From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from cn.fujitsu.com ([59.151.112.132]:31947 "EHLO heian.cn.fujitsu.com" rhost-flags-OK-FAIL-OK-FAIL) by vger.kernel.org with ESMTP id S1752301AbdESCn6 (ORCPT ); Thu, 18 May 2017 22:43:58 -0400 Message-ID: <591E5BEB.5070008@cn.fujitsu.com> Date: Fri, 19 May 2017 10:43:55 +0800 From: Xiao Yang MIME-Version: 1.0 Subject: Re: [PATCH V2] xfs/289: test fragmented multi-fsb readdir References: In-Reply-To: Content-Type: text/plain; charset="UTF-8"; format=flowed Sender: fstests-owner@vger.kernel.org Content-Transfer-Encoding: quoted-printable To: Eric Sandeen , Eryu Guan Cc: fstests , Zorro Lang List-ID: Hi Eric and Eryu I tested it on RHE7.3GA and RHEL7.4Alpha=EF=BC=8C i got ENOSPC error inst= ead of=20 hang. Is it expected behavior? Please see the following detailed message: =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D [root@RHEL7U4Alpha_Intel64 xfstests]# ./check xfs/294 FSTYP -- xfs (non-debug) PLATFORM -- Linux/x86_64 RHEL7U4Alpha_Intel64 3.10.0-657.el7.x86_64 MKFS_OPTIONS -- -f -bsize=3D4096 /dev/sda11 MOUNT_OPTIONS -- -o context=3Dsystem_u:object_r:root_t:s0 /dev/sda11=20 /mnt/xfstests/scratch xfs/294 10s ... - output mismatch (see=20 /var/lib/xfstests/results//xfs/294.out.bad) --- tests/xfs/294.out 2017-05-19 10:39:31.798010275 +0800 +++ /var/lib/xfstests/results//xfs/294.out.bad 2017-05-19=20 10:40:02.112008755 +0800 @@ -1,2 +1,1234 @@ QA output created by 294 -23accd029fad51ec02e4ec1f24799878 - +touch: cannot touch=20 '/mnt/xfstests/scratch/testdir/12345678901234567890169': No space left=20 on device +touch: cannot touch=20 '/mnt/xfstests/scratch/testdir/12345678901234567890170': No space left=20 on device +touch: cannot touch=20 '/mnt/xfstests/scratch/testdir/12345678901234567890171': No space left=20 on device +touch: cannot touch=20 '/mnt/xfstests/scratch/testdir/12345678901234567890172': No space left=20 on device +touch: cannot touch=20 '/mnt/xfstests/scratch/testdir/12345678901234567890173': No space left=20 on device ... (Run 'diff -u tests/xfs/294.out=20 /var/lib/xfstests/results//xfs/294.out.bad' to see the entire diff) Ran: xfs/294 Failures: xfs/294 Failed 1 of 1 tests =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D Thanks, Xiao Yang On 2017/05/04 8:21, Eric Sandeen wrote: > Regression test for kernel commit: > 023cc840 xfs: handle array index overrun in xfs_dir2_leaf_readbuf() > > See commit for detailed problem description. > > tl;dr: readahead on weirdly fragmented multi-block directories > was broken. > > Signed-off-by: Eric Sandeen > --- > > V2: address most review comments, esp. nuking ginormous .out > file, but kept touch vs. echo, because echo allocates > 1 byte -> 1 block and throws off all > the space calibration. Still runs in 28s here. > > 288->289. > > Thanks, > -Eric > > diff --git a/tests/xfs/289 b/tests/xfs/289 > new file mode 100755 > index 0000000..86e3dd6 > --- /dev/null > +++ b/tests/xfs/289 > @@ -0,0 +1,118 @@ > +#! /bin/bash > +# FS QA Test 289 > +# > +# Test readdir on fragmented multi-fsb dir blocks > +# > +# If the readahead map ends with a partial multi-fsb dir > +# block, the loop at the end of xfs_dir2_leaf_readbuf() may > +# walk off the end of the mapping array, read garbage, > +# corrupt the loop control counter, and never return. > +# > +# Failure is a hang; KASAN should also catch this. > +# > +#---------------------------------------------------------------------= -- > +# Copyright (c) 2017 Red Hat, Inc. All Rights Reserved. > +# Author: Eric Sandeen > +# > +# 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=3D`basename $0` > +seqres=3D$RESULT_DIR/$seq > +echo "QA output created by $seq" > + > +here=3D`pwd` > +tmp=3D/tmp/$$ > +status=3D1 # 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 > + > +# real QA test starts here > + > +# Modify as appropriate. > +_supported_fs xfs > +_supported_os Linux > +_require_scratch > +_require_test_program "punch-alternating" > + > +# We want to mkfs with a very specific geometry > +MKFS_OPTIONS=3D"" > +_scratch_mkfs "-d size=3D512m -n size=3D8192 -i size=3D1024">> $seqre= s.full 2>&1 \ > + || _fail "mkfs failed" > +_scratch_mount > + > +# Make a ton of mostly-empty inode clusters so we can always > +# make more inodes > +mkdir $SCRATCH_MNT/tmp > +for I in `seq 1 10000`; do touch $SCRATCH_MNT/tmp/$I; done > + > +# These mostly-empty clusters will live here: > +mkdir $SCRATCH_MNT/clusters > +for I in `seq 1 32 10000`; do > + mv $SCRATCH_MNT/tmp/$I $SCRATCH_MNT/clusters; > +done > +rm -rf $SCRATCH_MNT/tmp > + > +# Make our test dir with a couple blocks, should be contiguous > +mkdir $SCRATCH_MNT/testdir > +# roughly 20 chars per file > +for I in `seq 1 100`; do > + touch $SCRATCH_MNT/testdir/12345678901234567890$I; > +done > + > +# Now completely fragment freespace. > +# Consume most of it: > +$XFS_IO_PROG -f -c "falloc 0 400m" $SCRATCH_MNT/fillfile || > + _fail "Could not allocate space" > + > +# File to fragment: > +$XFS_IO_PROG -f -c "falloc 0 70m" $SCRATCH_MNT/fragfile || > + _fail "Could not allocate space" > + > +df -h $SCRATCH_MNT>> $seqres.full 2>&1 > + > +# Fill remaining space; let this run to failure > +dd if=3D/dev/zero of=3D$SCRATCH_MNT/spacefile1 oflag=3Ddirect>> $seqr= es.full 2>&1 > +# Fragment our all-consuming file > +./src/punch-alternating $SCRATCH_MNT/fragfile>> $seqres.full 2>&1 > + > +# Punching might have freed up large-ish swaths of metadata > +# Consume hopefully any remaining contiguous freespace > +# (and then some for good measure) > +dd conv=3Dfsync if=3D/dev/zero of=3D$SCRATCH_MNT/spacefile2 bs=3D1M co= unt=3D64>> $seqres.full 2>&1 > + > +# Now populate the directory so that it must allocate these > +# fragmented blocks > +for I in `seq 1 1400`; do > + touch $SCRATCH_MNT/testdir/12345678901234567890$I; > +done > + > +# Now traverse that ugly thing! > +find $SCRATCH_MNT/testdir | sort | _filter_scratch | md5sum > + > +status=3D0 > +exit > diff --git a/tests/xfs/289.out b/tests/xfs/289.out > new file mode 100644 > index 0000000..35f8c5e > --- /dev/null > +++ b/tests/xfs/289.out > @@ -0,0 +1,2 @@ > +QA output created by 289 > +23accd029fad51ec02e4ec1f24799878 - > diff --git a/tests/xfs/group b/tests/xfs/group > index 75769f9..1b7046b 100644 > --- a/tests/xfs/group > +++ b/tests/xfs/group > @@ -285,6 +285,7 @@ > 285 dangerous_fuzzers dangerous_scrub > 286 dangerous_fuzzers dangerous_scrub dangerous_online_repair > 287 auto dump quota quick > +289 auto dir metadata dangerous > 290 auto rw prealloc quick ioctl zero > 291 auto repair > 292 auto mkfs quick > > > -- > 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 > > > . >