Received-SPF: pass (google.com: best guess record for domain of selinux-refpolicy-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Subject: Re: systemd_user_runtime_dir_t
To:     Russell Coker <russell@coker.com.au>,
        Topi Miettinen <toiwoton@gmail.com>
Cc:     selinux-refpolicy@vger.kernel.org
References: <2379388.043O1XupSG@liv>
 <b4ec6d11-1f0d-1f3c-7ee9-d1f9fd99ce1b@gmail.com> <3285433.FrU4F1G8SV@liv>
From:   Chris PeBenito <pebenito@ieee.org>
Message-ID: <01acc6a9-8c66-2c64-bfe2-3b44274c366e@ieee.org>
Date:   Thu, 9 Apr 2020 09:47:54 -0400
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101
 Thunderbird/68.6.0
MIME-Version: 1.0
In-Reply-To: <3285433.FrU4F1G8SV@liv>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Sender: selinux-refpolicy-owner@vger.kernel.org
Precedence: bulk

On 4/8/20 2:48 AM, Russell Coker wrote:
> On Wednesday, 8 April 2020 4:43:32 PM AEST Topi Miettinen wrote:
>> On 8.4.2020 5.19, Russell Coker wrote:
>>> What kind of name is that?  It's a poor choice of name by systemd upstream
>>> and I don't think we are obliged to copy that bad idea.
>>
>> The name is related to environment variable XDG_USER_RUNTIME_DIR, which
>> is not systemd specific construct:
> 
> That's a fine name for an environment variable, clearly indicating that it
> points to a directory.
> 
> It's a poor name for a domain, making a casual observer think it's a type for
> a directory not a domain for a process.
> 
>> https://specifications.freedesktop.org/basedir-spec/basedir-spec-latest.html
>>
>> So if the name should change, xdg_user_runtime_dir_t might be more
>> accurate.
> 
> It would be more accurate to have a name that reflects the function of managing
> the directory in question.
> 
>> On the other hand, the current name reflects that it's
>> specifically the service supplied by systemd called
>> "runtime-user-dir@.service" which provides the directory. If the
>> provider would be something other (/etc/init.d/elogind-mkuserdir), it
>> could be equally accurate to label the directory with something else.
> 
> If the provider changed to a different process (theoretically we are supposed
> to be able to swap out parts of systemd for equivalents) then it should have
> the same domain name.
> 
>> Though if the result (temporary directory owned by the user) is no
>> different in either case from the point of view of SELinux policy, why
>> change?
> 
> SE Linux is difficult enough to understand without making things needlessly
> confusing.

You're right.  I wasn't pleased when I came across it.  However, the way I see 
it, the direction you suggest is confusing in a different way, since it doesn't 
obviously align with the systemd command name.  I see it as being between a rock 
and a hard place, so I'm inclined to keep it as-is. This isn't a domain that is 
going to show up all over the place, it is a microscopic niche of the byzantine 
set of processes under the systemd umbrella.


-- 
Chris PeBenito