Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 28 Jan 2002 08:51:05 +0100 (CET)
From:      "Patrick M. Hausen" <hausen@punkt.de>
To:        "M. Warner Losh" <imp@village.org>
Cc:        charon@seektruth.org, dsyphers@uchicago.edu, security-officer@FreeBSD.ORG, stable@FreeBSD.ORG
Subject:   Re: Firewall config non-intuitiveness
Message-ID:  <200201280751.g0S7p5414157@hugo10.ka.punkt.de>
In-Reply-To: <20020127.120138.07163985.imp@village.org>

next in thread | previous in thread | raw e-mail | index | archive | help
Hi all!

In a long and tiring thread Warner Losh once wrote:

> : > The current behavior fails safe.  The current behavior is documented.
> : > I relied on that documentation when setting up my firewall.  Now you
> : > are wanting to change that documented behavior.  It is that way
> : > specifically so we fail safe.
> : 
> : The current behavior also renders systems unusable.  What good is having my 
> : web/mail server safe doing me if it can't process any mail or http requests?  
> : The default rc.conf says next to firewall_enable "Set to YES to enable 
> : firewall functionality," which implies that NO disables firewall 
> : functionality.  Which is read "disables firewall", not "disables custom 
> : firewall scripts."  I view the kernel as containing stuff that's 
> : _potentially_ used - I can have support in it for an ethernet card
> : that's not installed.  But the system doesn't hang looking for it.
> 
> Rendering the system unusable is fail safe.
> 
> : Anyway, the default rc.conf could have firewall_enable set to YES, which 
> : would make it "fail safe."
> 
> No.  That's not fail safe.  My machine will still break in an
> unacceptible way by this change.
> 
> Please write up the exact details that you want to do so that those on
> security-officer know exactly what you are proposing.  It is my
> understanding that you want to make enable_firewall=NO totally dyke
> out the firewall that was compiled into the kernel and be a totally
> open realy.  I know that this breaks at least one machine that I have,
> but I also know that this breaks our current fail-safe behavior, which
> I'm strongly opposed to.
> 
> However, I think I've become too embroiled in this issue, which is why
> I want the fine folks at security-officer@ to evaluate it (since I'm
> on that list, I'll refrain from doing more than stating my position).
> It just doesn't seem right to me.

Partly I agree with you, partly I think the original author is correct
in the way that the current behaviour is confusing at least to
novices. apm_enable="NO" -> disable apm. sendmail_enable="NO" -> disable
sendmail. firewall_enable="NO" -> you get the idea.
I'd suggest changing the name of the parameter to firewall_rules_enable
or firewall_script_enable, or ...

Wouldn't we get rid of this entire argument, if IPFIREWALL_DEFAULT_TO_ACCEPT
was the default for the kernel part of ipfw and there was an option
IPFIREWALL_DEFAULT_TO_DENY for anyone preferring the "old" behaviour?

Since it is recommended to explicitly state every rule in your config
file - with Cisco access lists as well as with ipfw - I assume
almost everyone has a "add deny all from any to any" line at the end of
his ruleset already? I'm currently in a project where I build a bridging
firewall. Seems like I need DEFAULT_TO_ACCEPT to get arp through.
I wouldn't mind if it became the default.

Regards,

Patrick M. Hausen
Technical Director
-- 
punkt.de GmbH         Internet - Dienstleistungen - Beratung
Scheffelstr. 17 a     Tel. 0721 9109 -0 Fax: -100
76135 Karlsruhe       http://punkt.de

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




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200201280751.g0S7p5414157>