From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from relay.sgi.com (relay2.corp.sgi.com [137.38.102.29]) by oss.sgi.com (Postfix) with ESMTP id 57D397F3F for ; Tue, 23 Sep 2014 07:42:30 -0500 (CDT) Received: from cuda.sgi.com (cuda3.sgi.com [192.48.176.15]) by relay2.corp.sgi.com (Postfix) with ESMTP id 37BFD304053 for ; Tue, 23 Sep 2014 05:42:27 -0700 (PDT) Received: from mx1.redhat.com (mx1.redhat.com [209.132.183.28]) by cuda.sgi.com with ESMTP id yGaoQDtUyCU5QuL7 (version=TLSv1 cipher=AES256-SHA bits=256 verify=NO) for ; Tue, 23 Sep 2014 05:42:26 -0700 (PDT) From: Brian Foster Subject: [PATCH v3] generic/032: add xfs unwritten extent data corruption reproducer Date: Tue, 23 Sep 2014 08:42:24 -0400 Message-Id: <1411476144-51074-1-git-send-email-bfoster@redhat.com> List-Id: XFS Filesystem from SGI List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Errors-To: xfs-bounces@oss.sgi.com Sender: xfs-bounces@oss.sgi.com To: fstests@vger.kernel.org Cc: xfs@oss.sgi.com XFS had a data corruption problem where writeback of pages to unwritten extents would fail to run unwritten extent conversion at I/O completion. This causes subsequent reads of written, but unconverted regions to return zeroes. This occurs on sub-page block size filesystems when writeback contends for the inode lock (e.g., with a file writer). Add a test that creates the conditions to reproduce the data corruption and detect it by looking for unwritten extents after all said extents have been overwritten. Signed-off-by: Brian Foster --- v3: - Exit test (_fail) on unwritten extent detection. v2: http://oss.sgi.com/archives/xfs/2014-09/msg00315.html - Converted to generic test. - Use fiemap instead of xfs_bmap. - Added to rw group. - Various fixups: init/clean $tmp, loop syntax, redirect output to $seqres.full, use _scratch_remount. v1: http://oss.sgi.com/archives/xfs/2014-09/msg00296.html tests/generic/032 | 111 ++++++++++++++++++++++++++++++++++++++++++++++++++ tests/generic/032.out | 5 +++ tests/generic/group | 1 + 3 files changed, 117 insertions(+) create mode 100755 tests/generic/032 create mode 100644 tests/generic/032.out diff --git a/tests/generic/032 b/tests/generic/032 new file mode 100755 index 0000000..53fb3de --- /dev/null +++ b/tests/generic/032 @@ -0,0 +1,111 @@ +#! /bin/bash +# FS QA Test No. 032 +# +# This test implements a data corruption scenario on XFS filesystems with +# sub-page sized blocks and unwritten extents. Inode lock contention during +# writeback of pages to unwritten extents leads to failure to convert those +# extents on I/O completion. This causes data corruption as unwritten extents +# are always read back as zeroes. +# +#----------------------------------------------------------------------- +# Copyright (c) 2014 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 / + kill -9 $syncpid > /dev/null 2>&1 + wait + rm -f $tmp.* +} + +# get standard environment, filters and checks +. ./common/rc +. ./common/punch + +# real QA test starts here +rm -f $seqres.full + +_syncloop() +{ + while [ true ]; do + sync + done +} + +# Modify as appropriate. +_supported_fs generic +_supported_os Linux +_require_scratch +_require_xfs_io_command "falloc" +_require_xfs_io_command "fiemap" + +_scratch_mkfs >/dev/null 2>&1 +_scratch_mount + +# run background sync thread +_syncloop & +syncpid=$! + +for iters in $(seq 1 100) +do + rm -f $SCRATCH_MNT/file + + # create a delalloc block in each page of the first 64k of the file + for pgoff in $(seq 0 0x1000 0xf000); do + offset=$((pgoff + 0xc00)) + $XFS_IO_PROG -f \ + -c "pwrite $offset 0x1" \ + $SCRATCH_MNT/file >> $seqres.full 2>&1 + done + + # preallocate the first 64k and overwite, writing past 64k to contend + # with writeback + $XFS_IO_PROG \ + -c "falloc 0 0x10000" \ + -c "pwrite 0 0x100000" \ + -c "fsync" \ + $SCRATCH_MNT/file >> $seqres.full 2>&1 + + # Check for unwritten extents. We should have none since we wrote over + # the entire preallocated region and ran fsync. + $XFS_IO_PROG -c "fiemap -v" $SCRATCH_MNT/file | \ + tee -a $seqres.full | \ + _filter_fiemap | grep unwritten + [ $? == 0 ] && _fail "Unwritten extents found!" +done + +echo $iters iterations + +kill $syncpid +wait + +# clear page cache and dump the file +_scratch_remount +hexdump $SCRATCH_MNT/file + +status=0 +exit diff --git a/tests/generic/032.out b/tests/generic/032.out new file mode 100644 index 0000000..ca5376d --- /dev/null +++ b/tests/generic/032.out @@ -0,0 +1,5 @@ +QA output created by 032 +100 iterations +0000000 cdcd cdcd cdcd cdcd cdcd cdcd cdcd cdcd +* +0100000 diff --git a/tests/generic/group b/tests/generic/group index bdcfd9d..8e0c22a 100644 --- a/tests/generic/group +++ b/tests/generic/group @@ -31,6 +31,7 @@ 026 acl quick auto 027 auto enospc 028 auto quick +032 auto quick rw 053 acl repair auto quick 062 attr udf auto quick 068 other auto freeze dangerous stress -- 1.8.3.1 _______________________________________________ xfs mailing list xfs@oss.sgi.com http://oss.sgi.com/mailman/listinfo/xfs