On Wed, Apr 12, 2017 at 11:39:37PM +0800, Zorro Lang wrote:
> There was a bug during log replay, the attr/attr3 leaf verifier
> reported corruption when encountering a leaf attribute with a
> count of 0 in the header, as below:
>
> Metadata corruption detected at xfs_attr3_leaf block 0x480988/0x1000
>
> commit f714016 from xfsprogs has fixed this bug. This test case
> will emulate this corruption by xfs_db and use xfs_repair to fix
> it.
>
> Signed-off-by: Zorro Lang <zlang@xxxxxxxxxx>
> ---
>
> Hi,
>
> V2 did below changes:
> 1) remove the requirement for xfs_db's write and addr commands
> 2) check if xfs_db support write -d option
> 3) turn to get filesystem block size from _filter_mkfs output
> 4) add more comments
>
> Thanks,
> Zorro
>
> tests/xfs/288 | 110 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
> tests/xfs/288.out | 2 +
> tests/xfs/group | 1 +
> 3 files changed, 113 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..81e7b50
> --- /dev/null
> +++ b/tests/xfs/288
> @@ -0,0 +1,110 @@
> +#! /bin/bash
> +# FS QA Test 288
> +#
> +# When an attribute leaf block count is 0, xfs_repair should junk
> +# that leaf directly (as xfsprogs commit f714016).
> +#
> +#-----------------------------------------------------------------------
> +# 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()
> +{
> + cd /
> + rm -f $tmp.*
> +}
> +
> +# get standard environment, filters and checks
> +. ./common/rc
> +. ./common/filter
> +. ./common/attr
> +
> +# remove previous $seqres.full before test
> +rm -f $seqres.full
> +
> +# V4 XFS attr type is 'attr'
> +write_cmd="write"
> +
> +# Modify as appropriate.
> +_supported_fs xfs
> +_supported_os Linux
> +_require_scratch
> +_require_attrs
> +
> +_scratch_mkfs_xfs | _filter_mkfs 2>$tmp.mkfs >/dev/null
> +. $tmp.mkfs
> +
> +# If test on XFS with CRC enabled, after change attr's hdr.count we need
> +# to recalculate crc by use write -d option
> +if [ $_fs_has_crcs -eq 1 ]; then
> + # check if xfs_db support write -d option
> + _scratch_xfs_db -x -c "sb" -c "write -d" 2>&1 | \
> + grep -q "invalid option"
> + if [ $? -eq 0 ];then
> + _notrun "xfs_db doesn't support write -d option"
> + fi
> + # This's a trick to check if xfs_db support recalculating
> + # CRCs of attr3
> + _scratch_xfs_db -x -c "daddr 0" \
> + -c "type attr3" \
> + -c "write -d" 2>/dev/null | \
> + grep -q "Cannot recalculate CRCs on this type of object"
> + if [ $? -eq 0 ];then
> + _notrun "xfs_db can't recalculate CRCs of attr3"
> + fi
> + # V5 XFS attr type is 'attr3'
> + write_cmd="write -d"
> +fi
> +
> +_scratch_mount
> +
> +touch $SCRATCH_MNT/$seq.attrfile
> +inum=$(stat -c '%i' $SCRATCH_MNT/$seq.attrfile)
> +
> +# To get an attr block leaf, we need to extend attr format to extent
> +# format at least, and the max inode size is half of filesystem
> +# block size, so write half of block size attr to make sure attr
> +# out of local format.
> +maxisize=$((dbsize/2))
> +$SETFATTR_PROG -n "user.testattr${seq}" \
> + -v `python -c "print ($maxisize * 'a')"` \
Need to check for the existence of python, or use xfs_io to write
$maxisize bytes to a file and backtick-cat it into the command line.
> + $SCRATCH_MNT/$seq.attrfile
> +
> +_scratch_unmount
> +# manually corrupt the XFS, by set the header count of attr to 0
> +_scratch_xfs_db -x -c "inode $inum" \
> + -c "addr a.bmx[0].startblock" \
"ablock 0" ?
> + -c "$write_cmd hdr.count 0" >> $seqres.full
> +
Uh... we need an xfs_repair -n run here to validate that repair actually
finds a broken xattr block. Only then should we fix the fs. Remember
that xfs_repair only tells us if it thinks the fs is still broken.
--D
> +# This repair should junk above leaf attribute and fix this XFS, then
> +# SCRATCH_DEV shouldn't be corrupted after this case done.
> +_repair_scratch_fs >> $seqres.full
> +
> +echo "Silence is golden"
> +
> +# success, all done
> +status=0
> +exit
> diff --git a/tests/xfs/288.out b/tests/xfs/288.out
> new file mode 100644
> index 0000000..2958a5c
> --- /dev/null
> +++ b/tests/xfs/288.out
> @@ -0,0 +1,2 @@
> +QA output created by 288
> +Silence is golden
> diff --git a/tests/xfs/group b/tests/xfs/group
> index 75769f9..f27a7b6 100644
> --- a/tests/xfs/group
> +++ b/tests/xfs/group
> @@ -285,6 +285,7 @@
> 285 dangerous_fuzzers dangerous_scrub
> 286 dangerous_fuzzers dangerous_scrub dangerous_online_repair
> 287 auto dump quota quick
> +288 auto quick repair fuzzers
> 290 auto rw prealloc quick ioctl zero
> 291 auto repair
> 292 auto mkfs quick
> --
> 2.7.4
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
