btrfs: incremental send after replacing directory

Test that an incremental send operation works when in both snapshots
there are two directory inodes that have the same number but
different generations and have an entry with the same name that
corresponds to different inodes in each snapshot.

The btrfs issue is fixed by the following patch for the linux kernel:

  "Btrfs: incremental send, do not issue invalid rmdir operations"

Signed-off-by: Robbie Ko <robbieko@synology.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: Eryu Guan <eguan@redhat.com>
Signed-off-by: Eryu Guan <eguan@redhat.com>

3 files changed, 168 insertions, 0 deletions

diff --git a/tests/btrfs/135 b/tests/btrfs/135
new file mode 100755
index 0000000000..bdc70b3da7
--- /dev/null
+++ b/tests/btrfs/135
@@ -0,0 +1,159 @@
+#! /bin/bash
+# FS QA Test No. btrfs/135
+#
+# Test that an incremental send operation works when in both snapshots there are
+# two directory inodes that have the same number but different generations and
+# have an entry with the same name that corresponds to different inodes in each
+# snapshot.
+#
+#-----------------------------------------------------------------------
+# Copyright (C) 2017 Synology Inc. All Rights Reserved.
+# Author: Robbie Ko <robbieko@synology.com>
+#
+# 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"
+
+tmp=/tmp/$$
+status=1	# failure is the default!
+trap "_cleanup; exit \$status" 0 1 2 3 15
+
+_cleanup()
+{
+	cd /
+	rm -fr $send_files_dir
+	rm -f $tmp.*
+}
+
+# get standard environment, filters and checks
+. ./common/rc
+. ./common/filter
+
+# real QA test starts here
+_supported_fs btrfs
+_supported_os Linux
+_require_test
+_require_scratch
+_require_fssum
+
+send_files_dir=$TEST_DIR/btrfs-test-$seq
+
+rm -f $seqres.full
+rm -fr $send_files_dir
+mkdir $send_files_dir
+
+_scratch_mkfs >>$seqres.full 2>&1
+_scratch_mount
+
+touch $SCRATCH_MNT/f
+mkdir $SCRATCH_MNT/d1
+mkdir $SCRATCH_MNT/d259_old
+mv $SCRATCH_MNT/d1 $SCRATCH_MNT/d259_old/d1
+
+# Filesystem looks like:
+#
+# .                                                             (ino 256)
+# |--- f                                                        (ino 257)
+# |
+# |--- d259_old/                                                (ino 259)
+#         |--- d1/                                              (ino 258)
+#
+$BTRFS_UTIL_PROG subvolume snapshot -r $SCRATCH_MNT \
+	$SCRATCH_MNT/mysnap1 > /dev/null
+
+$BTRFS_UTIL_PROG send $SCRATCH_MNT/mysnap1 -f \
+	$send_files_dir/1.snap 2>&1 1>/dev/null | _filter_scratch
+
+_scratch_unmount
+_scratch_mkfs >>$seqres.full 2>&1
+_scratch_mount
+mkdir $SCRATCH_MNT/d1
+mkdir $SCRATCH_MNT/dir258
+mkdir $SCRATCH_MNT/dir259
+mv $SCRATCH_MNT/d1 $SCRATCH_MNT/dir259/d1
+
+# Filesystem now looks like:
+#
+# .                                                             (ino 256)
+# |--- dir258/                                                  (ino 258)
+# |
+# |--- dir259/                                                  (ino 259)
+#        |--- d1/                                               (ino 257)
+#
+# Notice that at this point all inodes have a generation with value 7, which is
+# the generation value for a brand new filesystem.
+
+# Now create the second snapshot. This makes the filesystem's current generation
+# value to increase to the value 8, due to a transaction commit performed by the
+# snapshot creation ioctl.
+$BTRFS_UTIL_PROG subvolume snapshot -r $SCRATCH_MNT \
+	$SCRATCH_MNT/mysnap2 > /dev/null
+
+# Now receive the first snapshot created in the first filesystem.
+# Before creating any inodes, the receive command creates the first snapshot,
+# which causes a transaction commit and therefore bumps the filesystem's current
+# generation to the value 9. All the inodes created end up getting a generation
+# value of 9 and the snapshot's root inode (256) gets a generation value of 8.
+$BTRFS_UTIL_PROG receive $SCRATCH_MNT -f $send_files_dir/1.snap > /dev/null
+rm $send_files_dir/1.snap
+
+$FSSUM_PROG -A -f -w $send_files_dir/1.fssum $SCRATCH_MNT/mysnap1
+$FSSUM_PROG -A -f -w $send_files_dir/2.fssum $SCRATCH_MNT/mysnap2
+
+$BTRFS_UTIL_PROG send $SCRATCH_MNT/mysnap1 -f \
+	$send_files_dir/1.snap 2>&1 1>/dev/null | _filter_scratch
+$BTRFS_UTIL_PROG send -p $SCRATCH_MNT/mysnap1 $SCRATCH_MNT/mysnap2 \
+	-f $send_files_dir/2.snap 2>&1 1>/dev/null | _filter_scratch
+
+# Now recreate the filesystem by receiving both send streams and verify we get
+# the same content that the original filesystem had.
+_scratch_unmount
+_scratch_mkfs >>$seqres.full 2>&1
+_scratch_mount
+
+$BTRFS_UTIL_PROG receive $SCRATCH_MNT -f $send_files_dir/1.snap > /dev/null
+$FSSUM_PROG -r $send_files_dir/1.fssum $SCRATCH_MNT/mysnap1
+# The receive of the incremental send stream used to fail because it contained
+# a rmdir operation with an invalid path. The output of the receive command,
+# with verbose mode (argument -vv), was the following:
+#
+# utimes
+# unlink f
+# mkdir o257-7-0
+# mkdir o259-7-0
+# rename o257-7-0 -> o259-7-0/d1
+# chown o259-7-0/d1 - uid=0, gid=0
+# chmod o259-7-0/d1 - mode=0755
+# utimes o259-7-0/d1
+# rmdir o258-9-0
+# ERROR: rmdir o258-9-0 failed: No such file or directory
+#
+# When the kernel was computing the send stream and processing inode 258, it
+# noticed that in both snapshots this inode is a direct child of inode 259, and
+# that in both snapshots inode 259 as an entry with the name "d1". Because of
+# that it decided (incorrectly) to issue an rmdir operation using the orphanized
+# name of the inode 258 from the parent snapshot (which has a generation of 9).
+# However that decision to send an rmdir operation due to the dentry collision
+# was incorrect because the inodes with number 259 in both snapshots are not the
+# same, they have different generations.
+#
+$BTRFS_UTIL_PROG receive $SCRATCH_MNT -f $send_files_dir/2.snap > /dev/null
+$FSSUM_PROG -r $send_files_dir/2.fssum $SCRATCH_MNT/mysnap2
+
+status=0
+exit
diff --git a/tests/btrfs/135.out b/tests/btrfs/135.out
new file mode 100644
index 0000000000..c917842dee
--- /dev/null
+++ b/tests/btrfs/135.out
@@ -0,0 +1,8 @@
+QA output created by 135
+At subvol SCRATCH_MNT/mysnap1
+At subvol mysnap1
+At subvol SCRATCH_MNT/mysnap1
+At subvol SCRATCH_MNT/mysnap2
+At subvol mysnap1
+OK
+OK
diff --git a/tests/btrfs/group b/tests/btrfs/group
index f33befb359..88fb8db420 100644
--- a/tests/btrfs/group
+++ b/tests/btrfs/group
@@ -136,3 +136,4 @@
 132 auto enospc
 133 auto quick send
 134 auto quick send
+135 auto quick send