[langsec-discuss] [oss-security] Thoughts on Shellshock and beyond
s.kieske at mittwald.de
Tue Oct 7 12:50:08 UTC 2014
On 07/10/14 13:59, David A. Wheeler wrote:
> I am still struggling with this one. I am trying to create that list here:
> But to be honest, that list is pretty pathetic. This is a challenging class of vulnerability to detect or prevent ahead of time. Ideas would be very welcome.
Well I think the core issue was:
A parser, which was written to just run trusted input
was exposed years later via third party configuration
( e.g. /bin/sh == bash & cgi using /bin/sh) to untrusted
I guess to avoid such cases in the future, you must be explicit
from which sources you accept input.
This whitelist should not contain arbitrary content like "stdin"
or any input source which itself has no strict whitelist on input
So in short: you need to design and implement interfaces
for every program which enforce explicit security boundaries.
Example: why should I be able to feed audio files
into graphics processing programs?
This is a lot about parser and language safety imho.
So I can't mention langsec.org enough.
Go there, read the papers, look up the videos from
I know this means re-implementing almost all code out there.
But if you can proof that the input for program A
constructs a turing complete grammar you have already lost.
You can not "fix" these bugs, as there is an infinite number of them.
You need to fix the parsers and the languages first.
Mit freundlichen Grüßen / Regards
Mittwald CM Service GmbH & Co. KG
Königsberger Straße 6
Geschäftsführer: Robert Meyer
St.Nr.: 331/5721/1033, USt-IdNr.: DE814773217, HRA 6640, AG Bad Oeynhausen
Komplementärin: Robert Meyer Verwaltungs GmbH, HRB 13260, AG Bad Oeynhausen
More information about the langsec-discuss