1 files changed, 32 insertions, 16 deletions

diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 1beb9458f622..9cc14ab35297 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -884,7 +884,7 @@ struct btrfs_root *btrfs_create_tree(struct btrfs_trans_handle *trans,
 	btrfs_set_root_used(&root->root_item, leaf->len);
 	btrfs_set_root_last_snapshot(&root->root_item, 0);
 	btrfs_set_root_dirid(&root->root_item, 0);
-	if (is_fstree(objectid))
+	if (btrfs_is_fstree(objectid))
 		generate_random_guid(root->root_item.uuid);
 	else
 		export_guid(root->root_item.uuid, &guid_null);
@@ -1104,7 +1104,7 @@ static int btrfs_init_fs_root(struct btrfs_root *root, dev_t anon_dev)
 
 	if (btrfs_root_id(root) != BTRFS_TREE_LOG_OBJECTID &&
 	    !btrfs_is_data_reloc_root(root) &&
-	    is_fstree(btrfs_root_id(root))) {
+	    btrfs_is_fstree(btrfs_root_id(root))) {
 		set_bit(BTRFS_ROOT_SHAREABLE, &root->state);
 		btrfs_check_and_init_root_item(&root->root_item);
 	}
@@ -1113,7 +1113,7 @@ static int btrfs_init_fs_root(struct btrfs_root *root, dev_t anon_dev)
 	 * Don't assign anonymous block device to roots that are not exposed to
 	 * userspace, the id pool is limited to 1M
 	 */
-	if (is_fstree(btrfs_root_id(root)) &&
+	if (btrfs_is_fstree(btrfs_root_id(root)) &&
 	    btrfs_root_refs(&root->root_item) > 0) {
 		if (!anon_dev) {
 			ret = get_anon_bdev(&root->anon_dev);
@@ -1246,6 +1246,8 @@ void btrfs_free_fs_info(struct btrfs_fs_info *fs_info)
 {
 	struct percpu_counter *em_counter = &fs_info->evictable_extent_maps;
 
+	if (fs_info->fs_devices)
+		btrfs_close_devices(fs_info->fs_devices);
 	percpu_counter_destroy(&fs_info->stats_read_blocks);
 	percpu_counter_destroy(&fs_info->dirty_metadata_bytes);
 	percpu_counter_destroy(&fs_info->delalloc_bytes);
@@ -1315,7 +1317,7 @@ static struct btrfs_root *btrfs_get_root_ref(struct btrfs_fs_info *fs_info,
 	 * This is namely for free-space-tree and quota tree, which can change
 	 * at runtime and should only be grabbed from fs_info.
 	 */
-	if (!is_fstree(objectid) && objectid != BTRFS_DATA_RELOC_TREE_OBJECTID)
+	if (!btrfs_is_fstree(objectid) && objectid != BTRFS_DATA_RELOC_TREE_OBJECTID)
 		return ERR_PTR(-ENOENT);
 again:
 	root = btrfs_lookup_fs_root(fs_info, objectid);
@@ -1835,6 +1837,8 @@ void btrfs_put_root(struct btrfs_root *root)
 	if (refcount_dec_and_test(&root->refs)) {
 		if (WARN_ON(!xa_empty(&root->inodes)))
 			xa_destroy(&root->inodes);
+		if (WARN_ON(!xa_empty(&root->delayed_nodes)))
+			xa_destroy(&root->delayed_nodes);
 		WARN_ON(test_bit(BTRFS_ROOT_DEAD_RELOC_TREE, &root->state));
 		if (root->anon_dev)
 			free_anon_bdev(root->anon_dev);
@@ -1945,7 +1949,6 @@ static void btrfs_init_qgroup(struct btrfs_fs_info *fs_info)
 	fs_info->qgroup_tree = RB_ROOT;
 	INIT_LIST_HEAD(&fs_info->dirty_qgroups);
 	fs_info->qgroup_seq = 1;
-	fs_info->qgroup_ulist = NULL;
 	fs_info->qgroup_rescan_running = false;
 	fs_info->qgroup_drop_subtree_thres = BTRFS_QGROUP_DROP_SUBTREE_THRES_DEFAULT;
 	mutex_init(&fs_info->qgroup_rescan_lock);
@@ -2156,8 +2159,7 @@ static int load_global_roots_objectid(struct btrfs_root *tree_root,
 		found = true;
 		root = read_tree_root_path(tree_root, path, &key);
 		if (IS_ERR(root)) {
-			if (!btrfs_test_opt(fs_info, IGNOREBADROOTS))
-				ret = PTR_ERR(root);
+			ret = PTR_ERR(root);
 			break;
 		}
 		set_bit(BTRFS_ROOT_TRACK_DIRTY, &root->state);
@@ -3395,6 +3397,7 @@ int __cold open_ctree(struct super_block *sb, struct btrfs_fs_devices *fs_device
 	fs_info->delalloc_batch = sectorsize * 512 * (1 + ilog2(nr_cpu_ids));
 
 	fs_info->nodesize = nodesize;
+	fs_info->nodesize_bits = ilog2(nodesize);
 	fs_info->sectorsize = sectorsize;
 	fs_info->sectorsize_bits = ilog2(sectorsize);
 	fs_info->csums_per_leaf = BTRFS_MAX_ITEM_SIZE(fs_info) / fs_info->csum_size;
@@ -3560,6 +3563,7 @@ int __cold open_ctree(struct super_block *sb, struct btrfs_fs_devices *fs_device
 		goto fail_sysfs;
 	}
 
+	btrfs_zoned_reserve_data_reloc_bg(fs_info);
 	btrfs_free_zone_cache(fs_info);
 
 	btrfs_check_active_zone_reservation(fs_info);
@@ -3680,7 +3684,6 @@ fail_alloc:
 
 	iput(fs_info->btree_inode);
 fail:
-	btrfs_close_devices(fs_info->fs_devices);
 	ASSERT(ret < 0);
 	return ret;
 }
@@ -3693,7 +3696,7 @@ static void btrfs_end_super_write(struct bio *bio)
 
 	bio_for_each_folio_all(fi, bio) {
 		if (bio->bi_status) {
-			btrfs_warn_rl_in_rcu(device->fs_info,
+			btrfs_warn_rl(device->fs_info,
 				"lost super block write due to IO error on %s (%d)",
 				btrfs_dev_name(device),
 				blk_status_to_errno(bio->bi_status));
@@ -3991,7 +3994,7 @@ int btrfs_get_num_tolerated_disk_barrier_failures(u64 flags)
 	}
 
 	if (min_tolerated == INT_MAX) {
-		pr_warn("BTRFS: unknown raid flag: %llu", flags);
+		btrfs_warn(NULL, "unknown raid flag: %llu", flags);
 		min_tolerated = 0;
 	}
 
@@ -4310,8 +4313,8 @@ void __cold close_ctree(struct btrfs_fs_info *fs_info)
 	 *
 	 * So wait for all ongoing ordered extents to complete and then run
 	 * delayed iputs. This works because once we reach this point no one
-	 * can either create new ordered extents nor create delayed iputs
-	 * through some other means.
+	 * can create new ordered extents, but delayed iputs can still be added
+	 * by a reclaim worker (see comments further below).
 	 *
 	 * Also note that btrfs_wait_ordered_roots() is not safe here, because
 	 * it waits for BTRFS_ORDERED_COMPLETE to be set on an ordered extent,
@@ -4322,15 +4325,29 @@ void __cold close_ctree(struct btrfs_fs_info *fs_info)
 	btrfs_flush_workqueue(fs_info->endio_write_workers);
 	/* Ordered extents for free space inodes. */
 	btrfs_flush_workqueue(fs_info->endio_freespace_worker);
+	/*
+	 * Run delayed iputs in case an async reclaim worker is waiting for them
+	 * to be run as mentioned above.
+	 */
 	btrfs_run_delayed_iputs(fs_info);
-	/* There should be no more workload to generate new delayed iputs. */
-	set_bit(BTRFS_FS_STATE_NO_DELAYED_IPUT, &fs_info->fs_state);
 
 	cancel_work_sync(&fs_info->async_reclaim_work);
 	cancel_work_sync(&fs_info->async_data_reclaim_work);
 	cancel_work_sync(&fs_info->preempt_reclaim_work);
 	cancel_work_sync(&fs_info->em_shrinker_work);
 
+	/*
+	 * Run delayed iputs again because an async reclaim worker may have
+	 * added new ones if it was flushing delalloc:
+	 *
+	 * shrink_delalloc() -> btrfs_start_delalloc_roots() ->
+	 *    start_delalloc_inodes() -> btrfs_add_delayed_iput()
+	 */
+	btrfs_run_delayed_iputs(fs_info);
+
+	/* There should be no more workload to generate new delayed iputs. */
+	set_bit(BTRFS_FS_STATE_NO_DELAYED_IPUT, &fs_info->fs_state);
+
 	/* Cancel or finish ongoing discard work */
 	btrfs_discard_cleanup(fs_info);
 
@@ -4413,7 +4430,6 @@ void __cold close_ctree(struct btrfs_fs_info *fs_info)
 	iput(fs_info->btree_inode);
 
 	btrfs_mapping_tree_free(fs_info);
-	btrfs_close_devices(fs_info->fs_devices);
 }
 
 void btrfs_mark_buffer_dirty(struct btrfs_trans_handle *trans,
@@ -4625,7 +4641,7 @@ static void btrfs_destroy_marked_extents(struct btrfs_fs_info *fs_info,
 
 	while (btrfs_find_first_extent_bit(dirty_pages, start, &start, &end,
 					   mark, NULL)) {
-		btrfs_clear_extent_bits(dirty_pages, start, end, mark);
+		btrfs_clear_extent_bit(dirty_pages, start, end, mark, NULL);
 		while (start <= end) {
 			eb = find_extent_buffer(fs_info, start);
 			start += fs_info->nodesize;