###LLVM slicer This is a static slicer based on the Mark Weiser's algorithm in [1]. It is augmented to perform an inter-procedural analysis.

How To Build

• mkdir LLVMSlicer.obj
• cd !$
• `CC=gcc CXX=g++ ../LLVMSlicer/configure --with-llvmsrc=<> --with-llvmobj=<>
• CXXFLAGS+=-std=c++11 make ENABLE_OPTIMIZED=1

###How To Run

• To get a mapping of binary to llvm-ir. This pass numbers the llvm-irs in a specific function (-mapping-function) and for each of them specifies the info of the corresponding binary mneumonics. This analysis pass used the metadata add by mcsema to infer this mapping.
  • Eg. opt -load=LVMSlicer.so -srcline-mapping -mapping-function=main -mapping-output=mapping.txt test/example1.ll -o /tmp/example1.analysed.bc
  • Eg. opt -load=LVMSlicer.so -srcline-mapping -mapping-function=sub_8048420 -mapping-output=mapping.txt test/example2.ll -o /tmp/example2.analysed.bc
    • -mapping-function: The function we want to focus on.
    • -mapping-output: The Output file with mapping. If not specifed the output will be dumped in stdout
• To obtain backward slicing using a function, line number of llvm-ir and variable name as slicing criterion.
  • The line number (specified as -criterion-line) can be chosen using the above mapping, as it specifies the line numbers of the llvm-ir.
  • Eg. opt -load=LLVMSlicer.so -criterion-function=main -criterion-line=43 -criterion-variable=product -slice-inter test/example1.ll -o /tmp/example1.sliced.bc
  • Eg. opt -load=LLVMSlicer.so -criterion-function=sub_8048420 -criterion-line=256 -criterion-variable=ESP_val -slice-inter test/example2.ll -o /tmp/example2.sliced.bc

###Acknowledgement

• This worked is borrowed from Jiri Slaby jirislaby@gmail.com
  • https://github.com/jirislaby/LLVMSlicer