generic: test data corruption on ext4 caused by written/delayed extent

This test excercises the problem with unwritten and delayed extents
in ext4 extent status tree where we might in some cases lose a block
worth of data. Even though this was a ext4 specific problem the
reproducer can be easily tun on any file system so let's do that just
in case.

This tests excercises the problem fixed in kernel with commit
"ext4: Fix data corruption caused by unwritten and delayed extents"

Signed-off-by: Lukas Czerner <lczerner@redhat.com>
---
 tests/generic/082     | 84 +++++++++++++++++++++++++++++++++++++++++++++++++++
 tests/generic/082.out | 14 +++++++++
 2 files changed, 98 insertions(+)
 create mode 100755 tests/generic/082
 create mode 100644 tests/generic/082.out

On Wed, Apr 22, 2015 at 04:55:55PM +0200, Lukas Czerner wrote:
> This test excercises the problem with unwritten and delayed extents
> in ext4 extent status tree where we might in some cases lose a block
> worth of data. Even though this was a ext4 specific problem the
> reproducer can be easily tun on any file system so let's do that just
> in case.
> 
> This tests excercises the problem fixed in kernel with commit
> "ext4: Fix data corruption caused by unwritten and delayed extents"
> 
> Signed-off-by: Lukas Czerner <lczerner@redhat.com>
> ---
>  tests/generic/082     | 84 +++++++++++++++++++++++++++++++++++++++++++++++++++
>  tests/generic/082.out | 14 +++++++++
>  2 files changed, 98 insertions(+)
>  create mode 100755 tests/generic/082
>  create mode 100644 tests/generic/082.out

update to tests/generic/group file is missing

> 
> diff --git a/tests/generic/082 b/tests/generic/082
> new file mode 100755
> index 0000000..2fa22c6
> --- /dev/null
> +++ b/tests/generic/082
> @@ -0,0 +1,84 @@
> +#! /bin/bash
> +# FS QA Test 082
> +#
> +# This test excercises the problem with unwritten and delayed extents
> +# in ext4 extent status tree where we might in some cases lose a block
> +# worth of data. Even though this was a ext4 specific problem the
> +# reproducer can be easily tun on any file system so let's do that just
> +# in case.
> +#
> +# This tests excercises the problem fixed in kernel with commit
> +# "ext4: Fix data corruption caused by unwritten and delayed extents"
> +#
> +#-----------------------------------------------------------------------
> +# Copyright (c) 2015 Red Hat, Inc., Lukas Czerner <lczerner@redhat.com>
> +#
> +# 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
> +
> +# real QA test starts here
> +
> +# Modify as appropriate.
> +_supported_fs generic
> +_supported_os IRIX Linux
> +_require_test

_require_xfs_io_command "falloc" too

I saw failure on cifs because it doesn't support fallocate so the file
size is not extended, and hexdump output differs

> +
> +test_file=${TEST_DIR}/testfile-082

use $seq not hardcoded 082?

> +
> +rm -f $test_file

rm -f $seqres.full too, otherwise the file keeps growing

Thanks,
Eryu
> +
> +# The first write creates a delayed extent, fallocate creates
> +# unwritten extent which will be marked as delayed in ext4
> +# extent status tree. Second write will convert unwritten/delayed
> +# block into written/delayed.
> +$XFS_IO_PROG -f -c "pwrite -S 0xaa 4096 2048" \
> +		-c "falloc 0 131072" \
> +		-c "pwrite -S 0xbb 65536 2048" \
> +		-c "fsync"  $test_file >> $seqres.full 2>&1
> +
> +# Drop the caches to evict dirty buffers from memory
> +echo 3 > /proc/sys/vm/drop_caches
> +
> +# Write into the second part of the block with 0xbb write from before
> +# will create new empty! buffer because the block is still marked as
> +# delayed even though it's already written - resulting in
> +# overwriting previous data.
> +$XFS_IO_PROG -c "pwrite -S 0xdd 67584 2048" $test_file >> $seqres.full 2>&1
> +
> +# On a faulty ext4 oxbb data will be missing, overwritten by zeroes.
> +hexdump -C $test_file
> +
> +# success, all done
> +status=0
> +exit
> diff --git a/tests/generic/082.out b/tests/generic/082.out
> new file mode 100644
> index 0000000..ef28b9a
> --- /dev/null
> +++ b/tests/generic/082.out
> @@ -0,0 +1,14 @@
> +QA output created by 082
> +00000000  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
> +*
> +00001000  aa aa aa aa aa aa aa aa  aa aa aa aa aa aa aa aa  |................|
> +*
> +00001800  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
> +*
> +00010000  bb bb bb bb bb bb bb bb  bb bb bb bb bb bb bb bb  |................|
> +*
> +00010800  dd dd dd dd dd dd dd dd  dd dd dd dd dd dd dd dd  |................|
> +*
> +00011000  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
> +*
> +00020000
> -- 
> 1.8.3.1
> 
> --
> 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
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

On Thu, 23 Apr 2015, Eryu Guan wrote:

> Date: Thu, 23 Apr 2015 11:49:09 +0800
> From: Eryu Guan <eguan@redhat.com>
> To: Lukas Czerner <lczerner@redhat.com>
> Cc: fstests@vger.kernel.org, linux-ext4@vger.kernel.org
> Subject: Re: [PATCH] generic: test data corruption on ext4 caused by
>     written/delayed extent
> 
> On Wed, Apr 22, 2015 at 04:55:55PM +0200, Lukas Czerner wrote:
> > This test excercises the problem with unwritten and delayed extents
> > in ext4 extent status tree where we might in some cases lose a block
> > worth of data. Even though this was a ext4 specific problem the
> > reproducer can be easily tun on any file system so let's do that just
> > in case.
> > 
> > This tests excercises the problem fixed in kernel with commit
> > "ext4: Fix data corruption caused by unwritten and delayed extents"
> > 
> > Signed-off-by: Lukas Czerner <lczerner@redhat.com>
> > ---
> >  tests/generic/082     | 84 +++++++++++++++++++++++++++++++++++++++++++++++++++
> >  tests/generic/082.out | 14 +++++++++
> >  2 files changed, 98 insertions(+)
> >  create mode 100755 tests/generic/082
> >  create mode 100644 tests/generic/082.out
> 
> update to tests/generic/group file is missing
> 
> > 
> > diff --git a/tests/generic/082 b/tests/generic/082
> > new file mode 100755
> > index 0000000..2fa22c6
> > --- /dev/null
> > +++ b/tests/generic/082
> > @@ -0,0 +1,84 @@
> > +#! /bin/bash
> > +# FS QA Test 082
> > +#
> > +# This test excercises the problem with unwritten and delayed extents
> > +# in ext4 extent status tree where we might in some cases lose a block
> > +# worth of data. Even though this was a ext4 specific problem the
> > +# reproducer can be easily tun on any file system so let's do that just
> > +# in case.
> > +#
> > +# This tests excercises the problem fixed in kernel with commit
> > +# "ext4: Fix data corruption caused by unwritten and delayed extents"
> > +#
> > +#-----------------------------------------------------------------------
> > +# Copyright (c) 2015 Red Hat, Inc., Lukas Czerner <lczerner@redhat.com>
> > +#
> > +# 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 $tmp.*
> > +}
> > +
> > +# get standard environment, filters and checks
> > +. ./common/rc
> > +. ./common/filter
> > +
> > +# real QA test starts here
> > +
> > +# Modify as appropriate.
> > +_supported_fs generic
> > +_supported_os IRIX Linux
> > +_require_test
> 
> _require_xfs_io_command "falloc" too
> 
> I saw failure on cifs because it doesn't support fallocate so the file
> size is not extended, and hexdump output differs
> 
> > +
> > +test_file=${TEST_DIR}/testfile-082
> 
> use $seq not hardcoded 082?
> 
> > +
> > +rm -f $test_file
> 
> rm -f $seqres.full too, otherwise the file keeps growing

Will resend.

Thanks!
-Lukas

> 
> Thanks,
> Eryu
> > +
> > +# The first write creates a delayed extent, fallocate creates
> > +# unwritten extent which will be marked as delayed in ext4
> > +# extent status tree. Second write will convert unwritten/delayed
> > +# block into written/delayed.
> > +$XFS_IO_PROG -f -c "pwrite -S 0xaa 4096 2048" \
> > +		-c "falloc 0 131072" \
> > +		-c "pwrite -S 0xbb 65536 2048" \
> > +		-c "fsync"  $test_file >> $seqres.full 2>&1
> > +
> > +# Drop the caches to evict dirty buffers from memory
> > +echo 3 > /proc/sys/vm/drop_caches
> > +
> > +# Write into the second part of the block with 0xbb write from before
> > +# will create new empty! buffer because the block is still marked as
> > +# delayed even though it's already written - resulting in
> > +# overwriting previous data.
> > +$XFS_IO_PROG -c "pwrite -S 0xdd 67584 2048" $test_file >> $seqres.full 2>&1
> > +
> > +# On a faulty ext4 oxbb data will be missing, overwritten by zeroes.
> > +hexdump -C $test_file
> > +
> > +# success, all done
> > +status=0
> > +exit
> > diff --git a/tests/generic/082.out b/tests/generic/082.out
> > new file mode 100644
> > index 0000000..ef28b9a
> > --- /dev/null
> > +++ b/tests/generic/082.out
> > @@ -0,0 +1,14 @@
> > +QA output created by 082
> > +00000000  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
> > +*
> > +00001000  aa aa aa aa aa aa aa aa  aa aa aa aa aa aa aa aa  |................|
> > +*
> > +00001800  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
> > +*
> > +00010000  bb bb bb bb bb bb bb bb  bb bb bb bb bb bb bb bb  |................|
> > +*
> > +00010800  dd dd dd dd dd dd dd dd  dd dd dd dd dd dd dd dd  |................|
> > +*
> > +00011000  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
> > +*
> > +00020000
> > -- 
> > 1.8.3.1
> > 
> > --
> > 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
> 
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html