Artifact

README.md

@@ -1 +1,59 @@
-# ValueCheck
+# ValueCheck
+
+This repo is for code release of our paper `Effective Bug Detection with Unused Definitions` in `EuroSys 2024`.   
+
+In the paper we propose to use cross-author unused definitions to detect bugs and prioritize bugs by its familarity.  
+
+Its workflow contains three steps:
+* With the bitcode as input, ValueCheck applies static analysis to identify unused definitions and prune false positives.
+* From the found snippets, ValueCheck extracts authorship information and capture cross-authorship relationship.
+* After calculating the code familiarity, ValueCheck prioritizes the detected unused definitions.
+
+This repo contains the source code, scrips, and other artifacts. These are required to reproduce the results we presented in the paper.
+It helps the reproduction of evaluation results in the section 8 of the paper. 
+
+The artifact is available on GitHub at https://github.com/floridsleeves/ValueCheck.
+
+## Software dependencies
+- Linux (tested on Ubuntu 20.04)
+- Python >=3.8
+- SVF >= 2.7
+- LLVM >= 12.0
+
+## Data sets
+- The artifact evaluates four open-source web applications. The scripts will automatically download their source code from GitHub and checkout the corresponding versions. 
+- The directory `bitcode` in the artifact includes the pre-compiled bitcode from each application by `wllvm` with flag `-fno-inline` `-O0` and `-g`. The bitcodes are broken into different modules to reduce the inter-procedural value analysis time of SVF.
+
+## Steps to reproduce 
+- We provide a script `./install.sh` to automatically install the dependencies and build the software. 
+- We provide a script `./run.sh` to automatically perform the evaluation. 
+
+```python
+./install.sh
+# Step 0: Clean the previous output files
+# Step 1: Install dependencies, compile SVF
+# Step 2: Compile ValueCheck
+
+./run.sh
+# Step 3: Run ValueCheck - the analysis tool. 
+# Step 4: Run ValueCheck and produce the result
+```
+
+## Evaluation and expected results
+We provide the scripts to automate the evaluation and generate the Tables and numbers in Section 4. 
+The output will be in the `result/` folder and contain the following key results:
+- `result/table_2_detected_bugs.csv`: 
+  - Total number of detected bugs from each application. (`Table 2`) 
+- `result/table_6_dok_effect.csv`: 
+  - The number of detected bugs within top 20 bugs under different DOK settings. (`Table 6`)
+- `result/figure_7_dist.pdf`, `result/figure_7_security.pdf`, `result/figure_7_days.pdf`: 
+  - The category of bugs based on distribution, security, and days before detected. (`Figure 7`)
+- `result/figure_9_detected_bug_dok.pdf`: 
+  - The figure of reported bugs when increasing DOK rank. (`Figure 9`)
+- `result/table_7_time_analysis.csv`: 
+  - Time (seconds) to run the analysis. (First column of `Table 7`)
+
+- In the `result/APP_NAME/` directory, `detected.csv` contains all the detected bugs.
+
+Note that some results involve random sampling (second and third columns in `Table 6`) and developers' confirmation (last column in `Table 2`), thus not included in the artifact. 
+Note that due to the differences in hardware environments and the optimization we later add to the tool, the performance results in `Table 7` can be different from the numbers reported in the paper.

SVF/.gitignore

@@ -0,0 +1,24 @@
+Release*/
+Debug*/
+build/
+html/
+Test-Suite
+Release+Asserts/
+Debug+Asserts/
+autoconf/
+tests/result/
+doxygen/
+.*
+!.gitignore
+*~
+*.o
+*.out
+*.ll
+*.bc
+*.opt
+*.log
+*.status
+*.obj
+*.svf
+cmake-build-debug/
+compile_commands.json

SVF/CMakeLists.txt

@@ -0,0 +1,90 @@
+cmake_minimum_required(VERSION 3.13.4)
+
+project("SVF")
+
+configure_file(${PROJECT_SOURCE_DIR}/.config.in
+               ${PROJECT_BINARY_DIR}/include/Util/config.h)
+
+# We need to match the build environment for LLVM: In particular, we need C++14
+# and the -fno-rtti flag
+set(CMAKE_CXX_STANDARD 14)
+set(CMAKE_CXX_STANDARD_REQUIRED ON)
+# add -std=gnu++14
+set(CMAKE_CXX_EXTENSIONS ON)
+
+add_compile_options("-fno-rtti")
+add_compile_options("-fno-exceptions")
+
+# Treat compiler warnings as errors
+add_compile_options("-Werror" "-Wall")
+add_compile_options("-fPIC")
+
+# Keep assertions enabled if requested
+option(SVF_ENABLE_ASSERTIONS "Always enable assertions")
+if(SVF_ENABLE_ASSERTIONS)
+  add_compile_options("-UNDEBUG")
+endif()
+
+# Turn this on if you need symbols (e.g., use them for backtrace debugging)
+# add_link_options("-rdynamic")
+
+option(SVF_COVERAGE "Create coverage build")
+if(SVF_COVERAGE OR DEFINED ENV{SVF_COVERAGE})
+  add_compile_options("-fprofile-arcs" "-ftest-coverage")
+  add_link_options("-fprofile-arcs" "-ftest-coverage")
+  message(STATUS "Enable coverage")
+endif()
+
+set(SVF_SANITIZE
+    ""
+    CACHE STRING "Create sanitizer build (address)")
+if(SVF_SANITIZE STREQUAL "address")
+  add_compile_options("-fno-omit-frame-pointer" "-fsanitize=address")
+  add_link_options("-fsanitize=address")
+  message(STATUS "Sanitizer build: ${SVF_SANITIZE}")
+elseif(SVF_SANITIZE STREQUAL "thread")
+  add_compile_options("-fsanitize=thread")
+  add_link_options("-fsanitize=thread")
+  message(STATUS "Sanitizer build: ${SVF_SANITIZE}")
+elseif(NOT SVF_SANITIZE STREQUAL "")
+  message(ERROR "Unknown sanitizer type: ${SVF_SANITIZE}")
+endif()
+
+find_library(
+  Z3_LIBRARIES
+  NAMES z3
+  HINTS ${Z3_DIR} ENV Z3_DIR
+  PATH_SUFFIXES bin lib)
+find_path(
+  Z3_INCLUDES
+  NAMES z3++.h
+  HINTS ${Z3_DIR} ENV Z3_DIR
+  PATH_SUFFIXES include z3)
+if(NOT Z3_LIBRARIES OR NOT Z3_INCLUDES)
+  message(FATAL_ERROR "Z3 not found!")
+endif()
+message(STATUS "Found Z3: ${Z3_LIBRARIES}")
+message(STATUS "Z3 include dir: ${Z3_INCLUDES}")
+
+include_directories(${PROJECT_SOURCE_DIR}/svf/include
+                    ${PROJECT_BINARY_DIR}/include ${Z3_INCLUDES})
+
+# checks if the test-suite is present, if it is then build bc files and add
+# testing to cmake build
+if(EXISTS "${PROJECT_SOURCE_DIR}/Test-Suite")
+  include_directories(${PROJECT_SOURCE_DIR}/Test-Suite)
+  enable_testing()
+  add_subdirectory(Test-Suite)
+  include(CTest)
+endif()
+
+add_subdirectory(svf)
+add_subdirectory(svf-llvm)
+
+install(
+  DIRECTORY ${PROJECT_SOURCE_DIR}/svf/include/
+            ${PROJECT_SOURCE_DIR}/svf-llvm/include/
+  COMPONENT devel
+  DESTINATION include/svf
+  FILES_MATCHING
+  PATTERN "**/*.h")

SVF/Dockerfile

@@ -0,0 +1,36 @@
+FROM ubuntu:20.04
+
+# Stop ubuntu-20 interactive options.
+ENV DEBIAN_FRONTEND noninteractive
+
+# Stop script if any individual command fails.
+RUN set -e
+
+# Define LLVM version.
+ENV llvm_version=14.0.0
+
+# Define home directory
+ENV HOME=/home/SVF-tools
+
+# Define dependencies.
+ENV lib_deps="make g++-8 gcc-8 git zlib1g-dev libncurses5-dev build-essential libssl-dev libpcre2-dev zip vim libtinfo5"
+ENV build_deps="wget xz-utils cmake python git gdb tcl"
+
+# Fetch dependencies.
+RUN apt-get update --fix-missing
+RUN apt-get install -y $build_deps $lib_deps
+
+# Fetch and build SVF source.
+RUN echo "Downloading LLVM and building SVF to " ${HOME}
+WORKDIR ${HOME}
+RUN git clone "https://github.com/SVF-tools/SVF.git"
+WORKDIR ${HOME}/SVF
+RUN echo "Building SVF ..."
+RUN bash ./build.sh 
+
+# Export SVF, llvm, z3 paths
+ENV PATH=${HOME}/SVF/Release-build/bin:$PATH
+ENV PATH=${HOME}/SVF/llvm-$llvm_version.obj/bin:$PATH
+ENV SVF_DIR=${HOME}/SVF
+ENV LLVM_DIR=${HOME}/SVF/llvm-$llvm_version.obj
+ENV Z3_DIR=${HOME}/SVF/z3.obj