Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 23 Aug 2001 18:27:49 +0100
From:      Brian Somers <brian@Awfulhak.org>
To:        Matt Dillon <dillon@earth.backplane.com>
Cc:        "Andrey A. Chernov" <ache@nagual.pp.ru>, Brian Somers <brian@Awfulhak.org>, Jun Kuriyama <kuriyama@FreeBSD.ORG>, cvs-committers@FreeBSD.ORG, cvs-all@FreeBSD.ORG, brian@freebsd-services.com, brian@freebsd-services.com
Subject:   Re: cvs commit: src/etc/defaults rc.conf src/etc/mtree BSD.var.dist src/etc/namedb named.conf 
Message-ID:  <200108231727.f7NHRng82227@hak.lan.Awfulhak.org>
In-Reply-To: Message from Matt Dillon <dillon@earth.backplane.com>  of "Thu, 23 Aug 2001 09:45:34 PDT." <200108231645.f7NGjYe86993@earth.backplane.com> 

next in thread | previous in thread | raw e-mail | index | archive | help
> 
>     I like the idea of, finally, invoking named in a sandbox.  I don't
>     understand why the pidfile location has to change, though.  named
>     creates its pidfile as root before it setuid's itself.
> 
>     While it is true that named cannot rescan interfaces when operating
>     in this mode, this restriction has never been an impediment to anything
>     I've ever done with it.  Most dialup users don't run named, they simply
>     allow ppp to setup /etc/resolv.conf for them.  Those who do will be savvy
>     enough to add the appropriate override to /etc/rc.conf (or won't have to
>     if they don't bother to mergemaster the new default rc files).
> 
>     I know it isn't a perfect solution, but we *REALLY* need to secure 
>     named this time around.  It is years past the time we should have done
>     it.

The only problem I have with the change is that people who currently 
have ``named_enable=YES'' in their /etc/rc.conf will have broken 
configurations when they update.

If this change goes in, I won't argue as long as it's *NOT* MFC'd.

> 						-Matt
> 
> :> > >   Invoke named with privilege of bind:bind.
> :> > >   Change pidfile location to /var/run/named/pid.
> :> > 
> :> > Is it discussed or I miss something? We already have an option to run it
> :> > in bind sandbox, but as non-default option. Some functions not works in
> :> > bind sandbox, I don't remember exactly at this moment.
> :> 
> :> named won't be able to listen on interface addresses that are not 
> :> configured when named is invoked.  This can break name services on a 
> :> dialup server quite badly.
> :
> :Yes, exact this thing.
> :
> :> I think this change should be reverted.
> :
> :I too.
> :If named allows root compromise, better fix named.
> :
> :-- 
> :Andrey A. Chernov
> :http://ache.pp.ru/

-- 
Brian <brian@freebsd-services.com>                <brian@Awfulhak.org>
      http://www.freebsd-services.com/        <brian@[uk.]FreeBSD.org>
Don't _EVER_ lose your sense of humour !      <brian@[uk.]OpenBSD.org>



To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe cvs-all" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200108231727.f7NHRng82227>