Git Product home page Git Product logo

libptrace's Introduction

libptrace

libptrace is an event driven process/thread debugging, tracing, and manipulation framework. It is written in C and has Python 2.7 and 3.7 bindings.

It is meant to be used as a library. The API has been designed with cross-platform support in mind. Although the current version only runs on 32- and 64-bit (wow64 is supported) versions of Windows the design and abstractions are such that other platforms and even remote debugging support can be added under the same API.

It is multiple core/thread aware, and can easily be used to run multiple event loops concurrently. Debugging process groups or pipelines or tracing a large number of processes can be scaled by increasing execution cores this way.

Foreword

I wrote the initial version of libptrace in 2006 while still in university, and re-licensed it for full use by Immunity Inc. in 2011, my employer at the time. Around this time the framework was meant to be integrated in Immunity Debugger, although I have always taken care to keep the framework separate, so it could be used in a stand-alone fashion.

With this goal in mind it saw a lot of development at Immunity, and I am grateful to have been able to actively develop on it during my working hours and see the framework become more mature. I redesigned most of it to be aware of multiple cores, to provide Python bindings, and create better abstractions.

With the eventual release of x64dbg it was decided we would no longer work on Immunity Debugger, and instead focus our efforts on other things. This also stalled libptrace development, and for the following years it collected dust in the internal git repository, until after the acquisition of Immunity Inc. by Cyxtera Technologies, now under Cyxtera Cybersecurity Inc., it was decided to release the project under the LGPL version 2.1 so it could hopefully see adoption and further development.

I want to kindly thank Immunity Inc., as well as Cyxtera Technologies, and Cyxtera Cybersecurity Inc. for making this release happen. Also I'd like to specifically thank Dave Aitel for pushing this release through layers of management and shielding me from the bureaucratic parts of the process. It would have been sad to see my labor lost and without purpose, and without their willingness to release the libptrace development that was done internally as open source this is what would have happened.

I also want to thank several coworkers that spent time using or working on the framework, most notably Massimiliano Oldani, Roderick Asselineau, and Christos Kalkanis. My thanks also extends to Lennert Buytenhek for his intrusive AVL tree and linked list implementations which I ended up using.

Although it has been over 4 years since any work has been done on the project, I hope people will find it useful. I believe the design and source code are relatively clean, and at this point allows the framework to be extended and ported to other platforms relatively easily. Hopefully one day this can turn into a true cross-platform debugging library under a single consistent API.

Finally, I'd like to dedicate this project to my lost daughter, Yuzuyu Huizer, with whom I hope to one day reunite. I love you.

-- Ronald Huizer [email protected]

Building

The current supported build method is by cross-compiling the project using MinGW and CMake. The build system has been tested on Ubuntu 19.04.

A 32-bit installer can be built using:

cmake -DCMAKE_BUILD_TYPE=Release -DCMAKE_TOOLCHAIN_FILE=mingw-w64-x86.cmake .
make
makensis libptrace-setup.nsi

A 64-bit bit installer can be built using:

cmake -DCMAKE_BUILD_TYPE=Release -DCMAKE_TOOLCHAIN_FILE=mingw-w64-x64.cmake .
make
makensis -DUSE64 libptrace-setup.nsi

Installation

After following the build steps above, the 32-bit installer will be at libptrace-setup32.exe and the 64-bit installer will be at libptrace-setup64.exe.

The installer will automatically determine what versions of Python are installed and whether they are 32-bit or 64-bit. The versions that were found will be offered by the installer, and versions that were not found will be grayed out.

Release downloads can be found here.

Examples

The doc/ directory is hopelessly outdated. It is included as a starting point for more current documentation. Please do not use it.

Several example scripts can be found in the python/scripts directory. They will be installed to Program Files\Immunity Inc\libptrace or in case of the 32-bit version Program Files (x86)\Immunity Inc\libptrace.

These examples do in no way show the framework exhaustively, but they should provide a good starting point for anyone that wants to use the framework. The C API has not been documented well, but some examples can be found in the unittests/windows directory. The python/ directory which contains the python bindings also provides a good overview. Finally, the include/libptrace directory contains the header files which are public, and as such provide a good list of API functions.

It should be noted the Python API is currently a fairly limited subset of the C API.

A python example

As an example, the general workflow to trace RegOpenKeyExW calls in an executable would be:

import _ptrace

def bp_handler(bp, thread):
    (key, subkey, options, sam, result) = _ptrace.cconv.args_get(thread, "%u%p%u%u%p")
    subkey = thread.process.read_utf16(subkey)
    print('T{}: RegOpenKeyEx({}, "{}", 0x{:08x}, 0x{:08x}, 0x{:08x})'
          .format(thread.id, key, subkey, options, sam, result), end='')

def attached_handler(process):
    bp = _ptrace.breakpoint_sw("advapi32!RegOpenKeyExW", bp_handler)
    process.breakpoint_set(bp)

handlers          = _ptrace.event_handlers()
handlers.attached = attached_handler

_ptrace.execv(r"C:\Windows\Notepad.exe", [], handlers, 0)
_ptrace.main()

If the return value of RegOpenKeyExW is also desired, it is easiest to set a breakpoint on the return address of the function call. This avoids the need for code analysis to determine function exit paths statically:

def bp_end_handler(bp, thread):
    print("=", _ptrace.cconv.retval_get(thread))

def bp_handler(bp, thread):
    (key, subkey, options, sam, result) = _ptrace.cconv.args_get(thread, "%u%p%u%u%p")
    subkey = thread.process.read_utf16(subkey)
    print('T{}: RegOpenKeyEx({}, "{}", 0x{:08x}, 0x{:08x}, 0x{:08x})'
          .format(thread.id, key, subkey, options, sam, result), end='')

    retaddr = _ptrace.cconv.retaddr_get(thread)
    if thread.process.breakpoint_find(retaddr) is None:
        bp_end = _ptrace.breakpoint_sw(retaddr, bp_end_handler)
        thread.process.breakpoint_set(bp_end)

Copyright

The copyright situation is made clear in each individual file. They are mostly held by Cyxtera Cybersecurity Inc., Ronald Huizer, or both at the same time. A select few files have third party copyright holders, such as Lennert Buytenhek.

License

Distributed under version 2.1 of the GNU Lesser General Public License. See COPYING and COPYING.LESSER for more information.

Authors

  • Ronald Huizer - project design and implementation
  • Massimiliano Oldani - symbol handling, bug fixes
  • Roderick Asselineau - pe improvements, interval tree, bug fixes
  • Christos Kalkanis - testing, improvements, bug fixes

Contact

Acknowledgments

  • Immunity Inc. and Cyxtera Cybersecurity Inc. for making this release happen.
  • Lennert Buytenhek for his intrusive AVL tree and linked list implementation.
  • Dave Aitel for dealing with the bureaucratic and legal parts of this release.

libptrace's People

Contributors

fengjixuchui avatar rhuizer 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.