On Tue, Oct 24, 2023 at 12:23:46PM +0100, fdmanana@xxxxxxxxxx wrote:
> From: Filipe Manana <fdmanana@xxxxxxxx>
>
> Test that snapshotting a new subvolume (created in the current transaction)
> that has a btree with a height > 1, works and does not result in a fs
> corruption.
>
> This exercises a regression introduced in kernel 6.5 by the kernel commit:
>
> 1b53e51a4a8f ("btrfs: don't commit transaction for every subvol create")
>
> Signed-off-by: Filipe Manana <fdmanana@xxxxxxxx>
> Reviewed-by: Josef Bacik <josef@xxxxxxxxxxxxxx>
> ---
>
> v2: Add git commit ID for the kernel fix, as it was merged yesterday into
> Linus' tree. Also fixed a typo and included Josef's Reviewed-by tag.
Thanks for doing this!
Reviewed-by: Zorro Lang <zlang@xxxxxxxxxx>
>
> tests/btrfs/302 | 61 +++++++++++++++++++++++++++++++++++++++++++++
> tests/btrfs/302.out | 4 +++
> 2 files changed, 65 insertions(+)
> create mode 100755 tests/btrfs/302
> create mode 100644 tests/btrfs/302.out
>
> diff --git a/tests/btrfs/302 b/tests/btrfs/302
> new file mode 100755
> index 00000000..f3e6044b
> --- /dev/null
> +++ b/tests/btrfs/302
> @@ -0,0 +1,61 @@
> +#! /bin/bash
> +# SPDX-License-Identifier: GPL-2.0
> +# Copyright (C) 2023 SUSE Linux Products GmbH. All Rights Reserved.
> +#
> +# FS QA Test 302
> +#
> +# Test that snapshotting a new subvolume (created in the current transaction)
> +# that has a btree with a height > 1, works and does not result in a filesystem
> +# corruption.
> +#
> +# This exercises a regression introduced in kernel 6.5 by the kernel commit:
> +#
> +# 1b53e51a4a8f ("btrfs: don't commit transaction for every subvol create")
> +#
> +. ./common/preamble
> +_begin_fstest auto quick snapshot subvol
> +
> +. ./common/filter
> +
> +_supported_fs btrfs
> +_require_scratch
> +_require_fssum
> +
> +_fixed_by_kernel_commit eb96e221937a \
> + "btrfs: fix unwritten extent buffer after snapshotting a new subvolume"
> +
> +# Use a filesystem with a 64K node size so that we have the same node size on
> +# every machine regardless of its page size (on x86_64 default node size is 16K
> +# due to the 4K page size, while on PPC it's 64K by default). This way we can
> +# make sure we are able to create a btree for the subvolume with a height of 2.
> +_scratch_mkfs -n 64K >> $seqres.full 2>&1 || _fail "mkfs failed"
> +_scratch_mount
> +
> +$BTRFS_UTIL_PROG subvolume create $SCRATCH_MNT/subvol | _filter_scratch
> +
> +# Create a few empty files on the subvolume, this bumps its btree height to 2
> +# (root node at level 1 and 2 leaves).
> +for ((i = 1; i <= 300; i++)); do
> + echo -n > $SCRATCH_MNT/subvol/file_$i
> +done
> +
> +# Create a checksum of the subvolume's content.
> +fssum_file="$SCRATCH_MNT/checksum.fssum"
> +$FSSUM_PROG -A -f -w $fssum_file $SCRATCH_MNT/subvol
> +
> +# Now create a snapshot of the subvolume and make it accessible from within the
> +# subvolume.
> +$BTRFS_UTIL_PROG subvolume snapshot -r $SCRATCH_MNT/subvol \
> + $SCRATCH_MNT/subvol/snap | _filter_scratch
> +
> +# Now unmount and mount again the fs. We want to verify we are able to read all
> +# metadata for the snapshot from disk (no IO failures, etc).
> +_scratch_cycle_mount
> +
> +# The snapshot's content should match the subvolume's content before we created
> +# the snapshot.
> +$FSSUM_PROG -r $fssum_file $SCRATCH_MNT/subvol/snap
> +
> +# success, all done
> +status=0
> +exit
> diff --git a/tests/btrfs/302.out b/tests/btrfs/302.out
> new file mode 100644
> index 00000000..8770aefc
> --- /dev/null
> +++ b/tests/btrfs/302.out
> @@ -0,0 +1,4 @@
> +QA output created by 302
> +Create subvolume 'SCRATCH_MNT/subvol'
> +Create a readonly snapshot of 'SCRATCH_MNT/subvol' in 'SCRATCH_MNT/subvol/snap'
> +OK
> --
> 2.40.1
>
