Re: systemd-devel Digest, Vol 155, Issue 8

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Sa, 11.03.23 08:29, Lal, Arun (arun.lal@xxxxxxxxx) wrote:

> 1) Dbus uses .conf files in /etc/dbus-1/system.d/ or /usr/share/dbus-1/system.d/ to allow and deny access to dbus method calls.
>     And what is the point of allowing a user in these conf files if
>     eventually systemd will block the call?

so, I think you are mixing up things. the caps thing is a red herring
i guess.

We definitely support polkit in systemd. but it has nothing to do with
caps.

dbus policy is mostly a useless concept, it's too static and riid. it
nowadays has been largely supplanted by polkit: thus the low-level
dbus policy for most services is just configured to be open, and the
services use polkit to authenticate the calls and refuse them if
polkit says no or cannot be reached.

> 2) Why is "busctl call" to slandered interfaces such as org.freedesktop.DBus.Peer still work even if caller is non-root.
>
> 3) I see that busctl commands such as "tree", "introspect" etc., are
> still allowed for non-root user. So why is there a restriction
> "call"?

I don't understand that question.

method calls systemd#s services provide are usually protected by at
least three levels: dbus policy (which as mentioned we mostly
configure to be entirely open), polkit, and then selinux if that's
available. Only if all three say "yes" we'll allow a call to go
through.

In none of the three cases process capabilities come into the mix
though, as mentioned in the other mail: we cannot use them for
authenticating in userspace securely.

Lennart

--
Lennart Poettering, Berlin



[Index of Archives]     [LARTC]     [Bugtraq]     [Yosemite Forum]     [Photo]

  Powered by Linux