linux-kernel - Re: [RFC PATCH 1/3] setgroups: new mode 'shadow' for /proc/PID/setgroups

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20210517160859.GA25644@mail.hallyn.com>
Date:   Mon, 17 May 2021 11:08:59 -0500
From:   "Serge E. Hallyn" <serge@...lyn.com>
To:     Giuseppe Scrivano <gscrivan@...hat.com>
Cc:     "Serge E. Hallyn" <serge@...lyn.com>, linux-kernel@...r.kernel.org,
        dwalsh@...hat.com, christian.brauner@...ntu.com,
        ebiederm@...ssion.com
Subject: Re: [RFC PATCH 1/3] setgroups: new mode 'shadow' for
 /proc/PID/setgroups

On Mon, May 17, 2021 at 03:30:16PM +0200, Giuseppe Scrivano wrote:
> Hi Serge,
> 
> thanks for the review.
> 
> "Serge E. Hallyn" <serge@...lyn.com> writes:
> >> diff --git a/kernel/user_namespace.c b/kernel/user_namespace.c
> >> index 8d62863721b0..b1940b63f7ac 100644
> >> --- a/kernel/user_namespace.c
> >> +++ b/kernel/user_namespace.c
> >> @@ -123,6 +123,7 @@ int create_user_ns(struct cred *new)
> >>  		ns->ucount_max[i] = INT_MAX;
> >>  	}
> >>  	ns->ucounts = ucounts;
> >> +	ns->shadow_group_info = get_current_groups();
> >
> > If userns u1 unshares u2 with shadow set, then when u2 unshares
> > u3, should u3 get the same shadowed set that u2 has, or should it
> > get all of u2's groups as u3's initial shadow set?
> 
> good question.  Thinking more of it, I think a reasonable interface is
> to expect a child userns to inherit the same shadow groups as its parent
> userns.  If "shadow" is written again to the /proc/PID/setgroups file
> then it grows shadow groups set to include the ones the userns had at
> creation time (which includes the parent shadow groups).  What do you
> think of it?  I'll play more with this idea and see if it works.

That's what I was thinking would make the most sense.