Re: [PATCH 5/5] xfs: test that we don't leak inodes and dquots during failed cow recovery

["Darrick J. Wong" <darrick.wong@oracle.com>]

On Wed, Oct 25, 2017 at 07:48:17PM +0800, Eryu Guan wrote:
> On Wed, Oct 18, 2017 at 04:38:01PM -0700, Darrick J. Wong wrote:
> > From: Darrick J. Wong <darrick.wong@oracle.com>
> > 
> > Add a couple of tests to check that we don't leak inodes or dquots
> > if CoW recovery fails and therefore the mount fails.
> > 
> > Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> > ---
> >  common/rc         |    1 
> >  tests/xfs/703     |  112 +++++++++++++++++++++++++++++++++++++++++++++++++++++
> >  tests/xfs/703.out |    9 ++++
> >  tests/xfs/704     |   90 +++++++++++++++++++++++++++++++++++++++++++
> >  tests/xfs/704.out |    5 ++
> >  tests/xfs/705     |  106 ++++++++++++++++++++++++++++++++++++++++++++++++++
> >  tests/xfs/705.out |    9 ++++
> >  tests/xfs/group   |    3 +
> >  8 files changed, 335 insertions(+)
> >  create mode 100755 tests/xfs/703
> >  create mode 100644 tests/xfs/703.out
> >  create mode 100755 tests/xfs/704
> >  create mode 100644 tests/xfs/704.out
> >  create mode 100755 tests/xfs/705
> >  create mode 100644 tests/xfs/705.out
> > 
> > 
> > diff --git a/common/rc b/common/rc
> > index 83aaced..fe68d67 100644
> > --- a/common/rc
> > +++ b/common/rc
> > @@ -3324,6 +3324,7 @@ _check_dmesg()
> >  	     -e "(INFO|ERR): suspicious RCU usage" \
> >  	     -e "INFO: possible circular locking dependency detected" \
> >  	     -e "general protection fault:" \
> > +	     -e "BUG .* remaining" \
> >  	     $seqres.dmesg
> >  	if [ $? -eq 0 ]; then
> >  		_dump_err "_check_dmesg: something found in dmesg (see $seqres.dmesg)"
> > diff --git a/tests/xfs/703 b/tests/xfs/703
> > new file mode 100755
> > index 0000000..f29b843
> > --- /dev/null
> > +++ b/tests/xfs/703
> > @@ -0,0 +1,112 @@
> > +#! /bin/bash
> > +# FS QA Test No. 703
> > +#
> > +# Ensure that we don't leak quota inodes when CoW recovery fails.
> > +#
> > +# Use xfs_fsr to inject bmap redo items in the log for a linked file and
> > +# an unlinked file; enable quota so that we always mount with the quota
> > +# inodes; and then corrupt the refcount btree to ensure that the CoW
> > +# garbage collection (and therefore the mount) fail.
> > +#
> > +# On a subsequent mount attempt, we should be able to replay the bmap
> > +# items for the linked and unlinked files without prematurely truncating
> > +# the unlinked inode and without leaking the linked inode, and we should
> > +# be able to release the quota inodes when we're aborting the mount.  We
> > +# also should not leak dquots.
> > +#
> > +#-----------------------------------------------------------------------
> > +# Copyright (c) 2017, Oracle and/or its affiliates.  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 -rf "$tmp".*
> > +}
> > +
> > +# get standard environment, filters and checks
> > +. ./common/rc
> > +. ./common/filter
> > +. ./common/attr
> > +. ./common/reflink
> > +. ./common/inject
> > +. ./common/quota
> > +
> > +# real QA test starts here
> > +_supported_os Linux
> > +_supported_fs xfs
> > +_require_quota
> > +_require_scratch_reflink
> > +_require_cp_reflink
> > +_require_command "$XFS_FSR_PROG" "xfs_fsr"
> > +_require_xfs_io_error_injection "bmap_finish_one"
> > +_require_xfs_scratch_rmapbt
> > +
> > +rm -f "$seqres.full"
> > +
> > +echo "Format and mount"
> > +_scratch_mkfs > "$seqres.full" 2>&1
> > +_scratch_mount -o noquota >> "$seqres.full" 2>&1
> > +
> > +testdir="$SCRATCH_MNT/test-$seq"
> > +blksz=65536
> > +blks=3
> > +mkdir "$testdir"
> > +
> > +echo "Create a many-block file"
> > +_pwrite_byte 0x62 0 $((blksz * blks)) $testdir/file1 >> $seqres.full
> > +_pwrite_byte 0x63 0 $blksz $testdir/file2 >> $seqres.full
> > +_reflink_range $testdir/file2 0 $testdir/file1 $blksz $blksz >> $seqres.full
> > +_scratch_cycle_mount noquota
> > +
> > +echo "Inject error"
> > +_scratch_inject_error "bmap_finish_one"
> > +
> > +echo "Defrag the file"
> > +$XFS_FSR_PROG -v -d $testdir/file1 >> $seqres.full 2>&1
> > +
> > +echo "FS should be shut down, touch will fail"
> > +touch $SCRATCH_MNT/badfs 2>&1 | _filter_scratch
> > +
> > +echo "Remount to replay log" | tee /dev/ttyprintk
> > +_scratch_unmount
> > +_scratch_dump_log >> $seqres.full
> > +_scratch_xfs_db -x -c 'agf 0' -c 'addr refcntroot' -c 'fuzz -d recs[1].startblock ones' >> $seqres.full
> > +_scratch_xfs_db -x -c 'agf 0' -c 'addr refcntroot' -c p >> $seqres.full
> > +
> > +# Suddenly enable quota to test if we can leak the quotacheck dquots!
> > +_scratch_mount -o quota >> $seqres.full 2>&1
> > +_scratch_unmount 2> /dev/null
> > +rm -f ${RESULT_DIR}/require_scratch
> > +
> > +echo "See if we leak"
> > +_test_unmount
> > +rmmod xfs
> 
> Test fails if xfs is built in or there's another mounted xfs, e.g.
> rootfs. There's a _require_btrfs_loadable btrfs-specific helper to check
> if we're able to unload btrfs as a module, and _notrun if not. I think
> we can make it generic easily, e.g. _require_module_loadable "$MOD_NAME"

Ok, I'll insert a new patch to refactor that in a generic way and then
rework this patch to use the generics.

> > +_test_mount
> > +_check_dmesg
> 
> check will do this for us :)

Ok.

--D

> 
> Thanks,
> Eryu
> --
> 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