On Sat, May 12, 2018 at 9:27 PM, Eryu Guan <guaneryu@xxxxxxxxx> wrote:
> On Mon, May 07, 2018 at 10:09:53PM +0800, Liu Bo wrote:
>> This is to reproduce a leak case of block reservation when bigalloc
>> and delalloc are enabled.
>
> Would you please provide more details about the leak? That helps to
> understand what the test wants to execrise, and people still could know
> the test purpose years later :)
>
Sure.
>>
>> Signed-off-by: Liu Bo <bo.liu@xxxxxxxxxxxxxxxxx>
>> ---
>> tests/ext4/033 | 89 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
>> tests/ext4/033.out | 14 +++++++++
>> tests/ext4/group | 1 +
>> 3 files changed, 104 insertions(+)
>> create mode 100755 tests/ext4/033
>> create mode 100644 tests/ext4/033.out
>>
>> diff --git a/tests/ext4/033 b/tests/ext4/033
>> new file mode 100755
>> index 00000000..74f1a83d
>> --- /dev/null
>> +++ b/tests/ext4/033
>> @@ -0,0 +1,89 @@
>> +#! /bin/bash
>> +# FS QA Test 033
>> +#
>> +# Reproduce ext4 block reservation leak when mkfs.ext4 -Obigalloc and
>> +# mount -odelalloc .
>> +#
>> +#-----------------------------------------------------------------------
>> +# Copyright (c) 2018 Alibaba Group. 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.*
>> +}
>> +
>> +# get standard environment, filters and checks
>> +. ./common/rc
>> +. ./common/filter
>> +
>> +# remove previous $seqres.full before test
>> +rm -f $seqres.full
>> +
>> +# real QA test starts here
>> +
>> +# Modify as appropriate.
>> +_supported_fs ext4
>> +_supported_os Linux
>> +
>> +_require_scratch
>> +_require_scratch_ext4_feature "bigalloc"
>
> Need "_require_xfs_io_command falloc"
>
>> +
>> +_scratch_mkfs -Obigalloc
>
> Need to drop the mkfs output or redirect them to $seqres.full, test
> fails currently because of the mkfs output.
>
OK, it did pass in my test though.
> Also, I think you need to specify the cluster size explicitly, i.e.
> "-C 64k", otherwise test passes unexpectedly on 2k block size ext4, as
> the default cluster size is not always 64k, but "16 times the block
> size" (mke2fs(8)).
>
Yes, make sense.
>> +
>> +# explicitly mount with -odelalloc since -onodelalloc won't trigger the
>> +# mentioned bug.
>> +_scratch_mount -odelalloc
>> +
>> +# case 1: two delayed extents result in leak.
>> +$XFS_IO_PROG -f -c "pwrite 0 1" -c "pwrite 16k 1" -c "fsync" $SCRATCH_MNT/file1 | _filter_xfs_io
>> +
>> +# Since the default cluster size is 64K, and all the writes we've made are in
>> +# one cluster, thus du is supposed to report 64K.
>> +echo "du $SCRATCH_MNT/file1 | $AWK_PROG '{print \$1}'" >> $seqres.full
>> +du $SCRATCH_MNT/file1 | $AWK_PROG '{print $1}' | tee -a $seqres.full
>> +
>> +# case 2: (one allocated extent + one delayed extent) result in leak.
>> +$XFS_IO_PROG -f -c "pwrite 0 1" -c "fsync" -c "pwrite 16k 1" -c "fsync" $SCRATCH_MNT/file2 | _filter_xfs_io
>> +
>> +# Since the default cluster size is 64K, and all the writes we've made are in
>> +# one cluster, thus du is supposed to report 64K.
>
> No need to repeat the same comments again and again :)
>
>> +echo "du $SCRATCH_MNT/file2 | $AWK_PROG '{print \$1}'" >> $seqres.full
>> +du $SCRATCH_MNT/file2 | $AWK_PROG '{print $1}' | tee -a $seqres.full
>
> Make this a local helper function?
>
Sure, will update.
Thanks a lot for the comments.
thanks,
liubo
> Thanks,
> Eryu
>
>> +
>> +# case 3: (one unwritten extent + one delayed extent) result in leak.
>> +$XFS_IO_PROG -f -c "falloc 0 1" -c "pwrite 16k 1" -c "fsync" $SCRATCH_MNT/file3 | _filter_xfs_io
>> +
>> +# Since the default cluster size is 64K, and all the writes we've made are in
>> +# one cluster, thus du is supposed to report 64K.
>> +echo "du $SCRATCH_MNT/file3 | $AWK_PROG '{print \$1}'" >> $seqres.full
>> +du $SCRATCH_MNT/file3 | $AWK_PROG '{print $1}' | tee -a $seqres.full
>> +
>> +
>> +# success, all done
>> +status=0
>> +exit
>> diff --git a/tests/ext4/033.out b/tests/ext4/033.out
>> new file mode 100644
>> index 00000000..7ff3d28f
>> --- /dev/null
>> +++ b/tests/ext4/033.out
>> @@ -0,0 +1,14 @@
>> +QA output created by 033
>> +wrote 1/1 bytes at offset 0
>> +XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
>> +wrote 1/1 bytes at offset 16384
>> +XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
>> +64
>> +wrote 1/1 bytes at offset 0
>> +XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
>> +wrote 1/1 bytes at offset 16384
>> +XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
>> +64
>> +wrote 1/1 bytes at offset 16384
>> +XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
>> +64
>> diff --git a/tests/ext4/group b/tests/ext4/group
>> index 5bd15f82..c0d9df28 100644
>> --- a/tests/ext4/group
>> +++ b/tests/ext4/group
>> @@ -35,6 +35,7 @@
>> 030 auto quick dax
>> 031 auto quick dax
>> 032 auto quick ioctl resize
>> +033 quick auto
>> 271 auto rw quick
>> 301 aio auto ioctl rw stress defrag
>> 302 aio auto ioctl rw stress defrag
>> --
>> 2.14.2
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe fstests" in
>> the body of a message to majordomo@xxxxxxxxxxxxxxx
>> 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@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
