[refpolicy] [PATCH] Make unconfined user run lvm programs in confined domain

[cpebenito@tresys.com (Christopher J. PeBenito)]

On 05/13/2014 09:55 AM, Daniel J Walsh wrote:
> 
> On 05/13/2014 08:46 AM, Christopher J. PeBenito wrote:
>> On 05/10/2014 10:45 AM, Nicolas Iooss wrote:
>>> When an unconfined user uses truecrypt to mount an encrypted file, dmsetup is
>>> called to setup a new device.  This program works with udev to configure the
>>> new device and uses SysV semaphores to synchronize states.  As udev runs
>>> dmsetup in lvm_t domain, the first dmsetup process needs to create lvm_t
>>> semaphores (not unconfined_t) and hence needs to run in lvm_t domain.
>>>
>>> More details are available in the archives on the ML:
>>> http://oss.tresys.com/pipermail/refpolicy/2014-May/007111.html
>>> ---
>>>  policy/modules/system/unconfined.te | 4 ++++
>>>  1 file changed, 4 insertions(+)
>>>
>>> diff --git a/policy/modules/system/unconfined.te b/policy/modules/system/unconfined.te
>>> index 472a39e..79f2909 100644
>>> --- a/policy/modules/system/unconfined.te
>>> +++ b/policy/modules/system/unconfined.te
>>> @@ -108,6 +108,10 @@ optional_policy(`
>>>  ')
>>>  
>>>  optional_policy(`
>>> +	lvm_run(unconfined_t, unconfined_r)
>>> +')
>>> +
>>> +optional_policy(`
>>>  	modutils_run_update_mods(unconfined_t, unconfined_r)
>>>  ')
>> Merged.
>>
> Why would we add a confinement to the unconfined domain?  I believe
> unconfined_t should stay unconfined as much as possible.
> 
> I wrote a blog about this.
> 
> https://danwalsh.livejournal.com/30084.html
> 
> The only reason to do this in the past was for correct labeling, but
> with file name transition rules, I believe almost all transitions from
> unconfined_t should be eliminated.
 
The file name transitions don't apply, as we're concerned about SysV semaphores in this case.

-- 
Chris PeBenito
Tresys Technology, LLC
www.tresys.com | oss.tresys.com