[Info-vax] VMS Privileges Versus Linux Capabilities

[Stephen Hoffman]

On 2016-06-16 23:03:46 +0000, lawrencedo99 at gmail.com said:

> On Friday, June 17, 2016 at 8:53:18 AM UTC+12, Simon Clubley wrote:
> 
>> The difference between the models is that in VMS there's no such thing 
>> as a fully privileged image or fully privileged user at least in the 
>> sense that is meant under Unix so you don't have to worry about 
>> emulating a root account or suid binaries under VMS.

Any user or installed image with an all-class privilege or any 
user-written system service or user-written system service, or any 
privileged server application, or any device driver or any execlet, or 
any hunk invoked from that context — which can potentially include a 
group-writable LOGIN.COM procedure of a privileged user, for instance — 
or ... whatever... is still a target, and the associated security needs 
to be reviewed.   UWSS and drivers and execlets and ACPs, and images 
installed with any ALL-class privilege — and other such constructs — 
are already or can become fully privileged, with complete system 
access.  Any code in any inner-mode is fully privileged.   Etc.

> Linux also has security options like SELinux or AppArmor. With one of 
> these enabled, even running as root will not give you unchecked access 
> to the system.

SEVMS was the mandatory access control variant of OpenVMS: 
http://h71000.www7.hp.com/openvms/products/sevms/info.html

OpenVMS lacks sandboxes or jails or a BSD-style pledge() mechanism, 
among other constructs.

What OpenVMS calls a subsystem identifier — an ACL-based entitlement 
for executables — can be quite useful.




-- 
Pure Personal Opinion | HoffmanLabs LLC