This repo. contains the evaluations I used for the project, Opaque Predicates Detective.
This is what I used for baseline evaluation. It is compiled at optimization levels O0, O1, O2, O3 – hence those folder names.
The trivial folder contains datasets pertaining to trivial opaque predicates. The content of weak and strong can be respectively inferred.
A trivial opaque predicate is constructed within a basic block.
A weak opaque predicate is constructed within a function.
A strong opaque predicate is constructed throughout multiple functions.
Christian Collberg is the first to classify opaque predicates into trivial/weak/strong in his paper "A taxonomy of obfuscating transformations."
Root directory of trivial/weak/strong folder contains the original source code files from github.com/tum-i22/obfuscation-benchmarks. Note that source code files inside the tigress-generated-programs folder in Banescu's benchmarks are not included since they are randomly generated, thus unrealistic.
tfinal folder contains the corresponding obfuscated binaries.
tout folder contains the corresponding obfuscated source code files.
tstats folder contains the virtual addresses of where the superfluous basic blocks are located in the obfuscated binaries.
tstats_topd folder contains the virtual addresses of where OpaquePredicatesDetective identifies the start of the superfluous basic blocks in the obfuscated binaries.