Re: GNOME keyring unlocking

[p10 <pditchev gmail com>]

Very good points . It seems I have a lot more to read on linux
security .

On Thu, 2013-10-10 at 19:01 +0100, Simon McVittie wrote:
> On 10/10/13 18:12, p10 wrote:
> > That's my current "security setup" - a user account that I use for
> > everything , and 'su' into root with a password I don't keep stored
> > anywhere
>
> If you type your root password into a gnome-terminal running with user
> privileges, a shell running with user privileges, etc., then your user
> account is root-equivalent to a determined attacker (for instance, other
> user processes could ptrace the gnome-terminal or shell, or put a
> keylogging 'su' wrapper in the $PATH).
>
> If you want real privilege separation, you'll need to log in as root (or
> as a separate administrative account) via something more privileged than
> your user account (e.g. gdm, or getty(8)/login(1) on a text-mode virtual
> console).
>
> > if a root service unlocks
> > the key-ring for all the user-space programs - there's no point in
> > having the system in the first place . So that is a problem that if I'm
> > not mistaken stands with the current setup too - if you unlock the
> > keyring every user-space app can access the stored passwords .
>
> gnome-keyring does not protect you from your own user session. Security
> is meaningless without a security model, and gnome-keyring's security
> model is <https://wiki.gnome.org/GnomeKeyring/SecurityPhilosophy>.
>
> When applications within a user session can be protected from each
> other, it will make sense to develop a new security model. I don't think
> we're there yet.
>
>     S
>
> _______________________________________________
> desktop-devel-list mailing list
> desktop-devel-list gnome org
> https://mail.gnome.org/mailman/listinfo/desktop-devel-list