Btrfs: fix corrupted metadata in the snapshot

When we delete a inode, we will remove all the delayed items including delayed inode update, and then truncate all the relative metadata. If there is lots of metadata, we will end the current transaction, and start a new transaction to truncate the left metadata. In this way, we will leave a inode item that its link counter is > 0, and also may leave some directory index items in fs/file tree after the current transaction ends. In other words, the metadata in this fs/file tree is inconsistent. If we create a snapshot for this tree now, we will find a inode with corrupted metadata in the new snapshot, and we won't continue to drop the left metadata, because its link counter is not 0.

We fix this problem by updating the inode item before the current transaction ends.

Signed-off-by Miao Xie <[email protected]>

---
 fs/btrfs/inode.c |    5 ++++-
 1 files changed, 4 insertions(+), 1 deletions(-)

fs/btrfs/inode.c +4 -1

index cae4c32..02eeecb 100644
@@ -3736,7 +3736,7 @@ void btrfs_evict_inode(struct inode *inode)
 	struct btrfs_trans_handle *trans;
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	struct btrfs_block_rsv *rsv, *global_rsv;
-	u64 min_size = btrfs_calc_trunc_metadata_size(root, 1);
+	u64 min_size = btrfs_calc_trunc_metadata_size(root, 2);
 	unsigned long nr;
 	int ret;
 

@@ -3818,6 +3818,9 @@ void btrfs_evict_inode(struct inode *inode)
 		if (ret != -EAGAIN)
 			break;
 
+		ret = btrfs_update_inode(trans, root, inode);
+		BUG_ON(ret);
+
 		nr = trans->blocks_used;
 		btrfs_end_transaction(trans, root);
 		trans = NULL;

Quoted 11 lines · Miao Xie

When we delete a inode, we will remove all the delayed items including delayed inode update, and then truncate all the relative metadata. If there is lots of metadata, we will end the current transaction, and start a new transaction to truncate the left metadata. In this way, we will leave a inode item that its link counter is > 0, and also may leave some directory index items in fs/file tree after the current transaction ends. In other words, the metadata in this fs/file tree is inconsistent. If we create a snapshot for this tree now, we will find a inode with corrupted metadata in the new snapshot, and we won't continue to drop the left metadata, because its link counter is not 0.

We fix this problem by updating the inode item before the current transaction ends.

A comment before the while() says

3780 /*

3781          * This is a bit simpler than btrfs_truncate since
3782          *
3783          * 1) We've already reserved our space for our orphan item in the
3784          *    unlink.
3785          * 2) We're going to delete the inode item, so we don't need to update
3786          *    it at all.
3787          *
3788          * So we just need to reserve some slack space in case we add bytes when
3789          * doing the truncate.
3790          */

Point 2 states that the inode update is not needed, but as you write in the changelog it can lead to inconsistent metadata. I can't say either way, but rather would like to hear Josef's oppinion on that, as the comment and related code comes from
4289a667a0d7c6b134898cac7bfbe950267c305c
(Btrfs: fix how we reserve space for deleting inodes)