[PATCH] xfs: scrub: Remove incorrect check executed on block format directories

[PATCH] xfs: scrub: Remove incorrect check executed on block format directories

[Chandan Babu R]

A directory with one directory block which in turns consists of two or more fs
blocks is incorrectly flagged as corrupt by scrub since it assumes that
"Block" format directories have a data fork single extent spanning the file
offset range of [0, Dir block size - 1].

This commit fixes the bug by removing the incorrect check.

Signed-off-by: Chandan Babu R <chandanrlinux@gmail.com>
---
 fs/xfs/scrub/dir.c | 9 ---------
 1 file changed, 9 deletions(-)

diff --git a/fs/xfs/scrub/dir.c b/fs/xfs/scrub/dir.c
index 178b3455a170..3ec6290c78bb 100644
--- a/fs/xfs/scrub/dir.c
+++ b/fs/xfs/scrub/dir.c
@@ -694,15 +694,6 @@ xchk_directory_blocks(
 	/* Iterate all the data extents in the directory... */
 	found = xfs_iext_lookup_extent(sc->ip, ifp, lblk, &icur, &got);
 	while (found && !(sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)) {
-		/* Block directories only have a single block at offset 0. */
-		if (is_block &&
-		    (got.br_startoff > 0 ||
-		     got.br_blockcount != args.geo->fsbcount)) {
-			xchk_fblock_set_corrupt(sc, XFS_DATA_FORK,
-					got.br_startoff);
-			break;
-		}
-
 		/* No more data blocks... */
 		if (got.br_startoff >= leaf_lblk)
 			break;
-- 
2.29.2

Re: [PATCH] xfs: scrub: Remove incorrect check executed on block format directories

[Darrick J. Wong]

On Fri, Mar 26, 2021 at 05:03:12PM +0530, Chandan Babu R wrote:
> A directory with one directory block which in turns consists of two or more fs
> blocks is incorrectly flagged as corrupt by scrub since it assumes that
> "Block" format directories have a data fork single extent spanning the file
> offset range of [0, Dir block size - 1].
> 
> This commit fixes the bug by removing the incorrect check.

It took me a minute to figure out that this isn't merely removing a
check that triggered somewhere; the check is incorrect /and/ redundant
because xfs_dir2_isblock sets is_block based on what it finds in the
iext tree, so it makes no sense to check that result against the iext
tree at all.

Reviewed-by: Darrick J. Wong <djwong@kernel.org>

--D


> Signed-off-by: Chandan Babu R <chandanrlinux@gmail.com>
> ---
>  fs/xfs/scrub/dir.c | 9 ---------
>  1 file changed, 9 deletions(-)
> 
> diff --git a/fs/xfs/scrub/dir.c b/fs/xfs/scrub/dir.c
> index 178b3455a170..3ec6290c78bb 100644
> --- a/fs/xfs/scrub/dir.c
> +++ b/fs/xfs/scrub/dir.c
> @@ -694,15 +694,6 @@ xchk_directory_blocks(
>  	/* Iterate all the data extents in the directory... */
>  	found = xfs_iext_lookup_extent(sc->ip, ifp, lblk, &icur, &got);
>  	while (found && !(sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)) {
> -		/* Block directories only have a single block at offset 0. */
> -		if (is_block &&
> -		    (got.br_startoff > 0 ||
> -		     got.br_blockcount != args.geo->fsbcount)) {
> -			xchk_fblock_set_corrupt(sc, XFS_DATA_FORK,
> -					got.br_startoff);
> -			break;
> -		}
> -
>  		/* No more data blocks... */
>  		if (got.br_startoff >= leaf_lblk)
>  			break;
> -- 
> 2.29.2
>