From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mail.kernel.org ([198.145.29.99]:39160 "EHLO mail.kernel.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754995AbeEHMqa (ORCPT ); Tue, 8 May 2018 08:46:30 -0400 Message-ID: <26f3c1c78d4ee94af5ef51af5c88fa25247efb90.camel@kernel.org> Subject: Re: [PATCH v3] generic: test for seeing unseen fsync errors on newly open files From: Jeff Layton Date: Tue, 08 May 2018 08:46:27 -0400 In-Reply-To: <20180502055051.GA29084@desktop> References: <20180428145937.6804-1-jlayton@kernel.org> <20180428230627.16791-1-jlayton@kernel.org> <20180502055051.GA29084@desktop> Content-Type: text/plain; charset="UTF-8" Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Sender: fstests-owner@vger.kernel.org To: Eryu Guan Cc: fstests@vger.kernel.org, willy@infradead.org, andres@anarazel.de, david@fromorbit.com, amir73il@gmail.com List-ID: On Wed, 2018-05-02 at 13:50 +0800, Eryu Guan wrote: > On Sat, Apr 28, 2018 at 07:06:27PM -0400, Jeff Layton wrote: > > From: Jeff Layton > > > > This adds a regression test for the following kernel patch: > > > > errseq: Always report a writeback error once > > > > This is motivated by some rather odd behavior done by the PostgreSQL > > project. The main database writers will offload the fsync calls to a > > separate process, which can open files after a writeback error has > > already occurred. > > > > This used to work with older kernels that reported the error to only > > one fd, but with the errseq_t changes we lost the ability to see > > errors that occurred before the open. The above patch restores that > > behavior. > > > > Signed-off-by: Jeff Layton > > Reviewed-by: Amir Goldstein > > --- > > tests/generic/999 | 95 +++++++++++++++++++++++++++++++++++++++++++++++++++ > > tests/generic/999.out | 5 +++ > > tests/generic/group | 1 + > > 3 files changed, 101 insertions(+) > > create mode 100755 tests/generic/999 > > create mode 100644 tests/generic/999.out > > > > diff --git a/tests/generic/999 b/tests/generic/999 > > new file mode 100755 > > index 000000000000..b816fd3bb423 > > --- /dev/null > > +++ b/tests/generic/999 > > @@ -0,0 +1,95 @@ > > +#! /bin/bash > > +# FS QA Test No. XXX > > ^^^ Need a test number here :) > > +# > > +# Open a file several times, write to it, fsync on all fds and make sure that > > +# they all return 0. Change the device to start throwing errors. Write again > > +# on all fds and fsync on all fds. Ensure that we get errors on all of them. > > +# Then fsync on all one last time and verify that all return 0. > > +# > > +#----------------------------------------------------------------------- > > +# Copyright (c) 2018, Jeff Layton > > +# > > +# 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 -rf $tmp.* $testdir > > + _dmerror_cleanup > > +} > > + > > +# get standard environment, filters and checks > > +. ./common/rc > > +. ./common/filter > > +. ./common/dmerror > > + > > +# real QA test starts here > > +_supported_os Linux > > +_require_scratch_nocheck > > +# This test uses "dm" without taking into account the data could be on > > +# realtime subvolume, thus the test will fail with rtinherit=1 > > +_require_no_rtinherit > > + > > +_require_dm_target error > > + > > +rm -f $seqres.full > > + > > +echo "Format and mount" > > +_scratch_mkfs > $seqres.full 2>&1 > > +_dmerror_init > > +_dmerror_mount > > + > > +datalen=65536 > > +_require_fs_space $SCRATCH_MNT $datalen > > + > > +# use fd 5 to hold file open > > +testfile=$SCRATCH_MNT/fsync-open-after-err > > +exec 5>$testfile > > + > > +# write some data to file and fsync it out > > +$XFS_IO_PROG -c "pwrite -q 0 $datalen" -c fsync $testfile > > + > > +# flip device to non-working mode > > +_dmerror_load_error_table > > + > > +# rewrite the data, call sync to ensure it's written back w/o scraping error > > +$XFS_IO_PROG -c "pwrite -q 0 $datalen" -c sync $testfile > > + > > +# heal the device error > > +_dmerror_load_working_table > > + > > +# open again and call fsync > > +echo "The following fsync should fail with EIO:" > > +$XFS_IO_PROG -c fsync $testfile > > +echo "done" > > > I built latest Linus tree, which should contain the mentioned fix, and I > saw different results on xfs, ext4 and btrfs. > > XFS fails the test as: > -fsync: Input/output error > +/mnt/scratch/fsync-open-after-err: Input/output error I think I see the problem. I was testing this with a logdev. I think we'll need a _require_logdev directive in the test. I'll send a v4 in a bit with that and some other corrections. > > > While btrfs fails as: > -fsync: Input/output error > +/mnt/scratch/fsync-open-after-err: Read-only file system > I think btrfs will require a specialized testcase (like we did for some of the earlier tests in this area). I'll send a separate patch for that one. > > > And both XFS and btrfs behave in the same way no matter I have the fix > applied or not (tested on v4.16 kernel and v4.17-rc3+ kernel). > > Only ext4 passes the test with v4.17-rc3+ kernel, and fails on v4.16 > kernel as expected: > -fsync: Input/output error > > Could you please take a look? > > Thanks, > Eryu > > > + > > +# close file > > +exec 5>&- > > + > > +# success, all done > > +_dmerror_unmount > > +_dmerror_cleanup > > + > > +status=0 > > +exit > > diff --git a/tests/generic/999.out b/tests/generic/999.out > > new file mode 100644 > > index 000000000000..38d2d7f6495f > > --- /dev/null > > +++ b/tests/generic/999.out > > @@ -0,0 +1,5 @@ > > +QA output created by 999 > > +Format and mount > > +The following fsync should fail with EIO: > > +fsync: Input/output error > > +done > > diff --git a/tests/generic/group b/tests/generic/group > > index ea8e51b35e79..48f491a5c32b 100644 > > --- a/tests/generic/group > > +++ b/tests/generic/group > > @@ -486,3 +486,4 @@ > > 481 auto quick log metadata > > 482 auto metadata replay > > 483 auto quick log metadata > > +999 auto quick > > -- > > 2.14.3 > > > > -- > > 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 -- Jeff Layton