[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <aYTQNBYBYWElNuan@nidhogg.toxiclabs.cc>
Date: Thu, 5 Feb 2026 18:15:40 +0100
From: Carlos Maiolino <cem@...nel.org>
To: "Darrick J. Wong" <djwong@...nel.org>
Cc: zlang@...hat.com, linux-ext4@...r.kernel.org, fstests@...r.kernel.org
Subject: Re: [PATCH 2/3] xfs/620: force xattr leaf format for this test
On Mon, Feb 02, 2026 at 11:11:28AM -0800, Darrick J. Wong wrote:
> From: Darrick J. Wong <djwong@...nel.org>
>
> Now that the kernel's parent pointer update code skips the attr intent
> mechanism when the attr fork is in local/shortform format, we need to
> bloat the attr fork to force the slow path for error injection testing.
>
> Signed-off-by: "Darrick J. Wong" <djwong@...nel.org>
Reviewed-by: Carlos Maiolino <cmaiolino@...hat.com>
> ---
> tests/xfs/620 | 7 +++++--
> 1 file changed, 5 insertions(+), 2 deletions(-)
>
>
> diff --git a/tests/xfs/620 b/tests/xfs/620
> index 42a30630f50ac0..47e042b937eb9c 100755
> --- a/tests/xfs/620
> +++ b/tests/xfs/620
> @@ -55,8 +55,11 @@ file5="file5"
>
> echo ""
>
> -# Create files
> -touch $SCRATCH_MNT/$testfolder1/$file4
> +# Create a file with a 53k xattr to force the attr structure out of short
> +# format. Parent pointer operations on a shortform attr structure can skip the
> +# attr intent mechanism and therefore do not trigger the larp knob.
> +truncate -s 53535 $SCRATCH_MNT/$testfolder1/$file4
> +$ATTR_PROG -s x $SCRATCH_MNT/$testfolder1/$file4 < $SCRATCH_MNT/$testfolder1/$file4 &>/dev/null
> _xfs_verify_parent "$testfolder1" "$file4" "$testfolder1/$file4"
>
> # Inject error
>
>
Powered by blists - more mailing lists