Code Monkey home page Code Monkey logo

nartoan / klee-examples Goto Github PK

View Code? Open in Web Editor NEW

This project forked from tracer-x/tracerx-examples

0.0 2.0 0.0 60.38 MB

Examples for Tracer-X KLEE

Home Page: http://www.comp.nus.edu.sg/~tracerx/

License: University of Illinois/NCSA Open Source License

Makefile 8.42% C 61.32% M4 2.29% Shell 6.76% Perl 1.14% C++ 2.46% TeX 0.77% Yacc 0.11% Logos 0.09% Roff 0.96% RPC 0.01% Objective-C 0.09% Perl 6 0.08% XS 1.31% Terra 0.06% HTML 10.89% Tcl 0.06% CSS 0.09% PHP 3.11%

klee-examples's Introduction

Simple examples for running with Tracer-X KLEE

Copyright 2015-2017 National University of Singapore. See LICENSE.md for license information.

This software includes third-party software, including some sample programs are from LLBMC 2013.1 distribution modified for running with KLEE, GNU Coreutils 6.10, ImageMagick 7.0.5-9, and portions of libPNG. Their license information is included in the license/LLBMC_LICENSE, license/COPYING, license/IMAGEMAGICK_LICENSE and license/LIBPNG_LICENSE.

LLBMC 2013.1 is Copyright (c) 2010-2013 Stephan Falke, Florian Merz, Carsten Sinz Karlsruhe Institute of Technology (KIT), Germany. All rights reserved. Coreutils 6.10 is Copyright (C) 1984-2008 Free Software Foundation. Refer to license/IMAGEMAGICK_NOTICE for the copyright notice of ImageMagick 7.0.5-9 and license/LIBPNG_LICENSE for the copyright notice of libPNG.

Each directory except include and utils contains a suite of simple examples.

  • basic - Basic examples.

  • binary-chop - Examples which test the binary chop algorithm for WCET analysis, including nsichneu.c and statemate.c which are also used to test scalability.

  • scalability - Examples that tests scalability, including Regexp.c, an example from KLEE tutorial at http://klee.github.io/tutorials.

  • include - Include files for running with various tools.

  • coreutils - Examples of 89 stand-alone programs in the GNU Coreutils-6.10 utility suite. In this version, KLEE was able to find some bugs and they are reported in the OSDI Paper. The coreutils directory contains coreutils-6.10 subdirectory of GNU Coreutils 6.10.

  • imagemagick - ImageMagick 7.0.5-9

  • utils - Scripts and other utilities. This contains, among others, cav17-coreutils.sh and cav17-scalability.sh for reproducing the experimental results of our CAV '17 paper submission.

The following directories are work in progress.

  • abstract - Examples for testing klee_abstract API.

  • join - Examples for testing Tracer-X klee_join API. This directory also contains academic programs focusing on program equivalence analysis, with an example taken from [KLEE OSDI paper 2008] (https://www.doc.ic.ac.uk/~cristic/papers/klee-osdi-08.pdf).

  • memleak - C programs with memory leak.

  • taint - Assortment of examples extracted from known vulnerabilities: some may require KLEE API introduced in a version of KLEE with taint propagation.

Configuring

To run any of the examples, first you need to execute the provided configure script. In particular, please set the following configure options:

  • The following options are relevant for all categories of examples:
  --with-klee             KLEE installation directory (defaults to
                          /usr/local/lib/tracerx)
  --with-z3               The directory of Z3 (defaults to /usr/local/lib/z3)
  --with-stp              The directory of STP (defaults to /usr/local)
  • The following option is relevant only for running with join examples
  --with-clpr             The directory of CLP(R) (defaults to /usr/local)
  • The following options are relevant only for running with coreutils examples
  --with-wllvm            The source directory of whole-program-llvm (defaults
                          to /usr/local/lib/whole-program-llvm)
  --with-texinfo          GNU texinfo 4.13 installation directory, i.e., the
                          info utility should be under the bin subdirectory of
                          the specified directory. Coreutils 6.10 requires GNU
                          texinfo no later than 4.13.
  • The following option is relevant for both coreutils and scalability examples
  --with-llbmc            LLBMC installation directory, i.e., the llbmc
                          executable should be under the bin subdirectory of
                          the specified directory (defaults to /usr/local).

Running with example programs other than join, coreutils, imagemagick and scalability

Instruction on running with the examples in the join, coreutils, and scalability directories can be found later in this document. Please first run the configure script as instructed above. Then to run the the example(s) in a particular directory, say basic, change your current directory to the basic directory.

The Makefile in each directory will create KLEE output directories <example-name>.tx which also contains the .dot files, and also <example-name>.inputs files that show the input values for all of the generated tests.

Sample usages:

  • To run a single example, e.g., addition_safe1.c in the basic directory with KLEE using Z3 solver and interpolation:

    make addition_safe1.tx

    Optionally, add ENABLE_COVERAGE=ON to the make command to enable coverage computation. If the file subsumption.dat exists, this would also perform a regression test on the number of subsumptions compared to reference data in subsumption.dat.

  • To run all examples with KLEE using Z3 solver and interpolation:

    make

    This is the same as executing make <example_name>.tx for all examples. Optionally, add ENABLE_COVERAGE=ON to the make command to enable coverage computation.

  • To run a single example, e.g., addition_safe1.c in the basic directory using Z3 solver but without interpolation:

    make addition_safe1.klee

    Optionally, add ENABLE_COVERAGE=ON to the make command to enable coverage computation.

  • To run a single example, e.g., addition_safe1.c in the basic directory using STP solver and without interpolation:

    make addition_safe1.stpklee

    Optionally, add ENABLE_COVERAGE=ON to the make command to enable coverage computation.

  • You can also execute LLBMC to analyze addition_safe1.c after properly configuring this repository using --with-llbmc. The execution can be done as in the following example:

    make addition_safe1.llbmc

  • To build an LLVM IR file of addition_safe1.c:

    make addition_safe1.ll

  • To clean the directory:

    make clean

Running with join examples

The examples in the join subdirectory requires Tracer-X KLEE to be compiled with CLP(R) 1.2l support. The Makefile tests can be run in the following way:

  • For running Tracer-X KLEE with join1.c to test the ability of -use-clpr option to load multiple CLP(R) files.

    make test-use-clpr

  • For running Tracer-X KLEE with join2.c to actually test the subsumption ability of Tracer-X KLEE with klee_join.

    make test-join

  • There are other programs such as count.c and sum.c in the join directory. Please consult join/Makefile, which contains test-* targets to run them.

Running with coreutils, imagemagick or scalability examples

  1. Run the configure script as instructed above.
  2. For coreutils or imagemagick, build it as follows:
    • Prerequisites:
      1. whole-program-llvm. This can be obtained by cloning it from GitHub, i.e.: git clone https://github.com/travitch/whole-program-llvm.git or installing it using PIP: sudo pip install wllvm.
      2. You may/may not need an older version of GNU texinfo. The build does not work with texinfo 5.2, but known to work with 4.13. Texinfo 4.13 can be found here.
    • Steps:
      1. cd coreutils for coreutils or cd imagemagick for imagemagick
      2. Run make build. This builds Coreutils 6.10 of ImageMagick 7.0.5-9 twice: one with for producing whole-program bitcode in coreutils/coreutils-6.10/obj-llvm/src (the only bitcode is imagemagick/ImageMagick-7.0.5-9/obj-llvm/utilities/.libs/magick-whole.bc for ImageMagick), and another with instrumentation for measuring coverage with llvm-cov in coreutils/coreutils-6.10/obj-cov/src (imagemagick/ImageMagick-7.0.5-9/obj-cov/utilities/.libs for ImageMagick).
  3. Set open file limits to 65536 in the following way: ulimit -n 65536. The number 65536 seems to work. Please note that this may not work depending on whether the number exceeds your hard limit. You can check the hard limit via ulimit -H -n.
  4. cd coreutils, cd imagemagick, or cd scalability
  5. Run make in any of the following ways:
    • Just make - Runs the test-case generation on all programs: in case of coreutils or imagemagick, this will build Coreutils 6.10 or respectively ImageMagick 7.0.5-9 if this is not done already.
    • make <program_name>.tx - Runs the test-case generation on <program_name>: it will build Coreutils 6.10 or ImageMagick 7.0.5-9 if this is not done already. Here, <program_name> is one of the programs whose executable file is found as coreutils/coreutils-6.10/obj-cov/src/<program_name> for Coreutils, or imagemagick/ImageMagick-7.0.5-9/obj-cov/utilities/.libs for ImageMagick. The output will be the result of Tracer-X KLEE run on the program with Z3 solver backend and interpolation, and its coverage information .The test cases are saved in <program_name>.tx directory.
    • make <program_name>.klee - Runs the test-case generation on <program_name> similar to make <program_name>.tx, however, Tracer-X KLEE will be invoked with -solver-backend=z3 and -no-interpolation to disable interpolation. This makes the run equivalent to running KLEE with -solver-backend=z3. The test cases are saved in <program_name>.klee directory.
    • make <program_name>.stpklee - Runs the test-case generation on <program_name> similar to make <program_name>.klee, however, Tracer-X KLEE will be invoked with -solver-backend=stp which disables interpolation and uses STP, resulting in standard KLEE run using STP. The test cases are saved in <program_name>.stpklee directory.
    • make <program_name>.llbmc - Runs LLBMC on <program_name>.
    • make experiment.csv or make llbmc-experiment.csv.
      • These are for executing with the example programs using KLEE, Tracer-X KLEE, and LLBMC, and collecting data for presentation. The targets will produce experiment.csv and llbmc-experiment.csv comma-separated values files, respectively, depending on which one was invoked. experiment.csv target will produce both experiment.csv and small-experiment.csv, which is the version with less data columns.
      • For coreutils benchmark only, llbmc-experiment.csv target will also produce experiment.csv (and small-experiment.csv) for the data for Tracer-X and KLEE (Z3 and STP) runs. This is because LLBMC can only be run on a smaller subset of Coreutils 6.10 programs, and on bitcode generated from the source of the main function only without linking with external files and libraries (libc). As such, for a more fair comparison, Tracer-X and KLEE must also be executed on the same bitcode. This is not the case for ImageMagick.
      • The experiments are run under different kinds of settings which can be seen in %.klee1, %.klee2, %.tx1, %.tx2, %.tx3, %.tx4, %.tx5, and %.tx6 targets defined in the coreutils/Makefile, imagemagick/Makefile, and scalability/Makefile, on the selected coreutils programs specified using the EXPERIMENT_SET variable in coreutils/Makefile, imagemagick/Makefile, and scalability/Makefile.

Note that by default, line coverage computation is performed using llvm-cov. To prevent, this, add ENABLE_COVERAGE=OFF to the make command, for example: make ENABLE_COVERAGE=OFF Regexp.tx.

klee-examples's People

Contributors

domainexpert avatar feliciahalim avatar nguyenquocbinh avatar rasoolmaghareh avatar

Watchers

 avatar  avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.