Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751478AbdIMHG7 (ORCPT ); Wed, 13 Sep 2017 03:06:59 -0400 Received: from mx1.redhat.com ([209.132.183.28]:53492 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750999AbdIMHG5 (ORCPT ); Wed, 13 Sep 2017 03:06:57 -0400 DMARC-Filter: OpenDMARC Filter v1.3.2 mx1.redhat.com 13FE58553E Authentication-Results: ext-mx04.extmail.prod.ext.phx2.redhat.com; dmarc=none (p=none dis=none) header.from=redhat.com Authentication-Results: ext-mx04.extmail.prod.ext.phx2.redhat.com; spf=fail smtp.mailfrom=eguan@redhat.com Date: Wed, 13 Sep 2017 15:06:54 +0800 From: Eryu Guan To: Ross Zwisler Cc: fstests@vger.kernel.org, Andrew Morton , linux-kernel@vger.kernel.org, "Darrick J. Wong" , "Theodore Ts'o" , Christoph Hellwig , Dan Williams , Dave Chinner , Jan Kara , linux-ext4@vger.kernel.org, linux-nvdimm@lists.01.org, linux-xfs@vger.kernel.org, Andreas Dilger Subject: Re: [PATCH 2/3] ext4: test for DAX + journaling corruption Message-ID: <20170913070654.GA8034@eguan.usersys.redhat.com> References: <20170912044521.22644-1-ross.zwisler@linux.intel.com> <20170912044521.22644-2-ross.zwisler@linux.intel.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20170912044521.22644-2-ross.zwisler@linux.intel.com> User-Agent: Mutt/1.8.3 (2017-05-23) X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.5.16 (mx1.redhat.com [10.5.110.28]); Wed, 13 Sep 2017 07:06:57 +0000 (UTC) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 8159 Lines: 276 On Mon, Sep 11, 2017 at 10:45:20PM -0600, Ross Zwisler wrote: > Add a regression test for the following kernel commit: > > ext4: prevent data corruption with journaling + DAX > > The test passes if either we successfully compare the data between the mmap > with journaling turned on and the one with journaling turned off, or if we > fail the chattr command to turn on or off journaling. The latter is how we > prevent this issue in the kernel. Yeah, I noticed that mounting ext4 with "-o dax,data=journal" is not allowed, enabling data journaling on a dax mount should be stopped too. > > Signed-off-by: Ross Zwisler > --- > .gitignore | 1 + > src/Makefile | 2 +- > src/t_ext4_dax_journal_corruption.c | 93 +++++++++++++++++++++++++++++++++++++ > tests/ext4/030 | 68 +++++++++++++++++++++++++++ > tests/ext4/030.out | 2 + > tests/ext4/group | 1 + > 6 files changed, 166 insertions(+), 1 deletion(-) > create mode 100644 src/t_ext4_dax_journal_corruption.c > create mode 100755 tests/ext4/030 > create mode 100644 tests/ext4/030.out > > diff --git a/.gitignore b/.gitignore > index 2accc37..4bdc5bf 100644 > --- a/.gitignore > +++ b/.gitignore > @@ -154,6 +154,7 @@ > /src/t_mmap_stale_pmd > /src/t_mmap_cow_race > /src/t_mmap_fallocate > +/src/t_ext4_dax_journal_corruption Better to add new entry in alphabetical order, I know there're already some out-of-order entries there, but this one is not affected and better to stop adding new ones :) > > # dmapi/ binaries > /dmapi/src/common/cmd/read_invis > diff --git a/src/Makefile b/src/Makefile > index b8aff49..e6558e2 100644 > --- a/src/Makefile > +++ b/src/Makefile > @@ -13,7 +13,7 @@ TARGETS = dirstress fill fill2 getpagesize holes lstat64 \ > multi_open_unlink dmiperf unwritten_sync genhashnames t_holes \ > t_mmap_writev t_truncate_cmtime dirhash_collide t_rename_overwrite \ > holetest t_truncate_self t_mmap_dio af_unix t_mmap_stale_pmd \ > - t_mmap_cow_race t_mmap_fallocate fsync-err > + t_mmap_cow_race t_mmap_fallocate fsync-err t_ext4_dax_journal_corruption > > LINUX_TARGETS = xfsctl bstat t_mtab getdevicesize preallo_rw_pattern_reader \ > preallo_rw_pattern_writer ftrunc trunc fs_perms testx looptest \ > diff --git a/src/t_ext4_dax_journal_corruption.c b/src/t_ext4_dax_journal_corruption.c > new file mode 100644 > index 0000000..e0d63f8 > --- /dev/null > +++ b/src/t_ext4_dax_journal_corruption.c > @@ -0,0 +1,93 @@ > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > +#include > + > +#define PAGE(a) ((a)*0x1000) > +#define STRLEN 256 > + > +void err_exit(char *op) > +{ > + fprintf(stderr, "%s: %s\n", op, strerror(errno)); > + exit(1); > +} > + > +void chattr_cmd(char *chattr, char *cmd, char *file) > +{ > + int ret; > + char command[STRLEN]; > + > + ret = snprintf(command, STRLEN, "%s %s %s 2>/dev/null", chattr, cmd, file); > + if (ret < 0) > + err_exit("snprintf"); > + > + ret = system(command); > + if (ret) /* Success - the kernel fix is to have this chattr fail */ > + exit(77); > +} > + > +int main(int argc, char *argv[]) > +{ > + int fd, err, len = PAGE(1); > + char *data, *dax_data, *chattr, *file; > + char string[STRLEN]; > + > + if (argc < 3) { > + printf("Usage: %s \n", basename(argv[0])); > + exit(0); > + } > + > + chattr = argv[1]; > + file = argv[2]; > + > + srand(time(NULL)); > + snprintf(string, STRLEN, "random number %d\n", rand()); > + > + fd = open(file, O_RDWR|O_CREAT, S_IRUSR|S_IWUSR); > + if (fd < 0) > + err_exit("fd"); > + > + /* begin with journaling off and DAX on */ > + chattr_cmd(chattr, "-j", file); > + > + ftruncate(fd, 0); > + fallocate(fd, 0, 0, len); > + > + dax_data = mmap(NULL, len, PROT_READ, MAP_SHARED, fd, 0); > + if (!dax_data) > + err_exit("mmap dax_data"); > + > + /* turns on journaling, and turns off DAX */ > + chattr_cmd(chattr, "+j", file); I'm a bit confused here, just from the test code, it's not obvious to me how DAX is turned off. I looked at the kernel code and there's a comment saying: "Update inode->i_flags after EXT4_INODE_JOURNAL_DATA was updated. E.g. S_DAX may get cleared / set." But isn't the per-inode dax flag proposal rejected? Anyway, some comments to explain how is DAX being turned off would be good here. > + > + data = mmap(NULL, len, PROT_READ|PROT_WRITE, MAP_SHARED, fd, 0); > + if (!data) > + err_exit("mmap data"); > + > + /* > + * Write the data using the non-DAX mapping, and try and read it back > + * using the DAX mapping. > + */ > + strcpy(data, string); > + if (strcmp(dax_data, string) != 0) > + printf("Data miscompare\n"); > + > + err = munmap(data, len); > + if (err < 0) > + err_exit("munmap data"); > + > + err = munmap(dax_data, len); > + if (err < 0) > + err_exit("munmap dax_data"); > + > + err = close(fd); > + if (err < 0) > + err_exit("close"); > + return 0; > +} > diff --git a/tests/ext4/030 b/tests/ext4/030 > new file mode 100755 > index 0000000..3ac4952 > --- /dev/null > +++ b/tests/ext4/030 > @@ -0,0 +1,68 @@ > +#! /bin/bash > +# FS QA Test ext4/030 > +# > +# This is a regression test for kernel patch: > +# ext4: prevent data corruption with journaling + DAX > +# created by Ross Zwisler > +# > +#----------------------------------------------------------------------- > +# Copyright (c) 2017 Intel Corporation. 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 > + > +# remove previous $seqres.full before test > +rm -f $seqres.full > + > +# Modify as appropriate. > +_supported_os Linux > +_supported_fs ext4 > +_require_scratch_dax > +_require_test_program "t_ext4_dax_journal_corruption" _require_command "$CHATTR_PROG" chattr > + > +# real QA test starts here > +_scratch_mkfs > $seqres.full 2>&1 > +_scratch_mount "-o dax,nodelalloc" >> $seqres.full 2>&1 Hmm, why do we need nodelalloc, need some comments too. > + > +src/t_ext4_dax_journal_corruption $CHATTR_PROG $SCRATCH_MNT/testfile Let's use "$here/src/t_...", Christoph pointed this out in another thread so I pay attention to it in review now :) Thanks, Eryu > + > +if [[ $? != 0 && $? != 77 ]]; then > + echo "Test failed, status $?" > + exit 1 > +fi > + > +# success, all done > +echo "Silence is golden" > +status=0 > +exit > diff --git a/tests/ext4/030.out b/tests/ext4/030.out > new file mode 100644 > index 0000000..06a1c8f > --- /dev/null > +++ b/tests/ext4/030.out > @@ -0,0 +1,2 @@ > +QA output created by 030 > +Silence is golden > diff --git a/tests/ext4/group b/tests/ext4/group > index 257bb64..ef768df 100644 > --- a/tests/ext4/group > +++ b/tests/ext4/group > @@ -32,6 +32,7 @@ > 027 auto quick fsmap > 028 auto quick fsmap > 029 auto quick fsmap > +030 auto quick > 271 auto rw quick > 301 aio auto ioctl rw stress defrag > 302 aio auto ioctl rw stress defrag > -- > 2.9.5 >