1

The latest openSSL library has a FIPS module which needs to be built with a fixed base address to be called FIPS compliant. The rationale behind this is that having a fixed base address enables them to do an integrity check of the library.

I personally believe that this is a security risk as attackers know the exact base address of the library and can use that location to run attacks. Especially attacks like return to libc attack or return oriented programming. Am I missing something here??

3

Having a known address is not a big risk. The only security measure that it defeats is ASLR, and ASLR is merely a technique to make exploits more difficult, it rarely makes them impossible.

The best defense against exploits is not to have a vulnerability in the first place!

This, not the fixed address, is the reason why using a FIPS-certified version of OpenSSL is detrimental to security. As new bugs are inevitably discovered, users must patch them as fast as possible, before exploits come out. As soon as you've patched the bug, you aren't running a certified version any longer. If you needed certification in the first place, you'll need to wait for a certified version with the bug fix, which typically takes months if the vendor even bothers to make one.

Given that the FIPS certification of OpenSSL is not really a security certification (FIPS 140 level 1 is little more than a check that the functions are computed correctly, the security aspects are virtually nil), the only reason to use the FIPS version rather than the latest stable release is if you need to tick the “FIPS validated” checkbox on a procurement form.

| improve this answer | |
  • In short, "No, and compliance != security." – Iszi May 25 '16 at 19:06
  • The best defense against exploits is not to have a vulnerability in the first place! The best way to fix crime in the community is to not have criminals! Quick, someone call the DOJ! Seriously though, this is the antithesis of defense in depth. The best defense against exploits is to assume everything is vulnerable and respond by using exploit mitigations for defense. Additionally, I would argue that disabling ASLR is a big risk. While it can be defeated by an attacker with arbitrary code execution, it is extremely effective at stopping scriptless exploits dead in their tracks. – forest Dec 16 '17 at 3:59
  • @forest My point is that disabling ASLR is not nearly as big a risk as not doing security upgrades. – Gilles 'SO- stop being evil' Dec 16 '17 at 12:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.