public inbox for fstests@vger.kernel.org
 help / color / mirror / Atom feed
From: Eryu Guan <guaneryu@gmail.com>
To: Jayashree <jaya@cs.utexas.edu>
Cc: fstests@vger.kernel.org, tytso@mit.edu, david@fromorbit.com,
	fdmanana@gmail.com, vijay@cs.utexas.edu
Subject: Re: [PATCH] fstest: CrashMonkey 'hard link' tests ported to xfstest
Date: Fri, 23 Nov 2018 18:52:14 +0800	[thread overview]
Message-ID: <20181123105214.GL3889@desktop> (raw)
In-Reply-To: <1542757864-14391-1-git-send-email-jaya@cs.utexas.edu>

On Tue, Nov 20, 2018 at 05:51:04PM -0600, Jayashree wrote:
> This patch aims to add more tests to the xfstest suite to check
> whether the target file system recovers correctly after a crash.
> These test cases are generated by CrashMonkey, a
> crash-consistency testing framework built at the SASLab at UT Austin.
> 
> This patch batches 37 crash-consistency tests into a xfstest test,
> each of which checks the hard link behavior under different scenarios.
> This test creates hard-links between files in the same directory or
> across directories, while allowing fsync of either the files involved,
> their parent directories, or unrelated sibling files. After each sub
> test, the metadata of the persisted file is checked for the correct
> link count. Additionally, each sub test is followed by fsck to check
> for inconsistencies. The tests run on a 256MB file system, and
> the working directory is cleaned up after every sub test.

Looks good to me! Thanks for the revision!

> 
> This test has been added to the “auto” group as advised by Eryu.
> Once a decision is made about the addition of the new “regress”
> group, this test can be moved to the appropriate group.

For now, I think we could also add 'quick' (it runs less than 30s on my
test vm) and 'metadata' group.

The two issues Filipe mentioned are easy to fix, I can fix them on
commit. But I'll let others take a look at this new version first, so
I'll leave it on the list for a while.

Thanks,
Eryu

> 
> Signed-off-by: Jayashree Mohan <jaya@cs.utexas.edu>
> ---
>  tests/generic/520     | 183 ++++++++++++++++++++++++++++++++++++++++++++++++++
>  tests/generic/520.out |  75 +++++++++++++++++++++
>  tests/generic/group   |   1 +
>  3 files changed, 259 insertions(+)
>  create mode 100755 tests/generic/520
>  create mode 100644 tests/generic/520.out
> 
> diff --git a/tests/generic/520 b/tests/generic/520
> new file mode 100755
> index 0000000..46e82ac
> --- /dev/null
> +++ b/tests/generic/520
> @@ -0,0 +1,183 @@
> +#! /bin/bash
> +# SPDX-License-Identifier: GPL-2.0
> +# Copyright (c) 2018 The University of Texas at Austin.  All Rights Reserved.
> +#
> +# FS QA Test 520
> +#
> +# Test case created by CrashMonkey
> +#
> +# Test if we create a hard link to a file and persist either of the files, all the names persist.
> +#
> +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()
> +{
> +	_cleanup_flakey
> +	cd /
> +	rm -f $tmp.*
> +}
> +
> +# get standard environment, filters and checks
> +. ./common/rc
> +. ./common/filter
> +. ./common/dmflakey
> +
> +# 256MB in byte
> +fssize=$((2**20 * 256))
> +
> +# remove previous $seqres.full before test
> +rm -f $seqres.full
> +
> +# real QA test starts here
> +_supported_fs generic
> +_supported_os Linux
> +_require_scratch_nocheck
> +_require_dm_target flakey
> +
> +# initialize scratch device
> +_scratch_mkfs_sized $fssize >> $seqres.full 2>&1
> +_require_metadata_journaling $SCRATCH_DEV
> +_init_flakey
> +
> +stat_opt='-c "blocks: %b size: %s inode: %i links: %h"'
> +before=""
> +after=""
> +
> +
> +# Using _scratch_mkfs instead of cleaning up the  working directory,
> +# adds about 10 seconds of delay in total for the 37 tests.
> +clean_dir()
> +{
> +	_mount_flakey
> +	rm -rf $SCRATCH_MNT/*
> +	sync
> +	_unmount_flakey
> +}
> +
> +check_consistency()
> +{
> +	_flakey_drop_and_remount | tee -a $seqres.full
> +
> +	if [ -f $1 ]; then
> +		after=`stat "$stat_opt" $1`
> +	fi
> +
> +	if [ "$before" != "$after" ] && [ $2 -ne 1 ]; then
> +		echo "Before: $before"
> +		echo "After: $after"
> +	fi
> +
> +	_unmount_flakey
> +	_check_scratch_fs $FLAKEY_DEV
> +	[ $? -ne 0 ] && _fatal "fsck failed"
> +}
> +
> +# create a hard link $2 to file $1, and fsync $3, followed by power-cut
> +test_link_fsync()
> +{
> +	local sibling=0
> +	local src=$SCRATCH_MNT/$1
> +	local dest=$SCRATCH_MNT/$2
> +	before=""
> +	after=""
> +
> +	if [ "$3" == "./" ]; then
> +		fsync=$SCRATCH_MNT
> +	else
> +		fsync=$SCRATCH_MNT/$3
> +	fi
> +
> +	echo -ne "\n=== link $src $dest  with fsync $fsync ===\n" | _filter_scratch
> +	_mount_flakey
> +
> +	# Now execute the workload
> +	# Create the directory in which the source and destination files
> +	# will be created
> +	mkdir -p "${src%/*}"
> +	mkdir -p "${dest%/*}"
> +	touch $src
> +	ln $src $dest
> +
> +	# If the file being persisted is a sibling, create it first
> +	if [ ! -f $fsync ]; then
> +		sibling=1
> +		touch $fsync
> +	fi
> +
> +	$XFS_IO_PROG -c "fsync" $fsync
> +
> +	if [ $sibling -ne 1 ]; then
> +		before=`stat "$stat_opt" $src`
> +	fi
> +
> +	check_consistency $src $sibling
> +	clean_dir
> +}
> +
> +# create a hard link $2 to file $1, and sync, followed by power-cut
> +test_link_sync()
> +{
> +	local src=$SCRATCH_MNT/$1
> +	local dest=$SCRATCH_MNT/$2
> +	before=""
> +	after=""
> +	echo -ne "\n=== link $src $dest  with sync ===\n" | _filter_scratch
> +	_mount_flakey
> +
> +	# now execute the workload
> +	# Create the directory in which the source and destination files
> +	# will be created
> +	mkdir -p "${src%/*}"
> +	mkdir -p "${dest%/*}"
> +	touch $src
> +	ln $src $dest
> +	sync
> +	before=`stat "$stat_opt" $src`
> +
> +	check_consistency $src 0
> +	clean_dir
> +}
> +
> +
> +# Create different combinations to run the link test
> +# Group 0: Both files within root directory
> +file_names[0]="foo bar"
> +fsync_names[0]="./ foo bar"
> +
> +# Group 1: Create hard link in a sub directory
> +file_names[1]="foo A/bar"
> +fsync_names[1]="./ foo bar A A/bar A/foo"
> +
> +# Group 2: Create hard link in parent directory
> +file_names[2]="A/foo bar"
> +fsync_names[2]="./ foo bar A A/bar A/foo"
> +
> +# Group 3: Both files within a directory other than root
> +file_names[3]="A/foo A/bar"
> +fsync_names[3]="./ A A/bar A/foo"
> +
> +#Group 4: Exercise name reuse : Link file in sub-directory
> +file_names[4]="bar A/bar"
> +fsync_names[4]="./ foo bar A A/bar A/foo"
> +
> +#Group 5: Exercise name reuse : Link file in parent directory
> +file_names[5]="A/bar bar"
> +fsync_names[5]="./ foo bar A A/bar A/foo"
> +
> +for ((test_group=0; test_group<6; test_group++)); do
> +	for file in ${fsync_names[$test_group]}; do
> +		test_link_fsync ${file_names[$test_group]} $file
> +	done
> +	test_link_sync ${file_names[$test_group]}
> +done
> +
> +# success, all done
> +status=0
> +exit
> diff --git a/tests/generic/520.out b/tests/generic/520.out
> new file mode 100644
> index 0000000..b0d75a1
> --- /dev/null
> +++ b/tests/generic/520.out
> @@ -0,0 +1,75 @@
> +QA output created by 520
> +
> +=== link SCRATCH_MNT/foo SCRATCH_MNT/bar  with fsync SCRATCH_MNT ===
> +
> +=== link SCRATCH_MNT/foo SCRATCH_MNT/bar  with fsync SCRATCH_MNT/foo ===
> +
> +=== link SCRATCH_MNT/foo SCRATCH_MNT/bar  with fsync SCRATCH_MNT/bar ===
> +
> +=== link SCRATCH_MNT/foo SCRATCH_MNT/bar  with sync ===
> +
> +=== link SCRATCH_MNT/foo SCRATCH_MNT/A/bar  with fsync SCRATCH_MNT ===
> +
> +=== link SCRATCH_MNT/foo SCRATCH_MNT/A/bar  with fsync SCRATCH_MNT/foo ===
> +
> +=== link SCRATCH_MNT/foo SCRATCH_MNT/A/bar  with fsync SCRATCH_MNT/bar ===
> +
> +=== link SCRATCH_MNT/foo SCRATCH_MNT/A/bar  with fsync SCRATCH_MNT/A ===
> +
> +=== link SCRATCH_MNT/foo SCRATCH_MNT/A/bar  with fsync SCRATCH_MNT/A/bar ===
> +
> +=== link SCRATCH_MNT/foo SCRATCH_MNT/A/bar  with fsync SCRATCH_MNT/A/foo ===
> +
> +=== link SCRATCH_MNT/foo SCRATCH_MNT/A/bar  with sync ===
> +
> +=== link SCRATCH_MNT/A/foo SCRATCH_MNT/bar  with fsync SCRATCH_MNT ===
> +
> +=== link SCRATCH_MNT/A/foo SCRATCH_MNT/bar  with fsync SCRATCH_MNT/foo ===
> +
> +=== link SCRATCH_MNT/A/foo SCRATCH_MNT/bar  with fsync SCRATCH_MNT/bar ===
> +
> +=== link SCRATCH_MNT/A/foo SCRATCH_MNT/bar  with fsync SCRATCH_MNT/A ===
> +
> +=== link SCRATCH_MNT/A/foo SCRATCH_MNT/bar  with fsync SCRATCH_MNT/A/bar ===
> +
> +=== link SCRATCH_MNT/A/foo SCRATCH_MNT/bar  with fsync SCRATCH_MNT/A/foo ===
> +
> +=== link SCRATCH_MNT/A/foo SCRATCH_MNT/bar  with sync ===
> +
> +=== link SCRATCH_MNT/A/foo SCRATCH_MNT/A/bar  with fsync SCRATCH_MNT ===
> +
> +=== link SCRATCH_MNT/A/foo SCRATCH_MNT/A/bar  with fsync SCRATCH_MNT/A ===
> +
> +=== link SCRATCH_MNT/A/foo SCRATCH_MNT/A/bar  with fsync SCRATCH_MNT/A/bar ===
> +
> +=== link SCRATCH_MNT/A/foo SCRATCH_MNT/A/bar  with fsync SCRATCH_MNT/A/foo ===
> +
> +=== link SCRATCH_MNT/A/foo SCRATCH_MNT/A/bar  with sync ===
> +
> +=== link SCRATCH_MNT/bar SCRATCH_MNT/A/bar  with fsync SCRATCH_MNT ===
> +
> +=== link SCRATCH_MNT/bar SCRATCH_MNT/A/bar  with fsync SCRATCH_MNT/foo ===
> +
> +=== link SCRATCH_MNT/bar SCRATCH_MNT/A/bar  with fsync SCRATCH_MNT/bar ===
> +
> +=== link SCRATCH_MNT/bar SCRATCH_MNT/A/bar  with fsync SCRATCH_MNT/A ===
> +
> +=== link SCRATCH_MNT/bar SCRATCH_MNT/A/bar  with fsync SCRATCH_MNT/A/bar ===
> +
> +=== link SCRATCH_MNT/bar SCRATCH_MNT/A/bar  with fsync SCRATCH_MNT/A/foo ===
> +
> +=== link SCRATCH_MNT/bar SCRATCH_MNT/A/bar  with sync ===
> +
> +=== link SCRATCH_MNT/A/bar SCRATCH_MNT/bar  with fsync SCRATCH_MNT ===
> +
> +=== link SCRATCH_MNT/A/bar SCRATCH_MNT/bar  with fsync SCRATCH_MNT/foo ===
> +
> +=== link SCRATCH_MNT/A/bar SCRATCH_MNT/bar  with fsync SCRATCH_MNT/bar ===
> +
> +=== link SCRATCH_MNT/A/bar SCRATCH_MNT/bar  with fsync SCRATCH_MNT/A ===
> +
> +=== link SCRATCH_MNT/A/bar SCRATCH_MNT/bar  with fsync SCRATCH_MNT/A/bar ===
> +
> +=== link SCRATCH_MNT/A/bar SCRATCH_MNT/bar  with fsync SCRATCH_MNT/A/foo ===
> +
> +=== link SCRATCH_MNT/A/bar SCRATCH_MNT/bar  with sync ===
> -- 
> 2.7.4
> 

  parent reply	other threads:[~2018-11-23 21:36 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-11-20 23:51 [PATCH] fstest: CrashMonkey 'hard link' tests ported to xfstest Jayashree
2018-11-21 10:47 ` Filipe Manana
2018-11-23 10:52 ` Eryu Guan [this message]
2018-11-28 14:31   ` Filipe Manana

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20181123105214.GL3889@desktop \
    --to=guaneryu@gmail.com \
    --cc=david@fromorbit.com \
    --cc=fdmanana@gmail.com \
    --cc=fstests@vger.kernel.org \
    --cc=jaya@cs.utexas.edu \
    --cc=tytso@mit.edu \
    --cc=vijay@cs.utexas.edu \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox