On Tue, 17 Dec 2002, Matt Zimmerman wrote:

> On Tue, Dec 17, 2002 at 04:08:26PM +0100, Susanne Oberhauser wrote:
>
> > Nevertheless would you agree with me that for systmes claiming to run on
> > *Linux*, relying on the existence of a user 'root' should be ok?  This
> > would allow portable software to have just *one* platform specific backend
> > for *all* flavors of Linux, and would ease porting of such software to
> > Linux.
>
> Certainly, a program which is not expected to work on anything but an
> LSB-compliant system can make this assumption, and many others.  But in this
> specific case, it is (in many cases) in fact easier to check for uid=0 than
> username="root" anyway, and in general, there are relatively few cases where
> it makes sense to test for root privileges rather than something more
> specific.  With the continuing development and proliferation of more
> fine-grained access control systems for Linux, root will become less magic,
> and could be removed or assigned reduced privileges.


Indeed. Engarde Linux has been around a while, and it's hardened with
LIDS. There's no gurantee that root can do anything you'd want to if
you're running LIDS-enabled.

I'm pretty sure you will encounter difficulty if you're using selinux or
Bastille.

Possessors of such systems won't care whether they're LSB-compliant,
though they mare care that tests for privilege fail.

--


Cheers
John.

Join the "Linux Support by Small Businesses" list at
http://mail.computerdatasafe.com.au/mailman/listinfo/lssb

Reply via email to