From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from relay.sgi.com (relay3.corp.sgi.com [198.149.34.15]) by oss.sgi.com (Postfix) with ESMTP id EBFD97F6A for ; Sat, 9 Mar 2013 12:08:10 -0600 (CST) Received: from cuda.sgi.com (cuda1.sgi.com [192.48.157.11]) by relay3.corp.sgi.com (Postfix) with ESMTP id 8A18AAC001 for ; Sat, 9 Mar 2013 10:08:07 -0800 (PST) Received: from sandeen.net (sandeen.net [63.231.237.45]) by cuda.sgi.com with ESMTP id fM3rNYCRg2T4nDKM for ; Sat, 09 Mar 2013 10:08:06 -0800 (PST) Message-ID: <513B7A85.6070606@sandeen.net> Date: Sat, 09 Mar 2013 12:08:05 -0600 From: Eric Sandeen MIME-Version: 1.0 Subject: Re: xfstests: 300: diskspace consumption of cloned files on btrfs References: <50F9C32A.6000600@oracle.com> In-Reply-To: <50F9C32A.6000600@oracle.com> List-Id: XFS Filesystem from SGI List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Errors-To: xfs-bounces@oss.sgi.com Sender: xfs-bounces@oss.sgi.com To: Koen De Wit Cc: linux-btrfs , xfs@oss.sgi.com On 1/18/13 3:48 PM, Koen De Wit wrote: > Signed-off-by: Koen De Wit > > --- > 300 | 84 > +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ > 300.out | 4 +++ > 2 files changed, 88 insertions(+), 0 deletions(-) > create mode 100644 300 > create mode 100644 300.out > > diff --git a/300 b/300 > new file mode 100644 > index 0000000..a6f706c > --- /dev/null > +++ b/300 > @@ -0,0 +1,84 @@ > +#! /bin/bash > +# FS QA Test No. 300 > +# > +# Diskspace consumption of sparse file clones ("reflinks") on btrfs > +# - Check that a reflink does not consume space > +# - Check that a reflink starts to consume space when the original file > +# is modified > +# - Check that diskspace is freed up after deleting all files > +# > +#----------------------------------------------------------------------- > +# Copyright (c) 2013, 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 > +#----------------------------------------------------------------------- > +# > +# creator > +owner=koen.de.wit@oracle.com > + > +seq=`basename $0` > +echo "QA output created by $seq" > + > +here=`pwd` > +tmp=/tmp/$$ > +status=1 # failure is the default! > +trap "_cleanup; exit \$status" 0 1 2 3 15 > + > +_cleanup() > +{ > + cd / > + rm -f $tmp.* > +} > + > +# get standard environment, filters and checks > +. ./common.rc > +. ./common.filter > + > +# real QA test starts here > +_supported_fs btrfs > +_supported_os Linux > + > +_require_cp_reflink need def'n > +TESTDIR1=$TEST_DIR/test-$seq.$$ > +mkdir $TESTDIR1 possible clash if pids cycle? > +_free() { > + _df_dir $TEST_DIR | $AWK_PROG '{ print $5 }' > +} > + > +free_before=`_free` > +$XFS_IO_PROG -f -c "pwrite -S 0x61 0 10m" -c "fsync" $TESTDIR1/original > > /dev/null patch wrapped. FWIW sometimes redirecting to $seq.full rather than /dev/null makes it easier to see what's happened if anything goes wrong. > +cp --reflink=always $TESTDIR1/original $TESTDIR1/copy1 > +cp --reflink=auto $TESTDIR1/copy1 $TESTDIR1/copy2 > +btrfs filesystem sync $TEST_DIR > /dev/null > +free_after_creating_reflinks=`_free` > +_within_tolerance "free space after creating 1 file and 2 reflinks" \ > + `expr $free_before - $free_after_creating_reflinks` 10252 512 -v > + A comment about what's being checked here might help, unfortunately the _within_tolerance helper is a little cryptic. why should the difference be 10252, out of curiosity? Does that depend on block size or anything? > +$XFS_IO_PROG -c "pwrite -S 0x62 3m 5m" -c "fsync" $TESTDIR1/original > > /dev/null patch wrapped :) > +btrfs filesystem sync $TEST_DIR > /dev/null > +free_after_overwriting=`_free` > +_within_tolerance "free space after overwriting original file" \ > + `expr $free_before - $free_after_overwriting` 15376 512 -v Ok, so you overwrote 5 megs of the original 10 meg file now. So this is checking that the difference is within 512 bytes newly-consumed space against . . . 15376 what? It'd just be nice to have a comment for future readers. I wonder if comparing to $free_after_creating_reflinks would be clearer, so the space used is just that space used for the new data? > +rm $TESTDIR1/original $TESTDIR1/copy1 $TESTDIR1/copy2 > +btrfs filesystem sync $TEST_DIR > /dev/null > +free_after_removing=`_free` > +_within_tolerance "free space after removing all files" \ > + `expr $free_before - $free_after_removing` 0 128 -v This makes sense, make sure we're back to the original. Is the 128 slop even necessary? Need to add this test to groups as well. Thanks, -Eric > + > +# success, all done > +status=0 > +exit > diff --git a/300.out b/300.out > new file mode 100644 > index 0000000..bbb94f4 > --- /dev/null > +++ b/300.out > @@ -0,0 +1,4 @@ > +QA output created by 300 > +free space after creating 1 file and 2 reflinks is in range > +free space after overwriting original file is in range > +free space after removing all files is in range > _______________________________________________ xfs mailing list xfs@oss.sgi.com http://oss.sgi.com/mailman/listinfo/xfs