linux-ext4 - Re: [PATCH] Fstest: ext4: regression test for a ENOSPC warning

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Sat, 3 Nov 2018 17:42:01 +0800
From:   Eryu Guan <guaneryu@...il.com>
To:     Liu Bo <bo.liu@...ux.alibaba.com>
Cc:     fstests@...r.kernel.org, linux-ext4@...r.kernel.org
Subject: Re: [PATCH] Fstest: ext4: regression test for a ENOSPC warning

On Fri, Nov 02, 2018 at 10:19:25AM +0800, Liu Bo wrote:
> This adds a regression test for a ENOSPC warning which can be triggered if
> '-odioread_nolock,nodelalloc' and quota are used.
> 
> The bug was fixed by patch
> "Ext4: fix ENOSPC when both quota and dioread_nolock are enabled".
> 
> Signed-off-by: Liu Bo <bo.liu@...ux.alibaba.com>
> ---
>  tests/ext4/034     | 75 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
>  tests/ext4/034.out | 17 +++++++++++++
>  tests/ext4/group   |  1 +
>  3 files changed, 93 insertions(+)
>  create mode 100755 tests/ext4/034
>  create mode 100644 tests/ext4/034.out
> 
> diff --git a/tests/ext4/034 b/tests/ext4/034
> new file mode 100755
> index 00000000..2449c3b7
> --- /dev/null
> +++ b/tests/ext4/034
> @@ -0,0 +1,75 @@
> +#! /bin/bash
> +# SPDX-License-Identifier: GPL-2.0
> +# Copyright (c) 2018 Liu Bo, Alibaba.  All Rights Reserved.
> +#
> +# FS QA Test 034
> +#
> +# Regression test for a ENOSPC warning when both quota and
> +# "-o dioread_nolock,nodelalloc" is used.
> +#
> +# The bug was fixed by commit
> +# "Ext4: fix ENOSPC when both quota and dioread_nolock are enabled"
> +#
> +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
> +. ./common/quota
> +
> +# remove previous $seqres.full before test
> +rm -f $seqres.full
> +
> +# real QA test starts here
> +
> +# Modify as appropriate.
> +_supported_fs generic

_supported_fs ext4

> +_supported_os Linux
> +_require_scratch
> +_require_quota
> +_require_nobody
> +
> +echo "Format and mount"
> +_scratch_mkfs "-O quota" > $seqres.full 2>&1

This needs a require rule to check quota feature support status, i.e.

_require_scratch_ext4_feature "quota"

> +_scratch_mount "-o dioread_nolock,nodelalloc" > $seqres.full 2>&1
> +quotaon $SCRATCH_MNT 2>/dev/null

And given that "quota" feature is enabled, quota is turned on
automatically on mount, seems there's no need to do quotaon.

> +
> +echo "Create the original file"
> +touch $SCRATCH_MNT/foobar >> $seqres.full
> +chown nobody $SCRATCH_MNT/foobar >> $seqres.full
> +
> +quotaon $SCRATCH_MNT 2>/dev/null

Duplicated quotaon?

> +echo "Set hard quota"
> +setquota -u nobody 0 10M 0 0 $SCRATCH_MNT
> +_report_quota_blocks $SCRATCH_MNT

Looks like this is not needed either, if it's only for debug purpose
just dump the output to $seqres.full

> +
> +echo "Create 2 level extent tree (btree) for foobar with a unwritten extent"
> +$XFS_IO_PROG -f -c "pwrite 0 4k" -c "falloc 4k 4k" -c "pwrite 8k 4k" -c "pwrite 20k 4k"  -c "pwrite 28k 4k" -c "pwrite 36k 4k"  -c "fsync" $SCRATCH_MNT/foobar >> $seqres.full

Too long line, wrap it :)

> +
> +xfs_io -c "fiemap -v" $SCRATCH_MNT/foobar >> $seqres.full

$XFS_IO_PROG

> +
> +echo "Convert unwritten extent to written and collapse extent tree to inode"
> +$XFS_IO_PROG -c "pwrite 4k 4k" $SCRATCH_MNT/foobar >> $seqres.full
> +
> +echo "Create a new file and do fsync to force a jbd2 commit"
> +$XFS_IO_PROG -f -c "pwrite 0 4k" -c "fsync" $SCRATCH_MNT/dummy >> $seqres.full
> +
> +echo "sync $SCRATCH_MNT to writeback"
> +$XFS_IO_PROG -c "syncfs" $SCRATCH_MNT >> $seqres.full

syncfs needs

_require_xfs_io_command "syncfs"

Perhaps _scratch_unmount or just sync to trigger the writeback?

Thanks,
Eryu

> +
> +# success, all done
> +status=0
> +exit
> diff --git a/tests/ext4/034.out b/tests/ext4/034.out
> new file mode 100644
> index 00000000..dd0c8a7c
> --- /dev/null
> +++ b/tests/ext4/034.out
> @@ -0,0 +1,17 @@
> +QA output created by 034
> +Format and mount
> +Create the original file
> +Set hard quota
> +root 20 0 0
> +nobody 0 0 10240
> +User soft hard grace
> +Block time: 7days; Inode
> +Block File limits 
> +----------------------------------------------------------------------   
> +*** for user quotas
> +   
> +   
> +Create 2 level extent tree (btree) for foobar with a unwritten extent
> +Convert unwritten extent to written and collapse extent tree to inode
> +Create a new file and do fsync to force a jbd2 commit
> +sync /mnt/scratch to writeback
> diff --git a/tests/ext4/group b/tests/ext4/group
> index b850f568..eb744a12 100644
> --- a/tests/ext4/group
> +++ b/tests/ext4/group
> @@ -36,6 +36,7 @@
>  031 auto quick dax
>  032 auto quick ioctl resize
>  033 auto ioctl resize
> +034 auto quick quota
>  271 auto rw quick
>  301 aio auto ioctl rw stress defrag
>  302 aio auto ioctl rw stress defrag
> -- 
> 2.14.2
>