linux-ext4 - Re: [PATCH 10/12] xfs/528: require a real SCRATCH

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20251210195220.GD94594@frogsfrogsfrogs>
Date: Wed, 10 Dec 2025 11:52:20 -0800
From: "Darrick J. Wong" <djwong@...nel.org>
To: Christoph Hellwig <hch@....de>
Cc: Zorro Lang <zlang@...nel.org>, Anand Jain <anand.jain@...cle.com>,
	Filipe Manana <fdmanana@...e.com>, fstests@...r.kernel.org,
	linux-ext4@...r.kernel.org, linux-xfs@...r.kernel.org
Subject: Re: [PATCH 10/12] xfs/528: require a real SCRATCH_RTDEV

On Wed, Dec 10, 2025 at 06:46:56AM +0100, Christoph Hellwig wrote:
> Require a real SCRATCH_RTDEV instead of faking one up using a loop
> device, as otherwise the options specified in MKFS_OPTIONS might
> not actually work the configuration.
> 
> Signed-off-by: Christoph Hellwig <hch@....de>
> ---
>  tests/xfs/528     | 34 ++++------------------------------
>  tests/xfs/528.out |  2 --
>  2 files changed, 4 insertions(+), 32 deletions(-)
> 
> diff --git a/tests/xfs/528 b/tests/xfs/528
> index a1efbbd27b96..c83545e959dc 100755
> --- a/tests/xfs/528
> +++ b/tests/xfs/528
> @@ -10,27 +10,16 @@
>  . ./common/preamble
>  _begin_fstest auto quick insert zero collapse punch rw realtime
>  
> -# Override the default cleanup function.
> -_cleanup()
> -{
> -	cd /
> -	_scratch_unmount >> $seqres.full 2>&1
> -	[ -n "$rt_loop_dev" ] && _destroy_loop_device $rt_loop_dev
> -	rm -f $tmp.* $TEST_DIR/$seq.rtvol
> -}
> -
> -# Import common functions.
>  . ./common/filter
>  
> -_require_loop
>  _require_command "$FILEFRAG_PROG" filefrag
>  _require_xfs_io_command "fpunch"
>  _require_xfs_io_command "fzero"
>  _require_xfs_io_command "fcollapse"
>  _require_xfs_io_command "finsert"
> -# Note that we don't _require_realtime because we synthesize a rt volume
> -# below.  This also means we cannot run the post-test check.
> -_require_scratch_nocheck
> +
> +_require_realtime
> +_require_scratch
>  
>  log() {
>  	echo "$@" | tee -a $seqres.full
> @@ -63,7 +52,6 @@ test_ops() {
>  	local lunaligned_off=$unaligned_sz
>  
>  	log "Format rtextsize=$rextsize"
> -	_scratch_unmount
>  	_scratch_mkfs -r extsize=$rextsize >> $seqres.full
>  	_try_scratch_mount || \
>  		_notrun "Could not mount rextsize=$rextsize with synthetic rt volume"
> @@ -150,30 +138,16 @@ test_ops() {
>  	check_file $SCRATCH_MNT/lpunch
>  
>  	log "Check everything, rextsize=$rextsize"
> +	_scratch_unmount
>  	_check_scratch_fs

Why does _scratch_unmount move to this part of the loop?  Unmounting the
filesystem means that _check_xfs_filesystem won't run xfs_scrub on it.

(Everything else looks ok.)

--D

>  }
>  
> -echo "Create fake rt volume"
> -$XFS_IO_PROG -f -c "truncate 400m" $TEST_DIR/$seq.rtvol
> -rt_loop_dev=$(_create_loop_device $TEST_DIR/$seq.rtvol)
> -
> -echo "Make sure synth rt volume works"
> -export USE_EXTERNAL=yes
> -export SCRATCH_RTDEV=$rt_loop_dev
> -_scratch_mkfs > $seqres.full
> -_try_scratch_mount || \
> -	_notrun "Could not mount with synthetic rt volume"
> -
>  # power of two
>  test_ops 262144
>  
>  # not a power of two
>  test_ops 327680
>  
> -_scratch_unmount
> -_destroy_loop_device $rt_loop_dev
> -unset rt_loop_dev
> -
>  # success, all done
>  status=0
>  exit
> diff --git a/tests/xfs/528.out b/tests/xfs/528.out
> index 0e081706618c..08de4c28b16c 100644
> --- a/tests/xfs/528.out
> +++ b/tests/xfs/528.out
> @@ -1,6 +1,4 @@
>  QA output created by 528
> -Create fake rt volume
> -Make sure synth rt volume works
>  Format rtextsize=262144
>  Test regular write, rextsize=262144
>  2dce060217cb2293dde96f7fdb3b9232  SCRATCH_MNT/write
> -- 
> 2.47.3
> 
>