2021-03-08 13:44:45

by Zorro Lang

[permalink] [raw]
Subject: [PATCH v2] xfstests: rename RENAME_WHITEOUT test on fs no enough sapce

This's a regression test for linux 6b4b8e6b4ad8 ("ext4: fix bug for
rename with RENAME_WHITEOUT"). Rename a file with RENAME_WHITEOUT
flag might cause corruption when there's not enough space to
complete this renaming operation.

Signed-off-by: Zorro Lang <[email protected]>
Signed-off-by: Sun Ke <[email protected]>
---

Hi,

Thanks for the reviewing from Eryu. V2 did below changes:
1) Import ./common/renameat2 and _require_renameat2 whiteout
2) Replace CHUNKS with NR_FILE
3) Reduce the number of test files from 64*64 to 4*64
4) Add to quick group

More details about the reviewing history, refer to:
https://patchwork.kernel.org/project/fstests/patch/[email protected]/

Thanks,
Zorro

tests/generic/626 | 74 +++++++++++++++++++++++++++++++++++++++++++
tests/generic/626.out | 2 ++
tests/generic/group | 1 +
3 files changed, 77 insertions(+)
create mode 100755 tests/generic/626
create mode 100644 tests/generic/626.out

diff --git a/tests/generic/626 b/tests/generic/626
new file mode 100755
index 00000000..1baa73f8
--- /dev/null
+++ b/tests/generic/626
@@ -0,0 +1,74 @@
+#! /bin/bash
+# SPDX-License-Identifier: GPL-2.0
+# Copyright (c) 2021 HUAWEI. All Rights Reserved.
+# Copyright (c) 2021 Red Hat Inc. All Rights Reserved.
+#
+# FS QA Test No. 626
+#
+# Test RENAME_WHITEOUT on filesystem without space to create one more inodes.
+# This is a regression test for kernel commit:
+# 6b4b8e6b4ad8 ("ext4: ext4: fix bug for rename with RENAME_WHITEOUT")
+#
+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
+. ./common/populate
+. ./common/renameat2
+
+# remove previous $seqres.full before test
+rm -f $seqres.full
+
+# real QA test starts here
+_supported_fs generic
+_require_scratch
+_require_renameat2 whiteout
+
+_scratch_mkfs_sized $((256 * 1024 * 1024)) >> $seqres.full 2>&1
+_scratch_mount
+
+# Create lots of files, to help to trigger the bug easily
+NR_FILE=$((4 * 64))
+for ((i=0; i<NR_FILE; i++));do
+ touch $SCRATCH_MNT/srcfile$i
+done
+# Try to fill the whole fs
+nr_free=$(stat -f -c '%f' $SCRATCH_MNT)
+blksz="$(_get_block_size $SCRATCH_MNT)"
+_fill_fs $((nr_free * blksz)) $SCRATCH_MNT/fill_space $blksz 0 >> $seqres.full 2>&1
+# Use empty files to fill the rest
+for ((i=0; i<10000; i++));do
+ touch $SCRATCH_MNT/fill_file$i 2>/dev/null
+ # Until no more files can be created
+ if [ $? -ne 0 ];then
+ break
+ fi
+done
+# ENOSPC is expected here
+for ((i=0; i<NR_FILE; i++));do
+ $here/src/renameat2 -w $SCRATCH_MNT/srcfile$i $SCRATCH_MNT/dstfile$i >> $seqres.full 2>&1
+done
+_scratch_cycle_mount
+# Expect no errors at here
+for ((i=0; i<NR_FILE; i++));do
+ ls -l $SCRATCH_MNT/srcfile$i >/dev/null
+done
+
+echo "Silence is golden"
+# success, all done
+status=0
+exit
diff --git a/tests/generic/626.out b/tests/generic/626.out
new file mode 100644
index 00000000..130b2fef
--- /dev/null
+++ b/tests/generic/626.out
@@ -0,0 +1,2 @@
+QA output created by 626
+Silence is golden
diff --git a/tests/generic/group b/tests/generic/group
index 84db3789..c3448fe3 100644
--- a/tests/generic/group
+++ b/tests/generic/group
@@ -628,3 +628,4 @@
623 auto quick shutdown
624 auto quick verity
625 auto quick verity
+626 auto quick rename enospc
--
2.29.2


2021-03-08 18:37:26

by Darrick J. Wong

[permalink] [raw]
Subject: Re: [PATCH v2] xfstests: rename RENAME_WHITEOUT test on fs no enough sapce

On Mon, Mar 08, 2021 at 09:43:27PM +0800, Zorro Lang wrote:
> This's a regression test for linux 6b4b8e6b4ad8 ("ext4: fix bug for
> rename with RENAME_WHITEOUT"). Rename a file with RENAME_WHITEOUT
> flag might cause corruption when there's not enough space to
> complete this renaming operation.
>
> Signed-off-by: Zorro Lang <[email protected]>
> Signed-off-by: Sun Ke <[email protected]>

Seems reasonable to me; does it pass on xfs?

If so,
Reviewed-by: Darrick J. Wong <[email protected]>

--D

> ---
>
> Hi,
>
> Thanks for the reviewing from Eryu. V2 did below changes:
> 1) Import ./common/renameat2 and _require_renameat2 whiteout
> 2) Replace CHUNKS with NR_FILE
> 3) Reduce the number of test files from 64*64 to 4*64
> 4) Add to quick group
>
> More details about the reviewing history, refer to:
> https://patchwork.kernel.org/project/fstests/patch/[email protected]/
>
> Thanks,
> Zorro
>
> tests/generic/626 | 74 +++++++++++++++++++++++++++++++++++++++++++
> tests/generic/626.out | 2 ++
> tests/generic/group | 1 +
> 3 files changed, 77 insertions(+)
> create mode 100755 tests/generic/626
> create mode 100644 tests/generic/626.out
>
> diff --git a/tests/generic/626 b/tests/generic/626
> new file mode 100755
> index 00000000..1baa73f8
> --- /dev/null
> +++ b/tests/generic/626
> @@ -0,0 +1,74 @@
> +#! /bin/bash
> +# SPDX-License-Identifier: GPL-2.0
> +# Copyright (c) 2021 HUAWEI. All Rights Reserved.
> +# Copyright (c) 2021 Red Hat Inc. All Rights Reserved.
> +#
> +# FS QA Test No. 626
> +#
> +# Test RENAME_WHITEOUT on filesystem without space to create one more inodes.
> +# This is a regression test for kernel commit:
> +# 6b4b8e6b4ad8 ("ext4: ext4: fix bug for rename with RENAME_WHITEOUT")
> +#
> +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
> +. ./common/populate
> +. ./common/renameat2
> +
> +# remove previous $seqres.full before test
> +rm -f $seqres.full
> +
> +# real QA test starts here
> +_supported_fs generic
> +_require_scratch
> +_require_renameat2 whiteout
> +
> +_scratch_mkfs_sized $((256 * 1024 * 1024)) >> $seqres.full 2>&1
> +_scratch_mount
> +
> +# Create lots of files, to help to trigger the bug easily
> +NR_FILE=$((4 * 64))
> +for ((i=0; i<NR_FILE; i++));do
> + touch $SCRATCH_MNT/srcfile$i
> +done
> +# Try to fill the whole fs
> +nr_free=$(stat -f -c '%f' $SCRATCH_MNT)
> +blksz="$(_get_block_size $SCRATCH_MNT)"
> +_fill_fs $((nr_free * blksz)) $SCRATCH_MNT/fill_space $blksz 0 >> $seqres.full 2>&1
> +# Use empty files to fill the rest
> +for ((i=0; i<10000; i++));do
> + touch $SCRATCH_MNT/fill_file$i 2>/dev/null
> + # Until no more files can be created
> + if [ $? -ne 0 ];then
> + break
> + fi
> +done
> +# ENOSPC is expected here
> +for ((i=0; i<NR_FILE; i++));do
> + $here/src/renameat2 -w $SCRATCH_MNT/srcfile$i $SCRATCH_MNT/dstfile$i >> $seqres.full 2>&1
> +done
> +_scratch_cycle_mount
> +# Expect no errors at here
> +for ((i=0; i<NR_FILE; i++));do
> + ls -l $SCRATCH_MNT/srcfile$i >/dev/null
> +done
> +
> +echo "Silence is golden"
> +# success, all done
> +status=0
> +exit
> diff --git a/tests/generic/626.out b/tests/generic/626.out
> new file mode 100644
> index 00000000..130b2fef
> --- /dev/null
> +++ b/tests/generic/626.out
> @@ -0,0 +1,2 @@
> +QA output created by 626
> +Silence is golden
> diff --git a/tests/generic/group b/tests/generic/group
> index 84db3789..c3448fe3 100644
> --- a/tests/generic/group
> +++ b/tests/generic/group
> @@ -628,3 +628,4 @@
> 623 auto quick shutdown
> 624 auto quick verity
> 625 auto quick verity
> +626 auto quick rename enospc
> --
> 2.29.2
>

2021-03-09 02:23:39

by Zorro Lang

[permalink] [raw]
Subject: Re: [PATCH v2] xfstests: rename RENAME_WHITEOUT test on fs no enough sapce

On Mon, Mar 08, 2021 at 10:36:10AM -0800, Darrick J. Wong wrote:
> On Mon, Mar 08, 2021 at 09:43:27PM +0800, Zorro Lang wrote:
> > This's a regression test for linux 6b4b8e6b4ad8 ("ext4: fix bug for
> > rename with RENAME_WHITEOUT"). Rename a file with RENAME_WHITEOUT
> > flag might cause corruption when there's not enough space to
> > complete this renaming operation.
> >
> > Signed-off-by: Zorro Lang <[email protected]>
> > Signed-off-by: Sun Ke <[email protected]>
>
> Seems reasonable to me; does it pass on xfs?

Yes, it's passed on xfs and latest ext4(fixed this bug).

Thanks,
Zorro

>
> If so,
> Reviewed-by: Darrick J. Wong <[email protected]>
>
> --D
>
> > ---
> >
> > Hi,
> >
> > Thanks for the reviewing from Eryu. V2 did below changes:
> > 1) Import ./common/renameat2 and _require_renameat2 whiteout
> > 2) Replace CHUNKS with NR_FILE
> > 3) Reduce the number of test files from 64*64 to 4*64
> > 4) Add to quick group
> >
> > More details about the reviewing history, refer to:
> > https://patchwork.kernel.org/project/fstests/patch/[email protected]/
> >
> > Thanks,
> > Zorro
> >
> > tests/generic/626 | 74 +++++++++++++++++++++++++++++++++++++++++++
> > tests/generic/626.out | 2 ++
> > tests/generic/group | 1 +
> > 3 files changed, 77 insertions(+)
> > create mode 100755 tests/generic/626
> > create mode 100644 tests/generic/626.out
> >
> > diff --git a/tests/generic/626 b/tests/generic/626
> > new file mode 100755
> > index 00000000..1baa73f8
> > --- /dev/null
> > +++ b/tests/generic/626
> > @@ -0,0 +1,74 @@
> > +#! /bin/bash
> > +# SPDX-License-Identifier: GPL-2.0
> > +# Copyright (c) 2021 HUAWEI. All Rights Reserved.
> > +# Copyright (c) 2021 Red Hat Inc. All Rights Reserved.
> > +#
> > +# FS QA Test No. 626
> > +#
> > +# Test RENAME_WHITEOUT on filesystem without space to create one more inodes.
> > +# This is a regression test for kernel commit:
> > +# 6b4b8e6b4ad8 ("ext4: ext4: fix bug for rename with RENAME_WHITEOUT")
> > +#
> > +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
> > +. ./common/populate
> > +. ./common/renameat2
> > +
> > +# remove previous $seqres.full before test
> > +rm -f $seqres.full
> > +
> > +# real QA test starts here
> > +_supported_fs generic
> > +_require_scratch
> > +_require_renameat2 whiteout
> > +
> > +_scratch_mkfs_sized $((256 * 1024 * 1024)) >> $seqres.full 2>&1
> > +_scratch_mount
> > +
> > +# Create lots of files, to help to trigger the bug easily
> > +NR_FILE=$((4 * 64))
> > +for ((i=0; i<NR_FILE; i++));do
> > + touch $SCRATCH_MNT/srcfile$i
> > +done
> > +# Try to fill the whole fs
> > +nr_free=$(stat -f -c '%f' $SCRATCH_MNT)
> > +blksz="$(_get_block_size $SCRATCH_MNT)"
> > +_fill_fs $((nr_free * blksz)) $SCRATCH_MNT/fill_space $blksz 0 >> $seqres.full 2>&1
> > +# Use empty files to fill the rest
> > +for ((i=0; i<10000; i++));do
> > + touch $SCRATCH_MNT/fill_file$i 2>/dev/null
> > + # Until no more files can be created
> > + if [ $? -ne 0 ];then
> > + break
> > + fi
> > +done
> > +# ENOSPC is expected here
> > +for ((i=0; i<NR_FILE; i++));do
> > + $here/src/renameat2 -w $SCRATCH_MNT/srcfile$i $SCRATCH_MNT/dstfile$i >> $seqres.full 2>&1
> > +done
> > +_scratch_cycle_mount
> > +# Expect no errors at here
> > +for ((i=0; i<NR_FILE; i++));do
> > + ls -l $SCRATCH_MNT/srcfile$i >/dev/null
> > +done
> > +
> > +echo "Silence is golden"
> > +# success, all done
> > +status=0
> > +exit
> > diff --git a/tests/generic/626.out b/tests/generic/626.out
> > new file mode 100644
> > index 00000000..130b2fef
> > --- /dev/null
> > +++ b/tests/generic/626.out
> > @@ -0,0 +1,2 @@
> > +QA output created by 626
> > +Silence is golden
> > diff --git a/tests/generic/group b/tests/generic/group
> > index 84db3789..c3448fe3 100644
> > --- a/tests/generic/group
> > +++ b/tests/generic/group
> > @@ -628,3 +628,4 @@
> > 623 auto quick shutdown
> > 624 auto quick verity
> > 625 auto quick verity
> > +626 auto quick rename enospc
> > --
> > 2.29.2
> >
>