On 08/24/2016 05:53 AM, Eryu Guan wrote:
On Tue, Aug 23, 2016 at 01:51:58PM -0400, Josef Bacik wrote:
This is a test for a problem I hit with send/receive. If you mount with -o
inode_cache it is possible to recycle inode numbers, and btrfs assumes that when
you have a recycled inode number that it's parent directory won't be needing a
rename, but this is not the case. Without the patch to fix this problem this
test will panic the box if ASSERT() is turned on, otherwise it'll simply fail
the fssum.
Signed-off-by: Josef Bacik <jbacik@xxxxxx>
---
tests/btrfs/131 | 91 +++++++++++++++++++++++++++++++++++++++++++++++++++++
tests/btrfs/131.out | 2 ++
tests/btrfs/group | 1 +
3 files changed, 94 insertions(+)
create mode 100644 tests/btrfs/131
create mode 100644 tests/btrfs/131.out
diff --git a/tests/btrfs/131 b/tests/btrfs/131
new file mode 100644
index 0000000..596adc1
--- /dev/null
+++ b/tests/btrfs/131
@@ -0,0 +1,91 @@
+#! /bin/bash
+# FS QA Test 131
+#
+# Make sure btrfs handles send/receive of a file that has been destroyed and
+# recreated with a recycled inode number and a renamed parent directory. This
+# is a test for a fix based on the patch
+#
A trailing whitespace in above line :)
+# Btrfs: handle pending renames with recycled inodes properly
+#
+#-----------------------------------------------------------------------
+# Copyright (c) 2016 Facebook. 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=`mktemp -d`
Usually we define "tmp=/tmp/$$", and tests and some common helpers use
$tmp.xxx for tmp files. I think we should keep the consistency across
tests.
$tmp is used as a tmp dir to store the btrfs send data, I think we can
send the data to some dir in $TEST_DIR, e.g.
send_dir=$TEST_DIR/btrfs_send_$seq
+status=1 # failure is the default!
+trap "_cleanup; exit \$status" 0 1 2 3 15
+
+_cleanup()
+{
+ rm -fr $tmp
Then clean the send dir here, as well as $tmp.*
cd /
rm -rf $send_dir
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 btrfs
+_supported_os Linux
+_require_scratch
+_require_fssum
+
+_scratch_mkfs > /dev/null 2>&1
+_scratch_mount "-o inode_cache"
Need some comments to explain why mount with "-o inode_cache" option.
+
+_run_btrfs_util_prog subvolume create $SCRATCH_MNT/base
+mkdir $SCRATCH_MNT/base/b
+mkdir $SCRATCH_MNT/base/a
+mv $SCRATCH_MNT/base/b $SCRATCH_MNT/base/a
+
+_run_btrfs_util_prog subvolume snapshot -r $SCRATCH_MNT/base \
+ $SCRATCH_MNT/snap1
+_run_btrfs_util_prog send $SCRATCH_MNT/snap1 -f $tmp/1.snap
+
+mv $SCRATCH_MNT/base/a $SCRATCH_MNT/base/c
+rmdir $SCRATCH_MNT/base/c/b
+
+# We need to sync to make sure the inode number is recycled
+sync
+mkdir $SCRATCH_MNT/base/c/b
+
+_run_btrfs_util_prog subvolume snapshot -r $SCRATCH_MNT/base $SCRATCH_MNT/snap2
+run_check $FSSUM_PROG -A -f -w $tmp/fssum $SCRATCH_MNT/snap2
+
+_run_btrfs_util_prog send -p $SCRATCH_MNT/snap1 $SCRATCH_MNT/snap2 -f $tmp/2.snap
+
+_scratch_unmount
+_scratch_mkfs > /dev/null 2>&1
+_scratch_mount
+
+_run_btrfs_util_prog receive $SCRATCH_MNT -f $tmp/1.snap
+_run_btrfs_util_prog receive $SCRATCH_MNT -f $tmp/2.snap
+run_check $FSSUM_PROG -r $tmp/fssum $SCRATCH_MNT/snap2
I did see assert failure (and kernel crash) on kernel with ASSERT turned
on, but I didn't see fssum reporting failures on kernel with ASSERT
turned off. Did I miss anything?
[root@dhcp-66-86-11 xfstests]# ./check -s btrfs btrfs/131
SECTION -- btrfs
RECREATING -- btrfs on /dev/sda5
FSTYP -- btrfs
PLATFORM -- Linux/x86_64 dhcp-66-86-11 4.8.0-rc2
MKFS_OPTIONS -- /dev/sda6
MOUNT_OPTIONS -- -o context=system_u:object_r:nfs_t:s0 /dev/sda6 /mnt/testarea/scratch
btrfs/131 1s ... 1s
Ran: btrfs/131
Passed all 1 tests
SECTION -- btrfs
=========================
Ran: btrfs/131
Passed all 1 tests
Huh I must have run when I had debugging stuff still in my script, you are right
this does work fine with ASSERT()'s turned off. That's fine, I still want to
have this test here since it did trigger that ASSERT(). I'll go through and fix
up the things you pointed out, thanks!
Josef
--
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
