On Fri, Jul 14, 2017 at 07:55:09PM -0300, Ernesto A. Fernández wrote:
> Check that the group permission bits of a file are not altered when setfacl
> fails. At the time of this patch the test fails for at least ext2, ext4 and
> jfs.
>
> Note that this test is not meaningful for all filesystems, because some
> will still succeed in setting the acls. This does not mean they don't have
> a bug in how they would handle an error.
>
> Signed-off-by: Ernesto A. Fernández <ernesto.mnd.fernandez@xxxxxxxxx>
> ---
> Changes in v3:
> - Several fixes requested by Eryu Guan (thank you for your review), most
> importantly:
> - Place the jfs support in a separate patch, already sent.
> - Make the test fail when the scratch device does not mount, so rootfs
> won't be filled.
> - Add the test to auto and quick groups.
> Changes in v2:
> - Use the scratch dev instead of the test dev so we can make a small fs and
> fill it faster. As suggested by Christoph Hellwig.
> - Make jfs work with _scratch_mkfs_sized, needed for the above.
> - Remove the mention to an issue with reiserfs, since I now believe the
> problem is on their side.
Thanks for the update! But I had a second thought about what to do when
setfacl succeeded, please see below.
>
> tests/generic/447 | 88 +++++++++++++++++++++++++++++++++++++++++++++++++++
> tests/generic/447.out | 2 ++
> tests/generic/group | 1 +
> 3 files changed, 91 insertions(+)
> create mode 100755 tests/generic/447
> create mode 100644 tests/generic/447.out
>
> diff --git a/tests/generic/447 b/tests/generic/447
> new file mode 100755
> index 0000000..c3062c8
> --- /dev/null
> +++ b/tests/generic/447
> @@ -0,0 +1,88 @@
> +#! /bin/bash
> +# FS QA Test 447
> +#
> +# Fill the device and set as many extended attributes to a file as
> +# possible. Then call setfacl on it and, if this fails for lack of
> +# space, test that the permissions remain the same.
> +#
> +#-----------------------------------------------------------------------
> +# Copyright (c) 2017 Ernesto A. Fernandez. All Rights Reserved.
> +#
> +# This program is free software; you can redistribute it and/or
> +# modify it under the terms of the GNU General Public License as
> +# published by the Free Software Foundation.
> +#
> +# This program is distributed in the hope that it would be useful,
> +# but WITHOUT ANY WARRANTY; without even the implied warranty of
> +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> +# GNU General Public License for more details.
> +#
> +# You should have received a copy of the GNU General Public License
> +# along with this program; if not, write the Free Software Foundation,
> +# Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
> +#-----------------------------------------------------------------------
> +#
> +
> +seq=`basename $0`
> +seqres=$RESULT_DIR/$seq
> +echo "QA output created by $seq"
> +
> +here=`pwd`
> +tmp=/tmp/$$
> +status=1 # failure is the default!
> +trap "_cleanup; exit \$status" 0 1 2 3 15
> +
> +_cleanup()
> +{
> + cd /
> + rm -f $tmp.*
> +}
> +
> +# get standard environment, filters and checks
> +. ./common/rc
> +. ./common/filter
> +. ./common/attr
> +
> +# remove previous $seqres.full before test
> +rm -f $seqres.full
> +
> +# real QA test starts here
> +
> +# Modify as appropriate.
> +_supported_fs generic
> +_supported_os Linux
> +_require_scratch
> +_require_test
> +_require_acls
> +_require_attrs
> +
> +_scratch_mkfs_sized $((50 * 1024 * 1024)) >> $seqres.full 2>&1
> +_scratch_mount || _fail "mount failed"
> +
> +TFILE=$SCRATCH_MNT/testfile.$seq
> +
> +# Create the test file and choose its permissions
> +touch $TFILE
> +chmod u+rwx $TFILE
> +chmod go-rwx $TFILE
> +
> +# Try to run out of space so setfacl will fail
> +$XFS_IO_PROG -c "pwrite 0 50m" $TFILE >>$seqres.full 2>&1
> +i=1
> +while setfattr -n user.$i $TFILE &>/dev/null; do
> + ((++i))
> +done
> +
> +if setfacl -m m:r $TFILE &>/dev/null; then
> + # setfacl succeeded, so the test was meaningless
> + # The filesystem might still have an issue
> + status=0
> + echo "-rwx------"
> + exit
I think we should call _notrun here instead of making up the correct
output. If the test is meaningless, don't run it makes more sense to me.
e.g.
if setfacl -m m:r $TFILE &>/dev/null; then
# setfacl succeeded, so the test was meaningless
# The filesystem might still have an issue
_notrun "<notrun reason here>"
fi
Sorry, I didn't bring this up in my first review..
Thanks,
Eryu
> +fi
> +
> +# Since setfacl failed, the permissions should not have changed
> +stat -c %A $TFILE
> +
> +status=0
> +exit
> diff --git a/tests/generic/447.out b/tests/generic/447.out
> new file mode 100644
> index 0000000..adec877
> --- /dev/null
> +++ b/tests/generic/447.out
> @@ -0,0 +1,2 @@
> +QA output created by 447
> +-rwx------
> diff --git a/tests/generic/group b/tests/generic/group
> index 8c1e21a..a3ec390 100644
> --- a/tests/generic/group
> +++ b/tests/generic/group
> @@ -449,3 +449,4 @@
> 444 auto quick acl
> 445 auto quick rw
> 446 auto quick rw dangerous
> +447 auto quick acl
> --
> 2.1.4
>
