Re: [PATCH] generic: test a scenario of power failure after renames and fsyncs

["Darrick J. Wong" <djwong@kernel.org>]

On Thu, Nov 27, 2025 at 06:17:44PM +0000, fdmanana@kernel.org wrote:
> From: Filipe Manana <fdmanana@suse.com>
> 
> Test moving a directory to another location, create a file in the old
> location of the directory and with the same name, fsync the file, then
> move the file elsewhere and fsync again the file. After a power failure
> we expect to be able to mount the fs and have the same content as before
> the power failure.
> 
> This exercises a bug fixed by the following kernel patch for btrfs:
> 
>  "btrfs: don't log conflicting inode if it's a dir moved in the current transaction"
> 
> Signed-off-by: Filipe Manana <fdmanana@suse.com>

Seems reasonable to me.
Reviewed-by: "Darrick J. Wong" <djwong@kernel.org>

--D

> ---
>  tests/generic/784     | 76 +++++++++++++++++++++++++++++++++++++++++++
>  tests/generic/784.out | 19 +++++++++++
>  2 files changed, 95 insertions(+)
>  create mode 100755 tests/generic/784
>  create mode 100644 tests/generic/784.out
> 
> diff --git a/tests/generic/784 b/tests/generic/784
> new file mode 100755
> index 00000000..8e01dff0
> --- /dev/null
> +++ b/tests/generic/784
> @@ -0,0 +1,76 @@
> +#! /bin/bash
> +# SPDX-License-Identifier: GPL-2.0
> +# Copyright (c) 2025 SUSE S.A.  All Rights Reserved.
> +#
> +# FS QA Test 784
> +#
> +# Test moving a directory to another location, create a file in the old location
> +# of the directory and with the same name, fsync the file, then move the file
> +# elsewhere and fsync again the file. After a power failure we expect to be able
> +# to mount the fs and have the same content as before the power failure.
> +#
> +. ./common/preamble
> +_begin_fstest auto quick log
> +
> +_cleanup()
> +{
> +	_cleanup_flakey
> +	cd /
> +	rm -r -f $tmp.*
> +}
> +
> +. ./common/filter
> +. ./common/dmflakey
> +
> +_require_scratch
> +_require_dm_target flakey
> +
> +[ "$FSTYP" = "btrfs" ] && _fixed_by_kernel_commit xxxxxxxxxxxx \
> +	"btrfs: don't log conflicting inode if it's a dir moved in the current transaction"
> +
> +_scratch_mkfs >>$seqres.full 2>&1 || _fail "mkfs failed"
> +_require_metadata_journaling $SCRATCH_DEV
> +_init_flakey
> +_mount_flakey
> +
> +list_fs_contents()
> +{
> +	# Remove the 'lost+found' directory which only exists in some
> +	# filesystems like extN and remove empty lines to have a consistent
> +	# output regardless of the existence of the 'lost+found' directory.
> +	ls -1R $SCRATCH_MNT/ | grep -v 'lost+found' | grep -v -e '^$' | \
> +		_filter_scratch
> +}
> +
> +mkdir $SCRATCH_MNT/dir1
> +mkdir $SCRATCH_MNT/dir2
> +
> +_scratch_sync
> +
> +mv $SCRATCH_MNT/dir1 $SCRATCH_MNT/dir2
> +
> +# Create a file with the old name of the first directory and persist it with
> +# a fsync.
> +$XFS_IO_PROG -f -c "fsync" $SCRATCH_MNT/dir1
> +
> +# Move the file to some other location.
> +mv $SCRATCH_MNT/dir1 $SCRATCH_MNT/dir2/foo
> +
> +# Fsync again the file.
> +$XFS_IO_PROG -c "fsync" $SCRATCH_MNT/dir2/foo
> +
> +echo -e "\nfs contents before power failure:\n"
> +list_fs_contents
> +
> +# Simulate a power failure and then mount again the filesystem to replay the
> +# journal/log. We expect the mount to succeed.
> +_flakey_drop_and_remount
> +
> +# We expect the fs contents to be the same as before the power failure.
> +echo -e "\nfs contents after power failure:\n"
> +list_fs_contents
> +
> +_unmount_flakey
> +
> +# success, all done
> +_exit 0
> diff --git a/tests/generic/784.out b/tests/generic/784.out
> new file mode 100644
> index 00000000..87942d04
> --- /dev/null
> +++ b/tests/generic/784.out
> @@ -0,0 +1,19 @@
> +QA output created by 784
> +
> +fs contents before power failure:
> +
> +SCRATCH_MNT/:
> +dir2
> +SCRATCH_MNT/dir2:
> +dir1
> +foo
> +SCRATCH_MNT/dir2/dir1:
> +
> +fs contents after power failure:
> +
> +SCRATCH_MNT/:
> +dir2
> +SCRATCH_MNT/dir2:
> +dir1
> +foo
> +SCRATCH_MNT/dir2/dir1:
> -- 
> 2.47.2
> 
>