diff mbox

[v2] fstests/xfs: xfs_repair should junk empty attribute leaf blocks

Message ID 1492011577-13436-1-git-send-email-zlang@redhat.com (mailing list archive)
State New, archived
Headers show

Commit Message

Zorro Lang April 12, 2017, 3:39 p.m. UTC
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@redhat.com>
---

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

Comments

Eric Sandeen April 12, 2017, 4:09 p.m. UTC | #1
On 4/12/17 10:39 AM, Zorro Lang wrote:
> +# 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"
> +

as mentioned on IRC, I don't think you have any validation here.

If you are running with an xfs_repair that does not find the corruption,
then neither the repair above nor the post-test xfs_repair -n will
find any problems, and the test will incorrectly pass.

-Eric
--
To unsubscribe from this list: send the line "unsubscribe fstests" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Darrick J. Wong April 12, 2017, 4:14 p.m. UTC | #2
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@redhat.com>
> ---
> 
> 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@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe fstests" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Eric Sandeen April 12, 2017, 4:20 p.m. UTC | #3
On 4/12/17 11:14 AM, Darrick J. Wong wrote:
> 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@redhat.com>
>> ---
>>
>> 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

Also, I think it might be a lot easier to just create the fs w/o CRCs,
I don't think we need CRCs to validate the xfs_repair fix.

>> +
>> +_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.

or don't depend on python at all, do something like:

dd if=/dev/zero bs=$maxsize count=1 | attr -s testattr${seq} $SCRATCH_MNT/$seq.attrfile

or /anything/ else that doesn't depend on python, please :) 

>> +	       $SCRATCH_MNT/$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

--
To unsubscribe from this list: send the line "unsubscribe fstests" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Zorro Lang April 12, 2017, 4:24 p.m. UTC | #4
On Wed, Apr 12, 2017 at 11:09:35AM -0500, Eric Sandeen wrote:
> 
> 
> On 4/12/17 10:39 AM, Zorro Lang wrote:
> > +# 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"
> > +
> 
> as mentioned on IRC, I don't think you have any validation here.
> 
> If you are running with an xfs_repair that does not find the corruption,
> then neither the repair above nor the post-test xfs_repair -n will
> find any problems, and the test will incorrectly pass.

Sure, you're right, I shouldn't validate the xfs_repair by xfs_repair.
I'll change that.

Thanks,
Zorro

> 
> -Eric
> --
> To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe fstests" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Zorro Lang April 12, 2017, 4:27 p.m. UTC | #5
On Wed, Apr 12, 2017 at 09:14:34AM -0700, Darrick J. Wong wrote:
> 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@redhat.com>
> > ---
> > 
> > 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.

As you and Eric all don't like this 'python' way, I'll change it :)

> 
> > +	       $SCRATCH_MNT/$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" ?

Anything weird ?

> 
> > +		   -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.

Yes, I'll change it.

Thanks,
Zorro

> 
> --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@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe fstests" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Darrick J. Wong April 12, 2017, 4:55 p.m. UTC | #6
On Thu, Apr 13, 2017 at 12:27:07AM +0800, Zorro Lang wrote:
> On Wed, Apr 12, 2017 at 09:14:34AM -0700, Darrick J. Wong wrote:
> > 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@redhat.com>
> > > ---
> > > 
> > > 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.
> 
> As you and Eric all don't like this 'python' way, I'll change it :)
> 
> > 
> > > +	       $SCRATCH_MNT/$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" ?
> 
> Anything weird ?

Well... if the attr fork should somehow be in btree format then 'ablock 0'
will handle it just fine, whereas directly reading the extents-format
extent array will break.

--D

> > 
> > > +		   -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.
> 
> Yes, I'll change it.
> 
> Thanks,
> Zorro
> 
> > 
> > --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@vger.kernel.org
> > > 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@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe fstests" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Zorro Lang April 12, 2017, 5:13 p.m. UTC | #7
On Wed, Apr 12, 2017 at 09:55:35AM -0700, Darrick J. Wong wrote:
> On Thu, Apr 13, 2017 at 12:27:07AM +0800, Zorro Lang wrote:
> > On Wed, Apr 12, 2017 at 09:14:34AM -0700, Darrick J. Wong wrote:
> > > 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@redhat.com>
> > > > ---
> > > > 
> > > > 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.
> > 
> > As you and Eric all don't like this 'python' way, I'll change it :)
> > 
> > > 
> > > > +	       $SCRATCH_MNT/$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" ?
> > 
> > Anything weird ?
> 
> Well... if the attr fork should somehow be in btree format then 'ablock 0'
> will handle it just fine, whereas directly reading the extents-format
> extent array will break.

Wow, yes, that's really helpful! Thanks so much :)

> 
> --D
> 
> > > 
> > > > +		   -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.
> > 
> > Yes, I'll change it.
> > 
> > Thanks,
> > Zorro
> > 
> > > 
> > > --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@vger.kernel.org
> > > > 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@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe fstests" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

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')"` \
+	       $SCRATCH_MNT/$seq.attrfile
+
+_scratch_unmount
+# manually corrupt the XFS, by set the header count of attr to 0
+_scratch_xfs_db -x -c "inode $inum" \
+		   -c "addr a.bmx[0].startblock" \
+		   -c "$write_cmd hdr.count 0" >> $seqres.full
+
+# 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