Re: [PATCH v2] ext4/309: add test for ext4_dir_entry2 wipe

[Leah Rumancik <leah.rumancik@gmail.com>]

On Thu, Jun 03, 2021 at 05:30:18AM +0000, xuyang2018.jy@fujitsu.com wrote:
> on 2021/5/17 22:48, Leah Rumancik wrote:
> > From: Leah Rumancik<lrumancik@google.com>
> > 
> > Check wiping of dir entry data upon removing a file, converting to an
> > htree, and splitting htree nodes.
> > 
> > Tests commit 6c0912739699d8e4b6a87086401bf3ad3c59502d ("ext4: wipe
> > ext4_dir_entry2 upon file deletion").
> > 
> > Signed-off-by: Leah Rumancik<leah.rumancik@gmail.com>
> > 
> > Changes in v2:
> > - fix formatting
> > - use _get_block_size instead of manually finding blocksize
> > - change scratch_dir to testdir to avoid confusion
> > ---
> >   tests/ext4/309     | 191 +++++++++++++++++++++++++++++++++++++++++++++
> >   tests/ext4/309.out |   5 ++
> >   tests/ext4/group   |   1 +
> >   3 files changed, 197 insertions(+)
> >   create mode 100755 tests/ext4/309
> >   create mode 100644 tests/ext4/309.out
> > 
> > diff --git a/tests/ext4/309 b/tests/ext4/309
> > new file mode 100755
> > index 00000000..a4f74e7f
> > --- /dev/null
> > +++ b/tests/ext4/309
> > @@ -0,0 +1,191 @@
> > +#!/bin/bash
> > +# SPDX-License-Identifier: GPL-2.0
> > +# Copyright (c) 2021 Google, Inc. All Rights Reserved.
> > +#
> > +# FS QA Test No. 309
> > +#
> > +# Test wiping of ext4_dir_entry2 data upon file removal, conversion
> > +# to htree, and splitting of htree nodes
> > +#
> > +seq=`basename $0`
> > +seqres=$RESULT_DIR/$seq
> > +echo "QA output created by $seq"
> > +
> > +status=1       # failure is the default!
> > +
> > +# get standard environment, filters and checks
> > +. ./common/rc
> > +. ./common/filter
> > +
> > +# remove previous $seqres.full before test
> > +rm -f $seqres.full
> > +
> > +# real QA test starts here
> > +_supported_fs ext4
> > +
> > +_require_scratch
> > +_require_command "$DEBUGFS_PROG" debugfs
> > +
> > +testdir="${SCRATCH_MNT}/testdir"
> > +
> > +# get block number filename's dir ent
> > +# argument 1: filename
> > +get_block() {
> > +	echo $($DEBUGFS_PROG $SCRATCH_DEV -R "dirsearch /testdir $1" 2>>  $seqres.full | grep -o -m 1 "phys [0-9]\+" | cut -c 6-)
> > +}
> > +
> > +# get offset of filename's dirent within the block
> > +# argument 1: filename
> > +get_offset() {
> > +	echo $($DEBUGFS_PROG $SCRATCH_DEV -R "dirsearch /testdir $1" 2>>  $seqres.full | grep -o -m 1 "offset [0-9]\+" | cut -c 8-)
> > +}
> > +
> > +# get record length of dir ent at specified block and offset
> > +# argument 1: block
> > +# argument 2: offset
> > +get_reclen() {
> > +	echo $(od $SCRATCH_DEV --skip-bytes=$(($1 * $blocksize + $2 + 4)) --read-bytes=2  -d -An  --endian=little | tr -d ' \t\n\r')
> When I test this case on centos7, it will report non-supported --endian
> option for od command because old od doesn't support this option before
> the following patch.
> https://github.com/coreutils/coreutils/commit/b370924c03adaef222859061c61be06fc30c9a3e#diff-1cfd938943be810271354b667b12b6ed6ec85481d3fabb6f85d94193bd201235
> 
> Is this option neccessary?
> 
> Best Regards
> Yang Xu.

I have added a check to skip the test if the od version is before the
endian flag was introduced. Please let me know if you still have issues.
-Leah