Message ID | 1440119834-31504-1-git-send-email-quwenruo@cn.fujitsu.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
On Fri, Aug 21, 2015 at 09:17:14AM +0800, Qu Wenruo wrote: > Btrfs qgroup reserve codes lacks check for rewrite dirty page, causing > every write, even rewriting a uncommitted dirty page, to reserve space. > > But only written data will free the reserved space, causing reserved > space leaking. > > The bug exists almost from the beginning of btrfs qgroup codes, but > nobody found it. > > For example: > > 1)Write [0, 12K) into file A > reserve 12K space > > File A: > 0 4K 8K 12K > |<--------dirty-------->| > reserved: 12K > > 2)Write [0,4K) into file A > 0 4K 8K 12K > |<--------dirty-------->| > reserved: 16K <<< Should be 12K > > 3) Commit transaction > Dirty pages [0,12) written to disk. > Free 12K reserved space. > reserved: 4K <<< Should be 0 > > This testcase will test such problem. > Kernel fix will need some huge change, so won't be soon. > > Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com> > Reviewed-by: Filipe Manana <fdmanana@suse.com> > --- > Changelog: > v2: > Use smaller write size inside loop, in case commit is trigger by dirty > page threshold, and ensure following write won't trigger EQUOT > v3: > Add more comments and fix some expression. > v4: > Rebase to latest fstests. > v5: > Use larger number to avoid conflicts with the test from Filipe Manana. > --- > tests/btrfs/099 | 86 +++++++++++++++++++++++++++++++++++++++++++++++++++++ > tests/btrfs/099.out | 13 ++++++++ > tests/btrfs/group | 1 + > 3 files changed, 100 insertions(+) > create mode 100755 tests/btrfs/099 > create mode 100644 tests/btrfs/099.out > > diff --git a/tests/btrfs/099 b/tests/btrfs/099 > new file mode 100755 > index 0000000..ba531e1 > --- /dev/null > +++ b/tests/btrfs/099 > @@ -0,0 +1,86 @@ > +#! /bin/bash > +# FS QA Test 099 > +# > +# Check for qgroup reserved space leaks caused by re-writing dirty ranges > +# This bug has been present in btrfs qgroup for a long time > +# > +#----------------------------------------------------------------------- > +# 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 > + > +_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 > +# space reserve. > +BLOCKSIZE=$(( 2 * 1024 * 1024 )) # 2M block size > +FILESIZE=$(( 128 * 1024 * 1024 )) # 128M file size > + > +_scratch_mkfs >> $seqres.full 2>&1 Just one minor issue here, please remove $seqres.full before test, otherwise $seqres.full will keep growing. 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 wrote on 2015/08/26 14:41 +0800: > On Fri, Aug 21, 2015 at 09:17:14AM +0800, Qu Wenruo wrote: >> Btrfs qgroup reserve codes lacks check for rewrite dirty page, causing >> every write, even rewriting a uncommitted dirty page, to reserve space. >> >> But only written data will free the reserved space, causing reserved >> space leaking. >> >> The bug exists almost from the beginning of btrfs qgroup codes, but >> nobody found it. >> >> For example: >> >> 1)Write [0, 12K) into file A >> reserve 12K space >> >> File A: >> 0 4K 8K 12K >> |<--------dirty-------->| >> reserved: 12K >> >> 2)Write [0,4K) into file A >> 0 4K 8K 12K >> |<--------dirty-------->| >> reserved: 16K <<< Should be 12K >> >> 3) Commit transaction >> Dirty pages [0,12) written to disk. >> Free 12K reserved space. >> reserved: 4K <<< Should be 0 >> >> This testcase will test such problem. >> Kernel fix will need some huge change, so won't be soon. >> >> Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com> >> Reviewed-by: Filipe Manana <fdmanana@suse.com> >> --- >> Changelog: >> v2: >> Use smaller write size inside loop, in case commit is trigger by dirty >> page threshold, and ensure following write won't trigger EQUOT >> v3: >> Add more comments and fix some expression. >> v4: >> Rebase to latest fstests. >> v5: >> Use larger number to avoid conflicts with the test from Filipe Manana. >> --- >> tests/btrfs/099 | 86 +++++++++++++++++++++++++++++++++++++++++++++++++++++ >> tests/btrfs/099.out | 13 ++++++++ >> tests/btrfs/group | 1 + >> 3 files changed, 100 insertions(+) >> create mode 100755 tests/btrfs/099 >> create mode 100644 tests/btrfs/099.out >> >> diff --git a/tests/btrfs/099 b/tests/btrfs/099 >> new file mode 100755 >> index 0000000..ba531e1 >> --- /dev/null >> +++ b/tests/btrfs/099 >> @@ -0,0 +1,86 @@ >> +#! /bin/bash >> +# FS QA Test 099 >> +# >> +# Check for qgroup reserved space leaks caused by re-writing dirty ranges >> +# This bug has been present in btrfs qgroup for a long time >> +# >> +#----------------------------------------------------------------------- >> +# 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 $tmp.* >> +} >> + >> +# get standard environment, filters and checks >> +. ./common/rc >> +. ./common/filter >> + >> +# real QA test starts here >> + >> +_supported_fs btrfs >> +_supported_os Linux >> +_require_scratch >> +_need_to_be_root >> + >> +# Use big blocksize to ensure there is still enough space left for metadata >> +# space reserve. >> +BLOCKSIZE=$(( 2 * 1024 * 1024 )) # 2M block size >> +FILESIZE=$(( 128 * 1024 * 1024 )) # 128M file size >> + >> +_scratch_mkfs >> $seqres.full 2>&1 > > Just one minor issue here, please remove $seqres.full before test, > otherwise $seqres.full will keep growing. > > Thanks, > Eryu > Oh, you're right. I did forgot to remove it. BTW, is it better to add 'rm $seqres.full' to the template? Thanks, Qu -- 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
On Wed, Aug 26, 2015 at 03:21:12PM +0800, Qu Wenruo wrote: > Eryu Guan wrote on 2015/08/26 14:41 +0800: > >Just one minor issue here, please remove $seqres.full before test, > >otherwise $seqres.full will keep growing. > > Oh, you're right. > I did forgot to remove it. > > BTW, is it better to add 'rm $seqres.full' to the template? Yes, that's a good idea. Separate patch, though. ;) Cheers, Dave.
diff --git a/tests/btrfs/099 b/tests/btrfs/099 new file mode 100755 index 0000000..ba531e1 --- /dev/null +++ b/tests/btrfs/099 @@ -0,0 +1,86 @@ +#! /bin/bash +# FS QA Test 099 +# +# Check for qgroup reserved space leaks caused by re-writing dirty ranges +# This bug has been present in btrfs qgroup for a long time +# +#----------------------------------------------------------------------- +# 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 $tmp.* +} + +# get standard environment, filters and checks +. ./common/rc +. ./common/filter + +# real QA test starts here + +_supported_fs btrfs +_supported_os Linux +_require_scratch +_need_to_be_root + +# Use big blocksize to ensure there is still enough space left for metadata +# space reserve. +BLOCKSIZE=$(( 2 * 1024 * 1024 )) # 2M block size +FILESIZE=$(( 128 * 1024 * 1024 )) # 128M file size + +_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 + +# loop 5 times without sync to ensure reserved space leak will happen +for i in `seq 1 5`; do + # Use 1/4 of the file size, to ensure even commit is trigger by + # dirty page threshold or commit interval, we should still be + # able to continue write + $XFS_IO_PROG -f -c "pwrite -b $BLOCKSIZE 0 $(($FILESIZE / 4))" \ + $SCRATCH_MNT/foo | _filter_xfs_io +done + +# Sync to make sure all the dirty pages are written to disk, which should +# free all the reserved space +sync + +# remove the file and sync, to ensure all quota space freed +rm $SCRATCH_MNT/foo +sync + +# We should be able to write $FILESIZE - $BLOCKSIZE data now +$XFS_IO_PROG -f -c "pwrite -b $BLOCKSIZE 0 $(($FILESIZE - $BLOCKSIZE))" \ + $SCRATCH_MNT/foo | _filter_xfs_io + +# success, all done +status=0 +exit diff --git a/tests/btrfs/099.out b/tests/btrfs/099.out new file mode 100644 index 0000000..858bd38 --- /dev/null +++ b/tests/btrfs/099.out @@ -0,0 +1,13 @@ +QA output created by 099 +wrote 33554432/33554432 bytes at offset 0 +XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) +wrote 33554432/33554432 bytes at offset 0 +XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) +wrote 33554432/33554432 bytes at offset 0 +XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) +wrote 33554432/33554432 bytes at offset 0 +XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) +wrote 33554432/33554432 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 0 +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 e13865a..68acbcf 100644 --- a/tests/btrfs/group +++ b/tests/btrfs/group @@ -100,3 +100,4 @@ 095 auto quick metadata 096 auto quick clone 097 auto quick send clone +099 auto quick qgroup