Re: [PATCH] capabilities: introduce per-process capability bounding set (v10)

[KaiGai Kohei <kaigai@ak.jp.nec.com>]

Andrew Morgan wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> KaiGai Kohei wrote:
>> Andrew Morgan wrote:
>>> -----BEGIN PGP SIGNED MESSAGE-----
>>> Hash: SHA1
>>>
>>> KaiGai Kohei wrote:
>>>>> +    if (!!cap_issubset(*inheritable,
>>>>> +               cap_combine(target->cap_inheritable,
>>>>> +                       current->cap_bset))) {
>>>>> +        /* no new pI capabilities outside bounding set */
>>>>> +        return -EPERM;
>>>>> +    }
>>>>>  
> 
>>> Yes, the !! was a bug. The correct check is a single !.
>> I was in trouble with getting -EPERM at pam_cap.so :-)
>>
>>> (Thus, the correct check says no 'new' pI bits can be outside cap_bset.)
>> If this condition intends to dominate 'new' pI bits by 'old' pI bits masked
>> with bounding set, we should not apply cap_combine() here.
>> I think applying cap_intersect() is correct for the purpose.
> 
> The check is not meant to limit existing pI bits.
> 
> The check is meant to limit what new bits can be 'added' to pI (in the
> case that pE & CAP_SETPCAP is true).

Thanks, I got understood as I wrote in the previous reply.

BTW, could you tell me your intention about pam_cap.c is implemented
with pam_sm_authenticate() and pam_sm_setcred()?
I think it can be done with pam_sm_open_session(), and this approach
enables to reduce the iteration of reading /etc/security/capability.conf.

How do you think the idea?
-- 
OSS Platform Development Division, NEC
KaiGai Kohei <kaigai@ak.jp.nec.com>