Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Subject: Re: [PATCH v5 5/5] sidechannel: Linux Security Module for sidechannel
To:     James Morris <jmorris@namei.org>,
        Casey Schaufler <casey.schaufler@intel.com>
Cc:     kristen@linux.intel.com, kernel-hardening@lists.openwall.com,
        deneen.t.dock@intel.com, linux-kernel@vger.kernel.org,
        dave.hansen@intel.com, linux-security-module@vger.kernel.org,
        selinux@tycho.nsa.gov, arjan@linux.intel.com
References: <20180926203446.2004-1-casey.schaufler@intel.com>
 <20180926203446.2004-6-casey.schaufler@intel.com>
 <alpine.LRH.2.21.1809280744560.8410@namei.org>
From:   Casey Schaufler <casey@schaufler-ca.com>
Message-ID: <025d4742-5947-545e-f603-502a0c5ee03f@schaufler-ca.com>
Date:   Thu, 27 Sep 2018 15:39:54 -0700
User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:52.0) Gecko/20100101
 Thunderbird/52.9.1
MIME-Version: 1.0
In-Reply-To: <alpine.LRH.2.21.1809280744560.8410@namei.org>
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit
Content-Language: en-US
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On 9/27/2018 2:45 PM, James Morris wrote:
> On Wed, 26 Sep 2018, Casey Schaufler wrote:
>
>> +	/*
>> +	 * Namespace checks. Considered safe if:
>> +	 *	cgroup namespace is the same
>> +	 *	User namespace is the same
>> +	 *	PID namespace is the same
>> +	 */
>> +	if (current->nsproxy)
>> +		ccgn = current->nsproxy->cgroup_ns;
>> +	if (p->nsproxy)
>> +		pcgn = p->nsproxy->cgroup_ns;
>> +	if (ccgn != pcgn)
>> +		return -EACCES;
>> +	if (current->cred->user_ns != p->cred->user_ns)
>> +		return -EACCES;
>> +	if (task_active_pid_ns(current) != task_active_pid_ns(p))
>> +		return -EACCES;
>> +	return 0;
> I really don't like the idea of hard-coding namespace security semantics 
> in an LSM.  Also, I'm not sure if these semantics make any sense.

Checks on namespaces where explicitly requested. I think
these are the most sensible, but I'm willing to be educated.
I was also requested to check on potential issues between containers,
but as there is no kernel concept of containers this is the
best I see we can do.

> It least make it user configurable.

Would you have a suggested granularity? I could have a
configuration option for each of cgroups, user and pid
namespaces but that's getting to be a lot of knobs to
twist.