diff mbox

[v2] ext4: fsmap tests

Message ID 20170602225220.GA5634@birch.djwong.org (mailing list archive)
State New, archived
Headers show

Commit Message

Darrick J. Wong June 2, 2017, 10:52 p.m. UTC
Test the GETFSMAP ioctl against ext4.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 tests/ext4/700     |   76 ++++++++++++++++++++++++++++++++++++++++++++++++++++
 tests/ext4/700.out |    7 +++++
 tests/ext4/701     |   65 ++++++++++++++++++++++++++++++++++++++++++++
 tests/ext4/701.out |    3 ++
 tests/ext4/702     |   66 +++++++++++++++++++++++++++++++++++++++++++++
 tests/ext4/702.out |    4 +++
 tests/ext4/group   |    3 ++
 7 files changed, 224 insertions(+)
 create mode 100755 tests/ext4/700
 create mode 100644 tests/ext4/700.out
 create mode 100755 tests/ext4/701
 create mode 100644 tests/ext4/701.out
 create mode 100755 tests/ext4/702
 create mode 100644 tests/ext4/702.out

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

Comments

Eryu Guan June 16, 2017, 6:34 a.m. UTC | #1
Hi Darrick,

On Fri, Jun 02, 2017 at 03:52:20PM -0700, Darrick J. Wong wrote:
> Test the GETFSMAP ioctl against ext4.
> 
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>

I'm planning to merge this patch after merging Jeff's patch that adds
external log dev support to ext4 (Jeff's patchset is still in its review
cycle), if you're not in a hurry. Please let me know if you'd like to
see ext4 fsmap tests sooner.

Thanks,
Eryu
--
To unsubscribe from this list: send the line "unsubscribe fstests" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Darrick J. Wong June 16, 2017, 6:07 p.m. UTC | #2
On Fri, Jun 16, 2017 at 02:34:45PM +0800, Eryu Guan wrote:
> Hi Darrick,
> 
> On Fri, Jun 02, 2017 at 03:52:20PM -0700, Darrick J. Wong wrote:
> > Test the GETFSMAP ioctl against ext4.
> > 
> > Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> 
> I'm planning to merge this patch after merging Jeff's patch that adds
> external log dev support to ext4 (Jeff's patchset is still in its review
> cycle), if you're not in a hurry. Please let me know if you'd like to
> see ext4 fsmap tests sooner.

I'm in no rush. :)

--D

> 
> Thanks,
> Eryu
--
To unsubscribe from this list: send the line "unsubscribe fstests" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Eryu Guan June 20, 2017, 3:15 p.m. UTC | #3
On Fri, Jun 02, 2017 at 03:52:20PM -0700, Darrick J. Wong wrote:
> Test the GETFSMAP ioctl against ext4.
> 
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> ---
>  tests/ext4/700     |   76 ++++++++++++++++++++++++++++++++++++++++++++++++++++
>  tests/ext4/700.out |    7 +++++
>  tests/ext4/701     |   65 ++++++++++++++++++++++++++++++++++++++++++++
>  tests/ext4/701.out |    3 ++
>  tests/ext4/702     |   66 +++++++++++++++++++++++++++++++++++++++++++++
>  tests/ext4/702.out |    4 +++
>  tests/ext4/group   |    3 ++
>  7 files changed, 224 insertions(+)
>  create mode 100755 tests/ext4/700
>  create mode 100644 tests/ext4/700.out
>  create mode 100755 tests/ext4/701
>  create mode 100644 tests/ext4/701.out
>  create mode 100755 tests/ext4/702
>  create mode 100644 tests/ext4/702.out
> 
> diff --git a/tests/ext4/700 b/tests/ext4/700
> new file mode 100755
> index 0000000..a4a1b83
> --- /dev/null
> +++ b/tests/ext4/700
> @@ -0,0 +1,76 @@
> +#! /bin/bash
> +# FS QA Test No. 700
> +#
> +# Check that getfsmap reports the BG metadata we're expecting.
> +#
> +#-----------------------------------------------------------------------
> +# Copyright (c) 2017, Oracle and/or its affiliates.  All Rights Reserved.
> +#
> +# 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".* $TEST_DIR/fsmap $TEST_DIR/testout
> +}
> +
> +# get standard environment, filters and checks
> +. ./common/rc
> +. ./common/filter
> +
> +# real QA test starts here
> +_supported_os Linux
> +_supported_fs ext4
> +_require_scratch
> +_require_xfs_io_command "fsmap"
> +
> +rm -f "$seqres.full"
> +
> +echo "Format and mount"
> +_scratch_mkfs > "$seqres.full" 2>&1
> +_scratch_mount
> +
> +echo "Get fsmap" | tee -a $seqres.full
> +$XFS_IO_PROG -c 'fsmap' $SCRATCH_MNT > $TEST_DIR/fsmap
> +cat $TEST_DIR/fsmap >> $seqres.full
> +
> +echo "Check fs metadata" | tee -a $seqres.full
> +x=$(grep -c 'static fs metadata' $TEST_DIR/fsmap)
> +test $x -gt 0 || echo "No fs metadata?"

This test fails if I use external log dev (with 4.12-rc5 kernel and
djwong-devel branch of your xfsprogs tree).

    --- tests/ext4/026.out      2017-06-20 20:44:11.307128655 +0800
    +++ /root/xfstests/results//ext4/026.out.bad        2017-06-20 21:02:31.361452548 +0800
    @@ -2,6 +2,10 @@
     Format and mount
     Get fsmap
     Check fs metadata
    +No fs metadata?
     Check block bitmap
    +No block bitmaps?
     Check inode bitmap
     ...

Seems this test assmues internal journal? Do we need a new
_require_no_logdev rule? (BTW, ext4/701 fails too if external log is
set).

> +
> +echo "Check block bitmap" | tee -a $seqres.full
> +x=$(grep -c 'special 102:1' $TEST_DIR/fsmap)
> +test $x -gt 0 || echo "No block bitmaps?"
> +
> +echo "Check inode bitmap" | tee -a $seqres.full
> +x=$(grep -c 'special 102:2' $TEST_DIR/fsmap)

It's not clear to me why "special 102:1" and "special 102:2" represent
"block bitmaps" and "inode bitmaps", do you mind adding some comments?

Also, the format of fsmap output seems not match the manpage? In
xfs_io.8 we have
extent: major:minor [startblock..endblock]: owner startoffset..endoffset length

and a fsmap output on ext4 filesystem is like
51: 8:33 [12582912..12583039]: special 102:3 128

So I assume 'special' is the 'owner' field, but '102:3' doesn't seem
like offsets?

> +test $x -gt 0 || echo "No inode bitmaps?"
> +
> +echo "Check inodes" | tee -a $seqres.full
> +x=$(grep -c 'inodes' $TEST_DIR/fsmap)
> +test $x -gt 0 || echo "No inodes?"
> +
> +# success, all done
> +status=0
> +exit
> diff --git a/tests/ext4/700.out b/tests/ext4/700.out
> new file mode 100644
> index 0000000..3984319
> --- /dev/null
> +++ b/tests/ext4/700.out
> @@ -0,0 +1,7 @@
> +QA output created by 700
> +Format and mount
> +Get fsmap
> +Check fs metadata
> +Check block bitmap
> +Check inode bitmap
> +Check inodes
> diff --git a/tests/ext4/701 b/tests/ext4/701
> new file mode 100755
> index 0000000..6ae93d7
> --- /dev/null
> +++ b/tests/ext4/701
> @@ -0,0 +1,65 @@
> +#! /bin/bash
> +# FS QA Test No. 701
> +#
> +# Populate filesystem, check that fsmap -n10000 matches fsmap -n1.
> +#
> +#-----------------------------------------------------------------------
> +# Copyright (c) 2017, Oracle and/or its affiliates.  All Rights Reserved.
> +#
> +# 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".* $TEST_DIR/a $TEST_DIR/b
> +}
> +
> +# get standard environment, filters and checks
> +. ./common/rc
> +. ./common/filter
> +. ./common/populate
> +
> +# real QA test starts here
> +_supported_os Linux
> +_supported_fs ext4
> +_require_scratch
> +_require_populate_commands
> +_require_xfs_io_command "fsmap"
> +
> +rm -f "$seqres.full"
> +
> +echo "Format and mount"
> +_scratch_populate_cached nofill > $seqres.full 2>&1

This test would take very long time for me to finish, it's been running
for 1 hour and it's still running. Seems like there're too many extents
to report.

[root@ibm-x3550m3-05 xfsprogs-dev]# du -hs /mnt/testarea/test/a
163M    /mnt/testarea/test/a
[root@ibm-x3550m3-05 xfsprogs-dev]# wc -l /mnt/testarea/test/a
3170842 /mnt/testarea/test/a

I digged into it a bit, it was _scratch_populate_cached that restored a
ext4 image with external journal to current test, which has no external
log used (because I ran this test first with external journal set), so
_scratch_mount failed to mount SCRATCH_DEV and fsmap works on rootfs,
which resulted in very long extents list.

I think we should take LOGDEV into account when deciding if cached fs
image should be used, e.g. I did this update to common/populate and test
worked fine for me.

-       meta_descr="FSTYP ${FSTYP} MKFS_OPTIONS ${MKFS_OPTIONS} SIZE $(blockdev --getsz "${SCRATCH_DEV}") ARGS $@"
+       meta_descr="FSTYP ${FSTYP} MKFS_OPTIONS ${MKFS_OPTIONS} SIZE $(blockdev --getsz "${SCRATCH_DEV}") LOGDEV ${SCRATCH_LOGDEV} USE_EXTERNAL ${USE_EXTERNAL} ARGS $@"

Thanks,
Eryu

> +
> +echo "Compare fsmap" | tee -a $seqres.full
> +_scratch_mount
> +$XFS_IO_PROG -c 'fsmap -n 65536' $SCRATCH_MNT | grep -v 'EXT:' > $TEST_DIR/a
> +$XFS_IO_PROG -c 'fsmap -n 1' $SCRATCH_MNT | grep -v 'EXT:' > $TEST_DIR/b
> +cat $TEST_DIR/a $TEST_DIR/b >> $seqres.full
> +
> +diff -uw $TEST_DIR/a $TEST_DIR/b
> +
> +# success, all done
> +status=0
> +exit
> diff --git a/tests/ext4/701.out b/tests/ext4/701.out
> new file mode 100644
> index 0000000..41ef1c6
> --- /dev/null
> +++ b/tests/ext4/701.out
> @@ -0,0 +1,3 @@
> +QA output created by 701
> +Format and mount
> +Compare fsmap
> diff --git a/tests/ext4/702 b/tests/ext4/702
> new file mode 100755
> index 0000000..ce3ae11
> --- /dev/null
> +++ b/tests/ext4/702
> @@ -0,0 +1,66 @@
> +#! /bin/bash
> +# FS QA Test No. 702
> +#
> +# Check that getfsmap reports external log devices
> +#
> +#-----------------------------------------------------------------------
> +# Copyright (c) 2017, Oracle and/or its affiliates.  All Rights Reserved.
> +#
> +# 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".* $TEST_DIR/fsmap $TEST_DIR/testout
> +}
> +
> +# get standard environment, filters and checks
> +. ./common/rc
> +. ./common/filter
> +
> +# real QA test starts here
> +_supported_os Linux
> +_supported_fs ext4
> +_require_logdev
> +_require_scratch
> +_require_xfs_io_command "fsmap"
> +
> +rm -f "$seqres.full"
> +
> +echo "Format and mount"
> +_scratch_mkfs > "$seqres.full" 2>&1
> +_scratch_mount
> +
> +echo "Get fsmap" | tee -a $seqres.full
> +$XFS_IO_PROG -c 'fsmap' $SCRATCH_MNT >> $seqres.full
> +$XFS_IO_PROG -c 'fsmap' $SCRATCH_MNT | tr '[]()' '    ' > $TEST_DIR/fsmap
> +
> +echo "Check device field of FS metadata and journalling log"
> +data_dev=$(grep 'static fs metadata' $TEST_DIR/fsmap | head -n 1 | awk '{print $2}')
> +journal_dev=$(grep 'journalling log' $TEST_DIR/fsmap | head -n 1 | awk '{print $2}')
> +test "${data_dev}" != "${journal_dev}" || echo "data ${data_dev} journal ${journal_dev}?"
> +
> +# success, all done
> +status=0
> +exit
> diff --git a/tests/ext4/702.out b/tests/ext4/702.out
> new file mode 100644
> index 0000000..1d3b7ef
> --- /dev/null
> +++ b/tests/ext4/702.out
> @@ -0,0 +1,4 @@
> +QA output created by 702
> +Format and mount
> +Get fsmap
> +Check device field of FS metadata and journalling log
> diff --git a/tests/ext4/group b/tests/ext4/group
> index 664c059..9007adf 100644
> --- a/tests/ext4/group
> +++ b/tests/ext4/group
> @@ -37,3 +37,6 @@
>  306 auto rw resize quick
>  307 auto ioctl rw defrag
>  308 auto ioctl rw prealloc quick defrag
> +700 auto quick fsmap
> +701 auto quick fsmap
> +702 auto quick fsmap
--
To unsubscribe from this list: send the line "unsubscribe fstests" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Darrick J. Wong June 20, 2017, 3:58 p.m. UTC | #4
On Tue, Jun 20, 2017 at 11:15:00PM +0800, Eryu Guan wrote:
> On Fri, Jun 02, 2017 at 03:52:20PM -0700, Darrick J. Wong wrote:
> > Test the GETFSMAP ioctl against ext4.
> > 
> > Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> > ---
> >  tests/ext4/700     |   76 ++++++++++++++++++++++++++++++++++++++++++++++++++++
> >  tests/ext4/700.out |    7 +++++
> >  tests/ext4/701     |   65 ++++++++++++++++++++++++++++++++++++++++++++
> >  tests/ext4/701.out |    3 ++
> >  tests/ext4/702     |   66 +++++++++++++++++++++++++++++++++++++++++++++
> >  tests/ext4/702.out |    4 +++
> >  tests/ext4/group   |    3 ++
> >  7 files changed, 224 insertions(+)
> >  create mode 100755 tests/ext4/700
> >  create mode 100644 tests/ext4/700.out
> >  create mode 100755 tests/ext4/701
> >  create mode 100644 tests/ext4/701.out
> >  create mode 100755 tests/ext4/702
> >  create mode 100644 tests/ext4/702.out
> > 
> > diff --git a/tests/ext4/700 b/tests/ext4/700
> > new file mode 100755
> > index 0000000..a4a1b83
> > --- /dev/null
> > +++ b/tests/ext4/700
> > @@ -0,0 +1,76 @@
> > +#! /bin/bash
> > +# FS QA Test No. 700
> > +#
> > +# Check that getfsmap reports the BG metadata we're expecting.
> > +#
> > +#-----------------------------------------------------------------------
> > +# Copyright (c) 2017, Oracle and/or its affiliates.  All Rights Reserved.
> > +#
> > +# 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".* $TEST_DIR/fsmap $TEST_DIR/testout
> > +}
> > +
> > +# get standard environment, filters and checks
> > +. ./common/rc
> > +. ./common/filter
> > +
> > +# real QA test starts here
> > +_supported_os Linux
> > +_supported_fs ext4
> > +_require_scratch
> > +_require_xfs_io_command "fsmap"
> > +
> > +rm -f "$seqres.full"
> > +
> > +echo "Format and mount"
> > +_scratch_mkfs > "$seqres.full" 2>&1
> > +_scratch_mount
> > +
> > +echo "Get fsmap" | tee -a $seqres.full
> > +$XFS_IO_PROG -c 'fsmap' $SCRATCH_MNT > $TEST_DIR/fsmap
> > +cat $TEST_DIR/fsmap >> $seqres.full
> > +
> > +echo "Check fs metadata" | tee -a $seqres.full
> > +x=$(grep -c 'static fs metadata' $TEST_DIR/fsmap)
> > +test $x -gt 0 || echo "No fs metadata?"
> 
> This test fails if I use external log dev (with 4.12-rc5 kernel and
> djwong-devel branch of your xfsprogs tree).
> 
>     --- tests/ext4/026.out      2017-06-20 20:44:11.307128655 +0800
>     +++ /root/xfstests/results//ext4/026.out.bad        2017-06-20 21:02:31.361452548 +0800
>     @@ -2,6 +2,10 @@
>      Format and mount
>      Get fsmap
>      Check fs metadata
>     +No fs metadata?
>      Check block bitmap
>     +No block bitmaps?

Uhoh. :)

Strangely, I couldn't reproduce this; maybe our software our out of
sync?  Can you send me results/ext4/026.out ?

>      Check inode bitmap
>      ...
> 
> Seems this test assmues internal journal? Do we need a new
> _require_no_logdev rule? (BTW, ext4/701 fails too if external log is
> set).
> 
> > +
> > +echo "Check block bitmap" | tee -a $seqres.full
> > +x=$(grep -c 'special 102:1' $TEST_DIR/fsmap)
> > +test $x -gt 0 || echo "No block bitmaps?"
> > +
> > +echo "Check inode bitmap" | tee -a $seqres.full
> > +x=$(grep -c 'special 102:2' $TEST_DIR/fsmap)
> 
> It's not clear to me why "special 102:1" and "special 102:2" represent
> "block bitmaps" and "inode bitmaps", do you mind adding some comments?

Those are ext4-specific owner codes, which xfs_io of course doesn't know
much about other than to print the raw number data.

> Also, the format of fsmap output seems not match the manpage? In
> xfs_io.8 we have
> extent: major:minor [startblock..endblock]: owner startoffset..endoffset length
> 
> and a fsmap output on ext4 filesystem is like
> 51: 8:33 [12582912..12583039]: special 102:3 128
> 
> So I assume 'special' is the 'owner' field, but '102:3' doesn't seem
> like offsets?

Yeah... Eric complained about the undiscoverable spaces in the owner
field too.  The whole field is 'special 102:2'.  Maybe we should just
use the (new) machine-readable format switch instead.  Sort of waiting
on Eric to pull all those patches into xfsprogs though...

> > +test $x -gt 0 || echo "No inode bitmaps?"
> > +
> > +echo "Check inodes" | tee -a $seqres.full
> > +x=$(grep -c 'inodes' $TEST_DIR/fsmap)
> > +test $x -gt 0 || echo "No inodes?"
> > +
> > +# success, all done
> > +status=0
> > +exit
> > diff --git a/tests/ext4/700.out b/tests/ext4/700.out
> > new file mode 100644
> > index 0000000..3984319
> > --- /dev/null
> > +++ b/tests/ext4/700.out
> > @@ -0,0 +1,7 @@
> > +QA output created by 700
> > +Format and mount
> > +Get fsmap
> > +Check fs metadata
> > +Check block bitmap
> > +Check inode bitmap
> > +Check inodes
> > diff --git a/tests/ext4/701 b/tests/ext4/701
> > new file mode 100755
> > index 0000000..6ae93d7
> > --- /dev/null
> > +++ b/tests/ext4/701
> > @@ -0,0 +1,65 @@
> > +#! /bin/bash
> > +# FS QA Test No. 701
> > +#
> > +# Populate filesystem, check that fsmap -n10000 matches fsmap -n1.
> > +#
> > +#-----------------------------------------------------------------------
> > +# Copyright (c) 2017, Oracle and/or its affiliates.  All Rights Reserved.
> > +#
> > +# 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".* $TEST_DIR/a $TEST_DIR/b
> > +}
> > +
> > +# get standard environment, filters and checks
> > +. ./common/rc
> > +. ./common/filter
> > +. ./common/populate
> > +
> > +# real QA test starts here
> > +_supported_os Linux
> > +_supported_fs ext4
> > +_require_scratch
> > +_require_populate_commands
> > +_require_xfs_io_command "fsmap"
> > +
> > +rm -f "$seqres.full"
> > +
> > +echo "Format and mount"
> > +_scratch_populate_cached nofill > $seqres.full 2>&1
> 
> This test would take very long time for me to finish, it's been running
> for 1 hour and it's still running. Seems like there're too many extents
> to report.
> 
> [root@ibm-x3550m3-05 xfsprogs-dev]# du -hs /mnt/testarea/test/a
> 163M    /mnt/testarea/test/a
> [root@ibm-x3550m3-05 xfsprogs-dev]# wc -l /mnt/testarea/test/a
> 3170842 /mnt/testarea/test/a
> 
> I digged into it a bit, it was _scratch_populate_cached that restored a
> ext4 image with external journal to current test, which has no external
> log used (because I ran this test first with external journal set), so
> _scratch_mount failed to mount SCRATCH_DEV and fsmap works on rootfs,
> which resulted in very long extents list.
> 
> I think we should take LOGDEV into account when deciding if cached fs
> image should be used, e.g. I did this update to common/populate and test
> worked fine for me.
> 
> -       meta_descr="FSTYP ${FSTYP} MKFS_OPTIONS ${MKFS_OPTIONS} SIZE $(blockdev --getsz "${SCRATCH_DEV}") ARGS $@"
> +       meta_descr="FSTYP ${FSTYP} MKFS_OPTIONS ${MKFS_OPTIONS} SIZE $(blockdev --getsz "${SCRATCH_DEV}") LOGDEV ${SCRATCH_LOGDEV} USE_EXTERNAL ${USE_EXTERNAL} ARGS $@"

Yes, though for XFS please add the rtdev configuration too.  I think
maybe something like this:

case "${FSTYP}" in
"ext4")
	extra_descr="LOGDEV ${SCRATCH_LOGDEV} USE_EXTERNAL ${USE_EXTERNAL}";;
"xfs")
	extra_descr="LOGDEV ${SCRATCH_LOGDEV} USE_EXTERNAL ${USE_EXTERNAL} RTDEV ${SCRATCH_RTDEV}";;
"btrfs")
	extra_descr="<multi-device configuration here>...";;
*)
	extra_descr=
	;;
esac

# Throw away cached image if it doesn't match our spec.
meta_descr="FSTYP ${FSTYP} MKFS_OPTIONS ${MKFS_OPTIONS} SIZE $(blockdev --getsz "${SCRATCH_DEV}") ${extra_descr} ARGS $@"
cmp -s "${POPULATE_METADUMP_DESCR}" <(echo "${meta_descr}") || rm -rf "${POPULATE_METADUMP}"

--D

> 
> Thanks,
> Eryu
> 
> > +
> > +echo "Compare fsmap" | tee -a $seqres.full
> > +_scratch_mount
> > +$XFS_IO_PROG -c 'fsmap -n 65536' $SCRATCH_MNT | grep -v 'EXT:' > $TEST_DIR/a
> > +$XFS_IO_PROG -c 'fsmap -n 1' $SCRATCH_MNT | grep -v 'EXT:' > $TEST_DIR/b
> > +cat $TEST_DIR/a $TEST_DIR/b >> $seqres.full
> > +
> > +diff -uw $TEST_DIR/a $TEST_DIR/b
> > +
> > +# success, all done
> > +status=0
> > +exit
> > diff --git a/tests/ext4/701.out b/tests/ext4/701.out
> > new file mode 100644
> > index 0000000..41ef1c6
> > --- /dev/null
> > +++ b/tests/ext4/701.out
> > @@ -0,0 +1,3 @@
> > +QA output created by 701
> > +Format and mount
> > +Compare fsmap
> > diff --git a/tests/ext4/702 b/tests/ext4/702
> > new file mode 100755
> > index 0000000..ce3ae11
> > --- /dev/null
> > +++ b/tests/ext4/702
> > @@ -0,0 +1,66 @@
> > +#! /bin/bash
> > +# FS QA Test No. 702
> > +#
> > +# Check that getfsmap reports external log devices
> > +#
> > +#-----------------------------------------------------------------------
> > +# Copyright (c) 2017, Oracle and/or its affiliates.  All Rights Reserved.
> > +#
> > +# 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".* $TEST_DIR/fsmap $TEST_DIR/testout
> > +}
> > +
> > +# get standard environment, filters and checks
> > +. ./common/rc
> > +. ./common/filter
> > +
> > +# real QA test starts here
> > +_supported_os Linux
> > +_supported_fs ext4
> > +_require_logdev
> > +_require_scratch
> > +_require_xfs_io_command "fsmap"
> > +
> > +rm -f "$seqres.full"
> > +
> > +echo "Format and mount"
> > +_scratch_mkfs > "$seqres.full" 2>&1
> > +_scratch_mount
> > +
> > +echo "Get fsmap" | tee -a $seqres.full
> > +$XFS_IO_PROG -c 'fsmap' $SCRATCH_MNT >> $seqres.full
> > +$XFS_IO_PROG -c 'fsmap' $SCRATCH_MNT | tr '[]()' '    ' > $TEST_DIR/fsmap
> > +
> > +echo "Check device field of FS metadata and journalling log"
> > +data_dev=$(grep 'static fs metadata' $TEST_DIR/fsmap | head -n 1 | awk '{print $2}')
> > +journal_dev=$(grep 'journalling log' $TEST_DIR/fsmap | head -n 1 | awk '{print $2}')
> > +test "${data_dev}" != "${journal_dev}" || echo "data ${data_dev} journal ${journal_dev}?"
> > +
> > +# success, all done
> > +status=0
> > +exit
> > diff --git a/tests/ext4/702.out b/tests/ext4/702.out
> > new file mode 100644
> > index 0000000..1d3b7ef
> > --- /dev/null
> > +++ b/tests/ext4/702.out
> > @@ -0,0 +1,4 @@
> > +QA output created by 702
> > +Format and mount
> > +Get fsmap
> > +Check device field of FS metadata and journalling log
> > diff --git a/tests/ext4/group b/tests/ext4/group
> > index 664c059..9007adf 100644
> > --- a/tests/ext4/group
> > +++ b/tests/ext4/group
> > @@ -37,3 +37,6 @@
> >  306 auto rw resize quick
> >  307 auto ioctl rw defrag
> >  308 auto ioctl rw prealloc quick defrag
> > +700 auto quick fsmap
> > +701 auto quick fsmap
> > +702 auto quick fsmap
> --
> To unsubscribe from this list: send the line "unsubscribe fstests" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe fstests" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Eryu Guan June 21, 2017, 7:43 a.m. UTC | #5
On Tue, Jun 20, 2017 at 08:58:48AM -0700, Darrick J. Wong wrote:
> On Tue, Jun 20, 2017 at 11:15:00PM +0800, Eryu Guan wrote:
> > On Fri, Jun 02, 2017 at 03:52:20PM -0700, Darrick J. Wong wrote:
> > > Test the GETFSMAP ioctl against ext4.
> > > 
> > > Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> > > ---
> > >  tests/ext4/700     |   76 ++++++++++++++++++++++++++++++++++++++++++++++++++++
> > >  tests/ext4/700.out |    7 +++++
> > >  tests/ext4/701     |   65 ++++++++++++++++++++++++++++++++++++++++++++
> > >  tests/ext4/701.out |    3 ++
> > >  tests/ext4/702     |   66 +++++++++++++++++++++++++++++++++++++++++++++
> > >  tests/ext4/702.out |    4 +++
> > >  tests/ext4/group   |    3 ++
> > >  7 files changed, 224 insertions(+)
> > >  create mode 100755 tests/ext4/700
> > >  create mode 100644 tests/ext4/700.out
> > >  create mode 100755 tests/ext4/701
> > >  create mode 100644 tests/ext4/701.out
> > >  create mode 100755 tests/ext4/702
> > >  create mode 100644 tests/ext4/702.out
> > > 
> > > diff --git a/tests/ext4/700 b/tests/ext4/700
> > > new file mode 100755
> > > index 0000000..a4a1b83
> > > --- /dev/null
> > > +++ b/tests/ext4/700
> > > @@ -0,0 +1,76 @@
> > > +#! /bin/bash
> > > +# FS QA Test No. 700
> > > +#
> > > +# Check that getfsmap reports the BG metadata we're expecting.
> > > +#
> > > +#-----------------------------------------------------------------------
> > > +# Copyright (c) 2017, Oracle and/or its affiliates.  All Rights Reserved.
> > > +#
> > > +# 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".* $TEST_DIR/fsmap $TEST_DIR/testout
> > > +}
> > > +
> > > +# get standard environment, filters and checks
> > > +. ./common/rc
> > > +. ./common/filter
> > > +
> > > +# real QA test starts here
> > > +_supported_os Linux
> > > +_supported_fs ext4
> > > +_require_scratch
> > > +_require_xfs_io_command "fsmap"
> > > +
> > > +rm -f "$seqres.full"
> > > +
> > > +echo "Format and mount"
> > > +_scratch_mkfs > "$seqres.full" 2>&1
> > > +_scratch_mount
> > > +
> > > +echo "Get fsmap" | tee -a $seqres.full
> > > +$XFS_IO_PROG -c 'fsmap' $SCRATCH_MNT > $TEST_DIR/fsmap
> > > +cat $TEST_DIR/fsmap >> $seqres.full
> > > +
> > > +echo "Check fs metadata" | tee -a $seqres.full
> > > +x=$(grep -c 'static fs metadata' $TEST_DIR/fsmap)
> > > +test $x -gt 0 || echo "No fs metadata?"
> > 
> > This test fails if I use external log dev (with 4.12-rc5 kernel and
> > djwong-devel branch of your xfsprogs tree).
> > 
> >     --- tests/ext4/026.out      2017-06-20 20:44:11.307128655 +0800
> >     +++ /root/xfstests/results//ext4/026.out.bad        2017-06-20 21:02:31.361452548 +0800
> >     @@ -2,6 +2,10 @@
> >      Format and mount
> >      Get fsmap
> >      Check fs metadata
> >     +No fs metadata?
> >      Check block bitmap
> >     +No block bitmaps?
> 
> Uhoh. :)
> 
> Strangely, I couldn't reproduce this; maybe our software our out of
> sync?  Can you send me results/ext4/026.out ?

You mean 026.full? Anyway, I attached both 026.full and 026.out.bad.

BTW, I'm using e2fsprogs 'next' branch, HEAD is "474f67a libext2fs: fix
fsync(2) detection", not sure if that matters.

> 
> >      Check inode bitmap
> >      ...
> > 
> > Seems this test assmues internal journal? Do we need a new
> > _require_no_logdev rule? (BTW, ext4/701 fails too if external log is
> > set).
> > 
> > > +
> > > +echo "Check block bitmap" | tee -a $seqres.full
> > > +x=$(grep -c 'special 102:1' $TEST_DIR/fsmap)
> > > +test $x -gt 0 || echo "No block bitmaps?"
> > > +
> > > +echo "Check inode bitmap" | tee -a $seqres.full
> > > +x=$(grep -c 'special 102:2' $TEST_DIR/fsmap)
> > 
> > It's not clear to me why "special 102:1" and "special 102:2" represent
> > "block bitmaps" and "inode bitmaps", do you mind adding some comments?
> 
> Those are ext4-specific owner codes, which xfs_io of course doesn't know
> much about other than to print the raw number data.
> 
> > Also, the format of fsmap output seems not match the manpage? In
> > xfs_io.8 we have
> > extent: major:minor [startblock..endblock]: owner startoffset..endoffset length
> > 
> > and a fsmap output on ext4 filesystem is like
> > 51: 8:33 [12582912..12583039]: special 102:3 128
> > 
> > So I assume 'special' is the 'owner' field, but '102:3' doesn't seem
> > like offsets?
> 
> Yeah... Eric complained about the undiscoverable spaces in the owner
> field too.  The whole field is 'special 102:2'.  Maybe we should just
> use the (new) machine-readable format switch instead.  Sort of waiting
> on Eric to pull all those patches into xfsprogs though...
> 
> > > +test $x -gt 0 || echo "No inode bitmaps?"
> > > +
> > > +echo "Check inodes" | tee -a $seqres.full
> > > +x=$(grep -c 'inodes' $TEST_DIR/fsmap)
> > > +test $x -gt 0 || echo "No inodes?"
> > > +
> > > +# success, all done
> > > +status=0
> > > +exit
> > > diff --git a/tests/ext4/700.out b/tests/ext4/700.out
> > > new file mode 100644
> > > index 0000000..3984319
> > > --- /dev/null
> > > +++ b/tests/ext4/700.out
> > > @@ -0,0 +1,7 @@
> > > +QA output created by 700
> > > +Format and mount
> > > +Get fsmap
> > > +Check fs metadata
> > > +Check block bitmap
> > > +Check inode bitmap
> > > +Check inodes
> > > diff --git a/tests/ext4/701 b/tests/ext4/701
> > > new file mode 100755
> > > index 0000000..6ae93d7
> > > --- /dev/null
> > > +++ b/tests/ext4/701
> > > @@ -0,0 +1,65 @@
> > > +#! /bin/bash
> > > +# FS QA Test No. 701
> > > +#
> > > +# Populate filesystem, check that fsmap -n10000 matches fsmap -n1.
> > > +#
> > > +#-----------------------------------------------------------------------
> > > +# Copyright (c) 2017, Oracle and/or its affiliates.  All Rights Reserved.
> > > +#
> > > +# 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".* $TEST_DIR/a $TEST_DIR/b
> > > +}
> > > +
> > > +# get standard environment, filters and checks
> > > +. ./common/rc
> > > +. ./common/filter
> > > +. ./common/populate
> > > +
> > > +# real QA test starts here
> > > +_supported_os Linux
> > > +_supported_fs ext4
> > > +_require_scratch
> > > +_require_populate_commands
> > > +_require_xfs_io_command "fsmap"
> > > +
> > > +rm -f "$seqres.full"
> > > +
> > > +echo "Format and mount"
> > > +_scratch_populate_cached nofill > $seqres.full 2>&1
> > 
> > This test would take very long time for me to finish, it's been running
> > for 1 hour and it's still running. Seems like there're too many extents
> > to report.
> > 
> > [root@ibm-x3550m3-05 xfsprogs-dev]# du -hs /mnt/testarea/test/a
> > 163M    /mnt/testarea/test/a
> > [root@ibm-x3550m3-05 xfsprogs-dev]# wc -l /mnt/testarea/test/a
> > 3170842 /mnt/testarea/test/a
> > 
> > I digged into it a bit, it was _scratch_populate_cached that restored a
> > ext4 image with external journal to current test, which has no external
> > log used (because I ran this test first with external journal set), so
> > _scratch_mount failed to mount SCRATCH_DEV and fsmap works on rootfs,
> > which resulted in very long extents list.
> > 
> > I think we should take LOGDEV into account when deciding if cached fs
> > image should be used, e.g. I did this update to common/populate and test
> > worked fine for me.
> > 
> > -       meta_descr="FSTYP ${FSTYP} MKFS_OPTIONS ${MKFS_OPTIONS} SIZE $(blockdev --getsz "${SCRATCH_DEV}") ARGS $@"
> > +       meta_descr="FSTYP ${FSTYP} MKFS_OPTIONS ${MKFS_OPTIONS} SIZE $(blockdev --getsz "${SCRATCH_DEV}") LOGDEV ${SCRATCH_LOGDEV} USE_EXTERNAL ${USE_EXTERNAL} ARGS $@"
> 
> Yes, though for XFS please add the rtdev configuration too.  I think
> maybe something like this:
> 
> case "${FSTYP}" in
> "ext4")
> 	extra_descr="LOGDEV ${SCRATCH_LOGDEV} USE_EXTERNAL ${USE_EXTERNAL}";;
> "xfs")
> 	extra_descr="LOGDEV ${SCRATCH_LOGDEV} USE_EXTERNAL ${USE_EXTERNAL} RTDEV ${SCRATCH_RTDEV}";;
> "btrfs")
> 	extra_descr="<multi-device configuration here>...";;
> *)
> 	extra_descr=
> 	;;
> esac

Yeah, this looks better. Would like to do this update in your next
version? :)

Thanks,
Eryu
Darrick J. Wong June 21, 2017, 6:01 p.m. UTC | #6
On Wed, Jun 21, 2017 at 03:43:39PM +0800, Eryu Guan wrote:
> On Tue, Jun 20, 2017 at 08:58:48AM -0700, Darrick J. Wong wrote:
> > On Tue, Jun 20, 2017 at 11:15:00PM +0800, Eryu Guan wrote:
> > > On Fri, Jun 02, 2017 at 03:52:20PM -0700, Darrick J. Wong wrote:
> > > > Test the GETFSMAP ioctl against ext4.
> > > > 
> > > > Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> > > > ---
> > > >  tests/ext4/700     |   76 ++++++++++++++++++++++++++++++++++++++++++++++++++++
> > > >  tests/ext4/700.out |    7 +++++
> > > >  tests/ext4/701     |   65 ++++++++++++++++++++++++++++++++++++++++++++
> > > >  tests/ext4/701.out |    3 ++
> > > >  tests/ext4/702     |   66 +++++++++++++++++++++++++++++++++++++++++++++
> > > >  tests/ext4/702.out |    4 +++
> > > >  tests/ext4/group   |    3 ++
> > > >  7 files changed, 224 insertions(+)
> > > >  create mode 100755 tests/ext4/700
> > > >  create mode 100644 tests/ext4/700.out
> > > >  create mode 100755 tests/ext4/701
> > > >  create mode 100644 tests/ext4/701.out
> > > >  create mode 100755 tests/ext4/702
> > > >  create mode 100644 tests/ext4/702.out
> > > > 
> > > > diff --git a/tests/ext4/700 b/tests/ext4/700
> > > > new file mode 100755
> > > > index 0000000..a4a1b83
> > > > --- /dev/null
> > > > +++ b/tests/ext4/700
> > > > @@ -0,0 +1,76 @@
> > > > +#! /bin/bash
> > > > +# FS QA Test No. 700
> > > > +#
> > > > +# Check that getfsmap reports the BG metadata we're expecting.
> > > > +#
> > > > +#-----------------------------------------------------------------------
> > > > +# Copyright (c) 2017, Oracle and/or its affiliates.  All Rights Reserved.
> > > > +#
> > > > +# 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".* $TEST_DIR/fsmap $TEST_DIR/testout
> > > > +}
> > > > +
> > > > +# get standard environment, filters and checks
> > > > +. ./common/rc
> > > > +. ./common/filter
> > > > +
> > > > +# real QA test starts here
> > > > +_supported_os Linux
> > > > +_supported_fs ext4
> > > > +_require_scratch
> > > > +_require_xfs_io_command "fsmap"
> > > > +
> > > > +rm -f "$seqres.full"
> > > > +
> > > > +echo "Format and mount"
> > > > +_scratch_mkfs > "$seqres.full" 2>&1
> > > > +_scratch_mount
> > > > +
> > > > +echo "Get fsmap" | tee -a $seqres.full
> > > > +$XFS_IO_PROG -c 'fsmap' $SCRATCH_MNT > $TEST_DIR/fsmap
> > > > +cat $TEST_DIR/fsmap >> $seqres.full
> > > > +
> > > > +echo "Check fs metadata" | tee -a $seqres.full
> > > > +x=$(grep -c 'static fs metadata' $TEST_DIR/fsmap)
> > > > +test $x -gt 0 || echo "No fs metadata?"
> > > 
> > > This test fails if I use external log dev (with 4.12-rc5 kernel and
> > > djwong-devel branch of your xfsprogs tree).
> > > 
> > >     --- tests/ext4/026.out      2017-06-20 20:44:11.307128655 +0800
> > >     +++ /root/xfstests/results//ext4/026.out.bad        2017-06-20 21:02:31.361452548 +0800
> > >     @@ -2,6 +2,10 @@
> > >      Format and mount
> > >      Get fsmap
> > >      Check fs metadata
> > >     +No fs metadata?
> > >      Check block bitmap
> > >     +No block bitmaps?
> > 
> > Uhoh. :)
> > 
> > Strangely, I couldn't reproduce this; maybe our software our out of
> > sync?  Can you send me results/ext4/026.out ?
> 
> You mean 026.full? Anyway, I attached both 026.full and 026.out.bad.
> 
> BTW, I'm using e2fsprogs 'next' branch, HEAD is "474f67a libext2fs: fix
> fsync(2) detection", not sure if that matters.

Aha, you're using 1024b blocks... which evidently is broken in the kernel.

> 
> > 
> > >      Check inode bitmap
> > >      ...
> > > 
> > > Seems this test assmues internal journal? Do we need a new
> > > _require_no_logdev rule? (BTW, ext4/701 fails too if external log is
> > > set).
> > > 
> > > > +
> > > > +echo "Check block bitmap" | tee -a $seqres.full
> > > > +x=$(grep -c 'special 102:1' $TEST_DIR/fsmap)
> > > > +test $x -gt 0 || echo "No block bitmaps?"
> > > > +
> > > > +echo "Check inode bitmap" | tee -a $seqres.full
> > > > +x=$(grep -c 'special 102:2' $TEST_DIR/fsmap)
> > > 
> > > It's not clear to me why "special 102:1" and "special 102:2" represent
> > > "block bitmaps" and "inode bitmaps", do you mind adding some comments?
> > 
> > Those are ext4-specific owner codes, which xfs_io of course doesn't know
> > much about other than to print the raw number data.
> > 
> > > Also, the format of fsmap output seems not match the manpage? In
> > > xfs_io.8 we have
> > > extent: major:minor [startblock..endblock]: owner startoffset..endoffset length
> > > 
> > > and a fsmap output on ext4 filesystem is like
> > > 51: 8:33 [12582912..12583039]: special 102:3 128
> > > 
> > > So I assume 'special' is the 'owner' field, but '102:3' doesn't seem
> > > like offsets?
> > 
> > Yeah... Eric complained about the undiscoverable spaces in the owner
> > field too.  The whole field is 'special 102:2'.  Maybe we should just
> > use the (new) machine-readable format switch instead.  Sort of waiting
> > on Eric to pull all those patches into xfsprogs though...
> > 
> > > > +test $x -gt 0 || echo "No inode bitmaps?"
> > > > +
> > > > +echo "Check inodes" | tee -a $seqres.full
> > > > +x=$(grep -c 'inodes' $TEST_DIR/fsmap)
> > > > +test $x -gt 0 || echo "No inodes?"
> > > > +
> > > > +# success, all done
> > > > +status=0
> > > > +exit
> > > > diff --git a/tests/ext4/700.out b/tests/ext4/700.out
> > > > new file mode 100644
> > > > index 0000000..3984319
> > > > --- /dev/null
> > > > +++ b/tests/ext4/700.out
> > > > @@ -0,0 +1,7 @@
> > > > +QA output created by 700
> > > > +Format and mount
> > > > +Get fsmap
> > > > +Check fs metadata
> > > > +Check block bitmap
> > > > +Check inode bitmap
> > > > +Check inodes
> > > > diff --git a/tests/ext4/701 b/tests/ext4/701
> > > > new file mode 100755
> > > > index 0000000..6ae93d7
> > > > --- /dev/null
> > > > +++ b/tests/ext4/701
> > > > @@ -0,0 +1,65 @@
> > > > +#! /bin/bash
> > > > +# FS QA Test No. 701
> > > > +#
> > > > +# Populate filesystem, check that fsmap -n10000 matches fsmap -n1.
> > > > +#
> > > > +#-----------------------------------------------------------------------
> > > > +# Copyright (c) 2017, Oracle and/or its affiliates.  All Rights Reserved.
> > > > +#
> > > > +# 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".* $TEST_DIR/a $TEST_DIR/b
> > > > +}
> > > > +
> > > > +# get standard environment, filters and checks
> > > > +. ./common/rc
> > > > +. ./common/filter
> > > > +. ./common/populate
> > > > +
> > > > +# real QA test starts here
> > > > +_supported_os Linux
> > > > +_supported_fs ext4
> > > > +_require_scratch
> > > > +_require_populate_commands
> > > > +_require_xfs_io_command "fsmap"
> > > > +
> > > > +rm -f "$seqres.full"
> > > > +
> > > > +echo "Format and mount"
> > > > +_scratch_populate_cached nofill > $seqres.full 2>&1
> > > 
> > > This test would take very long time for me to finish, it's been running
> > > for 1 hour and it's still running. Seems like there're too many extents
> > > to report.
> > > 
> > > [root@ibm-x3550m3-05 xfsprogs-dev]# du -hs /mnt/testarea/test/a
> > > 163M    /mnt/testarea/test/a
> > > [root@ibm-x3550m3-05 xfsprogs-dev]# wc -l /mnt/testarea/test/a
> > > 3170842 /mnt/testarea/test/a
> > > 
> > > I digged into it a bit, it was _scratch_populate_cached that restored a
> > > ext4 image with external journal to current test, which has no external
> > > log used (because I ran this test first with external journal set), so
> > > _scratch_mount failed to mount SCRATCH_DEV and fsmap works on rootfs,
> > > which resulted in very long extents list.
> > > 
> > > I think we should take LOGDEV into account when deciding if cached fs
> > > image should be used, e.g. I did this update to common/populate and test
> > > worked fine for me.
> > > 
> > > -       meta_descr="FSTYP ${FSTYP} MKFS_OPTIONS ${MKFS_OPTIONS} SIZE $(blockdev --getsz "${SCRATCH_DEV}") ARGS $@"
> > > +       meta_descr="FSTYP ${FSTYP} MKFS_OPTIONS ${MKFS_OPTIONS} SIZE $(blockdev --getsz "${SCRATCH_DEV}") LOGDEV ${SCRATCH_LOGDEV} USE_EXTERNAL ${USE_EXTERNAL} ARGS $@"
> > 
> > Yes, though for XFS please add the rtdev configuration too.  I think
> > maybe something like this:
> > 
> > case "${FSTYP}" in
> > "ext4")
> > 	extra_descr="LOGDEV ${SCRATCH_LOGDEV} USE_EXTERNAL ${USE_EXTERNAL}";;
> > "xfs")
> > 	extra_descr="LOGDEV ${SCRATCH_LOGDEV} USE_EXTERNAL ${USE_EXTERNAL} RTDEV ${SCRATCH_RTDEV}";;
> > "btrfs")
> > 	extra_descr="<multi-device configuration here>...";;
> > *)
> > 	extra_descr=
> > 	;;
> > esac
> 
> Yeah, this looks better. Would like to do this update in your next
> version? :)

I guess I can roll it up for my next patchbomb.

--D

> 
> Thanks,
> Eryu



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

Patch

diff --git a/tests/ext4/700 b/tests/ext4/700
new file mode 100755
index 0000000..a4a1b83
--- /dev/null
+++ b/tests/ext4/700
@@ -0,0 +1,76 @@ 
+#! /bin/bash
+# FS QA Test No. 700
+#
+# Check that getfsmap reports the BG metadata we're expecting.
+#
+#-----------------------------------------------------------------------
+# Copyright (c) 2017, Oracle and/or its affiliates.  All Rights Reserved.
+#
+# 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".* $TEST_DIR/fsmap $TEST_DIR/testout
+}
+
+# get standard environment, filters and checks
+. ./common/rc
+. ./common/filter
+
+# real QA test starts here
+_supported_os Linux
+_supported_fs ext4
+_require_scratch
+_require_xfs_io_command "fsmap"
+
+rm -f "$seqres.full"
+
+echo "Format and mount"
+_scratch_mkfs > "$seqres.full" 2>&1
+_scratch_mount
+
+echo "Get fsmap" | tee -a $seqres.full
+$XFS_IO_PROG -c 'fsmap' $SCRATCH_MNT > $TEST_DIR/fsmap
+cat $TEST_DIR/fsmap >> $seqres.full
+
+echo "Check fs metadata" | tee -a $seqres.full
+x=$(grep -c 'static fs metadata' $TEST_DIR/fsmap)
+test $x -gt 0 || echo "No fs metadata?"
+
+echo "Check block bitmap" | tee -a $seqres.full
+x=$(grep -c 'special 102:1' $TEST_DIR/fsmap)
+test $x -gt 0 || echo "No block bitmaps?"
+
+echo "Check inode bitmap" | tee -a $seqres.full
+x=$(grep -c 'special 102:2' $TEST_DIR/fsmap)
+test $x -gt 0 || echo "No inode bitmaps?"
+
+echo "Check inodes" | tee -a $seqres.full
+x=$(grep -c 'inodes' $TEST_DIR/fsmap)
+test $x -gt 0 || echo "No inodes?"
+
+# success, all done
+status=0
+exit
diff --git a/tests/ext4/700.out b/tests/ext4/700.out
new file mode 100644
index 0000000..3984319
--- /dev/null
+++ b/tests/ext4/700.out
@@ -0,0 +1,7 @@ 
+QA output created by 700
+Format and mount
+Get fsmap
+Check fs metadata
+Check block bitmap
+Check inode bitmap
+Check inodes
diff --git a/tests/ext4/701 b/tests/ext4/701
new file mode 100755
index 0000000..6ae93d7
--- /dev/null
+++ b/tests/ext4/701
@@ -0,0 +1,65 @@ 
+#! /bin/bash
+# FS QA Test No. 701
+#
+# Populate filesystem, check that fsmap -n10000 matches fsmap -n1.
+#
+#-----------------------------------------------------------------------
+# Copyright (c) 2017, Oracle and/or its affiliates.  All Rights Reserved.
+#
+# 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".* $TEST_DIR/a $TEST_DIR/b
+}
+
+# get standard environment, filters and checks
+. ./common/rc
+. ./common/filter
+. ./common/populate
+
+# real QA test starts here
+_supported_os Linux
+_supported_fs ext4
+_require_scratch
+_require_populate_commands
+_require_xfs_io_command "fsmap"
+
+rm -f "$seqres.full"
+
+echo "Format and mount"
+_scratch_populate_cached nofill > $seqres.full 2>&1
+
+echo "Compare fsmap" | tee -a $seqres.full
+_scratch_mount
+$XFS_IO_PROG -c 'fsmap -n 65536' $SCRATCH_MNT | grep -v 'EXT:' > $TEST_DIR/a
+$XFS_IO_PROG -c 'fsmap -n 1' $SCRATCH_MNT | grep -v 'EXT:' > $TEST_DIR/b
+cat $TEST_DIR/a $TEST_DIR/b >> $seqres.full
+
+diff -uw $TEST_DIR/a $TEST_DIR/b
+
+# success, all done
+status=0
+exit
diff --git a/tests/ext4/701.out b/tests/ext4/701.out
new file mode 100644
index 0000000..41ef1c6
--- /dev/null
+++ b/tests/ext4/701.out
@@ -0,0 +1,3 @@ 
+QA output created by 701
+Format and mount
+Compare fsmap
diff --git a/tests/ext4/702 b/tests/ext4/702
new file mode 100755
index 0000000..ce3ae11
--- /dev/null
+++ b/tests/ext4/702
@@ -0,0 +1,66 @@ 
+#! /bin/bash
+# FS QA Test No. 702
+#
+# Check that getfsmap reports external log devices
+#
+#-----------------------------------------------------------------------
+# Copyright (c) 2017, Oracle and/or its affiliates.  All Rights Reserved.
+#
+# 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".* $TEST_DIR/fsmap $TEST_DIR/testout
+}
+
+# get standard environment, filters and checks
+. ./common/rc
+. ./common/filter
+
+# real QA test starts here
+_supported_os Linux
+_supported_fs ext4
+_require_logdev
+_require_scratch
+_require_xfs_io_command "fsmap"
+
+rm -f "$seqres.full"
+
+echo "Format and mount"
+_scratch_mkfs > "$seqres.full" 2>&1
+_scratch_mount
+
+echo "Get fsmap" | tee -a $seqres.full
+$XFS_IO_PROG -c 'fsmap' $SCRATCH_MNT >> $seqres.full
+$XFS_IO_PROG -c 'fsmap' $SCRATCH_MNT | tr '[]()' '    ' > $TEST_DIR/fsmap
+
+echo "Check device field of FS metadata and journalling log"
+data_dev=$(grep 'static fs metadata' $TEST_DIR/fsmap | head -n 1 | awk '{print $2}')
+journal_dev=$(grep 'journalling log' $TEST_DIR/fsmap | head -n 1 | awk '{print $2}')
+test "${data_dev}" != "${journal_dev}" || echo "data ${data_dev} journal ${journal_dev}?"
+
+# success, all done
+status=0
+exit
diff --git a/tests/ext4/702.out b/tests/ext4/702.out
new file mode 100644
index 0000000..1d3b7ef
--- /dev/null
+++ b/tests/ext4/702.out
@@ -0,0 +1,4 @@ 
+QA output created by 702
+Format and mount
+Get fsmap
+Check device field of FS metadata and journalling log
diff --git a/tests/ext4/group b/tests/ext4/group
index 664c059..9007adf 100644
--- a/tests/ext4/group
+++ b/tests/ext4/group
@@ -37,3 +37,6 @@ 
 306 auto rw resize quick
 307 auto ioctl rw defrag
 308 auto ioctl rw prealloc quick defrag
+700 auto quick fsmap
+701 auto quick fsmap
+702 auto quick fsmap