public inbox for linux-xfs@vger.kernel.org
 help / color / mirror / Atom feed
* [PATCH] xfstests generic/320: heavy rm workload test
@ 2013-10-21 13:01 Eryu Guan
  2013-10-21 22:29 ` Dave Chinner
  2013-10-22  8:12 ` [PATCH v2] " Eryu Guan
  0 siblings, 2 replies; 4+ messages in thread
From: Eryu Guan @ 2013-10-21 13:01 UTC (permalink / raw)
  To: xfs; +Cc: Eryu Guan

This test is based on generic/273, a regression test for commit

9a3a5da xfs: check for stale inode before acquiring iflock on push

On unpatched kernel, rm processes would hang.

Signed-off-by: Eryu Guan <eguan@redhat.com>
---
 tests/generic/320     | 139 ++++++++++++++++++++++++++++++++++++++++++++++++++
 tests/generic/320.out |   2 +
 tests/generic/group   |   2 +-
 3 files changed, 142 insertions(+), 1 deletion(-)
 create mode 100755 tests/generic/320
 create mode 100644 tests/generic/320.out

diff --git a/tests/generic/320 b/tests/generic/320
new file mode 100755
index 0000000..afe35a3
--- /dev/null
+++ b/tests/generic/320
@@ -0,0 +1,139 @@
+#! /bin/bash
+# FS QA Test No. generic/320
+#
+# heavy rm workload
+#
+# Regression test for commit:
+# 9a3a5da xfs: check for stale inode before acquiring iflock on push
+#
+# Based on generic/273
+#
+#-----------------------------------------------------------------------
+# Copyright (c) 2011-2012 Fujitsu, Inc.  All Rights Reserved.
+# Copyright (c) 2013 Red Hat, Inc.  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 -rf $tmp.*
+}
+
+. ./common/rc
+. ./common/filter
+
+threads=50
+count=2
+
+_threads_set()
+{
+	_cpu_num=`grep -c processor /proc/cpuinfo`
+	threads=$(($_cpu_num * 50))
+	if [ $threads -gt 200 ]
+	then
+		threads=200
+	fi
+}
+
+_file_create()
+{
+	_i=0
+
+	if ! mkdir $SCRATCH_MNT/origin
+	then
+		echo "mkdir origin err"
+		status=1
+		exit
+	fi
+
+	cd $SCRATCH_MNT/origin
+
+	_disksize=`df --block-size=1 $SCRATCH_DEV | $AWK_PROG -v sd=$SCRATCH_DEV 'BEGIN{c=0}{for(i=1;i<=NF;++i){a[c]=$i;++c}}END{for(entry in a){if(a[entry] ~ sd){print a[entry + 3]; break}}}'`
+	_disksize=$(($_disksize / 3))
+	_num=$(($_disksize / $count / $threads / 4096))
+	_count=$count
+	while [ $_i -lt $_num ]
+	do
+		dd if=/dev/zero of=file_$_i bs=4096 count=$_count >/dev/null 2>&1
+		_i=$(($_i + 1))
+	done
+
+	cd $here
+}
+
+_worker()
+{
+	_suffix=$1
+
+	if ! mkdir $SCRATCH_MNT/sub_$_suffix
+	then
+		echo "mkdir sub_xxx err"
+		status=1
+		exit
+	fi
+
+	cp -r $SCRATCH_MNT/origin $SCRATCH_MNT/sub_$_suffix >>$seqres.full 2>&1
+	rm -rf $SCRATCH_MNT/sub_$_suffix
+}
+
+_do_workload()
+{
+	_pids=""
+	_pid=1
+
+	_threads_set
+	_file_create
+
+	_threads=$threads
+
+	while [ $_pid -lt $_threads ]
+	do
+		_worker $_pid &
+		_pids="$_pids $!"
+		_pid=$(($_pid + 1))
+	done
+
+	wait $_pids
+}
+
+# real QA test starts here
+_supported_fs generic
+_supported_os IRIX Linux
+_require_scratch
+
+echo "Silence is golden"
+
+rm -f $seqres.full
+
+_scratch_mkfs_sized $((2 * 1024 * 1024 * 1024)) >>$seqres.full 2>&1
+_scratch_mount >>$seqres.full 2>&1
+
+_do_workload
+
+_check_scratch_fs
+status=$?
+exit
diff --git a/tests/generic/320.out b/tests/generic/320.out
new file mode 100644
index 0000000..1c4165d
--- /dev/null
+++ b/tests/generic/320.out
@@ -0,0 +1,2 @@
+QA output created by 320
+Silence is golden
diff --git a/tests/generic/group b/tests/generic/group
index 3488c42..40da6dc 100644
--- a/tests/generic/group
+++ b/tests/generic/group
@@ -122,4 +122,4 @@
 317 auto metadata quick
 318 acl attr auto quick
 319 acl auto quick
-
+320 auto rw
-- 
1.8.3.1

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

^ permalink raw reply related	[flat|nested] 4+ messages in thread

* Re: [PATCH] xfstests generic/320: heavy rm workload test
  2013-10-21 13:01 [PATCH] xfstests generic/320: heavy rm workload test Eryu Guan
@ 2013-10-21 22:29 ` Dave Chinner
  2013-10-22  3:56   ` Eryu Guan
  2013-10-22  8:12 ` [PATCH v2] " Eryu Guan
  1 sibling, 1 reply; 4+ messages in thread
From: Dave Chinner @ 2013-10-21 22:29 UTC (permalink / raw)
  To: Eryu Guan; +Cc: xfs

On Mon, Oct 21, 2013 at 09:01:24PM +0800, Eryu Guan wrote:
> This test is based on generic/273, a regression test for commit
> 
> 9a3a5da xfs: check for stale inode before acquiring iflock on push
> 
> On unpatched kernel, rm processes would hang.
> 
> Signed-off-by: Eryu Guan <eguan@redhat.com>
> ---
>  tests/generic/320     | 139 ++++++++++++++++++++++++++++++++++++++++++++++++++
>  tests/generic/320.out |   2 +
>  tests/generic/group   |   2 +-
>  3 files changed, 142 insertions(+), 1 deletion(-)
>  create mode 100755 tests/generic/320
>  create mode 100644 tests/generic/320.out
> 
> diff --git a/tests/generic/320 b/tests/generic/320
> new file mode 100755
> index 0000000..afe35a3
> --- /dev/null
> +++ b/tests/generic/320
> @@ -0,0 +1,139 @@
> +#! /bin/bash
> +# FS QA Test No. generic/320
> +#
> +# heavy rm workload
> +#
> +# Regression test for commit:
> +# 9a3a5da xfs: check for stale inode before acquiring iflock on push
> +#
> +# Based on generic/273
> +#
> +#-----------------------------------------------------------------------
> +# Copyright (c) 2011-2012 Fujitsu, Inc.  All Rights Reserved.
> +# Copyright (c) 2013 Red Hat, Inc.  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 -rf $tmp.*
> +}
> +
> +. ./common/rc
> +. ./common/filter
> +
> +threads=50
> +count=2
> +
> +_threads_set()
> +{
> +	_cpu_num=`grep -c processor /proc/cpuinfo`
> +	threads=$(($_cpu_num * 50))
> +	if [ $threads -gt 200 ]
> +	then
> +		threads=200
> +	fi
> +}
> +
> +_file_create()
> +{
> +	_i=0
> +
> +	if ! mkdir $SCRATCH_MNT/origin
> +	then
> +		echo "mkdir origin err"
> +		status=1
> +		exit
> +	fi

mkdir on a scratch device you just created shouldn't ever fail. We
don't ibother adding noise to tests for such basic functionality
unless we are specifically testing basic functionality.

> +	cd $SCRATCH_MNT/origin

ORIGIN=$SCRATCH_MNT/origin

> +
> +	_disksize=`df --block-size=1 $SCRATCH_DEV | $AWK_PROG -v sd=$SCRATCH_DEV 'BEGIN{c=0}{for(i=1;i<=NF;++i){a[c]=$i;++c}}END{for(entry in a){if(a[entry] ~ sd){print a[entry + 3]; break}}}'`

What's all that about?

	df --block-size=1 $SCRATCH_DEV | awk '/^\// {print $4}'

gives the same result.

But given that you do a scratch_mkfs_sized call belowi and then you
reduce the "disksize" by a factor of 3, why are you even probing to
calculate the size of the filesystem?

> +	_disksize=$(($_disksize / 3))
> +	_num=$(($_disksize / $count / $threads / 4096))
> +	_count=$count
> +	while [ $_i -lt $_num ]
> +	do
> +		dd if=/dev/zero of=file_$_i bs=4096 count=$_count >/dev/null 2>&1
> +		_i=$(($_i + 1))
> +	done

Use XFS_IO_PROG, not dd. and use $ORIGIN/file.$_i so you don't need
to change directories.

> +
> +	cd $here
> +}
> +
> +_worker()
> +{
> +	_suffix=$1
> +
> +	if ! mkdir $SCRATCH_MNT/sub_$_suffix
> +	then
> +		echo "mkdir sub_xxx err"
> +		status=1
> +		exit
> +	fi

Again, no need to test this.
> +
> +	cp -r $SCRATCH_MNT/origin $SCRATCH_MNT/sub_$_suffix >>$seqres.full 2>&1
> +	rm -rf $SCRATCH_MNT/sub_$_suffix
> +}
> +
> +_do_workload()
> +{
> +	_pids=""
> +	_pid=1
> +
> +	_threads_set
> +	_file_create
> +
> +	_threads=$threads
> +
> +	while [ $_pid -lt $_threads ]
> +	do
> +		_worker $_pid &
> +		_pids="$_pids $!"
> +		_pid=$(($_pid + 1))
> +	done
> +
> +	wait $_pids
> +}

Can we get rid of all these leading "_" from the variables? That is
supposed to be used for library functions, not for variables local
to the test itself.

> +
> +# real QA test starts here
> +_supported_fs generic
> +_supported_os IRIX Linux
> +_require_scratch

these should be called before any local test functions.

> +echo "Silence is golden"
> +
> +rm -f $seqres.full
> +
> +_scratch_mkfs_sized $((2 * 1024 * 1024 * 1024)) >>$seqres.full 2>&1
> +_scratch_mount >>$seqres.full 2>&1
> +
> +_do_workload
> +
> +_check_scratch_fs
> +status=$?

_check_scratch_fs will exit if it fails, so status=0 is all you need
there.

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH] xfstests generic/320: heavy rm workload test
  2013-10-21 22:29 ` Dave Chinner
@ 2013-10-22  3:56   ` Eryu Guan
  0 siblings, 0 replies; 4+ messages in thread
From: Eryu Guan @ 2013-10-22  3:56 UTC (permalink / raw)
  To: Dave Chinner; +Cc: xfs

On Tue, Oct 22, 2013 at 09:29:55AM +1100, Dave Chinner wrote:
> On Mon, Oct 21, 2013 at 09:01:24PM +0800, Eryu Guan wrote:
> > This test is based on generic/273, a regression test for commit
> > 
> > 9a3a5da xfs: check for stale inode before acquiring iflock on push
> > 
> > On unpatched kernel, rm processes would hang.
> > 
> > Signed-off-by: Eryu Guan <eguan@redhat.com>
> > ---
> >  tests/generic/320     | 139 ++++++++++++++++++++++++++++++++++++++++++++++++++
> >  tests/generic/320.out |   2 +
> >  tests/generic/group   |   2 +-
> >  3 files changed, 142 insertions(+), 1 deletion(-)
> >  create mode 100755 tests/generic/320
> >  create mode 100644 tests/generic/320.out
> > 
> > diff --git a/tests/generic/320 b/tests/generic/320
> > new file mode 100755
> > index 0000000..afe35a3
> > --- /dev/null
> > +++ b/tests/generic/320
> > @@ -0,0 +1,139 @@
> > +#! /bin/bash
> > +# FS QA Test No. generic/320
> > +#
> > +# heavy rm workload
> > +#
> > +# Regression test for commit:
> > +# 9a3a5da xfs: check for stale inode before acquiring iflock on push
> > +#
> > +# Based on generic/273
> > +#
> > +#-----------------------------------------------------------------------
> > +# Copyright (c) 2011-2012 Fujitsu, Inc.  All Rights Reserved.
> > +# Copyright (c) 2013 Red Hat, Inc.  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 -rf $tmp.*
> > +}
> > +
> > +. ./common/rc
> > +. ./common/filter
> > +
> > +threads=50
> > +count=2
> > +
> > +_threads_set()
> > +{
> > +	_cpu_num=`grep -c processor /proc/cpuinfo`
> > +	threads=$(($_cpu_num * 50))
> > +	if [ $threads -gt 200 ]
> > +	then
> > +		threads=200
> > +	fi
> > +}
> > +
> > +_file_create()
> > +{
> > +	_i=0
> > +
> > +	if ! mkdir $SCRATCH_MNT/origin
> > +	then
> > +		echo "mkdir origin err"
> > +		status=1
> > +		exit
> > +	fi
> 
> mkdir on a scratch device you just created shouldn't ever fail. We
> don't ibother adding noise to tests for such basic functionality
> unless we are specifically testing basic functionality.

I copy from most of code from generic/273 and did minor modifications
& cleanups, I should have done major cleanups :)

Thanks for the detailed review, I'll address them all in next version.

Thanks!

Eryu Guan

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

^ permalink raw reply	[flat|nested] 4+ messages in thread

* [PATCH v2] xfstests generic/320: heavy rm workload test
  2013-10-21 13:01 [PATCH] xfstests generic/320: heavy rm workload test Eryu Guan
  2013-10-21 22:29 ` Dave Chinner
@ 2013-10-22  8:12 ` Eryu Guan
  1 sibling, 0 replies; 4+ messages in thread
From: Eryu Guan @ 2013-10-22  8:12 UTC (permalink / raw)
  To: xfs; +Cc: Eryu Guan

This test is based on generic/273, a regression test for commit

9a3a5da xfs: check for stale inode before acquiring iflock on push

On unpatched kernel, rm processes would hang.

Signed-off-by: Eryu Guan <eguan@redhat.com>
---
v2: Address Dave's comments
 - don't check return value of mkdir
 - don't cd to origin dir
 - don't calculate fs size, already have fs size
 - use $XFS_IO_PROG instead of dd
 - remove leading "_" from local vars
 - call _supported_fs etc. functions before any local functions
 - set status=0 before exit

 tests/generic/320     | 121 ++++++++++++++++++++++++++++++++++++++++++++++++++
 tests/generic/320.out |   2 +
 tests/generic/group   |   2 +-
 3 files changed, 124 insertions(+), 1 deletion(-)
 create mode 100755 tests/generic/320
 create mode 100644 tests/generic/320.out

diff --git a/tests/generic/320 b/tests/generic/320
new file mode 100755
index 0000000..09c533c
--- /dev/null
+++ b/tests/generic/320
@@ -0,0 +1,121 @@
+#! /bin/bash
+# FS QA Test No. generic/320
+#
+# heavy rm workload
+#
+# Regression test for commit:
+# 9a3a5da xfs: check for stale inode before acquiring iflock on push
+#
+# Based on generic/273
+#
+#-----------------------------------------------------------------------
+# Copyright (c) 2011-2012 Fujitsu, Inc.  All Rights Reserved.
+# Copyright (c) 2013 Red Hat, Inc.  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 -rf $tmp.*
+}
+
+# get standard environment, filters and checks
+. ./common/rc
+. ./common/filter
+
+# real QA test starts here
+_supported_fs generic
+_supported_os IRIX Linux
+_require_scratch
+
+threads=50
+count=2
+fs_size=$((2 * 1024 * 1024 * 1024))
+ORIGIN=$SCRATCH_MNT/origin
+
+threads_set()
+{
+	cpu_num=`grep -c processor /proc/cpuinfo`
+	threads=$(($cpu_num * 50))
+	if [ $threads -gt 200 ]
+	then
+		threads=200
+	fi
+}
+
+file_create()
+{
+	i=0
+	mkdir $ORIGIN
+
+	disksize=$(($fs_size / 3))
+	num=$(($disksize / $count / $threads / 4096))
+	while [ $i -lt $num ]; do
+		$XFS_IO_PROG -f -c "pwrite 0 $((4096*count))" $ORIGIN/file_$i >>$seqres.full 2>&1
+		i=$(($i + 1))
+	done
+}
+
+worker()
+{
+	suffix=$1
+
+	mkdir $SCRATCH_MNT/sub_$suffix
+
+	cp -r $ORIGIN/* $SCRATCH_MNT/sub_$suffix >>$seqres.full 2>&1
+	rm -rf $SCRATCH_MNT/sub_$suffix
+}
+
+do_workload()
+{
+	pids=""
+	loop=1
+
+	threads_set
+	file_create
+
+	while [ $loop -lt $threads ]; do
+		worker $loop &
+		pids="$pids $!"
+		loop=$(($loop + 1))
+	done
+
+	wait $pids
+}
+
+echo "Silence is golden"
+
+rm -f $seqres.full
+
+_scratch_mkfs_sized $fs_size >>$seqres.full 2>&1
+_scratch_mount >>$seqres.full 2>&1
+
+do_workload
+
+_check_scratch_fs
+status=0
+exit
diff --git a/tests/generic/320.out b/tests/generic/320.out
new file mode 100644
index 0000000..1c4165d
--- /dev/null
+++ b/tests/generic/320.out
@@ -0,0 +1,2 @@
+QA output created by 320
+Silence is golden
diff --git a/tests/generic/group b/tests/generic/group
index 3488c42..40da6dc 100644
--- a/tests/generic/group
+++ b/tests/generic/group
@@ -122,4 +122,4 @@
 317 auto metadata quick
 318 acl attr auto quick
 319 acl auto quick
-
+320 auto rw
-- 
1.8.3.1

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

^ permalink raw reply related	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2013-10-22  8:12 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2013-10-21 13:01 [PATCH] xfstests generic/320: heavy rm workload test Eryu Guan
2013-10-21 22:29 ` Dave Chinner
2013-10-22  3:56   ` Eryu Guan
2013-10-22  8:12 ` [PATCH v2] " Eryu Guan

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox