From: Eryu Guan Subject: Re: [PATCH] generic: test data corruption on ext4 caused by written/delayed extent Date: Thu, 23 Apr 2015 11:49:09 +0800 Message-ID: <20150423034909.GS17534@dhcp-13-216.nay.redhat.com> References: <1429714555-22133-1-git-send-email-lczerner@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: fstests@vger.kernel.org, linux-ext4@vger.kernel.org To: Lukas Czerner Return-path: Content-Disposition: inline In-Reply-To: <1429714555-22133-1-git-send-email-lczerner@redhat.com> Sender: fstests-owner@vger.kernel.org List-Id: linux-ext4.vger.kernel.org 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 > --- > 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 > +# > +# 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