Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 01 Sep 1995 11:32:09 -0700
From:      Eric Allman <eric@cs.berkeley.edu>
To:        Karl Strickland <karl@bagpuss.demon.co.uk>
Cc:        "Rodney W. Grimes" <rgrimes@gndrsh.aac.dev.com>, peter@haywire.dialix.com, freebsd-security@freebsd.org
Subject:   Re: Eric Allman's syslog.c fixes 
Message-ID:  <199509011832.LAA00429@mastodon.CS.Berkeley.EDU>
In-Reply-To: Mail from Karl Strickland <karl@bagpuss.demon.co.uk>  dated Fri, 01 Sep 1995 07:42:59 BST <199509010643.HAA08351@bagpuss.demon.co.uk>

next in thread | previous in thread | raw e-mail | index | archive | help
My take is that the copied fmt string should not be a problem,
because applications shouldn't be allowing the user to specify
it (if they did, users could drop in gratuitous `%'s).  The
only reason this is needed is to get %m, which wouldn't have
been necessary in the first case if strerror() had existed, or
if I had had the sense to implement that sort of routine instead
of the special-purpose %m hack.

Of course, it's always a good idea to bounds check -- I'm just
saying that I doubt that this will be a serious problem in real
life.

eric


============= In Reply To: ==============================================
: From:  Karl Strickland <karl@bagpuss.demon.co.uk>
: Subject:  Re: Eric Allman's syslog.c fixes
: Date:  Fri, 1 Sep 1995 07:42:59 +0100 (BST)

: > 
: > > 
: > > Eric Allman is running a new syslog.c through the mill at the
: > > moment. It'll be the one published in the RSN CERT advisory I presume.
: > > 
: > > It's thought to be bomproof on 4.4BSD systems (it uses vsnprintf), and
: > > the only holdup is portability to other OS's.
: > > 
: > > I keep a pretty close eye on this area, as it's sendmail related.  Is
: > > it worth bringing in the currently 'endorsed' version, and updating it
: > > to the CERT version if there are any changes later?
: > 
: > Yes, that would give Eric additional test data and eyes looking at
: > the solution.
: 
: I think that the fmt string should also be bounds checked - there is still
: no bounds check on the copy from the user supplied fmt string into the
: internal buffer.  Having said that, Im not aware of anything that lets
: the user mess with the fmt string, but I think it makes sense to fix it
: at this point.
: 
: Other than that, it looks good to me -- I did a make world (yes i know
: its extreme :)) with it and its been going without problems for > 24 hours..
: 
: -- 
: ------------------------------------------+----------------------------------
-
: Mailed using ELM on FreeBSD               |                    Karl Stricklan
d
: PGP 2.3a Public Key Available.            | Internet: karl@bagpuss.demon.co.u
k
:                                           |




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199509011832.LAA00429>