[v2] fstests: test for btrfs cloning of zero length ranges

Commit Message

Filipe Manana March 30, 2015, 11:04 p.m. UTC

Test cloning a file range with a length of zero into a destination offset
greater than zero.

This made btrfs create an extent state record with a start offset greater than
the end offset, resulting in chaos such as an infinite loop when evicting an
inode.

This issue was fixed by the following linux kernel patch:

   Btrfs: fix inode eviction infinite loop after cloning into it

Signed-off-by: Filipe Manana <[e-mail]>
---

V2: Use an offset of 64Kb so that the test can run on platforms with any
    page size. In btrfs the fs block size must be a multiple of the page
    size, so a 4Kb offset would make the test fail on machines with a
    page size > 4Kb because the clone ioctl only accepts offsets and lengths
    that are multiples of the block size.

 tests/btrfs/086     | 77 +++++++++++++++++++++++++++++++++++++++++++++++++++++
 tests/btrfs/086.out |  2 ++
 tests/btrfs/group   |  1 +
 3 files changed, 80 insertions(+)
 create mode 100755 tests/btrfs/086
 create mode 100644 tests/btrfs/086.out

Comments

David Sterba March 31, 2015, 5:15 p.m. UTC | #1

On Tue, Mar 31, 2015 at 12:04:58AM +0100, Filipe Manana wrote:
> Test cloning a file range with a length of zero into a destination offset
> greater than zero.
> 
> This made btrfs create an extent state record with a start offset greater than
> the end offset, resulting in chaos such as an infinite loop when evicting an
> inode.
> 
> This issue was fixed by the following linux kernel patch:
> 
>    Btrfs: fix inode eviction infinite loop after cloning into it
> 
> Signed-off-by: Filipe Manana <[e-mail]>

Reviewed-by: David Sterba <[e-mail]>

> V2: Use an offset of 64Kb so that the test can run on platforms with any
>     page size. In btrfs the fs block size must be a multiple of the page
>     size, so a 4Kb offset would make the test fail on machines with a
>     page size > 4Kb because the clone ioctl only accepts offsets and lengths
>     that are multiples of the block size.

Nice.
--
To unsubscribe from this list: send the line "unsubscribe fstests" in
the body of a message to [e-mail]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch

diff --git a/tests/btrfs/086 b/tests/btrfs/086
new file mode 100755
index 0000000..77c8da6
--- /dev/null
+++ b/tests/btrfs/086
@@ -0,0 +1,77 @@ 
+#! /bin/bash
+# FS QA Test No. btrfs/086
+#
+# Test cloning a file range with a length of zero into a destination offset
+# greater than zero.
+#
+# This made btrfs create an extent state record with a start offset greater than
+# the end offset, resulting in chaos such as an infinite loop when evicting an
+# inode.
+#
+# This issue was fixed by the following linux kernel patch:
+#
+#   Btrfs: fix inode eviction infinite loop after cloning into it
+#
+#-----------------------------------------------------------------------
+# Copyright (C) 2015 SUSE Linux Products GmbH. All Rights Reserved.
+# Author: Filipe Manana <[e-mail]>
+#
+# 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()
+{
+	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_scratch
+_require_cloner
+_need_to_be_root
+
+rm -f $seqres.full
+
+_scratch_mkfs >>$seqres.full 2>&1
+_scratch_mount
+
+touch $SCRATCH_MNT/foo
+touch $SCRATCH_MNT/bar
+
+# Now attempt to clone foo into bar. Because we pass a length of zero, the
+# clone ioctl will adjust the length to match the size of the file foo (minus
+# the source offset which is zero) - because the adjusted length value is
+# zero, it made btrfs create an extent state record for file bar with a start
+# offset (64k) greater then its end offset (64k - 1), which is something never
+# supposed to happen and for example it made inode eviction enter an infinite
+# loop that dumped a warning trace on each iteration.
+$CLONER_PROG -s 0 -d 65536 -l 0 $SCRATCH_MNT/foo $SCRATCH_MNT/bar
+echo "bar file size after clone operation: $(stat -c %s $SCRATCH_MNT/bar)"
+
+status=0
+exit
diff --git a/tests/btrfs/086.out b/tests/btrfs/086.out
new file mode 100644
index 0000000..1ae87ba
--- /dev/null
+++ b/tests/btrfs/086.out
@@ -0,0 +1,2 @@ 
+QA output created by 086
+bar file size after clone operation: 0
diff --git a/tests/btrfs/group b/tests/btrfs/group
index 81d462f..92bc0f9 100644
--- a/tests/btrfs/group
+++ b/tests/btrfs/group
@@ -88,3 +88,4 @@ 
 083 auto quick send
 084 auto quick send
 085 auto quick send
+086 auto quick clone