PhASAR a LLVM-based Static Analysis Framework

Version 2403

Secure Software Engineering Group

PhASAR is primarily developed and maintained by the Secure Software Engineering Group at Heinz Nixdorf Institute (University of Paderborn) and Fraunhofer IEM.

Lead developers of PhASAR are: Fabian Schiebel (@fabianbs96)(fabian.schiebel@iem.fraunhofer.de), Martin Mory (@MMory)(martin.mory@upb.de), Philipp Dominik Schubert (@pdschubert)(philipp.schubert@upb.de) and others.

Required Version of the C++ Standard

PhASAR requires C++-17.

However, building in C++20 mode is supported as an experimental feature. You may enable this setting the cmake variable CMAKE_CXX_STANDARD to 20. Although phasar currently does not make use of C++-20 features (except for some concepts behind an #ifdef border), your client application that just uses phasar as a library may want to use C++20 ealier.

Currently Supported Version of LLVM

PhASAR is currently set up to support LLVM-14.0.*

What is PhASAR?

PhASAR is a LLVM-based static analysis framework written in C++. It allows users to specify arbitrary data-flow problems which are then solved in a fully-automated manner on the specified LLVM IR target code. Computing points-to information, call-graph(s), etc. is done by the framework, thus you can focus on what matters.

Breaking Changes

To keep PhASAR in a state that it is well suited for state-of-the-art research in static analysis, as well as for productive use, we have to make breaking changes. Please refer to Breaking Changes for detailed information on what was broken recently and how to migrate.

How do I get started with PhASAR?

We have some documentation on PhASAR in our Wiki. You probably would like to read this README first.

Please also have a look on PhASAR's project directory and notice the project directory examples/ as well as the custom tool tools/example-tool/myphasartool.cpp.

Building PhASAR

It is recommended to compile PhASAR yourself in order to get the full C++ experience and to have full control over the build mode. However, you may also want to try out one of the pre-built versions of PhASAR or the Docker container.

As a shortcut for the very first PhASAR build on your system, you can use our bootstrap script. Please note that you must have python installed for the script to work properly.

./bootstrap.sh

Note: If you want to do changes within PhASAR, it is recommended to build it in Debug mode:

./bootstrap.sh -DCMAKE_BUILD_TYPE=Debug

The bootstrap script may ask for superuser permissions (to install the dependencies); however it is not recommended to start the whole script with sudo.

For subsequent builds, see Compiling PhASAR.

Compiling PhASAR (if not already done using the bootstrap script)

Set the system's variables for the C and C++ compiler to clang:

export CC=/usr/local/bin/clang
export CXX=/usr/local/bin/clang++

You may need to adjust the paths according to your system. When you cloned PhASAR from Github you need to initialize PhASAR's submodules before building it:

git submodule update --init

If you downloaded PhASAR as a compressed release (e.g. .zip or .tar.gz) you can use the init-submodules-release.sh script that manually clones the required submodules:

utils/init-submodules-release.sh

Navigate into the PhASAR directory. The following commands will do the job and compile the PhASAR framework:

mkdir build
cd build/
cmake -G Ninja -DCMAKE_BUILD_TYPE=Release ..
ninja -j $(nproc) # or use a different number of cores to compile it
sudo ninja install # only if you wish to install PhASAR system wide

When you have used the bootstrap.sh script to install PhASAR, the above steps are already done. Use them as a reference if you wish to modify PhASAR and recompile it.

After compilation using cmake the following two binaries can be found in the build/tools directory:

• phasar-cli - the PhASAR command-line tool (previously called phasar-llvm) that provides access to analyses that are already implemented within PhASAR. Use this if you don't want to build an own tool on top of PhASAR.
• myphasartool - an example tool that shows how tools can be build on top of PhASAR

Please be careful and check if errors occur during the compilation.

When using CMake to compile PhASAR the following optional parameters can be used:

Parameter : Type	Effect
BUILD_SHARED_LIBS : BOOL	Build shared libraries -- Not recommended anymore. You may want to use PHASAR_BUILD_DYNLIB instead (default is OFF)
PHASAR_BUILD_DYNLIB : BOOL	Build one fat shared library (default is OFF)
CMAKE_BUILD_TYPE : STRING	Build PhASAR in 'Debug', 'RelWithDebInfo' or 'Release' mode (default is 'Debug')
CMAKE_INSTALL_PREFIX : PATH	Path where PhASAR will be installed if "ninja install” is invoked or the “install” target is built (default is /usr/local/phasar)
PHASAR_CUSTOM_CONFIG_INSTALL_DIR : PATH	If set, customizes the directory, where configuration files for PhASAR are installed (default is /usr/local/.phasar-config)
PHASAR_ENABLE_DYNAMIC_LOG : BOOL	Makes it possible to switch the logger on and off at runtime (default is ON)
PHASAR_BUILD_DOC : BOOL	Build PhASAR documentation (default is OFF)
PHASAR_BUILD_UNITTESTS : BOOL	Build PhASAR unit tests (default is ON)
PHASAR_BUILD_IR : BOOL	Build PhASAR IR (required for running the unit tests) (default is ON)
PHASAR_BUILD_OPENSSL_TS_UNITTESTS : BOOL	Build PhASAR unit tests that require OpenSSL (default is OFF)
PHASAR_ENABLE_PAMM : STRING	Enable the performance measurement mechanism ('Off', 'Core' or 'Full', default is Off)
PHASAR_ENABLE_PIC : BOOL	Build Position-Independed Code (default is ON)
PHASAR_ENABLE_WARNINGS : BOOL	Enable compiler warnings (default is ON)
CMAKE_CXX_STANDARD : INT	Build phasar in C++17 or C++20 mode (default is 17)

You can use these parameters either directly or modify the installer-script bootstrap.sh

A Remark on Compile Time

C++'s long compile times are always a pain. As shown in the above, when using cmake the compilation can easily be run in parallel, resulting in shorter compilation times. Make use of it!

Running a Test Solver

To test if everything works as expected please run the following command:

$ phasar-cli -m test/llvm_test_code/basic/module_cpp.ll -D ifds-solvertest

You can find the phasar-cli tool in the build-tree under tools/phasar-cli.

If you obtain output other than a segmentation fault or an exception terminating the program abnormally everything works as expected.

Building PhASAR on a MacOS System

Due to unfortunate updates to MacOS and the handling of C++, especially on the newer M1 processors, we can't support native development on Mac. The easiest solution to develop PhASAR on a Mac right now is to use dockers development environments. Clone this repository as described in their documentation. Afterwards, you have to login once manually, as a root user by running docker exec -it -u root <container name> /bin/bash to complete the rest of the build process as described in this readme (install submodules, run bootstrap.sh, ...). Now you can just attach your docker container to VS Code or any other IDE, which supports remote development.

Installation

PhASAR can be installed using the installer scripts as explained in the following. However, you do not need to install PhASAR in order to use it.

Installing PhASAR on an Ubuntu System

In the following, we would like to give an complete example of how to install PhASAR using an Ubuntu or Unix-like system.

Therefore, we provide an installation script. To install PhASAR, just navigate to the top-level directory of PhASAR and use the following command:

./bootstrap.sh --install

The bootstrap script may ask for superuser permissions.

Done!

If You have already built phasar, you can just invoke

sudo ninja install

How to use PhASAR?

We recomment using phasar as a library with cmake.

If you already have installed phasar, Use-PhASAR-as-a-library may be a good start.

Otherwise, we recommend adding PhASAR as a git submodule to your repository. In this case, just add_subdirectory the phasar submodule directory within your CMakeLists.txt.

Assuming you have checked out phasar in external/phasar, the phasar-related cmake commands may look like this:

add_subdirectory(external/phasar EXCLUDE_FROM_ALL)            # Build phasar with your tool

...

target_link_libraries(yourphasartool
    ...
    phasar # Make your tool link against phasar
)

Depending on your use of PhASAR you also may need to add LLVM to your build.

For more information please consult our PhASAR wiki pages.

Please help us to improve PhASAR

You are using PhASAR and would like to help us in the future? Then please support us by filling out this web form.

By giving us feedback you help to decide in what direction PhASAR should stride in the future and give us clues about our user base. Thank you very much!

Installing PhASAR's Git pre-commit hook

You are very much welcome to contribute to the PhASAR project. Please make sure that you install our pre-commit hook that ensures your commit adheres to the most important coding rules of the PhASAR project. For more details please consult Coding Conventions and Contributing to PhASAR.

To install the pre-commit hook, please run the following commands in PhASAR's root directory:

pip install pre-commit
pre-commit install

Thanks. And have fun with the project.