1 files changed, 211 insertions, 140 deletions

diff --git a/fs/btrfs/block-group.c b/fs/btrfs/block-group.c
index c0a8f7d92acc..5b0cb04b2b93 100644
--- a/fs/btrfs/block-group.c
+++ b/fs/btrfs/block-group.c
@@ -191,21 +191,21 @@ static int btrfs_bg_start_cmp(const struct rb_node *new,
 /*
  * This adds the block group to the fs_info rb tree for the block group cache
  */
-static int btrfs_add_block_group_cache(struct btrfs_fs_info *info,
-				       struct btrfs_block_group *block_group)
+static int btrfs_add_block_group_cache(struct btrfs_block_group *block_group)
 {
+	struct btrfs_fs_info *fs_info = block_group->fs_info;
 	struct rb_node *exist;
 	int ret = 0;
 
 	ASSERT(block_group->length != 0);
 
-	write_lock(&info->block_group_cache_lock);
+	write_lock(&fs_info->block_group_cache_lock);
 
 	exist = rb_find_add_cached(&block_group->cache_node,
-			&info->block_group_cache_tree, btrfs_bg_start_cmp);
+			&fs_info->block_group_cache_tree, btrfs_bg_start_cmp);
 	if (exist)
 		ret = -EEXIST;
-	write_unlock(&info->block_group_cache_lock);
+	write_unlock(&fs_info->block_group_cache_lock);
 
 	return ret;
 }
@@ -525,10 +525,9 @@ int btrfs_add_new_free_space(struct btrfs_block_group *block_group, u64 start,
 		*total_added_ret = 0;
 
 	while (start < end) {
-		if (!find_first_extent_bit(&info->excluded_extents, start,
-					   &extent_start, &extent_end,
-					   EXTENT_DIRTY | EXTENT_UPTODATE,
-					   NULL))
+		if (!btrfs_find_first_extent_bit(&info->excluded_extents, start,
+						 &extent_start, &extent_end,
+						 EXTENT_DIRTY, NULL))
 			break;
 
 		if (extent_start <= start) {
@@ -584,7 +583,7 @@ static int sample_block_group_extent_item(struct btrfs_caching_control *caching_
 	struct btrfs_root *extent_root;
 	u64 search_offset;
 	u64 search_end = block_group->start + block_group->length;
-	struct btrfs_path *path;
+	BTRFS_PATH_AUTO_FREE(path);
 	struct btrfs_key search_key;
 	int ret = 0;
 
@@ -626,7 +625,6 @@ static int sample_block_group_extent_item(struct btrfs_caching_control *caching_
 
 	lockdep_assert_held(&caching_ctl->mutex);
 	lockdep_assert_held_read(&fs_info->commit_root_sem);
-	btrfs_free_path(path);
 	return ret;
 }
 
@@ -702,7 +700,7 @@ static int load_extent_tree_free(struct btrfs_caching_control *caching_ctl)
 	struct btrfs_block_group *block_group = caching_ctl->block_group;
 	struct btrfs_fs_info *fs_info = block_group->fs_info;
 	struct btrfs_root *extent_root;
-	struct btrfs_path *path;
+	BTRFS_PATH_AUTO_FREE(path);
 	struct extent_buffer *leaf;
 	struct btrfs_key key;
 	u64 total_found = 0;
@@ -738,8 +736,8 @@ static int load_extent_tree_free(struct btrfs_caching_control *caching_ctl)
 	path->reada = READA_FORWARD;
 
 	key.objectid = last;
-	key.offset = 0;
 	key.type = BTRFS_EXTENT_ITEM_KEY;
+	key.offset = 0;
 
 next:
 	ret = btrfs_search_slot(NULL, extent_root, &key, path, 0, 0);
@@ -785,8 +783,8 @@ next:
 
 		if (key.objectid < last) {
 			key.objectid = last;
-			key.offset = 0;
 			key.type = BTRFS_EXTENT_ITEM_KEY;
+			key.offset = 0;
 			btrfs_release_path(path);
 			goto next;
 		}
@@ -829,14 +827,13 @@ next:
 				       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);
+	btrfs_clear_extent_bits(&bg->fs_info->excluded_extents, bg->start,
+				bg->start + bg->length - 1, EXTENT_DIRTY);
 }
 
 static noinline void caching_thread(struct btrfs_work *work)
@@ -1421,9 +1418,8 @@ static bool clean_pinned_extents(struct btrfs_trans_handle *trans,
 	int ret;
 
 	spin_lock(&fs_info->trans_lock);
-	if (trans->transaction->list.prev != &fs_info->trans_list) {
-		prev_trans = list_last_entry(&trans->transaction->list,
-					     struct btrfs_transaction, list);
+	if (!list_is_first(&trans->transaction->list, &fs_info->trans_list)) {
+		prev_trans = list_prev_entry(trans->transaction, list);
 		refcount_inc(&prev_trans->use_count);
 	}
 	spin_unlock(&fs_info->trans_lock);
@@ -1440,14 +1436,14 @@ static bool clean_pinned_extents(struct btrfs_trans_handle *trans,
 	 */
 	mutex_lock(&fs_info->unused_bg_unpin_mutex);
 	if (prev_trans) {
-		ret = clear_extent_bits(&prev_trans->pinned_extents, start, end,
-					EXTENT_DIRTY);
+		ret = btrfs_clear_extent_bits(&prev_trans->pinned_extents, start, end,
+					      EXTENT_DIRTY);
 		if (ret)
 			goto out;
 	}
 
-	ret = clear_extent_bits(&trans->transaction->pinned_extents, start, end,
-				EXTENT_DIRTY);
+	ret = btrfs_clear_extent_bits(&trans->transaction->pinned_extents, start, end,
+				      EXTENT_DIRTY);
 out:
 	mutex_unlock(&fs_info->unused_bg_unpin_mutex);
 	if (prev_trans)
@@ -1457,6 +1453,32 @@ out:
 }
 
 /*
+ * Link the block_group to a list via bg_list.
+ *
+ * @bg:       The block_group to link to the list.
+ * @list:     The list to link it to.
+ *
+ * Use this rather than list_add_tail() directly to ensure proper respect
+ * to locking and refcounting.
+ *
+ * Returns: true if the bg was linked with a refcount bump and false otherwise.
+ */
+static bool btrfs_link_bg_list(struct btrfs_block_group *bg, struct list_head *list)
+{
+	struct btrfs_fs_info *fs_info = bg->fs_info;
+	bool added = false;
+
+	spin_lock(&fs_info->unused_bgs_lock);
+	if (list_empty(&bg->bg_list)) {
+		btrfs_get_block_group(bg);
+		list_add_tail(&bg->bg_list, list);
+		added = true;
+	}
+	spin_unlock(&fs_info->unused_bgs_lock);
+	return added;
+}
+
+/*
  * Process the unused_bgs list and remove any that don't have any allocated
  * space inside of them.
  */
@@ -1571,8 +1593,7 @@ void btrfs_delete_unused_bgs(struct btrfs_fs_info *fs_info)
 			 * drop under the "next" label for the
 			 * fs_info->unused_bgs list.
 			 */
-			btrfs_get_block_group(block_group);
-			list_add_tail(&block_group->bg_list, &retry_list);
+			btrfs_link_bg_list(block_group, &retry_list);
 
 			trace_btrfs_skip_unused_block_group(block_group);
 			spin_unlock(&block_group->lock);
@@ -1823,7 +1844,8 @@ void btrfs_reclaim_bgs_work(struct work_struct *work)
 	list_sort(NULL, &fs_info->reclaim_bgs, reclaim_bgs_cmp);
 	while (!list_empty(&fs_info->reclaim_bgs)) {
 		u64 zone_unusable;
-		u64 reclaimed;
+		u64 used;
+		u64 reserved;
 		int ret = 0;
 
 		bg = list_first_entry(&fs_info->reclaim_bgs,
@@ -1887,6 +1909,17 @@ void btrfs_reclaim_bgs_work(struct work_struct *work)
 			up_write(&space_info->groups_sem);
 			goto next;
 		}
+
+		/*
+		 * Cache the zone_unusable value before turning the block group
+		 * to read only. As soon as the block group is read only it's
+		 * zone_unusable value gets moved to the block group's read-only
+		 * bytes and isn't available for calculations anymore. We also
+		 * cache it before unlocking the block group, to prevent races
+		 * (reports from KCSAN and such tools) with tasks updating it.
+		 */
+		zone_unusable = bg->zone_unusable;
+
 		spin_unlock(&bg->lock);
 		spin_unlock(&space_info->lock);
 
@@ -1903,31 +1936,47 @@ void btrfs_reclaim_bgs_work(struct work_struct *work)
 			goto next;
 		}
 
-		/*
-		 * Cache the zone_unusable value before turning the block group
-		 * to read only. As soon as the blog group is read only it's
-		 * zone_unusable value gets moved to the block group's read-only
-		 * bytes and isn't available for calculations anymore.
-		 */
-		zone_unusable = bg->zone_unusable;
 		ret = inc_block_group_ro(bg, 0);
 		up_write(&space_info->groups_sem);
 		if (ret < 0)
 			goto next;
 
+		/*
+		 * The amount of bytes reclaimed corresponds to the sum of the
+		 * "used" and "reserved" counters. We have set the block group
+		 * to RO above, which prevents reservations from happening but
+		 * we may have existing reservations for which allocation has
+		 * not yet been done - btrfs_update_block_group() was not yet
+		 * called, which is where we will transfer a reserved extent's
+		 * size from the "reserved" counter to the "used" counter - this
+		 * happens when running delayed references. When we relocate the
+		 * chunk below, relocation first flushes dellaloc, waits for
+		 * ordered extent completion (which is where we create delayed
+		 * references for data extents) and commits the current
+		 * transaction (which runs delayed references), and only after
+		 * it does the actual work to move extents out of the block
+		 * group. So the reported amount of reclaimed bytes is
+		 * effectively the sum of the 'used' and 'reserved' counters.
+		 */
+		spin_lock(&bg->lock);
+		used = bg->used;
+		reserved = bg->reserved;
+		spin_unlock(&bg->lock);
+
 		btrfs_info(fs_info,
-			"reclaiming chunk %llu with %llu%% used %llu%% unusable",
+	"reclaiming chunk %llu with %llu%% used %llu%% reserved %llu%% unusable",
 				bg->start,
-				div64_u64(bg->used * 100, bg->length),
+				div64_u64(used * 100, bg->length),
+				div64_u64(reserved * 100, bg->length),
 				div64_u64(zone_unusable * 100, bg->length));
 		trace_btrfs_reclaim_block_group(bg);
-		reclaimed = bg->used;
 		ret = btrfs_relocate_chunk(fs_info, bg->start);
 		if (ret) {
 			btrfs_dec_block_group_ro(bg);
 			btrfs_err(fs_info, "error relocating chunk %llu",
 				  bg->start);
-			reclaimed = 0;
+			used = 0;
+			reserved = 0;
 			spin_lock(&space_info->lock);
 			space_info->reclaim_errors++;
 			if (READ_ONCE(space_info->periodic_reclaim))
@@ -1936,24 +1985,13 @@ void btrfs_reclaim_bgs_work(struct work_struct *work)
 		}
 		spin_lock(&space_info->lock);
 		space_info->reclaim_count++;
-		space_info->reclaim_bytes += reclaimed;
+		space_info->reclaim_bytes += used;
+		space_info->reclaim_bytes += reserved;
 		spin_unlock(&space_info->lock);
 
 next:
-		if (ret && !READ_ONCE(space_info->periodic_reclaim)) {
-			/* Refcount held by the reclaim_bgs list after splice. */
-			spin_lock(&fs_info->unused_bgs_lock);
-			/*
-			 * This block group might be added to the unused list
-			 * during the above process. Move it back to the
-			 * reclaim list otherwise.
-			 */
-			if (list_empty(&bg->bg_list)) {
-				btrfs_get_block_group(bg);
-				list_add_tail(&bg->bg_list, &retry_list);
-			}
-			spin_unlock(&fs_info->unused_bgs_lock);
-		}
+		if (ret && !READ_ONCE(space_info->periodic_reclaim))
+			btrfs_link_bg_list(bg, &retry_list);
 		btrfs_put_block_group(bg);
 
 		mutex_unlock(&fs_info->reclaim_bgs_lock);
@@ -1993,13 +2031,8 @@ void btrfs_mark_bg_to_reclaim(struct btrfs_block_group *bg)
 {
 	struct btrfs_fs_info *fs_info = bg->fs_info;
 
-	spin_lock(&fs_info->unused_bgs_lock);
-	if (list_empty(&bg->bg_list)) {
-		btrfs_get_block_group(bg);
+	if (btrfs_link_bg_list(bg, &fs_info->reclaim_bgs))
 		trace_btrfs_add_reclaim_block_group(bg);
-		list_add_tail(&bg->bg_list, &fs_info->reclaim_bgs);
-	}
-	spin_unlock(&fs_info->unused_bgs_lock);
 }
 
 static int read_bg_from_eb(struct btrfs_fs_info *fs_info, const struct btrfs_key *key,
@@ -2182,9 +2215,9 @@ 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 = set_extent_bit(&fs_info->excluded_extents, cache->start,
-				     cache->start + stripe_len - 1,
-				     EXTENT_UPTODATE, NULL);
+		ret = btrfs_set_extent_bit(&fs_info->excluded_extents, cache->start,
+					   cache->start + stripe_len - 1,
+					   EXTENT_DIRTY, NULL);
 		if (ret)
 			return ret;
 	}
@@ -2210,9 +2243,9 @@ static int exclude_super_stripes(struct btrfs_block_group *cache)
 				cache->start + cache->length - logical[nr]);
 
 			cache->bytes_super += len;
-			ret = set_extent_bit(&fs_info->excluded_extents, logical[nr],
-					     logical[nr] + len - 1,
-					     EXTENT_UPTODATE, NULL);
+			ret = btrfs_set_extent_bit(&fs_info->excluded_extents,
+						   logical[nr], logical[nr] + len - 1,
+						   EXTENT_DIRTY, NULL);
 			if (ret) {
 				kfree(logical);
 				return ret;
@@ -2337,6 +2370,7 @@ static int read_one_block_group(struct btrfs_fs_info *info,
 	cache->commit_used = cache->used;
 	cache->flags = btrfs_stack_block_group_flags(bgi);
 	cache->global_root_id = btrfs_stack_block_group_chunk_objectid(bgi);
+	cache->space_info = btrfs_find_space_info(info, cache->flags);
 
 	set_free_space_tree_thresholds(cache);
 
@@ -2410,11 +2444,12 @@ static int read_one_block_group(struct btrfs_fs_info *info,
 			goto error;
 	}
 
-	ret = btrfs_add_block_group_cache(info, cache);
+	ret = btrfs_add_block_group_cache(cache);
 	if (ret) {
 		btrfs_remove_free_space_cache(cache);
 		goto error;
 	}
+
 	trace_btrfs_add_block_group(info, cache, 0);
 	btrfs_add_bg_to_space_info(info, cache);
 
@@ -2459,7 +2494,8 @@ static int fill_dummy_bgs(struct btrfs_fs_info *fs_info)
 		bg->cached = BTRFS_CACHE_FINISHED;
 		bg->used = map->chunk_len;
 		bg->flags = map->type;
-		ret = btrfs_add_block_group_cache(fs_info, bg);
+		bg->space_info = btrfs_find_space_info(fs_info, bg->flags);
+		ret = btrfs_add_block_group_cache(bg);
 		/*
 		 * We may have some valid block group cache added already, in
 		 * that case we skip to the next one.
@@ -2509,8 +2545,8 @@ int btrfs_read_block_groups(struct btrfs_fs_info *info)
 		return fill_dummy_bgs(info);
 
 	key.objectid = 0;
-	key.offset = 0;
 	key.type = BTRFS_BLOCK_GROUP_ITEM_KEY;
+	key.offset = 0;
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
@@ -2641,7 +2677,7 @@ static int insert_dev_extent(struct btrfs_trans_handle *trans,
 {
 	struct btrfs_fs_info *fs_info = device->fs_info;
 	struct btrfs_root *root = fs_info->dev_root;
-	struct btrfs_path *path;
+	BTRFS_PATH_AUTO_FREE(path);
 	struct btrfs_dev_extent *extent;
 	struct extent_buffer *leaf;
 	struct btrfs_key key;
@@ -2658,7 +2694,7 @@ static int insert_dev_extent(struct btrfs_trans_handle *trans,
 	key.offset = start;
 	ret = btrfs_insert_empty_item(trans, root, path, &key, sizeof(*extent));
 	if (ret)
-		goto out;
+		return ret;
 
 	leaf = path->nodes[0];
 	extent = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_dev_extent);
@@ -2666,10 +2702,8 @@ static int insert_dev_extent(struct btrfs_trans_handle *trans,
 	btrfs_set_dev_extent_chunk_objectid(leaf, extent,
 					    BTRFS_FIRST_CHUNK_TREE_OBJECTID);
 	btrfs_set_dev_extent_chunk_offset(leaf, extent, chunk_offset);
-
 	btrfs_set_dev_extent_length(leaf, extent, num_bytes);
-out:
-	btrfs_free_path(path);
+
 	return ret;
 }
 
@@ -2771,8 +2805,12 @@ void btrfs_create_pending_block_groups(struct btrfs_trans_handle *trans)
 		/* Already aborted the transaction if it failed. */
 next:
 		btrfs_dec_delayed_refs_rsv_bg_inserts(fs_info);
+
+		spin_lock(&fs_info->unused_bgs_lock);
 		list_del_init(&block_group->bg_list);
 		clear_bit(BLOCK_GROUP_FLAG_NEW, &block_group->runtime_flags);
+		btrfs_put_block_group(block_group);
+		spin_unlock(&fs_info->unused_bgs_lock);
 
 		/*
 		 * If the block group is still unused, add it to the list of
@@ -2830,8 +2868,8 @@ static u64 calculate_global_root_id(const struct btrfs_fs_info *fs_info, u64 off
 }
 
 struct btrfs_block_group *btrfs_make_block_group(struct btrfs_trans_handle *trans,
-						 u64 type,
-						 u64 chunk_offset, u64 size)
+						 struct btrfs_space_info *space_info,
+						 u64 type, u64 chunk_offset, u64 size)
 {
 	struct btrfs_fs_info *fs_info = trans->fs_info;
 	struct btrfs_block_group *cache;
@@ -2885,10 +2923,10 @@ struct btrfs_block_group *btrfs_make_block_group(struct btrfs_trans_handle *tran
 	 * assigned to our block group. We want our bg to be added to the rbtree
 	 * with its ->space_info set.
 	 */
-	cache->space_info = btrfs_find_space_info(fs_info, cache->flags);
+	cache->space_info = space_info;
 	ASSERT(cache->space_info);
 
-	ret = btrfs_add_block_group_cache(fs_info, cache);
+	ret = btrfs_add_block_group_cache(cache);
 	if (ret) {
 		btrfs_remove_free_space_cache(cache);
 		btrfs_put_block_group(cache);
@@ -2910,7 +2948,7 @@ struct btrfs_block_group *btrfs_make_block_group(struct btrfs_trans_handle *tran
 	}
 #endif
 
-	list_add_tail(&cache->bg_list, &trans->new_bgs);
+	btrfs_link_bg_list(cache, &trans->new_bgs);
 	btrfs_inc_delayed_refs_rsv_bg_inserts(fs_info);
 
 	set_avail_alloc_bits(fs_info, type);
@@ -2930,6 +2968,7 @@ int btrfs_inc_block_group_ro(struct btrfs_block_group *cache,
 			     bool do_chunk_alloc)
 {
 	struct btrfs_fs_info *fs_info = cache->fs_info;
+	struct btrfs_space_info *space_info = cache->space_info;
 	struct btrfs_trans_handle *trans;
 	struct btrfs_root *root = btrfs_block_group_root(fs_info);
 	u64 alloc_flags;
@@ -2982,7 +3021,7 @@ int btrfs_inc_block_group_ro(struct btrfs_block_group *cache,
 		 */
 		alloc_flags = btrfs_get_alloc_profile(fs_info, cache->flags);
 		if (alloc_flags != cache->flags) {
-			ret = btrfs_chunk_alloc(trans, alloc_flags,
+			ret = btrfs_chunk_alloc(trans, space_info, alloc_flags,
 						CHUNK_ALLOC_FORCE);
 			/*
 			 * ENOSPC is allowed here, we may have enough space
@@ -3010,15 +3049,15 @@ int btrfs_inc_block_group_ro(struct btrfs_block_group *cache,
 	    (cache->flags & BTRFS_BLOCK_GROUP_SYSTEM))
 		goto unlock_out;
 
-	alloc_flags = btrfs_get_alloc_profile(fs_info, cache->space_info->flags);
-	ret = btrfs_chunk_alloc(trans, alloc_flags, CHUNK_ALLOC_FORCE);
+	alloc_flags = btrfs_get_alloc_profile(fs_info, space_info->flags);
+	ret = btrfs_chunk_alloc(trans, space_info, alloc_flags, CHUNK_ALLOC_FORCE);
 	if (ret < 0)
 		goto out;
 	/*
 	 * We have allocated a new chunk. We also need to activate that chunk to
 	 * grant metadata tickets for zoned filesystem.
 	 */
-	ret = btrfs_zoned_activate_one_bg(fs_info, cache->space_info, true);
+	ret = btrfs_zoned_activate_one_bg(fs_info, space_info, true);
 	if (ret < 0)
 		goto out;
 
@@ -3306,7 +3345,7 @@ int btrfs_setup_space_cache(struct btrfs_trans_handle *trans)
 	struct btrfs_fs_info *fs_info = trans->fs_info;
 	struct btrfs_block_group *cache, *tmp;
 	struct btrfs_transaction *cur_trans = trans->transaction;
-	struct btrfs_path *path;
+	BTRFS_PATH_AUTO_FREE(path);
 
 	if (list_empty(&cur_trans->dirty_bgs) ||
 	    !btrfs_test_opt(fs_info, SPACE_CACHE))
@@ -3323,7 +3362,6 @@ int btrfs_setup_space_cache(struct btrfs_trans_handle *trans)
 			cache_save_setup(cache, trans, path);
 	}
 
-	btrfs_free_path(path);
 	return 0;
 }
 
@@ -3346,7 +3384,7 @@ int btrfs_start_dirty_block_groups(struct btrfs_trans_handle *trans)
 	struct btrfs_transaction *cur_trans = trans->transaction;
 	int ret = 0;
 	int should_put;
-	struct btrfs_path *path = NULL;
+	BTRFS_PATH_AUTO_FREE(path);
 	LIST_HEAD(dirty);
 	struct list_head *io = &cur_trans->io_bgs;
 	int loops = 0;
@@ -3501,7 +3539,6 @@ out:
 		btrfs_cleanup_dirty_bgs(cur_trans, fs_info);
 	}
 
-	btrfs_free_path(path);
 	return ret;
 }
 
@@ -3512,7 +3549,7 @@ int btrfs_write_dirty_block_groups(struct btrfs_trans_handle *trans)
 	struct btrfs_transaction *cur_trans = trans->transaction;
 	int ret = 0;
 	int should_put;
-	struct btrfs_path *path;
+	BTRFS_PATH_AUTO_FREE(path);
 	struct list_head *io = &cur_trans->io_bgs;
 
 	path = btrfs_alloc_path();
@@ -3624,7 +3661,6 @@ int btrfs_write_dirty_block_groups(struct btrfs_trans_handle *trans)
 		btrfs_put_block_group(cache);
 	}
 
-	btrfs_free_path(path);
 	return ret;
 }
 
@@ -3703,8 +3739,8 @@ int btrfs_update_block_group(struct btrfs_trans_handle *trans,
 		spin_unlock(&cache->lock);
 		spin_unlock(&space_info->lock);
 
-		set_extent_bit(&trans->transaction->pinned_extents, bytenr,
-			       bytenr + num_bytes - 1, EXTENT_DIRTY, NULL);
+		btrfs_set_extent_bit(&trans->transaction->pinned_extents, bytenr,
+				     bytenr + num_bytes - 1, EXTENT_DIRTY, NULL);
 	}
 
 	spin_lock(&trans->transaction->dirty_bgs_lock);
@@ -3793,17 +3829,17 @@ out:
 /*
  * Update the block_group and space info counters.
  *
- * @cache:      The cache we are manipulating
- * @num_bytes:  The number of bytes in question
- * @delalloc:   The blocks are allocated for the delalloc write
+ * @cache:       The cache we are manipulating.
+ * @num_bytes:   The number of bytes in question.
+ * @is_delalloc: Whether the blocks are allocated for a delalloc write.
  *
  * This is called by somebody who is freeing space that was never actually used
  * on disk.  For example if you reserve some space for a new leaf in transaction
  * A and before transaction A commits you free that leaf, you call this with
  * reserve set to 0 in order to clear the reservation.
  */
-void btrfs_free_reserved_bytes(struct btrfs_block_group *cache,
-			       u64 num_bytes, int delalloc)
+void btrfs_free_reserved_bytes(struct btrfs_block_group *cache, u64 num_bytes,
+			       bool is_delalloc)
 {
 	struct btrfs_space_info *space_info = cache->space_info;
 
@@ -3817,7 +3853,7 @@ void btrfs_free_reserved_bytes(struct btrfs_block_group *cache,
 	space_info->bytes_reserved -= num_bytes;
 	space_info->max_extent_size = 0;
 
-	if (delalloc)
+	if (is_delalloc)
 		cache->delalloc_bytes -= num_bytes;
 	spin_unlock(&cache->lock);
 
@@ -3836,14 +3872,14 @@ static void force_metadata_allocation(struct btrfs_fs_info *info)
 	}
 }
 
-static int should_alloc_chunk(const struct btrfs_fs_info *fs_info,
-			      const struct btrfs_space_info *sinfo, int force)
+static bool should_alloc_chunk(const struct btrfs_fs_info *fs_info,
+			       const struct btrfs_space_info *sinfo, int force)
 {
 	u64 bytes_used = btrfs_space_info_used(sinfo, false);
 	u64 thresh;
 
 	if (force == CHUNK_ALLOC_FORCE)
-		return 1;
+		return true;
 
 	/*
 	 * in limited mode, we want to have some free space up to
@@ -3854,22 +3890,31 @@ static int should_alloc_chunk(const struct btrfs_fs_info *fs_info,
 		thresh = max_t(u64, SZ_64M, mult_perc(thresh, 1));
 
 		if (sinfo->total_bytes - bytes_used < thresh)
-			return 1;
+			return true;
 	}
 
 	if (bytes_used + SZ_2M < mult_perc(sinfo->total_bytes, 80))
-		return 0;
-	return 1;
+		return false;
+	return true;
 }
 
 int btrfs_force_chunk_alloc(struct btrfs_trans_handle *trans, u64 type)
 {
 	u64 alloc_flags = btrfs_get_alloc_profile(trans->fs_info, type);
+	struct btrfs_space_info *space_info;
 
-	return btrfs_chunk_alloc(trans, alloc_flags, CHUNK_ALLOC_FORCE);
+	space_info = btrfs_find_space_info(trans->fs_info, type);
+	if (!space_info) {
+		DEBUG_WARN();
+		return -EINVAL;
+	}
+
+	return btrfs_chunk_alloc(trans, space_info, alloc_flags, CHUNK_ALLOC_FORCE);
 }
 
-static struct btrfs_block_group *do_chunk_alloc(struct btrfs_trans_handle *trans, u64 flags)
+static struct btrfs_block_group *do_chunk_alloc(struct btrfs_trans_handle *trans,
+						struct btrfs_space_info *space_info,
+						u64 flags)
 {
 	struct btrfs_block_group *bg;
 	int ret;
@@ -3882,7 +3927,7 @@ static struct btrfs_block_group *do_chunk_alloc(struct btrfs_trans_handle *trans
 	 */
 	check_system_chunk(trans, flags);
 
-	bg = btrfs_create_chunk(trans, flags);
+	bg = btrfs_create_chunk(trans, space_info, flags);
 	if (IS_ERR(bg)) {
 		ret = PTR_ERR(bg);
 		goto out;
@@ -3930,8 +3975,16 @@ static struct btrfs_block_group *do_chunk_alloc(struct btrfs_trans_handle *trans
 	if (ret == -ENOSPC) {
 		const u64 sys_flags = btrfs_system_alloc_profile(trans->fs_info);
 		struct btrfs_block_group *sys_bg;
+		struct btrfs_space_info *sys_space_info;
+
+		sys_space_info = btrfs_find_space_info(trans->fs_info, sys_flags);
+		if (!sys_space_info) {
+			ret = -EINVAL;
+			btrfs_abort_transaction(trans, ret);
+			goto out;
+		}
 
-		sys_bg = btrfs_create_chunk(trans, sys_flags);
+		sys_bg = btrfs_create_chunk(trans, sys_space_info, sys_flags);
 		if (IS_ERR(sys_bg)) {
 			ret = PTR_ERR(sys_bg);
 			btrfs_abort_transaction(trans, ret);
@@ -4062,6 +4115,8 @@ out:
  *
  * This function, btrfs_chunk_alloc(), belongs to phase 1.
  *
+ * @space_info: specify which space_info the new chunk should belong to.
+ *
  * If @force is CHUNK_ALLOC_FORCE:
  *    - return 1 if it successfully allocates a chunk,
  *    - return errors including -ENOSPC otherwise.
@@ -4070,11 +4125,11 @@ out:
  *    - return 1 if it successfully allocates a chunk,
  *    - return errors including -ENOSPC otherwise.
  */
-int btrfs_chunk_alloc(struct btrfs_trans_handle *trans, u64 flags,
+int btrfs_chunk_alloc(struct btrfs_trans_handle *trans,
+		      struct btrfs_space_info *space_info, u64 flags,
 		      enum btrfs_chunk_alloc_enum force)
 {
 	struct btrfs_fs_info *fs_info = trans->fs_info;
-	struct btrfs_space_info *space_info;
 	struct btrfs_block_group *ret_bg;
 	bool wait_for_alloc = false;
 	bool should_alloc = false;
@@ -4113,9 +4168,6 @@ int btrfs_chunk_alloc(struct btrfs_trans_handle *trans, u64 flags,
 	if (flags & BTRFS_BLOCK_GROUP_SYSTEM)
 		return -ENOSPC;
 
-	space_info = btrfs_find_space_info(fs_info, flags);
-	ASSERT(space_info);
-
 	do {
 		spin_lock(&space_info->lock);
 		if (force < space_info->force_alloc)
@@ -4176,7 +4228,7 @@ int btrfs_chunk_alloc(struct btrfs_trans_handle *trans, u64 flags,
 			force_metadata_allocation(fs_info);
 	}
 
-	ret_bg = do_chunk_alloc(trans, flags);
+	ret_bg = do_chunk_alloc(trans, space_info, flags);
 	trans->allocating_chunk = false;
 
 	if (IS_ERR(ret_bg)) {
@@ -4252,6 +4304,10 @@ static void reserve_chunk_space(struct btrfs_trans_handle *trans,
 	if (left < bytes) {
 		u64 flags = btrfs_system_alloc_profile(fs_info);
 		struct btrfs_block_group *bg;
+		struct btrfs_space_info *space_info;
+
+		space_info = btrfs_find_space_info(fs_info, flags);
+		ASSERT(space_info);
 
 		/*
 		 * Ignore failure to create system chunk. We might end up not
@@ -4259,7 +4315,7 @@ static void reserve_chunk_space(struct btrfs_trans_handle *trans,
 		 * the paths we visit in the chunk tree (they were already COWed
 		 * or created in the current transaction for example).
 		 */
-		bg = btrfs_create_chunk(trans, flags);
+		bg = btrfs_create_chunk(trans, space_info, flags);
 		if (IS_ERR(bg)) {
 			ret = PTR_ERR(bg);
 		} else {
@@ -4367,6 +4423,43 @@ void btrfs_put_block_group_cache(struct btrfs_fs_info *info)
 	}
 }
 
+static void check_removing_space_info(struct btrfs_space_info *space_info)
+{
+	struct btrfs_fs_info *info = space_info->fs_info;
+
+	if (space_info->subgroup_id == BTRFS_SUB_GROUP_PRIMARY) {
+		/* This is a top space_info, proceed with its children first. */
+		for (int i = 0; i < BTRFS_SPACE_INFO_SUB_GROUP_MAX; i++) {
+			if (space_info->sub_group[i]) {
+				check_removing_space_info(space_info->sub_group[i]);
+				kfree(space_info->sub_group[i]);
+				space_info->sub_group[i] = NULL;
+			}
+		}
+	}
+
+	/*
+	 * Do not hide this behind enospc_debug, this is actually important and
+	 * indicates a real bug if this happens.
+	 */
+	if (WARN_ON(space_info->bytes_pinned > 0 || space_info->bytes_may_use > 0))
+		btrfs_dump_space_info(info, space_info, 0, 0);
+
+	/*
+	 * If there was a failure to cleanup a log tree, very likely due to an
+	 * IO failure on a writeback attempt of one or more of its extent
+	 * buffers, we could not do proper (and cheap) unaccounting of their
+	 * reserved space, so don't warn on bytes_reserved > 0 in that case.
+	 */
+	if (!(space_info->flags & BTRFS_BLOCK_GROUP_METADATA) ||
+	    !BTRFS_FS_LOG_CLEANUP_ERROR(info)) {
+		if (WARN_ON(space_info->bytes_reserved > 0))
+			btrfs_dump_space_info(info, space_info, 0, 0);
+	}
+
+	WARN_ON(space_info->reclaim_size > 0);
+}
+
 /*
  * Must be called only after stopping all workers, since we could have block
  * group caching kthreads running, and therefore they could race with us if we
@@ -4392,8 +4485,8 @@ int btrfs_free_block_groups(struct btrfs_fs_info *info)
 
 	write_lock(&info->block_group_cache_lock);
 	while (!list_empty(&info->caching_block_groups)) {
-		caching_ctl = list_entry(info->caching_block_groups.next,
-					 struct btrfs_caching_control, list);
+		caching_ctl = list_first_entry(&info->caching_block_groups,
+					       struct btrfs_caching_control, list);
 		list_del(&caching_ctl->list);
 		btrfs_put_caching_control(caching_ctl);
 	}
@@ -4464,32 +4557,10 @@ int btrfs_free_block_groups(struct btrfs_fs_info *info)
 	btrfs_release_global_block_rsv(info);
 
 	while (!list_empty(&info->space_info)) {
-		space_info = list_entry(info->space_info.next,
-					struct btrfs_space_info,
-					list);
-
-		/*
-		 * Do not hide this behind enospc_debug, this is actually
-		 * important and indicates a real bug if this happens.
-		 */
-		if (WARN_ON(space_info->bytes_pinned > 0 ||
-			    space_info->bytes_may_use > 0))
-			btrfs_dump_space_info(info, space_info, 0, 0);
-
-		/*
-		 * If there was a failure to cleanup a log tree, very likely due
-		 * to an IO failure on a writeback attempt of one or more of its
-		 * extent buffers, we could not do proper (and cheap) unaccounting
-		 * of their reserved space, so don't warn on bytes_reserved > 0 in
-		 * that case.
-		 */
-		if (!(space_info->flags & BTRFS_BLOCK_GROUP_METADATA) ||
-		    !BTRFS_FS_LOG_CLEANUP_ERROR(info)) {
-			if (WARN_ON(space_info->bytes_reserved > 0))
-				btrfs_dump_space_info(info, space_info, 0, 0);
-		}
+		space_info = list_first_entry(&info->space_info,
+					      struct btrfs_space_info, list);
 
-		WARN_ON(space_info->reclaim_size > 0);
+		check_removing_space_info(space_info);
 		list_del(&space_info->list);
 		btrfs_sysfs_remove_space_info(space_info);
 	}