On Wed, Aug 16, 2017 at 03:02:14PM +0800, Eryu Guan wrote:
> Hi Carlos,
>
> On Tue, Aug 15, 2017 at 02:55:41PM +0200, Carlos Maiolino wrote:
> > Tests the search algorithm for a free inode slot in a specific AG, done
> > in xfs_dialloc_ag_inobt().
> >
> > When finobt is not used, and agi->freecount is not 0, XFS will scan the AG inode
> > tree looking for a free inode slot, but if agi->freecount is corrupted,
> > and there is no free slot at all, it will end up in an infinite loop.
> >
> > This test checks for the infinite loop fix.
>
> The test passed for me with 4.13-rc kernel with latest upstream
> xfsprogs, but I expected it to fail (hang). Does it hang for you?
>
Hmm, yes, it passes to me too, weird, the bug is still there, I'll figure it out
and send a V2
> Some comments inline.
>
> >
> > Signed-off-by: Carlos Maiolino <cmaiolino@xxxxxxxxxx>
> > ---
> > tests/xfs/057 | 80 +++++++++++++++++++++++++++++++++++++++++++++++++++++++
> > tests/xfs/057.out | 2 ++
> > tests/xfs/group | 1 +
> > 3 files changed, 83 insertions(+)
> > create mode 100755 tests/xfs/057
> > create mode 100644 tests/xfs/057.out
> >
> > diff --git a/tests/xfs/057 b/tests/xfs/057
> > new file mode 100755
> > index 00000000..446ea1eb
> > --- /dev/null
> > +++ b/tests/xfs/057
> > @@ -0,0 +1,80 @@
> > +#! /bin/bash
> > +# FS QA Test 057
> > +#
> > +# Check if the filesystem will lockup when trying to allocate a new inode in
> > +# an AG with no free inodes but with a corrupted agi->freecount showing free inodes.
>
> Trailing whitespace in above line :)
>
> > +#
> > +# At the end of the test, the scratch device will purposely be in a corrupted
> > +# state, so there is no need for checking that.
> > +#-----------------------------------------------------------------------
> > +# Copyright (c) 2017 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 /
> > + rm -f $tmp.*
> > +}
> > +
> > +# get standard environment, filters and checks
> > +. ./common/rc
> > +. ./common/filter
> > +
> > +# remove previous $seqres.full before test
> > +rm -f $seqres.full
> > +
> > +# real QA test starts here
> > +
> > +# Modify as appropriate.
> > +_supported_fs generic
> > +_supported_os Linux
> > +#_require_scratch
>
> Can be removed.
>
> > +_require_scratch_nocheck
> > +
> > +_scratch_mkfs_xfs -mfinobt=0 >/dev/null 2>&1 \
> > + || _notrun "Failed to initialize filesystem."
>
> Hmm, I'd use _require_xfs_finobt to _notrun on old kernel/userspace
> where finobt feature is not supported, i.e.
>
> _require_scratch_nocheck
> _require_xfs_finobt
>
> _scratch_mkfs -mfinobt=0 >>$seqres.full 2>&1
>
> > +_scratch_mount >/dev/null 2>&1
>
> This mount should succeed, so I'd drop the I/O redirection, just let the
> error message fail the test if there's a mount failure.
>
> > +
> > +# At the end of filesystem's initialization, AG 0 will still have 61 free
> > +# inodes in the agi->freecount, create 61 extra dummy files to fill it up
> > +for i in `seq 1 61`; do
> > + touch $SCRATCH_MNT/dummy_file$1
> ^^^ $i not $1
> so I know why I can't reproduce the hang now, fixing this typo works for
> me :)
>
> > +done
>
> How about getting the freecount number via xfs_db (before
> _scratch_mount) instead of assmuing it always be 61?
>
> freecount=`_scratch_xfs_db -c "agi 0" -c "p freecount" | cut -d' ' -f 3`
>
> > +
> > +_scratch_unmount
> > +
> > +# agi->freecount is 0 here, corrupt it to show extra free inodes
> > +$XFS_DB_PROG -x -c "agi 0" -c "write -d freecount 10" $SCRATCH_DEV >/dev/null 2>&1
>
> redirect stdout and stderr to $seqres.full for debug purpose
>
> > +
> > +_scratch_mount >/dev/null 2>&1
>
> No need to drop stdout and stderr either.
>
> > +
> > +# Lock up a buggy kernel
> > +touch $SCRATCH_MNT/lockupfile
>
> Would this touch return any error message if the kernel bug is fixed? If
> so, I'd drop stdout/stderr of touch.
>
> > +
> > +# If we reach this point, the filesystem is fixed
> > +echo "Silence is golden"
> > +status=0
> > +exit
> > diff --git a/tests/xfs/057.out b/tests/xfs/057.out
> > new file mode 100644
> > index 00000000..185023c7
> > --- /dev/null
> > +++ b/tests/xfs/057.out
> > @@ -0,0 +1,2 @@
> > +QA output created by 057
> > +Silence is golden
> > diff --git a/tests/xfs/group b/tests/xfs/group
> > index cf876a29..3ca8e75a 100644
> > --- a/tests/xfs/group
> > +++ b/tests/xfs/group
> > @@ -54,6 +54,7 @@
> > 054 auto quick
> > 055 dump ioctl remote tape
> > 056 dump ioctl auto quick
> > +057 dangerous
>
> I think it's a good fit to auto and quick group too, and maybe fuzzers
> group.
>
> Thanks,
> Eryu
>
> > 059 dump ioctl auto quick
> > 060 dump ioctl auto quick
> > 061 dump ioctl auto quick
> > --
> > 2.13.5
> >
> > --
> > To unsubscribe from this list: send the line "unsubscribe fstests" in
> > the body of a message to majordomo@xxxxxxxxxxxxxxx
> > More majordomo info at http://vger.kernel.org/majordomo-info.html
--
Carlos
--
To unsubscribe from this list: send the line "unsubscribe fstests" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
