Date: Mon, 14 Apr 2014 10:47:23 +0100 From: Chris Rees <crees@bayofrum.net> To: Stefan Esser <se@freebsd.org>, Benjamin Podszun <dar@darklajid.de>, freebsd-ports@freebsd.org Cc: crees@freebsd.org Subject: Re: FreeBSD Port: security/sshguard-pf Message-ID: <022c964d-3478-4ad2-ac9d-a1749960c92e@email.android.com> In-Reply-To: <534BAC12.9090704@freebsd.org> References: <53453547.2070307@uni-bielefeld.de> <fe85c5ec-65aa-425d-9e58-f70f277ad491@darklajid.de> <534BAC12.9090704@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On 14 April 2014 10:36:18 BST, Stefan Esser <se@freebsd.org> wrote: >Am 14.04.2014 10:25, schrieb Benjamin Podszun: >> Looking at the rc script and the diff [1] the problem's easy enough: >> ${sshguard_pidfile} is passed as parameter to -i, but isn't set in >the >> script/has no default value. Either the related line from the >previous >> revision should be revived or the substitution should change to use >> ${pidfile}, which _is_ set. > >I just installed sshguard on one of my servers and noticed the same >problem. The program is not started due to several bugs: > >1) $sshguard_pidfile vs. $pidfile as noticed by you This one's my fault, sorry. >2) Pasing of log files to watch. They are correctly processed by > sshguard_prestart(), but the result is not pasted into the > command line. (You can manually add "-l <logfile>" options to > the command line in the rc script as a work around ...) Don't think this one is, but I'll investigate. Chris >There are other deficiencies: > >a) The documentation lacks details about the mechanism used to block > attacks. E.g. in case of IPFW, blocking rules are injected in lines > 55000 to 55050. You have to adapt your ruleset in such a way, that > any to-be-blocked service is only enabled at a later line, or the > blocking is ineffective. This port range should be mentioned at > least in the pkg message for ipfw. Better would be a section in > the man page, which explains the mechanism used by each backend. > >b) The security/sshguard-ipfw port is marked as NO_STAGE=no, while > security/sshguard seems to work just fine with staging enabled. > This is probably an oversight: when sshguard was fixed/verified > for staging, the sub-ports where not marked as staging clean. > >c) The MAKE_ARGS variable mention ACLOCAL, AUTOCONF and AUTOMAKE, but > no dependencies are registered for any of them. > >d) The master port's Makefile lists hosts, pf, and ipfw as possible > backends, selected by SSHGUARDFW, but does not mention ipfilter > as the fourth supported backend. > >I did not have time to check the code quality of the parser. I'm a >bit suspicious, that it might be possible to attack sshguard via >parameters passed under control of an attacker. > >If you create a PR, you may want to add these points to the PR ... > >Regards, STefan -- Sent from my Android phone with K-9 Mail. Please excuse my brevity. -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?022c964d-3478-4ad2-ac9d-a1749960c92e>