On Wed, Oct 18, 2017 at 04:38:01PM -0700, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@xxxxxxxxxx>
>
> 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@xxxxxxxxxx>
> ---
> 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"
> +_test_mount
> +_check_dmesg
check will do this for us :)
Thanks,
Eryu
--
To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
