Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Mon, 17 Nov 2014 22:37:27 -0500
From: Robert Watson <>
Subject: Re: Fuzzing findings (and maybe CVE requests) -
 Image/GraphicsMagick, elfutils, GIMP, gdk-pixbuf, file, ndisasm, less

Solutions to the most difficult problems are often found by approaching it
from the opposite direction...

What about using fuzzing to find those tools withOUT vulnerabilities and
"certifying them" in some way as safe for all inputs?

If a tool has too many problems or no champion to fix them, then there lies
a need to be filled for a programmer to produce a better tool.

On passing audit, the new tool would become the preferred tool for new
distro releases, especially for systems running websites.

Robert "DocSalvager" Watson

*Trust in truth keeps hope alive*

*     iCare for AffordableCare

* <>*

* <>*
* <>*
* <>*

<> <>

On Mon, Nov 17, 2014 at 4:30 PM, Daniel Kahn Gillmor <>

> On 11/16/2014 07:15 AM, Robert Święcki wrote:
> > To sum up: If somebody uses 'file' in an unconstrained OS environment
> > on untrusted inputs, and he gets pwnd in the result, then it's not a
> > security problem, it's an incompetence problem - and IMO it should be
> > discussed elsewhere.
> I think other people have made good points already that tools like
> "file" and "strings" are routinely used on untrusted input, and so
> deserve to be treated as part of the attack surface in a normal free
> software operating system (and therefore vulnerabilities in them warrant
> mention here on oss-security).
> I'd like to present one other argument against the kind of distinction
> that Robert suggests making here, though.
> If "file" or "strings" (or libmagic or libbfd, respectively) are
> considered as "private" tools that should only be run on trusted inputs,
> then we are effectively creating an entirely new class of
> vulnerabilities that we need to report and fix, which may not have any
> resolution.  In particular, we would need to report vulnerabilities in
> tools that use these "private" tools on public data.
> For example:
>  * roundcube (a webmail client) relies on libmagic1
>  * rox-filer (a graphical filesystem browser) relies on /usr/bin/file
>  * rkhunter (a tool for scanning potentially-malicious files for
> rootkits) relies on /usr/bin/file.
> The composable nature of unix-style tools means that a bug in one
> component is very likely to be a security vulnerability.
> And if our interest is in not overwhelming the list with vulnerability
> assignments, then declaring certain tools "off-limits" or "only to be
> run on trusted inputs" is actually likely to *increase* instead of
> decrease the total number of vulnerability counts (since we would now
> need to report vulnerabilities in packages like roundcube and rkhunter
> and rox-filer for exposing file and libmagic to untrusted input), while
> still not leaving our users any safer.
>         --dkg

Powered by blists - more mailing lists

Please check out the Open Source Software Security Wiki, which is counterpart to this mailing list.

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.