Subject: Re: kern.showallprocs implementation
To: Elad Efrat <elad@NetBSD.org>
From: Bill Studenmund <wrstuden@netbsd.org>
List: tech-kern
Date: 08/29/2005 14:20:06
--yQbNiKLmgenwUfTN
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Mon, Aug 29, 2005 at 11:39:51PM +0300, Elad Efrat wrote:
> Bill Studenmund wrote:
>=20
> [...]
> > So my vote is go with one knob now and see what folks experience. If we
> > have folks who find they really want to let a user see one thing or the
> > other, then we can revisit.
>=20
> Okay.
>=20
> Would you object to adding a sysctl node named ``security'' (either
> under ``kern'' or as a new main one), and underneath it another node,
> ``privacy'', and underneath it a knob similar to ``see_other_uids''?

Why not just use the FreeBSD name? If there is a good reason to use a=20
different name, so be it. But I am not sure that we have heard one. We=20
will (I expect) want to support it for FreeBSD compat, so if there is no=20
reason to be different, why not just make them the same?

I'm a bit uncomfortable with "privacy." I see "privacy" as an attribute of
an entity. I also see it as something controlled by the entity. But this
sysctl works a little differently. It's not that I, as user X, tell user
Y, "You can't see my stuff." It's the superuser tells user Y, "You can=20
only see your own stuff." And perhaps tells me the same thing. :-) The=20
main point is that it's for the system administrator to decide, where as I=
=20
see my privacy as something I should control. But that's just me.

Also, I am partially enamored of what FreeBSD has done with the naming
here. They are indicating. with the "bsd" node under "security", that
these controls/status are part of the "bsd" security model. Yes, "unix" =20
could have been used, as well as "uid/gid" (but that seems unwieldy). But=
=20
we leave space for future models, which our discussions indicate may soon=
=20
be coming (well, I hope so :-). I am also willing to have my, "don't=20
abstract too soon," arguement thrown back at me. :-)

Take care,

Bill

--yQbNiKLmgenwUfTN
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (NetBSD)

iD8DBQFDE3wGWz+3JHUci9cRAqqwAJ9XgPuemM5/BweNg+fMo71oB7dDLACeOaol
6BHoTY+MLop+pzCi4VKZHTs=
=jgud
-----END PGP SIGNATURE-----

--yQbNiKLmgenwUfTN--