Re: [PATCH v2] overlay: test memleak in copy-up error handling path

[Eryu Guan <eguan@redhat.com>]

On Tue, May 10, 2016 at 10:32:45AM +0800, Xiong Zhou wrote:
> This memleak leads to panic when unmount the underlying fs.
> Kernel commit ab79efab0 fixed this issue.
> 
> Signed-off-by: Xiong Zhou <xzhou@redhat.com>
> ---
> 
> Diff to v1:
> 	set up underlying fs in $SCRATCH_DEV instead of $SCRATCH_MNT
> 	xfs_io open test_file instead of src/writemod
> 	use little upper dir to trigger ENOSP failure instead of kill
> 	clean up in cleanup

Thanks for the update! And sorry for late review.. Some minor issues
inline

> 
> Thanks Eryu for reviewing!
> 
>  tests/overlay/005     | 112 ++++++++++++++++++++++++++++++++++++++++++++++++++
>  tests/overlay/005.out |   2 +
>  tests/overlay/group   |   1 +
>  3 files changed, 115 insertions(+)
>  create mode 100755 tests/overlay/005
>  create mode 100644 tests/overlay/005.out
> 
> diff --git a/tests/overlay/005 b/tests/overlay/005
> new file mode 100755
> index 0000000..3816bf1
> --- /dev/null
> +++ b/tests/overlay/005
> @@ -0,0 +1,112 @@
> +#! /bin/bash
> +# FS QA Test 005
> +#
> +# Memleak in overlayfs copy-up error handling path leads to panic
> +# when unmount the underlying fs.
> +# Kernel commit ab79efab0 fixed this issue.
> +#
> +#-----------------------------------------------------------------------
> +# Copyright (c) 2016 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()
> +{
> +	cd /
> +	rm -f $tmp.*
> +	_destroy_loop_device $fs_loop_dev
> +	_destroy_loop_device $upper_loop_dev
> +	rm -f $fs_img
> +	rm -f $upper_img
> +}
> +
> +# 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 overlay
> +_supported_os Linux
> +_require_scratch
> +_require_loop
> +

call _scratch_mkfs before creating fs images? It removes all existing
files in $SCRATCH_DEV and make sure no interference from previous test.

> +# setup loop images fs for overlayfs
> +fs_img=${SCRATCH_DEV}/${seq}.$$.img

"lower_img" seems better, we know what it's for from its name.

> +$XFS_IO_PROG -f -c "truncate 120m" $fs_img >>$seqres.full 2>&1
> +fs_loop_dev=`_create_loop_device $fs_img`
> +$MKFS_XFS_PROG -f -n ftype=1 $fs_loop_dev >>$seqres.full 2>&1
> +
> +# only 20m for upper dir
> +upper_img=${SCRATCH_DEV}/$$.${seq}.img
> +$XFS_IO_PROG -f -c "truncate 20m" $upper_img >>$seqres.full 2>&1
> +upper_loop_dev=`_create_loop_device $upper_img`
> +$MKFS_XFS_PROG -f -n ftype=1 $upper_loop_dev >>$seqres.full 2>&1
> +
> +# mount underlying xfs
> +mkdir -p ${SCRATCH_DEV}/xfsmnt

So "lowermnt" seems better here.

> +mkdir -p ${SCRATCH_DEV}/uppermnt
> +$MOUNT_PROG $fs_loop_dev ${SCRATCH_DEV}/xfsmnt
> +$MOUNT_PROG $upper_loop_dev ${SCRATCH_DEV}/uppermnt
> +
> +# prepare dirs
> +mkdir -p ${SCRATCH_DEV}/xfsmnt/lower
> +mkdir -p ${SCRATCH_DEV}/uppermnt/upper
> +mkdir -p ${SCRATCH_DEV}/uppermnt/work
> +
> +lowerd=${SCRATCH_DEV}/xfsmnt/lower
> +upperd=${SCRATCH_DEV}/uppermnt/upper
> +workd=${SCRATCH_DEV}/uppermnt/work
> +
> +# Create test file in lower dir, with too big a size for
> +# upper dir to copy up.
> +$XFS_IO_PROG -f -c "truncate 48m" ${lowerd}/test_file \
> +	>>$seqres.full 2>&1
> +
> +# mount new overlayfs
> +$MOUNT_PROG -t overlay $SELINUX_MOUNT_OPTIONS \
> +	    -o lowerdir=$lowerd -o upperdir=$upperd \
> +	    -o workdir=$workd \
> +	    $SCRATCH_DEV $SCRATCH_MNT
> +
> +# the open call triggers copy-up and it will fail ENOSP
                                                     ^^^^^ ENOSPC :)

> +$XFS_IO_PROG -f -c "o" ${SCRATCH_MNT}/test_file \
> +	>>$seqres.full 2>&1
> +
> +# unmount overlayfs
> +$UMOUNT_PROG $SCRATCH_MNT
> +
> +# unmount undelying xfs, this tiggers panic if memleak happens
> +$UMOUNT_PROG ${SCRATCH_DEV}/uppermnt
> +$UMOUNT_PROG ${SCRATCH_DEV}/xfsmnt
> +
> +# success, all done
> +echo "Silence is golden"
> +status=0
> +exit
> diff --git a/tests/overlay/005.out b/tests/overlay/005.out
> new file mode 100644
> index 0000000..a5027f1
> --- /dev/null
> +++ b/tests/overlay/005.out
> @@ -0,0 +1,2 @@
> +QA output created by 005
> +Silence is golden
> diff --git a/tests/overlay/group b/tests/overlay/group
> index fcf6d63..bc2474e 100644
> --- a/tests/overlay/group
> +++ b/tests/overlay/group
> @@ -7,3 +7,4 @@
>  002 auto quick metadata
>  003 auto quick whiteout
>  004 attr auto copyup quick
> +005 auto copyup dangerous quick

This bug has been fixed, 'dangerous' group can be removed.

Thanks,
Eryu