|
|
Decompilation of Selected C++ Constructions
Mihulka, Tomáš ; Zemek, Petr (referee) ; Matula, Peter (advisor)
This bachelor's thesis deals with the reconstruction of a hierarchy of classes and their virtual methods from programmes created by C++ language . The aim of this work is to extend a decompiler , which has been developed as a part of the Lissom project, by an analysis of those reconstructions for various decompilers . The reconstructions are created through detection of RTTI and virtual tables . The first part of this thesis involves a description of reverse engineering as well as of the Lissom project in terms of the decompiler . The following section of the paper explains the basics of C++ language , its structures and different possibilities of their decompilation . The final part of the paper deals with a design, implementation and testing of a recognition of RTTI and virtual tables .
|
|
|
Semi-Automatic Optimization Using Specialized Instructions
Mikó, Albert ; Hynek, Jiří (referee) ; Masařík, Karel (advisor)
The design of instruction sets for application specific processors is a difficult task. This thesis describes the tasks of selection, marking and creation of instruction set extensions for application specific processors. The presented semiautomatic method provides the user with a simple way to select instruction set extensions by marking a section of source code in the application. The creation of the new instruction in the modelling language itself is solved automatically. Thanks to this the user can concentrate his efforts on tasks where human ingenuity and experience can be used the most.
|
|
|
C Back-End for a Decompiler
Urban, Martin ; Ďurfina, Lukáš (referee) ; Zemek, Petr (advisor)
This thesis deals with the implementation of the back-end of the decompiler, which produces a code in C language. It contains basic information about the principals and using of the reverse engineering either in the area of information technology or apart from it. The main goal is to create the back-end of the decompiler which would generate a code that would be equivalent against the input and will be translatable into a binary code. Functionality of the output code will be conserved in state of the functionality of the source code. The output is the implementation of the classes in C++ language. It does described activity as a part of the general decompiler which is developed in terms of the project Lissom.
|
|
|
Decompiler Front-End Optimizations
Odaloš, Matej ; Ďurfina, Lukáš (referee) ; Křoustek, Jakub (advisor)
Decompiler is a reverse engineering tool for translation of binary codes into one of the higher level languages. This bachelor thesis describes such a tool paying special attention on decompiler of the Lissom project. There are proposed several techniques for translation optimalization like static LLVM IR code interpretation and memory for its results. Other optimalizations are conserning addition of platform dependent features like delay slot support or memory datalayout and endianness detection. Finally implemented techniques are demostrated on several examples.
|
|
|
Software Execution Acceleration Using Automatic Instruction Set Extensions
Melo, Stanislav ; Dolíhal, Luděk (referee) ; Masařík, Karel (advisor)
One of the important feature of application specific processors is performance. To maximize it, the processor must adapt to needs of application that it is going to perform. One of the ways to do that is to search for appropriate instructions that can be joined into one special instruction. This instruction is then implemented in hardware as a single function block. This bachelor's thesis describes problem of finding and selecting suitable candidates for instruction-set extension. It also provides brief overview of the few best known algorithms that solve this problem. Moreover the thesis deals with the single-cut algorithm and its implementation in more detail.
|
|
|
Automatic Searching of Instruction Extensions for Application Processors
Češka, Martin ; Křoustek, Jakub (referee) ; Masařík, Karel (advisor)
This thesis deals with the process of automatic searching of instruction-set extensions for application-specific instruction-set processors. This process uses slightly edited ISEGEN algorithm. At first, all important terms including this algorithm are described. Then there is a detailed description of implementation of whole process in C++ programming language. At last, newly created program is considered as useful or useless based on speed-up of processor at performing of input program using found extensions.
|
|
|
Reconstruction of Data Types for Decompilation
Matula, Peter ; Ďurfina, Lukáš (referee) ; Křoustek, Jakub (advisor)
This document describes methods for a reconstruction of data types in the decompilation problem. It defines the concept of reverse engineering and introduces decompiler developed by the Lissom project. It presents existing methods of reconstruction of the simple and complex data types, and explains in detail approaches based on data-flow analysis and analysis of the memory operation offsets. The core of this thesis is the design of a new technique of reconstructing simple and complex data types, suitable for deployment in a retargetable decompiler environment of the Lissom project. Basic principles of the new technique, its implementation and related changes in decompiler and intermediate language are described. The solution is tested and the conclusion discusses the achievements, shortcomings and direction of the further work.
|
|
|
Generic Decompilation of Bytecode into High-Level Representation
Mrázek, Petr ; Zemek, Petr (referee) ; Křoustek, Jakub (advisor)
The work describes methods and principles of decompilation, basic information about reverse engineering and its use in both software engineering and engineering in general. Furthermore, it introduces the decompiler developed within the Lissom project at BUT FIT. The goal of the work is to design and implement a retargetable decompiler for bytecode, which extends the original decompiler.
|
|
|
Source-Code Migration Using Decompilation
Korec, Tomáš ; Ďurfina, Lukáš (referee) ; Zemek, Petr (advisor)
This thesis deals with source-code migration of high-level programming languages using decompilation. A migration tool developed within the thesis is built on top of the middle-end and back-end parts of Lissom project decompiler. Several compilers generating LLVM IR code from input languages are discussed. Compilers suitable for integration to the migration tool were chosen. Compiled LLVM IR code is an input of the decompiler's optimizing middle-end. The output from the migration tool is a code in the C language or Python-like language generated by the back-end of the decompiler. The input languages are Fortran and its dialects, C/C++/Objective-C/Objective-C++, and D. The thesis describes problems connected with migration of these languages, their solutions, and ways to improve quality and readability of the produced source code.
|
|
|
Code Structuring in Decompiler Back-End
Hrbek, David ; Matula, Peter (referee) ; Zemek, Petr (advisor)
The goal of this thesis was to design and implement an algorithm for code structuring in Lissom decompiler's back-end. This algorithm eliminates indirect jumps (branch/goto) from low-level code with a use of high-level constructs, such as conditional statements (if, switch) and loops (for, while). This thesis contains an introduction into the topic of decompilation, some information about the Lissom project's decompiler, a proposal of the structuring algorithm, details of its implementation, testsuite description and results summary.
|
guest ::
