From: Benjamin Walkenhorst <krylon@gmx.net>
To: linux-admin@vger.kernel.org
Subject: Securing my machine
Date: Tue, 25 Nov 2003 18:43:38 +0100 [thread overview]
Message-ID: <200311251843.38934.krylon@gmx.net> (raw)
Hello everybody,
I use Slackware 9.1 on my desktop-machine. I do so quite happily, it
took only a week for Slackware to become my new primary OS. =)
I connect to the internet through a small server/gateway running
NetBSD-1.6.1. The gateway connects to my ISP via ISDN-dial-up
connection.
The NetBSD-machine runs ipf (NetBSD's packet filter, roughly equivalent
to iptables) and IPNAT.
I run gtk-gnutella on my desktop-machine from time to time. Since I want
others to be able to connect to my machine (also, for getting
push-connections), I decided to forward the corresponding port to my
Linux-machine.
This has even shown to work fine, thanks. =)
But I am getting a little concerned about letting others connect to my
machine. Since my desktop-machine is behind a firewall, also since I am
the only user on my home-network, I did not exactly take care to secure
my Linux-machine.
Now I am getting worried someone might break into my machine via
GNUtella. I don't think gtk-gnutella was written with security in
mind...
So I want to tighten the security on my Linux machine in a way that
includes minimal inconvenience. Of course, I am going to start with all
the usual stuff, like installing tripwire, shutting down unneeded
services (in fact, I do this by default after installation), taking
care of file-permissions, cleaning up unneeded suid/sgid-bits, and so
on.
But then I read, most of all network-attacks are done via
buffer-overflows, so this is what I am most concerned about. I hear,
there's basically two ways of handling this problem:
- Using MAC/RBAC for controlling the ressources an application can
access - if it's getting compromised, it won't be able to harm the
system (seLinux).
- Preventing buffer-overflows in the first place. There's several
options how to achieve this, the most important are a) applying
patches to the kernel (PaX, grSecurity) or to the GNU C Compiler
(ProPolice)
In general, preventing buffer-overflows at all seems preferrable to me,
since it does not seem to require that much work. Also, this is they
way the OpenBSD-project has been going, and OpenBSD surely has a
reputation for first-class security.
So I got several questions:
- Has anyone worked with these system-add-ons? Got any experiences to
share with me?
- ProPolice sounds nice. But using it would require lots of
recompiling... What exactly do I have to recompile in order to benefit
from it? Just the application in question? The libraries, too? The
kernel? The entire system?
I am going to evaluate CRUX and Gentoo on my desktop-machine, both of
which offer the option of recompiling the entire system. If I choose
to use one of these as my primary system, recompiling won't be a
problem, any more. As of now, it is, if system libraries or even the
base system are involved.
- PaX/grSecurity sound really sweet. But I see on the homepages, there
are patches available only for linux-kernels 2.4.22. Is 2.6 going to
be supported in the near future?
I am using 2.6.0-test8 right now, and I am rather happy with it, so I
would like to keep using 2.6, once the final version is out.
On the other hand, I can switch back to 2.4.22 if PaX/grSecurity
offers serious protection.
And a lot of grSec's features sound really neat. =) Right now, this
sounds like the best way to secure my machine, since it invloves only
minimal setup, just patching and recompiling the kernel, while
increasing system-security drastically. If I got things right, that
is...
- MAC/RBAC does not really sound like I need it. Then again, more
security never hurts.
But this also sounds like it is going to be a lot of learning plus a
lot of effort to get it working. Furthermore, the corresponding
kernel-patch is developed at the NSA, and I do not exactly trust the
NSA to contribute to my privacy.
In order for M/RB-AC to be really useful, I'm afraid, you have to take
a lot of time to set it up correctly. And, as I said, I do not know
terribly much about this topic.
If I get things right, seLinux and grSecurity are not mutually
exclusive.
So, in general, any information will be appreciated. If there are
further promising ways of protecting my system against
buffer-overflows, I would like to know, as well.
Of course, I like to read a lot, so any hints on where to look for
information will be appreciated as well (if there's something
useful/interesting to read, there).
I am aware of pageexec.virtualave.net (PaX's homepage) and
grsecurity.net, as well as the NSA's seLinux-page.
Anything I missed? Anything I should know?
Thank you very much in advance,
Kind regards,
Benjamin Walkenhorst
--
Benjamin Walkenhorst
eMail: krylon@gmx.net
http://www.krylon.de
next reply other threads:[~2003-11-25 17:43 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2003-11-25 17:43 Benjamin Walkenhorst [this message]
2003-12-01 10:54 ` Securing my machine Alexander Economou
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=200311251843.38934.krylon@gmx.net \
--to=krylon@gmx.net \
--cc=linux-admin@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).