Re: [PATCH v3 2/2] modules:capabilities: add a per-task modules autoload restriction

From: Djalal Harouni <tixxdz@gmail.com>
To: Andy Lutomirski <luto@kernel.org>
Cc: Kees Cook <keescook@chromium.org>,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	Andrew Morton <akpm@linux-foundation.org>,
	"Serge E. Hallyn" <serge@hallyn.com>,
	"kernel-hardening@lists.openwall.com"
	<kernel-hardening@lists.openwall.com>,
	LSM List <linux-security-module@vger.kernel.org>,
	Linux API <linux-api@vger.kernel.org>,
	Dongsu Park <dpark@posteo.net>,
	Casey Schaufler <casey@schaufler-ca.com>,
	James Morris <james.l.morris@oracle.com>,
	Paul Moore <paul@paul-moore.com>,
	Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>,
	Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	Jonathan Corbet <corbet@lwn.net>, Jessica Yu <jeyu@redhat.com>,
	Rusty Russell <rusty@rustcorp.com.au>,
	Arnaldo Carvalho de Melo <acme@redhat.com>,
	Mauro
Subject: Re: [PATCH v3 2/2] modules:capabilities: add a per-task modules autoload restriction
Date: Sat, 22 Apr 2017 02:12:29 +0200	[thread overview]
Message-ID: <CAEiveUcx8fwQgXdLPeMNsTjX2KPhQKH__a-XzcHko_1aCmh4sg@mail.gmail.com> (raw)
In-Reply-To: <CALCETrXzZwz4a9uhQfdHb4L2afApTvVLPwVx640vvV-fzbCR0w@mail.gmail.com>

On Sat, Apr 22, 2017 at 1:51 AM, Andy Lutomirski <luto@kernel.org> wrote:
[...]
>>> I personally like my implicit_rights idea, and it might be interesting
>>> to prototype it.
>>
>> I don't like blocking a needed feature behind a large super-feature
>> that doesn't exist yet. We'd be able to refactor this code into using
>> such a thing in the future, so I'd prefer to move ahead with this
>> since it would stop actual exploits.
>
> I don't think the super-feature is so hard, and I think we should not
> add the per-task thing the way it's done in this patch.  Let's not add
> per-task things where the best argument for their security is "not
> sure how it would be exploited".

Actually the XFRM framework CVE-2017-7184 [1] is one real example, of
course there are others. The exploit was used on a generic distro
during a security contest that distro is Ubuntu. That distro will
never provide a module autoloading restriction by default to not harm
it's users. Consumers or containers/sandboxes then can run their
confined apps using such facilities.

These bugs will stay in embedded devices that use these generic
distros for ever.

> Anyway, I think the sysctl is really the important bit.  The per-task
> setting is icing on the cake IMO.  One upon a time autoload was more
> important, but these days modaliases are supposed to do most of the
> work.  I bet that modern distros don't need unprivileged autoload at
> all.

Actually I think they do and we can't just change that. Users may
depend on it, it is a well established facility.

Now the other problem is CAP_NET_ADMIN which does lot of things, it is
more like the CAP_SYS_ADMIN.

This is a quick list that I got from only the past months, I'm pretty
sure there are more:

* DCCP use after free CVE-2017-6074
* n_hldc CVE-2017-2636
* XFRM framework CVE-2017-7184
* L2TPv3 CVE-2016-10200

Most of these need CAP_NET_ADMIN to be autoloaded, however we also
need CAP_NET_ADMIN for other things... therefore it is better to have
an extra facility that could coexist with CAP_NET_ADMIN and other
sandbox features.

[1] http://www.openwall.com/lists/oss-security/2017/03/29/2

-- 
tixxdz