[v2,6/7] btrfs-progs: mkfs: Workaround BUG_ON caused by rootdir option
diff mbox

Message ID 20170911063612.32114-7-quwenruo.btrfs@gmx.com
State New
Headers show

Commit Message

Qu Wenruo Sept. 11, 2017, 6:36 a.m. UTC
--rootdir option will start a transaction to fill the fs, however if
something goes wrong, from ENOSPC to lack of permission, we won't commit
transaction and cause BUG_ON trigger by uncommitted transaction:

extent buffer leak: start 29392896 len 16384
extent_io.c:579: free_extent_buffer: BUG_ON `eb->flags & EXTENT_DIRTY` triggered, value 1

The root fix is to introduce btrfs_abort_transaction() in btrfs-progs,
however in this particular case, we can workaround it by force
committing the transaction.

Since during mkfs, the magic of btrfs is set to an invalid one, without
setting fs_info->finalize_on_close() the fs is never able to be mounted.
So even we force to commit wrong transaction we won't screw up things

Signed-off-by: Qu Wenruo <quwenruo.btrfs@gmx.com>
 mkfs/main.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff mbox

diff --git a/mkfs/main.c b/mkfs/main.c
index 6561ac52..f78c24ce 100644
--- a/mkfs/main.c
+++ b/mkfs/main.c
@@ -1025,6 +1025,18 @@  static int make_image(const char *source_dir, struct btrfs_root *root)
 		printf("Making image is completed.\n");
 	return 0;
+	/*
+	 * XXX:
+	 * To avoid BUG_ON() triggered by uncommitted transaction,
+	 * here we must commit transaction before we have proper
+	 * btrfs_abort_transaction() in btrfs-progs.
+	 *
+	 * Don't worry, the magic number is not valid so the fs can't be
+	 * mounted by kernel even we commit the trans.
+	 * And we don't want to pollute the original error, so we ignore
+	 * the return value from btrfs_commit_transaction().
+	 */
+	btrfs_commit_transaction(trans, root);
 	while (!list_empty(&dir_head.list)) {
 		dir_entry = list_entry(dir_head.list.next,
 				       struct directory_name_entry, list);