Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   =?utf-8?Q?Lu=C3=ADs_Henriques?= <lhenriques@suse.de>
To:     Joseph Qi <joseph.qi@linux.alibaba.com>
Cc:     ocfs2-devel@oss.oracle.com, linux-kernel@vger.kernel.org,
        Mark Fasheh <mark@fasheh.com>,
        Joel Becker <jlbec@evilplan.org>,
        Heming Zhao <heming.zhao@suse.com>
Subject: Re: [PATCH] ocfs2: fix use-after-free when unmounting read-only
 filesystem
References: <20230522102506.9205-1-lhenriques@suse.de>
        <a57047d6-8f57-3dde-f15f-b24f63d675cd@linux.alibaba.com>
Date:   Mon, 22 May 2023 13:23:07 +0100
In-Reply-To: <a57047d6-8f57-3dde-f15f-b24f63d675cd@linux.alibaba.com> (Joseph
        Qi's message of "Mon, 22 May 2023 20:01:25 +0800")
Message-ID: <87h6s47dxw.fsf@brahms.olymp>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

Joseph Qi <joseph.qi@linux.alibaba.com> writes:

> On 5/22/23 6:25 PM, Lu=C3=ADs Henriques wrote:
>> It's trivial to trigger a use-after-free bug in the ocfs2 quotas code us=
ing
>> fstest generic/452.  After mounting a filesystem as read-only, quotas are
>
> generic/452 is for testing ext4 mounted with dax and ro.
> But ocfs2 doesn't support dax yet.

Right, but I think it's still useful to run the 'generic' test-suite in a
filesystem.  We can always find issues in the test itself or, in this
case, a bug in the filesystem.

>> suspended and ocfs2_mem_dqinfo is freed through ->ocfs2_local_free_info(=
).  When
>> unmounting the filesystem, an UAF access to the oinfo will eventually ca=
use a
>> crash.
>
> In ocfs2_fill_super(), it won't enable quota if is a readonly mount.
> Do you mean remount as readonly?

Yes, sorry. Instead of "mounting", the patch changelog should say

  "After remounting a filesystem as read-only..."

Cheers,
--=20
Lu=C3=ADs

>
> Thanks,
> Joseph
>
>>=20
>> Cc: <stable@vger.kernel.org>
>> Signed-off-by: Lu=C3=ADs Henriques <lhenriques@suse.de>
>> ---
>>  fs/ocfs2/super.c | 6 ++++--
>>  1 file changed, 4 insertions(+), 2 deletions(-)
>>=20
>> diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c
>> index 0b0e6a132101..988d1c076861 100644
>> --- a/fs/ocfs2/super.c
>> +++ b/fs/ocfs2/super.c
>> @@ -952,8 +952,10 @@ static void ocfs2_disable_quotas(struct ocfs2_super=
 *osb)
>>  	for (type =3D 0; type < OCFS2_MAXQUOTAS; type++) {
>>  		if (!sb_has_quota_loaded(sb, type))
>>  			continue;
>> -		oinfo =3D sb_dqinfo(sb, type)->dqi_priv;
>> -		cancel_delayed_work_sync(&oinfo->dqi_sync_work);
>> +		if (!sb_has_quota_suspended(sb, type)) {
>> +			oinfo =3D sb_dqinfo(sb, type)->dqi_priv;
>> +			cancel_delayed_work_sync(&oinfo->dqi_sync_work);
>> +		}
>>  		inode =3D igrab(sb->s_dquot.files[type]);
>>  		/* Turn off quotas. This will remove all dquot structures from
>>  		 * memory and so they will be automatically synced to global