Zum Inhalt springen

Details

The NSA was right: 20 ways your C++ can be hacked

C++ trusts the programmer and offers no safety net by default, which is why institutions like the NSA and CISA call it a "memory-unsafe" language.

In this talk, we will go through 20 weaknesses from MITRE's CWE list that are common in C++, see how realistic code can be exploited, and discuss how to avoid and detect them.

For each weakness, we will show:

- Realistic code with a subtle flaw and what an attacker could do with it
- Reasonable coding practices and modern C++ idioms that prevent it
- Which tools catch it (sanitizers, hardening, static analysis, fuzzing)
- Which tools stay silent, giving you a false sense of safety

Out-of-bounds writes, use after free, race conditions, type confusion and more are often only one innocent-looking line away.
Let's see how undefined behavior is a promise we keep breaking, and take away some practical ideas for writing C++ that is harder to exploit.

The Speaker:
Dimitris Platis calls himself a Software Engineer and a Maker. He works with embedded systems of all shapes and sizes running some kind of Linux, from Yocto to AOSP (Android).
In the evenings he likes to tinker on open source projects (github.com/platisd) and blog about them (platis.solutions/blog).
Dimitris works primarily with C++ and also offers C++ training courses for companies, with customers including large corporations from the automotive, semiconductor and defense sectors.

Das könnte dir auch gefallen