Re: max_dir_size_kb option list

[Dave Chinner <david@fromorbit.com>]

On Tue, Dec 23, 2014 at 03:06:32PM +0300, Alexander Tsvetkov wrote:
> Hello Dave,
> 
> Attached updated version of the test.

[snip entire quoted previous emails]

Closer, but still not quite all there.

A couple of things about posting patches, first. please don't top
post and when sending a new version of a patch, please send it with
an appropriate subject such as:

[PATCH v4] ext4: Add new test for max_dir_size_kb mount option

So it's clear that there's a new version of the patch been sent.
Also, the patch should be in-line, not an attachment.

> From 6f90894347d579e6cc6be9af159eb5d4a12c059e Mon Sep 17 00:00:00 2001
> From: Alexander Tsvetkov <alexander.tsvetkov@oracle.com>
> Date: Tue, 23 Dec 2014 14:58:13 +0300
> Subject: [PATCH] added test for max_dir_size_kb mount option
> 
> ---

The commit message should not be blank - it needs to explain why the
new test is needed.

Also, the change history of the patch (what's changed between each
version posted)  should be documented here below the first "---"
delimiter so people who have already commented on previous versions
know what you have and haven't changed.

Fo more information, please go and read
Documentation/SubmittingPatches from your local kernel source tree.

>  tests/ext4/005     | 137 +++++++++++++++++++++++++++++++++++++++++++++++++++++
>  tests/ext4/005.out |   2 +
>  tests/ext4/group   |   1 +
>  3 files changed, 140 insertions(+)
>  create mode 100755 tests/ext4/005
>  create mode 100644 tests/ext4/005.out
> 
> diff --git a/tests/ext4/005 b/tests/ext4/005
> new file mode 100755
> index 0000000..4cfcb25
> --- /dev/null
> +++ b/tests/ext4/005
> @@ -0,0 +1,137 @@
> +#! /bin/bash
> +# FS QA Test
> +#
> +# Test for mount option max_dir_size_kb
> +#
> +#-----------------------------------------------------------------------
> +# Copyright (c) 2014 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
> +#-----------------------------------------------------------------------
> +#
> +
> +seq=$(basename $0)
> +seqres=$RESULT_DIR/$seq
> +tmp=/tmp/$$
> +
> +testdir=$SCRATCH_MNT/dir1.$seq
> +testdir2=$testdir/dir2.$seq
> +testfile=$SCRATCH_MNT/testfile

I think I've already pointed out that "testdir" should not be used
as a local variable to point to something on the scratch device.
It's too easily confused with TEST_DIR. Please rename them to
something less confusing e.g. dir1, dir2, and fsimg_file.

> +
> +echo "QA output created by $seq"
> +echo "Silence is golden"
> +rm -f $seqres.full
> +
> +status=1 # failure is the default!
> +trap "_cleanup; exit \$status" 0 1 2 3 15
> +
> +_cleanup()
> +{
> +	cd /
> +	rm -rf $tmp.*
> +}
> +
> +# filter expected output with ENOSPC error
> +filter_enospc()
> +{
> +	sed -e "/^.*No space left on device$/d"
> +}
> +
> +# $1 - expected limit after filling
> +# $2 - where to create
> +create_items() 
> +{
> +	limit=$1
> +	dir=${2:-$testdir}

Just pass the variable at the call site.

> +	sync
> +	echo 3 > /proc/sys/vm/drop_caches

Why is this necessary given there's always a remount just before
this is called?

> +	MAX_INUM=$((limit * 1024 * 3 / 24))
> +	for i in $(seq 0 $MAX_INUM); do
> +		touch $dir/$i 2>&1 1>/dev/null | filter_enospc

touch is silent when it succeeds, so there's no need to redirect
anything to /dev/null.

> +		if [ ${PIPESTATUS[0]} -ne 0 ]; then
> +			break
> +		fi
> +	done
> +	size=$(stat -c %s $dir)
> +	size=$((size / 1024))
> +	if [ $size -gt $limit ]; then
> +		echo "FAIL! expected dir size: $limit, actually: $size"
> +	fi
> +}
> +
> +# $1 - low directory limit value
> +# $2 - high directory limit value 
> +# $3 - mkfs options
> +run_test()
> +{
> +	LIMIT1=$1
> +	LIMIT2=$2
> +	MKFS_OPT=$3
> +
> +	_scratch_mkfs $MKFS_OPT >>$seqres.full 2>&1
> +	_scratch_mount -o max_dir_size_kb=$LIMIT1
> +	mkdir $testdir
> +
> +	# Exceed with low limit
> +	create_items $LIMIT1
> +
> +	# Exceed with the same limit after remount
> +	_scratch_mount "-o remount,max_dir_size_kb=$LIMIT1"

Urk. We have _scratch_remount, so this is kinda confusing as to
be using _scratch_mount to do remounts. You need to document why
this is a safe thing to do.

> +	create_items $LIMIT1

And. well, after a remount, running sync and dropping caches is
unnecessary...

> +	# Exceed with high limit after remount 
> +	_scratch_mount "-o remount,max_dir_size_kb=$LIMIT2"
> +	create_items $LIMIT2
> +
> +	# Exceed with low limit after remount 
> +	_scratch_mount "-o remount,max_dir_size_kb=$LIMIT1"
> +	create_items $LIMIT2
> +
> +	# Exceed limits of two test dirs resided on different fs, 
> +	# second fs is mounted on nested test dir of the first fs
> +	rm -fr $testdir/*
> +	rmdir $testdir
> +	mkdir -p $testdir2
> +	touch $testfile
> +	$MKFS_EXT4_PROG -F $MKFS_OPT $testfile 4m >> $seqres.full 2>&1

_mkfs_dev $testfile 4m ?

> +	_mount -o loop,max_dir_size_kb=$LIMIT2 $testfile $testdir2
> +	create_items $LIMIT1
> +	create_items $LIMIT2 $testdir2
> +	_scratch_mount "-o remount,max_dir_size_kb=$LIMIT2"
> +	_mount -o remount,max_dir_size_kb=$LIMIT1 $testfile $testdir2
> +	create_items $LIMIT2
> +	create_items $LIMIT2 $testdir2
> +
> +	umount -d $testdir2
> +	_scratch_unmount >/dev/null 2>&1
> +}
> +
> +# get standard environment, filters and checks
> +. ./common/rc
> +
> +# real QA test starts here
> +
> +_supported_fs ext4
> +_supported_os Linux
> +_require_scratch
> +_require_loop

Put these at the top after the _cleanup function.

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com