Re: [PATCH v2] fstests: btrfs: Add regression test for reserved space leak.

[Qu Wenruo <quwenruo@cn.fujitsu.com>]

Tsutomu Itoh wrote on 2015/08/05 10:26 +0900:
> On 2015/08/05 10:08, Qu Wenruo wrote:
>> The regression is introduced in v4.2-rc1, with the big btrfs qgroup
>> change.
>> The problem is, qgroup reserved space is never freed, causing even we
>> increase the limit, we can still hit the EDQUOT much faster than it
>> should.
>>
>> Reported-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
>> Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
>> ---
>>    tests/btrfs/089     | 88 +++++++++++++++++++++++++++++++++++++++++++++++++++++
>>    tests/btrfs/089.out |  5 +++
>>    tests/btrfs/group   |  1 +
>>    3 files changed, 94 insertions(+)
>>    create mode 100755 tests/btrfs/089
>>    create mode 100644 tests/btrfs/089.out
>>
>> diff --git a/tests/btrfs/089 b/tests/btrfs/089
>> new file mode 100755
>> index 0000000..82db96c
>> --- /dev/null
>> +++ b/tests/btrfs/089
>> @@ -0,0 +1,88 @@
>> +#! /bin/bash
>> +# FS QA Test 089
>> +#
>> +# Regression test for btrfs qgroup reserved space leak.
>> +#
>> +# Due to qgroup reserved space leak, EDQUOT can be trigged even it's not
>> +# over limit after previous write.
>> +#
>> +#-----------------------------------------------------------------------
>> +# Copyright (c) 2015 Fujitsu. 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 $tmp.*
>> +}
>> +
>> +# get standard environment, filters and checks
>> +. ./common/rc
>> +. ./common/filter
>> +
>> +# real QA test starts here
>> +
>> +# Modify as appropriate.
>> +_supported_fs btrfs
>> +_supported_os Linux
>> +_require_scratch
>> +_need_to_be_root
>> +
>> +# Use big blocksize to ensure there is still enough space left
>> +# for metadata reserve after hitting EDQUOT
>> +BLOCKSIZE=$(( 2 * 1024 * 1024 ))
>> +FILESIZE=$(( 128 * 1024 * 1024 )) # 128Mbytes
>> +
>> +# The last block won't be able to finish write, as metadata takes
>> +# $NODESIZE space, causing the last block triggering EDQUOT
>> +LENGTH=$(( $FILESIZE - $BLOCKSIZE ))
>> +
>> +_scratch_mkfs >>$seqres.full 2>&1
>> +_scratch_mount
>> +_require_fs_space $SCRATCH_MNT $(($FILESIZE * 2 / 1024))
>> +
>> +_run_btrfs_util_prog quota enable $SCRATCH_MNT
>> +_run_btrfs_util_prog qgroup limit $FILESIZE 5 $SCRATCH_MNT
>> +
>> +$XFS_IO_PROG -f -c "pwrite -b $BLOCKSIZE 0 $LENGTH" \
>> +	$SCRATCH_MNT/foo | _filter_xfs_io
>> +
>> +# A sync is needed to trigger a commit_transaction.
>> +# As the reserved space freeing happens at commit_transaction time,
>> +# without a transaction commit, no reserved space needs freeing and
>> +# won't trigger the bug.
>> +sync
>
> Isn't '$BTRFS_UTIL_PROG filesystem sync' better instead of 'sync'?
>
> Thanks,
> Tsutomu
Hi, Tsutomu-san,

Yes, I did use such method before, but Dave said it's better to use 
unified interface to sync a filesystem other than the specialized one.

So I still use sync as Dave said.

Thanks,
Qu
>
>> +
>> +# Double the limit to allow further write
>> +_run_btrfs_util_prog qgroup limit $(($FILESIZE * 2)) 5 $SCRATCH_MNT
>> +
>> +# Test whether further write can succeed
>> +$XFS_IO_PROG -f -c "pwrite -b $BLOCKSIZE $LENGTH $LENGTH" \
>> +	$SCRATCH_MNT/foo | _filter_xfs_io
>> +
>> +# success, all done
>> +status=0
>> +exit
>> diff --git a/tests/btrfs/089.out b/tests/btrfs/089.out
>> new file mode 100644
>> index 0000000..396888f
>> --- /dev/null
>> +++ b/tests/btrfs/089.out
>> @@ -0,0 +1,5 @@
>> +QA output created by 089
>> +wrote 132120576/132120576 bytes at offset 0
>> +XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
>> +wrote 132120576/132120576 bytes at offset 132120576
>> +XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
>> diff --git a/tests/btrfs/group b/tests/btrfs/group
>> index ffe18bf..225b532 100644
>> --- a/tests/btrfs/group
>> +++ b/tests/btrfs/group
>> @@ -91,6 +91,7 @@
>>    086 auto quick clone
>>    087 auto quick send
>>    088 auto quick metadata
>> +089 auto quick qgroup
>>    090 auto quick metadata
>>    091 auto quick qgroup
>>    092 auto quick send
>>
>
>
> --
> 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
>