All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH v2 0/3] xfs/348: test handling of malformed inode mode
@ 2016-12-25 18:09 Amir Goldstein
  2016-12-25 18:09 ` [PATCH v2 1/3] xfs/348: test handling of invalid inode modes Amir Goldstein
                   ` (3 more replies)
  0 siblings, 4 replies; 10+ messages in thread
From: Amir Goldstein @ 2016-12-25 18:09 UTC (permalink / raw)
  To: Eryu Guan, Darrick J . Wong, Christoph Hellwig
  Cc: Brian Foster, fstests, linux-xfs

Eryu,

I beefed up the initial test sent earlier today with more checks
on mounted fs.

The original patch testing only xfs_repair remains patch 1 in this series.

Patch 2 adds fstat tests on mounted fs, which are safe on my test system.

Patch 3 is explosive. It exposes an XFS assert, but I left a "safety pin"
that needs to be commented out to reproduce the assert.

I will soon post my kernel configs and dmesg logs.

Amir.

v2:
- test fstat with malformed inode mode
- test readlink/readdir with malformed inode mode

v1:
- test xfs_repair with malformed inode mode

Amir Goldstein (3):
  xfs/348: test handling of invalid inode modes
  xfs/348: test fstat with malformed inode mode
  xfs/348: test readlink/readdir with malformed inode mode

 tests/xfs/348     | 154 +++++++++++++++++++++++++++
 tests/xfs/348.out | 312 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
 tests/xfs/group   |   1 +
 3 files changed, 467 insertions(+)
 create mode 100755 tests/xfs/348
 create mode 100644 tests/xfs/348.out

-- 
2.7.4


^ permalink raw reply	[flat|nested] 10+ messages in thread

* [PATCH v2 1/3] xfs/348: test handling of invalid inode modes
  2016-12-25 18:09 [PATCH v2 0/3] xfs/348: test handling of malformed inode mode Amir Goldstein
@ 2016-12-25 18:09 ` Amir Goldstein
  2017-01-03 20:36   ` Darrick J. Wong
  2016-12-25 18:09 ` [PATCH v2 2/3] xfs/348: test fstat with malformed inode mode Amir Goldstein
                   ` (2 subsequent siblings)
  3 siblings, 1 reply; 10+ messages in thread
From: Amir Goldstein @ 2016-12-25 18:09 UTC (permalink / raw)
  To: Eryu Guan, Darrick J . Wong, Christoph Hellwig
  Cc: Brian Foster, fstests, linux-xfs

Set all possible file type values for different types of files
and verify that xfs_repair detects the correct errors.

When setting invalid file type values (e.g. core.mode = 0170644),
all files are expected to have been junked by xfs_repair.

When setting valid file type values to non matching file types,
xfs_repair would either detect wrong format and junk the file, e.g.:
  would have junked entry "FILE" in directory PARENT_INO
or detect a ftype mismatch error, e.g.:
  would fix ftype mismatch (5/3) in directory/child PARENT_INO/FIFO_INO

If ftype feature is enabled, when setting file type to one of the
special types (i.e. FIFO(1), CHRDEV(2),BLKDEV(6),SOCKET(14)),
xfs_repair is expected to detect ftype mismatch error. Otherewise,
xfs_repair is not expected to detect ftype mismatch error.

Signed-off-by: Amir Goldstein <amir73il@gmail.com>
---
 tests/xfs/348     | 128 ++++++++++++++++++++++++++++++++
 tests/xfs/348.out | 216 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
 tests/xfs/group   |   1 +
 3 files changed, 345 insertions(+)
 create mode 100755 tests/xfs/348
 create mode 100644 tests/xfs/348.out

diff --git a/tests/xfs/348 b/tests/xfs/348
new file mode 100755
index 0000000..9ec696a
--- /dev/null
+++ b/tests/xfs/348
@@ -0,0 +1,128 @@
+#! /bin/bash
+# FSQA Test No. 348
+#
+# Test handling of invalid inode modes
+#
+# Set all possible file type values for different types of files
+# and verify that xfs_repair detects the correct errors.
+#
+#-----------------------------------------------------------------------
+#
+# Copyright (C) 2016 CTERA Networks. All Rights Reserved.
+# Author: Amir Goldstein <amir73il@gmail.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()
+{
+	rm -f $tmp.*
+}
+
+# get standard environment, filters and checks
+. ./common/rc
+. ./common/filter
+. ./common/repair
+
+# real QA test starts here
+_supported_fs xfs
+_supported_os Linux
+_require_scratch
+
+rm -f $seqres.full
+
+_scratch_mkfs >>$seqres.full 2>&1
+
+_scratch_mount
+
+# Create our test files.
+testdir=$SCRATCH_MNT/test
+mkdir -p $testdir
+mkdir $testdir/DIR
+echo 123 > $testdir/FILE
+ln -s $testdir/FILE $testdir/SYMLINK
+mknod $testdir/CHRDEV c 1 1
+mknod $testdir/BLKDEV b 1 1
+mknod $testdir/FIFO p
+
+xfs_info $SCRATCH_MNT | grep -q "ftype=1" && FTYPE_FEATURE=1
+
+# Record test dir inode for xfs_repair filter
+inode_filter=$tmp.sed
+rm -f $inode_filter
+pino=$(ls -id $testdir | awk '{print $1}')
+echo "s/inode $pino/PARENT_INO/" >> $inode_filter
+echo "s/directory $pino/directory PARENT_INO/" >> $inode_filter
+
+inodes=""
+# Record inode numbers for xfs_db commands and xfs_repair filter
+for f in DIR FILE SYMLINK CHRDEV BLKDEV FIFO; do
+	ino=$(ls -id $testdir/$f | awk '{print $1}')
+	inodes="$inodes $ino"
+	echo "s/inode $ino/${f}_INO/" >> $inode_filter
+done
+
+_scratch_unmount
+
+# Possible mode file type values (mode & S_IFMT) >> 12
+dtypes="0 1 2 3 4 5 6 7 10 11 12 13 14 15 16 17"
+echo "===== Find inode by file type:"
+for dt in $dtypes; do
+	# Set all our test files to dt value
+	for ino in $inodes; do
+		_scratch_xfs_db -c "inode $ino" -c "print core.mode" | \
+			grep -q "0${dt}0...$" && \
+			(echo "dt=$dt => inode $ino" | sed -f $inode_filter)
+	done
+done
+for dt in $dtypes; do
+	echo
+	echo "===== Setting dt=$dt to all files:"
+	# Set all our test files to dt value
+	for ino in $inodes; do
+		_scratch_xfs_db -x -c "inode $ino" -c "write core.mode 0${dt}0644"
+	done
+	# Repair should detect the inconsistencies
+	# For invalid dt values, all files would have been junked.
+	# For valid dt values, one test file is expected to be valid.
+	# The rest would either have wrong format or non matching dir ftype.
+	_scratch_xfs_repair -n 2>&1  | tee -a $seqres.full | \
+		_filter_repair | grep "^would have junked" | sed -f $inode_filter | sort -u
+	# If ftype feature is enabled, when setting file type to one of the
+	# special types (i.e. FIFO(1), CHRDEV(2),BLKDEV(6),SOCKET(14)),
+	# xfs_repair is expected to detect ftype mismatch error. Otherewise,
+	# xfs_repair is not expected to detect ftype mismatch error.
+	if [ "$FTYPE_FEATURE" = 1 ] && (echo ':1:2:6:14:' | grep -q ":$dt:"); then
+		_scratch_xfs_repair -n 2>&1 | grep -q "^would fix ftype mismatch" || \
+			_fail "xfs_repair should fix ftype mismatch"
+	else
+		_scratch_xfs_repair -n 2>&1 | grep -q -v "^would fix ftype mismatch" || \
+			_fail "xfs_repair should not fix ftype mismatch"
+	fi
+done
+
+# Repair should detect and junk all test files
+_scratch_xfs_repair 2>&1 >> $seqres.full 2>&1 || _fail "xfs_repair should not fail"
+
+status=0
+exit
diff --git a/tests/xfs/348.out b/tests/xfs/348.out
new file mode 100644
index 0000000..bf4924b
--- /dev/null
+++ b/tests/xfs/348.out
@@ -0,0 +1,216 @@
+QA output created by 348
+===== Find inode by file type:
+dt=1 => FIFO_INO
+dt=2 => CHRDEV_INO
+dt=4 => DIR_INO
+dt=6 => BLKDEV_INO
+dt=10 => FILE_INO
+dt=12 => SYMLINK_INO
+
+===== Setting dt=0 to all files:
+core.mode = 0644
+core.mode = 0644
+core.mode = 0644
+core.mode = 0644
+core.mode = 0644
+core.mode = 0644
+would have junked entry "BLKDEV" in directory PARENT_INO
+would have junked entry "CHRDEV" in directory PARENT_INO
+would have junked entry "DIR" in directory PARENT_INO
+would have junked entry "FIFO" in directory PARENT_INO
+would have junked entry "FILE" in directory PARENT_INO
+would have junked entry "SYMLINK" in directory PARENT_INO
+
+===== Setting dt=1 to all files:
+core.mode = 010644
+core.mode = 010644
+core.mode = 010644
+core.mode = 010644
+core.mode = 010644
+core.mode = 010644
+would have junked entry "DIR" in directory PARENT_INO
+would have junked entry "FILE" in directory PARENT_INO
+would have junked entry "SYMLINK" in directory PARENT_INO
+
+===== Setting dt=2 to all files:
+core.mode = 020644
+core.mode = 020644
+core.mode = 020644
+core.mode = 020644
+core.mode = 020644
+core.mode = 020644
+would have junked entry "DIR" in directory PARENT_INO
+would have junked entry "FILE" in directory PARENT_INO
+would have junked entry "SYMLINK" in directory PARENT_INO
+
+===== Setting dt=3 to all files:
+core.mode = 030644
+core.mode = 030644
+core.mode = 030644
+core.mode = 030644
+core.mode = 030644
+core.mode = 030644
+would have junked entry "BLKDEV" in directory PARENT_INO
+would have junked entry "CHRDEV" in directory PARENT_INO
+would have junked entry "DIR" in directory PARENT_INO
+would have junked entry "FIFO" in directory PARENT_INO
+would have junked entry "FILE" in directory PARENT_INO
+would have junked entry "SYMLINK" in directory PARENT_INO
+
+===== Setting dt=4 to all files:
+core.mode = 040644
+core.mode = 040644
+core.mode = 040644
+core.mode = 040644
+core.mode = 040644
+core.mode = 040644
+would have junked entry "BLKDEV" in directory PARENT_INO
+would have junked entry "CHRDEV" in directory PARENT_INO
+would have junked entry "FIFO" in directory PARENT_INO
+would have junked entry "FILE" in directory PARENT_INO
+
+===== Setting dt=5 to all files:
+core.mode = 050644
+core.mode = 050644
+core.mode = 050644
+core.mode = 050644
+core.mode = 050644
+core.mode = 050644
+would have junked entry "BLKDEV" in directory PARENT_INO
+would have junked entry "CHRDEV" in directory PARENT_INO
+would have junked entry "DIR" in directory PARENT_INO
+would have junked entry "FIFO" in directory PARENT_INO
+would have junked entry "FILE" in directory PARENT_INO
+would have junked entry "SYMLINK" in directory PARENT_INO
+
+===== Setting dt=6 to all files:
+core.mode = 060644
+core.mode = 060644
+core.mode = 060644
+core.mode = 060644
+core.mode = 060644
+core.mode = 060644
+would have junked entry "DIR" in directory PARENT_INO
+would have junked entry "FILE" in directory PARENT_INO
+would have junked entry "SYMLINK" in directory PARENT_INO
+
+===== Setting dt=7 to all files:
+core.mode = 070644
+core.mode = 070644
+core.mode = 070644
+core.mode = 070644
+core.mode = 070644
+core.mode = 070644
+would have junked entry "BLKDEV" in directory PARENT_INO
+would have junked entry "CHRDEV" in directory PARENT_INO
+would have junked entry "DIR" in directory PARENT_INO
+would have junked entry "FIFO" in directory PARENT_INO
+would have junked entry "FILE" in directory PARENT_INO
+would have junked entry "SYMLINK" in directory PARENT_INO
+
+===== Setting dt=10 to all files:
+core.mode = 0100644
+core.mode = 0100644
+core.mode = 0100644
+core.mode = 0100644
+core.mode = 0100644
+core.mode = 0100644
+would have junked entry "BLKDEV" in directory PARENT_INO
+would have junked entry "CHRDEV" in directory PARENT_INO
+would have junked entry "DIR" in directory PARENT_INO
+would have junked entry "FIFO" in directory PARENT_INO
+would have junked entry "SYMLINK" in directory PARENT_INO
+
+===== Setting dt=11 to all files:
+core.mode = 0110644
+core.mode = 0110644
+core.mode = 0110644
+core.mode = 0110644
+core.mode = 0110644
+core.mode = 0110644
+would have junked entry "BLKDEV" in directory PARENT_INO
+would have junked entry "CHRDEV" in directory PARENT_INO
+would have junked entry "DIR" in directory PARENT_INO
+would have junked entry "FIFO" in directory PARENT_INO
+would have junked entry "FILE" in directory PARENT_INO
+would have junked entry "SYMLINK" in directory PARENT_INO
+
+===== Setting dt=12 to all files:
+core.mode = 0120644
+core.mode = 0120644
+core.mode = 0120644
+core.mode = 0120644
+core.mode = 0120644
+core.mode = 0120644
+would have junked entry "BLKDEV" in directory PARENT_INO
+would have junked entry "CHRDEV" in directory PARENT_INO
+would have junked entry "DIR" in directory PARENT_INO
+would have junked entry "FIFO" in directory PARENT_INO
+would have junked entry "FILE" in directory PARENT_INO
+
+===== Setting dt=13 to all files:
+core.mode = 0130644
+core.mode = 0130644
+core.mode = 0130644
+core.mode = 0130644
+core.mode = 0130644
+core.mode = 0130644
+would have junked entry "BLKDEV" in directory PARENT_INO
+would have junked entry "CHRDEV" in directory PARENT_INO
+would have junked entry "DIR" in directory PARENT_INO
+would have junked entry "FIFO" in directory PARENT_INO
+would have junked entry "FILE" in directory PARENT_INO
+would have junked entry "SYMLINK" in directory PARENT_INO
+
+===== Setting dt=14 to all files:
+core.mode = 0140644
+core.mode = 0140644
+core.mode = 0140644
+core.mode = 0140644
+core.mode = 0140644
+core.mode = 0140644
+would have junked entry "DIR" in directory PARENT_INO
+would have junked entry "FILE" in directory PARENT_INO
+would have junked entry "SYMLINK" in directory PARENT_INO
+
+===== Setting dt=15 to all files:
+core.mode = 0150644
+core.mode = 0150644
+core.mode = 0150644
+core.mode = 0150644
+core.mode = 0150644
+core.mode = 0150644
+would have junked entry "BLKDEV" in directory PARENT_INO
+would have junked entry "CHRDEV" in directory PARENT_INO
+would have junked entry "DIR" in directory PARENT_INO
+would have junked entry "FIFO" in directory PARENT_INO
+would have junked entry "FILE" in directory PARENT_INO
+would have junked entry "SYMLINK" in directory PARENT_INO
+
+===== Setting dt=16 to all files:
+core.mode = 0160644
+core.mode = 0160644
+core.mode = 0160644
+core.mode = 0160644
+core.mode = 0160644
+core.mode = 0160644
+would have junked entry "BLKDEV" in directory PARENT_INO
+would have junked entry "CHRDEV" in directory PARENT_INO
+would have junked entry "DIR" in directory PARENT_INO
+would have junked entry "FIFO" in directory PARENT_INO
+would have junked entry "FILE" in directory PARENT_INO
+would have junked entry "SYMLINK" in directory PARENT_INO
+
+===== Setting dt=17 to all files:
+core.mode = 0170644
+core.mode = 0170644
+core.mode = 0170644
+core.mode = 0170644
+core.mode = 0170644
+core.mode = 0170644
+would have junked entry "BLKDEV" in directory PARENT_INO
+would have junked entry "CHRDEV" in directory PARENT_INO
+would have junked entry "DIR" in directory PARENT_INO
+would have junked entry "FIFO" in directory PARENT_INO
+would have junked entry "FILE" in directory PARENT_INO
+would have junked entry "SYMLINK" in directory PARENT_INO
diff --git a/tests/xfs/group b/tests/xfs/group
index c237b50..68c2bb6 100644
--- a/tests/xfs/group
+++ b/tests/xfs/group
@@ -334,3 +334,4 @@
 345 auto quick clone
 346 auto quick clone
 347 auto quick clone
+348 auto quick repair
-- 
2.7.4


^ permalink raw reply related	[flat|nested] 10+ messages in thread

* [PATCH v2 2/3] xfs/348: test fstat with malformed inode mode
  2016-12-25 18:09 [PATCH v2 0/3] xfs/348: test handling of malformed inode mode Amir Goldstein
  2016-12-25 18:09 ` [PATCH v2 1/3] xfs/348: test handling of invalid inode modes Amir Goldstein
@ 2016-12-25 18:09 ` Amir Goldstein
  2016-12-25 18:09 ` [PATCH v2 3/3] xfs/348: test readlink/readdir " Amir Goldstein
  2017-01-03  5:52 ` [PATCH v2 0/3] xfs/348: test handling of " Eryu Guan
  3 siblings, 0 replies; 10+ messages in thread
From: Amir Goldstein @ 2016-12-25 18:09 UTC (permalink / raw)
  To: Eryu Guan, Darrick J . Wong, Christoph Hellwig
  Cc: Brian Foster, fstests, linux-xfs

In addition to testing xfs_repair on inodes with malformed mode,
also test fstat of those inodes on a mounted fs.

This additional test is quite noisy with dmesg warnings, so
check dmesg has been disabled.

Signed-off-by: Amir Goldstein <amir73il@gmail.com>
---
 tests/xfs/348     | 16 ++++++++++
 tests/xfs/348.out | 96 +++++++++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 112 insertions(+)

diff --git a/tests/xfs/348 b/tests/xfs/348
index 9ec696a..2a54f9e 100755
--- a/tests/xfs/348
+++ b/tests/xfs/348
@@ -49,6 +49,10 @@ _supported_fs xfs
 _supported_os Linux
 _require_scratch
 
+# This test will corrupt fs intentionally, so there will be WARNINGs
+# in dmesg as expected
+_disable_dmesg_check
+
 rm -f $seqres.full
 
 _scratch_mkfs >>$seqres.full 2>&1
@@ -119,6 +123,18 @@ for dt in $dtypes; do
 		_scratch_xfs_repair -n 2>&1 | grep -q -v "^would fix ftype mismatch" || \
 			_fail "xfs_repair should not fix ftype mismatch"
 	fi
+
+	_scratch_mount
+	for file in DIR FILE SYMLINK CHRDEV BLKDEV FIFO; do
+		rm -f $tmp.stat.err
+		ftype=$(stat --printf=%F $testdir/$file 2>$tmp.stat.err)
+		if [ -s $tmp.stat.err ]; then
+			cat $tmp.stat.err
+		else
+			echo "stat: '$testdir/$file' is a $ftype"
+		fi | _filter_scratch
+	done
+	_scratch_unmount
 done
 
 # Repair should detect and junk all test files
diff --git a/tests/xfs/348.out b/tests/xfs/348.out
index bf4924b..cdb7d78 100644
--- a/tests/xfs/348.out
+++ b/tests/xfs/348.out
@@ -20,6 +20,12 @@ would have junked entry "DIR" in directory PARENT_INO
 would have junked entry "FIFO" in directory PARENT_INO
 would have junked entry "FILE" in directory PARENT_INO
 would have junked entry "SYMLINK" in directory PARENT_INO
+stat: cannot stat 'SCRATCH_MNT/test/DIR': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/FILE': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/SYMLINK': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/CHRDEV': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/BLKDEV': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/FIFO': Structure needs cleaning
 
 ===== Setting dt=1 to all files:
 core.mode = 010644
@@ -31,6 +37,12 @@ core.mode = 010644
 would have junked entry "DIR" in directory PARENT_INO
 would have junked entry "FILE" in directory PARENT_INO
 would have junked entry "SYMLINK" in directory PARENT_INO
+stat: cannot stat 'SCRATCH_MNT/test/DIR': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/FILE': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/SYMLINK': Structure needs cleaning
+stat: 'SCRATCH_MNT/test/CHRDEV' is a fifo
+stat: 'SCRATCH_MNT/test/BLKDEV' is a fifo
+stat: 'SCRATCH_MNT/test/FIFO' is a fifo
 
 ===== Setting dt=2 to all files:
 core.mode = 020644
@@ -42,6 +54,12 @@ core.mode = 020644
 would have junked entry "DIR" in directory PARENT_INO
 would have junked entry "FILE" in directory PARENT_INO
 would have junked entry "SYMLINK" in directory PARENT_INO
+stat: cannot stat 'SCRATCH_MNT/test/DIR': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/FILE': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/SYMLINK': Structure needs cleaning
+stat: 'SCRATCH_MNT/test/CHRDEV' is a character special file
+stat: 'SCRATCH_MNT/test/BLKDEV' is a character special file
+stat: 'SCRATCH_MNT/test/FIFO' is a character special file
 
 ===== Setting dt=3 to all files:
 core.mode = 030644
@@ -56,6 +74,12 @@ would have junked entry "DIR" in directory PARENT_INO
 would have junked entry "FIFO" in directory PARENT_INO
 would have junked entry "FILE" in directory PARENT_INO
 would have junked entry "SYMLINK" in directory PARENT_INO
+stat: cannot stat 'SCRATCH_MNT/test/DIR': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/FILE': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/SYMLINK': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/CHRDEV': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/BLKDEV': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/FIFO': Structure needs cleaning
 
 ===== Setting dt=4 to all files:
 core.mode = 040644
@@ -68,6 +92,12 @@ would have junked entry "BLKDEV" in directory PARENT_INO
 would have junked entry "CHRDEV" in directory PARENT_INO
 would have junked entry "FIFO" in directory PARENT_INO
 would have junked entry "FILE" in directory PARENT_INO
+stat: 'SCRATCH_MNT/test/DIR' is a directory
+stat: 'SCRATCH_MNT/test/FILE' is a directory
+stat: 'SCRATCH_MNT/test/SYMLINK' is a directory
+stat: cannot stat 'SCRATCH_MNT/test/CHRDEV': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/BLKDEV': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/FIFO': Structure needs cleaning
 
 ===== Setting dt=5 to all files:
 core.mode = 050644
@@ -82,6 +112,12 @@ would have junked entry "DIR" in directory PARENT_INO
 would have junked entry "FIFO" in directory PARENT_INO
 would have junked entry "FILE" in directory PARENT_INO
 would have junked entry "SYMLINK" in directory PARENT_INO
+stat: cannot stat 'SCRATCH_MNT/test/DIR': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/FILE': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/SYMLINK': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/CHRDEV': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/BLKDEV': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/FIFO': Structure needs cleaning
 
 ===== Setting dt=6 to all files:
 core.mode = 060644
@@ -93,6 +129,12 @@ core.mode = 060644
 would have junked entry "DIR" in directory PARENT_INO
 would have junked entry "FILE" in directory PARENT_INO
 would have junked entry "SYMLINK" in directory PARENT_INO
+stat: cannot stat 'SCRATCH_MNT/test/DIR': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/FILE': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/SYMLINK': Structure needs cleaning
+stat: 'SCRATCH_MNT/test/CHRDEV' is a block special file
+stat: 'SCRATCH_MNT/test/BLKDEV' is a block special file
+stat: 'SCRATCH_MNT/test/FIFO' is a block special file
 
 ===== Setting dt=7 to all files:
 core.mode = 070644
@@ -107,6 +149,12 @@ would have junked entry "DIR" in directory PARENT_INO
 would have junked entry "FIFO" in directory PARENT_INO
 would have junked entry "FILE" in directory PARENT_INO
 would have junked entry "SYMLINK" in directory PARENT_INO
+stat: cannot stat 'SCRATCH_MNT/test/DIR': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/FILE': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/SYMLINK': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/CHRDEV': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/BLKDEV': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/FIFO': Structure needs cleaning
 
 ===== Setting dt=10 to all files:
 core.mode = 0100644
@@ -120,6 +168,12 @@ would have junked entry "CHRDEV" in directory PARENT_INO
 would have junked entry "DIR" in directory PARENT_INO
 would have junked entry "FIFO" in directory PARENT_INO
 would have junked entry "SYMLINK" in directory PARENT_INO
+stat: cannot stat 'SCRATCH_MNT/test/DIR': Structure needs cleaning
+stat: 'SCRATCH_MNT/test/FILE' is a regular file
+stat: cannot stat 'SCRATCH_MNT/test/SYMLINK': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/CHRDEV': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/BLKDEV': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/FIFO': Structure needs cleaning
 
 ===== Setting dt=11 to all files:
 core.mode = 0110644
@@ -134,6 +188,12 @@ would have junked entry "DIR" in directory PARENT_INO
 would have junked entry "FIFO" in directory PARENT_INO
 would have junked entry "FILE" in directory PARENT_INO
 would have junked entry "SYMLINK" in directory PARENT_INO
+stat: cannot stat 'SCRATCH_MNT/test/DIR': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/FILE': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/SYMLINK': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/CHRDEV': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/BLKDEV': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/FIFO': Structure needs cleaning
 
 ===== Setting dt=12 to all files:
 core.mode = 0120644
@@ -147,6 +207,12 @@ would have junked entry "CHRDEV" in directory PARENT_INO
 would have junked entry "DIR" in directory PARENT_INO
 would have junked entry "FIFO" in directory PARENT_INO
 would have junked entry "FILE" in directory PARENT_INO
+stat: 'SCRATCH_MNT/test/DIR' is a symbolic link
+stat: 'SCRATCH_MNT/test/FILE' is a symbolic link
+stat: 'SCRATCH_MNT/test/SYMLINK' is a symbolic link
+stat: cannot stat 'SCRATCH_MNT/test/CHRDEV': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/BLKDEV': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/FIFO': Structure needs cleaning
 
 ===== Setting dt=13 to all files:
 core.mode = 0130644
@@ -161,6 +227,12 @@ would have junked entry "DIR" in directory PARENT_INO
 would have junked entry "FIFO" in directory PARENT_INO
 would have junked entry "FILE" in directory PARENT_INO
 would have junked entry "SYMLINK" in directory PARENT_INO
+stat: cannot stat 'SCRATCH_MNT/test/DIR': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/FILE': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/SYMLINK': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/CHRDEV': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/BLKDEV': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/FIFO': Structure needs cleaning
 
 ===== Setting dt=14 to all files:
 core.mode = 0140644
@@ -172,6 +244,12 @@ core.mode = 0140644
 would have junked entry "DIR" in directory PARENT_INO
 would have junked entry "FILE" in directory PARENT_INO
 would have junked entry "SYMLINK" in directory PARENT_INO
+stat: cannot stat 'SCRATCH_MNT/test/DIR': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/FILE': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/SYMLINK': Structure needs cleaning
+stat: 'SCRATCH_MNT/test/CHRDEV' is a socket
+stat: 'SCRATCH_MNT/test/BLKDEV' is a socket
+stat: 'SCRATCH_MNT/test/FIFO' is a socket
 
 ===== Setting dt=15 to all files:
 core.mode = 0150644
@@ -186,6 +264,12 @@ would have junked entry "DIR" in directory PARENT_INO
 would have junked entry "FIFO" in directory PARENT_INO
 would have junked entry "FILE" in directory PARENT_INO
 would have junked entry "SYMLINK" in directory PARENT_INO
+stat: cannot stat 'SCRATCH_MNT/test/DIR': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/FILE': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/SYMLINK': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/CHRDEV': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/BLKDEV': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/FIFO': Structure needs cleaning
 
 ===== Setting dt=16 to all files:
 core.mode = 0160644
@@ -200,6 +284,12 @@ would have junked entry "DIR" in directory PARENT_INO
 would have junked entry "FIFO" in directory PARENT_INO
 would have junked entry "FILE" in directory PARENT_INO
 would have junked entry "SYMLINK" in directory PARENT_INO
+stat: cannot stat 'SCRATCH_MNT/test/DIR': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/FILE': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/SYMLINK': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/CHRDEV': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/BLKDEV': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/FIFO': Structure needs cleaning
 
 ===== Setting dt=17 to all files:
 core.mode = 0170644
@@ -214,3 +304,9 @@ would have junked entry "DIR" in directory PARENT_INO
 would have junked entry "FIFO" in directory PARENT_INO
 would have junked entry "FILE" in directory PARENT_INO
 would have junked entry "SYMLINK" in directory PARENT_INO
+stat: cannot stat 'SCRATCH_MNT/test/DIR': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/FILE': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/SYMLINK': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/CHRDEV': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/BLKDEV': Structure needs cleaning
+stat: cannot stat 'SCRATCH_MNT/test/FIFO': Structure needs cleaning
-- 
2.7.4


^ permalink raw reply related	[flat|nested] 10+ messages in thread

* [PATCH v2 3/3] xfs/348: test readlink/readdir with malformed inode mode
  2016-12-25 18:09 [PATCH v2 0/3] xfs/348: test handling of malformed inode mode Amir Goldstein
  2016-12-25 18:09 ` [PATCH v2 1/3] xfs/348: test handling of invalid inode modes Amir Goldstein
  2016-12-25 18:09 ` [PATCH v2 2/3] xfs/348: test fstat with malformed inode mode Amir Goldstein
@ 2016-12-25 18:09 ` Amir Goldstein
  2017-01-03  5:52 ` [PATCH v2 0/3] xfs/348: test handling of " Eryu Guan
  3 siblings, 0 replies; 10+ messages in thread
From: Amir Goldstein @ 2016-12-25 18:09 UTC (permalink / raw)
  To: Eryu Guan, Darrick J . Wong, Christoph Hellwig
  Cc: Brian Foster, fstests, linux-xfs

In addition to testing xfs_repair on inodes with malformed mode,
and fstat of those inodes on a mounted fs, try to also list content
of mock directory and readlink of mock symlink.

Readdir of mock directory triggers XFS assertion on umount:
 XFS: Assertion failed: !rwsem_is_locked(&inode->i_rwsem), file:
 /home/amir/build/src/linux/fs/xfs/xfs_super.c, line: 951

To reproduce, need to comment out the variable 'safety'.

Signed-off-by: Amir Goldstein <amir73il@gmail.com>
---
 tests/xfs/348 | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/tests/xfs/348 b/tests/xfs/348
index 2a54f9e..cb0e3ea 100755
--- a/tests/xfs/348
+++ b/tests/xfs/348
@@ -132,6 +132,16 @@ for dt in $dtypes; do
 			cat $tmp.stat.err
 		else
 			echo "stat: '$testdir/$file' is a $ftype"
+			# Verify that readlink of a file posing as a symlink
+			# and ls of a file posing as a directory does not blow up
+			safety="-d"
+			# NOTE that ls DOES blow up with kernel 4.9 on malformed
+			# directory - remove the safety pin to explode
+			if [ -d $testdir/$file ]; then
+				ls "$safety" $testdir/$file &> /dev/null
+			elif [ -h $testdir/$file ]; then
+				readlink $testdir/$file &> /dev/null
+			fi
 		fi | _filter_scratch
 	done
 	_scratch_unmount
-- 
2.7.4


^ permalink raw reply related	[flat|nested] 10+ messages in thread

* Re: [PATCH v2 0/3] xfs/348: test handling of malformed inode mode
  2016-12-25 18:09 [PATCH v2 0/3] xfs/348: test handling of malformed inode mode Amir Goldstein
                   ` (2 preceding siblings ...)
  2016-12-25 18:09 ` [PATCH v2 3/3] xfs/348: test readlink/readdir " Amir Goldstein
@ 2017-01-03  5:52 ` Eryu Guan
  2017-01-11  6:45   ` Amir Goldstein
  3 siblings, 1 reply; 10+ messages in thread
From: Eryu Guan @ 2017-01-03  5:52 UTC (permalink / raw)
  To: Amir Goldstein
  Cc: Darrick J . Wong, Christoph Hellwig, Brian Foster, fstests, linux-xfs

On Sun, Dec 25, 2016 at 08:09:33PM +0200, Amir Goldstein wrote:
> Eryu,
> 
> I beefed up the initial test sent earlier today with more checks
> on mounted fs.
> 
> The original patch testing only xfs_repair remains patch 1 in this series.
> 
> Patch 2 adds fstat tests on mounted fs, which are safe on my test system.
> 
> Patch 3 is explosive. It exposes an XFS assert, but I left a "safety pin"
> that needs to be commented out to reproduce the assert.

My kenrel config doesn't turn DEBUG on, so I don't see a kernel crash :)

I can push this test out after the fix lands in upstream, then I think
there's no need to leave a switch in the test.

Otherwise tests look good to me. But I'd like to have Darrick to review
too, as he had written many fuzzer tests and suggested this test :)

Thanks,
Eryu

> 
> I will soon post my kernel configs and dmesg logs.
> 
> Amir.
> 
> v2:
> - test fstat with malformed inode mode
> - test readlink/readdir with malformed inode mode
> 
> v1:
> - test xfs_repair with malformed inode mode
> 
> Amir Goldstein (3):
>   xfs/348: test handling of invalid inode modes
>   xfs/348: test fstat with malformed inode mode
>   xfs/348: test readlink/readdir with malformed inode mode
> 
>  tests/xfs/348     | 154 +++++++++++++++++++++++++++
>  tests/xfs/348.out | 312 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
>  tests/xfs/group   |   1 +
>  3 files changed, 467 insertions(+)
>  create mode 100755 tests/xfs/348
>  create mode 100644 tests/xfs/348.out
> 
> -- 
> 2.7.4
> 

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [PATCH v2 1/3] xfs/348: test handling of invalid inode modes
  2016-12-25 18:09 ` [PATCH v2 1/3] xfs/348: test handling of invalid inode modes Amir Goldstein
@ 2017-01-03 20:36   ` Darrick J. Wong
  2017-01-04 10:36     ` Amir Goldstein
  0 siblings, 1 reply; 10+ messages in thread
From: Darrick J. Wong @ 2017-01-03 20:36 UTC (permalink / raw)
  To: Amir Goldstein
  Cc: Eryu Guan, Christoph Hellwig, Brian Foster, fstests, linux-xfs

On Sun, Dec 25, 2016 at 08:09:34PM +0200, Amir Goldstein wrote:
> Set all possible file type values for different types of files
> and verify that xfs_repair detects the correct errors.
> 
> When setting invalid file type values (e.g. core.mode = 0170644),
> all files are expected to have been junked by xfs_repair.
> 
> When setting valid file type values to non matching file types,
> xfs_repair would either detect wrong format and junk the file, e.g.:
>   would have junked entry "FILE" in directory PARENT_INO
> or detect a ftype mismatch error, e.g.:
>   would fix ftype mismatch (5/3) in directory/child PARENT_INO/FIFO_INO
> 
> If ftype feature is enabled, when setting file type to one of the
> special types (i.e. FIFO(1), CHRDEV(2),BLKDEV(6),SOCKET(14)),
> xfs_repair is expected to detect ftype mismatch error. Otherewise,

"Otherwise"

> xfs_repair is not expected to detect ftype mismatch error.
> 
> Signed-off-by: Amir Goldstein <amir73il@gmail.com>
> ---
>  tests/xfs/348     | 128 ++++++++++++++++++++++++++++++++
>  tests/xfs/348.out | 216 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
>  tests/xfs/group   |   1 +
>  3 files changed, 345 insertions(+)
>  create mode 100755 tests/xfs/348
>  create mode 100644 tests/xfs/348.out
> 
> diff --git a/tests/xfs/348 b/tests/xfs/348
> new file mode 100755
> index 0000000..9ec696a
> --- /dev/null
> +++ b/tests/xfs/348
> @@ -0,0 +1,128 @@
> +#! /bin/bash
> +# FSQA Test No. 348
> +#
> +# Test handling of invalid inode modes
> +#
> +# Set all possible file type values for different types of files
> +# and verify that xfs_repair detects the correct errors.
> +#
> +#-----------------------------------------------------------------------
> +#
> +# Copyright (C) 2016 CTERA Networks. All Rights Reserved.
> +# Author: Amir Goldstein <amir73il@gmail.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()
> +{
> +	rm -f $tmp.*
> +}
> +
> +# get standard environment, filters and checks
> +. ./common/rc
> +. ./common/filter
> +. ./common/repair
> +
> +# real QA test starts here
> +_supported_fs xfs
> +_supported_os Linux
> +_require_scratch
> +
> +rm -f $seqres.full
> +
> +_scratch_mkfs >>$seqres.full 2>&1
> +
> +_scratch_mount
> +
> +# Create our test files.
> +testdir=$SCRATCH_MNT/test
> +mkdir -p $testdir
> +mkdir $testdir/DIR
> +echo 123 > $testdir/FILE
> +ln -s $testdir/FILE $testdir/SYMLINK
> +mknod $testdir/CHRDEV c 1 1
> +mknod $testdir/BLKDEV b 1 1
> +mknod $testdir/FIFO p
> +
> +xfs_info $SCRATCH_MNT | grep -q "ftype=1" && FTYPE_FEATURE=1
> +
> +# Record test dir inode for xfs_repair filter
> +inode_filter=$tmp.sed
> +rm -f $inode_filter
> +pino=$(ls -id $testdir | awk '{print $1}')
> +echo "s/inode $pino/PARENT_INO/" >> $inode_filter
> +echo "s/directory $pino/directory PARENT_INO/" >> $inode_filter
> +
> +inodes=""
> +# Record inode numbers for xfs_db commands and xfs_repair filter
> +for f in DIR FILE SYMLINK CHRDEV BLKDEV FIFO; do
> +	ino=$(ls -id $testdir/$f | awk '{print $1}')
> +	inodes="$inodes $ino"
> +	echo "s/inode $ino/${f}_INO/" >> $inode_filter
> +done
> +
> +_scratch_unmount
> +
> +# Possible mode file type values (mode & S_IFMT) >> 12
> +dtypes="0 1 2 3 4 5 6 7 10 11 12 13 14 15 16 17"
> +echo "===== Find inode by file type:"
> +for dt in $dtypes; do
> +	# Set all our test files to dt value
> +	for ino in $inodes; do
> +		_scratch_xfs_db -c "inode $ino" -c "print core.mode" | \
> +			grep -q "0${dt}0...$" && \
> +			(echo "dt=$dt => inode $ino" | sed -f $inode_filter)
> +	done
> +done
> +for dt in $dtypes; do
> +	echo
> +	echo "===== Setting dt=$dt to all files:"
> +	# Set all our test files to dt value
> +	for ino in $inodes; do
> +		_scratch_xfs_db -x -c "inode $ino" -c "write core.mode 0${dt}0644"
> +	done
> +	# Repair should detect the inconsistencies
> +	# For invalid dt values, all files would have been junked.
> +	# For valid dt values, one test file is expected to be valid.
> +	# The rest would either have wrong format or non matching dir ftype.
> +	_scratch_xfs_repair -n 2>&1  | tee -a $seqres.full | \
> +		_filter_repair | grep "^would have junked" | sed -f $inode_filter | sort -u

I would also like to see a second test that scrambles the ftype field in
the directory entry (instead of changing the inode core.mode) but xfs_db
can't write to dir3 blocks because it doesn't know how to set the dir
block CRC.

I'm going to send a patch to add that as part of my xfsprogs 4.11
patchbomb (hopefully next week, but after the xfsprogs 4.9 release) so
that test can wait.

> +	# If ftype feature is enabled, when setting file type to one of the
> +	# special types (i.e. FIFO(1), CHRDEV(2),BLKDEV(6),SOCKET(14)),
> +	# xfs_repair is expected to detect ftype mismatch error. Otherewise,
> +	# xfs_repair is not expected to detect ftype mismatch error.
> +	if [ "$FTYPE_FEATURE" = 1 ] && (echo ':1:2:6:14:' | grep -q ":$dt:"); then
> +		_scratch_xfs_repair -n 2>&1 | grep -q "^would fix ftype mismatch" || \
> +			_fail "xfs_repair should fix ftype mismatch"
> +	else
> +		_scratch_xfs_repair -n 2>&1 | grep -q -v "^would fix ftype mismatch" || \
> +			_fail "xfs_repair should not fix ftype mismatch"

Just FYI the whole test will stop as soon as we hit a _fail.  Please
consider simply echoing a complaint to stdout so that the golden output
diff will catch this, and we can see all the failing cases.

--D

> +	fi
> +done
> +
> +# Repair should detect and junk all test files
> +_scratch_xfs_repair 2>&1 >> $seqres.full 2>&1 || _fail "xfs_repair should not fail"
> +
> +status=0
> +exit
> diff --git a/tests/xfs/348.out b/tests/xfs/348.out
> new file mode 100644
> index 0000000..bf4924b
> --- /dev/null
> +++ b/tests/xfs/348.out
> @@ -0,0 +1,216 @@
> +QA output created by 348
> +===== Find inode by file type:
> +dt=1 => FIFO_INO
> +dt=2 => CHRDEV_INO
> +dt=4 => DIR_INO
> +dt=6 => BLKDEV_INO
> +dt=10 => FILE_INO
> +dt=12 => SYMLINK_INO
> +
> +===== Setting dt=0 to all files:
> +core.mode = 0644
> +core.mode = 0644
> +core.mode = 0644
> +core.mode = 0644
> +core.mode = 0644
> +core.mode = 0644
> +would have junked entry "BLKDEV" in directory PARENT_INO
> +would have junked entry "CHRDEV" in directory PARENT_INO
> +would have junked entry "DIR" in directory PARENT_INO
> +would have junked entry "FIFO" in directory PARENT_INO
> +would have junked entry "FILE" in directory PARENT_INO
> +would have junked entry "SYMLINK" in directory PARENT_INO
> +
> +===== Setting dt=1 to all files:
> +core.mode = 010644
> +core.mode = 010644
> +core.mode = 010644
> +core.mode = 010644
> +core.mode = 010644
> +core.mode = 010644
> +would have junked entry "DIR" in directory PARENT_INO
> +would have junked entry "FILE" in directory PARENT_INO
> +would have junked entry "SYMLINK" in directory PARENT_INO
> +
> +===== Setting dt=2 to all files:
> +core.mode = 020644
> +core.mode = 020644
> +core.mode = 020644
> +core.mode = 020644
> +core.mode = 020644
> +core.mode = 020644
> +would have junked entry "DIR" in directory PARENT_INO
> +would have junked entry "FILE" in directory PARENT_INO
> +would have junked entry "SYMLINK" in directory PARENT_INO
> +
> +===== Setting dt=3 to all files:
> +core.mode = 030644
> +core.mode = 030644
> +core.mode = 030644
> +core.mode = 030644
> +core.mode = 030644
> +core.mode = 030644
> +would have junked entry "BLKDEV" in directory PARENT_INO
> +would have junked entry "CHRDEV" in directory PARENT_INO
> +would have junked entry "DIR" in directory PARENT_INO
> +would have junked entry "FIFO" in directory PARENT_INO
> +would have junked entry "FILE" in directory PARENT_INO
> +would have junked entry "SYMLINK" in directory PARENT_INO
> +
> +===== Setting dt=4 to all files:
> +core.mode = 040644
> +core.mode = 040644
> +core.mode = 040644
> +core.mode = 040644
> +core.mode = 040644
> +core.mode = 040644
> +would have junked entry "BLKDEV" in directory PARENT_INO
> +would have junked entry "CHRDEV" in directory PARENT_INO
> +would have junked entry "FIFO" in directory PARENT_INO
> +would have junked entry "FILE" in directory PARENT_INO
> +
> +===== Setting dt=5 to all files:
> +core.mode = 050644
> +core.mode = 050644
> +core.mode = 050644
> +core.mode = 050644
> +core.mode = 050644
> +core.mode = 050644
> +would have junked entry "BLKDEV" in directory PARENT_INO
> +would have junked entry "CHRDEV" in directory PARENT_INO
> +would have junked entry "DIR" in directory PARENT_INO
> +would have junked entry "FIFO" in directory PARENT_INO
> +would have junked entry "FILE" in directory PARENT_INO
> +would have junked entry "SYMLINK" in directory PARENT_INO
> +
> +===== Setting dt=6 to all files:
> +core.mode = 060644
> +core.mode = 060644
> +core.mode = 060644
> +core.mode = 060644
> +core.mode = 060644
> +core.mode = 060644
> +would have junked entry "DIR" in directory PARENT_INO
> +would have junked entry "FILE" in directory PARENT_INO
> +would have junked entry "SYMLINK" in directory PARENT_INO
> +
> +===== Setting dt=7 to all files:
> +core.mode = 070644
> +core.mode = 070644
> +core.mode = 070644
> +core.mode = 070644
> +core.mode = 070644
> +core.mode = 070644
> +would have junked entry "BLKDEV" in directory PARENT_INO
> +would have junked entry "CHRDEV" in directory PARENT_INO
> +would have junked entry "DIR" in directory PARENT_INO
> +would have junked entry "FIFO" in directory PARENT_INO
> +would have junked entry "FILE" in directory PARENT_INO
> +would have junked entry "SYMLINK" in directory PARENT_INO
> +
> +===== Setting dt=10 to all files:
> +core.mode = 0100644
> +core.mode = 0100644
> +core.mode = 0100644
> +core.mode = 0100644
> +core.mode = 0100644
> +core.mode = 0100644
> +would have junked entry "BLKDEV" in directory PARENT_INO
> +would have junked entry "CHRDEV" in directory PARENT_INO
> +would have junked entry "DIR" in directory PARENT_INO
> +would have junked entry "FIFO" in directory PARENT_INO
> +would have junked entry "SYMLINK" in directory PARENT_INO
> +
> +===== Setting dt=11 to all files:
> +core.mode = 0110644
> +core.mode = 0110644
> +core.mode = 0110644
> +core.mode = 0110644
> +core.mode = 0110644
> +core.mode = 0110644
> +would have junked entry "BLKDEV" in directory PARENT_INO
> +would have junked entry "CHRDEV" in directory PARENT_INO
> +would have junked entry "DIR" in directory PARENT_INO
> +would have junked entry "FIFO" in directory PARENT_INO
> +would have junked entry "FILE" in directory PARENT_INO
> +would have junked entry "SYMLINK" in directory PARENT_INO
> +
> +===== Setting dt=12 to all files:
> +core.mode = 0120644
> +core.mode = 0120644
> +core.mode = 0120644
> +core.mode = 0120644
> +core.mode = 0120644
> +core.mode = 0120644
> +would have junked entry "BLKDEV" in directory PARENT_INO
> +would have junked entry "CHRDEV" in directory PARENT_INO
> +would have junked entry "DIR" in directory PARENT_INO
> +would have junked entry "FIFO" in directory PARENT_INO
> +would have junked entry "FILE" in directory PARENT_INO
> +
> +===== Setting dt=13 to all files:
> +core.mode = 0130644
> +core.mode = 0130644
> +core.mode = 0130644
> +core.mode = 0130644
> +core.mode = 0130644
> +core.mode = 0130644
> +would have junked entry "BLKDEV" in directory PARENT_INO
> +would have junked entry "CHRDEV" in directory PARENT_INO
> +would have junked entry "DIR" in directory PARENT_INO
> +would have junked entry "FIFO" in directory PARENT_INO
> +would have junked entry "FILE" in directory PARENT_INO
> +would have junked entry "SYMLINK" in directory PARENT_INO
> +
> +===== Setting dt=14 to all files:
> +core.mode = 0140644
> +core.mode = 0140644
> +core.mode = 0140644
> +core.mode = 0140644
> +core.mode = 0140644
> +core.mode = 0140644
> +would have junked entry "DIR" in directory PARENT_INO
> +would have junked entry "FILE" in directory PARENT_INO
> +would have junked entry "SYMLINK" in directory PARENT_INO
> +
> +===== Setting dt=15 to all files:
> +core.mode = 0150644
> +core.mode = 0150644
> +core.mode = 0150644
> +core.mode = 0150644
> +core.mode = 0150644
> +core.mode = 0150644
> +would have junked entry "BLKDEV" in directory PARENT_INO
> +would have junked entry "CHRDEV" in directory PARENT_INO
> +would have junked entry "DIR" in directory PARENT_INO
> +would have junked entry "FIFO" in directory PARENT_INO
> +would have junked entry "FILE" in directory PARENT_INO
> +would have junked entry "SYMLINK" in directory PARENT_INO
> +
> +===== Setting dt=16 to all files:
> +core.mode = 0160644
> +core.mode = 0160644
> +core.mode = 0160644
> +core.mode = 0160644
> +core.mode = 0160644
> +core.mode = 0160644
> +would have junked entry "BLKDEV" in directory PARENT_INO
> +would have junked entry "CHRDEV" in directory PARENT_INO
> +would have junked entry "DIR" in directory PARENT_INO
> +would have junked entry "FIFO" in directory PARENT_INO
> +would have junked entry "FILE" in directory PARENT_INO
> +would have junked entry "SYMLINK" in directory PARENT_INO
> +
> +===== Setting dt=17 to all files:
> +core.mode = 0170644
> +core.mode = 0170644
> +core.mode = 0170644
> +core.mode = 0170644
> +core.mode = 0170644
> +core.mode = 0170644
> +would have junked entry "BLKDEV" in directory PARENT_INO
> +would have junked entry "CHRDEV" in directory PARENT_INO
> +would have junked entry "DIR" in directory PARENT_INO
> +would have junked entry "FIFO" in directory PARENT_INO
> +would have junked entry "FILE" in directory PARENT_INO
> +would have junked entry "SYMLINK" in directory PARENT_INO
> diff --git a/tests/xfs/group b/tests/xfs/group
> index c237b50..68c2bb6 100644
> --- a/tests/xfs/group
> +++ b/tests/xfs/group
> @@ -334,3 +334,4 @@
>  345 auto quick clone
>  346 auto quick clone
>  347 auto quick clone
> +348 auto quick repair
> -- 
> 2.7.4
> 

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [PATCH v2 1/3] xfs/348: test handling of invalid inode modes
  2017-01-03 20:36   ` Darrick J. Wong
@ 2017-01-04 10:36     ` Amir Goldstein
  2017-01-04 20:33       ` Darrick J. Wong
  0 siblings, 1 reply; 10+ messages in thread
From: Amir Goldstein @ 2017-01-04 10:36 UTC (permalink / raw)
  To: Darrick J. Wong
  Cc: Eryu Guan, Christoph Hellwig, Brian Foster, fstests, linux-xfs

On Tue, Jan 3, 2017 at 10:36 PM, Darrick J. Wong
<darrick.wong@oracle.com> wrote:
> On Sun, Dec 25, 2016 at 08:09:34PM +0200, Amir Goldstein wrote:
>> Set all possible file type values for different types of files
>> and verify that xfs_repair detects the correct errors.
>>
>> When setting invalid file type values (e.g. core.mode = 0170644),
>> all files are expected to have been junked by xfs_repair.
>>
>> When setting valid file type values to non matching file types,
>> xfs_repair would either detect wrong format and junk the file, e.g.:
>>   would have junked entry "FILE" in directory PARENT_INO
>> or detect a ftype mismatch error, e.g.:
>>   would fix ftype mismatch (5/3) in directory/child PARENT_INO/FIFO_INO
>>
>> If ftype feature is enabled, when setting file type to one of the
>> special types (i.e. FIFO(1), CHRDEV(2),BLKDEV(6),SOCKET(14)),
>> xfs_repair is expected to detect ftype mismatch error. Otherewise,
>
> "Otherwise"
>

thanks

> I would also like to see a second test that scrambles the ftype field in
> the directory entry (instead of changing the inode core.mode) but xfs_db
> can't write to dir3 blocks because it doesn't know how to set the dir
> block CRC.
>
> I'm going to send a patch to add that as part of my xfsprogs 4.11
> patchbomb (hopefully next week, but after the xfsprogs 4.9 release) so
> that test can wait.
>

Sure, I'll beef up the test when that code arrives.

>> +     # If ftype feature is enabled, when setting file type to one of the
>> +     # special types (i.e. FIFO(1), CHRDEV(2),BLKDEV(6),SOCKET(14)),
>> +     # xfs_repair is expected to detect ftype mismatch error. Otherewise,
>> +     # xfs_repair is not expected to detect ftype mismatch error.
>> +     if [ "$FTYPE_FEATURE" = 1 ] && (echo ':1:2:6:14:' | grep -q ":$dt:"); then
>> +             _scratch_xfs_repair -n 2>&1 | grep -q "^would fix ftype mismatch" || \
>> +                     _fail "xfs_repair should fix ftype mismatch"
>> +     else
>> +             _scratch_xfs_repair -n 2>&1 | grep -q -v "^would fix ftype mismatch" || \
>> +                     _fail "xfs_repair should not fix ftype mismatch"
>
> Just FYI the whole test will stop as soon as we hit a _fail.  Please
> consider simply echoing a complaint to stdout so that the golden output
> diff will catch this, and we can see all the failing cases.
>

Will do.

Can you please reply to v2 patch series.
It contain 2 more patches beyond this one with additional testing of access
to mounted fs with the malformed inodes.

My question is what to do WRT patch 3/3 which triggers an XFS Assert in the
kernel (readdir of a phony directory that is really a regular file or symlink).
http://www.spinics.net/lists/linux-xfs/msg03059.html

Is this something of concern to you? Can you instruct me whether the assertion
should be fixed or propose a fix yourself?
Or is it just a non issue and we should not add patch 3/3 to this test.

Amir.

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [PATCH v2 1/3] xfs/348: test handling of invalid inode modes
  2017-01-04 10:36     ` Amir Goldstein
@ 2017-01-04 20:33       ` Darrick J. Wong
  0 siblings, 0 replies; 10+ messages in thread
From: Darrick J. Wong @ 2017-01-04 20:33 UTC (permalink / raw)
  To: Amir Goldstein
  Cc: Eryu Guan, Christoph Hellwig, Brian Foster, fstests, linux-xfs

On Wed, Jan 04, 2017 at 12:36:24PM +0200, Amir Goldstein wrote:
> On Tue, Jan 3, 2017 at 10:36 PM, Darrick J. Wong
> <darrick.wong@oracle.com> wrote:
> > On Sun, Dec 25, 2016 at 08:09:34PM +0200, Amir Goldstein wrote:
> >> Set all possible file type values for different types of files
> >> and verify that xfs_repair detects the correct errors.
> >>
> >> When setting invalid file type values (e.g. core.mode = 0170644),
> >> all files are expected to have been junked by xfs_repair.
> >>
> >> When setting valid file type values to non matching file types,
> >> xfs_repair would either detect wrong format and junk the file, e.g.:
> >>   would have junked entry "FILE" in directory PARENT_INO
> >> or detect a ftype mismatch error, e.g.:
> >>   would fix ftype mismatch (5/3) in directory/child PARENT_INO/FIFO_INO
> >>
> >> If ftype feature is enabled, when setting file type to one of the
> >> special types (i.e. FIFO(1), CHRDEV(2),BLKDEV(6),SOCKET(14)),
> >> xfs_repair is expected to detect ftype mismatch error. Otherewise,
> >
> > "Otherwise"
> >
> 
> thanks
> 
> > I would also like to see a second test that scrambles the ftype field in
> > the directory entry (instead of changing the inode core.mode) but xfs_db
> > can't write to dir3 blocks because it doesn't know how to set the dir
> > block CRC.
> >
> > I'm going to send a patch to add that as part of my xfsprogs 4.11
> > patchbomb (hopefully next week, but after the xfsprogs 4.9 release) so
> > that test can wait.
> >
> 
> Sure, I'll beef up the test when that code arrives.
> 
> >> +     # If ftype feature is enabled, when setting file type to one of the
> >> +     # special types (i.e. FIFO(1), CHRDEV(2),BLKDEV(6),SOCKET(14)),
> >> +     # xfs_repair is expected to detect ftype mismatch error. Otherewise,
> >> +     # xfs_repair is not expected to detect ftype mismatch error.
> >> +     if [ "$FTYPE_FEATURE" = 1 ] && (echo ':1:2:6:14:' | grep -q ":$dt:"); then
> >> +             _scratch_xfs_repair -n 2>&1 | grep -q "^would fix ftype mismatch" || \
> >> +                     _fail "xfs_repair should fix ftype mismatch"
> >> +     else
> >> +             _scratch_xfs_repair -n 2>&1 | grep -q -v "^would fix ftype mismatch" || \
> >> +                     _fail "xfs_repair should not fix ftype mismatch"
> >
> > Just FYI the whole test will stop as soon as we hit a _fail.  Please
> > consider simply echoing a complaint to stdout so that the golden output
> > diff will catch this, and we can see all the failing cases.
> >
> 
> Will do.
> 
> Can you please reply to v2 patch series.
> It contain 2 more patches beyond this one with additional testing of access
> to mounted fs with the malformed inodes.
> 
> My question is what to do WRT patch 3/3 which triggers an XFS Assert in the
> kernel (readdir of a phony directory that is really a regular file or symlink).
> http://www.spinics.net/lists/linux-xfs/msg03059.html
> 
> Is this something of concern to you? Can you instruct me whether the assertion
> should be fixed or propose a fix yourself?
> Or is it just a non issue and we should not add patch 3/3 to this test.

ASSERTs by definition halt the system, so it's only a crasher bug if
CONFIG_XFS_DEBUG=y.  However, your test writes garbage to the
filesystem, so we must try to avoid interacting with corrupted junk when
possible.

I suggest bouncing corruption out to userspace by changing that ASSERT to:

if (rval == 0 || args->dp->i_d.di_size == args->geo->blksize)
	return -EFSCORRUPTED;

...and perhaps following it up with a check in _dinode_verify to reject
the inode if it has an obviously bogus i_size.

--D

> 
> Amir.
> --
> To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [PATCH v2 0/3] xfs/348: test handling of malformed inode mode
  2017-01-03  5:52 ` [PATCH v2 0/3] xfs/348: test handling of " Eryu Guan
@ 2017-01-11  6:45   ` Amir Goldstein
  2017-01-11  9:08     ` Eryu Guan
  0 siblings, 1 reply; 10+ messages in thread
From: Amir Goldstein @ 2017-01-11  6:45 UTC (permalink / raw)
  To: Eryu Guan
  Cc: Darrick J . Wong, Christoph Hellwig, Brian Foster, fstests, linux-xfs

On Tue, Jan 3, 2017 at 7:52 AM, Eryu Guan <eguan@redhat.com> wrote:
> On Sun, Dec 25, 2016 at 08:09:33PM +0200, Amir Goldstein wrote:
>> Eryu,
>>
>> I beefed up the initial test sent earlier today with more checks
>> on mounted fs.
>>
>> The original patch testing only xfs_repair remains patch 1 in this series.
>>
>> Patch 2 adds fstat tests on mounted fs, which are safe on my test system.
>>
>> Patch 3 is explosive. It exposes an XFS assert, but I left a "safety pin"
>> that needs to be commented out to reproduce the assert.
>
> My kenrel config doesn't turn DEBUG on, so I don't see a kernel crash :)
>
> I can push this test out after the fix lands in upstream, then I think
> there's no need to leave a switch in the test.
>
> Otherwise tests look good to me. But I'd like to have Darrick to review
> too, as he had written many fuzzer tests and suggested this test :)
>

Eryu,

Darrick has actually reviewed patch 1 v1 and gave only minor comments,
which I addressed in v3.

So how about merging patch 1 (v3) (sanity of xfs_repair) and leave patches 2-3
for later?

Amir.

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [PATCH v2 0/3] xfs/348: test handling of malformed inode mode
  2017-01-11  6:45   ` Amir Goldstein
@ 2017-01-11  9:08     ` Eryu Guan
  0 siblings, 0 replies; 10+ messages in thread
From: Eryu Guan @ 2017-01-11  9:08 UTC (permalink / raw)
  To: Amir Goldstein
  Cc: Darrick J . Wong, Christoph Hellwig, Brian Foster, fstests, linux-xfs

On Wed, Jan 11, 2017 at 08:45:34AM +0200, Amir Goldstein wrote:
> On Tue, Jan 3, 2017 at 7:52 AM, Eryu Guan <eguan@redhat.com> wrote:
> > On Sun, Dec 25, 2016 at 08:09:33PM +0200, Amir Goldstein wrote:
> >> Eryu,
> >>
> >> I beefed up the initial test sent earlier today with more checks
> >> on mounted fs.
> >>
> >> The original patch testing only xfs_repair remains patch 1 in this series.
> >>
> >> Patch 2 adds fstat tests on mounted fs, which are safe on my test system.
> >>
> >> Patch 3 is explosive. It exposes an XFS assert, but I left a "safety pin"
> >> that needs to be commented out to reproduce the assert.
> >
> > My kenrel config doesn't turn DEBUG on, so I don't see a kernel crash :)
> >
> > I can push this test out after the fix lands in upstream, then I think
> > there's no need to leave a switch in the test.
> >
> > Otherwise tests look good to me. But I'd like to have Darrick to review
> > too, as he had written many fuzzer tests and suggested this test :)
> >
> 
> Eryu,
> 
> Darrick has actually reviewed patch 1 v1 and gave only minor comments,
> which I addressed in v3.
> 
> So how about merging patch 1 (v3) (sanity of xfs_repair) and leave patches 2-3
> for later?

Yeah, I was thinking about this too, will give it a quick test and queue
it up if there's no issues found. Sorry I'm a bit late reviewing these
patches, I was busy with other work these days.

Thanks,
Eryu

^ permalink raw reply	[flat|nested] 10+ messages in thread

end of thread, other threads:[~2017-01-11  9:08 UTC | newest]

Thread overview: 10+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2016-12-25 18:09 [PATCH v2 0/3] xfs/348: test handling of malformed inode mode Amir Goldstein
2016-12-25 18:09 ` [PATCH v2 1/3] xfs/348: test handling of invalid inode modes Amir Goldstein
2017-01-03 20:36   ` Darrick J. Wong
2017-01-04 10:36     ` Amir Goldstein
2017-01-04 20:33       ` Darrick J. Wong
2016-12-25 18:09 ` [PATCH v2 2/3] xfs/348: test fstat with malformed inode mode Amir Goldstein
2016-12-25 18:09 ` [PATCH v2 3/3] xfs/348: test readlink/readdir " Amir Goldstein
2017-01-03  5:52 ` [PATCH v2 0/3] xfs/348: test handling of " Eryu Guan
2017-01-11  6:45   ` Amir Goldstein
2017-01-11  9:08     ` Eryu Guan

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.