Re: [PATCH] xfs/288: test xfs_growfs to ensure rejection if target isn't mounted

["Bill O'Donnell" <billodo@redhat.com>]

On Mon, Apr 24, 2017 at 02:54:59PM -0500, Eric Sandeen wrote:
> On 4/24/17 10:09 AM, Bill O'Donnell wrote:
> > The manpage for xfs_growfs specifies that the file/dir to be
> > grown must be mounted. This is a test to ensure that if the
> > target isn't mounted, the command is rejected. The test covers
> > absolute paths and relative paths, including symbolic links
> > and bind mounts.
> > 
> > Signed-off-by: Bill O'Donnell <billodo@redhat.com>
> > ---
> >  tests/xfs/288     | 121 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
> >  tests/xfs/288.out | 119 +++++++++++++++++++++++++++++++++++++++++++++++++++++
> >  tests/xfs/group   |   1 +
> >  3 files changed, 241 insertions(+)
> >  create mode 100755 tests/xfs/288
> >  create mode 100644 tests/xfs/288.out
> > 
> > diff --git a/tests/xfs/288 b/tests/xfs/288
> > new file mode 100755
> > index 0000000..628f753
> > --- /dev/null
> > +++ b/tests/xfs/288
> > @@ -0,0 +1,121 @@
> > +#! /bin/bash
> > +# FS QA Test 288
> > +#
> > +# Test to ensure xfs_growfs command rejects non-existent mount points
> > +#
> > +#-----------------------------------------------------------------------
> > +# Copyright (c) 2017 Red Hat, Inc.  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/$$
> > +status=1	# failure is the default!
> > +trap "_cleanup; exit \$status" 0 1 2 3 15
> > +
> > +_cleanup()
> > +{
> > +    umount $tmpdir
> > +    umount $tmpbind
> > +    rmdir $tmpdir
> > +    rm -f $tmpsymlink
> > +    rmdir $tmpbind
> > +    rm -f $tmp
> > +    rm -f $tmpfile
> > +}
> > +
> > +# get standard environment, filters and checks
> > +. ./common/rc
> > +. ./common/filter
> > +
> > +# remove previous $seqres.full before test
> > +rm -f $seqres.full
> > +
> > +# real QA test starts here
> > +
> > +# Modify as appropriate.
> > +_supported_fs xfs
> > +_supported_os Linux
> > +_require_test
> > +_require_loop
> > +
> > +tmpfile=$TEST_DIR/fsfile
> > +tmpdir=$TEST_DIR/tmpdir
> > +tmpsymlink=$TEST_DIR/tmpsymlink.$$
> > +tmpbind=$TEST_DIR/tmpbind.$$
> > +
> > +mkdir -p $tmpdir || _fail "!!! failed to create temp mount dir"
> > +
> > +echo "=== mkfs.xfs ==="
> > +mkfs.xfs -d file,name=$tmpfile,size=64m -f
> > +
> > +echo "=== truncate ==="
> > +truncate --size=128m $tmpfile
> > +
> > +echo "=== xfs_growfs - unmounted, command should be rejected ==="
> > +xfs_growfs $tmpdir
> > +
> > +echo "=== xfs_growfs - check relative path, unmounted ==="
> > +cd $TEST_DIR
> > +xfs_growfs ./tmpdir
> > +
> > +echo "=== xfs_growfs - no path, unmounted ==="
> > +cd $TEST_DIR
> 
> Is there a reason to keep cd'ing to the same dir?  Here and below
> as well.

Nah, just a cut and paste oversight.

> 
> > +xfs_growfs tmpdir
> > +
> > +echo "=== mount ==="
> > +mount -o loop $tmpfile $tmpdir || _fail "!!! failed to loopback mount"
> > +
> > +echo "=== xfs_growfs - mounted ==="
> > +xfs_growfs $tmpdir
> > +
> > +echo "=== xfs_growfs - check relative path ==="
> > +cd $TEST_DIR
> > +xfs_growfs ./tmpdir
> > +
> > +echo "=== xfs_growfs - no path ==="
> > +cd $TEST_DIR
> > +xfs_growfs tmpdir
> > +
> > +echo "=== test symbolic link ==="
> > +ln -s $tmpdir $tmpsymlink
> > +xfs_growfs $tmpsymlink
> > +
> > +echo "=== test symbolic link using relative path ==="
> > +xfs_growfs ./tmpsymlink.$$
> > +
> > +echo "=== test symbolic link using no path ==="
> > +xfs_growfs tmpsymlink.$$
> > +
> > +echo "=== test bind mount ==="
> > +mkdir $tmpbind
> > +mount -o bind $tmpdir $tmpbind
> > +xfs_growfs $tmpbind
> > +
> > +echo "=== test bind mount - relative path ==="
> > +
> > +xfs_growfs ./tmpbind.$$
> > +
> > +echo "=== test bind mount - no path ==="
> > +xfs_growfs tmpbind.$$
> 
> Testing a plain file as a target might be good too.  The implementation
> shouldn't care, but one mistake I've seen is to do "xfs_info imagefile.img"
> and they get an answer about the host fs, not the image file.  It'd
> be good to ensure that an argument of a regular file is also rejected.
> 
> You could just point xfs_growfs at the image file you create, before
> doing any of the other mount tests.

check.

> 
> > +
> > +# success, all done
> > +status=0
> > +exit
> > diff --git a/tests/xfs/288.out b/tests/xfs/288.out
> > new file mode 100644
> > index 0000000..a526048
> > --- /dev/null
> > +++ b/tests/xfs/288.out
> > @@ -0,0 +1,119 @@
> > +QA output created by 288
> > +=== mkfs.xfs ===
> > +meta-data=/mnt/test/fsfile       isize=512    agcount=4, agsize=4096 blks
> > +         =                       sectsz=512   attr=2, projid32bit=1
> > +         =                       crc=1        finobt=1, sparse=0, rmapbt=0, reflink=0
> > +data     =                       bsize=4096   blocks=16384, imaxpct=25
> > +         =                       sunit=0      swidth=0 blks
> > +naming   =version 2              bsize=4096   ascii-ci=0 ftype=1
> > +log      =internal log           bsize=4096   blocks=855, version=2
> > +         =                       sectsz=512   sunit=0 blks, lazy-count=1
> > +realtime =none                   extsz=4096   blocks=0, rtextents=0
> > +=== truncate ===
> > +=== xfs_growfs - unmounted, command should be rejected ===
> > +xfs_growfs: /mnt/test/tmpdir is not a mounted XFS filesystem
> > +=== xfs_growfs - check relative path, unmounted ===
> > +xfs_growfs: ./tmpdir is not a mounted XFS filesystem
> > +=== xfs_growfs - no path, unmounted ===
> > +xfs_growfs: tmpdir is not a mounted XFS filesystem
> > +=== mount ===
> > +=== xfs_growfs - mounted ===
> > +meta-data=/dev/loop0             isize=512    agcount=4, agsize=4096 blks
> 
> Hm, when I ran this I had a loop0 already set up, so got loop1, and this
> failed to match.  That may be another argument against actually checking
> the output too carefully.  

Agreed.

> 
> > +         =                       sectsz=512   attr=2, projid32bit=1
> > +         =                       crc=1        finobt=1 spinodes=0 rmapbt=0
> > +         =                       reflink=0
> > +data     =                       bsize=4096   blocks=16384, imaxpct=25
> > +         =                       sunit=0      swidth=0 blks
> > +naming   =version 2              bsize=4096   ascii-ci=0 ftype=1
> > +log      =internal               bsize=4096   blocks=855, version=2
> > +         =                       sectsz=512   sunit=0 blks, lazy-count=1
> > +realtime =none                   extsz=4096   blocks=0, rtextents=0
> > +data blocks changed from 16384 to 32768
> > +=== xfs_growfs - check relative path ===
> > +meta-data=/dev/loop0             isize=512    agcount=8, agsize=4096 blks
> > +         =                       sectsz=512   attr=2, projid32bit=1
> > +         =                       crc=1        finobt=1 spinodes=0 rmapbt=0
> > +         =                       reflink=0
> > +data     =                       bsize=4096   blocks=32768, imaxpct=25
> > +         =                       sunit=0      swidth=0 blks
> > +naming   =version 2              bsize=4096   ascii-ci=0 ftype=1
> > +log      =internal               bsize=4096   blocks=855, version=2
> > +         =                       sectsz=512   sunit=0 blks, lazy-count=1
> > +realtime =none                   extsz=4096   blocks=0, rtextents=0
> 
> yeah, so as Darrick mentioned I'd make sure that it either grows each
> time, or that you don't care if it grows at all, just for consistency.
> 
> A little odd to have only the first one grow, but have all the subsequent
> ones still "succeed."

Agreed.

> 
> You'll need to filter all this output somehow - if anyone runs with different
> options in MKFS_OPTS, much of this output may fail to match - but things like
> fs features, block size, inode size, etc are not relevant to the test.
> 
> I think that _filter_mkfs might actually filter growfs just fine, not sure.
> 
> Thanks,
> -Eric
>