[PATCH v2] xfstests: add test for btrfs incremental send infinite loop issue
Josef Bacik
jbacik at fb.com
Thu Jan 23 13:17:58 CST 2014
On 01/22/2014 05:05 AM, Filipe David Borba Manana wrote:
> Regression test for btrfs' incremental send feature:
>
> 1) Create several nested directories;
>
> 2) Create a read only snapshot;
>
> 3) Change the parentship of some of the deepest directories in a reverse
> way, so that parents become children and children become parents;
>
> 4) Create another read only snapshot and use it for an incremental send
> relative to the first snapshot.
>
> At step 4 btrfs' send entered an infinite loop, increasing the memory it
> used while building path strings until a krealloc was unable to allocate
> more memory, which caused a warning dump in dmesg.
>
> The following linux kernel patch fixes this issue.
>
> Btrfs: fix infinite path build loops in incremental send
> (https://patchwork.kernel.org/patch/3522361/)
>
> Signed-off-by: Filipe David Borba Manana <fdmanana at gmail.com>
> ---
>
> V2: Updated test to trigger one more code path in the corresponding
> btrfs linux kernel patch that fixes this issue.
>
> tests/btrfs/030 | 144 +++++++++++++++++++++++++++++++++++++++++++++++++++
> tests/btrfs/030.out | 9 ++++
> tests/btrfs/group | 1 +
> 3 files changed, 154 insertions(+)
> create mode 100755 tests/btrfs/030
> create mode 100644 tests/btrfs/030.out
>
> diff --git a/tests/btrfs/030 b/tests/btrfs/030
> new file mode 100755
> index 0000000..5e1b4fc
> --- /dev/null
> +++ b/tests/btrfs/030
> @@ -0,0 +1,144 @@
> +#! /bin/bash
> +# FS QA Test No. btrfs/030
> +#
> +# Regression test for btrfs' incremental send feature:
> +#
> +# 1) Create several nested directories;
> +# 2) Create a read only snapshot;
> +# 3) Change the parentship of some of the deepest directories in a reverse
> +# way, so that parents become children and children become parents;
> +# 4) Create another read only snapshot and use it for an incremental send
> +# relative to the first snapshot.
> +#
> +# At step 4 btrfs' send entered an infinite loop, increasing the memory it
> +# used while building path strings until a krealloc was unable to allocate
> +# more memory, which caused a warning dump in dmesg.
> +#
> +#-----------------------------------------------------------------------
> +# Copyright (c) 2014 Filipe Manana. 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=`mktemp -d`
> +status=1 # failure is the default!
> +trap "_cleanup; exit \$status" 0 1 2 3 15
> +
> +_cleanup()
> +{
> + rm -fr $tmp
> +}
> +
> +# get standard environment, filters and checks
> +. ./common/rc
> +. ./common/filter
> +
> +# real QA test starts here
> +_supported_fs btrfs
> +_supported_os Linux
> +_require_scratch
> +_need_to_be_root
> +
> +FSSUM_PROG=$here/src/fssum
> +[ -x $FSSUM_PROG ] || _notrun "fssum not built"
> +
> +rm -f $seqres.full
> +
> +_scratch_mkfs >/dev/null 2>&1
> +_scratch_mount
> +
> +
> +mkdir -p $SCRATCH_MNT/a/b/c
> +echo "hello" > $SCRATCH_MNT/a/b/c/file.txt
> +mkdir $SCRATCH_MNT/a/b/c/d
> +mkdir $SCRATCH_MNT/a/b/c2
> +mkdir $SCRATCH_MNT/a/b/www
> +mkdir -p $SCRATCH_MNT/a/b/c3/x/y
> +
> +# Directory tree looks like:
> +#
> +# . (ino 256)
> +# |-- a/ (ino 257)
> +# |-- b/ (ino 258)
> +# |-- c/ (ino 259)
> +# | |-- file.txt (ino 260)
> +# | |-- d/ (ino 261)
> +# |
> +# |-- c2/ (ino 262)
> +# |-- www/ (ino 263)
> +# |
> +# |-- c3/ (ino 264)
> +# |-- x/ (ino 265)
> +# |-- y/ (ino 266)
> +
> +$BTRFS_UTIL_PROG subvol snapshot -r $SCRATCH_MNT $SCRATCH_MNT/mysnap1 | \
> + _filter_scratch
> +
> +echo " world" >> $SCRATCH_MNT/a/b/c/file.txt
> +mv $SCRATCH_MNT/a/b/c/d $SCRATCH_MNT/a/b/c2/d2
> +mv $SCRATCH_MNT/a/b/c $SCRATCH_MNT/a/b/c2/d2/cc
> +mv $SCRATCH_MNT/a/b/c3/x/y $SCRATCH_MNT/a/b/c2/y2
> +mv $SCRATCH_MNT/a/b/c3/x $SCRATCH_MNT/a/b/c2/y2/x2
> +mv $SCRATCH_MNT/a/b/c3 $SCRATCH_MNT/a/b/c2/y2/x2/Z
> +mv $SCRATCH_MNT/a/b/www $SCRATCH_MNT/a/b/c2/y2/x2/WWW
> +ln $SCRATCH_MNT/a/b/c2/d2/cc/file.txt $SCRATCH_MNT/a/b/c2/y2/x2/Z/file_link.txt
> +mv $SCRATCH_MNT/a/b/c2/d2/cc/file.txt $SCRATCH_MNT/a/b/c2/y2/x2
> +
> +# Directory tree now looks like:
> +#
> +# . (ino 256)
> +# |-- a/ (ino 257)
> +# |-- b/ (ino 258)
> +# |-- c2/ (ino 262)
> +# |-- d2/ (ino 261)
> +# | |-- cc/ (ino 259)
> +# | |-- file.txt (ino 260)
> +# |-- y2/ (ino 266)
> +# |-- x2/ (ino 265)
> +# |-- WWW/ (ino 263)
> +# |-- Z/ (ino 264)
> +# |-- file_link.txt
> +
> +$BTRFS_UTIL_PROG subvol snapshot -r $SCRATCH_MNT $SCRATCH_MNT/mysnap2 | \
> + _filter_scratch
> +
> +$FSSUM_PROG -A -f -w $tmp/1.fssum $SCRATCH_MNT/mysnap1
> +$FSSUM_PROG -A -f -w $tmp/2.fssum -x $SCRATCH_MNT/mysnap2/mysnap1 \
> + $SCRATCH_MNT/mysnap2
> +
> +$BTRFS_UTIL_PROG send $SCRATCH_MNT/mysnap1 -f $tmp/1.snap 2>&1 | _filter_scratch
> +$BTRFS_UTIL_PROG send -p $SCRATCH_MNT/mysnap1 $SCRATCH_MNT/mysnap2 \
> + -f $tmp/2.snap 2>&1 | _filter_scratch
> +
> +_scratch_unmount
> +_check_btrfs_filesystem $SCRATCH_DEV
> +_scratch_mkfs >/dev/null 2>&1
> +_scratch_mount
> +
> +$BTRFS_UTIL_PROG receive $SCRATCH_MNT -f $tmp/1.snap
> +$FSSUM_PROG -r $tmp/1.fssum $SCRATCH_MNT/mysnap1
> +$BTRFS_UTIL_PROG receive $SCRATCH_MNT -f $tmp/2.snap
> +$FSSUM_PROG -r $tmp/2.fssum $SCRATCH_MNT/mysnap2
> +
You should do || _fail for the fssum stuff, and maybe redirect the
output to $seqres.full.
> +_scratch_unmount
> +_check_btrfs_filesystem $SCRATCH_DEV
> +
> +status=0
> +exit
> diff --git a/tests/btrfs/030.out b/tests/btrfs/030.out
> new file mode 100644
> index 0000000..2999f8c
> --- /dev/null
> +++ b/tests/btrfs/030.out
> @@ -0,0 +1,9 @@
> +QA output created by 030
> +Create a readonly snapshot of 'SCRATCH_MNT' in 'SCRATCH_MNT/mysnap1'
> +Create a readonly snapshot of 'SCRATCH_MNT' in 'SCRATCH_MNT/mysnap2'
> +At subvol SCRATCH_MNT/mysnap1
> +At subvol SCRATCH_MNT/mysnap2
> +At subvol mysnap1
> +OK
> +At snapshot mysnap2
> +OK
So this depends on the output of btrfs send not changing ever, which
historically is not a good assumption to make. It would be better if
you do like the other send tests and just test for failure like the
other send tests do. Thanks,
Josef
More information about the xfs
mailing list