Skip to contentSkip to conversation

overlay: Test constant d_ino feature

PATCH Superseded Chandan Rajendra 2 participants 4 messages Aug 31, 2017

v2 of 5 Superseded Latest is v5 →

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 overlay: Test constant d_ino feature

This commit adds a test to verify constant d_ino feature. The following scenarios are checked,
- Parent's (i.e. "..") d_ino must always be calculated because a pure dir can be residing inside a merged dir.
- d_ino for "." must always be calculated because the present directory can have a copy-up origin.
- Verify d_ino values corresponding to "." and ".." entries of a pure lower dir.
- Verify d_ino of ".." entry of a merged dir.
- Verify pure lower residing in dir which has another lower layer pure lower residing in dir which has another lower layer.

---
Changelog:
v1->v2:
1. Removed tests copied over from overlay/017.
2. Removed src/t_dir_ino.c.
3. Add code for testing,
- d_ino of . and .. entries of pure_lower_dir.
- d_ino of .. entry of merged_dir.
- Verify pure lower residing in dir which has another lower layer. Thanks to Amir for suggesting these tests.

 tests/overlay/037     | 144 ++++++++++++++++++++++++++++++++++++++++++++++++++
 tests/overlay/037.out |   2 +
 tests/overlay/group   |   1 +
 3 files changed, 147 insertions(+)
 create mode 100755 tests/overlay/037
 create mode 100644 tests/overlay/037.out
tests/overlay/037 new file +144 -0
new file mode 100755
index 0000000..225d2b9
@@ -0,0 +1,144 @@
+#! /bin/bash
+# FSQA Test No. 037
+#
+# Test constant d_ino numbers
+#
+#-----------------------------------------------------------------------
+#
+# Copyright (C) 2017 IBM Corporation. All Rights Reserved.
+# Author: Chandan Rajendra <[email protected]>
+#
+# 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 -f $tmp.*
+}
+
+# get standard environment, filters and checks
+. ./common/rc
+. ./common/filter
+
+# real QA test starts here
+_supported_fs overlay
+_supported_os Linux
+_require_scratch
+_require_test_program "af_unix"
+_require_test_program "t_dir_type"
+
+rm -f $seqres.full
+
+_scratch_mkfs >>$seqres.full 2>&1
+
+# Create our test files.
+lowerdir=$OVL_BASE_SCRATCH_MNT/$OVL_LOWER
+mkdir -p $lowerdir
+
+pure_lower_dir=$lowerdir/pure_lower_dir
+impure_dir=$lowerdir/impure_dir
+merged_dir=$lowerdir/merged_dir
+
+mkdir $pure_lower_dir
+mkdir $impure_dir
+mkdir $merged_dir
+
+impure_dir_st_ino=$(stat -c '%i' $impure_dir)
+pure_lower_dir_st_ino=$(stat -c '%i' $pure_lower_dir)
+
+_scratch_mount
+
+impure_dir=$SCRATCH_MNT/impure_dir
+pure_upper_dir=$impure_dir/pure_upper_dir
+
+mkdir -p $pure_upper_dir
+
+# Parent's (i.e. "..") d_ino must always be calculated because a pure
+# dir can be residing inside a merged dir.
+parent_d=$($here/src/t_dir_type $pure_upper_dir $impure_dir_st_ino)
+[[ $parent_d != ".. d" ]] && \
+	echo "Pure dir inside a merged dir: Invalid d_ino reported for .."
+
+# d_ino for "." must always be calculated because the present
+# directory can have a copy-up origin.
+current_d=$($here/src/t_dir_type $impure_dir $impure_dir_st_ino)
+[[ $current_d != ". d" ]] && echo "Impure dir: Invalid d_ino reported for ."
+
+# Verify d_ino values corresponding to "." and ".." entries of a
+# pure lower dir.
+ovl_root_st_ino=$(stat -c '%i' $SCRATCH_MNT/)
+pure_lower_dir=$SCRATCH_MNT/pure_lower_dir
+
+parent_d=$($here/src/t_dir_type $pure_lower_dir $ovl_root_st_ino)
+[[ $parent_d != ".. d" ]] && echo "Pure lower dir: Invalid d_ino reported for .."
+
+current_d=$($here/src/t_dir_type $pure_lower_dir $pure_lower_dir_st_ino)
+[[ $current_d != ". d" ]] && echo "Pure lower dir: Invalid d_ino reported for ."
+
+# Verify d_ino of ".." entry of a merged dir.
+merged_dir=$SCRATCH_MNT/merged_dir
+touch $merged_dir/file.bin
+
+parent_d=$($here/src/t_dir_type $merged_dir $ovl_root_st_ino)
+[[ $parent_d != ".. d" ]] && echo "Merged dir: Invalid d_ino reported for .."
+
+_scratch_unmount
+
+# Verify pure lower residing in dir which has another lower layer
+_scratch_mkfs
+
+OVL_LOWER0=${OVL_LOWER}0
+OVL_LOWER1=${OVL_LOWER}1
+
+mkdir -p $OVL_BASE_SCRATCH_MNT/$OVL_LOWER0
+mkdir -p $OVL_BASE_SCRATCH_MNT/$OVL_LOWER1
+mkdir -p $OVL_BASE_SCRATCH_MNT/$OVL_UPPER
+mkdir -p $OVL_BASE_SCRATCH_MNT/$OVL_WORK
+
+lowerdir0=$OVL_BASE_SCRATCH_MNT/$OVL_LOWER0
+lowerdir1=$OVL_BASE_SCRATCH_MNT/$OVL_LOWER1
+upperdir=$OVL_BASE_SCRATCH_MNT/$OVL_UPPER
+workdir=$OVL_BASE_SCRATCH_MNT/$OVL_WORK
+
+mkdir -p $lowerdir0/dir0
+mkdir -p $lowerdir1/dir0/dir1
+
+dir0_st_ino=$(stat -c '%i' $lowerdir0/dir0)
+
+$MOUNT_PROG -t overlay overlay -o lowerdir=$lowerdir0:$lowerdir1 \
+	    -o upperdir=$upperdir -o workdir=$workdir $SCRATCH_MNT
+
+# Copy-up dir0
+touch $SCRATCH_MNT/dir0/file.bin
+
+parent_d=$($here/src/t_dir_type $SCRATCH_MNT/dir0/dir1 $dir0_st_ino)
+[[ $parent_d != ".. d" ]] && \
+	echo "Pure lower in dir which has another lower layer: Invalid d_ino reported for .."
+
+$UMOUNT_PROG $SCRATCH_MNT
+
+echo "Silence is golden"
+status=0
+exit
tests/overlay/037.out new file +2 -0
new file mode 100644
index 0000000..5c3a30a
@@ -0,0 +1,2 @@
+QA output created by 037
+Silence is golden
tests/overlay/group +1 -0
index 4cc1d74..b2460e1 100644
@@ -39,3 +39,4 @@
 034 auto quick copyup hardlink
 035 auto quick mount
 036 auto quick mount
+037 auto quick copyup
Signature · 1 line
2.9.5
#1 lore raw
Quoted 13 lines · Chandan Rajendra

This commit adds a test to verify constant d_ino feature. The following scenarios are checked,
- Parent's (i.e. "..") d_ino must always be calculated because a pure dir can be residing inside a merged dir.
- d_ino for "." must always be calculated because the present directory can have a copy-up origin.
- Verify d_ino values corresponding to "." and ".." entries of a pure lower dir.
- Verify d_ino of ".." entry of a merged dir.
- Verify pure lower residing in dir which has another lower layer pure lower residing in dir which has another lower layer.

Quoted 10 lines

---
Changelog:
v1->v2:
1. Removed tests copied over from overlay/017.
2. Removed src/t_dir_ino.c.
3. Add code for testing,
- d_ino of . and .. entries of pure_lower_dir.
- d_ino of .. entry of merged_dir.
- Verify pure lower residing in dir which has another lower layer. Thanks to Amir for suggesting these tests.

Nice work!
See more comments below.

Quoted 65 lines
 tests/overlay/037     | 144 ++++++++++++++++++++++++++++++++++++++++++++++++++
 tests/overlay/037.out |   2 +
 tests/overlay/group   |   1 +
 3 files changed, 147 insertions(+)
 create mode 100755 tests/overlay/037
 create mode 100644 tests/overlay/037.out
tests/overlay/037 new file +51 -0
new file mode 100755
index 0000000..225d2b9
@@ -0,0 +1,144 @@
+#! /bin/bash
+# FSQA Test No. 037
+#
+# Test constant d_ino numbers
+#
+#-----------------------------------------------------------------------
+#
+# Copyright (C) 2017 IBM Corporation. All Rights Reserved.
+# Author: Chandan Rajendra <[email protected]>
+#
+# 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 -f $tmp.*
+}
+
+# get standard environment, filters and checks
+. ./common/rc
+. ./common/filter
+
+# real QA test starts here
+_supported_fs overlay
+_supported_os Linux
+_require_scratch
+_require_test_program "af_unix"
+_require_test_program "t_dir_type"
+

tests/overlay/037 line 61

Quoted 10 lines
tests/overlay/037 new file +10 -0
+rm -f $seqres.full
+
+_scratch_mkfs >>$seqres.full 2>&1
+
+# Create our test files.
+lowerdir=$OVL_BASE_SCRATCH_MNT/$OVL_LOWER
+mkdir -p $lowerdir
+
+pure_lower_dir=$lowerdir/pure_lower_dir
+impure_dir=$lowerdir/impure_dir

You name this dir "impure", but I don't see anything impure about it in the test.
Maybe it is a good idea to repeat the d_ino tests on impure_dir before and after dir becomes impure.
The readdir code logic is really different depending on dir impure xattr. although it should not affect results of d_ino for . and .. its a good idea to verify that.
Besides, its good to verify that readdir cache is correctly invalidated when pure
dir becomes impure and it would be also good to verify that cache is invalidated when impure dir becomes pure again (bonus for checking that impure xattr is removed)

tests/overlay/037 line 70

Quoted 9 lines
tests/overlay/037 new file +9 -0
+merged_dir=$lowerdir/merged_dir
+
+mkdir $pure_lower_dir
+mkdir $impure_dir
+mkdir $merged_dir
+
+impure_dir_st_ino=$(stat -c '%i' $impure_dir)
+pure_lower_dir_st_ino=$(stat -c '%i' $pure_lower_dir)
+

Its better that tests restrict themselves to verifying requirements. It is required that overlay d_ino == overlay st_ino.
It is NOT required that overlay d_ino == real st_ino and future implementation may change that, so please get all the values of $XXX_st_ino from stat on overlay mount of respective dir.

Quoted 11 lines
tests/overlay/037 new file +11 -0
+_scratch_mount
+
+impure_dir=$SCRATCH_MNT/impure_dir
+pure_upper_dir=$impure_dir/pure_upper_dir
+
+mkdir -p $pure_upper_dir
+
+# Parent's (i.e. "..") d_ino must always be calculated because a pure
+# dir can be residing inside a merged dir.
+parent_d=$($here/src/t_dir_type $pure_upper_dir $impure_dir_st_ino)
+[[ $parent_d != ".. d" ]] && \

tests/overlay/037 line 91

Quoted 10 lines
tests/overlay/037 new file +10 -0
+       echo "Pure dir inside a merged dir: Invalid d_ino reported for .."
+
+# d_ino for "." must always be calculated because the present
+# directory can have a copy-up origin.
+current_d=$($here/src/t_dir_type $impure_dir $impure_dir_st_ino)
+[[ $current_d != ". d" ]] && echo "Impure dir: Invalid d_ino reported for ."
+
+# Verify d_ino values corresponding to "." and ".." entries of a
+# pure lower dir.
+ovl_root_st_ino=$(stat -c '%i' $SCRATCH_MNT/)

It's a bit bold to assume that root inode behaves like any other inode. Code may change in the future to return constant ino (2) for root like many fs do. It would be less controversial for the test to work with a normal dir, so please add another level of dir like you did in multi lower test below. Also, not a must, but it would have been nicer to read the test if both test setups (multi and single lower) actually used the same dir structure.

Quoted 30 lines
tests/overlay/037 new file +30 -0
+pure_lower_dir=$SCRATCH_MNT/pure_lower_dir
+
+parent_d=$($here/src/t_dir_type $pure_lower_dir $ovl_root_st_ino)
+[[ $parent_d != ".. d" ]] && echo "Pure lower dir: Invalid d_ino reported for .."
+
+current_d=$($here/src/t_dir_type $pure_lower_dir $pure_lower_dir_st_ino)
+[[ $current_d != ". d" ]] && echo "Pure lower dir: Invalid d_ino reported for ."
+
+# Verify d_ino of ".." entry of a merged dir.
+merged_dir=$SCRATCH_MNT/merged_dir
+touch $merged_dir/file.bin
+
+parent_d=$($here/src/t_dir_type $merged_dir $ovl_root_st_ino)
+[[ $parent_d != ".. d" ]] && echo "Merged dir: Invalid d_ino reported for .."
+
+_scratch_unmount
+
+# Verify pure lower residing in dir which has another lower layer
+_scratch_mkfs
+
+OVL_LOWER0=${OVL_LOWER}0
+OVL_LOWER1=${OVL_LOWER}1
+
+mkdir -p $OVL_BASE_SCRATCH_MNT/$OVL_LOWER0
+mkdir -p $OVL_BASE_SCRATCH_MNT/$OVL_LOWER1
+mkdir -p $OVL_BASE_SCRATCH_MNT/$OVL_UPPER
+mkdir -p $OVL_BASE_SCRATCH_MNT/$OVL_WORK
+
+lowerdir0=$OVL_BASE_SCRATCH_MNT/$OVL_LOWER0
+lowerdir1=$OVL_BASE_SCRATCH_MNT/$OVL_LOWER1

tests/overlay/037 line 131

Quoted 10 lines
tests/overlay/037 new file +10 -0
+upperdir=$OVL_BASE_SCRATCH_MNT/$OVL_UPPER
+workdir=$OVL_BASE_SCRATCH_MNT/$OVL_WORK
+
+mkdir -p $lowerdir0/dir0
+mkdir -p $lowerdir1/dir0/dir1
+
+dir0_st_ino=$(stat -c '%i' $lowerdir0/dir0)
+
+$MOUNT_PROG -t overlay overlay -o lowerdir=$lowerdir0:$lowerdir1 \
+           -o upperdir=$upperdir -o workdir=$workdir $SCRATCH_MNT

I find it confusing to understand (an already confusing setup) when lower0 is the middle and not lowermost layer.
Myabe best to describe them as $lowerdir and $middir for clarity.

Quoted 33 lines
tests/overlay/037 new file +13 -0
+
+# Copy-up dir0
+touch $SCRATCH_MNT/dir0/file.bin
+
+parent_d=$($here/src/t_dir_type $SCRATCH_MNT/dir0/dir1 $dir0_st_ino)
+[[ $parent_d != ".. d" ]] && \
+       echo "Pure lower in dir which has another lower layer: Invalid d_ino reported for .."
+
+$UMOUNT_PROG $SCRATCH_MNT
+
+echo "Silence is golden"
+status=0
+exit
tests/overlay/037.out new file +2 -0
new file mode 100644
index 0000000..5c3a30a
@@ -0,0 +1,2 @@
+QA output created by 037
+Silence is golden
tests/overlay/group +1 -1
index 4cc1d74..b2460e1 100644
@@ -39,3 +39,4 @@
 034 auto quick copyup hardlink
 035 auto quick mount
 036 auto quick mount
+037 auto quick copyup
--

2.9.5

#2 lore raw
Quoted 109 lines · IST Amir Goldstein
Quoted 13 lines · Chandan Rajendra

This commit adds a test to verify constant d_ino feature. The following scenarios are checked,
- Parent's (i.e. "..") d_ino must always be calculated because a pure dir can be residing inside a merged dir.
- d_ino for "." must always be calculated because the present directory can have a copy-up origin.
- Verify d_ino values corresponding to "." and ".." entries of a pure lower dir.
- Verify d_ino of ".." entry of a merged dir.
- Verify pure lower residing in dir which has another lower layer pure lower residing in dir which has another lower layer.

Quoted 10 lines

---
Changelog:
v1->v2:
1. Removed tests copied over from overlay/017.
2. Removed src/t_dir_ino.c.
3. Add code for testing,
- d_ino of . and .. entries of pure_lower_dir.
- d_ino of .. entry of merged_dir.
- Verify pure lower residing in dir which has another lower layer. Thanks to Amir for suggesting these tests.

Nice work!
See more comments below.

Quoted 65 lines
 tests/overlay/037     | 144 ++++++++++++++++++++++++++++++++++++++++++++++++++
 tests/overlay/037.out |   2 +
 tests/overlay/group   |   1 +
 3 files changed, 147 insertions(+)
 create mode 100755 tests/overlay/037
 create mode 100644 tests/overlay/037.out
tests/overlay/037 new file +51 -0
new file mode 100755
index 0000000..225d2b9
@@ -0,0 +1,144 @@
+#! /bin/bash
+# FSQA Test No. 037
+#
+# Test constant d_ino numbers
+#
+#-----------------------------------------------------------------------
+#
+# Copyright (C) 2017 IBM Corporation. All Rights Reserved.
+# Author: Chandan Rajendra <[email protected]>
+#
+# 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 -f $tmp.*
+}
+
+# get standard environment, filters and checks
+. ./common/rc
+. ./common/filter
+
+# real QA test starts here
+_supported_fs overlay
+_supported_os Linux
+_require_scratch
+_require_test_program "af_unix"
+_require_test_program "t_dir_type"
+

tests/overlay/037 line 61

Quoted 10 lines
tests/overlay/037 new file +10 -0
+rm -f $seqres.full
+
+_scratch_mkfs >>$seqres.full 2>&1
+
+# Create our test files.
+lowerdir=$OVL_BASE_SCRATCH_MNT/$OVL_LOWER
+mkdir -p $lowerdir
+
+pure_lower_dir=$lowerdir/pure_lower_dir
+impure_dir=$lowerdir/impure_dir

You name this dir "impure", but I don't see anything impure about it in the test.
Maybe it is a good idea to repeat the d_ino tests on impure_dir before and after dir becomes impure.

Quoted 8 lines

The readdir code logic is really different depending on dir impure xattr. although it should not affect results of d_ino for . and .. its a good idea to verify that.
Besides, its good to verify that readdir cache is correctly invalidated when pure
dir becomes impure and it would be also good to verify that cache is invalidated when impure dir becomes pure again (bonus for checking that impure xattr is removed)

Amir, Since overlayfs does not maintain a readdir cache for pure directories, I guess you wanted the test to check for cache invalidation in the case of an impure dir becoming a pure dir.

Signature · 1 line
chandan
#3 lore raw
Quoted 119 lines · Chandan Rajendra
Quoted 109 lines · IST Amir Goldstein
Quoted 13 lines · Chandan Rajendra

This commit adds a test to verify constant d_ino feature. The following scenarios are checked,
- Parent's (i.e. "..") d_ino must always be calculated because a pure dir can be residing inside a merged dir.
- d_ino for "." must always be calculated because the present directory can have a copy-up origin.
- Verify d_ino values corresponding to "." and ".." entries of a pure lower dir.
- Verify d_ino of ".." entry of a merged dir.
- Verify pure lower residing in dir which has another lower layer pure lower residing in dir which has another lower layer.

Quoted 10 lines

---
Changelog:
v1->v2:
1. Removed tests copied over from overlay/017.
2. Removed src/t_dir_ino.c.
3. Add code for testing,
- d_ino of . and .. entries of pure_lower_dir.
- d_ino of .. entry of merged_dir.
- Verify pure lower residing in dir which has another lower layer. Thanks to Amir for suggesting these tests.

Nice work!
See more comments below.

Quoted 65 lines
 tests/overlay/037     | 144 ++++++++++++++++++++++++++++++++++++++++++++++++++
 tests/overlay/037.out |   2 +
 tests/overlay/group   |   1 +
 3 files changed, 147 insertions(+)
 create mode 100755 tests/overlay/037
 create mode 100644 tests/overlay/037.out
tests/overlay/037 new file +51 -0
new file mode 100755
index 0000000..225d2b9
@@ -0,0 +1,144 @@
+#! /bin/bash
+# FSQA Test No. 037
+#
+# Test constant d_ino numbers
+#
+#-----------------------------------------------------------------------
+#
+# Copyright (C) 2017 IBM Corporation. All Rights Reserved.
+# Author: Chandan Rajendra <[email protected]>
+#
+# 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 -f $tmp.*
+}
+
+# get standard environment, filters and checks
+. ./common/rc
+. ./common/filter
+
+# real QA test starts here
+_supported_fs overlay
+_supported_os Linux
+_require_scratch
+_require_test_program "af_unix"
+_require_test_program "t_dir_type"
+

tests/overlay/037 line 61

Quoted 10 lines
tests/overlay/037 new file +10 -0
+rm -f $seqres.full
+
+_scratch_mkfs >>$seqres.full 2>&1
+
+# Create our test files.
+lowerdir=$OVL_BASE_SCRATCH_MNT/$OVL_LOWER
+mkdir -p $lowerdir
+
+pure_lower_dir=$lowerdir/pure_lower_dir
+impure_dir=$lowerdir/impure_dir

You name this dir "impure", but I don't see anything impure about it in the test.
Maybe it is a good idea to repeat the d_ino tests on impure_dir before and after dir becomes impure.

Quoted 8 lines

The readdir code logic is really different depending on dir impure xattr. although it should not affect results of d_ino for . and .. its a good idea to verify that.
Besides, its good to verify that readdir cache is correctly invalidated when pure
dir becomes impure and it would be also good to verify that cache is invalidated when impure dir becomes pure again (bonus for checking that impure xattr is removed)

Quoted 4 lines

Amir, Since overlayfs does not maintain a readdir cache for pure directories, I guess you wanted the test to check for cache invalidation in the case of an impure dir becoming a pure dir.

Yes, you are right. I meant the new ovl_iterate_real() code handles both impure dirs and pure dirs with merge parent, so its good to check both cases and the transition between them.