Related Publications

Andrew Moss, Elisabeth Oswald, Dan Page, and Michael Tunstall. Compiler Assisted Masking. In Emmanuel Prouff and Patrick Schaumont, editors, CHES, volume 7428 of LNCS, pages 58–75. Springer, 2012
G. Agosta, A. Barenghi, M. Maggi and G. Pelosi, "Compiler-based side channel vulnerability analysis and optimized countermeasures application" Design Automation Conference (DAC), 2013 50th ACM/EDAC/IEEE, Austin, TX, 2013, p. 1-6.

Obfuscation

PhD thesis of Stéphanie Riaud (in French): http://dumas.ccsd.cnrs.fr/dumas-00636793/document
Obfuscator-LLVM:
- Bogus Control Flow (https://github.com/obfuscator-llvm/obfuscator/wiki/Bogus-Control-Flow) Average overhead performance ➡ 28 %
- Control Flow Flattening (https://github.com/obfuscator-llvm/obfuscator/wiki/Control-Flow-Flattening) Average overhead performance ➡ 860 %
- Instruction Substitution (https://github.com/obfuscator-llvm/obfuscator/wiki/Instructions-Substitution) Average overhead performance ➡ 13.5 %
- A commercial version of this project with more features is available at https://strong.codes/

Model of attacker

In the case of cyber attacks

Attacker can run programs on the same machine
- attack on the branch predictor.
  (in French, has links) https://interstices.info/jcms/c_25753/une-faille-de-securite-dans-les-processeurs
- attack on the Dcache (need ref)
Attacker interacts with remote machine
- buffer overflow, canari, ASLR, CFI
- Canaries with Clang: -fstack-protector, -fstack-protector-strong, -fstack-protector-all
- ASLR with Clang: -fpie -pie
- Safe stack with Clang: -fsanitize=safe-stack. More information : http://dslab.epfl.ch/proj/cpi/ (research prototype)
- CFI with Clang: -fsanitize=cfi. More information : http://clang.llvm.org/docs/ControlFlowIntegrity.html
- Global Offset Table (GOT) Protection with Clang: -Wl,-z,now -Wl,-z,relro

Implementations of automated Masking