From: Eryu Guan Subject: Re: [xfstests PATCH v2] generic: add test of file mode when setfacl fails Date: Fri, 14 Jul 2017 15:25:29 +0800 Message-ID: <20170714072529.GM2478@eguan.usersys.redhat.com> References: <20170714065925.GA6075@debian.home> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 8bit Cc: fstests@vger.kernel.org, Christoph Hellwig , linux-ext4@vger.kernel.org, jfs-discussion@lists.sourceforge.net, reiserfs-devel@vger.kernel.org To: Ernesto =?utf-8?Q?A=2E_Fern=C3=A1ndez?= Return-path: Content-Disposition: inline In-Reply-To: <20170714065925.GA6075@debian.home> Sender: reiserfs-devel-owner@vger.kernel.org List-Id: linux-ext4.vger.kernel.org On Fri, Jul 14, 2017 at 03:59:28AM -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. > > Also add support for jfs in _scratch_mkfs_sized, which is needed here. > > 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 > --- > 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 updated patch! Some minor comments inline. > > common/rc | 5 ++- > tests/generic/447 | 91 +++++++++++++++++++++++++++++++++++++++++++++++++++ > tests/generic/447.out | 2 ++ > tests/generic/group | 1 + > 4 files changed, 98 insertions(+), 1 deletion(-) > create mode 100755 tests/generic/447 > create mode 100644 tests/generic/447.out > > diff --git a/common/rc b/common/rc > index b505365..a2258be 100644 > --- a/common/rc > +++ b/common/rc > @@ -1020,7 +1020,7 @@ _scratch_mkfs_sized() > xfs) > def_blksz=`echo $MKFS_OPTIONS|sed -rn 's/.*-b ?size= ?+([0-9]+).*/\1/p'` > ;; > - ext2|ext3|ext4|ext4dev|udf|btrfs|reiser4|ocfs2) > + ext2|ext3|ext4|ext4dev|udf|btrfs|reiser4|ocfs2|jfs) > def_blksz=`echo $MKFS_OPTIONS| sed -rn 's/.*-b ?+([0-9]+).*/\1/p'` > ;; > esac > @@ -1068,6 +1068,9 @@ _scratch_mkfs_sized() > (( fssize <= 100 * 1024 * 1024 )) && mixed_opt='--mixed' > $MKFS_BTRFS_PROG $MKFS_OPTIONS $mixed_opt -b $fssize $SCRATCH_DEV > ;; > + jfs) > + ${MKFS_PROG}.$FSTYP -F $MKFS_OPTIONS -b $blocksize $SCRATCH_DEV $blocks > + ;; > reiser4) > # mkfs.resier4 requires size in KB as input for creating filesystem > $MKFS_REISER4_PROG $MKFS_OPTIONS -y -b $blocksize $SCRATCH_DEV \ Could you please add jfs support to _scratch_mkfs_sized in a separate patch? > diff --git a/tests/generic/447 b/tests/generic/447 > new file mode 100755 > index 0000000..ff81355 > --- /dev/null > +++ b/tests/generic/447 > @@ -0,0 +1,91 @@ > +#! /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 / > + _scratch_unmount This is not needed, SCRATCH_DEV will be unmounted by next test that calls _require_scratch. > + 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 Because this test is trying to consume all space on $SCRATCH_MNT, it'd be good to exit if _scratch_mount fails, otherwise it's filling rootfs. _scratch_mount || _fail "mount failed" > + > +cd $SCRATCH_MNT > +TFILE=testfile.$seq Hmm, fstests prefer: TFILE=$SCRATCH_MNT/testfile.$seq touch $TFILE chmod u_rwx $TFILE ... instead of cd to $SCRATCH_MNT. > + > +# Create the test file and choose its permissions > +rm -f $TFILE $SCRATCH_DEV was just created, no need to remove $TFILE here. > +touch $TFILE > +chmod u+rwx $TFILE > +chmod go-rwx $TFILE > + > +# Try to run out of space so setfacl will fail > +yes xxxxxxxxxx &> $TFILE How about just writing 50M to $TFILE? $XFS_IO_PROG -fc "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 > +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..f11d798 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 acl Should be in auto and quick group too. Basically auto group means this test is a well defined regression test and the run time is less than 300s, quick group is a subset of auto, and the run time is usually within 30s. Thanks, Eryu