Merge tag 'for-6.6-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux

Pull btrfs updates from David Sterba:
 "No new features, the bulk of the changes are fixes, refactoring and
  cleanups. The notable fix is the scrub performance restoration after
  rewrite in 6.4, though still only partial.

  Fixes:

   - scrub performance drop due to rewrite in 6.4 partially restored:
      - do IO grouping by blg_plug/blk_unplug again
      - avoid unnecessary tree searches when processing stripes, in
        extent and checksum trees
      - the drop is noticeable on fast PCIe devices, -66% and restored
        to -33% of the original
      - backports to 6.4 planned

   - handle more corner cases of transaction commit during orphan
     cleanup or delayed ref processing

   - use correct fsid/metadata_uuid when validating super block

   - copy directory permissions and time when creating a stub subvolume

  Core:

   - debugging feature integrity checker deprecated, to be removed in
     6.7

   - in zoned mode, zones are activated just before the write, making
     error handling easier, now the overcommit mechanism can be enabled
     again which improves performance by avoiding more frequent flushing

   - v0 extent handling completely removed, deprecated long time ago

   - error handling improvements

   - tests:
      - extent buffer bitmap tests
      - pinned extent splitting tests

   - cleanups and refactoring:
      - compression writeback
      - extent buffer bitmap
      - space flushing, ENOSPC handling"

* tag 'for-6.6-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux: (110 commits)
  btrfs: zoned: skip splitting and logical rewriting on pre-alloc write
  btrfs: tests: test invalid splitting when skipping pinned drop extent_map
  btrfs: tests: add a test for btrfs_add_extent_mapping
  btrfs: tests: add extent_map tests for dropping with odd layouts
  btrfs: scrub: move write back of repaired sectors to scrub_stripe_read_repair_worker()
  btrfs: scrub: don't go ordered workqueue for dev-replace
  btrfs: scrub: fix grouping of read IO
  btrfs: scrub: avoid unnecessary csum tree search preparing stripes
  btrfs: scrub: avoid unnecessary extent tree search preparing stripes
  btrfs: copy dir permission and time when creating a stub subvolume
  btrfs: remove pointless empty list check when reading delayed dir indexes
  btrfs: drop redundant check to use fs_devices::metadata_uuid
  btrfs: compare the correct fsid/metadata_uuid in btrfs_validate_super
  btrfs: use the correct superblock to compare fsid in btrfs_validate_super
  btrfs: simplify memcpy either of metadata_uuid or fsid
  btrfs: add a helper to read the superblock metadata_uuid
  btrfs: remove v0 extent handling
  btrfs: output extra debug info if we failed to find an inline backref
  btrfs: move the !zoned assert into run_delalloc_cow
  btrfs: consolidate the error handling in run_delalloc_nocow
  ...

44 files changed, 2259 insertions, 1617 deletions

diff --git a/fs/btrfs/Kconfig b/fs/btrfs/Kconfig
index 66fa9ab2c046..3282adc84d52 100644
--- a/fs/btrfs/Kconfig
+++ b/fs/btrfs/Kconfig
@@ -49,9 +49,11 @@ config BTRFS_FS_POSIX_ACL
 	  If you don't know what Access Control Lists are, say N
 
 config BTRFS_FS_CHECK_INTEGRITY
-	bool "Btrfs with integrity check tool compiled in (DANGEROUS)"
+	bool "Btrfs with integrity check tool compiled in (DEPRECATED)"
 	depends on BTRFS_FS
 	help
+	  This feature has been deprecated and will be removed in 6.7.
+
 	  Adds code that examines all block write requests (including
 	  writes of the super block). The goal is to verify that the
 	  state of the filesystem on disk is always consistent, i.e.,
diff --git a/fs/btrfs/accessors.h b/fs/btrfs/accessors.h
index ceadfc5d6c66..8cfc8214109c 100644
--- a/fs/btrfs/accessors.h
+++ b/fs/btrfs/accessors.h
@@ -3,6 +3,8 @@
 #ifndef BTRFS_ACCESSORS_H
 #define BTRFS_ACCESSORS_H
 
+#include <linux/stddef.h>
+
 struct btrfs_map_token {
 	struct extent_buffer *eb;
 	char *kaddr;
@@ -34,13 +36,13 @@ static inline void put_unaligned_le8(u8 val, void *p)
 	read_extent_buffer(eb, (char *)(result),			\
 			   ((unsigned long)(ptr)) +			\
 			    offsetof(type, member),			\
-			   sizeof(((type *)0)->member)))
+			    sizeof_field(type, member)))
 
 #define write_eb_member(eb, ptr, type, member, result) (\
 	write_extent_buffer(eb, (char *)(result),			\
 			   ((unsigned long)(ptr)) +			\
 			    offsetof(type, member),			\
-			   sizeof(((type *)0)->member)))
+			    sizeof_field(type, member)))
 
 #define DECLARE_BTRFS_SETGET_BITS(bits)					\
 u##bits btrfs_get_token_##bits(struct btrfs_map_token *token,		\
@@ -62,25 +64,25 @@ DECLARE_BTRFS_SETGET_BITS(64)
 static inline u##bits btrfs_##name(const struct extent_buffer *eb,	\
 				   const type *s)			\
 {									\
-	static_assert(sizeof(u##bits) == sizeof(((type *)0))->member);	\
+	static_assert(sizeof(u##bits) == sizeof_field(type, member));	\
 	return btrfs_get_##bits(eb, s, offsetof(type, member));		\
 }									\
 static inline void btrfs_set_##name(const struct extent_buffer *eb, type *s, \
 				    u##bits val)			\
 {									\
-	static_assert(sizeof(u##bits) == sizeof(((type *)0))->member);	\
+	static_assert(sizeof(u##bits) == sizeof_field(type, member));	\
 	btrfs_set_##bits(eb, s, offsetof(type, member), val);		\
 }									\
 static inline u##bits btrfs_token_##name(struct btrfs_map_token *token,	\
 					 const type *s)			\
 {									\
-	static_assert(sizeof(u##bits) == sizeof(((type *)0))->member);	\
+	static_assert(sizeof(u##bits) == sizeof_field(type, member));	\
 	return btrfs_get_token_##bits(token, s, offsetof(type, member));\
 }									\
 static inline void btrfs_set_token_##name(struct btrfs_map_token *token,\
 					  type *s, u##bits val)		\
 {									\
-	static_assert(sizeof(u##bits) == sizeof(((type *)0))->member);	\
+	static_assert(sizeof(u##bits) == sizeof_field(type, member));	\
 	btrfs_set_token_##bits(token, s, offsetof(type, member), val);	\
 }
 
@@ -111,17 +113,14 @@ static inline void btrfs_set_##name(type *s, u##bits val)		\
 static inline u64 btrfs_device_total_bytes(const struct extent_buffer *eb,
 					   struct btrfs_dev_item *s)
 {
-	static_assert(sizeof(u64) ==
-		      sizeof(((struct btrfs_dev_item *)0))->total_bytes);
-	return btrfs_get_64(eb, s, offsetof(struct btrfs_dev_item,
-					    total_bytes));
+	static_assert(sizeof(u64) == sizeof_field(struct btrfs_dev_item, total_bytes));
+	return btrfs_get_64(eb, s, offsetof(struct btrfs_dev_item, total_bytes));
 }
 static inline void btrfs_set_device_total_bytes(const struct extent_buffer *eb,
 						struct btrfs_dev_item *s,
 						u64 val)
 {
-	static_assert(sizeof(u64) ==
-		      sizeof(((struct btrfs_dev_item *)0))->total_bytes);
+	static_assert(sizeof(u64) == sizeof_field(struct btrfs_dev_item, total_bytes));
 	WARN_ON(!IS_ALIGNED(val, eb->fs_info->sectorsize));
 	btrfs_set_64(eb, s, offsetof(struct btrfs_dev_item, total_bytes), val);
 }
diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c
index 79336fa853db..b7d54efb4728 100644
--- a/fs/btrfs/backref.c
+++ b/fs/btrfs/backref.c
@@ -3373,7 +3373,6 @@ int btrfs_backref_add_tree_node(struct btrfs_backref_cache *cache,
 				struct btrfs_key *node_key,
 				struct btrfs_backref_node *cur)
 {
-	struct btrfs_fs_info *fs_info = cache->fs_info;
 	struct btrfs_backref_edge *edge;
 	struct btrfs_backref_node *exist;
 	int ret;
@@ -3462,25 +3461,21 @@ int btrfs_backref_add_tree_node(struct btrfs_backref_cache *cache,
 			ret = handle_direct_tree_backref(cache, &key, cur);
 			if (ret < 0)
 				goto out;
-			continue;
-		} else if (unlikely(key.type == BTRFS_EXTENT_REF_V0_KEY)) {
-			ret = -EINVAL;
-			btrfs_print_v0_err(fs_info);
-			btrfs_handle_fs_error(fs_info, ret, NULL);
-			goto out;
-		} else if (key.type != BTRFS_TREE_BLOCK_REF_KEY) {
-			continue;
+		} else if (key.type == BTRFS_TREE_BLOCK_REF_KEY) {
+			/*
+			 * key.type == BTRFS_TREE_BLOCK_REF_KEY, inline ref
+			 * offset means the root objectid. We need to search
+			 * the tree to get its parent bytenr.
+			 */
+			ret = handle_indirect_tree_backref(cache, path, &key, node_key,
+							   cur);
+			if (ret < 0)
+				goto out;
 		}
-
 		/*
-		 * key.type == BTRFS_TREE_BLOCK_REF_KEY, inline ref offset
-		 * means the root objectid. We need to search the tree to get
-		 * its parent bytenr.
+		 * Unrecognized tree backref items (if it can pass tree-checker)
+		 * would be ignored.
 		 */
-		ret = handle_indirect_tree_backref(cache, path, &key, node_key,
-						   cur);
-		if (ret < 0)
-			goto out;
 	}
 	ret = 0;
 	cur->checked = 1;
diff --git a/fs/btrfs/block-group.c b/fs/btrfs/block-group.c
index 82324c327a50..0cb1dee965a0 100644
--- a/fs/btrfs/block-group.c
+++ b/fs/btrfs/block-group.c
@@ -504,13 +504,20 @@ static void fragment_free_space(struct btrfs_block_group *block_group)
 #endif
 
 /*
- * This is only called by btrfs_cache_block_group, since we could have freed
- * extents we need to check the pinned_extents for any extents that can't be
- * used yet since their free space will be released as soon as the transaction
- * commits.
+ * Add a free space range to the in memory free space cache of a block group.
+ * This checks if the range contains super block locations and any such
+ * locations are not added to the free space cache.
+ *
+ * @block_group:      The target block group.
+ * @start:            Start offset of the range.
+ * @end:              End offset of the range (exclusive).
+ * @total_added_ret:  Optional pointer to return the total amount of space
+ *                    added to the block group's free space cache.
+ *
+ * Returns 0 on success or < 0 on error.
  */
-int add_new_free_space(struct btrfs_block_group *block_group, u64 start, u64 end,
-		       u64 *total_added_ret)
+int btrfs_add_new_free_space(struct btrfs_block_group *block_group, u64 start,
+			     u64 end, u64 *total_added_ret)
 {
 	struct btrfs_fs_info *info = block_group->fs_info;
 	u64 extent_start, extent_end, size;
@@ -520,11 +527,10 @@ int add_new_free_space(struct btrfs_block_group *block_group, u64 start, u64 end
 		*total_added_ret = 0;
 
 	while (start < end) {
-		ret = find_first_extent_bit(&info->excluded_extents, start,
-					    &extent_start, &extent_end,
-					    EXTENT_DIRTY | EXTENT_UPTODATE,
-					    NULL);
-		if (ret)
+		if (!find_first_extent_bit(&info->excluded_extents, start,
+					   &extent_start, &extent_end,
+					   EXTENT_DIRTY | EXTENT_UPTODATE,
+					   NULL))
 			break;
 
 		if (extent_start <= start) {
@@ -799,8 +805,8 @@ next:
 		    key.type == BTRFS_METADATA_ITEM_KEY) {
 			u64 space_added;
 
-			ret = add_new_free_space(block_group, last, key.objectid,
-						 &space_added);
+			ret = btrfs_add_new_free_space(block_group, last,
+						       key.objectid, &space_added);
 			if (ret)
 				goto out;
 			total_found += space_added;
@@ -821,14 +827,20 @@ next:
 		path->slots[0]++;
 	}
 
-	ret = add_new_free_space(block_group, last,
-				 block_group->start + block_group->length,
-				 NULL);
+	ret = btrfs_add_new_free_space(block_group, last,
+				       block_group->start + block_group->length,
+				       NULL);
 out:
 	btrfs_free_path(path);
 	return ret;
 }
 
+static inline void btrfs_free_excluded_extents(const struct btrfs_block_group *bg)
+{
+	clear_extent_bits(&bg->fs_info->excluded_extents, bg->start,
+			  bg->start + bg->length - 1, EXTENT_UPTODATE);
+}
+
 static noinline void caching_thread(struct btrfs_work *work)
 {
 	struct btrfs_block_group *block_group;
@@ -2098,8 +2110,9 @@ static int exclude_super_stripes(struct btrfs_block_group *cache)
 	if (cache->start < BTRFS_SUPER_INFO_OFFSET) {
 		stripe_len = BTRFS_SUPER_INFO_OFFSET - cache->start;
 		cache->bytes_super += stripe_len;
-		ret = btrfs_add_excluded_extent(fs_info, cache->start,
-						stripe_len);
+		ret = set_extent_bit(&fs_info->excluded_extents, cache->start,
+				     cache->start + stripe_len - 1,
+				     EXTENT_UPTODATE, NULL);
 		if (ret)
 			return ret;
 	}
@@ -2125,8 +2138,9 @@ static int exclude_super_stripes(struct btrfs_block_group *cache)
 				cache->start + cache->length - logical[nr]);
 
 			cache->bytes_super += len;
-			ret = btrfs_add_excluded_extent(fs_info, logical[nr],
-							len);
+			ret = set_extent_bit(&fs_info->excluded_extents, logical[nr],
+					     logical[nr] + len - 1,
+					     EXTENT_UPTODATE, NULL);
 			if (ret) {
 				kfree(logical);
 				return ret;
@@ -2319,8 +2333,8 @@ static int read_one_block_group(struct btrfs_fs_info *info,
 		btrfs_free_excluded_extents(cache);
 	} else if (cache->used == 0) {
 		cache->cached = BTRFS_CACHE_FINISHED;
-		ret = add_new_free_space(cache, cache->start,
-					 cache->start + cache->length, NULL);
+		ret = btrfs_add_new_free_space(cache, cache->start,
+					       cache->start + cache->length, NULL);
 		btrfs_free_excluded_extents(cache);
 		if (ret)
 			goto error;
@@ -2767,7 +2781,7 @@ struct btrfs_block_group *btrfs_make_block_group(struct btrfs_trans_handle *tran
 		return ERR_PTR(ret);
 	}
 
-	ret = add_new_free_space(cache, chunk_offset, chunk_offset + size, NULL);
+	ret = btrfs_add_new_free_space(cache, chunk_offset, chunk_offset + size, NULL);
 	btrfs_free_excluded_extents(cache);
 	if (ret) {
 		btrfs_put_block_group(cache);
@@ -4075,7 +4089,7 @@ int btrfs_chunk_alloc(struct btrfs_trans_handle *trans, u64 flags,
 
 	if (IS_ERR(ret_bg)) {
 		ret = PTR_ERR(ret_bg);
-	} else if (from_extent_allocation) {
+	} else if (from_extent_allocation && (flags & BTRFS_BLOCK_GROUP_DATA)) {
 		/*
 		 * New block group is likely to be used soon. Try to activate
 		 * it now. Failure is OK for now.
@@ -4273,6 +4287,17 @@ int btrfs_free_block_groups(struct btrfs_fs_info *info)
 	struct btrfs_caching_control *caching_ctl;
 	struct rb_node *n;
 
+	if (btrfs_is_zoned(info)) {
+		if (info->active_meta_bg) {
+			btrfs_put_block_group(info->active_meta_bg);
+			info->active_meta_bg = NULL;
+		}
+		if (info->active_system_bg) {
+			btrfs_put_block_group(info->active_system_bg);
+			info->active_system_bg = NULL;
+		}
+	}
+
 	write_lock(&info->block_group_cache_lock);
 	while (!list_empty(&info->caching_block_groups)) {
 		caching_ctl = list_entry(info->caching_block_groups.next,
diff --git a/fs/btrfs/block-group.h b/fs/btrfs/block-group.h
index 74b61e663028..2bdbcb834f95 100644
--- a/fs/btrfs/block-group.h
+++ b/fs/btrfs/block-group.h
@@ -291,8 +291,8 @@ int btrfs_cache_block_group(struct btrfs_block_group *cache, bool wait);
 void btrfs_put_caching_control(struct btrfs_caching_control *ctl);
 struct btrfs_caching_control *btrfs_get_caching_control(
 		struct btrfs_block_group *cache);
-int add_new_free_space(struct btrfs_block_group *block_group,
-		       u64 start, u64 end, u64 *total_added_ret);
+int btrfs_add_new_free_space(struct btrfs_block_group *block_group,
+			     u64 start, u64 end, u64 *total_added_ret);
 struct btrfs_trans_handle *btrfs_start_trans_remove_block_group(
 				struct btrfs_fs_info *fs_info,
 				const u64 chunk_offset);
diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h
index d47a927b3504..bda1fdbba666 100644
--- a/fs/btrfs/btrfs_inode.h
+++ b/fs/btrfs/btrfs_inode.h
@@ -498,12 +498,8 @@ int btrfs_prealloc_file_range_trans(struct inode *inode,
 				    u64 start, u64 num_bytes, u64 min_size,
 				    loff_t actual_len, u64 *alloc_hint);
 int btrfs_run_delalloc_range(struct btrfs_inode *inode, struct page *locked_page,
-			     u64 start, u64 end, int *page_started,
-			     unsigned long *nr_written, struct writeback_control *wbc);
+			     u64 start, u64 end, struct writeback_control *wbc);
 int btrfs_writepage_cow_fixup(struct page *page);
-void btrfs_writepage_endio_finish_ordered(struct btrfs_inode *inode,
-					  struct page *page, u64 start,
-					  u64 end, bool uptodate);
 int btrfs_encoded_io_compression_from_extent(struct btrfs_fs_info *fs_info,
 					     int compress_type);
 int btrfs_encoded_read_regular_fill_pages(struct btrfs_inode *inode,
diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c
index d37c066f2df7..53c1211dd60b 100644
--- a/fs/btrfs/delayed-inode.c
+++ b/fs/btrfs/delayed-inode.c
@@ -1736,9 +1736,6 @@ int btrfs_readdir_delayed_dir_index(struct dir_context *ctx,
 	int over = 0;
 	unsigned char d_type;
 
-	if (list_empty(ins_list))
-		return 0;
-
 	/*
 	 * Changing the data of the delayed item is impossible. So
 	 * we needn't lock them. And we have held i_mutex of the
diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c
index 5f10965fd72b..fff22ed55c42 100644
--- a/fs/btrfs/dev-replace.c
+++ b/fs/btrfs/dev-replace.c
@@ -792,9 +792,9 @@ static int btrfs_set_target_alloc_state(struct btrfs_device *srcdev,
 
 	lockdep_assert_held(&srcdev->fs_info->chunk_mutex);
 
-	while (!find_first_extent_bit(&srcdev->alloc_state, start,
-				      &found_start, &found_end,
-				      CHUNK_ALLOCATED, &cached_state)) {
+	while (find_first_extent_bit(&srcdev->alloc_state, start,
+				     &found_start, &found_end,
+				     CHUNK_ALLOCATED, &cached_state)) {
 		ret = set_extent_bit(&tgtdev->alloc_state, found_start,
 				     found_end, CHUNK_ALLOCATED, NULL);
 		if (ret)
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index a9a2c5446c18..0a96ea8c1d3a 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -313,21 +313,16 @@ static bool check_tree_block_fsid(struct extent_buffer *eb)
 	struct btrfs_fs_info *fs_info = eb->fs_info;
 	struct btrfs_fs_devices *fs_devices = fs_info->fs_devices, *seed_devs;
 	u8 fsid[BTRFS_FSID_SIZE];
-	u8 *metadata_uuid;
 
 	read_extent_buffer(eb, fsid, offsetof(struct btrfs_header, fsid),
 			   BTRFS_FSID_SIZE);
+
 	/*
-	 * Checking the incompat flag is only valid for the current fs. For
-	 * seed devices it's forbidden to have their uuid changed so reading
-	 * ->fsid in this case is fine
+	 * alloc_fs_devices() copies the fsid into metadata_uuid if the
+	 * metadata_uuid is unset in the superblock, including for a seed device.
+	 * So, we can use fs_devices->metadata_uuid.
 	 */
-	if (btrfs_fs_incompat(fs_info, METADATA_UUID))
-		metadata_uuid = fs_devices->metadata_uuid;
-	else
-		metadata_uuid = fs_devices->fsid;
-
-	if (!memcmp(fsid, metadata_uuid, BTRFS_FSID_SIZE))
+	if (memcmp(fsid, fs_info->fs_devices->metadata_uuid, BTRFS_FSID_SIZE) == 0)
 		return false;
 
 	list_for_each_entry(seed_devs, &fs_devices->seed_list, seed_list)
@@ -2384,21 +2379,18 @@ int btrfs_validate_super(struct btrfs_fs_info *fs_info,
 		ret = -EINVAL;
 	}
 
-	if (memcmp(fs_info->fs_devices->fsid, fs_info->super_copy->fsid,
-		   BTRFS_FSID_SIZE)) {
+	if (memcmp(fs_info->fs_devices->fsid, sb->fsid, BTRFS_FSID_SIZE) != 0) {
 		btrfs_err(fs_info,
 		"superblock fsid doesn't match fsid of fs_devices: %pU != %pU",
-			fs_info->super_copy->fsid, fs_info->fs_devices->fsid);
+			  sb->fsid, fs_info->fs_devices->fsid);
 		ret = -EINVAL;
 	}
 
-	if (btrfs_fs_incompat(fs_info, METADATA_UUID) &&
-	    memcmp(fs_info->fs_devices->metadata_uuid,
-		   fs_info->super_copy->metadata_uuid, BTRFS_FSID_SIZE)) {
+	if (memcmp(fs_info->fs_devices->metadata_uuid, btrfs_sb_fsid_ptr(sb),
+		   BTRFS_FSID_SIZE) != 0) {
 		btrfs_err(fs_info,
 "superblock metadata_uuid doesn't match metadata uuid of fs_devices: %pU != %pU",
-			fs_info->super_copy->metadata_uuid,
-			fs_info->fs_devices->metadata_uuid);
+			  btrfs_sb_fsid_ptr(sb), fs_info->fs_devices->metadata_uuid);
 		ret = -EINVAL;
 	}
 
@@ -2869,6 +2861,56 @@ static int btrfs_check_uuid_tree(struct btrfs_fs_info *fs_info)
 	return 0;
 }
 
+static int btrfs_cleanup_fs_roots(struct btrfs_fs_info *fs_info)
+{
+	u64 root_objectid = 0;
+	struct btrfs_root *gang[8];
+	int i = 0;
+	int err = 0;
+	unsigned int ret = 0;
+
+	while (1) {
+		spin_lock(&fs_info->fs_roots_radix_lock);
+		ret = radix_tree_gang_lookup(&fs_info->fs_roots_radix,
+					     (void **)gang, root_objectid,
+					     ARRAY_SIZE(gang));
+		if (!ret) {
+			spin_unlock(&fs_info->fs_roots_radix_lock);
+			break;
+		}
+		root_objectid = gang[ret - 1]->root_key.objectid + 1;
+
+		for (i = 0; i < ret; i++) {
+			/* Avoid to grab roots in dead_roots. */
+			if (btrfs_root_refs(&gang[i]->root_item) == 0) {
+				gang[i] = NULL;
+				continue;
+			}
+			/* Grab all the search result for later use. */
+			gang[i] = btrfs_grab_root(gang[i]);
+		}
+		spin_unlock(&fs_info->fs_roots_radix_lock);
+
+		for (i = 0; i < ret; i++) {
+			if (!gang[i])
+				continue;
+			root_objectid = gang[i]->root_key.objectid;
+			err = btrfs_orphan_cleanup(gang[i]);
+			if (err)
+				goto out;
+			btrfs_put_root(gang[i]);
+		}
+		root_objectid++;
+	}
+out:
+	/* Release the uncleaned roots due to error. */
+	for (; i < ret; i++) {
+		if (gang[i])
+			btrfs_put_root(gang[i]);
+	}
+	return err;
+}
+
 /*
  * Some options only have meaning at mount time and shouldn't persist across
  * remounts, or be displayed. Clear these at the end of mount and remount
@@ -3222,7 +3264,7 @@ int __cold open_ctree(struct super_block *sb, struct btrfs_fs_devices *fs_device
 
 	/* check FS state, whether FS is broken. */
 	if (btrfs_super_flags(disk_super) & BTRFS_SUPER_FLAG_ERROR)
-		set_bit(BTRFS_FS_STATE_ERROR, &fs_info->fs_state);
+		WRITE_ONCE(fs_info->fs_error, -EUCLEAN);
 
 	/*
 	 * In the long term, we'll store the compression type in the super
@@ -3417,6 +3459,8 @@ int __cold open_ctree(struct super_block *sb, struct btrfs_fs_devices *fs_device
 
 	btrfs_free_zone_cache(fs_info);
 
+	btrfs_check_active_zone_reservation(fs_info);
+
 	if (!sb_rdonly(sb) && fs_info->fs_devices->missing_devices &&
 	    !btrfs_check_rw_degradable(fs_info, NULL)) {
 		btrfs_warn(fs_info,
@@ -4136,56 +4180,6 @@ void btrfs_drop_and_free_fs_root(struct btrfs_fs_info *fs_info,
 		btrfs_put_root(root);
 }
 
-int btrfs_cleanup_fs_roots(struct btrfs_fs_info *fs_info)
-{
-	u64 root_objectid = 0;
-	struct btrfs_root *gang[8];
-	int i = 0;
-	int err = 0;
-	unsigned int ret = 0;
-
-	while (1) {
-		spin_lock(&fs_info->fs_roots_radix_lock);
-		ret = radix_tree_gang_lookup(&fs_info->fs_roots_radix,
-					     (void **)gang, root_objectid,
-					     ARRAY_SIZE(gang));
-		if (!ret) {
-			spin_unlock(&fs_info->fs_roots_radix_lock);
-			break;
-		}
-		root_objectid = gang[ret - 1]->root_key.objectid + 1;
-
-		for (i = 0; i < ret; i++) {
-			/* Avoid to grab roots in dead_roots */
-			if (btrfs_root_refs(&gang[i]->root_item) == 0) {
-				gang[i] = NULL;
-				continue;
-			}
-			/* grab all the search result for later use */
-			gang[i] = btrfs_grab_root(gang[i]);
-		}
-		spin_unlock(&fs_info->fs_roots_radix_lock);
-
-		for (i = 0; i < ret; i++) {
-			if (!gang[i])
-				continue;
-			root_objectid = gang[i]->root_key.objectid;
-			err = btrfs_orphan_cleanup(gang[i]);
-			if (err)
-				goto out;
-			btrfs_put_root(gang[i]);
-		}
-		root_objectid++;
-	}
-out:
-	/* release the uncleaned roots due to error */
-	for (; i < ret; i++) {
-		if (gang[i])
-			btrfs_put_root(gang[i]);
-	}
-	return err;
-}
-
 int btrfs_commit_super(struct btrfs_fs_info *fs_info)
 {
 	struct btrfs_root *root = fs_info->tree_root;
@@ -4228,7 +4222,7 @@ static void warn_about_uncommitted_trans(struct btrfs_fs_info *fs_info)
 		u64 found_end;
 
 		found = true;
-		while (!find_first_extent_bit(&trans->dirty_pages, cur,
+		while (find_first_extent_bit(&trans->dirty_pages, cur,
 			&found_start, &found_end, EXTENT_DIRTY, &cached)) {
 			dirty_bytes += found_end + 1 - found_start;
 			cur = found_end + 1;
@@ -4552,9 +4546,7 @@ static void btrfs_destroy_ordered_extents(struct btrfs_root *root)
 static void btrfs_destroy_all_ordered_extents(struct btrfs_fs_info *fs_info)
 {
 	struct btrfs_root *root;
-	struct list_head splice;
-
-	INIT_LIST_HEAD(&splice);
+	LIST_HEAD(splice);
 
 	spin_lock(&fs_info->ordered_root_lock);
 	list_splice_init(&fs_info->ordered_roots, &splice);
@@ -4660,9 +4652,7 @@ static void btrfs_destroy_delayed_refs(struct btrfs_transaction *trans,
 static void btrfs_destroy_delalloc_inodes(struct btrfs_root *root)
 {
 	struct btrfs_inode *btrfs_inode;
-	struct list_head splice;
-
-	INIT_LIST_HEAD(&splice);
+	LIST_HEAD(splice);
 
 	spin_lock(&root->delalloc_lock);
 	list_splice_init(&root->delalloc_inodes, &splice);
@@ -4695,9 +4685,7 @@ static void btrfs_destroy_delalloc_inodes(struct btrfs_root *root)
 static void btrfs_destroy_all_delalloc_inodes(struct btrfs_fs_info *fs_info)
 {
 	struct btrfs_root *root;
-	struct list_head splice;
-
-	INIT_LIST_HEAD(&splice);
+	LIST_HEAD(splice);
 
 	spin_lock(&fs_info->delalloc_root_lock);
 	list_splice_init(&fs_info->delalloc_roots, &splice);
@@ -4716,21 +4704,16 @@ static void btrfs_destroy_all_delalloc_inodes(struct btrfs_fs_info *fs_info)
 	spin_unlock(&fs_info->delalloc_root_lock);
 }
 
-static int btrfs_destroy_marked_extents(struct btrfs_fs_info *fs_info,
-					struct extent_io_tree *dirty_pages,
-					int mark)
+static void btrfs_destroy_marked_extents(struct btrfs_fs_info *fs_info,
+					 struct extent_io_tree *dirty_pages,
+					 int mark)
 {
-	int ret;
 	struct extent_buffer *eb;
 	u64 start = 0;
 	u64 end;
 
-	while (1) {
-		ret = find_first_extent_bit(dirty_pages, start, &start, &end,
-					    mark, NULL);
-		if (ret)
-			break;
-
+	while (find_first_extent_bit(dirty_pages, start, &start, &end,
+				     mark, NULL)) {
 		clear_extent_bits(dirty_pages, start, end, mark);
 		while (start <= end) {
 			eb = find_extent_buffer(fs_info, start);
@@ -4746,16 +4729,13 @@ static int btrfs_destroy_marked_extents(struct btrfs_fs_info *fs_info,
 			free_extent_buffer_stale(eb);
 		}
 	}
-
-	return ret;
 }
 
-static int btrfs_destroy_pinned_extent(struct btrfs_fs_info *fs_info,
-				       struct extent_io_tree *unpin)
+static void btrfs_destroy_pinned_extent(struct btrfs_fs_info *fs_info,
+					struct extent_io_tree *unpin)
 {
 	u64 start;
 	u64 end;
-	int ret;
 
 	while (1) {
 		struct extent_state *cached_state = NULL;
@@ -4767,9 +4747,8 @@ static int btrfs_destroy_pinned_extent(struct btrfs_fs_info *fs_info,
 		 * the same extent range.
 		 */
 		mutex_lock(&fs_info->unused_bg_unpin_mutex);
-		ret = find_first_extent_bit(unpin, 0, &start, &end,
-					    EXTENT_DIRTY, &cached_state);
-		if (ret) {
+		if (!find_first_extent_bit(unpin, 0, &start, &end,
+					   EXTENT_DIRTY, &cached_state)) {
 			mutex_unlock(&fs_info->unused_bg_unpin_mutex);
 			break;
 		}
@@ -4780,8 +4759,6 @@ static int btrfs_destroy_pinned_extent(struct btrfs_fs_info *fs_info,
 		mutex_unlock(&fs_info->unused_bg_unpin_mutex);
 		cond_resched();
 	}
-
-	return 0;
 }
 
 static void btrfs_cleanup_bg_io(struct btrfs_block_group *cache)
diff --git a/fs/btrfs/disk-io.h b/fs/btrfs/disk-io.h
index b03767f4d7ed..02b645744a82 100644
--- a/fs/btrfs/disk-io.h
+++ b/fs/btrfs/disk-io.h
@@ -77,7 +77,6 @@ struct btrfs_root *btrfs_extent_root(struct btrfs_fs_info *fs_info, u64 bytenr);
 struct btrfs_root *btrfs_block_group_root(struct btrfs_fs_info *fs_info);
 
 void btrfs_free_fs_info(struct btrfs_fs_info *fs_info);
-int btrfs_cleanup_fs_roots(struct btrfs_fs_info *fs_info);
 void btrfs_btree_balance_dirty(struct btrfs_fs_info *fs_info);
 void btrfs_btree_balance_dirty_nodelay(struct btrfs_fs_info *fs_info);
 void btrfs_drop_and_free_fs_root(struct btrfs_fs_info *fs_info,
diff --git a/fs/btrfs/extent-io-tree.c b/fs/btrfs/extent-io-tree.c
index a2315a4b8b75..ff8e117a1ace 100644
--- a/fs/btrfs/extent-io-tree.c
+++ b/fs/btrfs/extent-io-tree.c
@@ -831,15 +831,15 @@ static struct extent_state *find_first_extent_bit_state(struct extent_io_tree *t
  *
  * Note: If there are multiple bits set in @bits, any of them will match.
  *
- * Return 0 if we find something, and update @start_ret and @end_ret.
- * Return 1 if we found nothing.
+ * Return true if we find something, and update @start_ret and @end_ret.
+ * Return false if we found nothing.
  */
-int find_first_extent_bit(struct extent_io_tree *tree, u64 start,
-			  u64 *start_ret, u64 *end_ret, u32 bits,
-			  struct extent_state **cached_state)
+bool find_first_extent_bit(struct extent_io_tree *tree, u64 start,
+			   u64 *start_ret, u64 *end_ret, u32 bits,
+			   struct extent_state **cached_state)
 {
 	struct extent_state *state;
-	int ret = 1;
+	bool ret = false;
 
 	spin_lock(&tree->lock);
 	if (cached_state && *cached_state) {
@@ -863,7 +863,7 @@ got_it:
 		cache_state_if_flags(state, cached_state, 0);
 		*start_ret = state->start;
 		*end_ret = state->end;
-		ret = 0;
+		ret = true;
 	}
 out:
 	spin_unlock(&tree->lock);
diff --git a/fs/btrfs/extent-io-tree.h b/fs/btrfs/extent-io-tree.h
index fbd3b275ab1c..28c23a23d121 100644
--- a/fs/btrfs/extent-io-tree.h
+++ b/fs/btrfs/extent-io-tree.h
@@ -182,9 +182,9 @@ int convert_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
 		       u32 bits, u32 clear_bits,
 		       struct extent_state **cached_state);
 
-int find_first_extent_bit(struct extent_io_tree *tree, u64 start,
-			  u64 *start_ret, u64 *end_ret, u32 bits,
-			  struct extent_state **cached_state);
+bool find_first_extent_bit(struct extent_io_tree *tree, u64 start,
+			   u64 *start_ret, u64 *end_ret, u32 bits,
+			   struct extent_state **cached_state);
 void find_first_clear_extent_bit(struct extent_io_tree *tree, u64 start,
 				 u64 *start_ret, u64 *end_ret, u32 bits);
 int find_contiguous_extent_bit(struct extent_io_tree *tree, u64 start,
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index f396a9afa403..f356f08b55cb 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -69,27 +69,6 @@ static int block_group_bits(struct btrfs_block_group *cache, u64 bits)
 	return (cache->flags & bits) == bits;
 }
 
-int btrfs_add_excluded_extent(struct btrfs_fs_info *fs_info,
-			      u64 start, u64 num_bytes)
-{
-	u64 end = start + num_bytes - 1;
-	set_extent_bit(&fs_info->excluded_extents, start, end,
-		       EXTENT_UPTODATE, NULL);
-	return 0;
-}
-
-void btrfs_free_excluded_extents(struct btrfs_block_group *cache)
-{
-	struct btrfs_fs_info *fs_info = cache->fs_info;
-	u64 start, end;
-
-	start = cache->start;
-	end = start + cache->length - 1;
-
-	clear_extent_bits(&fs_info->excluded_extents, start, end,
-			  EXTENT_UPTODATE);
-}
-
 /* simple helper to search for an existing data extent at a given offset */
 int btrfs_lookup_data_extent(struct btrfs_fs_info *fs_info, u64 start, u64 len)
 {
@@ -187,8 +166,10 @@ search_again:
 			num_refs = btrfs_extent_refs(leaf, ei);
 			extent_flags = btrfs_extent_flags(leaf, ei);
 		} else {
-			ret = -EINVAL;
-			btrfs_print_v0_err(fs_info);
+			ret = -EUCLEAN;
+			btrfs_err(fs_info,
+			"unexpected extent item size, has %u expect >= %zu",
+				  item_size, sizeof(*ei));
 			if (trans)
 				btrfs_abort_transaction(trans, ret);
 			else
@@ -402,11 +383,11 @@ int btrfs_get_extent_inline_ref_type(const struct extent_buffer *eb,
 		}
 	}
 
+	WARN_ON(1);
 	btrfs_print_leaf(eb);
 	btrfs_err(eb->fs_info,
 		  "eb %llu iref 0x%lx invalid extent inline ref type %d",
 		  eb->start, (unsigned long)iref, type);
-	WARN_ON(1);
 
 	return BTRFS_REF_TYPE_INVALID;
 }
@@ -624,12 +605,12 @@ static noinline int remove_extent_data_ref(struct btrfs_trans_handle *trans,
 		ref2 = btrfs_item_ptr(leaf, path->slots[0],
 				      struct btrfs_shared_data_ref);
 		num_refs = btrfs_shared_data_ref_count(leaf, ref2);
-	} else if (unlikely(key.type == BTRFS_EXTENT_REF_V0_KEY)) {
-		btrfs_print_v0_err(trans->fs_info);
-		btrfs_abort_transaction(trans, -EINVAL);
-		return -EINVAL;
 	} else {
-		BUG();
+		btrfs_err(trans->fs_info,
+			  "unrecognized backref key (%llu %u %llu)",
+			  key.objectid, key.type, key.offset);
+		btrfs_abort_transaction(trans, -EUCLEAN);
+		return -EUCLEAN;
 	}
 
 	BUG_ON(num_refs < refs_to_drop);
@@ -660,7 +641,6 @@ static noinline u32 extent_data_ref_count(struct btrfs_path *path,
 	leaf = path->nodes[0];
 	btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
 
-	BUG_ON(key.type == BTRFS_EXTENT_REF_V0_KEY);
 	if (iref) {
 		/*
 		 * If type is invalid, we should have bailed out earlier than
@@ -869,6 +849,11 @@ again:
 		err = -ENOENT;
 		goto out;
 	} else if (WARN_ON(ret)) {
+		btrfs_print_leaf(path->nodes[0]);
+		btrfs_err(fs_info,
+"extent item not found for insert, bytenr %llu num_bytes %llu parent %llu root_objectid %llu owner %llu offset %llu",
+			  bytenr, num_bytes, parent, root_objectid, owner,
+			  offset);
 		err = -EIO;
 		goto out;
 	}
@@ -876,8 +861,10 @@ again:
 	leaf = path->nodes[0];
 	item_size = btrfs_item_size(leaf, path->slots[0]);
 	if (unlikely(item_size < sizeof(*ei))) {
-		err = -EINVAL;
-		btrfs_print_v0_err(fs_info);
+		err = -EUCLEAN;
+		btrfs_err(fs_info,
+			  "unexpected extent item size, has %llu expect >= %zu",
+			  item_size, sizeof(*ei));
 		btrfs_abort_transaction(trans, err);
 		goto out;
 	}
@@ -1079,13 +1066,13 @@ static int lookup_extent_backref(struct btrfs_trans_handle *trans,
 /*
  * helper to update/remove inline back ref
  */
-static noinline_for_stack
-void update_inline_extent_backref(struct btrfs_path *path,
+static noinline_for_stack int update_inline_extent_backref(struct btrfs_path *path,
 				  struct btrfs_extent_inline_ref *iref,
 				  int refs_to_mod,
 				  struct btrfs_delayed_extent_op *extent_op)
 {
 	struct extent_buffer *leaf = path->nodes[0];
+	struct btrfs_fs_info *fs_info = leaf->fs_info;
 	struct btrfs_extent_item *ei;
 	struct btrfs_extent_data_ref *dref = NULL;
 	struct btrfs_shared_data_ref *sref = NULL;
@@ -1098,18 +1085,33 @@ void update_inline_extent_backref(struct btrfs_path *path,
 
 	ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
 	refs = btrfs_extent_refs(leaf, ei);
-	WARN_ON(refs_to_mod < 0 && refs + refs_to_mod <= 0);
+	if (unlikely(refs_to_mod < 0 && refs + refs_to_mod <= 0)) {
+		struct btrfs_key key;
+		u32 extent_size;
+
+		btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
+		if (key.type == BTRFS_METADATA_ITEM_KEY)
+			extent_size = fs_info->nodesize;
+		else
+			extent_size = key.offset;
+		btrfs_print_leaf(leaf);
+		btrfs_err(fs_info,
+	"invalid refs_to_mod for extent %llu num_bytes %u, has %d expect >= -%llu",
+			  key.objectid, extent_size, refs_to_mod, refs);
+		return -EUCLEAN;
+	}
 	refs += refs_to_mod;
 	btrfs_set_extent_refs(leaf, ei, refs);
 	if (extent_op)
 		__run_delayed_extent_op(extent_op, leaf, ei);
 
+	type = btrfs_get_extent_inline_ref_type(leaf, iref, BTRFS_REF_TYPE_ANY);
 	/*
-	 * If type is invalid, we should have bailed out after
-	 * lookup_inline_extent_backref().
+	 * Function btrfs_get_extent_inline_ref_type() has already printed
+	 * error messages.
 	 */
-	type = btrfs_get_extent_inline_ref_type(leaf, iref, BTRFS_REF_TYPE_ANY);
-	ASSERT(type != BTRFS_REF_TYPE_INVALID);
+	if (unlikely(type == BTRFS_REF_TYPE_INVALID))
+		return -EUCLEAN;
 
 	if (type == BTRFS_EXTENT_DATA_REF_KEY) {
 		dref = (struct btrfs_extent_data_ref *)(&iref->offset);
@@ -1119,10 +1121,43 @@ void update_inline_extent_backref(struct btrfs_path *path,
 		refs = btrfs_shared_data_ref_count(leaf, sref);
 	} else {
 		refs = 1;
-		BUG_ON(refs_to_mod != -1);
+		/*
+		 * For tree blocks we can only drop one ref for it, and tree
+		 * blocks should not have refs > 1.
+		 *
+		 * Furthermore if we're inserting a new inline backref, we
+		 * won't reach this path either. That would be
+		 * setup_inline_extent_backref().
+		 */
+		if (unlikely(refs_to_mod != -1)) {
+			struct btrfs_key key;
+
+			btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
+
+			btrfs_print_leaf(leaf);
+			btrfs_err(fs_info,
+			"invalid refs_to_mod for tree block %llu, has %d expect -1",
+				  key.objectid, refs_to_mod);
+			return -EUCLEAN;
+		}
 	}
 
-	BUG_ON(refs_to_mod < 0 && refs < -refs_to_mod);
+	if (unlikely(refs_to_mod < 0 && refs < -refs_to_mod)) {
+		struct btrfs_key key;
+		u32 extent_size;
+
+		btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
+		if (key.type == BTRFS_METADATA_ITEM_KEY)
+			extent_size = fs_info->nodesize;
+		else
+			extent_size = key.offset;
+		btrfs_print_leaf(leaf);
+		btrfs_err(fs_info,
+"invalid refs_to_mod for backref entry, iref %lu extent %llu num_bytes %u, has %d expect >= -%llu",
+			  (unsigned long)iref, key.objectid, extent_size,
+			  refs_to_mod, refs);
+		return -EUCLEAN;
+	}
 	refs += refs_to_mod;
 
 	if (refs > 0) {
@@ -1142,6 +1177,7 @@ void update_inline_extent_backref(struct btrfs_path *path,
 		btrfs_truncate_item(path, item_size, 1);
 	}
 	btrfs_mark_buffer_dirty(leaf);
+	return 0;
 }
 
 static noinline_for_stack
@@ -1170,7 +1206,7 @@ int insert_inline_extent_backref(struct btrfs_trans_handle *trans,
 				   bytenr, num_bytes, root_objectid, path->slots[0]);
 			return -EUCLEAN;
 		}
-		update_inline_extent_backref(path, iref, refs_to_add, extent_op);
+		ret = update_inline_extent_backref(path, iref, refs_to_add, extent_op);
 	} else if (ret == -ENOENT) {
 		setup_inline_extent_backref(trans->fs_info, path, iref, parent,
 					    root_objectid, owner, offset,
@@ -1190,7 +1226,7 @@ static int remove_extent_backref(struct btrfs_trans_handle *trans,
 
 	BUG_ON(!is_data && refs_to_drop != 1);
 	if (iref)
-		update_inline_extent_backref(path, iref, -refs_to_drop, NULL);
+		ret = update_inline_extent_backref(path, iref, -refs_to_drop, NULL);
 	else if (is_data)
 		ret = remove_extent_data_ref(trans, root, path, refs_to_drop);
 	else
@@ -1629,8 +1665,10 @@ again:
 	item_size = btrfs_item_size(leaf, path->slots[0]);
 
 	if (unlikely(item_size < sizeof(*ei))) {
-		err = -EINVAL;
-		btrfs_print_v0_err(fs_info);
+		err = -EUCLEAN;
+		btrfs_err(fs_info,
+			  "unexpected extent item size, has %u expect >= %zu",
+			  item_size, sizeof(*ei));
 		btrfs_abort_transaction(trans, err);
 		goto out;
 	}
@@ -2751,9 +2789,8 @@ int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans)
 		struct extent_state *cached_state = NULL;
 
 		mutex_lock(&fs_info->unused_bg_unpin_mutex);
-		ret = find_first_extent_bit(unpin, 0, &start, &end,
-					    EXTENT_DIRTY, &cached_state);
-		if (ret) {
+		if (!find_first_extent_bit(unpin, 0, &start, &end,
+					   EXTENT_DIRTY, &cached_state)) {
 			mutex_unlock(&fs_info->unused_bg_unpin_mutex);
 			break;
 		}
@@ -3059,8 +3096,10 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 	leaf = path->nodes[0];
 	item_size = btrfs_item_size(leaf, extent_slot);
 	if (unlikely(item_size < sizeof(*ei))) {
-		ret = -EINVAL;
-		btrfs_print_v0_err(info);
+		ret = -EUCLEAN;
+		btrfs_err(trans->fs_info,
+			  "unexpected extent item size, has %u expect >= %zu",
+			  item_size, sizeof(*ei));
 		btrfs_abort_transaction(trans, ret);
 		goto out;
 	}
@@ -3351,11 +3390,38 @@ int btrfs_free_extent(struct btrfs_trans_handle *trans, struct btrfs_ref *ref)
 }
 
 enum btrfs_loop_type {
+	/*
+	 * Start caching block groups but do not wait for progress or for them
+	 * to be done.
+	 */
 	LOOP_CACHING_NOWAIT,
+
+	/*
+	 * Wait for the block group free_space >= the space we're waiting for if
+	 * the block group isn't cached.
+	 */
 	LOOP_CACHING_WAIT,
+
+	/*
+	 * Allow allocations to happen from block groups that do not yet have a
+	 * size classification.
+	 */
 	LOOP_UNSET_SIZE_CLASS,
+
+	/*
+	 * Allocate a chunk and then retry the allocation.
+	 */
 	LOOP_ALLOC_CHUNK,
+
+	/*
+	 * Ignore the size class restrictions for this allocation.
+	 */
 	LOOP_WRONG_SIZE_CLASS,
+
+	/*
+	 * Ignore the empty size, only try to allocate the number of bytes
+	 * needed for this allocation.
+	 */
 	LOOP_NO_EMPTY_SIZE,
 };
 
@@ -3427,7 +3493,6 @@ btrfs_release_block_group(struct btrfs_block_group *cache,
  * Helper function for find_free_extent().
  *
  * Return -ENOENT to inform caller that we need fallback to unclustered mode.
- * Return -EAGAIN to inform caller that we need to re-search this block group
  * Return >0 to inform caller that we find nothing
  * Return 0 means we have found a location and set ffe_ctl->found_offset.
  */
@@ -3508,14 +3573,6 @@ refill_cluster:
 			trace_btrfs_reserve_extent_cluster(bg, ffe_ctl);
 			return 0;
 		}
-	} else if (!ffe_ctl->cached && ffe_ctl->loop > LOOP_CACHING_NOWAIT &&
-		   !ffe_ctl->retry_clustered) {
-		spin_unlock(&last_ptr->refill_lock);
-
-		ffe_ctl->retry_clustered = true;
-		btrfs_wait_block_group_cache_progress(bg, ffe_ctl->num_bytes +
-				ffe_ctl->empty_cluster + ffe_ctl->empty_size);
-		return -EAGAIN;
 	}
 	/*
 	 * At this point we either didn't find a cluster or we weren't able to
@@ -3530,7 +3587,6 @@ refill_cluster:
 /*
  * Return >0 to inform caller that we find nothing
  * Return 0 when we found an free extent and set ffe_ctrl->found_offset
- * Return -EAGAIN to inform caller that we need to re-search this block group
  */
 static int find_free_extent_unclustered(struct btrfs_block_group *bg,
 					struct find_free_extent_ctl *ffe_ctl)
@@ -3568,25 +3624,8 @@ static int find_free_extent_unclustered(struct btrfs_block_group *bg,
 	offset = btrfs_find_space_for_alloc(bg, ffe_ctl->search_start,
 			ffe_ctl->num_bytes, ffe_ctl->empty_size,
 			&ffe_ctl->max_extent_size);
-
-	/*
-	 * If we didn't find a chunk, and we haven't failed on this block group
-	 * before, and this block group is in the middle of caching and we are
-	 * ok with waiting, then go ahead and wait for progress to be made, and
-	 * set @retry_unclustered to true.
-	 *
-	 * If @retry_unclustered is true then we've already waited on this
-	 * block group once and should move on to the next block group.
-	 */
-	if (!offset && !ffe_ctl->retry_unclustered && !ffe_ctl->cached &&
-	    ffe_ctl->loop > LOOP_CACHING_NOWAIT) {
-		btrfs_wait_block_group_cache_progress(bg, ffe_ctl->num_bytes +
-						      ffe_ctl->empty_size);
-		ffe_ctl->retry_unclustered = true;
-		return -EAGAIN;
-	} else if (!offset) {
+	if (!offset)
 		return 1;
-	}
 	ffe_ctl->found_offset = offset;
 	return 0;
 }
@@ -3600,7 +3639,7 @@ static int do_allocation_clustered(struct btrfs_block_group *block_group,
 	/* We want to try and use the cluster allocator, so lets look there */
 	if (ffe_ctl->last_ptr && ffe_ctl->use_cluster) {
 		ret = find_free_extent_clustered(block_group, ffe_ctl, bg_ret);
-		if (ret >= 0 || ret == -EAGAIN)
+		if (ret >= 0)
 			return ret;
 		/* ret == -ENOENT case falls through */
 	}
@@ -3685,7 +3724,9 @@ static int do_allocation_zoned(struct btrfs_block_group *block_group,
 	}
 	spin_unlock(&block_group->lock);
 
-	if (!ret && !btrfs_zone_activate(block_group)) {
+	/* Metadata block group is activated at write time. */
+	if (!ret && (block_group->flags & BTRFS_BLOCK_GROUP_DATA) &&
+	    !btrfs_zone_activate(block_group)) {
 		ret = 1;
 		/*
 		 * May need to clear fs_info->{treelog,data_reloc}_bg.
@@ -3709,7 +3750,8 @@ static int do_allocation_zoned(struct btrfs_block_group *block_group,
 	       fs_info->data_reloc_bg == 0);
 
 	if (block_group->ro ||
-	    test_bit(BLOCK_GROUP_FLAG_ZONED_DATA_RELOC, &block_group->runtime_flags)) {
+	    (!ffe_ctl->for_data_reloc &&
+	     test_bit(BLOCK_GROUP_FLAG_ZONED_DATA_RELOC, &block_group->runtime_flags))) {
 		ret = 1;
 		goto out;
 	}
@@ -3752,8 +3794,26 @@ static int do_allocation_zoned(struct btrfs_block_group *block_group,
 	if (ffe_ctl->for_treelog && !fs_info->treelog_bg)
 		fs_info->treelog_bg = block_group->start;
 
-	if (ffe_ctl->for_data_reloc && !fs_info->data_reloc_bg)
-		fs_info->data_reloc_bg = block_group->start;
+	if (ffe_ctl->for_data_reloc) {
+		if (!fs_info->data_reloc_bg)
+			fs_info->data_reloc_bg = block_group->start;
+		/*
+		 * Do not allow allocations from this block group, unless it is
+		 * for data relocation. Compared to increasing the ->ro, setting
+		 * the ->zoned_data_reloc_ongoing flag still allows nocow
+		 * writers to come in. See btrfs_inc_nocow_writers().
+		 *
+		 * We need to disable an allocation to avoid an allocation of
+		 * regular (non-relocation data) extent. With mix of relocation
+		 * extents and regular extents, we can dispatch WRITE commands
+		 * (for relocation extents) and ZONE APPEND commands (for
+		 * regular extents) at the same time to the same zone, which
+		 * easily break the write pointer.
+		 *
+		 * Also, this flag avoids this block group to be zone finished.
+		 */
+		set_bit(BLOCK_GROUP_FLAG_ZONED_DATA_RELOC, &block_group->runtime_flags);
+	}
 
 	ffe_ctl->found_offset = start + block_group->alloc_offset;
 	block_group->alloc_offset += num_bytes;
@@ -3771,24 +3831,8 @@ static int do_allocation_zoned(struct btrfs_block_group *block_group,
 out:
 	if (ret && ffe_ctl->for_treelog)
 		fs_info->treelog_bg = 0;
-	if (ret && ffe_ctl->for_data_reloc &&
-	    fs_info->data_reloc_bg == block_group->start) {
-		/*
-		 * Do not allow further allocations from this block group.
-		 * Compared to increasing the ->ro, setting the
-		 * ->zoned_data_reloc_ongoing flag still allows nocow
-		 *  writers to come in. See btrfs_inc_nocow_writers().
-		 *
-		 * We need to disable an allocation to avoid an allocation of
-		 * regular (non-relocation data) extent. With mix of relocation
-		 * extents and regular extents, we can dispatch WRITE commands
-		 * (for relocation extents) and ZONE APPEND commands (for
-		 * regular extents) at the same time to the same zone, which
-		 * easily break the write pointer.
-		 */
-		set_bit(BLOCK_GROUP_FLAG_ZONED_DATA_RELOC, &block_group->runtime_flags);
+	if (ret && ffe_ctl->for_data_reloc)
 		fs_info->data_reloc_bg = 0;
-	}
 	spin_unlock(&fs_info->relocation_bg_lock);
 	spin_unlock(&fs_info->treelog_bg_lock);
 	spin_unlock(&block_group->lock);
@@ -3816,8 +3860,7 @@ static void release_block_group(struct btrfs_block_group *block_group,
 {
 	switch (ffe_ctl->policy) {
 	case BTRFS_EXTENT_ALLOC_CLUSTERED:
-		ffe_ctl->retry_clustered = false;
-		ffe_ctl->retry_unclustered = false;
+		ffe_ctl->retry_uncached = false;
 		break;
 	case BTRFS_EXTENT_ALLOC_ZONED:
 		/* Nothing to do */
@@ -3861,6 +3904,10 @@ static void found_extent(struct find_free_extent_ctl *ffe_ctl,
 static int can_allocate_chunk_zoned(struct btrfs_fs_info *fs_info,
 				    struct find_free_extent_ctl *ffe_ctl)
 {
+	/* Block group's activeness is not a requirement for METADATA block groups. */
+	if (!(ffe_ctl->flags & BTRFS_BLOCK_GROUP_DATA))
+		return 0;
+
 	/* If we can activate new zone, just allocate a chunk and use it */
 	if (btrfs_can_activate_zone(fs_info->fs_devices, ffe_ctl->flags))
 		return 0;
@@ -3949,15 +3996,7 @@ static int find_free_extent_update_loop(struct btrfs_fs_info *fs_info,
 	if (ffe_ctl->index < BTRFS_NR_RAID_TYPES)
 		return 1;
 
-	/*
-	 * LOOP_CACHING_NOWAIT, search partially cached block groups, kicking
-	 *			caching kthreads as we move along
-	 * LOOP_CACHING_WAIT, search everything, and wait if our bg is caching
-	 * LOOP_UNSET_SIZE_CLASS, allow unset size class
-	 * LOOP_ALLOC_CHUNK, force a chunk allocation and try again
-	 * LOOP_NO_EMPTY_SIZE, set empty_size and empty_cluster to 0 and try
-	 *		       again
-	 */
+	/* See the comments for btrfs_loop_type for an explanation of the phases. */
 	if (ffe_ctl->loop < LOOP_NO_EMPTY_SIZE) {
 		ffe_ctl->index = 0;
 		/*
@@ -4168,9 +4207,7 @@ static noinline int find_free_extent(struct btrfs_root *root,
 	ffe_ctl->orig_have_caching_bg = false;
 	ffe_ctl->index = btrfs_bg_flags_to_raid_index(ffe_ctl->flags);
 	ffe_ctl->loop = 0;
-	/* For clustered allocation */
-	ffe_ctl->retry_clustered = false;
-	ffe_ctl->retry_unclustered = false;
+	ffe_ctl->retry_uncached = false;
 	ffe_ctl->cached = 0;
 	ffe_ctl->max_extent_size = 0;
 	ffe_ctl->total_free_space = 0;
@@ -4321,16 +4358,12 @@ have_block_group:
 
 		bg_ret = NULL;
 		ret = do_allocation(block_group, ffe_ctl, &bg_ret);
-		if (ret == 0) {
-			if (bg_ret && bg_ret != block_group) {
-				btrfs_release_block_group(block_group,
-							  ffe_ctl->delalloc);
-				block_group = bg_ret;
-			}
-		} else if (ret == -EAGAIN) {
-			goto have_block_group;
-		} else if (ret > 0) {
+		if (ret > 0)
 			goto loop;
+
+		if (bg_ret && bg_ret != block_group) {
+			btrfs_release_block_group(block_group, ffe_ctl->delalloc);
+			block_group = bg_ret;
 		}
 
 		/* Checks */
@@ -4371,6 +4404,15 @@ have_block_group:
 		btrfs_release_block_group(block_group, ffe_ctl->delalloc);
 		break;
 loop:
+		if (!ffe_ctl->cached && ffe_ctl->loop > LOOP_CACHING_NOWAIT &&
+		    !ffe_ctl->retry_uncached) {
+			ffe_ctl->retry_uncached = true;
+			btrfs_wait_block_group_cache_progress(block_group,
+						ffe_ctl->num_bytes +
+						ffe_ctl->empty_cluster +
+						ffe_ctl->empty_size);
+			goto have_block_group;
+		}
 		release_block_group(block_group, ffe_ctl, ffe_ctl->delalloc);
 		cond_resched();
 	}
diff --git a/fs/btrfs/extent-tree.h b/fs/btrfs/extent-tree.h
index 429d5c570061..88c249c37516 100644
--- a/fs/btrfs/extent-tree.h
+++ b/fs/btrfs/extent-tree.h
@@ -48,16 +48,11 @@ struct find_free_extent_ctl {
 	int loop;
 
 	/*
-	 * Whether we're refilling a cluster, if true we need to re-search
-	 * current block group but don't try to refill the cluster again.
+	 * Set to true if we're retrying the allocation on this block group
+	 * after waiting for caching progress, this is so that we retry only
+	 * once before moving on to another block group.
 	 */
-	bool retry_clustered;
-
-	/*
-	 * Whether we're updating free space cache, if true we need to re-search
-	 * current block group but don't try updating free space cache again.
-	 */
-	bool retry_unclustered;
+	bool retry_uncached;
 
 	/* If current block group is cached */
 	int cached;
@@ -96,9 +91,6 @@ int btrfs_get_extent_inline_ref_type(const struct extent_buffer *eb,
 				     enum btrfs_inline_ref_type is_data);
 u64 hash_extent_data_ref(u64 root_objectid, u64 owner, u64 offset);
 
-int btrfs_add_excluded_extent(struct btrfs_fs_info *fs_info,
-			      u64 start, u64 num_bytes);
-void btrfs_free_excluded_extents(struct btrfs_block_group *cache);
 int btrfs_run_delayed_refs(struct btrfs_trans_handle *trans, unsigned long count);
 void btrfs_cleanup_ref_head_accounting(struct btrfs_fs_info *fs_info,
 				  struct btrfs_delayed_ref_root *delayed_refs,
diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c
index 90ad3006ef3a..ac3fca5a5e41 100644
--- a/fs/btrfs/extent_io.c
+++ b/fs/btrfs/extent_io.c
@@ -181,34 +181,9 @@ void extent_range_clear_dirty_for_io(struct inode *inode, u64 start, u64 end)
 	}
 }
 
-void extent_range_redirty_for_io(struct inode *inode, u64 start, u64 end)
-{
-	struct address_space *mapping = inode->i_mapping;
-	unsigned long index = start >> PAGE_SHIFT;
-	unsigned long end_index = end >> PAGE_SHIFT;
-	struct folio *folio;
-
-	while (index <= end_index) {
-		folio = filemap_get_folio(mapping, index);
-		filemap_dirty_folio(mapping, folio);
-		folio_account_redirty(folio);
-		index += folio_nr_pages(folio);
-		folio_put(folio);
-	}
-}
-
-/*
- * Process one page for __process_pages_contig().
- *
- * Return >0 if we hit @page == @locked_page.
- * Return 0 if we updated the page status.
- * Return -EGAIN if the we need to try again.
- * (For PAGE_LOCK case but got dirty page or page not belong to mapping)
- */
-static int process_one_page(struct btrfs_fs_info *fs_info,
-			    struct address_space *mapping,
-			    struct page *page, struct page *locked_page,
-			    unsigned long page_ops, u64 start, u64 end)
+static void process_one_page(struct btrfs_fs_info *fs_info,
+			     struct page *page, struct page *locked_page,
+			     unsigned long page_ops, u64 start, u64 end)
 {
 	u32 len;
 
@@ -224,94 +199,36 @@ static int process_one_page(struct btrfs_fs_info *fs_info,
 	if (page_ops & PAGE_END_WRITEBACK)
 		btrfs_page_clamp_clear_writeback(fs_info, page, start, len);
 
-	if (page == locked_page)
-		return 1;
-
-	if (page_ops & PAGE_LOCK) {
-		int ret;
-
-		ret = btrfs_page_start_writer_lock(fs_info, page, start, len);
-		if (ret)
-			return ret;
-		if (!PageDirty(page) || page->mapping != mapping) {
-			btrfs_page_end_writer_lock(fs_info, page, start, len);
-			return -EAGAIN;
-		}
-	}
-	if (page_ops & PAGE_UNLOCK)
+	if (page != locked_page && (page_ops & PAGE_UNLOCK))
 		btrfs_page_end_writer_lock(fs_info, page, start, len);
-	return 0;
 }
 
-static int __process_pages_contig(struct address_space *mapping,
-				  struct page *locked_page,
-				  u64 start, u64 end, unsigned long page_ops,
-				  u64 *processed_end)
+static void __process_pages_contig(struct address_space *mapping,
+				   struct page *locked_page, u64 start, u64 end,
+				   unsigned long page_ops)
 {
 	struct btrfs_fs_info *fs_info = btrfs_sb(mapping->host->i_sb);
 	pgoff_t start_index = start >> PAGE_SHIFT;
 	pgoff_t end_index = end >> PAGE_SHIFT;
 	pgoff_t index = start_index;
-	unsigned long pages_processed = 0;
 	struct folio_batch fbatch;
-	int err = 0;
 	int i;
 
-	if (page_ops & PAGE_LOCK) {
-		ASSERT(page_ops == PAGE_LOCK);
-		ASSERT(processed_end && *processed_end == start);
-	}
-
 	folio_batch_init(&fbatch);
 	while (index <= end_index) {
 		int found_folios;
 
 		found_folios = filemap_get_folios_contig(mapping, &index,
 				end_index, &fbatch);
-
-		if (found_folios == 0) {
-			/*
-			 * Only if we're going to lock these pages, we can find
-			 * nothing at @index.
-			 */
-			ASSERT(page_ops & PAGE_LOCK);
-			err = -EAGAIN;
-			goto out;
-		}
-
 		for (i = 0; i < found_folios; i++) {
-			int process_ret;
 			struct folio *folio = fbatch.folios[i];
-			process_ret = process_one_page(fs_info, mapping,
-					&folio->page, locked_page, page_ops,
-					start, end);
-			if (process_ret < 0) {
-				err = -EAGAIN;
-				folio_batch_release(&fbatch);
-				goto out;
-			}
-			pages_processed += folio_nr_pages(folio);
+
+			process_one_page(fs_info, &folio->page, locked_page,
+					 page_ops, start, end);
 		}
 		folio_batch_release(&fbatch);
 		cond_resched();
 	}
-out:
-	if (err && processed_end) {
-		/*
-		 * Update @processed_end. I know this is awful since it has
-		 * two different return value patterns (inclusive vs exclusive).
-		 *
-		 * But the exclusive pattern is necessary if @start is 0, or we
-		 * underflow and check against processed_end won't work as
-		 * expected.
-		 */
-		if (pages_processed)
-			*processed_end = min(end,
-			((u64)(start_index + pages_processed) << PAGE_SHIFT) - 1);
-		else
-			*processed_end = start;
-	}
-	return err;
 }
 
 static noinline void __unlock_for_delalloc(struct inode *inode,
@@ -326,29 +243,63 @@ static noinline void __unlock_for_delalloc(struct inode *inode,
 		return;
 
 	__process_pages_contig(inode->i_mapping, locked_page, start, end,
-			       PAGE_UNLOCK, NULL);
+			       PAGE_UNLOCK);
 }
 
 static noinline int lock_delalloc_pages(struct inode *inode,
 					struct page *locked_page,
-					u64 delalloc_start,
-					u64 delalloc_end)
+					u64 start,
+					u64 end)
 {
-	unsigned long index = delalloc_start >> PAGE_SHIFT;
-	unsigned long end_index = delalloc_end >> PAGE_SHIFT;
-	u64 processed_end = delalloc_start;
-	int ret;
+	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
+	struct address_space *mapping = inode->i_mapping;
+	pgoff_t start_index = start >> PAGE_SHIFT;
+	pgoff_t end_index = end >> PAGE_SHIFT;
+	pgoff_t index = start_index;
+	u64 processed_end = start;
+	struct folio_batch fbatch;
 
-	ASSERT(locked_page);
 	if (index == locked_page->index && index == end_index)
 		return 0;
 
-	ret = __process_pages_contig(inode->i_mapping, locked_page, delalloc_start,
-				     delalloc_end, PAGE_LOCK, &processed_end);
-	if (ret == -EAGAIN && processed_end > delalloc_start)
-		__unlock_for_delalloc(inode, locked_page, delalloc_start,
-				      processed_end);
-	return ret;
+	folio_batch_init(&fbatch);
+	while (index <= end_index) {
+		unsigned int found_folios, i;
+
+		found_folios = filemap_get_folios_contig(mapping, &index,
+				end_index, &fbatch);
+		if (found_folios == 0)
+			goto out;
+
+		for (i = 0; i < found_folios; i++) {
+			struct page *page = &fbatch.folios[i]->page;
+			u32 len = end + 1 - start;
+
+			if (page == locked_page)
+				continue;
+
+			if (btrfs_page_start_writer_lock(fs_info, page, start,
+							 len))
+				goto out;
+
+			if (!PageDirty(page) || page->mapping != mapping) {
+				btrfs_page_end_writer_lock(fs_info, page, start,
+							   len);
+				goto out;
+			}
+
+			processed_end = page_offset(page) + PAGE_SIZE - 1;
+		}
+		folio_batch_release(&fbatch);
+		cond_resched();
+	}
+
+	return 0;
+out:
+	folio_batch_release(&fbatch);
+	if (processed_end > start)
+		__unlock_for_delalloc(inode, locked_page, start, processed_end);
+	return -EAGAIN;
 }
 
 /*
@@ -467,7 +418,7 @@ void extent_clear_unlock_delalloc(struct btrfs_inode *inode, u64 start, u64 end,
 	clear_extent_bit(&inode->io_tree, start, end, clear_bits, NULL);
 
 	__process_pages_contig(inode->vfs_inode.i_mapping, locked_page,
-			       start, end, page_ops, NULL);
+			       start, end, page_ops);
 }
 
 static bool btrfs_verify_page(struct page *page, u64 start)
@@ -497,31 +448,6 @@ static void end_page_read(struct page *page, bool uptodate, u64 start, u32 len)
 		btrfs_subpage_end_reader(fs_info, page, start, len);
 }
 
-/* lots and lots of room for performance fixes in the end_bio funcs */
-
-void end_extent_writepage(struct page *page, int err, u64 start, u64 end)
-{
-	struct btrfs_inode *inode;
-	const bool uptodate = (err == 0);
-	int ret = 0;
-
-	ASSERT(page && page->mapping);
-	inode = BTRFS_I(page->mapping->host);
-	btrfs_writepage_endio_finish_ordered(inode, page, start, end, uptodate);
-
-	if (!uptodate) {
-		const struct btrfs_fs_info *fs_info = inode->root->fs_info;
-		u32 len;
-
-		ASSERT(end + 1 - start <= U32_MAX);
-		len = end + 1 - start;
-
-		btrfs_page_clear_uptodate(fs_info, page, start, len);
-		ret = err < 0 ? err : -EIO;
-		mapping_set_error(page->mapping, ret);
-	}
-}
-
 /*
  * after a writepage IO is done, we need to:
  * clear the uptodate bits on error
@@ -1243,38 +1169,45 @@ static inline void contiguous_readpages(struct page *pages[], int nr_pages,
 static noinline_for_stack int writepage_delalloc(struct btrfs_inode *inode,
 		struct page *page, struct writeback_control *wbc)
 {
-	const u64 page_end = page_offset(page) + PAGE_SIZE - 1;
-	u64 delalloc_start = page_offset(page);
+	const u64 page_start = page_offset(page);
+	const u64 page_end = page_start + PAGE_SIZE - 1;
+	u64 delalloc_start = page_start;
+	u64 delalloc_end = page_end;
 	u64 delalloc_to_write = 0;
-	/* How many pages are started by btrfs_run_delalloc_range() */
-	unsigned long nr_written = 0;
-	int ret;
-	int page_started = 0;
+	int ret = 0;
 
 	while (delalloc_start < page_end) {
-		u64 delalloc_end = page_end;
-		bool found;
-
-		found = find_lock_delalloc_range(&inode->vfs_inode, page,
-					       &delalloc_start,
-					       &delalloc_end);
-		if (!found) {
+		delalloc_end = page_end;
+		if (!find_lock_delalloc_range(&inode->vfs_inode, page,
+					      &delalloc_start, &delalloc_end)) {
 			delalloc_start = delalloc_end + 1;
 			continue;
 		}
+
 		ret = btrfs_run_delalloc_range(inode, page, delalloc_start,
-				delalloc_end, &page_started, &nr_written, wbc);
-		if (ret)
+					       delalloc_end, wbc);
+		if (ret < 0)
 			return ret;
 
-		/*
-		 * delalloc_end is already one less than the total length, so
-		 * we don't subtract one from PAGE_SIZE
-		 */
-		delalloc_to_write += (delalloc_end - delalloc_start +
-				      PAGE_SIZE) >> PAGE_SHIFT;
 		delalloc_start = delalloc_end + 1;
 	}
+
+	/*
+	 * delalloc_end is already one less than the total length, so
+	 * we don't subtract one from PAGE_SIZE
+	 */
+	delalloc_to_write +=
+		DIV_ROUND_UP(delalloc_end + 1 - page_start, PAGE_SIZE);
+
+	/*
+	 * If btrfs_run_dealloc_range() already started I/O and unlocked
+	 * the pages, we just need to account for them here.
+	 */
+	if (ret == 1) {
+		wbc->nr_to_write -= delalloc_to_write;
+		return 1;
+	}
+
 	if (wbc->nr_to_write < delalloc_to_write) {
 		int thresh = 8192;
 
@@ -1284,16 +1217,6 @@ static noinline_for_stack int writepage_delalloc(struct btrfs_inode *inode,
 					 thresh);
 	}
 
-	/* Did btrfs_run_dealloc_range() already unlock and start the IO? */
-	if (page_started) {
-		/*
-		 * We've unlocked the page, so we can't update the mapping's
-		 * writeback index, just update nr_to_write.
-		 */
-		wbc->nr_to_write -= nr_written;
-		return 1;
-	}
-
 	return 0;
 }
 
@@ -1382,6 +1305,7 @@ static noinline_for_stack int __extent_writepage_io(struct btrfs_inode *inode,
 
 	bio_ctrl->end_io_func = end_bio_extent_writepage;
 	while (cur <= end) {
+		u32 len = end - cur + 1;
 		u64 disk_bytenr;
 		u64 em_end;
 		u64 dirty_range_start = cur;
@@ -1389,8 +1313,8 @@ static noinline_for_stack int __extent_writepage_io(struct btrfs_inode *inode,
 		u32 iosize;
 
 		if (cur >= i_size) {
-			btrfs_writepage_endio_finish_ordered(inode, page, cur,
-							     end, true);
+			btrfs_mark_ordered_io_finished(inode, page, cur, len,
+						       true);
 			/*
 			 * This range is beyond i_size, thus we don't need to
 			 * bother writing back.
@@ -1399,7 +1323,7 @@ static noinline_for_stack int __extent_writepage_io(struct btrfs_inode *inode,
 			 * writeback the sectors with subpage dirty bits,
 			 * causing writeback without ordered extent.
 			 */
-			btrfs_page_clear_dirty(fs_info, page, cur, end + 1 - cur);
+			btrfs_page_clear_dirty(fs_info, page, cur, len);
 			break;
 		}
 
@@ -1410,7 +1334,7 @@ static noinline_for_stack int __extent_writepage_io(struct btrfs_inode *inode,
 			continue;
 		}
 
-		em = btrfs_get_extent(inode, NULL, 0, cur, end - cur + 1);
+		em = btrfs_get_extent(inode, NULL, 0, cur, len);
 		if (IS_ERR(em)) {
 			ret = PTR_ERR_OR_ZERO(em);
 			goto out_error;
@@ -1486,7 +1410,6 @@ static int __extent_writepage(struct page *page, struct btrfs_bio_ctrl *bio_ctrl
 	struct folio *folio = page_folio(page);
 	struct inode *inode = page->mapping->host;
 	const u64 page_start = page_offset(page);
-	const u64 page_end = page_start + PAGE_SIZE - 1;
 	int ret;
 	int nr = 0;
 	size_t pg_offset;
@@ -1530,8 +1453,13 @@ done:
 		set_page_writeback(page);
 		end_page_writeback(page);
 	}
-	if (ret)
-		end_extent_writepage(page, ret, page_start, page_end);
+	if (ret) {
+		btrfs_mark_ordered_io_finished(BTRFS_I(inode), page, page_start,
+					       PAGE_SIZE, !ret);
+		btrfs_page_clear_uptodate(btrfs_sb(inode->i_sb), page,
+					  page_start, PAGE_SIZE);
+		mapping_set_error(page->mapping, ret);
+	}
 	unlock_page(page);
 	ASSERT(ret <= 0);
 	return ret;
@@ -1877,11 +1805,10 @@ static int submit_eb_subpage(struct page *page, struct writeback_control *wbc)
  * previous call.
  * Return <0 for fatal error.
  */
-static int submit_eb_page(struct page *page, struct writeback_control *wbc,
-			  struct extent_buffer **eb_context)
+static int submit_eb_page(struct page *page, struct btrfs_eb_write_context *ctx)
 {
+	struct writeback_control *wbc = ctx->wbc;
 	struct address_space *mapping = page->mapping;
-	struct btrfs_block_group *cache = NULL;
 	struct extent_buffer *eb;
 	int ret;
 
@@ -1908,7 +1835,7 @@ static int submit_eb_page(struct page *page, struct writeback_control *wbc,
 		return 0;
 	}
 
-	if (eb == *eb_context) {
+	if (eb == ctx->eb) {
 		spin_unlock(&mapping->private_lock);
 		return 0;
 	}
@@ -1917,34 +1844,25 @@ static int submit_eb_page(struct page *page, struct writeback_control *wbc,
 	if (!ret)
 		return 0;
 
-	if (!btrfs_check_meta_write_pointer(eb->fs_info, eb, &cache)) {
-		/*
-		 * If for_sync, this hole will be filled with
-		 * trasnsaction commit.
-		 */
-		if (wbc->sync_mode == WB_SYNC_ALL && !wbc->for_sync)
-			ret = -EAGAIN;
-		else
+	ctx->eb = eb;
+
+	ret = btrfs_check_meta_write_pointer(eb->fs_info, ctx);
+	if (ret) {
+		if (ret == -EBUSY)
 			ret = 0;
 		free_extent_buffer(eb);
 		return ret;
 	}
 
-	*eb_context = eb;
-
 	if (!lock_extent_buffer_for_io(eb, wbc)) {
-		btrfs_revert_meta_write_pointer(cache, eb);
-		if (cache)
-			btrfs_put_block_group(cache);
 		free_extent_buffer(eb);
 		return 0;
 	}
-	if (cache) {
-		/*
-		 * Implies write in zoned mode. Mark the last eb in a block group.
-		 */
-		btrfs_schedule_zone_finish_bg(cache, eb);
-		btrfs_put_block_group(cache);
+	/* Implies write in zoned mode. */
+	if (ctx->zoned_bg) {
+		/* Mark the last eb in the block group. */
+		btrfs_schedule_zone_finish_bg(ctx->zoned_bg, eb);
+		ctx->zoned_bg->meta_write_pointer += eb->len;
 	}
 	write_one_eb(eb, wbc);
 	free_extent_buffer(eb);
@@ -1954,7 +1872,7 @@ static int submit_eb_page(struct page *page, struct writeback_control *wbc,
 int btree_write_cache_pages(struct address_space *mapping,
 				   struct writeback_control *wbc)
 {
-	struct extent_buffer *eb_context = NULL;
+	struct btrfs_eb_write_context ctx = { .wbc = wbc };
 	struct btrfs_fs_info *fs_info = BTRFS_I(mapping->host)->root->fs_info;
 	int ret = 0;
 	int done = 0;
@@ -1996,7 +1914,7 @@ retry:
 		for (i = 0; i < nr_folios; i++) {
 			struct folio *folio = fbatch.folios[i];
 
-			ret = submit_eb_page(&folio->page, wbc, &eb_context);
+			ret = submit_eb_page(&folio->page, &ctx);
 			if (ret == 0)
 				continue;
 			if (ret < 0) {
@@ -2057,6 +1975,9 @@ retry:
 		ret = 0;
 	if (!ret && BTRFS_FS_ERROR(fs_info))
 		ret = -EROFS;
+
+	if (ctx.zoned_bg)
+		btrfs_put_block_group(ctx.zoned_bg);
 	btrfs_zoned_meta_io_unlock(fs_info);
 	return ret;
 }
@@ -2150,7 +2071,7 @@ retry:
 		for (i = 0; i < nr_folios; i++) {
 			struct folio *folio = fbatch.folios[i];
 
-			done_index = folio->index + folio_nr_pages(folio);
+			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
@@ -2233,11 +2154,11 @@ retry:
  * already been ran (aka, ordered extent inserted) and all pages are still
  * locked.
  */
-int extent_write_locked_range(struct inode *inode, u64 start, u64 end,
-			      struct writeback_control *wbc)
+void extent_write_locked_range(struct inode *inode, struct page *locked_page,
+			       u64 start, u64 end, struct writeback_control *wbc,
+			       bool pages_dirty)
 {
 	bool found_error = false;
-	int first_error = 0;
 	int ret = 0;
 	struct address_space *mapping = inode->i_mapping;
 	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
@@ -2256,18 +2177,16 @@ int extent_write_locked_range(struct inode *inode, u64 start, u64 end,
 
 	while (cur <= end) {
 		u64 cur_end = min(round_down(cur, PAGE_SIZE) + PAGE_SIZE - 1, end);
+		u32 cur_len = cur_end + 1 - cur;
 		struct page *page;
 		int nr = 0;
 
 		page = find_get_page(mapping, cur >> PAGE_SHIFT);
-		/*
-		 * All pages in the range are locked since
-		 * btrfs_run_delalloc_range(), thus there is no way to clear
-		 * the page dirty flag.
-		 */
 		ASSERT(PageLocked(page));
-		ASSERT(PageDirty(page));
-		clear_page_dirty_for_io(page);
+		if (pages_dirty && page != locked_page) {
+			ASSERT(PageDirty(page));
+			clear_page_dirty_for_io(page);
+		}
 
 		ret = __extent_writepage_io(BTRFS_I(inode), page, &bio_ctrl,
 					    i_size, &nr);
@@ -2279,23 +2198,21 @@ int extent_write_locked_range(struct inode *inode, u64 start, u64 end,
 			set_page_writeback(page);
 			end_page_writeback(page);
 		}
-		if (ret)
-			end_extent_writepage(page, ret, cur, cur_end);
-		btrfs_page_unlock_writer(fs_info, page, cur, cur_end + 1 - cur);
-		if (ret < 0) {
-			found_error = true;
-			first_error = ret;
+		if (ret) {
+			btrfs_mark_ordered_io_finished(BTRFS_I(inode), page,
+						       cur, cur_len, !ret);
+			btrfs_page_clear_uptodate(fs_info, page, cur, cur_len);
+			mapping_set_error(page->mapping, ret);
 		}
+		btrfs_page_unlock_writer(fs_info, page, cur, cur_len);
+		if (ret < 0)
+			found_error = true;
 next_page:
 		put_page(page);
 		cur = cur_end + 1;
 	}
 
 	submit_write_bio(&bio_ctrl, found_error ? ret : 0);
-
-	if (found_error)
-		return first_error;
-	return ret;
 }
 
 int extent_writepages(struct address_space *mapping,
@@ -3315,8 +3232,8 @@ struct extent_buffer *btrfs_clone_extent_buffer(const struct extent_buffer *src)
 			return NULL;
 		}
 		WARN_ON(PageDirty(p));
-		copy_page(page_address(p), page_address(src->pages[i]));
 	}
+	copy_extent_buffer_full(new, src);
 	set_extent_buffer_uptodate(new);
 
 	return new;
@@ -3559,6 +3476,7 @@ struct extent_buffer *alloc_extent_buffer(struct btrfs_fs_info *fs_info,
 	struct extent_buffer *exists = NULL;
 	struct page *p;
 	struct address_space *mapping = fs_info->btree_inode->i_mapping;
+	struct btrfs_subpage *prealloc = NULL;
 	u64 lockdep_owner = owner_root;
 	int uptodate = 1;
 	int ret;
@@ -3595,36 +3513,30 @@ struct extent_buffer *alloc_extent_buffer(struct btrfs_fs_info *fs_info,
 	btrfs_set_buffer_lockdep_class(lockdep_owner, eb, level);
 
 	num_pages = num_extent_pages(eb);
-	for (i = 0; i < num_pages; i++, index++) {
-		struct btrfs_subpage *prealloc = NULL;
 
+	/*
+	 * Preallocate page->private for subpage case, so that we won't
+	 * allocate memory with private_lock nor page lock hold.
+	 *
+	 * 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 (IS_ERR(prealloc)) {
+			exists = ERR_CAST(prealloc);
+			goto free_eb;
+		}
+	}
+
+	for (i = 0; i < num_pages; i++, index++) {
 		p = find_or_create_page(mapping, index, GFP_NOFS|__GFP_NOFAIL);
 		if (!p) {
 			exists = ERR_PTR(-ENOMEM);
+			btrfs_free_subpage(prealloc);
 			goto free_eb;
 		}
 
-		/*
-		 * Preallocate page->private for subpage case, so that we won't
-		 * allocate memory with private_lock hold.  The memory will be
-		 * freed by attach_extent_buffer_page() or freed manually if
-		 * we exit earlier.
-		 *
-		 * Although we have ensured one subpage eb can only have one
-		 * page, but it may change in the future for 16K page size
-		 * support, so we still preallocate the memory in the loop.
-		 */
-		if (fs_info->nodesize < PAGE_SIZE) {
-			prealloc = btrfs_alloc_subpage(fs_info, BTRFS_SUBPAGE_METADATA);
-			if (IS_ERR(prealloc)) {
-				ret = PTR_ERR(prealloc);
-				unlock_page(p);
-				put_page(p);
-				exists = ERR_PTR(ret);
-				goto free_eb;
-			}
-		}
-
 		spin_lock(&mapping->private_lock);
 		exists = grab_extent_buffer(fs_info, p);
 		if (exists) {
@@ -4210,30 +4122,9 @@ static void assert_eb_page_uptodate(const struct extent_buffer *eb,
 	}
 }
 
-void write_extent_buffer_chunk_tree_uuid(const struct extent_buffer *eb,
-		const void *srcv)
-{
-	char *kaddr;
-
-	assert_eb_page_uptodate(eb, eb->pages[0]);
-	kaddr = page_address(eb->pages[0]) +
-		get_eb_offset_in_page(eb, offsetof(struct btrfs_header,
-						   chunk_tree_uuid));
-	memcpy(kaddr, srcv, BTRFS_FSID_SIZE);
-}
-
-void write_extent_buffer_fsid(const struct extent_buffer *eb, const void *srcv)
-{
-	char *kaddr;
-
-	assert_eb_page_uptodate(eb, eb->pages[0]);
-	kaddr = page_address(eb->pages[0]) +
-		get_eb_offset_in_page(eb, offsetof(struct btrfs_header, fsid));
-	memcpy(kaddr, srcv, BTRFS_FSID_SIZE);
-}
-
-void write_extent_buffer(const struct extent_buffer *eb, const void *srcv,
-			 unsigned long start, unsigned long len)
+static void __write_extent_buffer(const struct extent_buffer *eb,
+				  const void *srcv, unsigned long start,
+				  unsigned long len, bool use_memmove)
 {
 	size_t cur;
 	size_t offset;
@@ -4241,6 +4132,8 @@ void write_extent_buffer(const struct extent_buffer *eb, const void *srcv,
 	char *kaddr;
 	char *src = (char *)srcv;
 	unsigned long i = get_eb_page_index(start);
+	/* For unmapped (dummy) ebs, no need to check their uptodate status. */
+	const bool check_uptodate = !test_bit(EXTENT_BUFFER_UNMAPPED, &eb->bflags);
 
 	WARN_ON(test_bit(EXTENT_BUFFER_NO_CHECK, &eb->bflags));
 
@@ -4251,11 +4144,15 @@ void write_extent_buffer(const struct extent_buffer *eb, const void *srcv,
 
 	while (len > 0) {
 		page = eb->pages[i];
-		assert_eb_page_uptodate(eb, page);
+		if (check_uptodate)
+			assert_eb_page_uptodate(eb, page);
 
 		cur = min(len, PAGE_SIZE - offset);
 		kaddr = page_address(page);
-		memcpy(kaddr + offset, src, cur);
+		if (use_memmove)
+			memmove(kaddr + offset, src, cur);
+		else
+			memcpy(kaddr + offset, src, cur);
 
 		src += cur;
 		len -= cur;
@@ -4264,55 +4161,54 @@ void write_extent_buffer(const struct extent_buffer *eb, const void *srcv,
 	}
 }
 
-void memzero_extent_buffer(const struct extent_buffer *eb, unsigned long start,
-		unsigned long len)
+void write_extent_buffer(const struct extent_buffer *eb, const void *srcv,
+			 unsigned long start, unsigned long len)
 {
-	size_t cur;
-	size_t offset;
-	struct page *page;
-	char *kaddr;
-	unsigned long i = get_eb_page_index(start);
+	return __write_extent_buffer(eb, srcv, start, len, false);
+}
 
-	if (check_eb_range(eb, start, len))
-		return;
+static void memset_extent_buffer(const struct extent_buffer *eb, int c,
+				 unsigned long start, unsigned long len)
+{
+	unsigned long cur = start;
 
-	offset = get_eb_offset_in_page(eb, start);
+	while (cur < start + len) {
+		unsigned long index = get_eb_page_index(cur);
+		unsigned int offset = get_eb_offset_in_page(eb, cur);
+		unsigned int cur_len = min(start + len - cur, PAGE_SIZE - offset);
+		struct page *page = eb->pages[index];
 
-	while (len > 0) {
-		page = eb->pages[i];
 		assert_eb_page_uptodate(eb, page);
+		memset(page_address(page) + offset, c, cur_len);
 
-		cur = min(len, PAGE_SIZE - offset);
-		kaddr = page_address(page);
-		memset(kaddr + offset, 0, cur);
-
-		len -= cur;
-		offset = 0;
-		i++;
+		cur += cur_len;
 	}
 }
 
+void memzero_extent_buffer(const struct extent_buffer *eb, unsigned long start,
+			   unsigned long len)
+{
+	if (check_eb_range(eb, start, len))
+		return;
+	return memset_extent_buffer(eb, 0, start, len);
+}
+
 void copy_extent_buffer_full(const struct extent_buffer *dst,
 			     const struct extent_buffer *src)
 {
-	int i;
-	int num_pages;
+	unsigned long cur = 0;
 
 	ASSERT(dst->len == src->len);
 
-	if (dst->fs_info->nodesize >= PAGE_SIZE) {
-		num_pages = num_extent_pages(dst);
-		for (i = 0; i < num_pages; i++)
-			copy_page(page_address(dst->pages[i]),
-				  page_address(src->pages[i]));
-	} else {
-		size_t src_offset = get_eb_offset_in_page(src, 0);
-		size_t dst_offset = get_eb_offset_in_page(dst, 0);
+	while (cur < src->len) {
+		unsigned long index = get_eb_page_index(cur);
+		unsigned long offset = get_eb_offset_in_page(src, cur);
+		unsigned long cur_len = min(src->len, PAGE_SIZE - offset);
+		void *addr = page_address(src->pages[index]) + offset;
+
+		write_extent_buffer(dst, addr, cur, cur_len);
 
-		ASSERT(src->fs_info->nodesize < PAGE_SIZE);
-		memcpy(page_address(dst->pages[0]) + dst_offset,
-		       page_address(src->pages[0]) + src_offset,
-		       src->len);
+		cur += cur_len;
 	}
 }
 
@@ -4406,6 +4302,15 @@ int extent_buffer_test_bit(const struct extent_buffer *eb, unsigned long start,
 	return 1U & (kaddr[offset] >> (nr & (BITS_PER_BYTE - 1)));
 }
 
+static u8 *extent_buffer_get_byte(const struct extent_buffer *eb, unsigned long bytenr)
+{
+	unsigned long index = get_eb_page_index(bytenr);
+
+	if (check_eb_range(eb, bytenr, 1))
+		return NULL;
+	return page_address(eb->pages[index]) + get_eb_offset_in_page(eb, bytenr);
+}
+
 /*
  * Set an area of a bitmap to 1.
  *
@@ -4417,35 +4322,28 @@ int extent_buffer_test_bit(const struct extent_buffer *eb, unsigned long start,
 void extent_buffer_bitmap_set(const struct extent_buffer *eb, unsigned long start,
 			      unsigned long pos, unsigned long len)
 {
+	unsigned int first_byte = start + BIT_BYTE(pos);
+	unsigned int last_byte = start + BIT_BYTE(pos + len - 1);
+	const bool same_byte = (first_byte == last_byte);
+	u8 mask = BITMAP_FIRST_BYTE_MASK(pos);
 	u8 *kaddr;
-	struct page *page;
-	unsigned long i;
-	size_t offset;
-	const unsigned int size = pos + len;
-	int bits_to_set = BITS_PER_BYTE - (pos % BITS_PER_BYTE);
-	u8 mask_to_set = BITMAP_FIRST_BYTE_MASK(pos);
 
-	eb_bitmap_offset(eb, start, pos, &i, &offset);
-	page = eb->pages[i];
-	assert_eb_page_uptodate(eb, page);
-	kaddr = page_address(page);
+	if (same_byte)
+		mask &= BITMAP_LAST_BYTE_MASK(pos + len);
 
-	while (len >= bits_to_set) {
-		kaddr[offset] |= mask_to_set;
-		len -= bits_to_set;
-		bits_to_set = BITS_PER_BYTE;
-		mask_to_set = ~0;
-		if (++offset >= PAGE_SIZE && len > 0) {
-			offset = 0;
-			page = eb->pages[++i];
-			assert_eb_page_uptodate(eb, page);
-			kaddr = page_address(page);
-		}
-	}
-	if (len) {
-		mask_to_set &= BITMAP_LAST_BYTE_MASK(size);
-		kaddr[offset] |= mask_to_set;
-	}
+	/* Handle the first byte. */
+	kaddr = extent_buffer_get_byte(eb, first_byte);
+	*kaddr |= mask;
+	if (same_byte)
+		return;
+
+	/* Handle the byte aligned part. */
+	ASSERT(first_byte + 1 <= last_byte);
+	memset_extent_buffer(eb, 0xff, first_byte + 1, last_byte - first_byte - 1);
+
+	/* Handle the last byte. */
+	kaddr = extent_buffer_get_byte(eb, last_byte);
+	*kaddr |= BITMAP_LAST_BYTE_MASK(pos + len);
 }
 
 
@@ -4461,35 +4359,28 @@ void extent_buffer_bitmap_clear(const struct extent_buffer *eb,
 				unsigned long start, unsigned long pos,
 				unsigned long len)
 {
+	unsigned int first_byte = start + BIT_BYTE(pos);
+	unsigned int last_byte = start + BIT_BYTE(pos + len - 1);
+	const bool same_byte = (first_byte == last_byte);
+	u8 mask = BITMAP_FIRST_BYTE_MASK(pos);
 	u8 *kaddr;
-	struct page *page;
-	unsigned long i;
-	size_t offset;
-	const unsigned int size = pos + len;
-	int bits_to_clear = BITS_PER_BYTE - (pos % BITS_PER_BYTE);
-	u8 mask_to_clear = BITMAP_FIRST_BYTE_MASK(pos);
 
-	eb_bitmap_offset(eb, start, pos, &i, &offset);
-	page = eb->pages[i];
-	assert_eb_page_uptodate(eb, page);
-	kaddr = page_address(page);
+	if (same_byte)
+		mask &= BITMAP_LAST_BYTE_MASK(pos + len);
 
-	while (len >= bits_to_clear) {
-		kaddr[offset] &= ~mask_to_clear;
-		len -= bits_to_clear;
-		bits_to_clear = BITS_PER_BYTE;
-		mask_to_clear = ~0;
-		if (++offset >= PAGE_SIZE && len > 0) {
-			offset = 0;
-			page = eb->pages[++i];
-			assert_eb_page_uptodate(eb, page);
-			kaddr = page_address(page);
-		}
-	}
-	if (len) {
-		mask_to_clear &= BITMAP_LAST_BYTE_MASK(size);
-		kaddr[offset] &= ~mask_to_clear;
-	}
+	/* Handle the first byte. */
+	kaddr = extent_buffer_get_byte(eb, first_byte);
+	*kaddr &= ~mask;
+	if (same_byte)
+		return;
+
+	/* Handle the byte aligned part. */
+	ASSERT(first_byte + 1 <= last_byte);
+	memset_extent_buffer(eb, 0, first_byte + 1, last_byte - first_byte - 1);
+
+	/* Handle the last byte. */
+	kaddr = extent_buffer_get_byte(eb, last_byte);
+	*kaddr &= ~BITMAP_LAST_BYTE_MASK(pos + len);
 }
 
 static inline bool areas_overlap(unsigned long src, unsigned long dst, unsigned long len)
@@ -4498,60 +4389,29 @@ static inline bool areas_overlap(unsigned long src, unsigned long dst, unsigned
 	return distance < len;
 }
 
-static void copy_pages(struct page *dst_page, struct page *src_page,
-		       unsigned long dst_off, unsigned long src_off,
-		       unsigned long len)
-{
-	char *dst_kaddr = page_address(dst_page);
-	char *src_kaddr;
-	int must_memmove = 0;
-
-	if (dst_page != src_page) {
-		src_kaddr = page_address(src_page);
-	} else {
-		src_kaddr = dst_kaddr;
-		if (areas_overlap(src_off, dst_off, len))
-			must_memmove = 1;
-	}
-
-	if (must_memmove)
-		memmove(dst_kaddr + dst_off, src_kaddr + src_off, len);
-	else
-		memcpy(dst_kaddr + dst_off, src_kaddr + src_off, len);
-}
-
 void memcpy_extent_buffer(const struct extent_buffer *dst,
 			  unsigned long dst_offset, unsigned long src_offset,
 			  unsigned long len)
 {
-	size_t cur;
-	size_t dst_off_in_page;
-	size_t src_off_in_page;
-	unsigned long dst_i;
-	unsigned long src_i;
+	unsigned long cur_off = 0;
 
 	if (check_eb_range(dst, dst_offset, len) ||
 	    check_eb_range(dst, src_offset, len))
 		return;
 
-	while (len > 0) {
-		dst_off_in_page = get_eb_offset_in_page(dst, dst_offset);
-		src_off_in_page = get_eb_offset_in_page(dst, src_offset);
-
-		dst_i = get_eb_page_index(dst_offset);
-		src_i = get_eb_page_index(src_offset);
-
-		cur = min(len, (unsigned long)(PAGE_SIZE -
-					       src_off_in_page));
-		cur = min_t(unsigned long, cur,
-			(unsigned long)(PAGE_SIZE - dst_off_in_page));
-
-		copy_pages(dst->pages[dst_i], dst->pages[src_i],
-			   dst_off_in_page, src_off_in_page, cur);
-
-		src_offset += cur;
-		dst_offset += cur;
-		len -= cur;
+	while (cur_off < len) {
+		unsigned long cur_src = cur_off + src_offset;
+		unsigned long pg_index = get_eb_page_index(cur_src);
+		unsigned long pg_off = get_eb_offset_in_page(dst, cur_src);
+		unsigned long cur_len = min(src_offset + len - cur_src,
+					    PAGE_SIZE - pg_off);
+		void *src_addr = page_address(dst->pages[pg_index]) + pg_off;
+		const bool use_memmove = areas_overlap(src_offset + cur_off,
+						       dst_offset + cur_off, cur_len);
+
+		__write_extent_buffer(dst, src_addr, dst_offset + cur_off, cur_len,
+				      use_memmove);
+		cur_off += cur_len;
 	}
 }
 
@@ -4559,23 +4419,26 @@ void memmove_extent_buffer(const struct extent_buffer *dst,
 			   unsigned long dst_offset, unsigned long src_offset,
 			   unsigned long len)
 {
-	size_t cur;
-	size_t dst_off_in_page;
-	size_t src_off_in_page;
 	unsigned long dst_end = dst_offset + len - 1;
 	unsigned long src_end = src_offset + len - 1;
-	unsigned long dst_i;
-	unsigned long src_i;
 
 	if (check_eb_range(dst, dst_offset, len) ||
 	    check_eb_range(dst, src_offset, len))
 		return;
+
 	if (dst_offset < src_offset) {
 		memcpy_extent_buffer(dst, dst_offset, src_offset, len);
 		return;
 	}
+
 	while (len > 0) {
-		dst_i = get_eb_page_index(dst_end);
+		unsigned long src_i;
+		size_t cur;
+		size_t dst_off_in_page;
+		size_t src_off_in_page;
+		void *src_addr;
+		bool use_memmove;
+
 		src_i = get_eb_page_index(src_end);
 
 		dst_off_in_page = get_eb_offset_in_page(dst, dst_end);
@@ -4583,9 +4446,14 @@ void memmove_extent_buffer(const struct extent_buffer *dst,
 
 		cur = min_t(unsigned long, len, src_off_in_page + 1);
 		cur = min(cur, dst_off_in_page + 1);
-		copy_pages(dst->pages[dst_i], dst->pages[src_i],
-			   dst_off_in_page - cur + 1,
-			   src_off_in_page - cur + 1, cur);
+
+		src_addr = page_address(dst->pages[src_i]) + src_off_in_page -
+					cur + 1;
+		use_memmove = areas_overlap(src_end - cur + 1, dst_end - cur + 1,
+					    cur);
+
+		__write_extent_buffer(dst, src_addr, dst_end - cur + 1, cur,
+				      use_memmove);
 
 		dst_end -= cur;
 		src_end -= cur;
diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h
index c5fae3a7d911..68368ba99321 100644
--- a/fs/btrfs/extent_io.h
+++ b/fs/btrfs/extent_io.h
@@ -40,7 +40,6 @@ enum {
 	ENUM_BIT(PAGE_START_WRITEBACK),
 	ENUM_BIT(PAGE_END_WRITEBACK),
 	ENUM_BIT(PAGE_SET_ORDERED),
-	ENUM_BIT(PAGE_LOCK),
 };
 
 /*
@@ -94,6 +93,13 @@ struct extent_buffer {
 #endif
 };
 
+struct btrfs_eb_write_context {
+	struct writeback_control *wbc;
+	struct extent_buffer *eb;
+	/* Block group @eb resides in. Only used for zoned mode. */
+	struct btrfs_block_group *zoned_bg;
+};
+
 /*
  * Get the correct offset inside the page of extent buffer.
  *
@@ -178,8 +184,9 @@ int try_release_extent_mapping(struct page *page, gfp_t mask);
 int try_release_extent_buffer(struct page *page);
 
 int btrfs_read_folio(struct file *file, struct folio *folio);
-int extent_write_locked_range(struct inode *inode, u64 start, u64 end,
-			      struct writeback_control *wbc);
+void extent_write_locked_range(struct inode *inode, struct page *locked_page,
+			       u64 start, u64 end, struct writeback_control *wbc,
+			       bool pages_dirty);
 int extent_writepages(struct address_space *mapping,
 		      struct writeback_control *wbc);
 int btree_write_cache_pages(struct address_space *mapping,
@@ -236,11 +243,24 @@ void read_extent_buffer(const struct extent_buffer *eb, void *dst,
 int read_extent_buffer_to_user_nofault(const struct extent_buffer *eb,
 				       void __user *dst, unsigned long start,
 				       unsigned long len);
-void write_extent_buffer_fsid(const struct extent_buffer *eb, const void *src);
-void write_extent_buffer_chunk_tree_uuid(const struct extent_buffer *eb,
-		const void *src);
 void write_extent_buffer(const struct extent_buffer *eb, const void *src,
 			 unsigned long start, unsigned long len);
+
+static inline void write_extent_buffer_chunk_tree_uuid(
+		const struct extent_buffer *eb, const void *chunk_tree_uuid)
+{
+	write_extent_buffer(eb, chunk_tree_uuid,
+			    offsetof(struct btrfs_header, chunk_tree_uuid),
+			    BTRFS_FSID_SIZE);
+}
+
+static inline void write_extent_buffer_fsid(const struct extent_buffer *eb,
+					    const void *fsid)
+{
+	write_extent_buffer(eb, fsid, offsetof(struct btrfs_header, fsid),
+			    BTRFS_FSID_SIZE);
+}
+
 void copy_extent_buffer_full(const struct extent_buffer *dst,
 			     const struct extent_buffer *src);
 void copy_extent_buffer(const struct extent_buffer *dst,
@@ -266,7 +286,6 @@ void set_extent_buffer_dirty(struct extent_buffer *eb);
 void set_extent_buffer_uptodate(struct extent_buffer *eb);
 void clear_extent_buffer_uptodate(struct extent_buffer *eb);
 void extent_range_clear_dirty_for_io(struct inode *inode, u64 start, u64 end);
-void extent_range_redirty_for_io(struct inode *inode, u64 start, u64 end);
 void extent_clear_unlock_delalloc(struct btrfs_inode *inode, u64 start, u64 end,
 				  struct page *locked_page,
 				  u32 bits_to_clear, unsigned long page_ops);
@@ -277,8 +296,6 @@ void btrfs_clear_buffer_dirty(struct btrfs_trans_handle *trans,
 
 int btrfs_alloc_page_array(unsigned int nr_pages, struct page **page_array);
 
-void end_extent_writepage(struct page *page, int err, u64 start, u64 end);
-
 #ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS
 bool find_lock_delalloc_range(struct inode *inode,
 			     struct page *locked_page, u64 *start,
diff --git a/fs/btrfs/file-item.c b/fs/btrfs/file-item.c
index 696bf695d8eb..1ce5dd154499 100644
--- a/fs/btrfs/file-item.c
+++ b/fs/btrfs/file-item.c
@@ -597,29 +597,37 @@ fail:
  * Each bit represents a sector. Thus caller should ensure @csum_buf passed
  * in is large enough to contain all csums.
  */
-int btrfs_lookup_csums_bitmap(struct btrfs_root *root, u64 start, u64 end,
-			      u8 *csum_buf, unsigned long *csum_bitmap,
-			      bool search_commit)
+int btrfs_lookup_csums_bitmap(struct btrfs_root *root, struct btrfs_path *path,
+			      u64 start, u64 end, u8 *csum_buf,
+			      unsigned long *csum_bitmap)
 {
 	struct btrfs_fs_info *fs_info = root->fs_info;
 	struct btrfs_key key;
-	struct btrfs_path *path;
 	struct extent_buffer *leaf;
 	struct btrfs_csum_item *item;
 	const u64 orig_start = start;
+	bool free_path = false;
 	int ret;
 
 	ASSERT(IS_ALIGNED(start, fs_info->sectorsize) &&
 	       IS_ALIGNED(end + 1, fs_info->sectorsize));
 
-	path = btrfs_alloc_path();
-	if (!path)
-		return -ENOMEM;
+	if (!path) {
+		path = btrfs_alloc_path();
+		if (!path)
+			return -ENOMEM;
+		free_path = true;
+	}
 
-	if (search_commit) {
-		path->skip_locking = 1;
-		path->reada = READA_FORWARD;
-		path->search_commit_root = 1;
+	/* Check if we can reuse the previous path. */
+	if (path->nodes[0]) {
+		btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
+
+		if (key.objectid == BTRFS_EXTENT_CSUM_OBJECTID &&
+		    key.type == BTRFS_EXTENT_CSUM_KEY &&
+		    key.offset <= start)
+			goto search_forward;
+		btrfs_release_path(path);
 	}
 
 	key.objectid = BTRFS_EXTENT_CSUM_OBJECTID;
@@ -656,6 +664,7 @@ int btrfs_lookup_csums_bitmap(struct btrfs_root *root, u64 start, u64 end,
 		}
 	}
 
+search_forward:
 	while (start <= end) {
 		u64 csum_end;
 
@@ -712,7 +721,8 @@ int btrfs_lookup_csums_bitmap(struct btrfs_root *root, u64 start, u64 end,
 	}
 	ret = 0;
 fail:
-	btrfs_free_path(path);
+	if (free_path)
+		btrfs_free_path(path);
 	return ret;
 }
 
diff --git a/fs/btrfs/file-item.h b/fs/btrfs/file-item.h
index 4ec669b69008..04bd2d34efb1 100644
--- a/fs/btrfs/file-item.h
+++ b/fs/btrfs/file-item.h
@@ -57,9 +57,9 @@ int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end,
 int btrfs_lookup_csums_list(struct btrfs_root *root, u64 start, u64 end,
 			    struct list_head *list, int search_commit,
 			    bool nowait);
-int btrfs_lookup_csums_bitmap(struct btrfs_root *root, u64 start, u64 end,
-			      u8 *csum_buf, unsigned long *csum_bitmap,
-			      bool search_commit);
+int btrfs_lookup_csums_bitmap(struct btrfs_root *root, struct btrfs_path *path,
+			      u64 start, u64 end, u8 *csum_buf,
+			      unsigned long *csum_bitmap);
 void btrfs_extent_item_to_extent_map(struct btrfs_inode *inode,
 				     const struct btrfs_path *path,
 				     struct btrfs_file_extent_item *fi,
diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c
index 3887a8e1c964..ca46a529d56b 100644
--- a/fs/btrfs/file.c
+++ b/fs/btrfs/file.c
@@ -2999,7 +2999,7 @@ static long btrfs_fallocate(struct file *file, int mode,
 	struct extent_changeset *data_reserved = NULL;
 	struct falloc_range *range;
 	struct falloc_range *tmp;
-	struct list_head reserve_list;
+	LIST_HEAD(reserve_list);
 	u64 cur_offset;
 	u64 last_byte;
 	u64 alloc_start;
@@ -3091,7 +3091,6 @@ static long btrfs_fallocate(struct file *file, int mode,
 	btrfs_assert_inode_range_clean(BTRFS_I(inode), alloc_start, locked_end);
 
 	/* First, check if we exceed the qgroup limit */
-	INIT_LIST_HEAD(&reserve_list);
 	while (cur_offset < alloc_end) {
 		em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, cur_offset,
 				      alloc_end - cur_offset);
diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c
index 880800418075..27fad70451aa 100644
--- a/fs/btrfs/free-space-cache.c
+++ b/fs/btrfs/free-space-cache.c
@@ -1219,10 +1219,9 @@ static noinline_for_stack int write_pinned_extent_entries(
 	start = block_group->start;
 
 	while (start < block_group->start + block_group->length) {
-		ret = find_first_extent_bit(unpin, start,
-					    &extent_start, &extent_end,
-					    EXTENT_DIRTY, NULL);
-		if (ret)
+		if (!find_first_extent_bit(unpin, start,
+					   &extent_start, &extent_end,
+					   EXTENT_DIRTY, NULL))
 			return 0;
 
 		/* This pinned extent is out of our range */
@@ -2705,13 +2704,8 @@ static int __btrfs_add_free_space_zoned(struct btrfs_block_group *block_group,
 		bg_reclaim_threshold = READ_ONCE(sinfo->bg_reclaim_threshold);
 
 	spin_lock(&ctl->tree_lock);
-	/* Count initial region as zone_unusable until it gets activated. */
 	if (!used)
 		to_free = size;
-	else if (initial &&
-		 test_bit(BTRFS_FS_ACTIVE_ZONE_TRACKING, &block_group->fs_info->flags) &&
-		 (block_group->flags & (BTRFS_BLOCK_GROUP_METADATA | BTRFS_BLOCK_GROUP_SYSTEM)))
-		to_free = 0;
 	else if (initial)
 		to_free = block_group->zone_capacity;
 	else if (offset >= block_group->alloc_offset)
@@ -2739,8 +2733,7 @@ static int __btrfs_add_free_space_zoned(struct btrfs_block_group *block_group,
 	reclaimable_unusable = block_group->zone_unusable -
 			       (block_group->length - block_group->zone_capacity);
 	/* All the region is now unusable. Mark it as unused and reclaim */
-	if (block_group->zone_unusable == block_group->length &&
-	    block_group->alloc_offset) {
+	if (block_group->zone_unusable == block_group->length) {
 		btrfs_mark_bg_unused(block_group);
 	} else if (bg_reclaim_threshold &&
 		   reclaimable_unusable >=
@@ -2944,7 +2937,8 @@ void btrfs_dump_free_space(struct btrfs_block_group *block_group,
 	btrfs_info(fs_info, "block group has cluster?: %s",
 	       list_empty(&block_group->cluster_list) ? "no" : "yes");
 	btrfs_info(fs_info,
-		   "%d blocks of free space at or bigger than bytes is", count);
+		   "%d free space entries at or bigger than %llu bytes",
+		   count, bytes);
 }
 
 void btrfs_init_free_space_ctl(struct btrfs_block_group *block_group,
diff --git a/fs/btrfs/free-space-tree.c b/fs/btrfs/free-space-tree.c
index f169378e2ca6..c0e734082dcc 100644
--- a/fs/btrfs/free-space-tree.c
+++ b/fs/btrfs/free-space-tree.c
@@ -1517,8 +1517,10 @@ static int load_free_space_bitmaps(struct btrfs_caching_control *caching_ctl,
 			} else if (prev_bit == 1 && bit == 0) {
 				u64 space_added;
 
-				ret = add_new_free_space(block_group, extent_start,
-							 offset, &space_added);
+				ret = btrfs_add_new_free_space(block_group,
+							       extent_start,
+							       offset,
+							       &space_added);
 				if (ret)
 					goto out;
 				total_found += space_added;
@@ -1533,7 +1535,7 @@ static int load_free_space_bitmaps(struct btrfs_caching_control *caching_ctl,
 		}
 	}
 	if (prev_bit == 1) {
-		ret = add_new_free_space(block_group, extent_start, end, NULL);
+		ret = btrfs_add_new_free_space(block_group, extent_start, end, NULL);
 		if (ret)
 			goto out;
 		extent_count++;
@@ -1590,8 +1592,9 @@ static int load_free_space_extents(struct btrfs_caching_control *caching_ctl,
 		ASSERT(key.type == BTRFS_FREE_SPACE_EXTENT_KEY);
 		ASSERT(key.objectid < end && key.objectid + key.offset <= end);
 
-		ret = add_new_free_space(block_group, key.objectid,
-					 key.objectid + key.offset, &space_added);
+		ret = btrfs_add_new_free_space(block_group, key.objectid,
+					       key.objectid + key.offset,
+					       &space_added);
 		if (ret)
 			goto out;
 		total_found += space_added;
diff --git a/fs/btrfs/fs.h b/fs/btrfs/fs.h
index 203d2a267828..a523d64d5491 100644
--- a/fs/btrfs/fs.h
+++ b/fs/btrfs/fs.h
@@ -46,8 +46,6 @@ static_assert(sizeof(struct btrfs_super_block) == BTRFS_SUPER_INFO_SIZE);
  * Runtime (in-memory) states of filesystem
  */
 enum {
-	/* Global indicator of serious filesystem errors */
-	BTRFS_FS_STATE_ERROR,
 	/*
 	 * Filesystem is being remounted, allow to skip some operations, like
 	 * defrag
@@ -686,6 +684,12 @@ struct btrfs_fs_info {
 	bool qgroup_rescan_running;
 	u8 qgroup_drop_subtree_thres;
 
+	/*
+	 * If this is not 0, then it indicates a serious filesystem error has
+	 * happened and it contains that error (negative errno value).
+	 */
+	int fs_error;
+
 	/* Filesystem state */
 	unsigned long fs_state;
 
@@ -766,6 +770,9 @@ struct btrfs_fs_info {
 	u64 data_reloc_bg;
 	struct mutex zoned_data_reloc_io_lock;
 
+	struct btrfs_block_group *active_meta_bg;
+	struct btrfs_block_group *active_system_bg;
+
 	u64 nr_global_roots;
 
 	spinlock_t zone_active_bgs_lock;
@@ -962,8 +969,8 @@ static inline void btrfs_wake_unfinished_drop(struct btrfs_fs_info *fs_info)
 	clear_and_wake_up_bit(BTRFS_FS_UNFINISHED_DROPS, &fs_info->flags);
 }
 
-#define BTRFS_FS_ERROR(fs_info)	(unlikely(test_bit(BTRFS_FS_STATE_ERROR, \
-						   &(fs_info)->fs_state)))
+#define BTRFS_FS_ERROR(fs_info)	(READ_ONCE((fs_info)->fs_error))
+
 #define BTRFS_FS_LOG_CLEANUP_ERROR(fs_info)				\
 	(unlikely(test_bit(BTRFS_FS_STATE_LOG_CLEANUP_ERROR,		\
 			   &(fs_info)->fs_state)))
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 4e987fe6d648..f09fbdc43f0f 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -124,11 +124,11 @@ static struct kmem_cache *btrfs_inode_cachep;
 
 static int btrfs_setsize(struct inode *inode, struct iattr *attr);
 static int btrfs_truncate(struct btrfs_inode *inode, bool skip_writeback);
-static noinline int cow_file_range(struct btrfs_inode *inode,
-				   struct page *locked_page,
-				   u64 start, u64 end, int *page_started,
-				   unsigned long *nr_written, int unlock,
-				   u64 *done_offset);
+
+static noinline int run_delalloc_cow(struct btrfs_inode *inode,
+				     struct page *locked_page, u64 start,
+				     u64 end, struct writeback_control *wbc,
+				     bool pages_dirty);
 static struct extent_map *create_io_em(struct btrfs_inode *inode, u64 start,
 				       u64 len, u64 orig_start, u64 block_start,
 				       u64 block_len, u64 orig_block_len,
@@ -423,11 +423,10 @@ static inline void btrfs_cleanup_ordered_extents(struct btrfs_inode *inode,
 
 	while (index <= end_index) {
 		/*
-		 * For locked page, we will call end_extent_writepage() on it
-		 * in run_delalloc_range() for the error handling.  That
-		 * end_extent_writepage() function will call
-		 * btrfs_mark_ordered_io_finished() to clear page Ordered and
-		 * run the ordered extent accounting.
+		 * For locked page, we will call btrfs_mark_ordered_io_finished
+		 * through btrfs_mark_ordered_io_finished() on it
+		 * in run_delalloc_range() for the error handling, which will
+		 * clear page Ordered and run the ordered extent accounting.
 		 *
 		 * Here we can't just clear the Ordered bit, or
 		 * btrfs_mark_ordered_io_finished() would skip the accounting
@@ -815,24 +814,22 @@ static inline void inode_should_defrag(struct btrfs_inode *inode,
 }
 
 /*
- * we create compressed extents in two phases.  The first
- * phase compresses a range of pages that have already been
- * locked (both pages and state bits are locked).
+ * Work queue call back to started compression on a file and pages.
  *
- * This is done inside an ordered work queue, and the compression
- * is spread across many cpus.  The actual IO submission is step
- * two, and the ordered work queue takes care of making sure that
- * happens in the same order things were put onto the queue by
- * writepages and friends.
+ * This is done inside an ordered work queue, and the compression is spread
+ * across many cpus.  The actual IO submission is step two, and the ordered work
+ * queue takes care of making sure that happens in the same order things were
+ * put onto the queue by writepages and friends.
  *
- * If this code finds it can't get good compression, it puts an
- * entry onto the work queue to write the uncompressed bytes.  This
- * makes sure that both compressed inodes and uncompressed inodes
- * are written in the same order that the flusher thread sent them
- * down.
+ * If this code finds it can't get good compression, it puts an entry onto the
+ * work queue to write the uncompressed bytes.  This makes sure that both
+ * compressed inodes and uncompressed inodes are written in the same order that
+ * the flusher thread sent them down.
  */
-static noinline int compress_file_range(struct async_chunk *async_chunk)
+static void compress_file_range(struct btrfs_work *work)
 {
+	struct async_chunk *async_chunk =
+		container_of(work, struct async_chunk, work);
 	struct btrfs_inode *inode = async_chunk->inode;
 	struct btrfs_fs_info *fs_info = inode->root->fs_info;
 	struct address_space *mapping = inode->vfs_inode.i_mapping;
@@ -842,19 +839,24 @@ static noinline int compress_file_range(struct async_chunk *async_chunk)
 	u64 actual_end;
 	u64 i_size;
 	int ret = 0;
-	struct page **pages = NULL;
+	struct page **pages;
 	unsigned long nr_pages;
 	unsigned long total_compressed = 0;
 	unsigned long total_in = 0;
+	unsigned int poff;
 	int i;
-	int will_compress;
 	int compress_type = fs_info->compress_type;
-	int compressed_extents = 0;
-	int redirty = 0;
 
 	inode_should_defrag(inode, start, end, end - start + 1, SZ_16K);
 
 	/*
+	 * We need to call clear_page_dirty_for_io on each page in the range.
+	 * Otherwise applications with the file mmap'd can wander in and change
+	 * the page contents while we are compressing them.
+	 */
+	extent_range_clear_dirty_for_io(&inode->vfs_inode, start, end);
+
+	/*
 	 * We need to save i_size before now because it could change in between
 	 * us evaluating the size and assigning it.  This is because we lock and
 	 * unlock the page in truncate and fallocate, and then modify the i_size
@@ -868,7 +870,7 @@ static noinline int compress_file_range(struct async_chunk *async_chunk)
 	barrier();
 	actual_end = min_t(u64, i_size, end + 1);
 again:
-	will_compress = 0;
+	pages = NULL;
 	nr_pages = (end >> PAGE_SHIFT) - (start >> PAGE_SHIFT) + 1;
 	nr_pages = min_t(unsigned long, nr_pages, BTRFS_MAX_COMPRESSED_PAGES);
 
@@ -912,78 +914,57 @@ again:
 	ret = 0;
 
 	/*
-	 * we do compression for mount -o compress and when the
-	 * inode has not been flagged as nocompress.  This flag can
-	 * change at any time if we discover bad compression ratios.
+	 * We do compression for mount -o compress and when the inode has not
+	 * been flagged as NOCOMPRESS.  This flag can change at any time if we
+	 * discover bad compression ratios.
 	 */
-	if (inode_need_compress(inode, start, end)) {
-		WARN_ON(pages);
-		pages = kcalloc(nr_pages, sizeof(struct page *), GFP_NOFS);
-		if (!pages) {
-			/* just bail out to the uncompressed code */
-			nr_pages = 0;
-			goto cont;
-		}
-
-		if (inode->defrag_compress)
-			compress_type = inode->defrag_compress;
-		else if (inode->prop_compress)
-			compress_type = inode->prop_compress;
+	if (!inode_need_compress(inode, start, end))
+		goto cleanup_and_bail_uncompressed;
 
+	pages = kcalloc(nr_pages, sizeof(struct page *), GFP_NOFS);
+	if (!pages) {
 		/*
-		 * we need to call clear_page_dirty_for_io on each
-		 * page in the range.  Otherwise applications with the file
-		 * mmap'd can wander in and change the page contents while
-		 * we are compressing them.
-		 *
-		 * If the compression fails for any reason, we set the pages
-		 * dirty again later on.
-		 *
-		 * Note that the remaining part is redirtied, the start pointer
-		 * has moved, the end is the original one.
+		 * Memory allocation failure is not a fatal error, we can fall
+		 * back to uncompressed code.
 		 */
-		if (!redirty) {
-			extent_range_clear_dirty_for_io(&inode->vfs_inode, start, end);
-			redirty = 1;
-		}
+		goto cleanup_and_bail_uncompressed;
+	}
 
-		/* Compression level is applied here and only here */
-		ret = btrfs_compress_pages(
-			compress_type | (fs_info->compress_level << 4),
-					   mapping, start,
-					   pages,
-					   &nr_pages,
-					   &total_in,
-					   &total_compressed);
+	if (inode->defrag_compress)
+		compress_type = inode->defrag_compress;
+	else if (inode->prop_compress)
+		compress_type = inode->prop_compress;
+
+	/* Compression level is applied here. */
+	ret = btrfs_compress_pages(compress_type | (fs_info->compress_level << 4),
+				   mapping, start, pages, &nr_pages, &total_in,
+				   &total_compressed);
+	if (ret)
+		goto mark_incompressible;
 
-		if (!ret) {
-			unsigned long offset = offset_in_page(total_compressed);
-			struct page *page = pages[nr_pages - 1];
+	/*
+	 * Zero the tail end of the last page, as we might be sending it down
+	 * to disk.
+	 */
+	poff = offset_in_page(total_compressed);
+	if (poff)
+		memzero_page(pages[nr_pages - 1], poff, PAGE_SIZE - poff);
 
-			/* zero the tail end of the last page, we might be
-			 * sending it down to disk
-			 */
-			if (offset)
-				memzero_page(page, offset, PAGE_SIZE - offset);
-			will_compress = 1;
-		}
-	}
-cont:
 	/*
+	 * Try to create an inline extent.
+	 *
+	 * If we didn't compress the entire range, try to create an uncompressed
+	 * inline extent, else a compressed one.
+	 *
 	 * Check cow_file_range() for why we don't even try to create inline
-	 * extent for subpage case.
+	 * extent for the subpage case.
 	 */
 	if (start == 0 && fs_info->sectorsize == PAGE_SIZE) {
-		/* lets try to make an inline extent */
-		if (ret || total_in < actual_end) {
-			/* we didn't compress the entire range, try
-			 * to make an uncompressed inline extent.
-			 */
-			ret = cow_file_range_inline(inode, actual_end,
-						    0, BTRFS_COMPRESS_NONE,
-						    NULL, false);
+		if (total_in < actual_end) {
+			ret = cow_file_range_inline(inode, actual_end, 0,
+						    BTRFS_COMPRESS_NONE, NULL,
+						    false);
 		} else {
-			/* try making a compressed inline extent */
 			ret = cow_file_range_inline(inode, actual_end,
 						    total_compressed,
 						    compress_type, pages,
@@ -1013,99 +994,52 @@ cont:
 						     PAGE_UNLOCK |
 						     PAGE_START_WRITEBACK |
 						     PAGE_END_WRITEBACK);
-
-			/*
-			 * Ensure we only free the compressed pages if we have
-			 * them allocated, as we can still reach here with
-			 * inode_need_compress() == false.
-			 */
-			if (pages) {
-				for (i = 0; i < nr_pages; i++) {
-					WARN_ON(pages[i]->mapping);
-					put_page(pages[i]);
-				}
-				kfree(pages);
-			}
-			return 0;
+			goto free_pages;
 		}
 	}
 
-	if (will_compress) {
-		/*
-		 * we aren't doing an inline extent round the compressed size
-		 * up to a block size boundary so the allocator does sane
-		 * things
-		 */
-		total_compressed = ALIGN(total_compressed, blocksize);
+	/*
+	 * We aren't doing an inline extent. Round the compressed size up to a
+	 * block size boundary so the allocator does sane things.
+	 */
+	total_compressed = ALIGN(total_compressed, blocksize);
 
-		/*
-		 * one last check to make sure the compression is really a
-		 * win, compare the page count read with the blocks on disk,
-		 * compression must free at least one sector size
-		 */
-		total_in = round_up(total_in, fs_info->sectorsize);
-		if (total_compressed + blocksize <= total_in) {
-			compressed_extents++;
+	/*
+	 * One last check to make sure the compression is really a win, compare
+	 * the page count read with the blocks on disk, compression must free at
+	 * least one sector.
+	 */
+	total_in = round_up(total_in, fs_info->sectorsize);
+	if (total_compressed + blocksize > total_in)
+		goto mark_incompressible;
 
-			/*
-			 * The async work queues will take care of doing actual
-			 * allocation on disk for these compressed pages, and
-			 * will submit them to the elevator.
-			 */
-			add_async_extent(async_chunk, start, total_in,
-					total_compressed, pages, nr_pages,
-					compress_type);
-
-			if (start + total_in < end) {
-				start += total_in;
-				pages = NULL;
-				cond_resched();
-				goto again;
-			}
-			return compressed_extents;
-		}
+	/*
+	 * The async work queues will take care of doing actual allocation on
+	 * disk for these compressed pages, and will submit the bios.
+	 */
+	add_async_extent(async_chunk, start, total_in, total_compressed, pages,
+			 nr_pages, compress_type);
+	if (start + total_in < end) {
+		start += total_in;
+		cond_resched();
+		goto again;
 	}
+	return;
+
+mark_incompressible:
+	if (!btrfs_test_opt(fs_info, FORCE_COMPRESS) && !inode->prop_compress)
+		inode->flags |= BTRFS_INODE_NOCOMPRESS;
+cleanup_and_bail_uncompressed:
+	add_async_extent(async_chunk, start, end - start + 1, 0, NULL, 0,
+			 BTRFS_COMPRESS_NONE);
+free_pages:
 	if (pages) {
-		/*
-		 * the compression code ran but failed to make things smaller,
-		 * free any pages it allocated and our page pointer array
-		 */
 		for (i = 0; i < nr_pages; i++) {
 			WARN_ON(pages[i]->mapping);
 			put_page(pages[i]);
 		}
 		kfree(pages);
-		pages = NULL;
-		total_compressed = 0;
-		nr_pages = 0;
-
-		/* flag the file so we don't compress in the future */
-		if (!btrfs_test_opt(fs_info, FORCE_COMPRESS) &&
-		    !(inode->prop_compress)) {
-			inode->flags |= BTRFS_INODE_NOCOMPRESS;
-		}
-	}
-cleanup_and_bail_uncompressed:
-	/*
-	 * No compression, but we still need to write the pages in the file
-	 * we've been given so far.  redirty the locked page if it corresponds
-	 * to our extent and set things up for the async work queue to run
-	 * cow_file_range to do the normal delalloc dance.
-	 */
-	if (async_chunk->locked_page &&
-	    (page_offset(async_chunk->locked_page) >= start &&
-	     page_offset(async_chunk->locked_page)) <= end) {
-		__set_page_dirty_nobuffers(async_chunk->locked_page);
-		/* unlocked later on in the async handlers */
 	}
-
-	if (redirty)
-		extent_range_redirty_for_io(&inode->vfs_inode, start, end);
-	add_async_extent(async_chunk, start, end - start + 1, 0, NULL, 0,
-			 BTRFS_COMPRESS_NONE);
-	compressed_extents++;
-
-	return compressed_extents;
 }
 
 static void free_async_extent_pages(struct async_extent *async_extent)
@@ -1124,14 +1058,12 @@ static void free_async_extent_pages(struct async_extent *async_extent)
 	async_extent->pages = NULL;
 }
 
-static int submit_uncompressed_range(struct btrfs_inode *inode,
-				     struct async_extent *async_extent,
-				     struct page *locked_page)
+static void submit_uncompressed_range(struct btrfs_inode *inode,
+				      struct async_extent *async_extent,
+				      struct page *locked_page)
 {
 	u64 start = async_extent->start;
 	u64 end = async_extent->start + async_extent->ram_size - 1;
-	unsigned long nr_written = 0;
-	int page_started = 0;
 	int ret;
 	struct writeback_control wbc = {
 		.sync_mode		= WB_SYNC_ALL,
@@ -1140,45 +1072,33 @@ static int submit_uncompressed_range(struct btrfs_inode *inode,
 		.no_cgroup_owner	= 1,
 	};
 
-	/*
-	 * Call cow_file_range() to run the delalloc range directly, since we
-	 * won't go to NOCOW or async path again.
-	 *
-	 * Also we call cow_file_range() with @unlock_page == 0, so that we
-	 * can directly submit them without interruption.
-	 */
-	ret = cow_file_range(inode, locked_page, start, end, &page_started,
-			     &nr_written, 0, NULL);
-	/* Inline extent inserted, page gets unlocked and everything is done */
-	if (page_started)
-		return 0;
-
+	wbc_attach_fdatawrite_inode(&wbc, &inode->vfs_inode);
+	ret = run_delalloc_cow(inode, locked_page, start, end, &wbc, false);
+	wbc_detach_inode(&wbc);
 	if (ret < 0) {
 		btrfs_cleanup_ordered_extents(inode, locked_page, start, end - start + 1);
 		if (locked_page) {
 			const u64 page_start = page_offset(locked_page);
-			const u64 page_end = page_start + PAGE_SIZE - 1;
 
 			set_page_writeback(locked_page);
 			end_page_writeback(locked_page);
-			end_extent_writepage(locked_page, ret, page_start, page_end);
+			btrfs_mark_ordered_io_finished(inode, locked_page,
+						       page_start, PAGE_SIZE,
+						       !ret);
+			btrfs_page_clear_uptodate(inode->root->fs_info,
+						  locked_page, page_start,
+						  PAGE_SIZE);
+			mapping_set_error(locked_page->mapping, ret);
 			unlock_page(locked_page);
 		}
-		return ret;
 	}
-
-	/* All pages will be unlocked, including @locked_page */
-	wbc_attach_fdatawrite_inode(&wbc, &inode->vfs_inode);
-	ret = extent_write_locked_range(&inode->vfs_inode, start, end, &wbc);
-	wbc_detach_inode(&wbc);
-	return ret;
 }
 
-static int submit_one_async_extent(struct btrfs_inode *inode,
-				   struct async_chunk *async_chunk,
-				   struct async_extent *async_extent,
-				   u64 *alloc_hint)
+static void submit_one_async_extent(struct async_chunk *async_chunk,
+				    struct async_extent *async_extent,
+				    u64 *alloc_hint)
 {
+	struct btrfs_inode *inode = async_chunk->inode;
 	struct extent_io_tree *io_tree = &inode->io_tree;
 	struct btrfs_root *root = inode->root;
 	struct btrfs_fs_info *fs_info = root->fs_info;
@@ -1206,9 +1126,8 @@ static int submit_one_async_extent(struct btrfs_inode *inode,
 	}
 	lock_extent(io_tree, start, end, NULL);
 
-	/* We have fall back to uncompressed write */
-	if (!async_extent->pages) {
-		ret = submit_uncompressed_range(inode, async_extent, locked_page);
+	if (async_extent->compress_type == BTRFS_COMPRESS_NONE) {
+		submit_uncompressed_range(inode, async_extent, locked_page);
 		goto done;
 	}
 
@@ -1217,7 +1136,6 @@ static int submit_one_async_extent(struct btrfs_inode *inode,
 				   async_extent->compressed_size,
 				   0, *alloc_hint, &ins, 1, 1);
 	if (ret) {
-		free_async_extent_pages(async_extent);
 		/*
 		 * Here we used to try again by going back to non-compressed
 		 * path for ENOSPC.  But we can't reserve space even for
@@ -1272,7 +1190,7 @@ done:
 	if (async_chunk->blkcg_css)
 		kthread_associate_blkcg(NULL);
 	kfree(async_extent);
-	return ret;
+	return;
 
 out_free_reserve:
 	btrfs_dec_block_group_reservations(fs_info, ins.objectid);
@@ -1286,39 +1204,13 @@ out_free:
 				     PAGE_UNLOCK | PAGE_START_WRITEBACK |
 				     PAGE_END_WRITEBACK);
 	free_async_extent_pages(async_extent);
-	goto done;
-}
-
-/*
- * Phase two of compressed writeback.  This is the ordered portion of the code,
- * which only gets called in the order the work was queued.  We walk all the
- * async extents created by compress_file_range and send them down to the disk.
- */
-static noinline void submit_compressed_extents(struct async_chunk *async_chunk)
-{
-	struct btrfs_inode *inode = async_chunk->inode;
-	struct btrfs_fs_info *fs_info = inode->root->fs_info;
-	struct async_extent *async_extent;
-	u64 alloc_hint = 0;
-	int ret = 0;
-
-	while (!list_empty(&async_chunk->extents)) {
-		u64 extent_start;
-		u64 ram_size;
-
-		async_extent = list_entry(async_chunk->extents.next,
-					  struct async_extent, list);
-		list_del(&async_extent->list);
-		extent_start = async_extent->start;
-		ram_size = async_extent->ram_size;
-
-		ret = submit_one_async_extent(inode, async_chunk, async_extent,
-					      &alloc_hint);
-		btrfs_debug(fs_info,
+	if (async_chunk->blkcg_css)
+		kthread_associate_blkcg(NULL);
+	btrfs_debug(fs_info,
 "async extent submission failed root=%lld inode=%llu start=%llu len=%llu ret=%d",
-			    inode->root->root_key.objectid,
-			    btrfs_ino(inode), extent_start, ram_size, ret);
-	}
+		    root->root_key.objectid, btrfs_ino(inode), start,
+		    async_extent->ram_size, ret);
+	kfree(async_extent);
 }
 
 static u64 get_extent_allocation_hint(struct btrfs_inode *inode, u64 start,
@@ -1362,25 +1254,18 @@ static u64 get_extent_allocation_hint(struct btrfs_inode *inode, u64 start,
  * locked_page is the page that writepage had locked already.  We use
  * it to make sure we don't do extra locks or unlocks.
  *
- * *page_started is set to one if we unlock locked_page and do everything
- * required to start IO on it.  It may be clean and already done with
- * IO when we return.
- *
- * When unlock == 1, we unlock the pages in successfully allocated regions.
- * When unlock == 0, we leave them locked for writing them out.
+ * When this function fails, it unlocks all pages except @locked_page.
  *
- * However, we unlock all the pages except @locked_page in case of failure.
+ * When this function successfully creates an inline extent, it returns 1 and
+ * unlocks all pages including locked_page and starts I/O on them.
+ * (In reality inline extents are limited to a single page, so locked_page is
+ * the only page handled anyway).
  *
- * In summary, page locking state will be as follow:
+ * When this function succeed and creates a normal extent, the page locking
+ * status depends on the passed in flags:
  *
- * - page_started == 1 (return value)
- *     - All the pages are unlocked. IO is started.
- *     - Note that this can happen only on success
- * - unlock == 1
- *     - All the pages except @locked_page are unlocked in any case
- * - unlock == 0
- *     - On success, all the pages are locked for writing out them
- *     - On failure, all the pages except @locked_page are unlocked
+ * - If @keep_locked is set, all pages are kept locked.
+ * - Else all pages except for @locked_page 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
@@ -1389,10 +1274,9 @@ static u64 get_extent_allocation_hint(struct btrfs_inode *inode, u64 start,
  * example.
  */
 static noinline int cow_file_range(struct btrfs_inode *inode,
-				   struct page *locked_page,
-				   u64 start, u64 end, int *page_started,
-				   unsigned long *nr_written, int unlock,
-				   u64 *done_offset)
+				   struct page *locked_page, u64 start, u64 end,
+				   u64 *done_offset,
+				   bool keep_locked, bool no_inline)
 {
 	struct btrfs_root *root = inode->root;
 	struct btrfs_fs_info *fs_info = root->fs_info;
@@ -1431,7 +1315,7 @@ static noinline int cow_file_range(struct btrfs_inode *inode,
 	 * This means we can trigger inline extent even if we didn't want to.
 	 * So here we skip inline extent creation completely.
 	 */
-	if (start == 0 && fs_info->sectorsize == PAGE_SIZE) {
+	if (start == 0 && fs_info->sectorsize == PAGE_SIZE && !no_inline) {
 		u64 actual_end = min_t(u64, i_size_read(&inode->vfs_inode),
 				       end + 1);
 
@@ -1451,9 +1335,6 @@ static noinline int cow_file_range(struct btrfs_inode *inode,
 				     EXTENT_DELALLOC_NEW | EXTENT_DEFRAG |
 				     EXTENT_DO_ACCOUNTING, PAGE_UNLOCK |
 				     PAGE_START_WRITEBACK | PAGE_END_WRITEBACK);
-			*nr_written = *nr_written +
-			     (end - start + PAGE_SIZE) / PAGE_SIZE;
-			*page_started = 1;
 			/*
 			 * locked_page is locked by the caller of
 			 * writepage_delalloc(), not locked by
@@ -1463,11 +1344,12 @@ static noinline int cow_file_range(struct btrfs_inode *inode,
 			 * as it doesn't have any subpage::writers recorded.
 			 *
 			 * Here we manually unlock the page, since the caller
-			 * can't use page_started to determine if it's an
-			 * inline extent or a compressed extent.
+			 * can't determine if it's an inline extent or a
+			 * compressed extent.
 			 */
 			unlock_page(locked_page);
-			goto out;
+			ret = 1;
+			goto done;
 		} else if (ret < 0) {
 			goto out_unlock;
 		}
@@ -1498,6 +1380,31 @@ static noinline int cow_file_range(struct btrfs_inode *inode,
 		ret = btrfs_reserve_extent(root, cur_alloc_size, cur_alloc_size,
 					   min_alloc_size, 0, alloc_hint,
 					   &ins, 1, 1);
+		if (ret == -EAGAIN) {
+			/*
+			 * btrfs_reserve_extent only returns -EAGAIN for zoned
+			 * file systems, which is an indication that there are
+			 * no active zones to allocate from at the moment.
+			 *
+			 * If this is the first loop iteration, wait for at
+			 * least one zone to finish before retrying the
+			 * allocation.  Otherwise ask the caller to write out
+			 * the already allocated blocks before coming back to
+			 * us, or return -ENOSPC if it can't handle retries.
+			 */
+			ASSERT(btrfs_is_zoned(fs_info));
+			if (start == orig_start) {
+				wait_on_bit_io(&inode->root->fs_info->flags,
+					       BTRFS_FS_NEED_ZONE_FINISH,
+					       TASK_UNINTERRUPTIBLE);
+				continue;
+			}
+			if (done_offset) {
+				*done_offset = start - 1;
+				return 0;
+			}
+			ret = -ENOSPC;
+		}
 		if (ret < 0)
 			goto out_unlock;
 		cur_alloc_size = ins.offset;
@@ -1558,7 +1465,7 @@ static noinline int cow_file_range(struct btrfs_inode *inode,
 		 * Do set the Ordered (Private2) bit so we know this page was
 		 * properly setup for writepage.
 		 */
-		page_ops = unlock ? PAGE_UNLOCK : 0;
+		page_ops = (keep_locked ? 0 : PAGE_UNLOCK);
 		page_ops |= PAGE_SET_ORDERED;
 
 		extent_clear_unlock_delalloc(inode, start, start + ram_size - 1,
@@ -1581,7 +1488,9 @@ static noinline int cow_file_range(struct btrfs_inode *inode,
 		if (ret)
 			goto out_unlock;
 	}
-out:
+done:
+	if (done_offset)
+		*done_offset = end;
 	return ret;
 
 out_drop_extent_cache:
@@ -1591,21 +1500,6 @@ out_reserve:
 	btrfs_free_reserved_extent(fs_info, ins.objectid, ins.offset, 1);
 out_unlock:
 	/*
-	 * If done_offset is non-NULL and ret == -EAGAIN, we expect the
-	 * caller to write out the successfully allocated region and retry.
-	 */
-	if (done_offset && ret == -EAGAIN) {
-		if (orig_start < start)
-			*done_offset = start - 1;
-		else
-			*done_offset = start;
-		return ret;
-	} else if (ret == -EAGAIN) {
-		/* Convert to -ENOSPC since the caller cannot retry. */
-		ret = -ENOSPC;
-	}
-
-	/*
 	 * Now, we have three regions to clean up:
 	 *
 	 * |-------(1)----|---(2)---|-------------(3)----------|
@@ -1627,10 +1521,10 @@ out_unlock:
 	 * EXTENT_DEFRAG | EXTENT_CLEAR_META_RESV are handled by the cleanup
 	 * function.
 	 *
-	 * However, in case of unlock == 0, we still need to unlock the pages
+	 * However, in case of @keep_locked, we still need to unlock the pages
 	 * (except @locked_page) to ensure all the pages are unlocked.
 	 */
-	if (!unlock && orig_start < start) {
+	if (keep_locked && orig_start < start) {
 		if (!locked_page)
 			mapping_set_error(inode->vfs_inode.i_mapping, ret);
 		extent_clear_unlock_delalloc(inode, orig_start, start - 1,
@@ -1671,43 +1565,28 @@ out_unlock:
 }
 
 /*
- * work queue call back to started compression on a file and pages
- */
-static noinline void async_cow_start(struct btrfs_work *work)
-{
-	struct async_chunk *async_chunk;
-	int compressed_extents;
-
-	async_chunk = container_of(work, struct async_chunk, work);
-
-	compressed_extents = compress_file_range(async_chunk);
-	if (compressed_extents == 0) {
-		btrfs_add_delayed_iput(async_chunk->inode);
-		async_chunk->inode = NULL;
-	}
-}
-
-/*
- * work queue call back to submit previously compressed pages
+ * Phase two of compressed writeback.  This is the ordered portion of the code,
+ * which only gets called in the order the work was queued.  We walk all the
+ * async extents created by compress_file_range and send them down to the disk.
  */
-static noinline void async_cow_submit(struct btrfs_work *work)
+static noinline void submit_compressed_extents(struct btrfs_work *work)
 {
 	struct async_chunk *async_chunk = container_of(work, struct async_chunk,
 						     work);
 	struct btrfs_fs_info *fs_info = btrfs_work_owner(work);
+	struct async_extent *async_extent;
 	unsigned long nr_pages;
+	u64 alloc_hint = 0;
 
 	nr_pages = (async_chunk->end - async_chunk->start + PAGE_SIZE) >>
 		PAGE_SHIFT;
 
-	/*
-	 * ->inode could be NULL if async_chunk_start has failed to compress,
-	 * in which case we don't have anything to submit, yet we need to
-	 * always adjust ->async_delalloc_pages as its paired with the init
-	 * happening in run_delalloc_compressed
-	 */
-	if (async_chunk->inode)
-		submit_compressed_extents(async_chunk);
+	while (!list_empty(&async_chunk->extents)) {
+		async_extent = list_entry(async_chunk->extents.next,
+					  struct async_extent, list);
+		list_del(&async_extent->list);
+		submit_one_async_extent(async_chunk, async_extent, &alloc_hint);
+	}
 
 	/* atomic_sub_return implies a barrier */
 	if (atomic_sub_return(nr_pages, &fs_info->async_delalloc_pages) <
@@ -1721,8 +1600,7 @@ static noinline void async_cow_free(struct btrfs_work *work)
 	struct async_cow *async_cow;
 
 	async_chunk = container_of(work, struct async_chunk, work);
-	if (async_chunk->inode)
-		btrfs_add_delayed_iput(async_chunk->inode);
+	btrfs_add_delayed_iput(async_chunk->inode);
 	if (async_chunk->blkcg_css)
 		css_put(async_chunk->blkcg_css);
 
@@ -1732,10 +1610,8 @@ static noinline void async_cow_free(struct btrfs_work *work)
 }
 
 static bool run_delalloc_compressed(struct btrfs_inode *inode,
-				    struct writeback_control *wbc,
-				    struct page *locked_page,
-				    u64 start, u64 end, int *page_started,
-				    unsigned long *nr_written)
+				    struct page *locked_page, u64 start,
+				    u64 end, struct writeback_control *wbc)
 {
 	struct btrfs_fs_info *fs_info = inode->root->fs_info;
 	struct cgroup_subsys_state *blkcg_css = wbc_blkcg_css(wbc);
@@ -1809,65 +1685,42 @@ static bool run_delalloc_compressed(struct btrfs_inode *inode,
 			async_chunk[i].blkcg_css = NULL;
 		}
 
-		btrfs_init_work(&async_chunk[i].work, async_cow_start,
-				async_cow_submit, async_cow_free);
+		btrfs_init_work(&async_chunk[i].work, compress_file_range,
+				submit_compressed_extents, async_cow_free);
 
 		nr_pages = DIV_ROUND_UP(cur_end - start, PAGE_SIZE);
 		atomic_add(nr_pages, &fs_info->async_delalloc_pages);
 
 		btrfs_queue_work(fs_info->delalloc_workers, &async_chunk[i].work);
 
-		*nr_written += nr_pages;
 		start = cur_end + 1;
 	}
-	*page_started = 1;
 	return true;
 }
 
-static noinline int run_delalloc_zoned(struct btrfs_inode *inode,
-				       struct page *locked_page, u64 start,
-				       u64 end, int *page_started,
-				       unsigned long *nr_written,
-				       struct writeback_control *wbc)
+/*
+ * Run the delalloc range from start to end, and write back any dirty pages
+ * covered by the range.
+ */
+static noinline int run_delalloc_cow(struct btrfs_inode *inode,
+				     struct page *locked_page, u64 start,
+				     u64 end, struct writeback_control *wbc,
+				     bool pages_dirty)
 {
 	u64 done_offset = end;
 	int ret;
-	bool locked_page_done = false;
 
 	while (start <= end) {
-		ret = cow_file_range(inode, locked_page, start, end, page_started,
-				     nr_written, 0, &done_offset);
-		if (ret && ret != -EAGAIN)
+		ret = cow_file_range(inode, locked_page, start, end, &done_offset,
+				     true, false);
+		if (ret)
 			return ret;
-
-		if (*page_started) {
-			ASSERT(ret == 0);
-			return 0;
-		}
-
-		if (ret == 0)
-			done_offset = end;
-
-		if (done_offset == start) {
-			wait_on_bit_io(&inode->root->fs_info->flags,
-				       BTRFS_FS_NEED_ZONE_FINISH,
-				       TASK_UNINTERRUPTIBLE);
-			continue;
-		}
-
-		if (!locked_page_done) {
-			__set_page_dirty_nobuffers(locked_page);
-			account_page_redirty(locked_page);
-		}
-		locked_page_done = true;
-		extent_write_locked_range(&inode->vfs_inode, start, done_offset,
-					  wbc);
+		extent_write_locked_range(&inode->vfs_inode, locked_page, start,
+					  done_offset, wbc, pages_dirty);
 		start = done_offset + 1;
 	}
 
-	*page_started = 1;
-
-	return 0;
+	return 1;
 }
 
 static noinline int csum_exist_in_range(struct btrfs_fs_info *fs_info,
@@ -1894,8 +1747,7 @@ static noinline int csum_exist_in_range(struct btrfs_fs_info *fs_info,
 }
 
 static int fallback_to_cow(struct btrfs_inode *inode, struct page *locked_page,
-			   const u64 start, const u64 end,
-			   int *page_started, unsigned long *nr_written)
+			   const u64 start, const u64 end)
 {
 	const bool is_space_ino = btrfs_is_free_space_inode(inode);
 	const bool is_reloc_ino = btrfs_is_data_reloc_root(inode->root);
@@ -1903,6 +1755,7 @@ static int fallback_to_cow(struct btrfs_inode *inode, struct page *locked_page,
 	struct extent_io_tree *io_tree = &inode->io_tree;
 	u64 range_start = start;
 	u64 count;
+	int ret;
 
 	/*
 	 * If EXTENT_NORESERVE is set it means that when the buffered write was
@@ -1955,8 +1808,14 @@ static int fallback_to_cow(struct btrfs_inode *inode, struct page *locked_page,
 					 NULL);
 	}
 
-	return cow_file_range(inode, locked_page, start, end, page_started,
-			      nr_written, 1, NULL);
+	/*
+	 * Don't try to create inline extents, as a mix of inline extent that
+	 * is written out and unlocked directly and a normal NOCOW extent
+	 * doesn't work.
+	 */
+	ret = cow_file_range(inode, locked_page, start, end, NULL, false, true);
+	ASSERT(ret != 1);
+	return ret;
 }
 
 struct can_nocow_file_extent_args {
@@ -2105,9 +1964,7 @@ static int can_nocow_file_extent(struct btrfs_path *path,
  */
 static noinline int run_delalloc_nocow(struct btrfs_inode *inode,
 				       struct page *locked_page,
-				       const u64 start, const u64 end,
-				       int *page_started,
-				       unsigned long *nr_written)
+				       const u64 start, const u64 end)
 {
 	struct btrfs_fs_info *fs_info = inode->root->fs_info;
 	struct btrfs_root *root = inode->root;
@@ -2117,25 +1974,26 @@ static noinline int run_delalloc_nocow(struct btrfs_inode *inode,
 	int ret;
 	bool check_prev = true;
 	u64 ino = btrfs_ino(inode);
-	struct btrfs_block_group *bg;
-	bool nocow = false;
 	struct can_nocow_file_extent_args nocow_args = { 0 };
 
+	/*
+	 * Normally on a zoned device we're only doing COW writes, but in case
+	 * of relocation on a zoned filesystem serializes I/O so that we're only
+	 * writing sequentially and can end up here as well.
+	 */
+	ASSERT(!btrfs_is_zoned(fs_info) || btrfs_is_data_reloc_root(root));
+
 	path = btrfs_alloc_path();
 	if (!path) {
-		extent_clear_unlock_delalloc(inode, start, end, locked_page,
-					     EXTENT_LOCKED | EXTENT_DELALLOC |
-					     EXTENT_DO_ACCOUNTING |
-					     EXTENT_DEFRAG, PAGE_UNLOCK |
-					     PAGE_START_WRITEBACK |
-					     PAGE_END_WRITEBACK);
-		return -ENOMEM;
+		ret = -ENOMEM;
+		goto error;
 	}
 
 	nocow_args.end = end;
 	nocow_args.writeback_path = true;
 
 	while (1) {
+		struct btrfs_block_group *nocow_bg = NULL;
 		struct btrfs_ordered_extent *ordered;
 		struct btrfs_key found_key;
 		struct btrfs_file_extent_item *fi;
@@ -2146,8 +2004,6 @@ static noinline int run_delalloc_nocow(struct btrfs_inode *inode,
 		int extent_type;
 		bool is_prealloc;
 
-		nocow = false;
-
 		ret = btrfs_lookup_file_extent(NULL, root, path, ino,
 					       cur_offset, 0);
 		if (ret < 0)
@@ -2172,11 +2028,8 @@ next_slot:
 		leaf = path->nodes[0];
 		if (path->slots[0] >= btrfs_header_nritems(leaf)) {
 			ret = btrfs_next_leaf(root, path);
-			if (ret < 0) {
-				if (cow_start != (u64)-1)
-					cur_offset = cow_start;
+			if (ret < 0)
 				goto error;
-			}
 			if (ret > 0)
 				break;
 			leaf = path->nodes[0];
@@ -2209,7 +2062,7 @@ next_slot:
 		if (found_key.offset > cur_offset) {
 			extent_end = found_key.offset;
 			extent_type = 0;
-			goto out_check;
+			goto must_cow;
 		}
 
 		/*
@@ -2239,24 +2092,22 @@ next_slot:
 
 		nocow_args.start = cur_offset;
 		ret = can_nocow_file_extent(path, &found_key, inode, &nocow_args);
-		if (ret < 0) {
-			if (cow_start != (u64)-1)
-				cur_offset = cow_start;
+		if (ret < 0)
 			goto error;
-		} else if (ret == 0) {
-			goto out_check;
-		}
+		if (ret == 0)
+			goto must_cow;
 
 		ret = 0;
-		bg = btrfs_inc_nocow_writers(fs_info, nocow_args.disk_bytenr);
-		if (bg)
-			nocow = true;
-out_check:
-		/*
-		 * If nocow is false then record the beginning of the range
-		 * that needs to be COWed
-		 */
-		if (!nocow) {
+		nocow_bg = btrfs_inc_nocow_writers(fs_info, nocow_args.disk_bytenr);
+		if (!nocow_bg) {
+must_cow:
+			/*
+			 * If we can't perform NOCOW writeback for the range,
+			 * then record the beginning of the range that needs to
+			 * be COWed.  It will be written out before the next
+			 * NOCOW range if we find one, or when exiting this
+			 * loop.
+			 */
 			if (cow_start == (u64)-1)
 				cow_start = cur_offset;
 			cur_offset = extent_end;
@@ -2275,11 +2126,12 @@ out_check:
 		 */
 		if (cow_start != (u64)-1) {
 			ret = fallback_to_cow(inode, locked_page,
-					      cow_start, found_key.offset - 1,
-					      page_started, nr_written);
-			if (ret)
-				goto error;
+					      cow_start, found_key.offset - 1);
 			cow_start = (u64)-1;
+			if (ret) {
+				btrfs_dec_nocow_writers(nocow_bg);
+				goto error;
+			}
 		}
 
 		nocow_end = cur_offset + nocow_args.num_bytes - 1;
@@ -2296,6 +2148,7 @@ out_check:
 					  ram_bytes, BTRFS_COMPRESS_NONE,
 					  BTRFS_ORDERED_PREALLOC);
 			if (IS_ERR(em)) {
+				btrfs_dec_nocow_writers(nocow_bg);
 				ret = PTR_ERR(em);
 				goto error;
 			}
@@ -2309,6 +2162,7 @@ out_check:
 				? (1 << BTRFS_ORDERED_PREALLOC)
 				: (1 << BTRFS_ORDERED_NOCOW),
 				BTRFS_COMPRESS_NONE);
+		btrfs_dec_nocow_writers(nocow_bg);
 		if (IS_ERR(ordered)) {
 			if (is_prealloc) {
 				btrfs_drop_extent_map_range(inode, cur_offset,
@@ -2318,11 +2172,6 @@ out_check:
 			goto error;
 		}
 
-		if (nocow) {
-			btrfs_dec_nocow_writers(bg);
-			nocow = false;
-		}
-
 		if (btrfs_is_data_reloc_root(root))
 			/*
 			 * Error handled later, as we must prevent
@@ -2357,17 +2206,24 @@ out_check:
 
 	if (cow_start != (u64)-1) {
 		cur_offset = end;
-		ret = fallback_to_cow(inode, locked_page, cow_start, end,
-				      page_started, nr_written);
+		ret = fallback_to_cow(inode, locked_page, cow_start, end);
+		cow_start = (u64)-1;
 		if (ret)
 			goto error;
 	}
 
-error:
-	if (nocow)
-		btrfs_dec_nocow_writers(bg);
+	btrfs_free_path(path);
+	return 0;
 
-	if (ret && cur_offset < end)
+error:
+	/*
+	 * If an error happened while a COW region is outstanding, cur_offset
+	 * needs to be reset to cow_start to ensure the COW region is unlocked
+	 * as well.
+	 */
+	if (cow_start != (u64)-1)
+		cur_offset = cow_start;
+	if (cur_offset < end)
 		extent_clear_unlock_delalloc(inode, cur_offset, end,
 					     locked_page, EXTENT_LOCKED |
 					     EXTENT_DELALLOC | EXTENT_DEFRAG |
@@ -2395,49 +2251,37 @@ static bool should_nocow(struct btrfs_inode *inode, u64 start, u64 end)
  * being touched for the first time.
  */
 int btrfs_run_delalloc_range(struct btrfs_inode *inode, struct page *locked_page,
-		u64 start, u64 end, int *page_started, unsigned long *nr_written,
-		struct writeback_control *wbc)
+			     u64 start, u64 end, struct writeback_control *wbc)
 {
-	int ret = 0;
 	const bool zoned = btrfs_is_zoned(inode->root->fs_info);
+	int ret;
 
 	/*
-	 * The range must cover part of the @locked_page, or the returned
-	 * @page_started can confuse the caller.
+	 * The range must cover part of the @locked_page, or a return of 1
+	 * can confuse the caller.
 	 */
 	ASSERT(!(end <= page_offset(locked_page) ||
 		 start >= page_offset(locked_page) + PAGE_SIZE));
 
 	if (should_nocow(inode, start, end)) {
-		/*
-		 * Normally on a zoned device we're only doing COW writes, but
-		 * in case of relocation on a zoned filesystem we have taken
-		 * precaution, that we're only writing sequentially. It's safe
-		 * to use run_delalloc_nocow() here, like for  regular
-		 * preallocated inodes.
-		 */
-		ASSERT(!zoned || btrfs_is_data_reloc_root(inode->root));
-		ret = run_delalloc_nocow(inode, locked_page, start, end,
-					 page_started, nr_written);
+		ret = run_delalloc_nocow(inode, locked_page, start, end);
 		goto out;
 	}
 
 	if (btrfs_inode_can_compress(inode) &&
 	    inode_need_compress(inode, start, end) &&
-	    run_delalloc_compressed(inode, wbc, locked_page, start,
-				    end, page_started, nr_written))
-		goto out;
+	    run_delalloc_compressed(inode, locked_page, start, end, wbc))
+		return 1;
 
 	if (zoned)
-		ret = run_delalloc_zoned(inode, locked_page, start, end,
-					 page_started, nr_written, wbc);
+		ret = run_delalloc_cow(inode, locked_page, start, end, wbc,
+				       true);
 	else
-		ret = cow_file_range(inode, locked_page, start, end,
-				     page_started, nr_written, 1, NULL);
+		ret = cow_file_range(inode, locked_page, start, end, NULL,
+				     false, false);
 
 out:
-	ASSERT(ret <= 0);
-	if (ret)
+	if (ret < 0)
 		btrfs_cleanup_ordered_extents(inode, locked_page, start,
 					      end - start + 1);
 	return ret;
@@ -2840,23 +2684,19 @@ struct btrfs_writepage_fixup {
 
 static void btrfs_writepage_fixup_worker(struct btrfs_work *work)
 {
-	struct btrfs_writepage_fixup *fixup;
+	struct btrfs_writepage_fixup *fixup =
+		container_of(work, struct btrfs_writepage_fixup, work);
 	struct btrfs_ordered_extent *ordered;
 	struct extent_state *cached_state = NULL;
 	struct extent_changeset *data_reserved = NULL;
-	struct page *page;
-	struct btrfs_inode *inode;
-	u64 page_start;
-	u64 page_end;
+	struct page *page = fixup->page;
+	struct btrfs_inode *inode = fixup->inode;
+	struct btrfs_fs_info *fs_info = inode->root->fs_info;
+	u64 page_start = page_offset(page);
+	u64 page_end = page_offset(page) + PAGE_SIZE - 1;
 	int ret = 0;
 	bool free_delalloc_space = true;
 
-	fixup = container_of(work, struct btrfs_writepage_fixup, work);
-	page = fixup->page;
-	inode = fixup->inode;
-	page_start = page_offset(page);
-	page_end = page_offset(page) + PAGE_SIZE - 1;
-
 	/*
 	 * This is similar to page_mkwrite, we need to reserve the space before
 	 * we take the page lock.
@@ -2949,10 +2789,12 @@ out_page:
 		 * to reflect the errors and clean the page.
 		 */
 		mapping_set_error(page->mapping, ret);
-		end_extent_writepage(page, ret, page_start, page_end);
+		btrfs_mark_ordered_io_finished(inode, page, page_start,
+					       PAGE_SIZE, !ret);
+		btrfs_page_clear_uptodate(fs_info, page, page_start, PAGE_SIZE);
 		clear_page_dirty_for_io(page);
 	}
-	btrfs_page_clear_checked(inode->root->fs_info, page, page_start, PAGE_SIZE);
+	btrfs_page_clear_checked(fs_info, page, page_start, PAGE_SIZE);
 	unlock_page(page);
 	put_page(page);
 	kfree(fixup);
@@ -3359,6 +3201,13 @@ out:
 			btrfs_free_reserved_extent(fs_info,
 					ordered_extent->disk_bytenr,
 					ordered_extent->disk_num_bytes, 1);
+			/*
+			 * Actually free the qgroup rsv which was released when
+			 * the ordered extent was created.
+			 */
+			btrfs_qgroup_free_refroot(fs_info, inode->root->root_key.objectid,
+						  ordered_extent->qgroup_rsv,
+						  BTRFS_QGROUP_RSV_DATA);
 		}
 	}
 
@@ -3384,15 +3233,6 @@ int btrfs_finish_ordered_io(struct btrfs_ordered_extent *ordered)
 	return btrfs_finish_one_ordered(ordered);
 }
 
-void btrfs_writepage_endio_finish_ordered(struct btrfs_inode *inode,
-					  struct page *page, u64 start,
-					  u64 end, bool uptodate)
-{
-	trace_btrfs_writepage_end_io_hook(inode, start, end, uptodate);
-
-	btrfs_mark_ordered_io_finished(inode, page, start, end + 1 - start, uptodate);
-}
-
 /*
  * Verify the checksum for a single sector without any extra action that depend
  * on the type of I/O.
@@ -3662,9 +3502,16 @@ int btrfs_orphan_cleanup(struct btrfs_root *root)
 		 */
 
 		if (found_key.offset == last_objectid) {
+			/*
+			 * We found the same inode as before. This means we were
+			 * not able to remove its items via eviction triggered
+			 * by an iput(). A transaction abort may have happened,
+			 * due to -ENOSPC for example, so try to grab the error
+			 * that lead to a transaction abort, if any.
+			 */
 			btrfs_err(fs_info,
 				  "Error removing orphan entry, stopping orphan cleanup");
-			ret = -EINVAL;
+			ret = BTRFS_FS_ERROR(fs_info) ?: -EINVAL;
 			goto out;
 		}
 
@@ -5735,11 +5582,11 @@ struct inode *btrfs_iget(struct super_block *s, u64 ino, struct btrfs_root *root
 	return btrfs_iget_path(s, ino, root, NULL);
 }
 
-static struct inode *new_simple_dir(struct super_block *s,
+static struct inode *new_simple_dir(struct inode *dir,
 				    struct btrfs_key *key,
 				    struct btrfs_root *root)
 {
-	struct inode *inode = new_inode(s);
+	struct inode *inode = new_inode(dir->i_sb);
 
 	if (!inode)
 		return ERR_PTR(-ENOMEM);
@@ -5758,8 +5605,10 @@ static struct inode *new_simple_dir(struct super_block *s,
 	inode->i_fop = &simple_dir_operations;
 	inode->i_mode = S_IFDIR | S_IRUGO | S_IWUSR | S_IXUGO;
 	inode->i_mtime = inode_set_ctime_current(inode);
-	inode->i_atime = inode->i_mtime;
+	inode->i_atime = dir->i_atime;
 	BTRFS_I(inode)->i_otime = inode->i_mtime;
+	inode->i_uid = dir->i_uid;
+	inode->i_gid = dir->i_gid;
 
 	return inode;
 }
@@ -5818,7 +5667,7 @@ struct inode *btrfs_lookup_dentry(struct inode *dir, struct dentry *dentry)
 		if (ret != -ENOENT)
 			inode = ERR_PTR(ret);
 		else
-			inode = new_simple_dir(dir->i_sb, &location, root);
+			inode = new_simple_dir(dir, &location, root);
 	} else {
 		inode = btrfs_iget(dir->i_sb, location.objectid, sub_root);
 		btrfs_put_root(sub_root);
@@ -6003,8 +5852,8 @@ static int btrfs_real_readdir(struct file *file, struct dir_context *ctx)
 	struct btrfs_key found_key;
 	struct btrfs_path *path;
 	void *addr;
-	struct list_head ins_list;
-	struct list_head del_list;
+	LIST_HEAD(ins_list);
+	LIST_HEAD(del_list);
 	int ret;
 	char *name_ptr;
 	int name_len;
@@ -6023,8 +5872,6 @@ static int btrfs_real_readdir(struct file *file, struct dir_context *ctx)
 	addr = private->filldir_buf;
 	path->reada = READA_FORWARD;
 
-	INIT_LIST_HEAD(&ins_list);
-	INIT_LIST_HEAD(&del_list);
 	put = btrfs_readdir_get_delayed_items(inode, private->last_index,
 					      &ins_list, &del_list);
 
@@ -9364,14 +9211,11 @@ static int start_delalloc_inodes(struct btrfs_root *root,
 	struct btrfs_inode *binode;
 	struct inode *inode;
 	struct btrfs_delalloc_work *work, *next;
-	struct list_head works;
-	struct list_head splice;
+	LIST_HEAD(works);
+	LIST_HEAD(splice);
 	int ret = 0;
 	bool full_flush = wbc->nr_to_write == LONG_MAX;
 
-	INIT_LIST_HEAD(&works);
-	INIT_LIST_HEAD(&splice);
-
 	mutex_lock(&root->delalloc_mutex);
 	spin_lock(&root->delalloc_lock);
 	list_splice_init(&root->delalloc_inodes, &splice);
@@ -9459,14 +9303,12 @@ int btrfs_start_delalloc_roots(struct btrfs_fs_info *fs_info, long nr,
 		.range_end = LLONG_MAX,
 	};
 	struct btrfs_root *root;
-	struct list_head splice;
+	LIST_HEAD(splice);
 	int ret;
 
 	if (BTRFS_FS_ERROR(fs_info))
 		return -EROFS;
 
-	INIT_LIST_HEAD(&splice);
-
 	mutex_lock(&fs_info->delalloc_root_mutex);
 	spin_lock(&fs_info->delalloc_root_lock);
 	list_splice_init(&fs_info->delalloc_roots, &splice);
diff --git a/fs/btrfs/messages.c b/fs/btrfs/messages.c
index 23fc11af498a..7695decc7243 100644
--- a/fs/btrfs/messages.c
+++ b/fs/btrfs/messages.c
@@ -10,14 +10,13 @@
 #ifdef CONFIG_PRINTK
 
 #define STATE_STRING_PREFACE	": state "
-#define STATE_STRING_BUF_LEN	(sizeof(STATE_STRING_PREFACE) + BTRFS_FS_STATE_COUNT)
+#define STATE_STRING_BUF_LEN	(sizeof(STATE_STRING_PREFACE) + BTRFS_FS_STATE_COUNT + 1)
 
 /*
  * Characters to print to indicate error conditions or uncommon filesystem state.
  * RO is not an error.
  */
 static const char fs_state_chars[] = {
-	[BTRFS_FS_STATE_ERROR]			= 'E',
 	[BTRFS_FS_STATE_REMOUNTING]		= 'M',
 	[BTRFS_FS_STATE_RO]			= 0,
 	[BTRFS_FS_STATE_TRANS_ABORTED]		= 'A',
@@ -37,6 +36,11 @@ static void btrfs_state_to_string(const struct btrfs_fs_info *info, char *buf)
 	memcpy(curr, STATE_STRING_PREFACE, sizeof(STATE_STRING_PREFACE));
 	curr += sizeof(STATE_STRING_PREFACE) - 1;
 
+	if (BTRFS_FS_ERROR(info)) {
+		*curr++ = 'E';
+		states_printed = true;
+	}
+
 	for_each_set_bit(bit, &fs_state, sizeof(fs_state)) {
 		WARN_ON_ONCE(bit >= BTRFS_FS_STATE_COUNT);
 		if ((bit < BTRFS_FS_STATE_COUNT) && fs_state_chars[bit]) {
@@ -155,7 +159,7 @@ void __btrfs_handle_fs_error(struct btrfs_fs_info *fs_info, const char *function
 	 * Today we only save the error info to memory.  Long term we'll also
 	 * send it down to the disk.
 	 */
-	set_bit(BTRFS_FS_STATE_ERROR, &fs_info->fs_state);
+	WRITE_ONCE(fs_info->fs_error, errno);
 
 	/* Don't go through full error handling during mount. */
 	if (!(sb->s_flags & SB_BORN))
@@ -252,12 +256,6 @@ void __cold _btrfs_printk(const struct btrfs_fs_info *fs_info, const char *fmt,
 }
 #endif
 
-void __cold btrfs_print_v0_err(struct btrfs_fs_info *fs_info)
-{
-	btrfs_err(fs_info,
-"Unsupported V0 extent filesystem detected. Aborting. Please re-create your filesystem with a newer kernel");
-}
-
 #if BITS_PER_LONG == 32
 void __cold btrfs_warn_32bit_limit(struct btrfs_fs_info *fs_info)
 {
diff --git a/fs/btrfs/messages.h b/fs/btrfs/messages.h
index deedc1a168e2..1ae6f8e23e07 100644
--- a/fs/btrfs/messages.h
+++ b/fs/btrfs/messages.h
@@ -181,8 +181,6 @@ do {								\
 #define ASSERT(expr)	(void)(expr)
 #endif
 
-void __cold btrfs_print_v0_err(struct btrfs_fs_info *fs_info);
-
 __printf(5, 6)
 __cold
 void __btrfs_handle_fs_error(struct btrfs_fs_info *fs_info, const char *function,
diff --git a/fs/btrfs/ordered-data.c b/fs/btrfs/ordered-data.c
index a629532283bc..b46ab348e8e5 100644
--- a/fs/btrfs/ordered-data.c
+++ b/fs/btrfs/ordered-data.c
@@ -410,6 +410,10 @@ void btrfs_mark_ordered_io_finished(struct btrfs_inode *inode,
 	unsigned long flags;
 	u64 cur = file_offset;
 
+	trace_btrfs_writepage_end_io_hook(inode, file_offset,
+					  file_offset + num_bytes - 1,
+					  uptodate);
+
 	spin_lock_irqsave(&tree->lock, flags);
 	while (cur < file_offset + num_bytes) {
 		u64 entry_end;
@@ -736,11 +740,9 @@ void btrfs_wait_ordered_roots(struct btrfs_fs_info *fs_info, u64 nr,
 			     const u64 range_start, const u64 range_len)
 {
 	struct btrfs_root *root;
-	struct list_head splice;
+	LIST_HEAD(splice);
 	u64 done;
 
-	INIT_LIST_HEAD(&splice);
-
 	mutex_lock(&fs_info->ordered_operations_mutex);
 	spin_lock(&fs_info->ordered_root_lock);
 	list_splice_init(&fs_info->ordered_roots, &splice);
diff --git a/fs/btrfs/print-tree.c b/fs/btrfs/print-tree.c
index aa06d9ca911d..0c93439e929f 100644
--- a/fs/btrfs/print-tree.c
+++ b/fs/btrfs/print-tree.c
@@ -95,8 +95,10 @@ static void print_extent_item(const struct extent_buffer *eb, int slot, int type
 	int ref_index = 0;
 
 	if (unlikely(item_size < sizeof(*ei))) {
-		btrfs_print_v0_err(eb->fs_info);
-		btrfs_handle_fs_error(eb->fs_info, -EINVAL, NULL);
+		btrfs_err(eb->fs_info,
+			  "unexpected extent item size, has %u expect >= %zu",
+			  item_size, sizeof(*ei));
+		btrfs_handle_fs_error(eb->fs_info, -EUCLEAN, NULL);
 	}
 
 	ei = btrfs_item_ptr(eb, slot, struct btrfs_extent_item);
@@ -291,10 +293,6 @@ void btrfs_print_leaf(const struct extent_buffer *l)
 			       btrfs_file_extent_num_bytes(l, fi),
 			       btrfs_file_extent_ram_bytes(l, fi));
 			break;
-		case BTRFS_EXTENT_REF_V0_KEY:
-			btrfs_print_v0_err(fs_info);
-			btrfs_handle_fs_error(fs_info, -EINVAL, NULL);
-			break;
 		case BTRFS_BLOCK_GROUP_ITEM_KEY:
 			bi = btrfs_item_ptr(l, i,
 					    struct btrfs_block_group_item);
diff --git a/fs/btrfs/qgroup.c b/fs/btrfs/qgroup.c
index 2637d6b157ff..b99230db3c82 100644
--- a/fs/btrfs/qgroup.c
+++ b/fs/btrfs/qgroup.c
@@ -3590,15 +3590,16 @@ btrfs_qgroup_rescan(struct btrfs_fs_info *fs_info)
 	 * going to clear all tracking information for a clean start.
 	 */
 
-	trans = btrfs_join_transaction(fs_info->fs_root);
-	if (IS_ERR(trans)) {
+	trans = btrfs_attach_transaction_barrier(fs_info->fs_root);
+	if (IS_ERR(trans) && trans != ERR_PTR(-ENOENT)) {
 		fs_info->qgroup_flags &= ~BTRFS_QGROUP_STATUS_FLAG_RESCAN;
 		return PTR_ERR(trans);
-	}
-	ret = btrfs_commit_transaction(trans);
-	if (ret) {
-		fs_info->qgroup_flags &= ~BTRFS_QGROUP_STATUS_FLAG_RESCAN;
-		return ret;
+	} else if (trans != ERR_PTR(-ENOENT)) {
+		ret = btrfs_commit_transaction(trans);
+		if (ret) {
+			fs_info->qgroup_flags &= ~BTRFS_QGROUP_STATUS_FLAG_RESCAN;
+			return ret;
+		}
 	}
 
 	qgroup_rescan_zero_tracking(fs_info);
@@ -3757,9 +3758,11 @@ static int try_flush_qgroup(struct btrfs_root *root)
 		goto out;
 	btrfs_wait_ordered_extents(root, U64_MAX, 0, (u64)-1);
 
-	trans = btrfs_join_transaction(root);
+	trans = btrfs_attach_transaction_barrier(root);
 	if (IS_ERR(trans)) {
 		ret = PTR_ERR(trans);
+		if (ret == -ENOENT)
+			ret = 0;
 		goto out;
 	}
 
diff --git a/fs/btrfs/raid56.c b/fs/btrfs/raid56.c
index 0249ea52bb80..3e014b9370a3 100644
--- a/fs/btrfs/raid56.c
+++ b/fs/btrfs/raid56.c
@@ -584,8 +584,7 @@ static int rbio_can_merge(struct btrfs_raid_bio *last,
 	if (last->operation == BTRFS_RBIO_PARITY_SCRUB)
 		return 0;
 
-	if (last->operation == BTRFS_RBIO_REBUILD_MISSING ||
-	    last->operation == BTRFS_RBIO_READ_REBUILD)
+	if (last->operation == BTRFS_RBIO_READ_REBUILD)
 		return 0;
 
 	return 1;
@@ -784,10 +783,7 @@ static noinline void unlock_stripe(struct btrfs_raid_bio *rbio)
 			spin_unlock(&rbio->bio_list_lock);
 			spin_unlock(&h->lock);
 
-			if (next->operation == BTRFS_RBIO_READ_REBUILD)
-				start_async_work(next, recover_rbio_work_locked);
-			else if (next->operation == BTRFS_RBIO_REBUILD_MISSING) {
-				steal_rbio(rbio, next);
+			if (next->operation == BTRFS_RBIO_READ_REBUILD) {
 				start_async_work(next, recover_rbio_work_locked);
 			} else if (next->operation == BTRFS_RBIO_WRITE) {
 				steal_rbio(rbio, next);
@@ -1517,11 +1513,11 @@ static void submit_read_wait_bio_list(struct btrfs_raid_bio *rbio,
 	while ((bio = bio_list_pop(bio_list))) {
 		bio->bi_end_io = raid_wait_read_end_io;
 
-		if (trace_raid56_scrub_read_recover_enabled()) {
+		if (trace_raid56_read_enabled()) {
 			struct raid56_bio_trace_info trace_info = { 0 };
 
 			bio_get_trace_info(rbio, bio, &trace_info);
-			trace_raid56_scrub_read_recover(rbio, bio, &trace_info);
+			trace_raid56_read(rbio, bio, &trace_info);
 		}
 		submit_bio(bio);
 	}
@@ -1698,8 +1694,7 @@ static int verify_one_sector(struct btrfs_raid_bio *rbio,
 	 * If we're rebuilding a read, we have to use pages from the
 	 * bio list if possible.
 	 */
-	if ((rbio->operation == BTRFS_RBIO_READ_REBUILD ||
-	     rbio->operation == BTRFS_RBIO_REBUILD_MISSING)) {
+	if (rbio->operation == BTRFS_RBIO_READ_REBUILD) {
 		sector = sector_in_rbio(rbio, stripe_nr, sector_nr, 0);
 	} else {
 		sector = rbio_stripe_sector(rbio, stripe_nr, sector_nr);
@@ -1763,8 +1758,7 @@ static int recover_vertical(struct btrfs_raid_bio *rbio, int sector_nr,
 		 * If we're rebuilding a read, we have to use pages from the
 		 * bio list if possible.
 		 */
-		if ((rbio->operation == BTRFS_RBIO_READ_REBUILD ||
-		     rbio->operation == BTRFS_RBIO_REBUILD_MISSING)) {
+		if (rbio->operation == BTRFS_RBIO_READ_REBUILD) {
 			sector = sector_in_rbio(rbio, stripe_nr, sector_nr, 0);
 		} else {
 			sector = rbio_stripe_sector(rbio, stripe_nr, sector_nr);
@@ -1897,8 +1891,7 @@ static int recover_sectors(struct btrfs_raid_bio *rbio)
 		goto out;
 	}
 
-	if (rbio->operation == BTRFS_RBIO_READ_REBUILD ||
-	    rbio->operation == BTRFS_RBIO_REBUILD_MISSING) {
+	if (rbio->operation == BTRFS_RBIO_READ_REBUILD) {
 		spin_lock(&rbio->bio_list_lock);
 		set_bit(RBIO_RMW_LOCKED_BIT, &rbio->flags);
 		spin_unlock(&rbio->bio_list_lock);
@@ -2112,8 +2105,8 @@ static void fill_data_csums(struct btrfs_raid_bio *rbio)
 		goto error;
 	}
 
-	ret = btrfs_lookup_csums_bitmap(csum_root, start, start + len - 1,
-					rbio->csum_buf, rbio->csum_bitmap, false);
+	ret = btrfs_lookup_csums_bitmap(csum_root, NULL, start, start + len - 1,
+					rbio->csum_buf, rbio->csum_bitmap);
 	if (ret < 0)
 		goto error;
 	if (bitmap_empty(rbio->csum_bitmap, len >> fs_info->sectorsize_bits))
@@ -2198,11 +2191,11 @@ static void submit_write_bios(struct btrfs_raid_bio *rbio,
 	while ((bio = bio_list_pop(bio_list))) {
 		bio->bi_end_io = raid_wait_write_end_io;
 
-		if (trace_raid56_write_stripe_enabled()) {
+		if (trace_raid56_write_enabled()) {
 			struct raid56_bio_trace_info trace_info = { 0 };
 
 			bio_get_trace_info(rbio, bio, &trace_info);
-			trace_raid56_write_stripe(rbio, bio, &trace_info);
+			trace_raid56_write(rbio, bio, &trace_info);
 		}
 		submit_bio(bio);
 	}
diff --git a/fs/btrfs/raid56.h b/fs/btrfs/raid56.h
index 0e84c9c9293f..45e6ff78316f 100644
--- a/fs/btrfs/raid56.h
+++ b/fs/btrfs/raid56.h
@@ -14,7 +14,6 @@ enum btrfs_rbio_ops {
 	BTRFS_RBIO_WRITE,
 	BTRFS_RBIO_READ_REBUILD,
 	BTRFS_RBIO_PARITY_SCRUB,
-	BTRFS_RBIO_REBUILD_MISSING,
 };
 
 struct btrfs_raid_bio {
diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index 46c3c1d57266..9951a0caf5bb 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -3006,9 +3006,6 @@ static int relocate_one_page(struct inode *inode, struct file_ra_state *ra,
 		if (!page)
 			return -ENOMEM;
 	}
-	ret = set_page_extent_mapped(page);
-	if (ret < 0)
-		goto release_page;
 
 	if (PageReadahead(page))
 		page_cache_async_readahead(inode->i_mapping, ra, NULL,
@@ -3024,6 +3021,15 @@ static int relocate_one_page(struct inode *inode, struct file_ra_state *ra,
 		}
 	}
 
+	/*
+	 * We could have lost page private when we dropped the lock to read the
+	 * page above, make sure we set_page_extent_mapped here so we have any
+	 * of the subpage blocksize stuff we need in place.
+	 */
+	ret = set_page_extent_mapped(page);
+	if (ret < 0)
+		goto release_page;
+
 	page_start = page_offset(page);
 	page_end = page_start + PAGE_SIZE - 1;
 
@@ -3250,12 +3256,13 @@ static int add_tree_block(struct reloc_control *rc,
 			if (type == BTRFS_TREE_BLOCK_REF_KEY)
 				owner = btrfs_extent_inline_ref_offset(eb, iref);
 		}
-	} else if (unlikely(item_size == sizeof(struct btrfs_extent_item_v0))) {
-		btrfs_print_v0_err(eb->fs_info);
-		btrfs_handle_fs_error(eb->fs_info, -EINVAL, NULL);
-		return -EINVAL;
 	} else {
-		BUG();
+		btrfs_print_leaf(eb);
+		btrfs_err(rc->block_group->fs_info,
+			  "unrecognized tree backref at tree block %llu slot %u",
+			  eb->start, path->slots[0]);
+		btrfs_release_path(path);
+		return -EUCLEAN;
 	}
 
 	btrfs_release_path(path);
@@ -3498,6 +3505,8 @@ int find_next_extent(struct reloc_control *rc, struct btrfs_path *path,
 
 	last = rc->block_group->start + rc->block_group->length;
 	while (1) {
+		bool block_found;
+
 		cond_resched();
 		if (rc->search_start >= last) {
 			ret = 1;
@@ -3548,11 +3557,11 @@ next:
 			goto next;
 		}
 
-		ret = find_first_extent_bit(&rc->processed_blocks,
-					    key.objectid, &start, &end,
-					    EXTENT_DIRTY, NULL);
+		block_found = find_first_extent_bit(&rc->processed_blocks,
+						    key.objectid, &start, &end,
+						    EXTENT_DIRTY, NULL);
 
-		if (ret == 0 && start <= key.objectid) {
+		if (block_found && start <= key.objectid) {
 			btrfs_release_path(path);
 			rc->search_start = end + 1;
 		} else {
diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c
index 7289f5bff397..b877203f1dc5 100644
--- a/fs/btrfs/scrub.c
+++ b/fs/btrfs/scrub.c
@@ -43,9 +43,20 @@ struct scrub_ctx;
 /*
  * The following value only influences the performance.
  *
- * This determines the batch size for stripe submitted in one go.
+ * This detemines how many stripes would be submitted in one go,
+ * which is 512KiB (BTRFS_STRIPE_LEN * SCRUB_STRIPES_PER_GROUP).
  */
-#define SCRUB_STRIPES_PER_SCTX	8	/* That would be 8 64K stripe per-device. */
+#define SCRUB_STRIPES_PER_GROUP		8
+
+/*
+ * How many groups we have for each sctx.
+ *
+ * This would be 8M per device, the same value as the old scrub in-flight bios
+ * size limit.
+ */
+#define SCRUB_GROUPS_PER_SCTX		16
+
+#define SCRUB_TOTAL_STRIPES		(SCRUB_GROUPS_PER_SCTX * SCRUB_STRIPES_PER_GROUP)
 
 /*
  * The following value times PAGE_SIZE needs to be large enough to match the
@@ -172,9 +183,11 @@ struct scrub_stripe {
 };
 
 struct scrub_ctx {
-	struct scrub_stripe	stripes[SCRUB_STRIPES_PER_SCTX];
+	struct scrub_stripe	stripes[SCRUB_TOTAL_STRIPES];
 	struct scrub_stripe	*raid56_data_stripes;
 	struct btrfs_fs_info	*fs_info;
+	struct btrfs_path	extent_path;
+	struct btrfs_path	csum_path;
 	int			first_free;
 	int			cur_stripe;
 	atomic_t		cancel_req;
@@ -315,10 +328,10 @@ static noinline_for_stack void scrub_free_ctx(struct scrub_ctx *sctx)
 	if (!sctx)
 		return;
 
-	for (i = 0; i < SCRUB_STRIPES_PER_SCTX; i++)
+	for (i = 0; i < SCRUB_TOTAL_STRIPES; i++)
 		release_scrub_stripe(&sctx->stripes[i]);
 
-	kfree(sctx);
+	kvfree(sctx);
 }
 
 static void scrub_put_ctx(struct scrub_ctx *sctx)
@@ -333,13 +346,20 @@ static noinline_for_stack struct scrub_ctx *scrub_setup_ctx(
 	struct scrub_ctx *sctx;
 	int		i;
 
-	sctx = kzalloc(sizeof(*sctx), GFP_KERNEL);
+	/* Since sctx has inline 128 stripes, it can go beyond 64K easily.  Use
+	 * kvzalloc().
+	 */
+	sctx = kvzalloc(sizeof(*sctx), GFP_KERNEL);
 	if (!sctx)
 		goto nomem;
 	refcount_set(&sctx->refs, 1);
 	sctx->is_dev_replace = is_dev_replace;
 	sctx->fs_info = fs_info;
-	for (i = 0; i < SCRUB_STRIPES_PER_SCTX; i++) {
+	sctx->extent_path.search_commit_root = 1;
+	sctx->extent_path.skip_locking = 1;
+	sctx->csum_path.search_commit_root = 1;
+	sctx->csum_path.skip_locking = 1;
+	for (i = 0; i < SCRUB_TOTAL_STRIPES; i++) {
 		int ret;
 
 		ret = init_scrub_stripe(fs_info, &sctx->stripes[i]);
@@ -970,6 +990,9 @@ skip:
 	spin_unlock(&sctx->stat_lock);
 }
 
+static void scrub_write_sectors(struct scrub_ctx *sctx, struct scrub_stripe *stripe,
+				unsigned long write_bitmap, bool dev_replace);
+
 /*
  * The main entrance for all read related scrub work, including:
  *
@@ -978,13 +1001,16 @@ skip:
  * - Go through the remaining mirrors and try to read as large blocksize as
  *   possible
  * - Go through all mirrors (including the failed mirror) sector-by-sector
+ * - Submit writeback for repaired sectors
  *
- * Writeback does not happen here, it needs extra synchronization.
+ * Writeback for dev-replace does not happen here, it needs extra
+ * synchronization for zoned devices.
  */
 static void scrub_stripe_read_repair_worker(struct work_struct *work)
 {
 	struct scrub_stripe *stripe = container_of(work, struct scrub_stripe, work);
-	struct btrfs_fs_info *fs_info = stripe->bg->fs_info;
+	struct scrub_ctx *sctx = stripe->sctx;
+	struct btrfs_fs_info *fs_info = sctx->fs_info;
 	int num_copies = btrfs_num_copies(fs_info, stripe->bg->start,
 					  stripe->bg->length);
 	int mirror;
@@ -1049,7 +1075,23 @@ static void scrub_stripe_read_repair_worker(struct work_struct *work)
 			goto out;
 	}
 out:
-	scrub_stripe_report_errors(stripe->sctx, stripe);
+	/*
+	 * Submit the repaired sectors.  For zoned case, we cannot do repair
+	 * in-place, but queue the bg to be relocated.
+	 */
+	if (btrfs_is_zoned(fs_info)) {
+		if (!bitmap_empty(&stripe->error_bitmap, stripe->nr_sectors))
+			btrfs_repair_one_zone(fs_info, sctx->stripes[0].bg->start);
+	} else if (!sctx->readonly) {
+		unsigned long repaired;
+
+		bitmap_andnot(&repaired, &stripe->init_error_bitmap,
+			      &stripe->error_bitmap, stripe->nr_sectors);
+		scrub_write_sectors(sctx, stripe, repaired, false);
+		wait_scrub_stripe_io(stripe);
+	}
+
+	scrub_stripe_report_errors(sctx, stripe);
 	set_bit(SCRUB_STRIPE_FLAG_REPAIR_DONE, &stripe->state);
 	wake_up(&stripe->repair_wait);
 }
@@ -1262,7 +1304,6 @@ static int get_raid56_logic_offset(u64 physical, int num,
 
 		/* Work out the disk rotation on this stripe-set */
 		rot = stripe_nr % map->num_stripes;
-		stripe_nr /= map->num_stripes;
 		/* calculate which stripe this data locates */
 		rot += i;
 		stripe_index = rot % map->num_stripes;
@@ -1468,6 +1509,8 @@ static void scrub_stripe_reset_bitmaps(struct scrub_stripe *stripe)
  * Return <0 for error.
  */
 static int scrub_find_fill_first_stripe(struct btrfs_block_group *bg,
+					struct btrfs_path *extent_path,
+					struct btrfs_path *csum_path,
 					struct btrfs_device *dev, u64 physical,
 					int mirror_num, u64 logical_start,
 					u32 logical_len,
@@ -1477,7 +1520,6 @@ static int scrub_find_fill_first_stripe(struct btrfs_block_group *bg,
 	struct btrfs_root *extent_root = btrfs_extent_root(fs_info, bg->start);
 	struct btrfs_root *csum_root = btrfs_csum_root(fs_info, bg->start);
 	const u64 logical_end = logical_start + logical_len;
-	struct btrfs_path path = { 0 };
 	u64 cur_logical = logical_start;
 	u64 stripe_end;
 	u64 extent_start;
@@ -1493,14 +1535,13 @@ static int scrub_find_fill_first_stripe(struct btrfs_block_group *bg,
 	/* The range must be inside the bg. */
 	ASSERT(logical_start >= bg->start && logical_end <= bg->start + bg->length);
 
-	path.search_commit_root = 1;
-	path.skip_locking = 1;
-
-	ret = find_first_extent_item(extent_root, &path, logical_start, logical_len);
+	ret = find_first_extent_item(extent_root, extent_path, logical_start,
+				     logical_len);
 	/* Either error or not found. */
 	if (ret)
 		goto out;
-	get_extent_info(&path, &extent_start, &extent_len, &extent_flags, &extent_gen);
+	get_extent_info(extent_path, &extent_start, &extent_len, &extent_flags,
+			&extent_gen);
 	if (extent_flags & BTRFS_EXTENT_FLAG_TREE_BLOCK)
 		stripe->nr_meta_extents++;
 	if (extent_flags & BTRFS_EXTENT_FLAG_DATA)
@@ -1528,7 +1569,7 @@ static int scrub_find_fill_first_stripe(struct btrfs_block_group *bg,
 
 	/* Fill the extent info for the remaining sectors. */
 	while (cur_logical <= stripe_end) {
-		ret = find_first_extent_item(extent_root, &path, cur_logical,
+		ret = find_first_extent_item(extent_root, extent_path, cur_logical,
 					     stripe_end - cur_logical + 1);
 		if (ret < 0)
 			goto out;
@@ -1536,7 +1577,7 @@ static int scrub_find_fill_first_stripe(struct btrfs_block_group *bg,
 			ret = 0;
 			break;
 		}
-		get_extent_info(&path, &extent_start, &extent_len,
+		get_extent_info(extent_path, &extent_start, &extent_len,
 				&extent_flags, &extent_gen);
 		if (extent_flags & BTRFS_EXTENT_FLAG_TREE_BLOCK)
 			stripe->nr_meta_extents++;
@@ -1561,9 +1602,9 @@ static int scrub_find_fill_first_stripe(struct btrfs_block_group *bg,
 		 */
 		ASSERT(BITS_PER_LONG >= BTRFS_STRIPE_LEN >> fs_info->sectorsize_bits);
 
-		ret = btrfs_lookup_csums_bitmap(csum_root, stripe->logical,
-						stripe_end, stripe->csums,
-						&csum_bitmap, true);
+		ret = btrfs_lookup_csums_bitmap(csum_root, csum_path,
+						stripe->logical, stripe_end,
+						stripe->csums, &csum_bitmap);
 		if (ret < 0)
 			goto out;
 		if (ret > 0)
@@ -1576,7 +1617,6 @@ static int scrub_find_fill_first_stripe(struct btrfs_block_group *bg,
 	}
 	set_bit(SCRUB_STRIPE_FLAG_INITIALIZED, &stripe->state);
 out:
-	btrfs_release_path(&path);
 	return ret;
 }
 
@@ -1654,6 +1694,28 @@ static bool stripe_has_metadata_error(struct scrub_stripe *stripe)
 	return false;
 }
 
+static void submit_initial_group_read(struct scrub_ctx *sctx,
+				      unsigned int first_slot,
+				      unsigned int nr_stripes)
+{
+	struct blk_plug plug;
+
+	ASSERT(first_slot < SCRUB_TOTAL_STRIPES);
+	ASSERT(first_slot + nr_stripes <= SCRUB_TOTAL_STRIPES);
+
+	scrub_throttle_dev_io(sctx, sctx->stripes[0].dev,
+			      btrfs_stripe_nr_to_offset(nr_stripes));
+	blk_start_plug(&plug);
+	for (int i = 0; i < nr_stripes; i++) {
+		struct scrub_stripe *stripe = &sctx->stripes[first_slot + i];
+
+		/* Those stripes should be initialized. */
+		ASSERT(test_bit(SCRUB_STRIPE_FLAG_INITIALIZED, &stripe->state));
+		scrub_submit_initial_read(sctx, stripe);
+	}
+	blk_finish_plug(&plug);
+}
+
 static int flush_scrub_stripes(struct scrub_ctx *sctx)
 {
 	struct btrfs_fs_info *fs_info = sctx->fs_info;
@@ -1666,11 +1728,11 @@ static int flush_scrub_stripes(struct scrub_ctx *sctx)
 
 	ASSERT(test_bit(SCRUB_STRIPE_FLAG_INITIALIZED, &sctx->stripes[0].state));
 
-	scrub_throttle_dev_io(sctx, sctx->stripes[0].dev,
-			      btrfs_stripe_nr_to_offset(nr_stripes));
-	for (int i = 0; i < nr_stripes; i++) {
-		stripe = &sctx->stripes[i];
-		scrub_submit_initial_read(sctx, stripe);
+	/* Submit the stripes which are populated but not submitted. */
+	if (nr_stripes % SCRUB_STRIPES_PER_GROUP) {
+		const int first_slot = round_down(nr_stripes, SCRUB_STRIPES_PER_GROUP);
+
+		submit_initial_group_read(sctx, first_slot, nr_stripes - first_slot);
 	}
 
 	for (int i = 0; i < nr_stripes; i++) {
@@ -1680,32 +1742,6 @@ static int flush_scrub_stripes(struct scrub_ctx *sctx)
 			   test_bit(SCRUB_STRIPE_FLAG_REPAIR_DONE, &stripe->state));
 	}
 
-	/*
-	 * Submit the repaired sectors.  For zoned case, we cannot do repair
-	 * in-place, but queue the bg to be relocated.
-	 */
-	if (btrfs_is_zoned(fs_info)) {
-		for (int i = 0; i < nr_stripes; i++) {
-			stripe = &sctx->stripes[i];
-
-			if (!bitmap_empty(&stripe->error_bitmap, stripe->nr_sectors)) {
-				btrfs_repair_one_zone(fs_info,
-						      sctx->stripes[0].bg->start);
-				break;
-			}
-		}
-	} else if (!sctx->readonly) {
-		for (int i = 0; i < nr_stripes; i++) {
-			unsigned long repaired;
-
-			stripe = &sctx->stripes[i];
-
-			bitmap_andnot(&repaired, &stripe->init_error_bitmap,
-				      &stripe->error_bitmap, stripe->nr_sectors);
-			scrub_write_sectors(sctx, stripe, repaired, false);
-		}
-	}
-
 	/* Submit for dev-replace. */
 	if (sctx->is_dev_replace) {
 		/*
@@ -1750,28 +1786,40 @@ static void raid56_scrub_wait_endio(struct bio *bio)
 
 static int queue_scrub_stripe(struct scrub_ctx *sctx, struct btrfs_block_group *bg,
 			      struct btrfs_device *dev, int mirror_num,
-			      u64 logical, u32 length, u64 physical)
+			      u64 logical, u32 length, u64 physical,
+			      u64 *found_logical_ret)
 {
 	struct scrub_stripe *stripe;
 	int ret;
 
-	/* No available slot, submit all stripes and wait for them. */
-	if (sctx->cur_stripe >= SCRUB_STRIPES_PER_SCTX) {
-		ret = flush_scrub_stripes(sctx);
-		if (ret < 0)
-			return ret;
-	}
+	/*
+	 * There should always be one slot left, as caller filling the last
+	 * slot should flush them all.
+	 */
+	ASSERT(sctx->cur_stripe < SCRUB_TOTAL_STRIPES);
 
 	stripe = &sctx->stripes[sctx->cur_stripe];
-
-	/* We can queue one stripe using the remaining slot. */
 	scrub_reset_stripe(stripe);
-	ret = scrub_find_fill_first_stripe(bg, dev, physical, mirror_num,
-					   logical, length, stripe);
+	ret = scrub_find_fill_first_stripe(bg, &sctx->extent_path,
+					   &sctx->csum_path, dev, physical,
+					   mirror_num, logical, length, stripe);
 	/* Either >0 as no more extents or <0 for error. */
 	if (ret)
 		return ret;
+	if (found_logical_ret)
+		*found_logical_ret = stripe->logical;
 	sctx->cur_stripe++;
+
+	/* We filled one group, submit it. */
+	if (sctx->cur_stripe % SCRUB_STRIPES_PER_GROUP == 0) {
+		const int first_slot = sctx->cur_stripe - SCRUB_STRIPES_PER_GROUP;
+
+		submit_initial_group_read(sctx, first_slot, SCRUB_STRIPES_PER_GROUP);
+	}
+
+	/* Last slot used, flush them all. */
+	if (sctx->cur_stripe == SCRUB_TOTAL_STRIPES)
+		return flush_scrub_stripes(sctx);
 	return 0;
 }
 
@@ -1785,6 +1833,8 @@ static int scrub_raid56_parity_stripe(struct scrub_ctx *sctx,
 	struct btrfs_fs_info *fs_info = sctx->fs_info;
 	struct btrfs_raid_bio *rbio;
 	struct btrfs_io_context *bioc = NULL;
+	struct btrfs_path extent_path = { 0 };
+	struct btrfs_path csum_path = { 0 };
 	struct bio *bio;
 	struct scrub_stripe *stripe;
 	bool all_empty = true;
@@ -1795,6 +1845,16 @@ static int scrub_raid56_parity_stripe(struct scrub_ctx *sctx,
 
 	ASSERT(sctx->raid56_data_stripes);
 
+	/*
+	 * For data stripe search, we cannot re-use the same extent/csum paths,
+	 * as the data stripe bytenr may be smaller than previous extent.  Thus
+	 * we have to use our own extent/csum paths.
+	 */
+	extent_path.search_commit_root = 1;
+	extent_path.skip_locking = 1;
+	csum_path.search_commit_root = 1;
+	csum_path.skip_locking = 1;
+
 	for (int i = 0; i < data_stripes; i++) {
 		int stripe_index;
 		int rot;
@@ -1809,7 +1869,7 @@ static int scrub_raid56_parity_stripe(struct scrub_ctx *sctx,
 
 		scrub_reset_stripe(stripe);
 		set_bit(SCRUB_STRIPE_FLAG_NO_REPORT, &stripe->state);
-		ret = scrub_find_fill_first_stripe(bg,
+		ret = scrub_find_fill_first_stripe(bg, &extent_path, &csum_path,
 				map->stripes[stripe_index].dev, physical, 1,
 				full_stripe_start + btrfs_stripe_nr_to_offset(i),
 				BTRFS_STRIPE_LEN, stripe);
@@ -1854,24 +1914,6 @@ static int scrub_raid56_parity_stripe(struct scrub_ctx *sctx,
 	/* For now, no zoned support for RAID56. */
 	ASSERT(!btrfs_is_zoned(sctx->fs_info));
 
-	/* Writeback for the repaired sectors. */
-	for (int i = 0; i < data_stripes; i++) {
-		unsigned long repaired;
-
-		stripe = &sctx->raid56_data_stripes[i];
-
-		bitmap_andnot(&repaired, &stripe->init_error_bitmap,
-			      &stripe->error_bitmap, stripe->nr_sectors);
-		scrub_write_sectors(sctx, stripe, repaired, false);
-	}
-
-	/* Wait for the above writebacks to finish. */
-	for (int i = 0; i < data_stripes; i++) {
-		stripe = &sctx->raid56_data_stripes[i];
-
-		wait_scrub_stripe_io(stripe);
-	}
-
 	/*
 	 * Now all data stripes are properly verified. Check if we have any
 	 * unrepaired, if so abort immediately or we could further corrupt the
@@ -1937,6 +1979,8 @@ static int scrub_raid56_parity_stripe(struct scrub_ctx *sctx,
 	bio_put(bio);
 	btrfs_bio_counter_dec(fs_info);
 
+	btrfs_release_path(&extent_path);
+	btrfs_release_path(&csum_path);
 out:
 	return ret;
 }
@@ -1958,18 +2002,15 @@ static int scrub_simple_mirror(struct scrub_ctx *sctx,
 {
 	struct btrfs_fs_info *fs_info = sctx->fs_info;
 	const u64 logical_end = logical_start + logical_length;
-	/* An artificial limit, inherit from old scrub behavior */
-	struct btrfs_path path = { 0 };
 	u64 cur_logical = logical_start;
 	int ret;
 
 	/* The range must be inside the bg */
 	ASSERT(logical_start >= bg->start && logical_end <= bg->start + bg->length);
 
-	path.search_commit_root = 1;
-	path.skip_locking = 1;
 	/* Go through each extent items inside the logical range */
 	while (cur_logical < logical_end) {
+		u64 found_logical;
 		u64 cur_physical = physical + cur_logical - logical_start;
 
 		/* Canceled? */
@@ -1994,7 +2035,7 @@ static int scrub_simple_mirror(struct scrub_ctx *sctx,
 
 		ret = queue_scrub_stripe(sctx, bg, device, mirror_num,
 					 cur_logical, logical_end - cur_logical,
-					 cur_physical);
+					 cur_physical, &found_logical);
 		if (ret > 0) {
 			/* No more extent, just update the accounting */
 			sctx->stat.last_physical = physical + logical_length;
@@ -2004,14 +2045,11 @@ static int scrub_simple_mirror(struct scrub_ctx *sctx,
 		if (ret < 0)
 			break;
 
-		ASSERT(sctx->cur_stripe > 0);
-		cur_logical = sctx->stripes[sctx->cur_stripe - 1].logical
-			      + BTRFS_STRIPE_LEN;
+		cur_logical = found_logical + BTRFS_STRIPE_LEN;
 
 		/* Don't hold CPU for too long time */
 		cond_resched();
 	}
-	btrfs_release_path(&path);
 	return ret;
 }
 
@@ -2109,6 +2147,9 @@ static noinline_for_stack int scrub_stripe(struct scrub_ctx *sctx,
 	u64 stripe_logical;
 	int stop_loop = 0;
 
+	/* Extent_path should be released by now. */
+	ASSERT(sctx->extent_path.nodes[0] == NULL);
+
 	scrub_blocked_if_needed(fs_info);
 
 	if (sctx->is_dev_replace &&
@@ -2227,6 +2268,9 @@ out:
 	ret2 = flush_scrub_stripes(sctx);
 	if (!ret)
 		ret = ret2;
+	btrfs_release_path(&sctx->extent_path);
+	btrfs_release_path(&sctx->csum_path);
+
 	if (sctx->raid56_data_stripes) {
 		for (int i = 0; i < nr_data_stripes(map); i++)
 			release_scrub_stripe(&sctx->raid56_data_stripes[i]);
@@ -2711,8 +2755,7 @@ static void scrub_workers_put(struct btrfs_fs_info *fs_info)
 /*
  * get a reference count on fs_info->scrub_workers. start worker if necessary
  */
-static noinline_for_stack int scrub_workers_get(struct btrfs_fs_info *fs_info,
-						int is_dev_replace)
+static noinline_for_stack int scrub_workers_get(struct btrfs_fs_info *fs_info)
 {
 	struct workqueue_struct *scrub_workers = NULL;
 	unsigned int flags = WQ_FREEZABLE | WQ_UNBOUND;
@@ -2722,10 +2765,7 @@ static noinline_for_stack int scrub_workers_get(struct btrfs_fs_info *fs_info,
 	if (refcount_inc_not_zero(&fs_info->scrub_workers_refcnt))
 		return 0;
 
-	if (is_dev_replace)
-		scrub_workers = alloc_ordered_workqueue("btrfs-scrub", flags);
-	else
-		scrub_workers = alloc_workqueue("btrfs-scrub", flags, max_active);
+	scrub_workers = alloc_workqueue("btrfs-scrub", flags, max_active);
 	if (!scrub_workers)
 		return -ENOMEM;
 
@@ -2777,7 +2817,7 @@ int btrfs_scrub_dev(struct btrfs_fs_info *fs_info, u64 devid, u64 start,
 	if (IS_ERR(sctx))
 		return PTR_ERR(sctx);
 
-	ret = scrub_workers_get(fs_info, is_dev_replace);
+	ret = scrub_workers_get(fs_info);
 	if (ret)
 		goto out_free_ctx;
 
diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c
index 8bfd44750efe..3a566150c531 100644
--- a/fs/btrfs/send.c
+++ b/fs/btrfs/send.c
@@ -3685,7 +3685,7 @@ static void tail_append_pending_moves(struct send_ctx *sctx,
 static int apply_children_dir_moves(struct send_ctx *sctx)
 {
 	struct pending_dir_move *pm;
-	struct list_head stack;
+	LIST_HEAD(stack);
 	u64 parent_ino = sctx->cur_ino;
 	int ret = 0;
 
@@ -3693,7 +3693,6 @@ static int apply_children_dir_moves(struct send_ctx *sctx)
 	if (!pm)
 		return 0;
 
-	INIT_LIST_HEAD(&stack);
 	tail_append_pending_moves(sctx, pm, &stack);
 
 	while (!list_empty(&stack)) {
@@ -4165,7 +4164,7 @@ static int process_recorded_refs(struct send_ctx *sctx, int *pending_move)
 	int ret = 0;
 	struct recorded_ref *cur;
 	struct recorded_ref *cur2;
-	struct list_head check_dirs;
+	LIST_HEAD(check_dirs);
 	struct fs_path *valid_path = NULL;
 	u64 ow_inode = 0;
 	u64 ow_gen;
@@ -4184,7 +4183,6 @@ static int process_recorded_refs(struct send_ctx *sctx, int *pending_move)
 	 * which is always '..'
 	 */
 	BUG_ON(sctx->cur_ino <= BTRFS_FIRST_FREE_OBJECTID);
-	INIT_LIST_HEAD(&check_dirs);
 
 	valid_path = fs_path_alloc();
 	if (!valid_path) {
diff --git a/fs/btrfs/space-info.c b/fs/btrfs/space-info.c
index 75e7fa337e66..d7e8cd4f140c 100644
--- a/fs/btrfs/space-info.c
+++ b/fs/btrfs/space-info.c
@@ -389,11 +389,7 @@ int btrfs_can_overcommit(struct btrfs_fs_info *fs_info,
 		return 0;
 
 	used = btrfs_space_info_used(space_info, true);
-	if (test_bit(BTRFS_FS_ACTIVE_ZONE_TRACKING, &fs_info->flags) &&
-	    (space_info->flags & BTRFS_BLOCK_GROUP_METADATA))
-		avail = 0;
-	else
-		avail = calc_available_free_space(fs_info, space_info, flush);
+	avail = calc_available_free_space(fs_info, space_info, flush);
 
 	if (used + bytes < space_info->total_bytes + avail)
 		return 1;
@@ -510,6 +506,7 @@ void btrfs_dump_space_info(struct btrfs_fs_info *fs_info,
 			   int dump_block_groups)
 {
 	struct btrfs_block_group *cache;
+	u64 total_avail = 0;
 	int index = 0;
 
 	spin_lock(&info->lock);
@@ -523,18 +520,27 @@ void btrfs_dump_space_info(struct btrfs_fs_info *fs_info,
 	down_read(&info->groups_sem);
 again:
 	list_for_each_entry(cache, &info->block_groups[index], list) {
+		u64 avail;
+
 		spin_lock(&cache->lock);
+		avail = cache->length - cache->used - cache->pinned -
+			cache->reserved - cache->delalloc_bytes -
+			cache->bytes_super - cache->zone_unusable;
 		btrfs_info(fs_info,
-			"block group %llu has %llu bytes, %llu used %llu pinned %llu reserved %llu zone_unusable %s",
-			cache->start, cache->length, cache->used, cache->pinned,
-			cache->reserved, cache->zone_unusable,
-			cache->ro ? "[readonly]" : "");
+"block group %llu has %llu bytes, %llu used %llu pinned %llu reserved %llu delalloc %llu super %llu zone_unusable (%llu bytes available) %s",
+			   cache->start, cache->length, cache->used, cache->pinned,
+			   cache->reserved, cache->delalloc_bytes,
+			   cache->bytes_super, cache->zone_unusable,
+			   avail, cache->ro ? "[readonly]" : "");
 		spin_unlock(&cache->lock);
 		btrfs_dump_free_space(cache, bytes);
+		total_avail += avail;
 	}
 	if (++index < BTRFS_NR_RAID_TYPES)
 		goto again;
 	up_read(&info->groups_sem);
+
+	btrfs_info(fs_info, "%llu bytes available across all block groups", total_avail);
 }
 
 static inline u64 calc_reclaim_items_nr(const struct btrfs_fs_info *fs_info,
@@ -715,9 +721,11 @@ static void flush_space(struct btrfs_fs_info *fs_info,
 		else
 			nr = -1;
 
-		trans = btrfs_join_transaction(root);
+		trans = btrfs_join_transaction_nostart(root);
 		if (IS_ERR(trans)) {
 			ret = PTR_ERR(trans);
+			if (ret == -ENOENT)
+				ret = 0;
 			break;
 		}
 		ret = btrfs_run_delayed_items_nr(trans, nr);
@@ -733,9 +741,11 @@ static void flush_space(struct btrfs_fs_info *fs_info,
 		break;
 	case FLUSH_DELAYED_REFS_NR:
 	case FLUSH_DELAYED_REFS:
-		trans = btrfs_join_transaction(root);
+		trans = btrfs_join_transaction_nostart(root);
 		if (IS_ERR(trans)) {
 			ret = PTR_ERR(trans);
+			if (ret == -ENOENT)
+				ret = 0;
 			break;
 		}
 		if (state == FLUSH_DELAYED_REFS_NR)
@@ -747,18 +757,6 @@ static void flush_space(struct btrfs_fs_info *fs_info,
 		break;
 	case ALLOC_CHUNK:
 	case ALLOC_CHUNK_FORCE:
-		/*
-		 * For metadata space on zoned filesystem, reaching here means we
-		 * don't have enough space left in active_total_bytes. Try to
-		 * activate a block group first, because we may have inactive
-		 * block group already allocated.
-		 */
-		ret = btrfs_zoned_activate_one_bg(fs_info, space_info, false);
-		if (ret < 0)
-			break;
-		else if (ret == 1)
-			break;
-
 		trans = btrfs_join_transaction(root);
 		if (IS_ERR(trans)) {
 			ret = PTR_ERR(trans);
@@ -770,22 +768,6 @@ static void flush_space(struct btrfs_fs_info *fs_info,
 					CHUNK_ALLOC_FORCE);
 		btrfs_end_transaction(trans);
 
-		/*
-		 * For metadata space on zoned filesystem, allocating a new chunk
-		 * is not enough. We still need to activate the block * group.
-		 * Active the newly allocated block group by (maybe) finishing
-		 * a block group.
-		 */
-		if (ret == 1) {
-			ret = btrfs_zoned_activate_one_bg(fs_info, space_info, true);
-			/*
-			 * Revert to the original ret regardless we could finish
-			 * one block group or not.
-			 */
-			if (ret >= 0)
-				ret = 1;
-		}
-
 		if (ret > 0 || ret == -ENOSPC)
 			ret = 0;
 		break;
@@ -800,9 +782,18 @@ static void flush_space(struct btrfs_fs_info *fs_info,
 		break;
 	case COMMIT_TRANS:
 		ASSERT(current->journal_info == NULL);
-		trans = btrfs_join_transaction(root);
+		/*
+		 * We don't want to start a new transaction, just attach to the
+		 * current one or wait it fully commits in case its commit is
+		 * happening at the moment. Note: we don't use a nostart join
+		 * because that does not wait for a transaction to fully commit
+		 * (only for it to be unblocked, state TRANS_STATE_UNBLOCKED).
+		 */
+		trans = btrfs_attach_transaction_barrier(root);
 		if (IS_ERR(trans)) {
 			ret = PTR_ERR(trans);
+			if (ret == -ENOENT)
+				ret = 0;
 			break;
 		}
 		ret = btrfs_commit_transaction(trans);
@@ -1408,8 +1399,18 @@ static void priority_reclaim_metadata_space(struct btrfs_fs_info *fs_info,
 		}
 	}
 
-	/* Attempt to steal from the global rsv if we can. */
-	if (!steal_from_global_rsv(fs_info, space_info, ticket)) {
+	/*
+	 * Attempt to steal from the global rsv if we can, except if the fs was
+	 * turned into error mode due to a transaction abort when flushing space
+	 * above, in that case fail with the abort error instead of returning
+	 * success to the caller if we can steal from the global rsv - this is
+	 * just to have caller fail immeditelly instead of later when trying to
+	 * modify the fs, making it easier to debug -ENOSPC problems.
+	 */
+	if (BTRFS_FS_ERROR(fs_info)) {
+		ticket->error = BTRFS_FS_ERROR(fs_info);
+		remove_ticket(space_info, ticket);
+	} else if (!steal_from_global_rsv(fs_info, space_info, ticket)) {
 		ticket->error = -ENOSPC;
 		remove_ticket(space_info, ticket);
 	}
diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c
index 8eda51b095c9..09bfe68d2ea3 100644
--- a/fs/btrfs/super.c
+++ b/fs/btrfs/super.c
@@ -709,12 +709,16 @@ int btrfs_parse_options(struct btrfs_fs_info *info, char *options,
 			break;
 #ifdef CONFIG_BTRFS_FS_CHECK_INTEGRITY
 		case Opt_check_integrity_including_extent_data:
+			btrfs_warn(info,
+	"integrity checker is deprecated and will be removed in 6.7");
 			btrfs_info(info,
 				   "enabling check integrity including extent data");
 			btrfs_set_opt(info->mount_opt, CHECK_INTEGRITY_DATA);
 			btrfs_set_opt(info->mount_opt, CHECK_INTEGRITY);
 			break;
 		case Opt_check_integrity:
+			btrfs_warn(info,
+	"integrity checker is deprecated and will be removed in 6.7");
 			btrfs_info(info, "enabling check integrity");
 			btrfs_set_opt(info->mount_opt, CHECK_INTEGRITY);
 			break;
@@ -727,6 +731,8 @@ int btrfs_parse_options(struct btrfs_fs_info *info, char *options,
 				goto out;
 			}
 			info->check_integrity_print_mask = intarg;
+			btrfs_warn(info,
+	"integrity checker is deprecated and will be removed in 6.7");
 			btrfs_info(info, "check_integrity_print_mask 0x%x",
 				   info->check_integrity_print_mask);
 			break;
diff --git a/fs/btrfs/sysfs.c b/fs/btrfs/sysfs.c
index 25294e624851..b1d1ac25237b 100644
--- a/fs/btrfs/sysfs.c
+++ b/fs/btrfs/sysfs.c
@@ -414,6 +414,12 @@ static ssize_t supported_sectorsizes_show(struct kobject *kobj,
 BTRFS_ATTR(static_feature, supported_sectorsizes,
 	   supported_sectorsizes_show);
 
+static ssize_t acl_show(struct kobject *kobj, struct kobj_attribute *a, char *buf)
+{
+	return sysfs_emit(buf, "%d\n", !!IS_ENABLED(CONFIG_BTRFS_FS_POSIX_ACL));
+}
+BTRFS_ATTR(static_feature, acl, acl_show);
+
 /*
  * Features which only depend on kernel version.
  *
@@ -421,6 +427,7 @@ BTRFS_ATTR(static_feature, supported_sectorsizes,
  * btrfs_supported_feature_attrs.
  */
 static struct attribute *btrfs_supported_static_feature_attrs[] = {
+	BTRFS_ATTR_PTR(static_feature, acl),
 	BTRFS_ATTR_PTR(static_feature, rmdir_subvol),
 	BTRFS_ATTR_PTR(static_feature, supported_checksums),
 	BTRFS_ATTR_PTR(static_feature, send_stream_version),
diff --git a/fs/btrfs/tests/extent-io-tests.c b/fs/btrfs/tests/extent-io-tests.c
index f6bc6d738555..1cc86af97dc6 100644
--- a/fs/btrfs/tests/extent-io-tests.c
+++ b/fs/btrfs/tests/extent-io-tests.c
@@ -319,86 +319,139 @@ out:
 	return ret;
 }
 
-static int check_eb_bitmap(unsigned long *bitmap, struct extent_buffer *eb,
-			   unsigned long len)
+static int check_eb_bitmap(unsigned long *bitmap, struct extent_buffer *eb)
 {
 	unsigned long i;
 
-	for (i = 0; i < len * BITS_PER_BYTE; i++) {
+	for (i = 0; i < eb->len * BITS_PER_BYTE; i++) {
 		int bit, bit1;
 
 		bit = !!test_bit(i, bitmap);
 		bit1 = !!extent_buffer_test_bit(eb, 0, i);
 		if (bit1 != bit) {
-			test_err("bits do not match");
+			u8 has;
+			u8 expect;
+
+			read_extent_buffer(eb, &has, i / BITS_PER_BYTE, 1);
+			expect = bitmap_get_value8(bitmap, ALIGN(i, BITS_PER_BYTE));
+
+			test_err(
+		"bits do not match, start byte 0 bit %lu, byte %lu has 0x%02x expect 0x%02x",
+				 i, i / BITS_PER_BYTE, has, expect);
 			return -EINVAL;
 		}
 
 		bit1 = !!extent_buffer_test_bit(eb, i / BITS_PER_BYTE,
 						i % BITS_PER_BYTE);
 		if (bit1 != bit) {
-			test_err("offset bits do not match");
+			u8 has;
+			u8 expect;
+
+			read_extent_buffer(eb, &has, i / BITS_PER_BYTE, 1);
+			expect = bitmap_get_value8(bitmap, ALIGN(i, BITS_PER_BYTE));
+
+			test_err(
+		"bits do not match, start byte %lu bit %lu, byte %lu has 0x%02x expect 0x%02x",
+				 i / BITS_PER_BYTE, i % BITS_PER_BYTE,
+				 i / BITS_PER_BYTE, has, expect);
 			return -EINVAL;
 		}
 	}
 	return 0;
 }
 
-static int __test_eb_bitmaps(unsigned long *bitmap, struct extent_buffer *eb,
-			     unsigned long len)
+static int test_bitmap_set(const char *name, unsigned long *bitmap,
+			   struct extent_buffer *eb,
+			   unsigned long byte_start, unsigned long bit_start,
+			   unsigned long bit_len)
+{
+	int ret;
+
+	bitmap_set(bitmap, byte_start * BITS_PER_BYTE + bit_start, bit_len);
+	extent_buffer_bitmap_set(eb, byte_start, bit_start, bit_len);
+	ret = check_eb_bitmap(bitmap, eb);
+	if (ret < 0)
+		test_err("%s test failed", name);
+	return ret;
+}
+
+static int test_bitmap_clear(const char *name, unsigned long *bitmap,
+			     struct extent_buffer *eb,
+			     unsigned long byte_start, unsigned long bit_start,
+			     unsigned long bit_len)
+{
+	int ret;
+
+	bitmap_clear(bitmap, byte_start * BITS_PER_BYTE + bit_start, bit_len);
+	extent_buffer_bitmap_clear(eb, byte_start, bit_start, bit_len);
+	ret = check_eb_bitmap(bitmap, eb);
+	if (ret < 0)
+		test_err("%s test failed", name);
+	return ret;
+}
+static int __test_eb_bitmaps(unsigned long *bitmap, struct extent_buffer *eb)
 {
 	unsigned long i, j;
+	unsigned long byte_len = eb->len;
 	u32 x;
 	int ret;
 
-	memset(bitmap, 0, len);
-	memzero_extent_buffer(eb, 0, len);
-	if (memcmp_extent_buffer(eb, bitmap, 0, len) != 0) {
-		test_err("bitmap was not zeroed");
-		return -EINVAL;
-	}
+	ret = test_bitmap_clear("clear all run 1", bitmap, eb, 0, 0,
+				byte_len * BITS_PER_BYTE);
+	if (ret < 0)
+		return ret;
 
-	bitmap_set(bitmap, 0, len * BITS_PER_BYTE);
-	extent_buffer_bitmap_set(eb, 0, 0, len * BITS_PER_BYTE);
-	ret = check_eb_bitmap(bitmap, eb, len);
-	if (ret) {
-		test_err("setting all bits failed");
+	ret = test_bitmap_set("set all", bitmap, eb, 0, 0, byte_len * BITS_PER_BYTE);
+	if (ret < 0)
 		return ret;
-	}
 
-	bitmap_clear(bitmap, 0, len * BITS_PER_BYTE);
-	extent_buffer_bitmap_clear(eb, 0, 0, len * BITS_PER_BYTE);
-	ret = check_eb_bitmap(bitmap, eb, len);
-	if (ret) {
-		test_err("clearing all bits failed");
+	ret = test_bitmap_clear("clear all run 2", bitmap, eb, 0, 0,
+				byte_len * BITS_PER_BYTE);
+	if (ret < 0)
+		return ret;
+
+	ret = test_bitmap_set("same byte set", bitmap, eb, 0, 2, 4);
+	if (ret < 0)
+		return ret;
+
+	ret = test_bitmap_clear("same byte partial clear", bitmap, eb, 0, 4, 1);
+	if (ret < 0)
+		return ret;
+
+	ret = test_bitmap_set("cross byte set", bitmap, eb, 2, 4, 8);
+	if (ret < 0)
+		return ret;
+
+	ret = test_bitmap_set("cross multi byte set", bitmap, eb, 4, 4, 24);
+	if (ret < 0)
+		return ret;
+
+	ret = test_bitmap_clear("cross byte clear", bitmap, eb, 2, 6, 4);
+	if (ret < 0)
+		return ret;
+
+	ret = test_bitmap_clear("cross multi byte clear", bitmap, eb, 4, 6, 20);
+	if (ret < 0)
 		return ret;
-	}
 
 	/* Straddling pages test */
-	if (len > PAGE_SIZE) {
-		bitmap_set(bitmap,
-			(PAGE_SIZE - sizeof(long) / 2) * BITS_PER_BYTE,
-			sizeof(long) * BITS_PER_BYTE);
-		extent_buffer_bitmap_set(eb, PAGE_SIZE - sizeof(long) / 2, 0,
-					sizeof(long) * BITS_PER_BYTE);
-		ret = check_eb_bitmap(bitmap, eb, len);
-		if (ret) {
-			test_err("setting straddling pages failed");
+	if (byte_len > PAGE_SIZE) {
+		ret = test_bitmap_set("cross page set", bitmap, eb,
+				      PAGE_SIZE - sizeof(long) / 2, 0,
+				      sizeof(long) * BITS_PER_BYTE);
+		if (ret < 0)
+			return ret;
+
+		ret = test_bitmap_set("cross page set all", bitmap, eb, 0, 0,
+				      byte_len * BITS_PER_BYTE);
+		if (ret < 0)
 			return ret;
-		}
 
-		bitmap_set(bitmap, 0, len * BITS_PER_BYTE);
-		bitmap_clear(bitmap,
-			(PAGE_SIZE - sizeof(long) / 2) * BITS_PER_BYTE,
-			sizeof(long) * BITS_PER_BYTE);
-		extent_buffer_bitmap_set(eb, 0, 0, len * BITS_PER_BYTE);
-		extent_buffer_bitmap_clear(eb, PAGE_SIZE - sizeof(long) / 2, 0,
+		ret = test_bitmap_clear("cross page clear", bitmap, eb,
+					PAGE_SIZE - sizeof(long) / 2, 0,
 					sizeof(long) * BITS_PER_BYTE);
-		ret = check_eb_bitmap(bitmap, eb, len);
-		if (ret) {
-			test_err("clearing straddling pages failed");
+		if (ret < 0)
 			return ret;
-		}
 	}
 
 	/*
@@ -406,9 +459,12 @@ static int __test_eb_bitmaps(unsigned long *bitmap, struct extent_buffer *eb,
 	 * something repetitive that could miss some hypothetical off-by-n bug.
 	 */
 	x = 0;
-	bitmap_clear(bitmap, 0, len * BITS_PER_BYTE);
-	extent_buffer_bitmap_clear(eb, 0, 0, len * BITS_PER_BYTE);
-	for (i = 0; i < len * BITS_PER_BYTE / 32; i++) {
+	ret = test_bitmap_clear("clear all run 3", bitmap, eb, 0, 0,
+				byte_len * BITS_PER_BYTE);
+	if (ret < 0)
+		return ret;
+
+	for (i = 0; i < byte_len * BITS_PER_BYTE / 32; i++) {
 		x = (0x19660dULL * (u64)x + 0x3c6ef35fULL) & 0xffffffffU;
 		for (j = 0; j < 32; j++) {
 			if (x & (1U << j)) {
@@ -418,7 +474,7 @@ static int __test_eb_bitmaps(unsigned long *bitmap, struct extent_buffer *eb,
 		}
 	}
 
-	ret = check_eb_bitmap(bitmap, eb, len);
+	ret = check_eb_bitmap(bitmap, eb);
 	if (ret) {
 		test_err("random bit pattern failed");
 		return ret;
@@ -456,7 +512,7 @@ static int test_eb_bitmaps(u32 sectorsize, u32 nodesize)
 		goto out;
 	}
 
-	ret = __test_eb_bitmaps(bitmap, eb, nodesize);
+	ret = __test_eb_bitmaps(bitmap, eb);
 	if (ret)
 		goto out;
 
@@ -473,7 +529,7 @@ static int test_eb_bitmaps(u32 sectorsize, u32 nodesize)
 		goto out;
 	}
 
-	ret = __test_eb_bitmaps(bitmap, eb, nodesize);
+	ret = __test_eb_bitmaps(bitmap, eb);
 out:
 	free_extent_buffer(eb);
 	kfree(bitmap);
@@ -592,6 +648,146 @@ out:
 	return ret;
 }
 
+static void dump_eb_and_memory_contents(struct extent_buffer *eb, void *memory,
+					const char *test_name)
+{
+	for (int i = 0; i < eb->len; i++) {
+		struct page *page = eb->pages[i >> PAGE_SHIFT];
+		void *addr = page_address(page) + offset_in_page(i);
+
+		if (memcmp(addr, memory + i, 1) != 0) {
+			test_err("%s failed", test_name);
+			test_err("eb and memory diffs at byte %u, eb has 0x%02x memory has 0x%02x",
+				 i, *(u8 *)addr, *(u8 *)(memory + i));
+			return;
+		}
+	}
+}
+
+static int verify_eb_and_memory(struct extent_buffer *eb, void *memory,
+				const char *test_name)
+{
+	for (int i = 0; i < (eb->len >> PAGE_SHIFT); i++) {
+		void *eb_addr = page_address(eb->pages[i]);
+
+		if (memcmp(memory + (i << PAGE_SHIFT), eb_addr, PAGE_SIZE) != 0) {
+			dump_eb_and_memory_contents(eb, memory, test_name);
+			return -EUCLEAN;
+		}
+	}
+	return 0;
+}
+
+/*
+ * Init both memory and extent buffer contents to the same randomly generated
+ * contents.
+ */
+static void init_eb_and_memory(struct extent_buffer *eb, void *memory)
+{
+	get_random_bytes(memory, eb->len);
+	write_extent_buffer(eb, memory, 0, eb->len);
+}
+
+static int test_eb_mem_ops(u32 sectorsize, u32 nodesize)
+{
+	struct btrfs_fs_info *fs_info;
+	struct extent_buffer *eb = NULL;
+	void *memory = NULL;
+	int ret;
+
+	test_msg("running extent buffer memory operation tests");
+
+	fs_info = btrfs_alloc_dummy_fs_info(nodesize, sectorsize);
+	if (!fs_info) {
+		test_std_err(TEST_ALLOC_FS_INFO);
+		return -ENOMEM;
+	}
+
+	memory = kvzalloc(nodesize, GFP_KERNEL);
+	if (!memory) {
+		test_err("failed to allocate memory");
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	eb = __alloc_dummy_extent_buffer(fs_info, SZ_1M, nodesize);
+	if (!eb) {
+		test_std_err(TEST_ALLOC_EXTENT_BUFFER);
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	init_eb_and_memory(eb, memory);
+	ret = verify_eb_and_memory(eb, memory, "full eb write");
+	if (ret < 0)
+		goto out;
+
+	memcpy(memory, memory + 16, 16);
+	memcpy_extent_buffer(eb, 0, 16, 16);
+	ret = verify_eb_and_memory(eb, memory, "same page non-overlapping memcpy 1");
+	if (ret < 0)
+		goto out;
+
+	memcpy(memory, memory + 2048, 16);
+	memcpy_extent_buffer(eb, 0, 2048, 16);
+	ret = verify_eb_and_memory(eb, memory, "same page non-overlapping memcpy 2");
+	if (ret < 0)
+		goto out;
+	memcpy(memory, memory + 2048, 2048);
+	memcpy_extent_buffer(eb, 0, 2048, 2048);
+	ret = verify_eb_and_memory(eb, memory, "same page non-overlapping memcpy 3");
+	if (ret < 0)
+		goto out;
+
+	memmove(memory + 512, memory + 256, 512);
+	memmove_extent_buffer(eb, 512, 256, 512);
+	ret = verify_eb_and_memory(eb, memory, "same page overlapping memcpy 1");
+	if (ret < 0)
+		goto out;
+
+	memmove(memory + 2048, memory + 512, 2048);
+	memmove_extent_buffer(eb, 2048, 512, 2048);
+	ret = verify_eb_and_memory(eb, memory, "same page overlapping memcpy 2");
+	if (ret < 0)
+		goto out;
+	memmove(memory + 512, memory + 2048, 2048);
+	memmove_extent_buffer(eb, 512, 2048, 2048);
+	ret = verify_eb_and_memory(eb, memory, "same page overlapping memcpy 3");
+	if (ret < 0)
+		goto out;
+
+	if (nodesize > PAGE_SIZE) {
+		memcpy(memory, memory + 4096 - 128, 256);
+		memcpy_extent_buffer(eb, 0, 4096 - 128, 256);
+		ret = verify_eb_and_memory(eb, memory, "cross page non-overlapping memcpy 1");
+		if (ret < 0)
+			goto out;
+
+		memcpy(memory + 4096 - 128, memory + 4096 + 128, 256);
+		memcpy_extent_buffer(eb, 4096 - 128, 4096 + 128, 256);
+		ret = verify_eb_and_memory(eb, memory, "cross page non-overlapping memcpy 2");
+		if (ret < 0)
+			goto out;
+
+		memmove(memory + 4096 - 128, memory + 4096 - 64, 256);
+		memmove_extent_buffer(eb, 4096 - 128, 4096 - 64, 256);
+		ret = verify_eb_and_memory(eb, memory, "cross page overlapping memcpy 1");
+		if (ret < 0)
+			goto out;
+
+		memmove(memory + 4096 - 64, memory + 4096 - 128, 256);
+		memmove_extent_buffer(eb, 4096 - 64, 4096 - 128, 256);
+		ret = verify_eb_and_memory(eb, memory, "cross page overlapping memcpy 2");
+		if (ret < 0)
+			goto out;
+	}
+out:
+	free_extent_buffer(eb);
+	kvfree(memory);
+	btrfs_free_dummy_fs_info(fs_info);
+	return ret;
+}
+
 int btrfs_test_extent_io(u32 sectorsize, u32 nodesize)
 {
 	int ret;
@@ -607,6 +803,10 @@ int btrfs_test_extent_io(u32 sectorsize, u32 nodesize)
 		goto out;
 
 	ret = test_eb_bitmaps(sectorsize, nodesize);
+	if (ret)
+		goto out;
+
+	ret = test_eb_mem_ops(sectorsize, nodesize);
 out:
 	return ret;
 }
diff --git a/fs/btrfs/tests/extent-map-tests.c b/fs/btrfs/tests/extent-map-tests.c
index ed0f36ae5346..29bdd08b241f 100644
--- a/fs/btrfs/tests/extent-map-tests.c
+++ b/fs/btrfs/tests/extent-map-tests.c
@@ -6,6 +6,7 @@
 #include <linux/types.h>
 #include "btrfs-tests.h"
 #include "../ctree.h"
+#include "../btrfs_inode.h"
 #include "../volumes.h"
 #include "../disk-io.h"
 #include "../block-group.h"
@@ -442,6 +443,406 @@ static int test_case_4(struct btrfs_fs_info *fs_info,
 	return ret;
 }
 
+static int add_compressed_extent(struct extent_map_tree *em_tree,
+				 u64 start, u64 len, u64 block_start)
+{
+	struct extent_map *em;
+	int ret;
+
+	em = alloc_extent_map();
+	if (!em) {
+		test_std_err(TEST_ALLOC_EXTENT_MAP);
+		return -ENOMEM;
+	}
+
+	em->start = start;
+	em->len = len;
+	em->block_start = block_start;
+	em->block_len = SZ_4K;
+	set_bit(EXTENT_FLAG_COMPRESSED, &em->flags);
+	write_lock(&em_tree->lock);
+	ret = add_extent_mapping(em_tree, em, 0);
+	write_unlock(&em_tree->lock);
+	free_extent_map(em);
+	if (ret < 0) {
+		test_err("cannot add extent map [%llu, %llu)", start, start + len);
+		return ret;
+	}
+
+	return 0;
+}
+
+struct extent_range {
+	u64 start;
+	u64 len;
+};
+
+/* The valid states of the tree after every drop, as described below. */
+struct extent_range valid_ranges[][7] = {
+	{
+	  { .start = 0,			.len = SZ_8K },		/* [0, 8K) */
+	  { .start = SZ_4K * 3,		.len = SZ_4K * 3},	/* [12k, 24k) */
+	  { .start = SZ_4K * 6,		.len = SZ_4K * 3},	/* [24k, 36k) */
+	  { .start = SZ_32K + SZ_4K,	.len = SZ_4K},		/* [36k, 40k) */
+	  { .start = SZ_4K * 10,	.len = SZ_4K * 6},	/* [40k, 64k) */
+	},
+	{
+	  { .start = 0,			.len = SZ_8K },		/* [0, 8K) */
+	  { .start = SZ_4K * 5,		.len = SZ_4K},		/* [20k, 24k) */
+	  { .start = SZ_4K * 6,		.len = SZ_4K * 3},	/* [24k, 36k) */
+	  { .start = SZ_32K + SZ_4K,	.len = SZ_4K},		/* [36k, 40k) */
+	  { .start = SZ_4K * 10,	.len = SZ_4K * 6},	/* [40k, 64k) */
+	},
+	{
+	  { .start = 0,			.len = SZ_8K },		/* [0, 8K) */
+	  { .start = SZ_4K * 5,		.len = SZ_4K},		/* [20k, 24k) */
+	  { .start = SZ_4K * 6,		.len = SZ_4K},		/* [24k, 28k) */
+	  { .start = SZ_32K,		.len = SZ_4K},		/* [32k, 36k) */
+	  { .start = SZ_32K + SZ_4K,	.len = SZ_4K},		/* [36k, 40k) */
+	  { .start = SZ_4K * 10,	.len = SZ_4K * 6},	/* [40k, 64k) */
+	},
+	{
+	  { .start = 0,			.len = SZ_8K},		/* [0, 8K) */
+	  { .start = SZ_4K * 5,		.len = SZ_4K},		/* [20k, 24k) */
+	  { .start = SZ_4K * 6,		.len = SZ_4K},		/* [24k, 28k) */
+	}
+};
+
+static int validate_range(struct extent_map_tree *em_tree, int index)
+{
+	struct rb_node *n;
+	int i;
+
+	for (i = 0, n = rb_first_cached(&em_tree->map);
+	     valid_ranges[index][i].len && n;
+	     i++, n = rb_next(n)) {
+		struct extent_map *entry = rb_entry(n, struct extent_map, rb_node);
+
+		if (entry->start != valid_ranges[index][i].start) {
+			test_err("mapping has start %llu expected %llu",
+				 entry->start, valid_ranges[index][i].start);
+			return -EINVAL;
+		}
+
+		if (entry->len != valid_ranges[index][i].len) {
+			test_err("mapping has len %llu expected %llu",
+				 entry->len, valid_ranges[index][i].len);
+			return -EINVAL;
+		}
+	}
+
+	/*
+	 * We exited because we don't have any more entries in the extent_map
+	 * but we still expect more valid entries.
+	 */
+	if (valid_ranges[index][i].len) {
+		test_err("missing an entry");
+		return -EINVAL;
+	}
+
+	/* We exited the loop but still have entries in the extent map. */
+	if (n) {
+		test_err("we have a left over entry in the extent map we didn't expect");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+/*
+ * Test scenario:
+ *
+ * Test the various edge cases of btrfs_drop_extent_map_range, create the
+ * following ranges
+ *
+ * [0, 12k)[12k, 24k)[24k, 36k)[36k, 40k)[40k,64k)
+ *
+ * And then we'll drop:
+ *
+ * [8k, 12k) - test the single front split
+ * [12k, 20k) - test the single back split
+ * [28k, 32k) - test the double split
+ * [32k, 64k) - test whole em dropping
+ *
+ * They'll have the EXTENT_FLAG_COMPRESSED flag set to keep the em tree from
+ * merging the em's.
+ */
+static int test_case_5(void)
+{
+	struct extent_map_tree *em_tree;
+	struct inode *inode;
+	u64 start, end;
+	int ret;
+
+	test_msg("Running btrfs_drop_extent_map_range tests");
+
+	inode = btrfs_new_test_inode();
+	if (!inode) {
+		test_std_err(TEST_ALLOC_INODE);
+		return -ENOMEM;
+	}
+
+	em_tree = &BTRFS_I(inode)->extent_tree;
+
+	/* [0, 12k) */
+	ret = add_compressed_extent(em_tree, 0, SZ_4K * 3, 0);
+	if (ret) {
+		test_err("cannot add extent range [0, 12K)");
+		goto out;
+	}
+
+	/* [12k, 24k) */
+	ret = add_compressed_extent(em_tree, SZ_4K * 3, SZ_4K * 3, SZ_4K);
+	if (ret) {
+		test_err("cannot add extent range [12k, 24k)");
+		goto out;
+	}
+
+	/* [24k, 36k) */
+	ret = add_compressed_extent(em_tree, SZ_4K * 6, SZ_4K * 3, SZ_8K);
+	if (ret) {
+		test_err("cannot add extent range [12k, 24k)");
+		goto out;
+	}
+
+	/* [36k, 40k) */
+	ret = add_compressed_extent(em_tree, SZ_32K + SZ_4K, SZ_4K, SZ_4K * 3);
+	if (ret) {
+		test_err("cannot add extent range [12k, 24k)");
+		goto out;
+	}
+
+	/* [40k, 64k) */
+	ret = add_compressed_extent(em_tree, SZ_4K * 10, SZ_4K * 6, SZ_16K);
+	if (ret) {
+		test_err("cannot add extent range [12k, 24k)");
+		goto out;
+	}
+
+	/* Drop [8k, 12k) */
+	start = SZ_8K;
+	end = (3 * SZ_4K) - 1;
+	btrfs_drop_extent_map_range(BTRFS_I(inode), start, end, false);
+	ret = validate_range(&BTRFS_I(inode)->extent_tree, 0);
+	if (ret)
+		goto out;
+
+	/* Drop [12k, 20k) */
+	start = SZ_4K * 3;
+	end = SZ_16K + SZ_4K - 1;
+	btrfs_drop_extent_map_range(BTRFS_I(inode), start, end, false);
+	ret = validate_range(&BTRFS_I(inode)->extent_tree, 1);
+	if (ret)
+		goto out;
+
+	/* Drop [28k, 32k) */
+	start = SZ_32K - SZ_4K;
+	end = SZ_32K - 1;
+	btrfs_drop_extent_map_range(BTRFS_I(inode), start, end, false);
+	ret = validate_range(&BTRFS_I(inode)->extent_tree, 2);
+	if (ret)
+		goto out;
+
+	/* Drop [32k, 64k) */
+	start = SZ_32K;
+	end = SZ_64K - 1;
+	btrfs_drop_extent_map_range(BTRFS_I(inode), start, end, false);
+	ret = validate_range(&BTRFS_I(inode)->extent_tree, 3);
+	if (ret)
+		goto out;
+out:
+	iput(inode);
+	return ret;
+}
+
+/*
+ * Test the btrfs_add_extent_mapping helper which will attempt to create an em
+ * for areas between two existing ems.  Validate it doesn't do this when there
+ * are two unmerged em's side by side.
+ */
+static int test_case_6(struct btrfs_fs_info *fs_info, struct extent_map_tree *em_tree)
+{
+	struct extent_map *em = NULL;
+	int ret;
+
+	ret = add_compressed_extent(em_tree, 0, SZ_4K, 0);
+	if (ret)
+		goto out;
+
+	ret = add_compressed_extent(em_tree, SZ_4K, SZ_4K, 0);
+	if (ret)
+		goto out;
+
+	em = alloc_extent_map();
+	if (!em) {
+		test_std_err(TEST_ALLOC_EXTENT_MAP);
+		return -ENOMEM;
+	}
+
+	em->start = SZ_4K;
+	em->len = SZ_4K;
+	em->block_start = SZ_16K;
+	em->block_len = SZ_16K;
+	write_lock(&em_tree->lock);
+	ret = btrfs_add_extent_mapping(fs_info, em_tree, &em, 0, SZ_8K);
+	write_unlock(&em_tree->lock);
+
+	if (ret != 0) {
+		test_err("got an error when adding our em: %d", ret);
+		goto out;
+	}
+
+	ret = -EINVAL;
+	if (em->start != 0) {
+		test_err("unexpected em->start at %llu, wanted 0", em->start);
+		goto out;
+	}
+	if (em->len != SZ_4K) {
+		test_err("unexpected em->len %llu, expected 4K", em->len);
+		goto out;
+	}
+	ret = 0;
+out:
+	free_extent_map(em);
+	free_extent_map_tree(em_tree);
+	return ret;
+}
+
+/*
+ * Regression test for btrfs_drop_extent_map_range.  Calling with skip_pinned ==
+ * true would mess up the start/end calculations and subsequent splits would be
+ * incorrect.
+ */
+static int test_case_7(void)
+{
+	struct extent_map_tree *em_tree;
+	struct extent_map *em;
+	struct inode *inode;
+	int ret;
+
+	test_msg("Running btrfs_drop_extent_cache with pinned");
+
+	inode = btrfs_new_test_inode();
+	if (!inode) {
+		test_std_err(TEST_ALLOC_INODE);
+		return -ENOMEM;
+	}
+
+	em_tree = &BTRFS_I(inode)->extent_tree;
+
+	em = alloc_extent_map();
+	if (!em) {
+		test_std_err(TEST_ALLOC_EXTENT_MAP);
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	/* [0, 16K), pinned */
+	em->start = 0;
+	em->len = SZ_16K;
+	em->block_start = 0;
+	em->block_len = SZ_4K;
+	set_bit(EXTENT_FLAG_PINNED, &em->flags);
+	write_lock(&em_tree->lock);
+	ret = add_extent_mapping(em_tree, em, 0);
+	write_unlock(&em_tree->lock);
+	if (ret < 0) {
+		test_err("couldn't add extent map");
+		goto out;
+	}
+	free_extent_map(em);
+
+	em = alloc_extent_map();
+	if (!em) {
+		test_std_err(TEST_ALLOC_EXTENT_MAP);
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	/* [32K, 48K), not pinned */
+	em->start = SZ_32K;
+	em->len = SZ_16K;
+	em->block_start = SZ_32K;
+	em->block_len = SZ_16K;
+	write_lock(&em_tree->lock);
+	ret = add_extent_mapping(em_tree, em, 0);
+	write_unlock(&em_tree->lock);
+	if (ret < 0) {
+		test_err("couldn't add extent map");
+		goto out;
+	}
+	free_extent_map(em);
+
+	/*
+	 * Drop [0, 36K) This should skip the [0, 4K) extent and then split the
+	 * [32K, 48K) extent.
+	 */
+	btrfs_drop_extent_map_range(BTRFS_I(inode), 0, (36 * SZ_1K) - 1, true);
+
+	/* Make sure our extent maps look sane. */
+	ret = -EINVAL;
+
+	em = lookup_extent_mapping(em_tree, 0, SZ_16K);
+	if (!em) {
+		test_err("didn't find an em at 0 as expected");
+		goto out;
+	}
+
+	if (em->start != 0) {
+		test_err("em->start is %llu, expected 0", em->start);
+		goto out;
+	}
+
+	if (em->len != SZ_16K) {
+		test_err("em->len is %llu, expected 16K", em->len);
+		goto out;
+	}
+
+	free_extent_map(em);
+
+	read_lock(&em_tree->lock);
+	em = lookup_extent_mapping(em_tree, SZ_16K, SZ_16K);
+	read_unlock(&em_tree->lock);
+	if (em) {
+		test_err("found an em when we weren't expecting one");
+		goto out;
+	}
+
+	read_lock(&em_tree->lock);
+	em = lookup_extent_mapping(em_tree, SZ_32K, SZ_16K);
+	read_unlock(&em_tree->lock);
+	if (!em) {
+		test_err("didn't find an em at 32K as expected");
+		goto out;
+	}
+
+	if (em->start != (36 * SZ_1K)) {
+		test_err("em->start is %llu, expected 36K", em->start);
+		goto out;
+	}
+
+	if (em->len != (12 * SZ_1K)) {
+		test_err("em->len is %llu, expected 12K", em->len);
+		goto out;
+	}
+
+	free_extent_map(em);
+
+	read_lock(&em_tree->lock);
+	em = lookup_extent_mapping(em_tree, 48 * SZ_1K, (u64)-1);
+	read_unlock(&em_tree->lock);
+	if (em) {
+		test_err("found an unexpected em above 48K");
+		goto out;
+	}
+
+	ret = 0;
+out:
+	free_extent_map(em);
+	iput(inode);
+	return ret;
+}
+
 struct rmap_test_vector {
 	u64 raid_type;
 	u64 physical_start;
@@ -619,6 +1020,17 @@ int btrfs_test_extent_map(void)
 	if (ret)
 		goto out;
 	ret = test_case_4(fs_info, em_tree);
+	if (ret)
+		goto out;
+	ret = test_case_5();
+	if (ret)
+		goto out;
+	ret = test_case_6(fs_info, em_tree);
+	if (ret)
+		goto out;
+	ret = test_case_7();
+	if (ret)
+		goto out;
 
 	test_msg("running rmap tests");
 	for (i = 0; i < ARRAY_SIZE(rmap_tests); i++) {
diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c
index a10e38c2609f..874e4394df86 100644
--- a/fs/btrfs/transaction.c
+++ b/fs/btrfs/transaction.c
@@ -292,10 +292,11 @@ loop:
 	spin_unlock(&fs_info->trans_lock);
 
 	/*
-	 * If we are ATTACH, we just want to catch the current transaction,
-	 * and commit it. If there is no transaction, just return ENOENT.
+	 * If we are ATTACH or TRANS_JOIN_NOSTART, we just want to catch the
+	 * current transaction, and commit it. If there is no transaction, just
+	 * return ENOENT.
 	 */
-	if (type == TRANS_ATTACH)
+	if (type == TRANS_ATTACH || type == TRANS_JOIN_NOSTART)
 		return -ENOENT;
 
 	/*
@@ -591,8 +592,13 @@ start_transaction(struct btrfs_root *root, unsigned int num_items,
 		u64 delayed_refs_bytes = 0;
 
 		qgroup_reserved = num_items * fs_info->nodesize;
-		ret = btrfs_qgroup_reserve_meta_pertrans(root, qgroup_reserved,
-				enforce_qgroups);
+		/*
+		 * Use prealloc for now, as there might be a currently running
+		 * transaction that could free this reserved space prematurely
+		 * by committing.
+		 */
+		ret = btrfs_qgroup_reserve_meta_prealloc(root, qgroup_reserved,
+							 enforce_qgroups, false);
 		if (ret)
 			return ERR_PTR(ret);
 
@@ -705,6 +711,14 @@ again:
 		h->reloc_reserved = reloc_reserved;
 	}
 
+	/*
+	 * Now that we have found a transaction to be a part of, convert the
+	 * qgroup reservation from prealloc to pertrans. A different transaction
+	 * can't race in and free our pertrans out from under us.
+	 */
+	if (qgroup_reserved)
+		btrfs_qgroup_convert_reserved_meta(root, qgroup_reserved);
+
 got_it:
 	if (!current->journal_info)
 		current->journal_info = h;
@@ -752,7 +766,7 @@ alloc_fail:
 		btrfs_block_rsv_release(fs_info, &fs_info->trans_block_rsv,
 					num_bytes, NULL);
 reserve_fail:
-	btrfs_qgroup_free_meta_pertrans(root, qgroup_reserved);
+	btrfs_qgroup_free_meta_prealloc(root, qgroup_reserved);
 	return ERR_PTR(ret);
 }
 
@@ -785,7 +799,10 @@ struct btrfs_trans_handle *btrfs_join_transaction_spacecache(struct btrfs_root *
 
 /*
  * Similar to regular join but it never starts a transaction when none is
- * running or after waiting for the current one to finish.
+ * running or when there's a running one at a state >= TRANS_STATE_UNBLOCKED.
+ * This is similar to btrfs_attach_transaction() but it allows the join to
+ * happen if the transaction commit already started but it's not yet in the
+ * "doing" phase (the state is < TRANS_STATE_COMMIT_DOING).
  */
 struct btrfs_trans_handle *btrfs_join_transaction_nostart(struct btrfs_root *root)
 {
@@ -1060,8 +1077,8 @@ int btrfs_write_marked_extents(struct btrfs_fs_info *fs_info,
 	u64 start = 0;
 	u64 end;
 
-	while (!find_first_extent_bit(dirty_pages, start, &start, &end,
-				      mark, &cached_state)) {
+	while (find_first_extent_bit(dirty_pages, start, &start, &end,
+				     mark, &cached_state)) {
 		bool wait_writeback = false;
 
 		err = convert_extent_bit(dirty_pages, start, end,
@@ -1114,8 +1131,8 @@ static int __btrfs_wait_marked_extents(struct btrfs_fs_info *fs_info,
 	u64 start = 0;
 	u64 end;
 
-	while (!find_first_extent_bit(dirty_pages, start, &start, &end,
-				      EXTENT_NEED_WAIT, &cached_state)) {
+	while (find_first_extent_bit(dirty_pages, start, &start, &end,
+				     EXTENT_NEED_WAIT, &cached_state)) {
 		/*
 		 * Ignore -ENOMEM errors returned by clear_extent_bit().
 		 * When committing the transaction, we'll remove any entries
diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c
index ffcff7188170..d1e46b839519 100644
--- a/fs/btrfs/tree-log.c
+++ b/fs/btrfs/tree-log.c
@@ -4841,13 +4841,11 @@ static int btrfs_log_changed_extents(struct btrfs_trans_handle *trans,
 	struct btrfs_ordered_extent *ordered;
 	struct btrfs_ordered_extent *tmp;
 	struct extent_map *em, *n;
-	struct list_head extents;
+	LIST_HEAD(extents);
 	struct extent_map_tree *tree = &inode->extent_tree;
 	int ret = 0;
 	int num = 0;
 
-	INIT_LIST_HEAD(&extents);
-
 	write_lock(&tree->lock);
 
 	list_for_each_entry_safe(em, n, &tree->modified_extents, list) {
@@ -6794,8 +6792,8 @@ static int log_new_ancestors(struct btrfs_trans_handle *trans,
 
 	while (true) {
 		struct btrfs_fs_info *fs_info = root->fs_info;
-		struct extent_buffer *leaf = path->nodes[0];
-		int slot = path->slots[0];
+		struct extent_buffer *leaf;
+		int slot;
 		struct btrfs_key search_key;
 		struct inode *inode;
 		u64 ino;
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index f6718999d183..9621455edebc 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -681,6 +681,14 @@ error_free_page:
 	return -EINVAL;
 }
 
+u8 *btrfs_sb_fsid_ptr(struct btrfs_super_block *sb)
+{
+	bool has_metadata_uuid = (btrfs_super_incompat_flags(sb) &
+				  BTRFS_FEATURE_INCOMPAT_METADATA_UUID);
+
+	return has_metadata_uuid ? sb->metadata_uuid : sb->fsid;
+}
+
 /*
  * Handle scanned device having its CHANGING_FSID_V2 flag set and the fs_devices
  * being created with a disk that has already completed its fsid change. Such
@@ -833,15 +841,8 @@ static noinline struct btrfs_device *device_list_add(const char *path,
 		    found_transid > fs_devices->latest_generation) {
 			memcpy(fs_devices->fsid, disk_super->fsid,
 					BTRFS_FSID_SIZE);
-
-			if (has_metadata_uuid)
-				memcpy(fs_devices->metadata_uuid,
-				       disk_super->metadata_uuid,
-				       BTRFS_FSID_SIZE);
-			else
-				memcpy(fs_devices->metadata_uuid,
-				       disk_super->fsid, BTRFS_FSID_SIZE);
-
+			memcpy(fs_devices->metadata_uuid,
+			       btrfs_sb_fsid_ptr(disk_super), BTRFS_FSID_SIZE);
 			fs_devices->fsid_change = false;
 		}
 	}
@@ -851,8 +852,9 @@ static noinline struct btrfs_device *device_list_add(const char *path,
 
 		if (fs_devices->opened) {
 			btrfs_err(NULL,
-		"device %s belongs to fsid %pU, and the fs is already mounted",
-				  path, fs_devices->fsid);
+"device %s belongs to fsid %pU, and the fs is already mounted, scanned by %s (%d)",
+				  path, fs_devices->fsid, current->comm,
+				  task_pid_nr(current));
 			mutex_unlock(&fs_devices->device_list_mutex);
 			return ERR_PTR(-EBUSY);
 		}
@@ -1424,9 +1426,9 @@ static bool contains_pending_extent(struct btrfs_device *device, u64 *start,
 
 	lockdep_assert_held(&device->fs_info->chunk_mutex);
 
-	if (!find_first_extent_bit(&device->alloc_state, *start,
-				   &physical_start, &physical_end,
-				   CHUNK_ALLOCATED, NULL)) {
+	if (find_first_extent_bit(&device->alloc_state, *start,
+				  &physical_start, &physical_end,
+				  CHUNK_ALLOCATED, NULL)) {
 
 		if (in_range(physical_start, *start, len) ||
 		    in_range(*start, physical_start,
@@ -1438,18 +1440,18 @@ static bool contains_pending_extent(struct btrfs_device *device, u64 *start,
 	return false;
 }
 
-static u64 dev_extent_search_start(struct btrfs_device *device, u64 start)
+static u64 dev_extent_search_start(struct btrfs_device *device)
 {
 	switch (device->fs_devices->chunk_alloc_policy) {
 	case BTRFS_CHUNK_ALLOC_REGULAR:
-		return max_t(u64, start, BTRFS_DEVICE_RANGE_RESERVED);
+		return BTRFS_DEVICE_RANGE_RESERVED;
 	case BTRFS_CHUNK_ALLOC_ZONED:
 		/*
 		 * We don't care about the starting region like regular
 		 * allocator, because we anyway use/reserve the first two zones
 		 * for superblock logging.
 		 */
-		return ALIGN(start, device->zone_info->zone_size);
+		return 0;
 	default:
 		BUG();
 	}
@@ -1581,15 +1583,15 @@ static bool dev_extent_hole_check(struct btrfs_device *device, u64 *hole_start,
  * correct usable device space, as device extent freed in current transaction
  * is not reported as available.
  */
-static int find_free_dev_extent_start(struct btrfs_device *device,
-				u64 num_bytes, u64 search_start, u64 *start,
-				u64 *len)
+static int find_free_dev_extent(struct btrfs_device *device, u64 num_bytes,
+				u64 *start, u64 *len)
 {
 	struct btrfs_fs_info *fs_info = device->fs_info;
 	struct btrfs_root *root = fs_info->dev_root;
 	struct btrfs_key key;
 	struct btrfs_dev_extent *dev_extent;
 	struct btrfs_path *path;
+	u64 search_start;
 	u64 hole_size;
 	u64 max_hole_start;
 	u64 max_hole_size;
@@ -1599,7 +1601,7 @@ static int find_free_dev_extent_start(struct btrfs_device *device,
 	int slot;
 	struct extent_buffer *l;
 
-	search_start = dev_extent_search_start(device, search_start);
+	search_start = dev_extent_search_start(device);
 
 	WARN_ON(device->zone_info &&
 		!IS_ALIGNED(num_bytes, device->zone_info->zone_size));
@@ -1725,13 +1727,6 @@ out:
 	return ret;
 }
 
-int find_free_dev_extent(struct btrfs_device *device, u64 num_bytes,
-			 u64 *start, u64 *len)
-{
-	/* FIXME use last free of some kind */
-	return find_free_dev_extent_start(device, num_bytes, 0, start, len);
-}
-
 static int btrfs_free_dev_extent(struct btrfs_trans_handle *trans,
 			  struct btrfs_device *device,
 			  u64 start, u64 *dev_extent_len)
@@ -6217,6 +6212,45 @@ static void set_io_stripe(struct btrfs_io_stripe *dst, const struct map_lookup *
 			stripe_offset + btrfs_stripe_nr_to_offset(stripe_nr);
 }
 
+/*
+ * Map one logical range to one or more physical ranges.
+ *
+ * @length:		(Mandatory) mapped length of this run.
+ *			One logical range can be split into different segments
+ *			due to factors like zones and RAID0/5/6/10 stripe
+ *			boundaries.
+ *
+ * @bioc_ret:		(Mandatory) returned btrfs_io_context structure.
+ *			which has one or more physical ranges (btrfs_io_stripe)
+ *			recorded inside.
+ *			Caller should call btrfs_put_bioc() to free it after use.
+ *
+ * @smap:		(Optional) single physical range optimization.
+ *			If the map request can be fulfilled by one single
+ *			physical range, and this is parameter is not NULL,
+ *			then @bioc_ret would be NULL, and @smap would be
+ *			updated.
+ *
+ * @mirror_num_ret:	(Mandatory) returned mirror number if the original
+ *			value is 0.
+ *
+ *			Mirror number 0 means to choose any live mirrors.
+ *
+ *			For non-RAID56 profiles, non-zero mirror_num means
+ *			the Nth mirror. (e.g. mirror_num 1 means the first
+ *			copy).
+ *
+ *			For RAID56 profile, mirror 1 means rebuild from P and
+ *			the remaining data stripes.
+ *
+ *			For RAID6 profile, mirror > 2 means mark another
+ *			data/P stripe error and rebuild from the remaining
+ *			stripes..
+ *
+ * @need_raid_map:	(Used only for integrity checker) whether the map wants
+ *                      a full stripe map (including all data and P/Q stripes)
+ *                      for RAID56. Should always be 1 except integrity checker.
+ */
 int btrfs_map_block(struct btrfs_fs_info *fs_info, enum btrfs_map_op op,
 		    u64 logical, u64 *length,
 		    struct btrfs_io_context **bioc_ret,
@@ -6391,9 +6425,7 @@ int btrfs_map_block(struct btrfs_fs_info *fs_info, enum btrfs_map_op op,
 	 * I/O context structure.
 	 */
 	if (smap && num_alloc_stripes == 1 &&
-	    !((map->type & BTRFS_BLOCK_GROUP_RAID56_MASK) && mirror_num > 1) &&
-	    (op == BTRFS_MAP_READ || !dev_replace_is_ongoing ||
-	     !dev_replace->tgtdev)) {
+	    !((map->type & BTRFS_BLOCK_GROUP_RAID56_MASK) && mirror_num > 1)) {
 		set_io_stripe(smap, map, stripe_index, stripe_offset, stripe_nr);
 		if (mirror_num_ret)
 			*mirror_num_ret = mirror_num;
diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h
index b8c51f16ba86..2128a032c3b7 100644
--- a/fs/btrfs/volumes.h
+++ b/fs/btrfs/volumes.h
@@ -650,8 +650,6 @@ int btrfs_cancel_balance(struct btrfs_fs_info *fs_info);
 int btrfs_create_uuid_tree(struct btrfs_fs_info *fs_info);
 int btrfs_uuid_scan_kthread(void *data);
 bool btrfs_chunk_writeable(struct btrfs_fs_info *fs_info, u64 chunk_offset);
-int find_free_dev_extent(struct btrfs_device *device, u64 num_bytes,
-			 u64 *start, u64 *max_avail);
 void btrfs_dev_stat_inc_and_print(struct btrfs_device *dev, int index);
 int btrfs_get_dev_stats(struct btrfs_fs_info *fs_info,
 			struct btrfs_ioctl_get_dev_stats *stats);
@@ -749,5 +747,6 @@ int btrfs_verify_dev_extents(struct btrfs_fs_info *fs_info);
 bool btrfs_repair_one_zone(struct btrfs_fs_info *fs_info, u64 logical);
 
 bool btrfs_pinned_by_swapfile(struct btrfs_fs_info *fs_info, void *ptr);
+u8 *btrfs_sb_fsid_ptr(struct btrfs_super_block *sb);
 
 #endif
diff --git a/fs/btrfs/zoned.c b/fs/btrfs/zoned.c
index 72b90bc19a19..09bc325d075d 100644
--- a/fs/btrfs/zoned.c
+++ b/fs/btrfs/zoned.c
@@ -65,6 +65,9 @@
 
 #define SUPER_INFO_SECTORS	((u64)BTRFS_SUPER_INFO_SIZE >> SECTOR_SHIFT)
 
+static void wait_eb_writebacks(struct btrfs_block_group *block_group);
+static int do_zone_finish(struct btrfs_block_group *block_group, bool fully_written);
+
 static inline bool sb_zone_is_full(const struct blk_zone *zone)
 {
 	return (zone->cond == BLK_ZONE_COND_FULL) ||
@@ -465,8 +468,8 @@ int btrfs_get_dev_zone_info(struct btrfs_device *device, bool populate_cache)
 	 * use the cache.
 	 */
 	if (populate_cache && bdev_is_zoned(device->bdev)) {
-		zone_info->zone_cache = vzalloc(sizeof(struct blk_zone) *
-						zone_info->nr_zones);
+		zone_info->zone_cache = vcalloc(zone_info->nr_zones,
+						sizeof(struct blk_zone));
 		if (!zone_info->zone_cache) {
 			btrfs_err_in_rcu(device->fs_info,
 				"zoned: failed to allocate zone cache for %s",
@@ -1583,19 +1586,9 @@ void btrfs_calc_zone_unusable(struct btrfs_block_group *cache)
 		return;
 
 	WARN_ON(cache->bytes_super != 0);
-
-	/* Check for block groups never get activated */
-	if (test_bit(BTRFS_FS_ACTIVE_ZONE_TRACKING, &cache->fs_info->flags) &&
-	    cache->flags & (BTRFS_BLOCK_GROUP_METADATA | BTRFS_BLOCK_GROUP_SYSTEM) &&
-	    !test_bit(BLOCK_GROUP_FLAG_ZONE_IS_ACTIVE, &cache->runtime_flags) &&
-	    cache->alloc_offset == 0) {
-		unusable = cache->length;
-		free = 0;
-	} else {
-		unusable = (cache->alloc_offset - cache->used) +
-			   (cache->length - cache->zone_capacity);
-		free = cache->zone_capacity - cache->alloc_offset;
-	}
+	unusable = (cache->alloc_offset - cache->used) +
+		   (cache->length - cache->zone_capacity);
+	free = cache->zone_capacity - cache->alloc_offset;
 
 	/* We only need ->free_space in ALLOC_SEQ block groups */
 	cache->cached = BTRFS_CACHE_FINISHED;
@@ -1707,10 +1700,21 @@ void btrfs_finish_ordered_zoned(struct btrfs_ordered_extent *ordered)
 {
 	struct btrfs_inode *inode = BTRFS_I(ordered->inode);
 	struct btrfs_fs_info *fs_info = inode->root->fs_info;
-	struct btrfs_ordered_sum *sum =
-		list_first_entry(&ordered->list, typeof(*sum), list);
-	u64 logical = sum->logical;
-	u64 len = sum->len;
+	struct btrfs_ordered_sum *sum;
+	u64 logical, len;
+
+	/*
+	 * Write to pre-allocated region is for the data relocation, and so
+	 * it should use WRITE operation. No split/rewrite are necessary.
+	 */
+	if (test_bit(BTRFS_ORDERED_PREALLOC, &ordered->flags))
+		return;
+
+	ASSERT(!list_empty(&ordered->list));
+	/* The ordered->list can be empty in the above pre-alloc case. */
+	sum = list_first_entry(&ordered->list, struct btrfs_ordered_sum, list);
+	logical = sum->logical;
+	len = sum->len;
 
 	while (len < ordered->disk_num_bytes) {
 		sum = list_next_entry(sum, list);
@@ -1747,41 +1751,121 @@ out:
 	}
 }
 
-bool btrfs_check_meta_write_pointer(struct btrfs_fs_info *fs_info,
-				    struct extent_buffer *eb,
-				    struct btrfs_block_group **cache_ret)
+static bool check_bg_is_active(struct btrfs_eb_write_context *ctx,
+			       struct btrfs_block_group **active_bg)
 {
-	struct btrfs_block_group *cache;
-	bool ret = true;
+	const struct writeback_control *wbc = ctx->wbc;
+	struct btrfs_block_group *block_group = ctx->zoned_bg;
+	struct btrfs_fs_info *fs_info = block_group->fs_info;
 
-	if (!btrfs_is_zoned(fs_info))
+	if (test_bit(BLOCK_GROUP_FLAG_ZONE_IS_ACTIVE, &block_group->runtime_flags))
 		return true;
 
-	cache = btrfs_lookup_block_group(fs_info, eb->start);
-	if (!cache)
-		return true;
+	if (fs_info->treelog_bg == block_group->start) {
+		if (!btrfs_zone_activate(block_group)) {
+			int ret_fin = btrfs_zone_finish_one_bg(fs_info);
 
-	if (cache->meta_write_pointer != eb->start) {
-		btrfs_put_block_group(cache);
-		cache = NULL;
-		ret = false;
-	} else {
-		cache->meta_write_pointer = eb->start + eb->len;
-	}
+			if (ret_fin != 1 || !btrfs_zone_activate(block_group))
+				return false;
+		}
+	} else if (*active_bg != block_group) {
+		struct btrfs_block_group *tgt = *active_bg;
 
-	*cache_ret = cache;
+		/* zoned_meta_io_lock protects fs_info->active_{meta,system}_bg. */
+		lockdep_assert_held(&fs_info->zoned_meta_io_lock);
 
-	return ret;
+		if (tgt) {
+			/*
+			 * If there is an unsent IO left in the allocated area,
+			 * we cannot wait for them as it may cause a deadlock.
+			 */
+			if (tgt->meta_write_pointer < tgt->start + tgt->alloc_offset) {
+				if (wbc->sync_mode == WB_SYNC_NONE ||
+				    (wbc->sync_mode == WB_SYNC_ALL && !wbc->for_sync))
+					return false;
+			}
+
+			/* Pivot active metadata/system block group. */
+			btrfs_zoned_meta_io_unlock(fs_info);
+			wait_eb_writebacks(tgt);
+			do_zone_finish(tgt, true);
+			btrfs_zoned_meta_io_lock(fs_info);
+			if (*active_bg == tgt) {
+				btrfs_put_block_group(tgt);
+				*active_bg = NULL;
+			}
+		}
+		if (!btrfs_zone_activate(block_group))
+			return false;
+		if (*active_bg != block_group) {
+			ASSERT(*active_bg == NULL);
+			*active_bg = block_group;
+			btrfs_get_block_group(block_group);
+		}
+	}
+
+	return true;
 }
 
-void btrfs_revert_meta_write_pointer(struct btrfs_block_group *cache,
-				     struct extent_buffer *eb)
+/*
+ * Check if @ctx->eb is aligned to the write pointer.
+ *
+ * Return:
+ *   0:        @ctx->eb is at the write pointer. You can write it.
+ *   -EAGAIN:  There is a hole. The caller should handle the case.
+ *   -EBUSY:   There is a hole, but the caller can just bail out.
+ */
+int btrfs_check_meta_write_pointer(struct btrfs_fs_info *fs_info,
+				   struct btrfs_eb_write_context *ctx)
 {
-	if (!btrfs_is_zoned(eb->fs_info) || !cache)
-		return;
+	const struct writeback_control *wbc = ctx->wbc;
+	const struct extent_buffer *eb = ctx->eb;
+	struct btrfs_block_group *block_group = ctx->zoned_bg;
+
+	if (!btrfs_is_zoned(fs_info))
+		return 0;
+
+	if (block_group) {
+		if (block_group->start > eb->start ||
+		    block_group->start + block_group->length <= eb->start) {
+			btrfs_put_block_group(block_group);
+			block_group = NULL;
+			ctx->zoned_bg = NULL;
+		}
+	}
+
+	if (!block_group) {
+		block_group = btrfs_lookup_block_group(fs_info, eb->start);
+		if (!block_group)
+			return 0;
+		ctx->zoned_bg = block_group;
+	}
+
+	if (block_group->meta_write_pointer == eb->start) {
+		struct btrfs_block_group **tgt;
+
+		if (!test_bit(BTRFS_FS_ACTIVE_ZONE_TRACKING, &fs_info->flags))
+			return 0;
 
-	ASSERT(cache->meta_write_pointer == eb->start + eb->len);
-	cache->meta_write_pointer = eb->start;
+		if (block_group->flags & BTRFS_BLOCK_GROUP_SYSTEM)
+			tgt = &fs_info->active_system_bg;
+		else
+			tgt = &fs_info->active_meta_bg;
+		if (check_bg_is_active(ctx, tgt))
+			return 0;
+	}
+
+	/*
+	 * Since we may release fs_info->zoned_meta_io_lock, someone can already
+	 * start writing this eb. In that case, we can just bail out.
+	 */
+	if (block_group->meta_write_pointer > eb->start)
+		return -EBUSY;
+
+	/* If for_sync, this hole will be filled with trasnsaction commit. */
+	if (wbc->sync_mode == WB_SYNC_ALL && !wbc->for_sync)
+		return -EAGAIN;
+	return -EBUSY;
 }
 
 int btrfs_zoned_issue_zeroout(struct btrfs_device *device, u64 physical, u64 length)
@@ -1879,10 +1963,10 @@ int btrfs_sync_zone_write_pointer(struct btrfs_device *tgt_dev, u64 logical,
 bool btrfs_zone_activate(struct btrfs_block_group *block_group)
 {
 	struct btrfs_fs_info *fs_info = block_group->fs_info;
-	struct btrfs_space_info *space_info = block_group->space_info;
 	struct map_lookup *map;
 	struct btrfs_device *device;
 	u64 physical;
+	const bool is_data = (block_group->flags & BTRFS_BLOCK_GROUP_DATA);
 	bool ret;
 	int i;
 
@@ -1891,7 +1975,6 @@ bool btrfs_zone_activate(struct btrfs_block_group *block_group)
 
 	map = block_group->physical_map;
 
-	spin_lock(&space_info->lock);
 	spin_lock(&block_group->lock);
 	if (test_bit(BLOCK_GROUP_FLAG_ZONE_IS_ACTIVE, &block_group->runtime_flags)) {
 		ret = true;
@@ -1904,30 +1987,44 @@ bool btrfs_zone_activate(struct btrfs_block_group *block_group)
 		goto out_unlock;
 	}
 
+	spin_lock(&fs_info->zone_active_bgs_lock);
 	for (i = 0; i < map->num_stripes; i++) {
+		struct btrfs_zoned_device_info *zinfo;
+		int reserved = 0;
+
 		device = map->stripes[i].dev;
 		physical = map->stripes[i].physical;
+		zinfo = device->zone_info;
 
-		if (device->zone_info->max_active_zones == 0)
+		if (zinfo->max_active_zones == 0)
 			continue;
 
+		if (is_data)
+			reserved = zinfo->reserved_active_zones;
+		/*
+		 * For the data block group, leave active zones for one
+		 * metadata block group and one system block group.
+		 */
+		if (atomic_read(&zinfo->active_zones_left) <= reserved) {
+			ret = false;
+			spin_unlock(&fs_info->zone_active_bgs_lock);
+			goto out_unlock;
+		}
+
 		if (!btrfs_dev_set_active_zone(device, physical)) {
 			/* Cannot activate the zone */
 			ret = false;
+			spin_unlock(&fs_info->zone_active_bgs_lock);
 			goto out_unlock;
 		}
+		if (!is_data)
+			zinfo->reserved_active_zones--;
 	}
+	spin_unlock(&fs_info->zone_active_bgs_lock);
 
 	/* Successfully activated all the zones */
 	set_bit(BLOCK_GROUP_FLAG_ZONE_IS_ACTIVE, &block_group->runtime_flags);
-	WARN_ON(block_group->alloc_offset != 0);
-	if (block_group->zone_unusable == block_group->length) {
-		block_group->zone_unusable = block_group->length - block_group->zone_capacity;
-		space_info->bytes_zone_unusable -= block_group->zone_capacity;
-	}
 	spin_unlock(&block_group->lock);
-	btrfs_try_granting_tickets(fs_info, space_info);
-	spin_unlock(&space_info->lock);
 
 	/* For the active block group list */
 	btrfs_get_block_group(block_group);
@@ -1940,7 +2037,6 @@ bool btrfs_zone_activate(struct btrfs_block_group *block_group)
 
 out_unlock:
 	spin_unlock(&block_group->lock);
-	spin_unlock(&space_info->lock);
 	return ret;
 }
 
@@ -2006,6 +2102,10 @@ static int do_zone_finish(struct btrfs_block_group *block_group, bool fully_writ
 	 * and block_group->meta_write_pointer for metadata.
 	 */
 	if (!fully_written) {
+		if (test_bit(BLOCK_GROUP_FLAG_ZONED_DATA_RELOC, &block_group->runtime_flags)) {
+			spin_unlock(&block_group->lock);
+			return -EAGAIN;
+		}
 		spin_unlock(&block_group->lock);
 
 		ret = btrfs_inc_block_group_ro(block_group, false);
@@ -2034,7 +2134,9 @@ static int do_zone_finish(struct btrfs_block_group *block_group, bool fully_writ
 			return 0;
 		}
 
-		if (block_group->reserved) {
+		if (block_group->reserved ||
+		    test_bit(BLOCK_GROUP_FLAG_ZONED_DATA_RELOC,
+			     &block_group->runtime_flags)) {
 			spin_unlock(&block_group->lock);
 			btrfs_dec_block_group_ro(block_group);
 			return -EAGAIN;
@@ -2043,6 +2145,9 @@ static int do_zone_finish(struct btrfs_block_group *block_group, bool fully_writ
 
 	clear_bit(BLOCK_GROUP_FLAG_ZONE_IS_ACTIVE, &block_group->runtime_flags);
 	block_group->alloc_offset = block_group->zone_capacity;
+	if (block_group->flags & (BTRFS_BLOCK_GROUP_METADATA | BTRFS_BLOCK_GROUP_SYSTEM))
+		block_group->meta_write_pointer = block_group->start +
+						  block_group->zone_capacity;
 	block_group->free_space_ctl->free_space = 0;
 	btrfs_clear_treelog_bg(block_group);
 	btrfs_clear_data_reloc_bg(block_group);
@@ -2052,18 +2157,21 @@ static int do_zone_finish(struct btrfs_block_group *block_group, bool fully_writ
 	for (i = 0; i < map->num_stripes; i++) {
 		struct btrfs_device *device = map->stripes[i].dev;
 		const u64 physical = map->stripes[i].physical;
+		struct btrfs_zoned_device_info *zinfo = device->zone_info;
 
-		if (device->zone_info->max_active_zones == 0)
+		if (zinfo->max_active_zones == 0)
 			continue;
 
 		ret = blkdev_zone_mgmt(device->bdev, REQ_OP_ZONE_FINISH,
 				       physical >> SECTOR_SHIFT,
-				       device->zone_info->zone_size >> SECTOR_SHIFT,
+				       zinfo->zone_size >> SECTOR_SHIFT,
 				       GFP_NOFS);
 
 		if (ret)
 			return ret;
 
+		if (!(block_group->flags & BTRFS_BLOCK_GROUP_DATA))
+			zinfo->reserved_active_zones++;
 		btrfs_dev_clear_active_zone(device, physical);
 	}
 
@@ -2102,8 +2210,10 @@ bool btrfs_can_activate_zone(struct btrfs_fs_devices *fs_devices, u64 flags)
 
 	/* Check if there is a device with active zones left */
 	mutex_lock(&fs_info->chunk_mutex);
+	spin_lock(&fs_info->zone_active_bgs_lock);
 	list_for_each_entry(device, &fs_devices->alloc_list, dev_alloc_list) {
 		struct btrfs_zoned_device_info *zinfo = device->zone_info;
+		int reserved = 0;
 
 		if (!device->bdev)
 			continue;
@@ -2113,17 +2223,21 @@ bool btrfs_can_activate_zone(struct btrfs_fs_devices *fs_devices, u64 flags)
 			break;
 		}
 
+		if (flags & BTRFS_BLOCK_GROUP_DATA)
+			reserved = zinfo->reserved_active_zones;
+
 		switch (flags & BTRFS_BLOCK_GROUP_PROFILE_MASK) {
 		case 0: /* single */
-			ret = (atomic_read(&zinfo->active_zones_left) >= 1);
+			ret = (atomic_read(&zinfo->active_zones_left) >= (1 + reserved));
 			break;
 		case BTRFS_BLOCK_GROUP_DUP:
-			ret = (atomic_read(&zinfo->active_zones_left) >= 2);
+			ret = (atomic_read(&zinfo->active_zones_left) >= (2 + reserved));
 			break;
 		}
 		if (ret)
 			break;
 	}
+	spin_unlock(&fs_info->zone_active_bgs_lock);
 	mutex_unlock(&fs_info->chunk_mutex);
 
 	if (!ret)
@@ -2265,7 +2379,10 @@ void btrfs_zoned_release_data_reloc_bg(struct btrfs_fs_info *fs_info, u64 logica
 
 	/* All relocation extents are written. */
 	if (block_group->start + block_group->alloc_offset == logical + length) {
-		/* Now, release this block group for further allocations. */
+		/*
+		 * Now, release this block group for further allocations and
+		 * zone finish.
+		 */
 		clear_bit(BLOCK_GROUP_FLAG_ZONED_DATA_RELOC,
 			  &block_group->runtime_flags);
 	}
@@ -2289,7 +2406,8 @@ int btrfs_zone_finish_one_bg(struct btrfs_fs_info *fs_info)
 
 		spin_lock(&block_group->lock);
 		if (block_group->reserved || block_group->alloc_offset == 0 ||
-		    (block_group->flags & BTRFS_BLOCK_GROUP_SYSTEM)) {
+		    (block_group->flags & BTRFS_BLOCK_GROUP_SYSTEM) ||
+		    test_bit(BLOCK_GROUP_FLAG_ZONED_DATA_RELOC, &block_group->runtime_flags)) {
 			spin_unlock(&block_group->lock);
 			continue;
 		}
@@ -2365,3 +2483,55 @@ int btrfs_zoned_activate_one_bg(struct btrfs_fs_info *fs_info,
 
 	return 0;
 }
+
+/*
+ * Reserve zones for one metadata block group, one tree-log block group, and one
+ * system block group.
+ */
+void btrfs_check_active_zone_reservation(struct btrfs_fs_info *fs_info)
+{
+	struct btrfs_fs_devices *fs_devices = fs_info->fs_devices;
+	struct btrfs_block_group *block_group;
+	struct btrfs_device *device;
+	/* Reserve zones for normal SINGLE metadata and tree-log block group. */
+	unsigned int metadata_reserve = 2;
+	/* Reserve a zone for SINGLE system block group. */
+	unsigned int system_reserve = 1;
+
+	if (!test_bit(BTRFS_FS_ACTIVE_ZONE_TRACKING, &fs_info->flags))
+		return;
+
+	/*
+	 * This function is called from the mount context. So, there is no
+	 * parallel process touching the bits. No need for read_seqretry().
+	 */
+	if (fs_info->avail_metadata_alloc_bits & BTRFS_BLOCK_GROUP_DUP)
+		metadata_reserve = 4;
+	if (fs_info->avail_system_alloc_bits & BTRFS_BLOCK_GROUP_DUP)
+		system_reserve = 2;
+
+	/* Apply the reservation on all the devices. */
+	mutex_lock(&fs_devices->device_list_mutex);
+	list_for_each_entry(device, &fs_devices->devices, dev_list) {
+		if (!device->bdev)
+			continue;
+
+		device->zone_info->reserved_active_zones =
+			metadata_reserve + system_reserve;
+	}
+	mutex_unlock(&fs_devices->device_list_mutex);
+
+	/* Release reservation for currently active block groups. */
+	spin_lock(&fs_info->zone_active_bgs_lock);
+	list_for_each_entry(block_group, &fs_info->zone_active_bgs, active_bg_list) {
+		struct map_lookup *map = block_group->physical_map;
+
+		if (!(block_group->flags &
+		      (BTRFS_BLOCK_GROUP_METADATA | BTRFS_BLOCK_GROUP_SYSTEM)))
+			continue;
+
+		for (int i = 0; i < map->num_stripes; i++)
+			map->stripes[i].dev->zone_info->reserved_active_zones--;
+	}
+	spin_unlock(&fs_info->zone_active_bgs_lock);
+}
diff --git a/fs/btrfs/zoned.h b/fs/btrfs/zoned.h
index 27322b926038..b9cec523b778 100644
--- a/fs/btrfs/zoned.h
+++ b/fs/btrfs/zoned.h
@@ -22,6 +22,11 @@ struct btrfs_zoned_device_info {
 	u8  zone_size_shift;
 	u32 nr_zones;
 	unsigned int max_active_zones;
+	/*
+	 * Reserved active zones for one metadata and one system block group.
+	 * It can vary per-device depending on the allocation status.
+	 */
+	int reserved_active_zones;
 	atomic_t active_zones_left;
 	unsigned long *seq_zones;
 	unsigned long *empty_zones;
@@ -58,11 +63,8 @@ void btrfs_redirty_list_add(struct btrfs_transaction *trans,
 			    struct extent_buffer *eb);
 bool btrfs_use_zone_append(struct btrfs_bio *bbio);
 void btrfs_record_physical_zoned(struct btrfs_bio *bbio);
-bool btrfs_check_meta_write_pointer(struct btrfs_fs_info *fs_info,
-				    struct extent_buffer *eb,
-				    struct btrfs_block_group **cache_ret);
-void btrfs_revert_meta_write_pointer(struct btrfs_block_group *cache,
-				     struct extent_buffer *eb);
+int btrfs_check_meta_write_pointer(struct btrfs_fs_info *fs_info,
+				   struct btrfs_eb_write_context *ctx);
 int btrfs_zoned_issue_zeroout(struct btrfs_device *device, u64 physical, u64 length);
 int btrfs_sync_zone_write_pointer(struct btrfs_device *tgt_dev, u64 logical,
 				  u64 physical_start, u64 physical_pos);
@@ -81,6 +83,7 @@ void btrfs_zoned_release_data_reloc_bg(struct btrfs_fs_info *fs_info, u64 logica
 int btrfs_zone_finish_one_bg(struct btrfs_fs_info *fs_info);
 int btrfs_zoned_activate_one_bg(struct btrfs_fs_info *fs_info,
 				struct btrfs_space_info *space_info, bool do_finish);
+void btrfs_check_active_zone_reservation(struct btrfs_fs_info *fs_info);
 #else /* CONFIG_BLK_DEV_ZONED */
 static inline int btrfs_get_dev_zone(struct btrfs_device *device, u64 pos,
 				     struct blk_zone *zone)
@@ -189,17 +192,10 @@ static inline void btrfs_record_physical_zoned(struct btrfs_bio *bbio)
 {
 }
 
-static inline bool btrfs_check_meta_write_pointer(struct btrfs_fs_info *fs_info,
-			       struct extent_buffer *eb,
-			       struct btrfs_block_group **cache_ret)
-{
-	return true;
-}
-
-static inline void btrfs_revert_meta_write_pointer(
-						struct btrfs_block_group *cache,
-						struct extent_buffer *eb)
+static inline int btrfs_check_meta_write_pointer(struct btrfs_fs_info *fs_info,
+						 struct btrfs_eb_write_context *ctx)
 {
+	return 0;
 }
 
 static inline int btrfs_zoned_issue_zeroout(struct btrfs_device *device,
@@ -262,6 +258,8 @@ static inline int btrfs_zoned_activate_one_bg(struct btrfs_fs_info *fs_info,
 	return 0;
 }
 
+static inline void btrfs_check_active_zone_reservation(struct btrfs_fs_info *fs_info) { }
+
 #endif
 
 static inline bool btrfs_dev_is_sequential(struct btrfs_device *device, u64 pos)