALIGN_DOWN(pos_in, bs), wb_len);
 	if (ret < 0)
 		return ret;
-	ret = btrfs_wait_ordered_range(BTRFS_I(inode_out), ALIGN_DOWN(pos_out, bs),
-				       wb_len);
+	ret = btrfs_wait_ordered_range(inode_out, ALIGN_DOWN(pos_out, bs), wb_len);
 	if (ret < 0)
 		return ret;
 
···
 		struct file *dst_file, loff_t destoff, loff_t len,
 		unsigned int remap_flags)
 {
-	struct inode *src_inode = file_inode(src_file);
-	struct inode *dst_inode = file_inode(dst_file);
+	struct btrfs_inode *src_inode = BTRFS_I(file_inode(src_file));
+	struct btrfs_inode *dst_inode = BTRFS_I(file_inode(dst_file));
 	bool same_inode = dst_inode == src_inode;
 	int ret;
 
···
 		return -EINVAL;
 
 	if (same_inode) {
-		btrfs_inode_lock(BTRFS_I(src_inode), BTRFS_ILOCK_MMAP);
+		btrfs_inode_lock(src_inode, BTRFS_ILOCK_MMAP);
 	} else {
-		lock_two_nondirectories(src_inode, dst_inode);
+		lock_two_nondirectories(&src_inode->vfs_inode, &dst_inode->vfs_inode);
 		btrfs_double_mmap_lock(src_inode, dst_inode);
 	}
 
···
 		goto out_unlock;
 
 	if (remap_flags & REMAP_FILE_DEDUP)
-		ret = btrfs_extent_same(src_inode, off, len, dst_inode, destoff);
+		ret = btrfs_extent_same(&src_inode->vfs_inode, off, len,
+					&dst_inode->vfs_inode, destoff);
 	else
 		ret = btrfs_clone_files(dst_file, src_file, off, len, destoff);
 
 out_unlock:
 	if (same_inode) {
-		btrfs_inode_unlock(BTRFS_I(src_inode), BTRFS_ILOCK_MMAP);
+		btrfs_inode_unlock(src_inode, BTRFS_ILOCK_MMAP);
 	} else {
 		btrfs_double_mmap_unlock(src_inode, dst_inode);
-		unlock_two_nondirectories(src_inode, dst_inode);
+		unlock_two_nondirectories(&src_inode->vfs_inode,
+					  &dst_inode->vfs_inode);
 	}
 
 	/*

fs/btrfs/relocation.c

···
 	return ret;
 }
 
-static int delete_block_group_cache(struct btrfs_fs_info *fs_info,
-				    struct btrfs_block_group *block_group,
+static int delete_block_group_cache(struct btrfs_block_group *block_group,
 				    struct inode *inode,
 				    u64 ino)
 {
+	struct btrfs_fs_info *fs_info = block_group->fs_info;
 	struct btrfs_root *root = fs_info->tree_root;
 	struct btrfs_trans_handle *trans;
+	struct btrfs_inode *btrfs_inode;
 	int ret = 0;
 
 	if (inode)
 		goto truncate;
 
-	inode = btrfs_iget(ino, root);
-	if (IS_ERR(inode))
+	btrfs_inode = btrfs_iget(ino, root);
+	if (IS_ERR(btrfs_inode))
 		return -ENOENT;
+	inode = &btrfs_inode->vfs_inode;
 
 truncate:
 	ret = btrfs_check_trunc_cache_free_space(fs_info,
···
 	}
 	if (!found)
 		return -ENOENT;
-	ret = delete_block_group_cache(leaf->fs_info, block_group, NULL,
-					space_cache_ino);
+	ret = delete_block_group_cache(block_group, NULL, space_cache_ino);
 	return ret;
 }
 
···
  * the inode is in data relocation tree and its link count is 0
  */
 static noinline_for_stack struct inode *create_reloc_inode(
-					struct btrfs_fs_info *fs_info,
 					const struct btrfs_block_group *group)
 {
-	struct inode *inode = NULL;
+	struct btrfs_fs_info *fs_info = group->fs_info;
+	struct btrfs_inode *inode = NULL;
 	struct btrfs_trans_handle *trans;
 	struct btrfs_root *root;
 	u64 objectid;
···
 		inode = NULL;
 		goto out;
 	}
-	BTRFS_I(inode)->reloc_block_group_start = group->start;
+	inode->reloc_block_group_start = group->start;
 
-	ret = btrfs_orphan_add(trans, BTRFS_I(inode));
+	ret = btrfs_orphan_add(trans, inode);
 out:
 	btrfs_put_root(root);
 	btrfs_end_transaction(trans);
 	btrfs_btree_balance_dirty(fs_info);
 	if (ret) {
-		iput(inode);
+		if (inode)
+			iput(&inode->vfs_inode);
 		inode = ERR_PTR(ret);
 	}
-	return inode;
+	return &inode->vfs_inode;
 }
 
 /*
···
 	btrfs_free_path(path);
 
 	if (!IS_ERR(inode))
-		ret = delete_block_group_cache(fs_info, rc->block_group, inode, 0);
+		ret = delete_block_group_cache(rc->block_group, inode, 0);
 	else
 		ret = PTR_ERR(inode);
 
···
 		goto out;
 	}
 
-	rc->data_inode = create_reloc_inode(fs_info, rc->block_group);
+	rc->data_inode = create_reloc_inode(rc->block_group);
 	if (IS_ERR(rc->data_inode)) {
 		err = PTR_ERR(rc->data_inode);
 		rc->data_inode = NULL;

fs/btrfs/scrub.c

···
 	if (path->nodes[0])
 		goto search_forward;
 
+	key.objectid = search_start;
 	if (btrfs_fs_incompat(fs_info, SKINNY_METADATA))
 		key.type = BTRFS_METADATA_ITEM_KEY;
 	else
 		key.type = BTRFS_EXTENT_ITEM_KEY;
-	key.objectid = search_start;
 	key.offset = (u64)-1;
 
 	ret = btrfs_search_slot(NULL, extent_root, &key, path, 0, 0);
···
 	path->skip_locking = 1;
 
 	key.objectid = scrub_dev->devid;
-	key.offset = 0ull;
 	key.type = BTRFS_DEV_EXTENT_KEY;
+	key.offset = 0ull;
 
 	while (1) {
 		u64 dev_extent_len;

fs/btrfs/send.c

···
 #include <linux/compat.h>
 #include <linux/crc32c.h>
 #include <linux/fsverity.h>
-
 #include "send.h"
 #include "ctree.h"
 #include "backref.h"
···
 	u64 cur_inode_rdev;
 	u64 cur_inode_last_extent;
 	u64 cur_inode_next_write_offset;
+	struct fs_path cur_inode_path;
 	bool cur_inode_new;
 	bool cur_inode_new_gen;
 	bool cur_inode_deleted;
···
 
 static void fs_path_reset(struct fs_path *p)
 {
-	if (p->reversed) {
+	if (p->reversed)
 		p->start = p->buf + p->buf_len - 1;
-		p->end = p->start;
-		*p->start = 0;
-	} else {
+	else
 		p->start = p->buf;
-		p->end = p->start;
-		*p->start = 0;
-	}
+
+	p->end = p->start;
+	*p->start = 0;
+}
+
+static void init_path(struct fs_path *p)
+{
+	p->reversed = 0;
+	p->buf = p->inline_buf;
+	p->buf_len = FS_PATH_INLINE_SIZE;
+	fs_path_reset(p);
 }
 
 static struct fs_path *fs_path_alloc(void)
···
 	p = kmalloc(sizeof(*p), GFP_KERNEL);
 	if (!p)
 		return NULL;
-	p->reversed = 0;
-	p->buf = p->inline_buf;
-	p->buf_len = FS_PATH_INLINE_SIZE;
-	fs_path_reset(p);
+	init_path(p);
 	return p;
 }
 
···
 	kfree(p);
 }
 
-static int fs_path_len(struct fs_path *p)
+static inline int fs_path_len(const struct fs_path *p)
 {
 	return p->end - p->start;
 }
···
 	if (p->buf_len >= len)
 		return 0;
 
-	if (len > PATH_MAX) {
-		WARN_ON(1);
-		return -ENOMEM;
-	}
+	if (WARN_ON(len > PATH_MAX))
+		return -ENAMETOOLONG;
 
-	path_len = p->end - p->start;
+	path_len = fs_path_len(p);
 	old_buf_len = p->buf_len;
 
 	/*
···
 	int ret;
 	int new_len;
 
-	new_len = p->end - p->start + name_len;
+	new_len = fs_path_len(p) + name_len;
 	if (p->start != p->end)
 		new_len++;
 	ret = fs_path_ensure_buf(p, new_len);
 	if (ret < 0)
-		goto out;
+		return ret;
 
 	if (p->reversed) {
 		if (p->start != p->end)
···
 		*p->end = 0;
 	}
 
-out:
-	return ret;
+	return 0;
 }
 
 static int fs_path_add(struct fs_path *p, const char *name, int name_len)
···
 
 	ret = fs_path_prepare_for_add(p, name_len, &prepared);
 	if (ret < 0)
-		goto out;
+		return ret;
 	memcpy(prepared, name, name_len);
 
-out:
-	return ret;
+	return 0;
 }
 
-static int fs_path_add_path(struct fs_path *p, struct fs_path *p2)
+static inline int fs_path_add_path(struct fs_path *p, const struct fs_path *p2)
 {
-	int ret;
-	char *prepared;
-
-	ret = fs_path_prepare_for_add(p, p2->end - p2->start, &prepared);
-	if (ret < 0)
-		goto out;
-	memcpy(prepared, p2->start, p2->end - p2->start);
-
-out:
-	return ret;
+	return fs_path_add(p, p2->start, fs_path_len(p2));
 }
 
 static int fs_path_add_from_extent_buffer(struct fs_path *p,
···
 
 	ret = fs_path_prepare_for_add(p, len, &prepared);
 	if (ret < 0)
-		goto out;
+		return ret;
 
 	read_extent_buffer(eb, prepared, off, len);
 
-out:
-	return ret;
+	return 0;
 }
 
 static int fs_path_copy(struct fs_path *p, struct fs_path *from)
···
 		return;
 
 	tmp = p->start;
-	len = p->end - p->start;
+	len = fs_path_len(p);
 	p->start = p->buf;
 	p->end = p->start + len;
 	memmove(p->start, tmp, len + 1);
 	p->reversed = 0;
+}
+
+static inline bool is_current_inode_path(const struct send_ctx *sctx,
+					 const struct fs_path *path)
+{
+	const struct fs_path *cur = &sctx->cur_inode_path;
+
+	return (strncmp(path->start, cur->start, fs_path_len(cur)) == 0);
 }
 
 static struct btrfs_path *alloc_path_for_send(void)
···
 #define TLV_PUT_PATH(sctx, attrtype, p) \
 	do { \
 		ret = tlv_put_string(sctx, attrtype, p->start, \
-			p->end - p->start); \
+				     fs_path_len((p)));	       \
 		if (ret < 0) \
 			goto tlv_put_failure; \
 	} while(0)
···
 
 	ret = begin_cmd(sctx, BTRFS_SEND_C_RENAME);
 	if (ret < 0)
-		goto out;
+		return ret;
 
 	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, from);
 	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH_TO, to);
···
 	ret = send_cmd(sctx);
 
 tlv_put_failure:
-out:
 	return ret;
 }
 
···
 
 	ret = begin_cmd(sctx, BTRFS_SEND_C_LINK);
 	if (ret < 0)
-		goto out;
+		return ret;
 
 	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, path);
 	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH_LINK, lnk);
···
 	ret = send_cmd(sctx);
 
 tlv_put_failure:
-out:
 	return ret;
 }
 
···
 
 	ret = begin_cmd(sctx, BTRFS_SEND_C_UNLINK);
 	if (ret < 0)
-		goto out;
+		return ret;
 
 	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, path);
 
 	ret = send_cmd(sctx);
 
 tlv_put_failure:
-out:
 	return ret;
 }
 
···
 
 	ret = begin_cmd(sctx, BTRFS_SEND_C_RMDIR);
 	if (ret < 0)
-		goto out;
+		return ret;
 
 	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, path);
 
 	ret = send_cmd(sctx);
 
 tlv_put_failure:
-out:
 	return ret;
 }
 
···
 
 	ret = get_inode_info(sctx->send_root, ino, &info);
 	if (ret < 0 && ret != -ENOENT)
-		goto out;
+		return ret;
 	left_ret = (info.nlink == 0) ? -ENOENT : ret;
 	left_gen = info.gen;
 	if (send_gen)
···
 	} else {
 		ret = get_inode_info(sctx->parent_root, ino, &info);
 		if (ret < 0 && ret != -ENOENT)
-			goto out;
+			return ret;
 		right_ret = (info.nlink == 0) ? -ENOENT : ret;
 		right_gen = info.gen;
 		if (parent_gen)
···
 		ret = -ENOENT;
 	}
 
-out:
 	return ret;
 }
 
···
 
 	ret = get_cur_inode_state(sctx, ino, gen, send_gen, parent_gen);
 	if (ret < 0)
-		goto out;
+		return ret;
 
 	if (ret == inode_state_no_change ||
 	    ret == inode_state_did_create ||
 	    ret == inode_state_will_delete)
-		ret = 1;
-	else
-		ret = 0;
+		return 1;
 
-out:
-	return ret;
+	return 0;
 }
 
 /*
···
 			*parent_gen = nce->parent_gen;
 			ret = fs_path_add(dest, nce->name, nce->name_len);
 			if (ret < 0)
-				goto out;
-			ret = nce->ret;
-			goto out;
+				return ret;
+			return nce->ret;
 		}
 	}
 
···
 	 */
 	ret = is_inode_existent(sctx, ino, gen, NULL, NULL);
 	if (ret < 0)
-		goto out;
+		return ret;
 
 	if (!ret) {
 		ret = gen_unique_name(sctx, ino, gen, dest);
 		if (ret < 0)
-			goto out;
+			return ret;
 		ret = 1;
 		goto out_cache;
 	}
···
 		ret = get_first_ref(sctx->parent_root, ino,
 				    parent_ino, parent_gen, dest);
 	if (ret < 0)
-		goto out;
+		return ret;
 
 	/*
 	 * Check if the ref was overwritten by an inode's ref that was processed
 	 * earlier. If yes, treat as orphan and return 1.
 	 */
 	ret = did_overwrite_ref(sctx, *parent_ino, *parent_gen, ino, gen,
-			dest->start, dest->end - dest->start);
+				dest->start, fs_path_len(dest));
 	if (ret < 0)
-		goto out;
+		return ret;
 	if (ret) {
 		fs_path_reset(dest);
 		ret = gen_unique_name(sctx, ino, gen, dest);
 		if (ret < 0)
-			goto out;
+			return ret;
 		ret = 1;
 	}
 
···
 	 * Store the result of the lookup in the name cache.
 	 */
 	nce = kmalloc(sizeof(*nce) + fs_path_len(dest), GFP_KERNEL);
-	if (!nce) {
-		ret = -ENOMEM;
-		goto out;
-	}
+	if (!nce)
+		return -ENOMEM;
 
 	nce->entry.key = ino;
 	nce->entry.gen = gen;
···
 	nce_ret = btrfs_lru_cache_store(&sctx->name_cache, &nce->entry, GFP_KERNEL);
 	if (nce_ret < 0) {
 		kfree(nce);
-		ret = nce_ret;
+		return nce_ret;
 	}
 
-out:
 	return ret;
 }
 
···
 	u64 parent_inode = 0;
 	u64 parent_gen = 0;
 	int stop = 0;
+	const bool is_cur_inode = (ino == sctx->cur_ino && gen == sctx->cur_inode_gen);
+
+	if (is_cur_inode && fs_path_len(&sctx->cur_inode_path) > 0) {
+		if (dest != &sctx->cur_inode_path)
+			return fs_path_copy(dest, &sctx->cur_inode_path);
+
+		return 0;
+	}
 
 	name = fs_path_alloc();
 	if (!name) {
···
 
 out:
 	fs_path_free(name);
-	if (!ret)
+	if (!ret) {
 		fs_path_unreverse(dest);
+		if (is_cur_inode && dest != &sctx->cur_inode_path)
+			ret = fs_path_copy(&sctx->cur_inode_path, dest);
+	}
+
 	return ret;
 }
 
···
 	return ret;
 }
 
+static struct fs_path *get_cur_inode_path(struct send_ctx *sctx)
+{
+	if (fs_path_len(&sctx->cur_inode_path) == 0) {
+		int ret;
+
+		ret = get_cur_path(sctx, sctx->cur_ino, sctx->cur_inode_gen,
+				   &sctx->cur_inode_path);
+		if (ret < 0)
+			return ERR_PTR(ret);
+	}
+
+	return &sctx->cur_inode_path;
+}
+
+static struct fs_path *get_path_for_command(struct send_ctx *sctx, u64 ino, u64 gen)
+{
+	struct fs_path *path;
+	int ret;
+
+	if (ino == sctx->cur_ino && gen == sctx->cur_inode_gen)
+		return get_cur_inode_path(sctx);
+
+	path = fs_path_alloc();
+	if (!path)
+		return ERR_PTR(-ENOMEM);
+
+	ret = get_cur_path(sctx, ino, gen, path);
+	if (ret < 0) {
+		fs_path_free(path);
+		return ERR_PTR(ret);
+	}
+
+	return path;
+}
+
+static void free_path_for_command(const struct send_ctx *sctx, struct fs_path *path)
+{
+	if (path != &sctx->cur_inode_path)
+		fs_path_free(path);
+}
+
 static int send_truncate(struct send_ctx *sctx, u64 ino, u64 gen, u64 size)
 {
 	struct btrfs_fs_info *fs_info = sctx->send_root->fs_info;
···
 
 	btrfs_debug(fs_info, "send_truncate %llu size=%llu", ino, size);
 
-	p = fs_path_alloc();
-	if (!p)
-		return -ENOMEM;
+	p = get_path_for_command(sctx, ino, gen);
+	if (IS_ERR(p))
+		return PTR_ERR(p);
 
 	ret = begin_cmd(sctx, BTRFS_SEND_C_TRUNCATE);
 	if (ret < 0)
 		goto out;
 
-	ret = get_cur_path(sctx, ino, gen, p);
-	if (ret < 0)
-		goto out;
 	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, p);
 	TLV_PUT_U64(sctx, BTRFS_SEND_A_SIZE, size);
 
···
 
 tlv_put_failure:
 out:
-	fs_path_free(p);
+	free_path_for_command(sctx, p);
 	return ret;
 }
 
···
 
 	btrfs_debug(fs_info, "send_chmod %llu mode=%llu", ino, mode);
 
-	p = fs_path_alloc();
-	if (!p)
-		return -ENOMEM;
+	p = get_path_for_command(sctx, ino, gen);
+	if (IS_ERR(p))
+		return PTR_ERR(p);
 
 	ret = begin_cmd(sctx, BTRFS_SEND_C_CHMOD);
 	if (ret < 0)
 		goto out;
 
-	ret = get_cur_path(sctx, ino, gen, p);
-	if (ret < 0)
-		goto out;
 	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, p);
 	TLV_PUT_U64(sctx, BTRFS_SEND_A_MODE, mode & 07777);
 
···
 
 tlv_put_failure:
 out:
-	fs_path_free(p);
+	free_path_for_command(sctx, p);
 	return ret;
 }
 
···
 
 	btrfs_debug(fs_info, "send_fileattr %llu fileattr=%llu", ino, fileattr);
 
-	p = fs_path_alloc();
-	if (!p)
-		return -ENOMEM;
+	p = get_path_for_command(sctx, ino, gen);
+	if (IS_ERR(p))
+		return PTR_ERR(p);
 
 	ret = begin_cmd(sctx, BTRFS_SEND_C_FILEATTR);
 	if (ret < 0)
 		goto out;
 
-	ret = get_cur_path(sctx, ino, gen, p);
-	if (ret < 0)
-		goto out;
 	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, p);
 	TLV_PUT_U64(sctx, BTRFS_SEND_A_FILEATTR, fileattr);
 
···
 
 tlv_put_failure:
 out:
-	fs_path_free(p);
+	free_path_for_command(sctx, p);
 	return ret;
 }
 
···
 	btrfs_debug(fs_info, "send_chown %llu uid=%llu, gid=%llu",
 		    ino, uid, gid);
 
-	p = fs_path_alloc();
-	if (!p)
-		return -ENOMEM;
+	p = get_path_for_command(sctx, ino, gen);
+	if (IS_ERR(p))
+		return PTR_ERR(p);
 
 	ret = begin_cmd(sctx, BTRFS_SEND_C_CHOWN);
 	if (ret < 0)
 		goto out;
 
-	ret = get_cur_path(sctx, ino, gen, p);
-	if (ret < 0)
-		goto out;
 	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, p);
 	TLV_PUT_U64(sctx, BTRFS_SEND_A_UID, uid);
 	TLV_PUT_U64(sctx, BTRFS_SEND_A_GID, gid);
···
 
 tlv_put_failure:
 out:
-	fs_path_free(p);
+	free_path_for_command(sctx, p);
 	return ret;
 }
 
···
 
 	btrfs_debug(fs_info, "send_utimes %llu", ino);
 
-	p = fs_path_alloc();
-	if (!p)
-		return -ENOMEM;
+	p = get_path_for_command(sctx, ino, gen);
+	if (IS_ERR(p))
+		return PTR_ERR(p);
 
 	path = alloc_path_for_send();
 	if (!path) {
···
 	if (ret < 0)
 		goto out;
 
-	ret = get_cur_path(sctx, ino, gen, p);
-	if (ret < 0)
-		goto out;
 	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, p);
 	TLV_PUT_BTRFS_TIMESPEC(sctx, BTRFS_SEND_A_ATIME, eb, &ii->atime);
 	TLV_PUT_BTRFS_TIMESPEC(sctx, BTRFS_SEND_A_MTIME, eb, &ii->mtime);
···
 
 tlv_put_failure:
 out:
-	fs_path_free(p);
+	free_path_for_command(sctx, p);
 	btrfs_free_path(path);
 	return ret;
 }
···
 		goto out;
 
 	ret = send_rename(sctx, path, orphan);
+	if (ret < 0)
+		goto out;
+
+	if (ino == sctx->cur_ino && gen == sctx->cur_inode_gen)
+		ret = fs_path_copy(&sctx->cur_inode_path, orphan);
 
 out:
 	fs_path_free(orphan);
···
 	return ret;
 }
 
+static int rename_current_inode(struct send_ctx *sctx,
+				struct fs_path *current_path,
+				struct fs_path *new_path)
+{
+	int ret;
+
+	ret = send_rename(sctx, current_path, new_path);
+	if (ret < 0)
+		return ret;
+
+	ret = fs_path_copy(&sctx->cur_inode_path, new_path);
+	if (ret < 0)
+		return ret;
+
+	return fs_path_copy(current_path, new_path);
+}
+
 /*
  * This does all the move/link/unlink/rmdir magic.
  */
···
 	u64 ow_inode = 0;
 	u64 ow_gen;
 	u64 ow_mode;
-	int did_overwrite = 0;
-	int is_orphan = 0;
 	u64 last_dir_ino_rm = 0;
+	bool did_overwrite = false;
+	bool is_orphan = false;
 	bool can_rename = true;
 	bool orphanized_dir = false;
 	bool orphanized_ancestor = false;
···
 		if (ret < 0)
 			goto out;
 		if (ret)
-			did_overwrite = 1;
+			did_overwrite = true;
 	}
 	if (sctx->cur_inode_new || did_overwrite) {
 		ret = gen_unique_name(sctx, sctx->cur_ino,
 				sctx->cur_inode_gen, valid_path);
 		if (ret < 0)
 			goto out;
-		is_orphan = 1;
+		is_orphan = true;
 	} else {
 		ret = get_cur_path(sctx, sctx->cur_ino, sctx->cur_inode_gen,
 				valid_path);
···
 				if (ret > 0) {
 					orphanized_ancestor = true;
 					fs_path_reset(valid_path);
+					fs_path_reset(&sctx->cur_inode_path);
 					ret = get_cur_path(sctx, sctx->cur_ino,
 							   sctx->cur_inode_gen,
 							   valid_path);
···
 		 * it depending on the inode mode.
 		 */
 		if (is_orphan && can_rename) {
-			ret = send_rename(sctx, valid_path, cur->full_path);
+			ret = rename_current_inode(sctx, valid_path, cur->full_path);
 			if (ret < 0)
 				goto out;
-			is_orphan = 0;
-			ret = fs_path_copy(valid_path, cur->full_path);
-			if (ret < 0)
-				goto out;
+			is_orphan = false;
 		} else if (can_rename) {
 			if (S_ISDIR(sctx->cur_inode_mode)) {
 				/*
···
 				 * dirs, we always have one new and one deleted
 				 * ref. The deleted ref is ignored later.
 				 */
-				ret = send_rename(sctx, valid_path,
-						  cur->full_path);
-				if (!ret)
-					ret = fs_path_copy(valid_path,
+				ret = rename_current_inode(sctx, valid_path,
 							   cur->full_path);
 				if (ret < 0)
 					goto out;
···
 					sctx->cur_inode_gen, valid_path);
 			if (ret < 0)
 				goto out;
-			is_orphan = 1;
+			is_orphan = true;
 		}
 
 		list_for_each_entry(cur, &sctx->deleted_refs, list) {
···
 				ret = send_unlink(sctx, cur->full_path);
 				if (ret < 0)
 					goto out;
+				if (is_current_inode_path(sctx, cur->full_path))
+					fs_path_reset(&sctx->cur_inode_path);
 			}
 			ret = dup_ref(cur, &check_dirs);
 			if (ret < 0)
···
 
 static int record_new_ref_if_needed(u64 dir, struct fs_path *name, void *ctx)
 {
-	int ret = 0;
+	int ret;
 	struct send_ctx *sctx = ctx;
 	struct rb_node *node = NULL;
 	struct recorded_ref data;
···
 
 	ret = get_inode_gen(sctx->send_root, dir, &dir_gen);
 	if (ret < 0)
-		goto out;
+		return ret;
 
 	data.dir = dir;
 	data.dir_gen = dir_gen;
···
 					 &sctx->new_refs, name, dir, dir_gen,
 					 sctx);
 	}
-out:
+
 	return ret;
 }
 
 static int record_deleted_ref_if_needed(u64 dir, struct fs_path *name, void *ctx)
 {
-	int ret = 0;
+	int ret;
 	struct send_ctx *sctx = ctx;
 	struct rb_node *node = NULL;
 	struct recorded_ref data;
···
 
 	ret = get_inode_gen(sctx->parent_root, dir, &dir_gen);
 	if (ret < 0)
-		goto out;
+		return ret;
 
 	data.dir = dir;
 	data.dir_gen = dir_gen;
···
 					 &sctx->deleted_refs, name, dir,
 					 dir_gen, sctx);
 	}
-out:
+
 	return ret;
 }
 
···
 	ret = iterate_inode_ref(sctx->send_root, sctx->left_path,
 				sctx->cmp_key, 0, record_new_ref_if_needed, sctx);
 	if (ret < 0)
-		goto out;
-	ret = 0;
+		return ret;
 
-out:
-	return ret;
+	return 0;
 }
 
 static int record_deleted_ref(struct send_ctx *sctx)
···
 				sctx->cmp_key, 0, record_deleted_ref_if_needed,
 				sctx);
 	if (ret < 0)
-		goto out;
-	ret = 0;
+		return ret;
 
-out:
-	return ret;
+	return 0;
 }
 
 static int record_changed_ref(struct send_ctx *sctx)
 {
-	int ret = 0;
+	int ret;
 
 	ret = iterate_inode_ref(sctx->send_root, sctx->left_path,
 			sctx->cmp_key, 0, record_new_ref_if_needed, sctx);
 	if (ret < 0)
-		goto out;
+		return ret;
 	ret = iterate_inode_ref(sctx->parent_root, sctx->right_path,
 			sctx->cmp_key, 0, record_deleted_ref_if_needed, sctx);
 	if (ret < 0)
-		goto out;
-	ret = 0;
+		return ret;
 
-out:
-	return ret;
+	return 0;
 }
 
 /*
···
 }
 
 static int send_set_xattr(struct send_ctx *sctx,
-			  struct fs_path *path,
 			  const char *name, int name_len,
 			  const char *data, int data_len)
 {
-	int ret = 0;
+	struct fs_path *path;
+	int ret;
+
+	path = get_cur_inode_path(sctx);
+	if (IS_ERR(path))
+		return PTR_ERR(path);
 
 	ret = begin_cmd(sctx, BTRFS_SEND_C_SET_XATTR);
 	if (ret < 0)
-		goto out;
+		return ret;
 
 	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, path);
 	TLV_PUT_STRING(sctx, BTRFS_SEND_A_XATTR_NAME, name, name_len);
···
 	ret = send_cmd(sctx);
 
 tlv_put_failure:
-out:
 	return ret;
 }
 
···
 			  struct fs_path *path,
 			  const char *name, int name_len)
 {
-	int ret = 0;
+	int ret;
 
 	ret = begin_cmd(sctx, BTRFS_SEND_C_REMOVE_XATTR);
 	if (ret < 0)
-		goto out;
+		return ret;
 
 	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, path);
 	TLV_PUT_STRING(sctx, BTRFS_SEND_A_XATTR_NAME, name, name_len);
···
 	ret = send_cmd(sctx);
 
 tlv_put_failure:
-out:
 	return ret;
 }
 
···
 			       const char *name, int name_len, const char *data,
 			       int data_len, void *ctx)
 {
-	int ret;
 	struct send_ctx *sctx = ctx;
-	struct fs_path *p;
 	struct posix_acl_xattr_header dummy_acl;
 
 	/* Capabilities are emitted by finish_inode_if_needed */
 	if (!strncmp(name, XATTR_NAME_CAPS, name_len))
 		return 0;
-
-	p = fs_path_alloc();
-	if (!p)
-		return -ENOMEM;
 
 	/*
 	 * This hack is needed because empty acls are stored as zero byte
···
 		}
 	}
 
-	ret = get_cur_path(sctx, sctx->cur_ino, sctx->cur_inode_gen, p);
-	if (ret < 0)
-		goto out;
-
-	ret = send_set_xattr(sctx, p, name, name_len, data, data_len);
-
-out:
-	fs_path_free(p);
-	return ret;
+	return send_set_xattr(sctx, name, name_len, data, data_len);
 }
 
 static int __process_deleted_xattr(int num, struct btrfs_key *di_key,
 				   const char *name, int name_len,
 				   const char *data, int data_len, void *ctx)
 {
-	int ret;
 	struct send_ctx *sctx = ctx;
 	struct fs_path *p;
 
-	p = fs_path_alloc();
-	if (!p)
-		return -ENOMEM;
+	p = get_cur_inode_path(sctx);
+	if (IS_ERR(p))
+		return PTR_ERR(p);
 
-	ret = get_cur_path(sctx, sctx->cur_ino, sctx->cur_inode_gen, p);
-	if (ret < 0)
-		goto out;
-
-	ret = send_remove_xattr(sctx, p, name, name_len);
-
-out:
-	fs_path_free(p);
-	return ret;
+	return send_remove_xattr(sctx, p, name, name_len);
 }
 
 static int process_new_xattr(struct send_ctx *sctx)
 {
-	int ret = 0;
-
-	ret = iterate_dir_item(sctx->send_root, sctx->left_path,
-			       __process_new_xattr, sctx);
-
-	return ret;
+	return iterate_dir_item(sctx->send_root, sctx->left_path,
+				__process_new_xattr, sctx);
 }
 
 static int process_deleted_xattr(struct send_ctx *sctx)
···
 
 static int process_changed_xattr(struct send_ctx *sctx)
 {
-	int ret = 0;
+	int ret;
 
 	ret = iterate_dir_item(sctx->send_root, sctx->left_path,
 			__process_changed_new_xattr, sctx);
 	if (ret < 0)
-		goto out;
-	ret = iterate_dir_item(sctx->parent_root, sctx->right_path,
-			__process_changed_deleted_xattr, sctx);
+		return ret;
 
-out:
-	return ret;
+	return iterate_dir_item(sctx->parent_root, sctx->right_path,
+				__process_changed_deleted_xattr, sctx);
 }
 
 static int process_all_new_xattrs(struct send_ctx *sctx)
···
 
 	ret = begin_cmd(sctx, BTRFS_SEND_C_ENABLE_VERITY);
 	if (ret < 0)
-		goto out;
+		return ret;
 
 	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, path);
 	TLV_PUT_U8(sctx, BTRFS_SEND_A_VERITY_ALGORITHM,
···
 	ret = send_cmd(sctx);
 
 tlv_put_failure:
-out:
 	return ret;
 }
 
 static int process_verity(struct send_ctx *sctx)
 {
 	int ret = 0;
-	struct inode *inode;
+	struct btrfs_inode *inode;
 	struct fs_path *p;
 
 	inode = btrfs_iget(sctx->cur_ino, sctx->send_root);
 	if (IS_ERR(inode))
 		return PTR_ERR(inode);
 
-	ret = btrfs_get_verity_descriptor(inode, NULL, 0);
+	ret = btrfs_get_verity_descriptor(&inode->vfs_inode, NULL, 0);
 	if (ret < 0)
 		goto iput;
 
···
 		}
 	}
 
-	ret = btrfs_get_verity_descriptor(inode, sctx->verity_descriptor, ret);
+	ret = btrfs_get_verity_descriptor(&inode->vfs_inode, sctx->verity_descriptor, ret);
 	if (ret < 0)
 		goto iput;
 
-	p = fs_path_alloc();
-	if (!p) {
-		ret = -ENOMEM;
+	p = get_cur_inode_path(sctx);
+	if (IS_ERR(p)) {
+		ret = PTR_ERR(p);
 		goto iput;
 	}
-	ret = get_cur_path(sctx, sctx->cur_ino, sctx->cur_inode_gen, p);
-	if (ret < 0)
-		goto free_path;
 
 	ret = send_verity(sctx, p, sctx->verity_descriptor);
-	if (ret < 0)
-		goto free_path;
-
-free_path:
-	fs_path_free(p);
 iput:
-	iput(inode);
+	iput(&inode->vfs_inode);
 	return ret;
 }
 
···
 	int ret = 0;
 	struct fs_path *p;
 
-	p = fs_path_alloc();
-	if (!p)
-		return -ENOMEM;
-
 	btrfs_debug(fs_info, "send_write offset=%llu, len=%d", offset, len);
+
+	p = get_cur_inode_path(sctx);
+	if (IS_ERR(p))
+		return PTR_ERR(p);
 
 	ret = begin_cmd(sctx, BTRFS_SEND_C_WRITE);
 	if (ret < 0)
-		goto out;
-
-	ret = get_cur_path(sctx, sctx->cur_ino, sctx->cur_inode_gen, p);
-	if (ret < 0)
-		goto out;
+		return ret;
 
 	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, p);
 	TLV_PUT_U64(sctx, BTRFS_SEND_A_FILE_OFFSET, offset);
 	ret = put_file_data(sctx, offset, len);
 	if (ret < 0)
-		goto out;
+		return ret;
 
 	ret = send_cmd(sctx);
 
 tlv_put_failure:
-out:
-	fs_path_free(p);
 	return ret;
 }
 
···
 {
 	int ret = 0;
 	struct fs_path *p;
+	struct fs_path *cur_inode_path;
 	u64 gen;
 
 	btrfs_debug(sctx->send_root->fs_info,
 		    "send_clone offset=%llu, len=%d, clone_root=%llu, clone_inode=%llu, clone_offset=%llu",
 		    offset, len, btrfs_root_id(clone_root->root),
 		    clone_root->ino, clone_root->offset);
+
+	cur_inode_path = get_cur_inode_path(sctx);
+	if (IS_ERR(cur_inode_path))
+		return PTR_ERR(cur_inode_path);
 
 	p = fs_path_alloc();
 	if (!p)
···
 	if (ret < 0)
 		goto out;
 
-	ret = get_cur_path(sctx, sctx->cur_ino, sctx->cur_inode_gen, p);
-	if (ret < 0)
-		goto out;
-
 	TLV_PUT_U64(sctx, BTRFS_SEND_A_FILE_OFFSET, offset);
 	TLV_PUT_U64(sctx, BTRFS_SEND_A_CLONE_LEN, len);
-	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, p);
+	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, cur_inode_path);
 
 	if (clone_root->root == sctx->send_root) {
 		ret = get_inode_gen(sctx->send_root, clone_root->ino, &gen);
···
 	int ret = 0;
 	struct fs_path *p;
 
-	p = fs_path_alloc();
-	if (!p)
-		return -ENOMEM;
+	p = get_cur_inode_path(sctx);
+	if (IS_ERR(p))
+		return PTR_ERR(p);
 
 	ret = begin_cmd(sctx, BTRFS_SEND_C_UPDATE_EXTENT);
 	if (ret < 0)
-		goto out;
-
-	ret = get_cur_path(sctx, sctx->cur_ino, sctx->cur_inode_gen, p);
-	if (ret < 0)
-		goto out;
+		return ret;
 
 	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, p);
 	TLV_PUT_U64(sctx, BTRFS_SEND_A_FILE_OFFSET, offset);
···
 	ret = send_cmd(sctx);
 
 tlv_put_failure:
-out:
-	fs_path_free(p);
 	return ret;
 }
 
···
 	if (sctx->flags & BTRFS_SEND_FLAG_NO_FILE_DATA)
 		return send_update_extent(sctx, offset, end - offset);
 
-	p = fs_path_alloc();
-	if (!p)
-		return -ENOMEM;
-	ret = get_cur_path(sctx, sctx->cur_ino, sctx->cur_inode_gen, p);
-	if (ret < 0)
-		goto tlv_put_failure;
+	p = get_cur_inode_path(sctx);
+	if (IS_ERR(p))
+		return PTR_ERR(p);
+
 	while (offset < end) {
 		u64 len = min(end - offset, read_size);
 
···
 	}
 	sctx->cur_inode_next_write_offset = offset;
 tlv_put_failure:
-	fs_path_free(p);
 	return ret;
 }
 
···
 				      struct btrfs_path *path, u64 offset,
 				      u64 len)
 {
-	struct btrfs_root *root = sctx->send_root;
-	struct btrfs_fs_info *fs_info = root->fs_info;
-	struct inode *inode;
+	struct btrfs_fs_info *fs_info = sctx->send_root->fs_info;
 	struct fs_path *fspath;
 	struct extent_buffer *leaf = path->nodes[0];
 	struct btrfs_key key;
···
 	size_t inline_size;
 	int ret;
 
-	inode = btrfs_iget(sctx->cur_ino, root);
-	if (IS_ERR(inode))
-		return PTR_ERR(inode);
-
-	fspath = fs_path_alloc();
-	if (!fspath) {
-		ret = -ENOMEM;
-		goto out;
-	}
+	fspath = get_cur_inode_path(sctx);
+	if (IS_ERR(fspath))
+		return PTR_ERR(fspath);
 
 	ret = begin_cmd(sctx, BTRFS_SEND_C_ENCODED_WRITE);
 	if (ret < 0)
-		goto out;
-
-	ret = get_cur_path(sctx, sctx->cur_ino, sctx->cur_inode_gen, fspath);
-	if (ret < 0)
-		goto out;
+		return ret;
 
 	btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
 	ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_file_extent_item);
···
 	ret = btrfs_encoded_io_compression_from_extent(fs_info,
 				btrfs_file_extent_compression(leaf, ei));
 	if (ret < 0)
-		goto out;
+		return ret;
 	TLV_PUT_U32(sctx, BTRFS_SEND_A_COMPRESSION, ret);
 
 	ret = put_data_header(sctx, inline_size);
 	if (ret < 0)
-		goto out;
+		return ret;
 	read_extent_buffer(leaf, sctx->send_buf + sctx->send_size,
 			   btrfs_file_extent_inline_start(ei), inline_size);
 	sctx->send_size += inline_size;
···
 	ret = send_cmd(sctx);
 
 tlv_put_failure:
-out:
-	fs_path_free(fspath);
-	iput(inode);
 	return ret;
 }
 
···
 {
 	struct btrfs_root *root = sctx->send_root;
 	struct btrfs_fs_info *fs_info = root->fs_info;
-	struct inode *inode;
+	struct btrfs_inode *inode;
 	struct fs_path *fspath;
 	struct extent_buffer *leaf = path->nodes[0];
 	struct btrfs_key key;
···
 	if (IS_ERR(inode))
 		return PTR_ERR(inode);
 
-	fspath = fs_path_alloc();
-	if (!fspath) {
-		ret = -ENOMEM;
+	fspath = get_cur_inode_path(sctx);
+	if (IS_ERR(fspath)) {
+		ret = PTR_ERR(fspath);
 		goto out;
 	}
 
 	ret = begin_cmd(sctx, BTRFS_SEND_C_ENCODED_WRITE);
-	if (ret < 0)
-		goto out;
-
-	ret = get_cur_path(sctx, sctx->cur_ino, sctx->cur_inode_gen, fspath);
 	if (ret < 0)
 		goto out;
 
···
 	 * Note that send_buf is a mapping of send_buf_pages, so this is really
 	 * reading into send_buf.
 	 */
-	ret = btrfs_encoded_read_regular_fill_pages(BTRFS_I(inode),
+	ret = btrfs_encoded_read_regular_fill_pages(inode,
 						    disk_bytenr, disk_num_bytes,
 						    sctx->send_buf_pages +
 						    (data_offset >> PAGE_SHIFT),
···
 
 tlv_put_failure:
 out:
-	fs_path_free(fspath);
-	iput(inode);
+	iput(&inode->vfs_inode);
 	return ret;
 }
 
···
 	}
 
 	if (sctx->cur_inode == NULL) {
+		struct btrfs_inode *btrfs_inode;
 		struct btrfs_root *root = sctx->send_root;
 
-		sctx->cur_inode = btrfs_iget(sctx->cur_ino, root);
-		if (IS_ERR(sctx->cur_inode)) {
-			int err = PTR_ERR(sctx->cur_inode);
+		btrfs_inode = btrfs_iget(sctx->cur_ino, root);
+		if (IS_ERR(btrfs_inode))
+			return PTR_ERR(btrfs_inode);
 
-			sctx->cur_inode = NULL;
-			return err;
-		}
+		sctx->cur_inode = &btrfs_inode->vfs_inode;
 		memset(&sctx->ra, 0, sizeof(struct file_ra_state));
 		file_ra_state_init(&sctx->ra, sctx->cur_inode->i_mapping);
 
···
  */
 static int send_capabilities(struct send_ctx *sctx)
 {
-	struct fs_path *fspath = NULL;
 	struct btrfs_path *path;
 	struct btrfs_dir_item *di;
 	struct extent_buffer *leaf;
···
 	leaf = path->nodes[0];
 	buf_len = btrfs_dir_data_len(leaf, di);
 
-	fspath = fs_path_alloc();
 	buf = kmalloc(buf_len, GFP_KERNEL);
-	if (!fspath || !buf) {
+	if (!buf) {
 		ret = -ENOMEM;
 		goto out;
 	}
 
-	ret = get_cur_path(sctx, sctx->cur_ino, sctx->cur_inode_gen, fspath);
-	if (ret < 0)
-		goto out;
-
 	data_ptr = (unsigned long)(di + 1) + btrfs_dir_name_len(leaf, di);
 	read_extent_buffer(leaf, buf, data_ptr, buf_len);
 
-	ret = send_set_xattr(sctx, fspath, XATTR_NAME_CAPS,
+	ret = send_set_xattr(sctx, XATTR_NAME_CAPS,
 			strlen(XATTR_NAME_CAPS), buf, buf_len);
 out:
 	kfree(buf);
-	fs_path_free(fspath);
 	btrfs_free_path(path);
 	return ret;
 }
···
 	sctx->cur_inode_last_extent = (u64)-1;
 	sctx->cur_inode_next_write_offset = 0;
 	sctx->ignore_cur_inode = false;
+	fs_path_reset(&sctx->cur_inode_path);
 
 	/*
 	 * Set send_progress to current inode. This will tell all get_cur_xxx
···
 		      btrfs_root_id(root), root->dedupe_in_progress);
 }
 
-long btrfs_ioctl_send(struct btrfs_inode *inode, const struct btrfs_ioctl_send_args *arg)
+long btrfs_ioctl_send(struct btrfs_root *send_root, const struct btrfs_ioctl_send_args *arg)
 {
 	int ret = 0;
-	struct btrfs_root *send_root = inode->root;
 	struct btrfs_fs_info *fs_info = send_root->fs_info;
 	struct btrfs_root *clone_root;
 	struct send_ctx *sctx = NULL;
···
 		goto out;
 	}
 
+	init_path(&sctx->cur_inode_path);
 	INIT_LIST_HEAD(&sctx->new_refs);
 	INIT_LIST_HEAD(&sctx->deleted_refs);
 
···
 		btrfs_lru_cache_clear(&sctx->backref_cache);
 		btrfs_lru_cache_clear(&sctx->dir_created_cache);
 		btrfs_lru_cache_clear(&sctx->dir_utimes_cache);
+
+		if (sctx->cur_inode_path.buf != sctx->cur_inode_path.inline_buf)
+			kfree(sctx->cur_inode_path.buf);
 
 		kfree(sctx);
 	}

fs/btrfs/send.h

···
 #include <linux/sizes.h>
 #include <linux/align.h>
 
-struct btrfs_inode;
+struct btrfs_root;
 struct btrfs_ioctl_send_args;
 
 #define BTRFS_SEND_STREAM_MAGIC "btrfs-stream"
···
 	__BTRFS_SEND_A_MAX		= 35,
 };
 
-long btrfs_ioctl_send(struct btrfs_inode *inode, const struct btrfs_ioctl_send_args *arg);
+long btrfs_ioctl_send(struct btrfs_root *send_root, const struct btrfs_ioctl_send_args *arg);
 
 #endif

fs/btrfs/space-info.c

···
 // SPDX-License-Identifier: GPL-2.0
 
-#include "linux/spinlock.h"
+#include <linux/spinlock.h>
 #include <linux/minmax.h>
 #include "misc.h"
 #include "ctree.h"

fs/btrfs/subpage.c

···
 
 #include <linux/slab.h>
 #include "messages.h"
-#include "ctree.h"
 #include "subpage.h"
 #include "btrfs_inode.h"
 
 /*
- * Subpage (sectorsize < PAGE_SIZE) support overview:
+ * Subpage (block size < folio size) support overview:
  *
  * Limitations:
  *
···
  *   This means a slightly higher tree locking latency.
  */
 
-#if PAGE_SIZE > SZ_4K
-bool btrfs_is_subpage(const struct btrfs_fs_info *fs_info, struct address_space *mapping)
-{
-	if (fs_info->sectorsize >= PAGE_SIZE)
-		return false;
-
-	/*
-	 * Only data pages (either through DIO or compression) can have no
-	 * mapping. And if page->mapping->host is data inode, it's subpage.
-	 * As we have ruled our sectorsize >= PAGE_SIZE case already.
-	 */
-	if (!mapping || !mapping->host || is_data_inode(BTRFS_I(mapping->host)))
-		return true;
-
-	/*
-	 * Now the only remaining case is metadata, which we only go subpage
-	 * routine if nodesize < PAGE_SIZE.
-	 */
-	if (fs_info->nodesize < PAGE_SIZE)
-		return true;
-	return false;
-}
-#endif
-
 int btrfs_attach_subpage(const struct btrfs_fs_info *fs_info,
 			 struct folio *folio, enum btrfs_subpage_type type)
 {
 	struct btrfs_subpage *subpage;
+
+	/* For metadata we don't support large folio yet. */
+	ASSERT(!folio_test_large(folio));
 
 	/*
 	 * We have cases like a dummy extent buffer page, which is not mapped
···
 		ASSERT(folio_test_locked(folio));
 
 	/* Either not subpage, or the folio already has private attached. */
-	if (!btrfs_is_subpage(fs_info, folio->mapping) || folio_test_private(folio))
+	if (folio_test_private(folio))
+		return 0;
+	if (type == BTRFS_SUBPAGE_METADATA && !btrfs_meta_is_subpage(fs_info))
+		return 0;
+	if (type == BTRFS_SUBPAGE_DATA && !btrfs_is_subpage(fs_info, folio))
 		return 0;
 
-	subpage = btrfs_alloc_subpage(fs_info, type);
+	subpage = btrfs_alloc_subpage(fs_info, folio_size(folio), type);
 	if (IS_ERR(subpage))
 		return  PTR_ERR(subpage);
 
···
 	return 0;
 }
 
-void btrfs_detach_subpage(const struct btrfs_fs_info *fs_info, struct folio *folio)
+void btrfs_detach_subpage(const struct btrfs_fs_info *fs_info, struct folio *folio,
+			  enum btrfs_subpage_type type)
 {
 	struct btrfs_subpage *subpage;
 
 	/* Either not subpage, or the folio already has private attached. */
-	if (!btrfs_is_subpage(fs_info, folio->mapping) || !folio_test_private(folio))
+	if (!folio_test_private(folio))
+		return;
+	if (type == BTRFS_SUBPAGE_METADATA && !btrfs_meta_is_subpage(fs_info))
+		return;
+	if (type == BTRFS_SUBPAGE_DATA && !btrfs_is_subpage(fs_info, folio))
 		return;
 
 	subpage = folio_detach_private(folio);
···
 }
 
 struct btrfs_subpage *btrfs_alloc_subpage(const struct btrfs_fs_info *fs_info,
-					  enum btrfs_subpage_type type)
+				size_t fsize, enum btrfs_subpage_type type)
 {
 	struct btrfs_subpage *ret;
 	unsigned int real_size;
 
-	ASSERT(fs_info->sectorsize < PAGE_SIZE);
+	ASSERT(fs_info->sectorsize < fsize);
 
 	real_size = struct_size(ret, bitmaps,
-			BITS_TO_LONGS(btrfs_bitmap_nr_max * fs_info->sectors_per_page));
+			BITS_TO_LONGS(btrfs_bitmap_nr_max *
+				      (fsize >> fs_info->sectorsize_bits)));
 	ret = kzalloc(real_size, GFP_NOFS);
 	if (!ret)
 		return ERR_PTR(-ENOMEM);
···
 {
 	struct btrfs_subpage *subpage;
 
-	if (!btrfs_is_subpage(fs_info, folio->mapping))
+	if (!btrfs_meta_is_subpage(fs_info))
 		return;
 
 	ASSERT(folio_test_private(folio) && folio->mapping);
···
 {
 	struct btrfs_subpage *subpage;
 
-	if (!btrfs_is_subpage(fs_info, folio->mapping))
+	if (!btrfs_meta_is_subpage(fs_info))
 		return;
 
 	ASSERT(folio_test_private(folio) && folio->mapping);
···
 	 */
 	if (folio->mapping)
 		ASSERT(folio_pos(folio) <= start &&
-		       start + len <= folio_pos(folio) + PAGE_SIZE);
+		       start + len <= folio_pos(folio) + folio_size(folio));
 }
 
 #define subpage_calc_start_bit(fs_info, folio, name, start, len)	\
 ({									\
-	unsigned int __start_bit;						\
+	unsigned int __start_bit;					\
+	const unsigned int blocks_per_folio =				\
+			   btrfs_blocks_per_folio(fs_info, folio);	\
 									\
 	btrfs_subpage_assert(fs_info, folio, start, len);		\
 	__start_bit = offset_in_page(start) >> fs_info->sectorsize_bits; \
-	__start_bit += fs_info->sectors_per_page * btrfs_bitmap_nr_##name; \
+	__start_bit += blocks_per_folio * btrfs_bitmap_nr_##name;	\
 	__start_bit;							\
 })
 
···
 	if (folio_pos(folio) >= orig_start + orig_len)
 		*len = 0;
 	else
-		*len = min_t(u64, folio_pos(folio) + PAGE_SIZE,
+		*len = min_t(u64, folio_pos(folio) + folio_size(folio),
 			     orig_start + orig_len) - *start;
 }
 
···
 
 	ASSERT(folio_test_locked(folio));
 
-	if (unlikely(!fs_info) || !btrfs_is_subpage(fs_info, folio->mapping)) {
+	if (unlikely(!fs_info) || !btrfs_is_subpage(fs_info, folio)) {
 		folio_unlock(folio);
 		return;
 	}
···
 				 struct folio *folio, unsigned long bitmap)
 {
 	struct btrfs_subpage *subpage = folio_get_private(folio);
-	const int start_bit = fs_info->sectors_per_page * btrfs_bitmap_nr_locked;
+	const unsigned int blocks_per_folio = btrfs_blocks_per_folio(fs_info, folio);
+	const int start_bit = blocks_per_folio * btrfs_bitmap_nr_locked;
 	unsigned long flags;
 	bool last = false;
 	int cleared = 0;
 	int bit;
 
-	if (!btrfs_is_subpage(fs_info, folio->mapping)) {
+	if (!btrfs_is_subpage(fs_info, folio)) {
 		folio_unlock(folio);
 		return;
 	}
···
 	}
 
 	spin_lock_irqsave(&subpage->lock, flags);
-	for_each_set_bit(bit, &bitmap, fs_info->sectors_per_page) {
+	for_each_set_bit(bit, &bitmap, blocks_per_folio) {
 		if (test_and_clear_bit(bit + start_bit, subpage->bitmaps))
 			cleared++;
 	}
···
 		folio_unlock(folio);
 }
 
-#define subpage_test_bitmap_all_set(fs_info, subpage, name)		\
+#define subpage_test_bitmap_all_set(fs_info, folio, name)		\
+({									\
+	struct btrfs_subpage *subpage = folio_get_private(folio);	\
+	const unsigned int blocks_per_folio =				\
+				btrfs_blocks_per_folio(fs_info, folio); \
+									\
 	bitmap_test_range_all_set(subpage->bitmaps,			\
-			fs_info->sectors_per_page * btrfs_bitmap_nr_##name, \
-			fs_info->sectors_per_page)
+			blocks_per_folio * btrfs_bitmap_nr_##name,	\
+			blocks_per_folio);				\
+})
 
-#define subpage_test_bitmap_all_zero(fs_info, subpage, name)		\
+#define subpage_test_bitmap_all_zero(fs_info, folio, name)		\
+({									\
+	struct btrfs_subpage *subpage = folio_get_private(folio);	\
+	const unsigned int blocks_per_folio =				\
+				btrfs_blocks_per_folio(fs_info, folio); \
+									\
 	bitmap_test_range_all_zero(subpage->bitmaps,			\
-			fs_info->sectors_per_page * btrfs_bitmap_nr_##name, \
-			fs_info->sectors_per_page)
+			blocks_per_folio * btrfs_bitmap_nr_##name,	\
+			blocks_per_folio);				\
+})
 
 void btrfs_subpage_set_uptodate(const struct btrfs_fs_info *fs_info,
 				struct folio *folio, u64 start, u32 len)
···
 
 	spin_lock_irqsave(&subpage->lock, flags);
 	bitmap_set(subpage->bitmaps, start_bit, len >> fs_info->sectorsize_bits);
-	if (subpage_test_bitmap_all_set(fs_info, subpage, uptodate))
+	if (subpage_test_bitmap_all_set(fs_info, folio, uptodate))
 		folio_mark_uptodate(folio);
 	spin_unlock_irqrestore(&subpage->lock, flags);
 }
···
 
 	spin_lock_irqsave(&subpage->lock, flags);
 	bitmap_clear(subpage->bitmaps, start_bit, len >> fs_info->sectorsize_bits);
-	if (subpage_test_bitmap_all_zero(fs_info, subpage, dirty))
+	if (subpage_test_bitmap_all_zero(fs_info, folio, dirty))
 		last = true;
 	spin_unlock_irqrestore(&subpage->lock, flags);
 	return last;
···
 
 	spin_lock_irqsave(&subpage->lock, flags);
 	bitmap_clear(subpage->bitmaps, start_bit, len >> fs_info->sectorsize_bits);
-	if (subpage_test_bitmap_all_zero(fs_info, subpage, writeback)) {
+	if (subpage_test_bitmap_all_zero(fs_info, folio, writeback)) {
 		ASSERT(folio_test_writeback(folio));
 		folio_end_writeback(folio);
 	}
···
 
 	spin_lock_irqsave(&subpage->lock, flags);
 	bitmap_clear(subpage->bitmaps, start_bit, len >> fs_info->sectorsize_bits);
-	if (subpage_test_bitmap_all_zero(fs_info, subpage, ordered))
+	if (subpage_test_bitmap_all_zero(fs_info, folio, ordered))
 		folio_clear_ordered(folio);
 	spin_unlock_irqrestore(&subpage->lock, flags);
 }
···
 
 	spin_lock_irqsave(&subpage->lock, flags);
 	bitmap_set(subpage->bitmaps, start_bit, len >> fs_info->sectorsize_bits);
-	if (subpage_test_bitmap_all_set(fs_info, subpage, checked))
+	if (subpage_test_bitmap_all_set(fs_info, folio, checked))
 		folio_set_checked(folio);
 	spin_unlock_irqrestore(&subpage->lock, flags);
 }
···
 			    struct folio *folio, u64 start, u32 len)	\
 {									\
 	if (unlikely(!fs_info) ||					\
-	    !btrfs_is_subpage(fs_info, folio->mapping)) {		\
+	    !btrfs_is_subpage(fs_info, folio)) {			\
 		folio_set_func(folio);					\
 		return;							\
 	}								\
···
 			      struct folio *folio, u64 start, u32 len)	\
 {									\
 	if (unlikely(!fs_info) ||					\
-	    !btrfs_is_subpage(fs_info, folio->mapping)) {		\
+	    !btrfs_is_subpage(fs_info, folio)) {			\
 		folio_clear_func(folio);				\
 		return;							\
 	}								\
···
 			     struct folio *folio, u64 start, u32 len)	\
 {									\
 	if (unlikely(!fs_info) ||					\
-	    !btrfs_is_subpage(fs_info, folio->mapping))			\
+	    !btrfs_is_subpage(fs_info, folio))				\
 		return folio_test_func(folio);				\
 	return btrfs_subpage_test_##name(fs_info, folio, start, len);	\
 }									\
···
 				  struct folio *folio, u64 start, u32 len) \
 {									\
 	if (unlikely(!fs_info) ||					\
-	    !btrfs_is_subpage(fs_info, folio->mapping)) {		\
+	    !btrfs_is_subpage(fs_info, folio)) {			\
 		folio_set_func(folio);					\
 		return;							\
 	}								\
···
 				    struct folio *folio, u64 start, u32 len) \
 {									\
 	if (unlikely(!fs_info) ||					\
-	    !btrfs_is_subpage(fs_info, folio->mapping)) {		\
+	    !btrfs_is_subpage(fs_info, folio)) {			\
 		folio_clear_func(folio);				\
 		return;							\
 	}								\
···
 				   struct folio *folio, u64 start, u32 len) \
 {									\
 	if (unlikely(!fs_info) ||					\
-	    !btrfs_is_subpage(fs_info, folio->mapping))			\
+	    !btrfs_is_subpage(fs_info, folio))				\
 		return folio_test_func(folio);				\
 	btrfs_subpage_clamp_range(folio, &start, &len);			\
 	return btrfs_subpage_test_##name(fs_info, folio, start, len);	\
+}									\
+void btrfs_meta_folio_set_##name(struct folio *folio, const struct extent_buffer *eb) \
+{									\
+	if (!btrfs_meta_is_subpage(eb->fs_info)) {			\
+		folio_set_func(folio);					\
+		return;							\
+	}								\
+	btrfs_subpage_set_##name(eb->fs_info, folio, eb->start, eb->len); \
+}									\
+void btrfs_meta_folio_clear_##name(struct folio *folio, const struct extent_buffer *eb) \
+{									\
+	if (!btrfs_meta_is_subpage(eb->fs_info)) {			\
+		folio_clear_func(folio);				\
+		return;							\
+	}								\
+	btrfs_subpage_clear_##name(eb->fs_info, folio, eb->start, eb->len); \
+}									\
+bool btrfs_meta_folio_test_##name(struct folio *folio, const struct extent_buffer *eb) \
+{									\
+	if (!btrfs_meta_is_subpage(eb->fs_info))			\
+		return folio_test_func(folio);				\
+	return btrfs_subpage_test_##name(eb->fs_info, folio, eb->start, eb->len); \
 }
 IMPLEMENT_BTRFS_PAGE_OPS(uptodate, folio_mark_uptodate, folio_clear_uptodate,
 			 folio_test_uptodate);
···
 IMPLEMENT_BTRFS_PAGE_OPS(checked, folio_set_checked, folio_clear_checked,
 			 folio_test_checked);
 
-#define GET_SUBPAGE_BITMAP(subpage, fs_info, name, dst)			\
+#define GET_SUBPAGE_BITMAP(fs_info, folio, name, dst)			\
 {									\
-	const int sectors_per_page = fs_info->sectors_per_page;		\
+	const unsigned int blocks_per_folio =				\
+				btrfs_blocks_per_folio(fs_info, folio);	\
+	const struct btrfs_subpage *subpage = folio_get_private(folio);	\
 									\
-	ASSERT(sectors_per_page < BITS_PER_LONG);			\
+	ASSERT(blocks_per_folio < BITS_PER_LONG);			\
 	*dst = bitmap_read(subpage->bitmaps,				\
-			   sectors_per_page * btrfs_bitmap_nr_##name,	\
-			   sectors_per_page);				\
+			   blocks_per_folio * btrfs_bitmap_nr_##name,	\
+			   blocks_per_folio);				\
 }
 
 #define SUBPAGE_DUMP_BITMAP(fs_info, folio, name, start, len)		\
 {									\
-	const struct btrfs_subpage *subpage = folio_get_private(folio);	\
 	unsigned long bitmap;						\
+	const unsigned int blocks_per_folio =				\
+				btrfs_blocks_per_folio(fs_info, folio);	\
 									\
-	GET_SUBPAGE_BITMAP(subpage, fs_info, name, &bitmap);		\
+	GET_SUBPAGE_BITMAP(fs_info, folio, name, &bitmap);		\
 	btrfs_warn(fs_info,						\
 	"dumpping bitmap start=%llu len=%u folio=%llu " #name "_bitmap=%*pbl", \
 		   start, len, folio_pos(folio),			\
-		   fs_info->sectors_per_page, &bitmap);			\
+		   blocks_per_folio, &bitmap);				\
 }
 
 /*
···
 	if (!IS_ENABLED(CONFIG_BTRFS_ASSERT))
 		return;
 
-	if (!btrfs_is_subpage(fs_info, folio->mapping)) {
+	if (!btrfs_is_subpage(fs_info, folio)) {
 		ASSERT(!folio_test_dirty(folio));
 		return;
 	}
···
 	int ret;
 
 	ASSERT(folio_test_locked(folio));
-	if (unlikely(!fs_info) || !btrfs_is_subpage(fs_info, folio->mapping))
+	if (unlikely(!fs_info) || !btrfs_is_subpage(fs_info, folio))
 		return;
 
 	subpage = folio_get_private(folio);
···
 	}
 	bitmap_set(subpage->bitmaps, start_bit, nbits);
 	ret = atomic_add_return(nbits, &subpage->nr_locked);
-	ASSERT(ret <= fs_info->sectors_per_page);
+	ASSERT(ret <= btrfs_blocks_per_folio(fs_info, folio));
 	spin_unlock_irqrestore(&subpage->lock, flags);
+}
+
+/*
+ * Clear the dirty flag for the folio.
+ *
+ * If the affected folio is no longer dirty, return true. Otherwise return false.
+ */
+bool btrfs_meta_folio_clear_and_test_dirty(struct folio *folio, const struct extent_buffer *eb)
+{
+	bool last;
+
+	if (!btrfs_meta_is_subpage(eb->fs_info)) {
+		folio_clear_dirty_for_io(folio);
+		return true;
+	}
+
+	last = btrfs_subpage_clear_and_test_dirty(eb->fs_info, folio, eb->start, eb->len);
+	if (last) {
+		folio_clear_dirty_for_io(folio);
+		return true;
+	}
+	return false;
 }
 
 void __cold btrfs_subpage_dump_bitmap(const struct btrfs_fs_info *fs_info,
 				      struct folio *folio, u64 start, u32 len)
 {
 	struct btrfs_subpage *subpage;
-	const u32 sectors_per_page = fs_info->sectors_per_page;
+	const unsigned int blocks_per_folio = btrfs_blocks_per_folio(fs_info, folio);
 	unsigned long uptodate_bitmap;
 	unsigned long dirty_bitmap;
 	unsigned long writeback_bitmap;
···
 	unsigned long flags;
 
 	ASSERT(folio_test_private(folio) && folio_get_private(folio));
-	ASSERT(sectors_per_page > 1);
+	ASSERT(blocks_per_folio > 1);
 	subpage = folio_get_private(folio);
 
 	spin_lock_irqsave(&subpage->lock, flags);
-	GET_SUBPAGE_BITMAP(subpage, fs_info, uptodate, &uptodate_bitmap);
-	GET_SUBPAGE_BITMAP(subpage, fs_info, dirty, &dirty_bitmap);
-	GET_SUBPAGE_BITMAP(subpage, fs_info, writeback, &writeback_bitmap);
-	GET_SUBPAGE_BITMAP(subpage, fs_info, ordered, &ordered_bitmap);
-	GET_SUBPAGE_BITMAP(subpage, fs_info, checked, &checked_bitmap);
-	GET_SUBPAGE_BITMAP(subpage, fs_info, locked, &locked_bitmap);
+	GET_SUBPAGE_BITMAP(fs_info, folio, uptodate, &uptodate_bitmap);
+	GET_SUBPAGE_BITMAP(fs_info, folio, dirty, &dirty_bitmap);
+	GET_SUBPAGE_BITMAP(fs_info, folio, writeback, &writeback_bitmap);
+	GET_SUBPAGE_BITMAP(fs_info, folio, ordered, &ordered_bitmap);
+	GET_SUBPAGE_BITMAP(fs_info, folio, checked, &checked_bitmap);
+	GET_SUBPAGE_BITMAP(fs_info, folio, locked, &locked_bitmap);
 	spin_unlock_irqrestore(&subpage->lock, flags);
 
 	dump_page(folio_page(folio, 0), "btrfs subpage dump");
 	btrfs_warn(fs_info,
 "start=%llu len=%u page=%llu, bitmaps uptodate=%*pbl dirty=%*pbl locked=%*pbl writeback=%*pbl ordered=%*pbl checked=%*pbl",
 		    start, len, folio_pos(folio),
-		    sectors_per_page, &uptodate_bitmap,
-		    sectors_per_page, &dirty_bitmap,
-		    sectors_per_page, &locked_bitmap,
-		    sectors_per_page, &writeback_bitmap,
-		    sectors_per_page, &ordered_bitmap,
-		    sectors_per_page, &checked_bitmap);
+		    blocks_per_folio, &uptodate_bitmap,
+		    blocks_per_folio, &dirty_bitmap,
+		    blocks_per_folio, &locked_bitmap,
+		    blocks_per_folio, &writeback_bitmap,
+		    blocks_per_folio, &ordered_bitmap,
+		    blocks_per_folio, &checked_bitmap);
 }
 
 void btrfs_get_subpage_dirty_bitmap(struct btrfs_fs_info *fs_info,
···
 	unsigned long flags;
 
 	ASSERT(folio_test_private(folio) && folio_get_private(folio));
-	ASSERT(fs_info->sectors_per_page > 1);
+	ASSERT(btrfs_blocks_per_folio(fs_info, folio) > 1);
 	subpage = folio_get_private(folio);
 
 	spin_lock_irqsave(&subpage->lock, flags);
-	GET_SUBPAGE_BITMAP(subpage, fs_info, dirty, ret_bitmap);
+	GET_SUBPAGE_BITMAP(fs_info, folio, dirty, ret_bitmap);
 	spin_unlock_irqrestore(&subpage->lock, flags);
 }

fs/btrfs/subpage.h

···
 #include <linux/spinlock.h>
 #include <linux/atomic.h>
 #include <linux/sizes.h>
+#include "btrfs_inode.h"
+#include "fs.h"
 
 struct address_space;
 struct folio;
-struct btrfs_fs_info;
 
 /*
  * Extra info for subpapge bitmap.
···
 	BTRFS_SUBPAGE_DATA,
 };
 
-#if PAGE_SIZE > SZ_4K
-bool btrfs_is_subpage(const struct btrfs_fs_info *fs_info, struct address_space *mapping);
-#else
-static inline bool btrfs_is_subpage(const struct btrfs_fs_info *fs_info,
-				    struct address_space *mapping)
+#if PAGE_SIZE > BTRFS_MIN_BLOCKSIZE
+/*
+ * Subpage support for metadata is more complex, as we can have dummy extent
+ * buffers, where folios have no mapping to determine the owning inode.
+ *
+ * Thankfully we only need to check if node size is smaller than page size.
+ * Even with larger folio support, we will only allocate a folio as large as
+ * node size.
+ * Thus if nodesize < PAGE_SIZE, we know metadata needs need to subpage routine.
+ */
+static inline bool btrfs_meta_is_subpage(const struct btrfs_fs_info *fs_info)
 {
+	return fs_info->nodesize < PAGE_SIZE;
+}
+static inline bool btrfs_is_subpage(const struct btrfs_fs_info *fs_info,
+				    struct folio *folio)
+{
+	if (folio->mapping && folio->mapping->host)
+		ASSERT(is_data_inode(BTRFS_I(folio->mapping->host)));
+	return fs_info->sectorsize < folio_size(folio);
+}
+#else
+static inline bool btrfs_meta_is_subpage(const struct btrfs_fs_info *fs_info)
+{
+	return false;
+}
+static inline bool btrfs_is_subpage(const struct btrfs_fs_info *fs_info,
+				    struct folio *folio)
+{
+	if (folio->mapping && folio->mapping->host)
+		ASSERT(is_data_inode(BTRFS_I(folio->mapping->host)));
 	return false;
 }
 #endif
 
 int btrfs_attach_subpage(const struct btrfs_fs_info *fs_info,
 			 struct folio *folio, enum btrfs_subpage_type type);
-void btrfs_detach_subpage(const struct btrfs_fs_info *fs_info, struct folio *folio);
+void btrfs_detach_subpage(const struct btrfs_fs_info *fs_info, struct folio *folio,
+			  enum btrfs_subpage_type type);
 
 /* Allocate additional data where page represents more than one sector */
 struct btrfs_subpage *btrfs_alloc_subpage(const struct btrfs_fs_info *fs_info,
-					  enum btrfs_subpage_type type);
+				size_t fsize, enum btrfs_subpage_type type);
 void btrfs_free_subpage(struct btrfs_subpage *subpage);
 
 void btrfs_folio_inc_eb_refs(const struct btrfs_fs_info *fs_info, struct folio *folio);
···
  * btrfs_folio_clamp_*() are similar to btrfs_folio_*(), except the range doesn't
  * need to be inside the page. Those functions will truncate the range
  * automatically.
+ *
+ * Both btrfs_folio_*() and btrfs_folio_clamp_*() are for data folios.
+ *
+ * For metadata, one should use btrfs_meta_folio_*() helpers instead, and there
+ * is no clamp version for metadata helpers, as we either go subpage
+ * (nodesize < PAGE_SIZE) or go regular folio helpers (nodesize >= PAGE_SIZE,
+ * and our folio is never larger than nodesize).
  */
 #define DECLARE_BTRFS_SUBPAGE_OPS(name)					\
 void btrfs_subpage_set_##name(const struct btrfs_fs_info *fs_info,	\
···
 void btrfs_folio_clamp_clear_##name(const struct btrfs_fs_info *fs_info,	\
 		struct folio *folio, u64 start, u32 len);			\
 bool btrfs_folio_clamp_test_##name(const struct btrfs_fs_info *fs_info,	\
-		struct folio *folio, u64 start, u32 len);
+		struct folio *folio, u64 start, u32 len);		\
+void btrfs_meta_folio_set_##name(struct folio *folio, const struct extent_buffer *eb); \
+void btrfs_meta_folio_clear_##name(struct folio *folio, const struct extent_buffer *eb); \
+bool btrfs_meta_folio_test_##name(struct folio *folio, const struct extent_buffer *eb);
 
 DECLARE_BTRFS_SUBPAGE_OPS(uptodate);
 DECLARE_BTRFS_SUBPAGE_OPS(dirty);
···
 
 void btrfs_folio_assert_not_dirty(const struct btrfs_fs_info *fs_info,
 				  struct folio *folio, u64 start, u32 len);
+bool btrfs_meta_folio_clear_and_test_dirty(struct folio *folio, const struct extent_buffer *eb);
 void btrfs_get_subpage_dirty_bitmap(struct btrfs_fs_info *fs_info,
 				    struct folio *folio,
 				    unsigned long *ret_bitmap);

fs/btrfs/super.c

···
 	u32 thread_pool_size;
 	unsigned long long mount_opt;
 	unsigned long compress_type:4;
-	unsigned int compress_level;
+	int compress_level;
 	refcount_t refs;
 };
 
···
 static int btrfs_fill_super(struct super_block *sb,
 			    struct btrfs_fs_devices *fs_devices)
 {
-	struct inode *inode;
+	struct btrfs_inode *inode;
 	struct btrfs_fs_info *fs_info = btrfs_sb(sb);
 	int err;
 
···
 		goto fail_close;
 	}
 
-	sb->s_root = d_make_root(inode);
+	sb->s_root = d_make_root(&inode->vfs_inode);
 	if (!sb->s_root) {
 		err = -ENOMEM;
 		goto fail_close;

fs/btrfs/sysfs.c

···
 {
 	ssize_t ret = 0;
 
-	/* An artificial limit to only support 4K and PAGE_SIZE */
+	if (BTRFS_MIN_BLOCKSIZE != SZ_4K && BTRFS_MIN_BLOCKSIZE != PAGE_SIZE)
+		ret += sysfs_emit_at(buf, ret, "%u ", BTRFS_MIN_BLOCKSIZE);
 	if (PAGE_SIZE > SZ_4K)
 		ret += sysfs_emit_at(buf, ret, "%u ", SZ_4K);
 	ret += sysfs_emit_at(buf, ret, "%lu\n", PAGE_SIZE);
···
 	/* Separate value from input in policy:value format. */
 	value_str = strchr(param, ':');
 	if (value_str) {
-		int ret;
+		char *retptr;
 
 		*value_str = 0;
 		value_str++;
 		if (!value_ret)
 			return -EINVAL;
-		ret = kstrtos64(value_str, 10, value_ret);
-		if (ret)
+
+		*value_ret = memparse(value_str, &retptr);
+		/* There could be any trailing typos after the value. */
+		retptr = skip_spaces(retptr);
+		if (*retptr != 0 || *value_ret <= 0)
 			return -EINVAL;
-		if (*value_ret < 0)
-			return -ERANGE;
 	}
 #endif
 

fs/btrfs/sysfs.h

···
 #include <linux/compiler_types.h>
 #include <linux/kobject.h>
 
+struct block_device;
 struct btrfs_fs_info;
 struct btrfs_device;
 struct btrfs_fs_devices;

fs/btrfs/tests/extent-io-tests.c

···
 		goto out;
 	}
 
-	eb = __alloc_dummy_extent_buffer(fs_info, 0, nodesize);
+	eb = alloc_dummy_extent_buffer(fs_info, 0);
 	if (!eb) {
 		test_std_err(TEST_ALLOC_ROOT);
 		ret = -ENOMEM;
···
 	 * Test again for case where the tree block is sectorsize aligned but
 	 * not nodesize aligned.
 	 */
-	eb = __alloc_dummy_extent_buffer(fs_info, sectorsize, nodesize);
+	eb = alloc_dummy_extent_buffer(fs_info, sectorsize);
 	if (!eb) {
 		test_std_err(TEST_ALLOC_ROOT);
 		ret = -ENOMEM;
···
 		goto out;
 	}
 
-	eb = __alloc_dummy_extent_buffer(fs_info, SZ_1M, nodesize);
+	eb = alloc_dummy_extent_buffer(fs_info, SZ_1M);
 	if (!eb) {
 		test_std_err(TEST_ALLOC_EXTENT_BUFFER);
 		ret = -ENOMEM;

fs/btrfs/tests/extent-map-tests.c

···
 	ret = btrfs_add_chunk_map(fs_info, map);
 	if (ret) {
 		test_err("error adding chunk map to mapping tree");
+		btrfs_free_chunk_map(map);
 		goto out_free;
 	}
 

fs/btrfs/transaction.c

···
 			cache = list_first_entry(&transaction->deleted_bgs,
 						 struct btrfs_block_group,
 						 bg_list);
+			/*
+			 * Not strictly necessary to lock, as no other task will be using a
+			 * block_group on the deleted_bgs list during a transaction abort.
+			 */
+			spin_lock(&transaction->fs_info->unused_bgs_lock);
 			list_del_init(&cache->bg_list);
+			spin_unlock(&transaction->fs_info->unused_bgs_lock);
 			btrfs_unfreeze_block_group(cache);
 			btrfs_put_block_group(cache);
 		}
···
 	struct btrfs_root *root = pending->root;
 	struct btrfs_root *parent_root;
 	struct btrfs_block_rsv *rsv;
-	struct inode *parent_inode = &pending->dir->vfs_inode;
+	struct btrfs_inode *parent_inode = pending->dir;
 	struct btrfs_path *path;
 	struct btrfs_dir_item *dir_item;
 	struct extent_buffer *tmp;
···
 	 * filesystem.
 	 */
 	nofs_flags = memalloc_nofs_save();
-	pending->error = fscrypt_setup_filename(parent_inode,
+	pending->error = fscrypt_setup_filename(&parent_inode->vfs_inode,
 						&pending->dentry->d_name, 0,
 						&fname);
 	memalloc_nofs_restore(nofs_flags);
···
 	}
 
 	key.objectid = objectid;
-	key.offset = (u64)-1;
 	key.type = BTRFS_ROOT_ITEM_KEY;
+	key.offset = (u64)-1;
 
 	rsv = trans->block_rsv;
 	trans->block_rsv = &pending->block_rsv;
···
 	trace_btrfs_space_reservation(fs_info, "transaction",
 				      trans->transid,
 				      trans->bytes_reserved, 1);
-	parent_root = BTRFS_I(parent_inode)->root;
+	parent_root = parent_inode->root;
 	ret = record_root_in_trans(trans, parent_root, 0);
 	if (ret)
 		goto fail;
-	cur_time = current_time(parent_inode);
+	cur_time = current_time(&parent_inode->vfs_inode);
 
 	/*
 	 * insert the directory item
 	 */
-	ret = btrfs_set_inode_index(BTRFS_I(parent_inode), &index);
+	ret = btrfs_set_inode_index(parent_inode, &index);
 	if (ret) {
 		btrfs_abort_transaction(trans, ret);
 		goto fail;
···
 
 	/* check if there is a file/dir which has the same name. */
 	dir_item = btrfs_lookup_dir_item(NULL, parent_root, path,
-					 btrfs_ino(BTRFS_I(parent_inode)),
+					 btrfs_ino(parent_inode),
 					 &fname.disk_name, 0);
 	if (dir_item != NULL && !IS_ERR(dir_item)) {
 		pending->error = -EEXIST;
···
 	 */
 	ret = btrfs_add_root_ref(trans, objectid,
 				 btrfs_root_id(parent_root),
-				 btrfs_ino(BTRFS_I(parent_inode)), index,
+				 btrfs_ino(parent_inode), index,
 				 &fname.disk_name);
 	if (ret) {
 		btrfs_abort_transaction(trans, ret);
···
 		goto fail;
 
 	ret = btrfs_insert_dir_item(trans, &fname.disk_name,
-				    BTRFS_I(parent_inode), &key, BTRFS_FT_DIR,
+				    parent_inode, &key, BTRFS_FT_DIR,
 				    index);
 	if (ret) {
 		btrfs_abort_transaction(trans, ret);
 		goto fail;
 	}
 
-	btrfs_i_size_write(BTRFS_I(parent_inode), parent_inode->i_size +
+	btrfs_i_size_write(parent_inode, parent_inode->vfs_inode.i_size +
 						  fname.disk_name.len * 2);
-	inode_set_mtime_to_ts(parent_inode,
-			      inode_set_ctime_current(parent_inode));
-	ret = btrfs_update_inode_fallback(trans, BTRFS_I(parent_inode));
+	inode_set_mtime_to_ts(&parent_inode->vfs_inode,
+			      inode_set_ctime_current(&parent_inode->vfs_inode));
+	ret = btrfs_update_inode_fallback(trans, parent_inode);
 	if (ret) {
 		btrfs_abort_transaction(trans, ret);
 		goto fail;
···
 
        list_for_each_entry_safe(block_group, tmp, &trans->new_bgs, bg_list) {
                btrfs_dec_delayed_refs_rsv_bg_inserts(fs_info);
+		/*
+		* Not strictly necessary to lock, as no other task will be using a
+		* block_group on the new_bgs list during a transaction abort.
+		*/
+	       spin_lock(&fs_info->unused_bgs_lock);
                list_del_init(&block_group->bg_list);
+	       btrfs_put_block_group(block_group);
+	       spin_unlock(&fs_info->unused_bgs_lock);
        }
 }
 

fs/btrfs/tree-log.c

···
  * and once to do all the other items.
  */
 
-static struct inode *btrfs_iget_logging(u64 objectid, struct btrfs_root *root)
+static struct btrfs_inode *btrfs_iget_logging(u64 objectid, struct btrfs_root *root)
 {
 	unsigned int nofs_flag;
-	struct inode *inode;
+	struct btrfs_inode *inode;
 
 	/*
 	 * We're holding a transaction handle whether we are logging or
···
 }
 
 /*
- * Item overwrite used by replay and tree logging.  eb, slot and key all refer
- * to the src data we are copying out.
+ * Item overwrite used by log replay. The given eb, slot and key all refer to
+ * the source data we are copying out.
  *
- * root is the tree we are copying into, and path is a scratch
- * path for use in this function (it should be released on entry and
- * will be released on exit).
+ * The given root is for the tree we are copying into, and path is a scratch
+ * path for use in this function (it should be released on entry and will be
+ * released on exit).
  *
  * If the key is already in the destination tree the existing item is
  * overwritten.  If the existing item isn't big enough, it is extended.
···
 	int save_old_i_size = 0;
 	unsigned long src_ptr;
 	unsigned long dst_ptr;
+	struct extent_buffer *dst_eb;
+	int dst_slot;
 	bool inode_item = key->type == BTRFS_INODE_ITEM_KEY;
 
 	/*
···
 	if (ret < 0)
 		return ret;
 
+	dst_eb = path->nodes[0];
+	dst_slot = path->slots[0];
+
 	if (ret == 0) {
 		char *src_copy;
-		char *dst_copy;
-		u32 dst_size = btrfs_item_size(path->nodes[0],
-						  path->slots[0]);
+		const u32 dst_size = btrfs_item_size(dst_eb, dst_slot);
+
 		if (dst_size != item_size)
 			goto insert;
 
···
 			btrfs_release_path(path);
 			return 0;
 		}
-		dst_copy = kmalloc(item_size, GFP_NOFS);
 		src_copy = kmalloc(item_size, GFP_NOFS);
-		if (!dst_copy || !src_copy) {
+		if (!src_copy) {
 			btrfs_release_path(path);
-			kfree(dst_copy);
-			kfree(src_copy);
 			return -ENOMEM;
 		}
 
 		read_extent_buffer(eb, src_copy, src_ptr, item_size);
+		dst_ptr = btrfs_item_ptr_offset(dst_eb, dst_slot);
+		ret = memcmp_extent_buffer(dst_eb, src_copy, dst_ptr, item_size);
 
-		dst_ptr = btrfs_item_ptr_offset(path->nodes[0], path->slots[0]);
-		read_extent_buffer(path->nodes[0], dst_copy, dst_ptr,
-				   item_size);
-		ret = memcmp(dst_copy, src_copy, item_size);
-
-		kfree(dst_copy);
 		kfree(src_copy);
 		/*
 		 * they have the same contents, just return, this saves
···
 			u64 nbytes;
 			u32 mode;
 
-			item = btrfs_item_ptr(path->nodes[0], path->slots[0],
+			item = btrfs_item_ptr(dst_eb, dst_slot,
 					      struct btrfs_inode_item);
-			nbytes = btrfs_inode_nbytes(path->nodes[0], item);
+			nbytes = btrfs_inode_nbytes(dst_eb, item);
 			item = btrfs_item_ptr(eb, slot,
 					      struct btrfs_inode_item);
 			btrfs_set_inode_nbytes(eb, item, nbytes);
···
 				      key, item_size);
 	path->skip_release_on_error = 0;
 
+	dst_eb = path->nodes[0];
+	dst_slot = path->slots[0];
+
 	/* make sure any existing item is the correct size */
 	if (ret == -EEXIST || ret == -EOVERFLOW) {
-		u32 found_size;
-		found_size = btrfs_item_size(path->nodes[0],
-						path->slots[0]);
+		const u32 found_size = btrfs_item_size(dst_eb, dst_slot);
+
 		if (found_size > item_size)
 			btrfs_truncate_item(trans, path, item_size, 1);
 		else if (found_size < item_size)
···
 	} else if (ret) {
 		return ret;
 	}
-	dst_ptr = btrfs_item_ptr_offset(path->nodes[0],
-					path->slots[0]);
+	dst_ptr = btrfs_item_ptr_offset(dst_eb, dst_slot);
 
 	/* don't overwrite an existing inode if the generation number
 	 * was logged as zero.  This is done when the tree logging code
···
 		dst_item = (struct btrfs_inode_item *)dst_ptr;
 
 		if (btrfs_inode_generation(eb, src_item) == 0) {
-			struct extent_buffer *dst_eb = path->nodes[0];
 			const u64 ino_size = btrfs_inode_size(eb, src_item);
 
 			/*
···
 		}
 
 		if (S_ISDIR(btrfs_inode_mode(eb, src_item)) &&
-		    S_ISDIR(btrfs_inode_mode(path->nodes[0], dst_item))) {
+		    S_ISDIR(btrfs_inode_mode(dst_eb, dst_item))) {
 			save_old_i_size = 1;
-			saved_i_size = btrfs_inode_size(path->nodes[0],
-							dst_item);
+			saved_i_size = btrfs_inode_size(dst_eb, dst_item);
 		}
 	}
 
-	copy_extent_buffer(path->nodes[0], eb, dst_ptr,
-			   src_ptr, item_size);
+	copy_extent_buffer(dst_eb, eb, dst_ptr, src_ptr, item_size);
 
 	if (save_old_i_size) {
 		struct btrfs_inode_item *dst_item;
+
 		dst_item = (struct btrfs_inode_item *)dst_ptr;
-		btrfs_set_inode_size(path->nodes[0], dst_item, saved_i_size);
+		btrfs_set_inode_size(dst_eb, dst_item, saved_i_size);
 	}
 
 	/* make sure the generation is filled in */
 	if (key->type == BTRFS_INODE_ITEM_KEY) {
 		struct btrfs_inode_item *dst_item;
+
 		dst_item = (struct btrfs_inode_item *)dst_ptr;
-		if (btrfs_inode_generation(path->nodes[0], dst_item) == 0) {
-			btrfs_set_inode_generation(path->nodes[0], dst_item,
-						   trans->transid);
-		}
+		if (btrfs_inode_generation(dst_eb, dst_item) == 0)
+			btrfs_set_inode_generation(dst_eb, dst_item, trans->transid);
 	}
 no_copy:
 	btrfs_release_path(path);
···
  * simple helper to read an inode off the disk from a given root
  * This can only be called for subvolume roots and not for the log
  */
-static noinline struct inode *read_one_inode(struct btrfs_root *root,
-					     u64 objectid)
+static noinline struct btrfs_inode *read_one_inode(struct btrfs_root *root,
+						   u64 objectid)
 {
-	struct inode *inode;
+	struct btrfs_inode *inode;
 
 	inode = btrfs_iget_logging(objectid, root);
 	if (IS_ERR(inode))
-		inode = NULL;
+		return NULL;
 	return inode;
 }
 
···
 	u64 start = key->offset;
 	u64 nbytes = 0;
 	struct btrfs_file_extent_item *item;
-	struct inode *inode = NULL;
+	struct btrfs_inode *inode = NULL;
 	unsigned long size;
 	int ret = 0;
 
···
 	 * file.  This must be done before the btrfs_drop_extents run
 	 * so we don't try to drop this extent.
 	 */
-	ret = btrfs_lookup_file_extent(trans, root, path,
-			btrfs_ino(BTRFS_I(inode)), start, 0);
+	ret = btrfs_lookup_file_extent(trans, root, path, btrfs_ino(inode), start, 0);
 
 	if (ret == 0 &&
 	    (found_type == BTRFS_FILE_EXTENT_REG ||
 	     found_type == BTRFS_FILE_EXTENT_PREALLOC)) {
-		struct btrfs_file_extent_item cmp1;
-		struct btrfs_file_extent_item cmp2;
-		struct btrfs_file_extent_item *existing;
-		struct extent_buffer *leaf;
+		struct btrfs_file_extent_item existing;
+		unsigned long ptr;
 
-		leaf = path->nodes[0];
-		existing = btrfs_item_ptr(leaf, path->slots[0],
-					  struct btrfs_file_extent_item);
-
-		read_extent_buffer(eb, &cmp1, (unsigned long)item,
-				   sizeof(cmp1));
-		read_extent_buffer(leaf, &cmp2, (unsigned long)existing,
-				   sizeof(cmp2));
+		ptr = btrfs_item_ptr_offset(path->nodes[0], path->slots[0]);
+		read_extent_buffer(path->nodes[0], &existing, ptr, sizeof(existing));
 
 		/*
 		 * we already have a pointer to this exact extent,
 		 * we don't have to do anything
 		 */
-		if (memcmp(&cmp1, &cmp2, sizeof(cmp1)) == 0) {
+		if (memcmp_extent_buffer(eb, &existing, (unsigned long)item,
+					 sizeof(existing)) == 0) {
 			btrfs_release_path(path);
 			goto out;
 		}
···
 	drop_args.start = start;
 	drop_args.end = extent_end;
 	drop_args.drop_cache = true;
-	ret = btrfs_drop_extents(trans, root, BTRFS_I(inode), &drop_args);
+	ret = btrfs_drop_extents(trans, root, inode, &drop_args);
 	if (ret)
 		goto out;
 
···
 				(unsigned long)item,  sizeof(*item));
 
 		ins.objectid = btrfs_file_extent_disk_bytenr(eb, item);
-		ins.offset = btrfs_file_extent_disk_num_bytes(eb, item);
 		ins.type = BTRFS_EXTENT_ITEM_KEY;
+		ins.offset = btrfs_file_extent_disk_num_bytes(eb, item);
 		offset = key->offset - btrfs_file_extent_offset(eb, item);
 
 		/*
···
 			goto out;
 	}
 
-	ret = btrfs_inode_set_file_extent_range(BTRFS_I(inode), start,
-						extent_end - start);
+	ret = btrfs_inode_set_file_extent_range(inode, start, extent_end - start);
 	if (ret)
 		goto out;
 
 update_inode:
-	btrfs_update_inode_bytes(BTRFS_I(inode), nbytes, drop_args.bytes_found);
-	ret = btrfs_update_inode(trans, BTRFS_I(inode));
+	btrfs_update_inode_bytes(inode, nbytes, drop_args.bytes_found);
+	ret = btrfs_update_inode(trans, inode);
 out:
-	iput(inode);
+	iput(&inode->vfs_inode);
 	return ret;
 }
 
···
 				      struct btrfs_dir_item *di)
 {
 	struct btrfs_root *root = dir->root;
-	struct inode *inode;
+	struct btrfs_inode *inode;
 	struct fscrypt_str name;
 	struct extent_buffer *leaf;
 	struct btrfs_key location;
···
 	if (ret)
 		goto out;
 
-	ret = unlink_inode_for_log_replay(trans, dir, BTRFS_I(inode), &name);
+	ret = unlink_inode_for_log_replay(trans, dir, inode, &name);
 out:
 	kfree(name.name);
-	iput(inode);
+	iput(&inode->vfs_inode);
 	return ret;
 }
 
···
 		u32 item_size;
 		u32 cur_offset = 0;
 		unsigned long base;
-		struct inode *victim_parent;
+		struct btrfs_inode *victim_parent;
 
 		leaf = path->nodes[0];
 
···
 					btrfs_release_path(path);
 
 					ret = unlink_inode_for_log_replay(trans,
-							BTRFS_I(victim_parent),
+							victim_parent,
 							inode, &victim_name);
 				}
-				iput(victim_parent);
+				iput(&victim_parent->vfs_inode);
 				kfree(victim_name.name);
 				if (ret)
 					return ret;
···
 			ret = !!btrfs_find_name_in_backref(log_eb, log_slot, &name);
 
 		if (!ret) {
-			struct inode *dir;
+			struct btrfs_inode *dir;
 
 			btrfs_release_path(path);
 			dir = read_one_inode(root, parent_id);
···
 				kfree(name.name);
 				goto out;
 			}
-			ret = unlink_inode_for_log_replay(trans, BTRFS_I(dir),
-						 inode, &name);
+			ret = unlink_inode_for_log_replay(trans, dir, inode, &name);
 			kfree(name.name);
-			iput(dir);
+			iput(&dir->vfs_inode);
 			if (ret)
 				goto out;
 			goto again;
···
 				  struct extent_buffer *eb, int slot,
 				  struct btrfs_key *key)
 {
-	struct inode *dir = NULL;
-	struct inode *inode = NULL;
+	struct btrfs_inode *dir = NULL;
+	struct btrfs_inode *inode = NULL;
 	unsigned long ref_ptr;
 	unsigned long ref_end;
 	struct fscrypt_str name = { 0 };
···
 		if (ret)
 			goto out;
 
-		ret = inode_in_dir(root, path, btrfs_ino(BTRFS_I(dir)),
-				   btrfs_ino(BTRFS_I(inode)), ref_index, &name);
+		ret = inode_in_dir(root, path, btrfs_ino(dir), btrfs_ino(inode),
+				   ref_index, &name);
 		if (ret < 0) {
 			goto out;
 		} else if (ret == 0) {
···
 			 * overwrite any existing back reference, and we don't
 			 * want to create dangling pointers in the directory.
 			 */
-			ret = __add_inode_ref(trans, root, path, log,
-					      BTRFS_I(dir), BTRFS_I(inode),
+			ret = __add_inode_ref(trans, root, path, log, dir, inode,
 					      inode_objectid, parent_objectid,
 					      ref_index, &name);
 			if (ret) {
···
 			}
 
 			/* insert our name */
-			ret = btrfs_add_link(trans, BTRFS_I(dir), BTRFS_I(inode),
-					     &name, 0, ref_index);
+			ret = btrfs_add_link(trans, dir, inode, &name, 0, ref_index);
 			if (ret)
 				goto out;
 
-			ret = btrfs_update_inode(trans, BTRFS_I(inode));
+			ret = btrfs_update_inode(trans, inode);
 			if (ret)
 				goto out;
 		}
···
 		kfree(name.name);
 		name.name = NULL;
 		if (log_ref_ver) {
-			iput(dir);
+			iput(&dir->vfs_inode);
 			dir = NULL;
 		}
 	}
···
 	 * dir index entries exist for a name but there is no inode reference
 	 * item with the same name.
 	 */
-	ret = unlink_old_inode_refs(trans, root, path, BTRFS_I(inode), eb, slot,
-				    key);
+	ret = unlink_old_inode_refs(trans, root, path, inode, eb, slot, key);
 	if (ret)
 		goto out;
 
···
 out:
 	btrfs_release_path(path);
 	kfree(name.name);
-	iput(dir);
-	iput(inode);
+	if (dir)
+		iput(&dir->vfs_inode);
+	if (inode)
+		iput(&inode->vfs_inode);
 	return ret;
 }
 
···
  * will free the inode.
  */
 static noinline int fixup_inode_link_count(struct btrfs_trans_handle *trans,
-					   struct inode *inode)
+					   struct btrfs_inode *inode)
 {
-	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_root *root = inode->root;
 	struct btrfs_path *path;
 	int ret;
 	u64 nlink = 0;
-	u64 ino = btrfs_ino(BTRFS_I(inode));
+	const u64 ino = btrfs_ino(inode);
 
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
 
-	ret = count_inode_refs(BTRFS_I(inode), path);
+	ret = count_inode_refs(inode, path);
 	if (ret < 0)
 		goto out;
 
 	nlink = ret;
 
-	ret = count_inode_extrefs(BTRFS_I(inode), path);
+	ret = count_inode_extrefs(inode, path);
 	if (ret < 0)
 		goto out;
 
···
 
 	ret = 0;
 
-	if (nlink != inode->i_nlink) {
-		set_nlink(inode, nlink);
-		ret = btrfs_update_inode(trans, BTRFS_I(inode));
+	if (nlink != inode->vfs_inode.i_nlink) {
+		set_nlink(&inode->vfs_inode, nlink);
+		ret = btrfs_update_inode(trans, inode);
 		if (ret)
 			goto out;
 	}
-	if (S_ISDIR(inode->i_mode))
-		BTRFS_I(inode)->index_cnt = (u64)-1;
+	if (S_ISDIR(inode->vfs_inode.i_mode))
+		inode->index_cnt = (u64)-1;
 
-	if (inode->i_nlink == 0) {
-		if (S_ISDIR(inode->i_mode)) {
+	if (inode->vfs_inode.i_nlink == 0) {
+		if (S_ISDIR(inode->vfs_inode.i_mode)) {
 			ret = replay_dir_deletes(trans, root, NULL, path,
 						 ino, 1);
 			if (ret)
···
 {
 	int ret;
 	struct btrfs_key key;
-	struct inode *inode;
 
 	key.objectid = BTRFS_TREE_LOG_FIXUP_OBJECTID;
 	key.type = BTRFS_ORPHAN_ITEM_KEY;
 	key.offset = (u64)-1;
 	while (1) {
+		struct btrfs_inode *inode;
+
 		ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
 		if (ret < 0)
 			break;
···
 		}
 
 		ret = fixup_inode_link_count(trans, inode);
-		iput(inode);
+		iput(&inode->vfs_inode);
 		if (ret)
 			break;
 
···
 {
 	struct btrfs_key key;
 	int ret = 0;
-	struct inode *inode;
+	struct btrfs_inode *inode;
+	struct inode *vfs_inode;
 
 	inode = read_one_inode(root, objectid);
 	if (!inode)
 		return -EIO;
 
+	vfs_inode = &inode->vfs_inode;
 	key.objectid = BTRFS_TREE_LOG_FIXUP_OBJECTID;
 	key.type = BTRFS_ORPHAN_ITEM_KEY;
 	key.offset = objectid;
···
 
 	btrfs_release_path(path);
 	if (ret == 0) {
-		if (!inode->i_nlink)
-			set_nlink(inode, 1);
+		if (!vfs_inode->i_nlink)
+			set_nlink(vfs_inode, 1);
 		else
-			inc_nlink(inode);
-		ret = btrfs_update_inode(trans, BTRFS_I(inode));
+			inc_nlink(vfs_inode);
+		ret = btrfs_update_inode(trans, inode);
 	} else if (ret == -EEXIST) {
 		ret = 0;
 	}
-	iput(inode);
+	iput(vfs_inode);
 
 	return ret;
 }
···
 				    const struct fscrypt_str *name,
 				    struct btrfs_key *location)
 {
-	struct inode *inode;
-	struct inode *dir;
+	struct btrfs_inode *inode;
+	struct btrfs_inode *dir;
 	int ret;
 
 	inode = read_one_inode(root, location->objectid);
···
 
 	dir = read_one_inode(root, dirid);
 	if (!dir) {
-		iput(inode);
+		iput(&inode->vfs_inode);
 		return -EIO;
 	}
 
-	ret = btrfs_add_link(trans, BTRFS_I(dir), BTRFS_I(inode), name,
-			     1, index);
+	ret = btrfs_add_link(trans, dir, inode, name, 1, index);
 
 	/* FIXME, put inode into FIXUP list */
 
-	iput(inode);
-	iput(dir);
+	iput(&inode->vfs_inode);
+	iput(&dir->vfs_inode);
 	return ret;
 }
 
···
 	bool index_dst_matches = false;
 	struct btrfs_key log_key;
 	struct btrfs_key search_key;
-	struct inode *dir;
+	struct btrfs_inode *dir;
 	u8 log_flags;
 	bool exists;
 	int ret;
···
 		ret = PTR_ERR(dir_dst_di);
 		goto out;
 	} else if (dir_dst_di) {
-		ret = delete_conflicting_dir_entry(trans, BTRFS_I(dir), path,
-						   dir_dst_di, &log_key,
-						   log_flags, exists);
+		ret = delete_conflicting_dir_entry(trans, dir, path, dir_dst_di,
+						   &log_key, log_flags, exists);
 		if (ret < 0)
 			goto out;
 		dir_dst_matches = (ret == 1);
···
 		ret = PTR_ERR(index_dst_di);
 		goto out;
 	} else if (index_dst_di) {
-		ret = delete_conflicting_dir_entry(trans, BTRFS_I(dir), path,
-						   index_dst_di, &log_key,
-						   log_flags, exists);
+		ret = delete_conflicting_dir_entry(trans, dir, path, index_dst_di,
+						   &log_key, log_flags, exists);
 		if (ret < 0)
 			goto out;
 		index_dst_matches = (ret == 1);
···
 
 out:
 	if (!ret && update_size) {
-		btrfs_i_size_write(BTRFS_I(dir), dir->i_size + name.len * 2);
-		ret = btrfs_update_inode(trans, BTRFS_I(dir));
+		btrfs_i_size_write(dir, dir->vfs_inode.i_size + name.len * 2);
+		ret = btrfs_update_inode(trans, dir);
 	}
 	kfree(name.name);
-	iput(dir);
+	iput(&dir->vfs_inode);
 	if (!ret && name_added)
 		ret = 1;
 	return ret;
···
 				      struct btrfs_root *log,
 				      struct btrfs_path *path,
 				      struct btrfs_path *log_path,
-				      struct inode *dir,
+				      struct btrfs_inode *dir,
 				      struct btrfs_key *dir_key)
 {
-	struct btrfs_root *root = BTRFS_I(dir)->root;
+	struct btrfs_root *root = dir->root;
 	int ret;
 	struct extent_buffer *eb;
 	int slot;
 	struct btrfs_dir_item *di;
 	struct fscrypt_str name = { 0 };
-	struct inode *inode = NULL;
+	struct btrfs_inode *inode = NULL;
 	struct btrfs_key location;
 
 	/*
···
 	if (ret)
 		goto out;
 
-	inc_nlink(inode);
-	ret = unlink_inode_for_log_replay(trans, BTRFS_I(dir), BTRFS_I(inode),
-					  &name);
+	inc_nlink(&inode->vfs_inode);
+	ret = unlink_inode_for_log_replay(trans, dir, inode, &name);
 	/*
 	 * Unlike dir item keys, dir index keys can only have one name (entry) in
 	 * them, as there are no key collisions since each key has a unique offset
···
 	btrfs_release_path(path);
 	btrfs_release_path(log_path);
 	kfree(name.name);
-	iput(inode);
+	if (inode)
+		iput(&inode->vfs_inode);
 	return ret;
 }
 
···
 	struct btrfs_key dir_key;
 	struct btrfs_key found_key;
 	struct btrfs_path *log_path;
-	struct inode *dir;
+	struct btrfs_inode *dir;
 
 	dir_key.objectid = dirid;
 	dir_key.type = BTRFS_DIR_INDEX_KEY;
···
 out:
 	btrfs_release_path(path);
 	btrfs_free_path(log_path);
-	iput(dir);
+	iput(&dir->vfs_inode);
 	return ret;
 }
 
···
 			 */
 			if (S_ISREG(mode)) {
 				struct btrfs_drop_extents_args drop_args = { 0 };
-				struct inode *inode;
+				struct btrfs_inode *inode;
 				u64 from;
 
 				inode = read_one_inode(root, key.objectid);
···
 					ret = -EIO;
 					break;
 				}
-				from = ALIGN(i_size_read(inode),
+				from = ALIGN(i_size_read(&inode->vfs_inode),
 					     root->fs_info->sectorsize);
 				drop_args.start = from;
 				drop_args.end = (u64)-1;
 				drop_args.drop_cache = true;
-				ret = btrfs_drop_extents(wc->trans, root,
-							 BTRFS_I(inode),
+				ret = btrfs_drop_extents(wc->trans, root, inode,
 							 &drop_args);
 				if (!ret) {
-					inode_sub_bytes(inode,
+					inode_sub_bytes(&inode->vfs_inode,
 							drop_args.bytes_found);
 					/* Update the inode's nbytes. */
-					ret = btrfs_update_inode(wc->trans,
-								 BTRFS_I(inode));
+					ret = btrfs_update_inode(wc->trans, inode);
 				}
-				iput(inode);
+				iput(&inode->vfs_inode);
 				if (ret)
 					break;
 			}
···
 	struct btrfs_dir_log_item *item;
 
 	key.objectid = dirid;
-	key.offset = first_offset;
 	key.type = BTRFS_DIR_LOG_INDEX_KEY;
+	key.offset = first_offset;
 	ret = btrfs_insert_empty_item(trans, log, path, &key, sizeof(*item));
 	/*
 	 * -EEXIST is fine and can happen sporadically when we are logging a
···
 	ihold(&curr_inode->vfs_inode);
 
 	while (true) {
-		struct inode *vfs_inode;
 		struct btrfs_key key;
 		struct btrfs_key found_key;
 		u64 next_index;
···
 			struct extent_buffer *leaf = path->nodes[0];
 			struct btrfs_dir_item *di;
 			struct btrfs_key di_key;
-			struct inode *di_inode;
+			struct btrfs_inode *di_inode;
 			int log_mode = LOG_INODE_EXISTS;
 			int type;
 
···
 				goto out;
 			}
 
-			if (!need_log_inode(trans, BTRFS_I(di_inode))) {
-				btrfs_add_delayed_iput(BTRFS_I(di_inode));
+			if (!need_log_inode(trans, di_inode)) {
+				btrfs_add_delayed_iput(di_inode);
 				break;
 			}
 
 			ctx->log_new_dentries = false;
 			if (type == BTRFS_FT_DIR)
 				log_mode = LOG_INODE_ALL;
-			ret = btrfs_log_inode(trans, BTRFS_I(di_inode),
-					      log_mode, ctx);
-			btrfs_add_delayed_iput(BTRFS_I(di_inode));
+			ret = btrfs_log_inode(trans, di_inode, log_mode, ctx);
+			btrfs_add_delayed_iput(di_inode);
 			if (ret)
 				goto out;
 			if (ctx->log_new_dentries) {
···
 		kfree(dir_elem);
 
 		btrfs_add_delayed_iput(curr_inode);
-		curr_inode = NULL;
 
-		vfs_inode = btrfs_iget_logging(ino, root);
-		if (IS_ERR(vfs_inode)) {
-			ret = PTR_ERR(vfs_inode);
+		curr_inode = btrfs_iget_logging(ino, root);
+		if (IS_ERR(curr_inode)) {
+			ret = PTR_ERR(curr_inode);
+			curr_inode = NULL;
 			break;
 		}
-		curr_inode = BTRFS_I(vfs_inode);
 	}
 out:
 	btrfs_free_path(path);
···
 				 struct btrfs_log_ctx *ctx)
 {
 	struct btrfs_ino_list *ino_elem;
-	struct inode *inode;
+	struct btrfs_inode *inode;
 
 	/*
 	 * It's rare to have a lot of conflicting inodes, in practice it is not
···
 	 * inode in LOG_INODE_EXISTS mode and rename operations update the log,
 	 * so that the log ends up with the new name and without the old name.
 	 */
-	if (!need_log_inode(trans, BTRFS_I(inode))) {
-		btrfs_add_delayed_iput(BTRFS_I(inode));
+	if (!need_log_inode(trans, inode)) {
+		btrfs_add_delayed_iput(inode);
 		return 0;
 	}
 
-	btrfs_add_delayed_iput(BTRFS_I(inode));
+	btrfs_add_delayed_iput(inode);
 
 	ino_elem = kmalloc(sizeof(*ino_elem), GFP_NOFS);
 	if (!ino_elem)
···
 	 */
 	while (!list_empty(&ctx->conflict_inodes)) {
 		struct btrfs_ino_list *curr;
-		struct inode *inode;
+		struct btrfs_inode *inode;
 		u64 ino;
 		u64 parent;
 
···
 			 * dir index key range logged for the directory. So we
 			 * must make sure the deletion is recorded.
 			 */
-			ret = btrfs_log_inode(trans, BTRFS_I(inode),
-					      LOG_INODE_ALL, ctx);
-			btrfs_add_delayed_iput(BTRFS_I(inode));
+			ret = btrfs_log_inode(trans, inode, LOG_INODE_ALL, ctx);
+			btrfs_add_delayed_iput(inode);
 			if (ret)
 				break;
 			continue;
···
 		 * it again because if some other task logged the inode after
 		 * that, we can avoid doing it again.
 		 */
-		if (!need_log_inode(trans, BTRFS_I(inode))) {
-			btrfs_add_delayed_iput(BTRFS_I(inode));
+		if (!need_log_inode(trans, inode)) {
+			btrfs_add_delayed_iput(inode);
 			continue;
 		}
 
···
 		 * well because during a rename we pin the log and update the
 		 * log with the new name before we unpin it.
 		 */
-		ret = btrfs_log_inode(trans, BTRFS_I(inode), LOG_INODE_EXISTS, ctx);
-		btrfs_add_delayed_iput(BTRFS_I(inode));
+		ret = btrfs_log_inode(trans, inode, LOG_INODE_EXISTS, ctx);
+		btrfs_add_delayed_iput(inode);
 		if (ret)
 			break;
 	}
···
 
 	list_for_each_entry(item, delayed_ins_list, log_list) {
 		struct btrfs_dir_item *dir_item;
-		struct inode *di_inode;
+		struct btrfs_inode *di_inode;
 		struct btrfs_key key;
 		int log_mode = LOG_INODE_EXISTS;
 
···
 			break;
 		}
 
-		if (!need_log_inode(trans, BTRFS_I(di_inode))) {
-			btrfs_add_delayed_iput(BTRFS_I(di_inode));
+		if (!need_log_inode(trans, di_inode)) {
+			btrfs_add_delayed_iput(di_inode);
 			continue;
 		}
 
···
 			log_mode = LOG_INODE_ALL;
 
 		ctx->log_new_dentries = false;
-		ret = btrfs_log_inode(trans, BTRFS_I(di_inode), log_mode, ctx);
+		ret = btrfs_log_inode(trans, di_inode, log_mode, ctx);
 
 		if (!ret && ctx->log_new_dentries)
-			ret = log_new_dir_dentries(trans, BTRFS_I(di_inode), ctx);
+			ret = log_new_dir_dentries(trans, di_inode, ctx);
 
-		btrfs_add_delayed_iput(BTRFS_I(di_inode));
+		btrfs_add_delayed_iput(di_inode);
 
 		if (ret)
 			break;
···
 		ptr = btrfs_item_ptr_offset(leaf, slot);
 		while (cur_offset < item_size) {
 			struct btrfs_key inode_key;
-			struct inode *dir_inode;
+			struct btrfs_inode *dir_inode;
 
 			inode_key.type = BTRFS_INODE_ITEM_KEY;
 			inode_key.offset = 0;
···
 				goto out;
 			}
 
-			if (!need_log_inode(trans, BTRFS_I(dir_inode))) {
-				btrfs_add_delayed_iput(BTRFS_I(dir_inode));
+			if (!need_log_inode(trans, dir_inode)) {
+				btrfs_add_delayed_iput(dir_inode);
 				continue;
 			}
 
 			ctx->log_new_dentries = false;
-			ret = btrfs_log_inode(trans, BTRFS_I(dir_inode),
-					      LOG_INODE_ALL, ctx);
+			ret = btrfs_log_inode(trans, dir_inode, LOG_INODE_ALL, ctx);
 			if (!ret && ctx->log_new_dentries)
-				ret = log_new_dir_dentries(trans,
-						   BTRFS_I(dir_inode), ctx);
-			btrfs_add_delayed_iput(BTRFS_I(dir_inode));
+				ret = log_new_dir_dentries(trans, dir_inode, ctx);
+			btrfs_add_delayed_iput(dir_inode);
 			if (ret)
 				goto out;
 		}
···
 		struct extent_buffer *leaf;
 		int slot;
 		struct btrfs_key search_key;
-		struct inode *inode;
+		struct btrfs_inode *inode;
 		u64 ino;
 		int ret = 0;
 
···
 		if (IS_ERR(inode))
 			return PTR_ERR(inode);
 
-		if (BTRFS_I(inode)->generation >= trans->transid &&
-		    need_log_inode(trans, BTRFS_I(inode)))
-			ret = btrfs_log_inode(trans, BTRFS_I(inode),
-					      LOG_INODE_EXISTS, ctx);
-		btrfs_add_delayed_iput(BTRFS_I(inode));
+		if (inode->generation >= trans->transid &&
+		    need_log_inode(trans, inode))
+			ret = btrfs_log_inode(trans, inode, LOG_INODE_EXISTS, ctx);
+		btrfs_add_delayed_iput(inode);
 		if (ret)
 			return ret;
 
···
 	struct btrfs_root *root = inode->root;
 	struct btrfs_fs_info *fs_info = root->fs_info;
 	int ret = 0;
-	bool log_dentries = false;
+	bool log_dentries;
 
-	if (btrfs_test_opt(fs_info, NOTREELOG)) {
-		ret = BTRFS_LOG_FORCE_COMMIT;
-		goto end_no_trans;
-	}
+	if (btrfs_test_opt(fs_info, NOTREELOG))
+		return BTRFS_LOG_FORCE_COMMIT;
 
-	if (btrfs_root_refs(&root->root_item) == 0) {
-		ret = BTRFS_LOG_FORCE_COMMIT;
-		goto end_no_trans;
-	}
+	if (btrfs_root_refs(&root->root_item) == 0)
+		return BTRFS_LOG_FORCE_COMMIT;
 
 	/*
 	 * If we're logging an inode from a subvolume created in the current
 	 * transaction we must force a commit since the root is not persisted.
 	 */
-	if (btrfs_root_generation(&root->root_item) == trans->transid) {
-		ret = BTRFS_LOG_FORCE_COMMIT;
-		goto end_no_trans;
-	}
+	if (btrfs_root_generation(&root->root_item) == trans->transid)
+		return BTRFS_LOG_FORCE_COMMIT;
 
 	/*
 	 * Skip already logged inodes or inodes corresponding to tmpfiles
···
 	 */
 	if ((btrfs_inode_in_log(inode, trans->transid) &&
 	     list_empty(&ctx->ordered_extents)) ||
-	    inode->vfs_inode.i_nlink == 0) {
-		ret = BTRFS_NO_LOG_SYNC;
-		goto end_no_trans;
-	}
+	    inode->vfs_inode.i_nlink == 0)
+		return BTRFS_NO_LOG_SYNC;
 
 	ret = start_log_trans(trans, root, ctx);
 	if (ret)
-		goto end_no_trans;
+		return ret;
 
 	ret = btrfs_log_inode(trans, inode, inode_only, ctx);
 	if (ret)
···
 		goto end_trans;
 	}
 
-	if (S_ISDIR(inode->vfs_inode.i_mode) && ctx->log_new_dentries)
-		log_dentries = true;
+	/*
+	 * Track if we need to log dentries because ctx->log_new_dentries can
+	 * be modified in the call chains below.
+	 */
+	log_dentries = ctx->log_new_dentries;
 
 	/*
 	 * On unlink we must make sure all our current and old parent directory
···
 
 	if (log_dentries)
 		ret = log_new_dir_dentries(trans, inode, ctx);
-	else
-		ret = 0;
 end_trans:
 	if (ret < 0) {
 		btrfs_set_log_full_commit(trans);
···
 	if (ret)
 		btrfs_remove_log_ctx(root, ctx);
 	btrfs_end_log_trans(root);
-end_no_trans:
+
 	return ret;
 }
 
···
 
 again:
 	key.objectid = BTRFS_TREE_LOG_OBJECTID;
-	key.offset = (u64)-1;
 	key.type = BTRFS_ROOT_ITEM_KEY;
+	key.offset = (u64)-1;
 
 	while (1) {
 		ret = btrfs_search_slot(NULL, log_root_tree, &key, path, 0, 0);

fs/btrfs/verity.c

···
 		goto out;
 	}
 	inode->ro_flags &= ~BTRFS_INODE_RO_VERITY;
-	btrfs_sync_inode_flags_to_i_flags(&inode->vfs_inode);
+	btrfs_sync_inode_flags_to_i_flags(inode);
 	ret = btrfs_update_inode(trans, inode);
 	if (ret) {
 		btrfs_abort_transaction(trans, ret);
···
 		goto out;
 	}
 	inode->ro_flags |= BTRFS_INODE_RO_VERITY;
-	btrfs_sync_inode_flags_to_i_flags(&inode->vfs_inode);
+	btrfs_sync_inode_flags_to_i_flags(inode);
 	ret = btrfs_update_inode(trans, inode);
 	if (ret)
 		goto end_trans;

fs/btrfs/volumes.c

···
 	path->skip_locking = 1;
 
 	key.objectid = device->devid;
-	key.offset = search_start;
 	key.type = BTRFS_DEV_EXTENT_KEY;
+	key.offset = search_start;
 
 	ret = btrfs_search_backwards(root, &key, path);
 	if (ret < 0)
···
 		return -ENOMEM;
 
 	key.objectid = device->devid;
-	key.offset = start;
 	key.type = BTRFS_DEV_EXTENT_KEY;
+	key.offset = start;
 again:
 	ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
 	if (ret > 0) {
···
 		return -ENOMEM;
 
 	key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
-	key.offset = 0;
 	key.type = BTRFS_DEV_ITEM_KEY;
+	key.offset = 0;
 
 	while (1) {
 		btrfs_reserve_chunk_metadata(trans, false);
···
 		return -ENOMEM;
 
 	key.objectid = BTRFS_FIRST_CHUNK_TREE_OBJECTID;
-	key.offset = chunk_offset;
 	key.type = BTRFS_CHUNK_ITEM_KEY;
+	key.offset = chunk_offset;
 
 	ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
 	if (ret < 0)
···
 
 again:
 	key.objectid = BTRFS_FIRST_CHUNK_TREE_OBJECTID;
-	key.offset = (u64)-1;
 	key.type = BTRFS_CHUNK_ITEM_KEY;
+	key.offset = (u64)-1;
 
 	while (1) {
 		mutex_lock(&fs_info->reclaim_bgs_lock);
···
 		bctl->sys.limit = limit_sys;
 	}
 	key.objectid = BTRFS_FIRST_CHUNK_TREE_OBJECTID;
-	key.offset = (u64)-1;
 	key.type = BTRFS_CHUNK_ITEM_KEY;
+	key.offset = (u64)-1;
 
 	while (1) {
 		if ((!counting && atomic_read(&fs_info->balance_pause_req)) ||
···
 
 again:
 	key.objectid = device->devid;
-	key.offset = (u64)-1;
 	key.type = BTRFS_DEV_EXTENT_KEY;
+	key.offset = (u64)-1;
 
 	do {
 		mutex_lock(&fs_info->reclaim_bgs_lock);
···
 	 * item - BTRFS_FIRST_CHUNK_TREE_OBJECTID).
 	 */
 	key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
-	key.offset = 0;
 	key.type = 0;
+	key.offset = 0;
 	btrfs_for_each_slot(root, &key, &found_key, path, iter_ret) {
 		struct extent_buffer *node = path->nodes[1];
 

fs/btrfs/volumes.h

···
 #define BTRFS_VOLUMES_H
 
 #include <linux/blk_types.h>
+#include <linux/blkdev.h>
 #include <linux/sizes.h>
 #include <linux/atomic.h>
 #include <linux/sort.h>
···
 #include <linux/completion.h>
 #include <linux/rbtree.h>
 #include <uapi/linux/btrfs.h>
+#include <uapi/linux/btrfs_tree.h>
 #include "messages.h"
 #include "rcu-string.h"
+#include "extent-io-tree.h"
 
 struct block_device;
 struct bdev_handle;
 struct btrfs_fs_info;
 struct btrfs_block_group;
 struct btrfs_trans_handle;
+struct btrfs_transaction;
 struct btrfs_zoned_device_info;
 
 #define BTRFS_MAX_DATA_CHUNK_SIZE	(10ULL * SZ_1G)

fs/btrfs/xattr.h

···
 #ifndef BTRFS_XATTR_H
 #define BTRFS_XATTR_H
 
+#include <linux/types.h>
+
 struct dentry;
 struct inode;
 struct qstr;

fs/btrfs/zlib.c

···
 	return ERR_PTR(-ENOMEM);
 }
 
+/*
+ * Helper for S390x with hardware zlib compression support.
+ *
+ * That hardware acceleration requires a buffer size larger than a single page
+ * to get ideal performance, thus we need to do the memory copy rather than
+ * use the page cache directly as input buffer.
+ */
+static int copy_data_into_buffer(struct address_space *mapping,
+				 struct workspace *workspace, u64 filepos,
+				 unsigned long length)
+{
+	u64 cur = filepos;
+
+	/* It's only for hardware accelerated zlib code. */
+	ASSERT(zlib_deflate_dfltcc_enabled());
+
+	while (cur < filepos + length) {
+		struct folio *folio;
+		void *data_in;
+		unsigned int offset;
+		unsigned long copy_length;
+		int ret;
+
+		ret = btrfs_compress_filemap_get_folio(mapping, cur, &folio);
+		if (ret < 0)
+			return ret;
+		/* No large folio support yet. */
+		ASSERT(!folio_test_large(folio));
+
+		offset = offset_in_folio(folio, cur);
+		copy_length = min(folio_size(folio) - offset,
+				  filepos + length - cur);
+
+		data_in = kmap_local_folio(folio, offset);
+		memcpy(workspace->buf + cur - filepos, data_in, copy_length);
+		kunmap_local(data_in);
+		cur += copy_length;
+	}
+	return 0;
+}
+
 int zlib_compress_folios(struct list_head *ws, struct address_space *mapping,
 			 u64 start, struct folio **folios, unsigned long *out_folios,
 			 unsigned long *total_in, unsigned long *total_out)
···
 	int nr_folios = 0;
 	struct folio *in_folio = NULL;
 	struct folio *out_folio = NULL;
-	unsigned long bytes_left;
-	unsigned int in_buf_folios;
 	unsigned long len = *total_out;
 	unsigned long nr_dest_folios = *out_folios;
 	const unsigned long max_out = nr_dest_folios * PAGE_SIZE;
···
 		 * the workspace buffer if required.
 		 */
 		if (workspace->strm.avail_in == 0) {
-			bytes_left = len - workspace->strm.total_in;
-			in_buf_folios = min(DIV_ROUND_UP(bytes_left, PAGE_SIZE),
-					    workspace->buf_size / PAGE_SIZE);
-			if (in_buf_folios > 1) {
-				int i;
+			unsigned long bytes_left = len - workspace->strm.total_in;
+			unsigned int copy_length = min(bytes_left, workspace->buf_size);
 
-				/* S390 hardware acceleration path, not subpage. */
-				ASSERT(!btrfs_is_subpage(
-						inode_to_fs_info(mapping->host),
-						mapping));
-				for (i = 0; i < in_buf_folios; i++) {
-					if (data_in) {
-						kunmap_local(data_in);
-						folio_put(in_folio);
-						data_in = NULL;
-					}
-					ret = btrfs_compress_filemap_get_folio(mapping,
-							start, &in_folio);
-					if (ret < 0)
-						goto out;
-					data_in = kmap_local_folio(in_folio, 0);
-					copy_page(workspace->buf + i * PAGE_SIZE,
-						  data_in);
-					start += PAGE_SIZE;
-				}
+			/*
+			 * This can only happen when hardware zlib compression is
+			 * enabled.
+			 */
+			if (copy_length > PAGE_SIZE) {
+				ret = copy_data_into_buffer(mapping, workspace,
+							    start, copy_length);
+				if (ret < 0)
+					goto out;
+				start += copy_length;
 				workspace->strm.next_in = workspace->buf;
-				workspace->strm.avail_in = min(bytes_left,
-							       in_buf_folios << PAGE_SHIFT);
+				workspace->strm.avail_in = copy_length;
 			} else {
 				unsigned int pg_off;
 				unsigned int cur_len;
···
 
 const struct btrfs_compress_op btrfs_zlib_compress = {
 	.workspace_manager	= &wsm,
+	.min_level		= 1,
 	.max_level		= 9,
 	.default_level		= BTRFS_ZLIB_DEFAULT_LEVEL,
 };

fs/btrfs/zoned.c

···
 		physical = map->stripes[i].physical;
 		zinfo = device->zone_info;
 
+		if (!device->bdev)
+			continue;
+
 		if (zinfo->max_active_zones == 0)
 			continue;
 
···
 		struct btrfs_zoned_device_info *zinfo = device->zone_info;
 		unsigned int nofs_flags;
 
+		if (!device->bdev)
+			continue;
+
 		if (zinfo->max_active_zones == 0)
 			continue;
 
···
 
 	if (!btrfs_is_zoned(fs_info))
 		return true;
+
+	if (test_bit(BTRFS_FS_NEED_ZONE_FINISH, &fs_info->flags))
+		return false;
 
 	/* Check if there is a device with active zones left */
 	mutex_lock(&fs_info->chunk_mutex);

fs/btrfs/zstd.c

···
 #define ZSTD_BTRFS_MAX_WINDOWLOG 17
 #define ZSTD_BTRFS_MAX_INPUT (1 << ZSTD_BTRFS_MAX_WINDOWLOG)
 #define ZSTD_BTRFS_DEFAULT_LEVEL 3
+#define ZSTD_BTRFS_MIN_LEVEL -15
 #define ZSTD_BTRFS_MAX_LEVEL 15
 /* 307s to avoid pathologically clashing with transaction commit */
 #define ZSTD_BTRFS_RECLAIM_JIFFIES (307 * HZ)
 
-static zstd_parameters zstd_get_btrfs_parameters(unsigned int level,
+static zstd_parameters zstd_get_btrfs_parameters(int level,
 						 size_t src_len)
 {
 	zstd_parameters params = zstd_get_params(level, src_len);
···
 	void *mem;
 	size_t size;
 	char *buf;
-	unsigned int level;
-	unsigned int req_level;
+	int level;
+	int req_level;
 	unsigned long last_used; /* jiffies */
 	struct list_head list;
 	struct list_head lru_list;
 	zstd_in_buffer in_buf;
 	zstd_out_buffer out_buf;
+	zstd_parameters params;
 };
 
 /*
···
 	return container_of(list, struct workspace, list);
 }
 
-void zstd_free_workspace(struct list_head *ws);
-struct list_head *zstd_alloc_workspace(unsigned int level);
+static inline int clip_level(int level)
+{
+	return max(0, level - 1);
+}
 
 /*
  * Timer callback to free unused workspaces.
···
 	list_for_each_prev_safe(pos, next, &wsm.lru_list) {
 		struct workspace *victim = container_of(pos, struct workspace,
 							lru_list);
-		unsigned int level;
+		int level;
 
 		if (time_after(victim->last_used, reclaim_threshold))
 			break;
···
 		list_del(&victim->list);
 		zstd_free_workspace(&victim->list);
 
-		if (list_empty(&wsm.idle_ws[level - 1]))
-			clear_bit(level - 1, &wsm.active_map);
+		if (list_empty(&wsm.idle_ws[level]))
+			clear_bit(level, &wsm.active_map);
 
 	}
 
···
 static void zstd_calc_ws_mem_sizes(void)
 {
 	size_t max_size = 0;
-	unsigned int level;
+	int level;
 
-	for (level = 1; level <= ZSTD_BTRFS_MAX_LEVEL; level++) {
+	for (level = ZSTD_BTRFS_MIN_LEVEL; level <= ZSTD_BTRFS_MAX_LEVEL; level++) {
+		if (level == 0)
+			continue;
 		zstd_parameters params =
 			zstd_get_btrfs_parameters(level, ZSTD_BTRFS_MAX_INPUT);
 		size_t level_size =
···
 			      zstd_dstream_workspace_bound(ZSTD_BTRFS_MAX_INPUT));
 
 		max_size = max_t(size_t, max_size, level_size);
-		zstd_ws_mem_sizes[level - 1] = max_size;
+		/* Use level 1 workspace size for all the fast mode negative levels. */
+		zstd_ws_mem_sizes[clip_level(level)] = max_size;
 	}
 }
 
···
  * offer the opportunity to reclaim the workspace in favor of allocating an
  * appropriately sized one in the future.
  */
-static struct list_head *zstd_find_workspace(unsigned int level)
+static struct list_head *zstd_find_workspace(int level)
 {
 	struct list_head *ws;
 	struct workspace *workspace;
-	int i = level - 1;
+	int i = clip_level(level);
 
 	spin_lock_bh(&wsm.lock);
 	for_each_set_bit_from(i, &wsm.active_map, ZSTD_BTRFS_MAX_LEVEL) {
···
 			list_del_init(ws);
 			/* keep its place if it's a lower level using this */
 			workspace->req_level = level;
-			if (level == workspace->level)
+			if (clip_level(level) == workspace->level)
 				list_del(&workspace->lru_list);
 			if (list_empty(&wsm.idle_ws[i]))
 				clear_bit(i, &wsm.active_map);
···
  * attempt to allocate a new workspace.  If we fail to allocate one due to
  * memory pressure, go to sleep waiting for the max level workspace to free up.
  */
-struct list_head *zstd_get_workspace(unsigned int level)
+struct list_head *zstd_get_workspace(int level)
 {
 	struct list_head *ws;
 	unsigned int nofs_flag;
···
 	spin_lock_bh(&wsm.lock);
 
 	/* A node is only taken off the lru if we are the corresponding level */
-	if (workspace->req_level == workspace->level) {
+	if (clip_level(workspace->req_level) == workspace->level) {
 		/* Hide a max level workspace from reclaim */
 		if (list_empty(&wsm.idle_ws[ZSTD_BTRFS_MAX_LEVEL - 1])) {
 			INIT_LIST_HEAD(&workspace->lru_list);
···
 		}
 	}
 
-	set_bit(workspace->level - 1, &wsm.active_map);
-	list_add(&workspace->list, &wsm.idle_ws[workspace->level - 1]);
+	set_bit(workspace->level, &wsm.active_map);
+	list_add(&workspace->list, &wsm.idle_ws[workspace->level]);
 	workspace->req_level = 0;
 
 	spin_unlock_bh(&wsm.lock);
 
-	if (workspace->level == ZSTD_BTRFS_MAX_LEVEL)
+	if (workspace->level == clip_level(ZSTD_BTRFS_MAX_LEVEL))
 		cond_wake_up(&wsm.wait);
 }
 
···
 	kfree(workspace);
 }
 
-struct list_head *zstd_alloc_workspace(unsigned int level)
+struct list_head *zstd_alloc_workspace(int level)
 {
 	struct workspace *workspace;
 
···
 	if (!workspace)
 		return ERR_PTR(-ENOMEM);
 
-	workspace->size = zstd_ws_mem_sizes[level - 1];
-	workspace->level = level;
+	/* Use level 1 workspace size for all the fast mode negative levels. */
+	workspace->size = zstd_ws_mem_sizes[clip_level(level)];
+	workspace->level = clip_level(level);
 	workspace->req_level = level;
 	workspace->last_used = jiffies;
 	workspace->mem = kvmalloc(workspace->size, GFP_KERNEL | __GFP_NOWARN);
···
 	const unsigned long nr_dest_folios = *out_folios;
 	const u64 orig_end = start + len;
 	unsigned long max_out = nr_dest_folios * PAGE_SIZE;
-	unsigned int pg_off;
 	unsigned int cur_len;
-	zstd_parameters params = zstd_get_btrfs_parameters(workspace->req_level,
-							   len);
 
+	workspace->params = zstd_get_btrfs_parameters(workspace->req_level, len);
 	*out_folios = 0;
 	*total_out = 0;
 	*total_in = 0;
 
 	/* Initialize the stream */
-	stream = zstd_init_cstream(&params, len, workspace->mem,
+	stream = zstd_init_cstream(&workspace->params, len, workspace->mem,
 			workspace->size);
 	if (unlikely(!stream)) {
 		struct btrfs_inode *inode = BTRFS_I(mapping->host);
···
 	ret = btrfs_compress_filemap_get_folio(mapping, start, &in_folio);
 	if (ret < 0)
 		goto out;
-	pg_off = offset_in_page(start);
 	cur_len = btrfs_calc_input_length(orig_end, start);
-	workspace->in_buf.src = kmap_local_folio(in_folio, pg_off);
+	workspace->in_buf.src = kmap_local_folio(in_folio, offset_in_page(start));
 	workspace->in_buf.pos = 0;
 	workspace->in_buf.size = cur_len;
 
···
 			ret = btrfs_compress_filemap_get_folio(mapping, start, &in_folio);
 			if (ret < 0)
 				goto out;
-			pg_off = offset_in_page(start);
 			cur_len = btrfs_calc_input_length(orig_end, start);
-			workspace->in_buf.src = kmap_local_folio(in_folio, pg_off);
+			workspace->in_buf.src = kmap_local_folio(in_folio,
+							 offset_in_page(start));
 			workspace->in_buf.pos = 0;
 			workspace->in_buf.size = cur_len;
 		}
···
 const struct btrfs_compress_op btrfs_zstd_compress = {
 	/* ZSTD uses own workspace manager */
 	.workspace_manager = NULL,
+	.min_level	= ZSTD_BTRFS_MIN_LEVEL,
 	.max_level	= ZSTD_BTRFS_MAX_LEVEL,
 	.default_level	= ZSTD_BTRFS_DEFAULT_LEVEL,
 };

include/uapi/linux/btrfs.h

···
  */
 #define BTRFS_DEFRAG_RANGE_COMPRESS 1
 #define BTRFS_DEFRAG_RANGE_START_IO 2
+#define BTRFS_DEFRAG_RANGE_COMPRESS_LEVEL 4
 #define BTRFS_DEFRAG_RANGE_FLAGS_SUPP	(BTRFS_DEFRAG_RANGE_COMPRESS |		\
+					 BTRFS_DEFRAG_RANGE_COMPRESS_LEVEL |	\
 					 BTRFS_DEFRAG_RANGE_START_IO)
 
 struct btrfs_ioctl_defrag_range_args {
···
 
 	/*
 	 * which compression method to use if turning on compression
-	 * for this defrag operation.  If unspecified, zlib will
-	 * be used
+	 * for this defrag operation. If unspecified, zlib will be
+	 * used. If compression level is also being specified, set the
+	 * BTRFS_DEFRAG_RANGE_COMPRESS_LEVEL flag and fill the compress
+	 * member structure instead of the compress_type field.
 	 */
-	__u32 compress_type;
+	union {
+		__u32 compress_type;
+		struct {
+			__u8 type;
+			__s8 level;
+		} compress;
+	};
 
 	/* spare for later */
 	__u32 unused[4];