All of lore.kernel.org
 help / color / mirror / Atom feed
From: Xiao Yang <yangx.jy@cn.fujitsu.com>
To: "Darrick J. Wong" <darrick.wong@oracle.com>,
	Eryu Guan <guaneryu@gmail.com>
Cc: fstests@vger.kernel.org
Subject: Re: [PATCH v2] xfs: Regression test for vulnerable directory integrity check
Date: Mon, 4 Jun 2018 12:55:18 +0800	[thread overview]
Message-ID: <5B14C636.6090901@cn.fujitsu.com> (raw)
In-Reply-To: <20180603225655.GY12940@magnolia>

On 2018/06/04 6:56, Darrick J. Wong wrote:
> On Sun, Jun 03, 2018 at 09:37:18PM +0800, Eryu Guan wrote:
>> On Wed, May 30, 2018 at 04:53:43PM +0800, Xiao Yang wrote:
>>> If a malicious XFS contains a block+ format directory wherein the
>>> directory inode's core.mode is corrupted, and there are subdirectories
>>> of the corrupted directory, an attempt to traverse up the directory
>>> tree by running xfs_scrub will crash the kernel in __xfs_dir3_data_check.
>>>
>>> Signed-off-by: Xiao Yang<yangx.jy@cn.fujitsu.com>
>> Thanks for the update!
>>
>>> ---
>>>   tests/xfs/448     | 96 +++++++++++++++++++++++++++++++++++++++++++++++++++++++
>>>   tests/xfs/448.out |  2 ++
>>>   tests/xfs/group   |  1 +
>>>   3 files changed, 99 insertions(+)
>>>   create mode 100755 tests/xfs/448
>>>   create mode 100644 tests/xfs/448.out
>>>
>>> diff --git a/tests/xfs/448 b/tests/xfs/448
>>> new file mode 100755
>>> index 0000000..9ea9295
>>> --- /dev/null
>>> +++ b/tests/xfs/448
>>> @@ -0,0 +1,96 @@
>>> +#! /bin/bash
>>> +# FS QA Test No. 448
>>> +#
>>> +# Regression test for commit:
>>> +# 46c59736d809 ("xfs: harden directory integrity checks some more")
>> But I couldn't triger the test failre (kernel crash?) with this commit
>> reverted, with either v1 nor v2 patch. I was testing with 4.17-rc5 based
>> kernel, could you please take a look?
>>
>>> +#
>>> +# If a malicious XFS contains a block+ format directory wherein
>>> +# the directory inode's core.mode is corrupted, and there are
>>> +# subdirectories of the corrupted directory, an attempt to traverse
>>> +# up the directory tree by running xfs_scrub will crash the
>>> +# kernel in __xfs_dir3_data_check.
>>> +#
>>> +# Notice:
>>> +# we should have non fatal asserts configured, because assert
>>> +# failures triggered by the intentional corrupt would crash system.
>>> +#
>>> +#-----------------------------------------------------------------------
>>> +# Copyright (c) 2018 FUJITSU LIMITED. 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 $tmp.*
>>> +}
>>> +
>>> +# get standard environment, filters and checks
>>> +. ./common/rc
>>> +. ./common/filter
>>> +. ./common/populate
>>> +
>>> +# real QA test starts here
>>> +_supported_os Linux
>>> +_supported_fs xfs
>>> +_require_scratch_nocheck
>>> +_require_xfs_io_command "scrub"
>>> +# Corrupt XFS on purpose, and skip if assert failures would crash system.
>>> +_require_no_xfs_bug_on_assert
>>> +
>>> +rm -f "$seqres.full"
>>> +
>>> +# Format and mount
>>> +_scratch_mkfs | _filter_mkfs>  $seqres.full 2>  $tmp.mkfs || _fail "mkfs failed"
>>> +_scratch_mount
>>> +
>>> +# Get directory block size
>>> +. $tmp.mkfs
>>> +
>>> +# Create a block+(e.g. leaf) format directory
>>> +__populate_create_dir "${SCRATCH_MNT}/dir_leaf" "$((dirbsize / 12))"
>>> +
>>> +# Get the block+ directory inode and a subdirectory inode of it
>>> +subdino=$(stat -c "%i" "$(find ${SCRATCH_MNT}/dir_leaf/* -type d | head -1)")
>>> +dino=$(stat -c "%i" "${SCRATCH_MNT}/dir_leaf")
>>> +
>>> +# Get the subdirectory's generation number
>>> +_scratch_unmount
>>> +subdgen=$(_scratch_xfs_get_metadata_field "core.gen" "inode $subdino")
>>> +
>>> +# Corrupt the directory inode's core.mode
>>> +setmode="0100755"
>>> +_scratch_xfs_set_metadata_field "core.mode" "$setmode" "inode $dino">>  $seqres.full
>>> +getmode=$(_scratch_xfs_get_metadata_field "core.mode" "inode $dino")
>>> +[ "$getmode" != "$setmode" ]&&  _fail "failed to set core.mode"
>>> +
>>> +# Scrub parent directory in subdirectory (online)
>>> +_scratch_mount
>>> +$XFS_IO_PROG -x -c "scrub parent $subdino $subdgen" ${SCRATCH_MNT}>>  $seqres.full
>> And I always see "+scrub: Inappropriate ioctl for device" here with this
>> v2 patch.
> Your kernel probably needs CONFIG_XFS_ONLINE_SCRUB=y
Agreed.
Eryu, you probably need to bulid kernel with CONFIG_XFS_ONLINE_SCRUB=y.
> This test probably needs:
>
> . ./config/fuzzy
> _supports_xfs_scrub || _fail "scrub not supported"
>
> (or make a _requires_xfs_scrub_enabled helper to encapsulate that?)
_require_xfs_io_command "scrub" should get NOTRUN if kernel does not 
support scrub ioctls, but an invalid parameter
passed to "scrub probe" hit "No parameters allowed" rather than 
"Inappropriate ioctl for device" unfortunately.
Currently, the xfs_io scrub command doesn't allow the probe function to 
have any parameter, so we need to remove the
invalid parameter in _require_xfs_io_command.

Thanks,
Xiao Yang
> --D
>
>> Thanks,
>> Eryu
>>
>>> +
>>> +echo "Silence is golden"
>>> +
>>> +# success, all done
>>> +status=0
>>> +exit
>>> diff --git a/tests/xfs/448.out b/tests/xfs/448.out
>>> new file mode 100644
>>> index 0000000..b6f0a53
>>> --- /dev/null
>>> +++ b/tests/xfs/448.out
>>> @@ -0,0 +1,2 @@
>>> +QA output created by 448
>>> +Silence is golden
>>> diff --git a/tests/xfs/group b/tests/xfs/group
>>> index 51326d9..dd39d08 100644
>>> --- a/tests/xfs/group
>>> +++ b/tests/xfs/group
>>> @@ -445,3 +445,4 @@
>>>   445 auto quick filestreams
>>>   446 auto quick
>>>   447 auto mount
>>> +448 auto quick fuzzers
>>> -- 
>>> 1.8.3.1
>>>
>>>
>>>
>
> .
>




      parent reply	other threads:[~2018-06-04  4:55 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-05-24  9:32 [PATCH] xfs: Regression test for vulnerable directory integrity check Xiao Yang
2018-05-25  4:37 ` Darrick J. Wong
2018-05-25  6:33   ` Xiao Yang
2018-05-29 17:53     ` Darrick J. Wong
2018-05-30  4:04       ` Xiao Yang
2018-05-30  4:52         ` Darrick J. Wong
2018-05-30  6:58         ` Eryu Guan
2018-05-30  8:53           ` [PATCH v2] " Xiao Yang
2018-05-30 14:56             ` Darrick J. Wong
2018-06-03 13:37             ` Eryu Guan
2018-06-03 22:56               ` Darrick J. Wong
2018-06-04  4:54                 ` Eryu Guan
2018-06-04  4:50                   ` [PATCH] common/rc: Fix _require_xfs_io_command for scrub probe Xiao Yang
2018-06-04 15:44                     ` Darrick J. Wong
2018-06-04  5:00                   ` [PATCH v2] xfs: Regression test for vulnerable directory integrity check Xiao Yang
2018-06-04  4:55                 ` Xiao Yang [this message]

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=5B14C636.6090901@cn.fujitsu.com \
    --to=yangx.jy@cn.fujitsu.com \
    --cc=darrick.wong@oracle.com \
    --cc=fstests@vger.kernel.org \
    --cc=guaneryu@gmail.com \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.