Giter Site home page Giter Site logo

chan9yan9 / nyx Goto Github PK

View Code? Open in Web Editor NEW

This project forked from rub-syssec/nyx

0.0 0.0 0.0 47.44 MB

USENIX 2021 - Nyx: Greybox Hypervisor Fuzzing using Fast Snapshots and Affine Types

License: GNU General Public License v2.0

Shell 0.56% Python 2.39% C 88.20% Assembly 0.42% Rust 8.27% Makefile 0.16%

nyx's Introduction

Nyx: Greybox Hypervisor Fuzzing using Fast Snapshots and Affine Types

Nyx is fast full-VM snapshot fuzzer for type-2 hypervisors. It's built upon kAFL, Redqueen and Hyper-Cube. The fuzzer is based on our paper (slides, recording):

@inproceedings {nyx,
author = {Sergej Schumilo and Cornelius Aschermann and Ali Abbasi and Simon W{\"o}r-ner and Thorsten Holz},
title = {Nyx: Greybox Hypervisor Fuzzing using Fast Snapshots and Affine Types},
booktitle = {30th {USENIX} Security Symposium ({USENIX} Security 21)},
year = {2021},
url = {https://www.usenix.org/conference/usenixsecurity21/presentation/schumilo},
}

Please note:

This release is based on our academic prototype implementation from early 2020 and it is quite outdated and has several known bugs, specific limitations and gets unstable in specific use cases (e.g. this version does not use the more stabile libxdc Intel-PT decoder).

If you are interested in using Nyx or one of Nyx's components for bug hunting or to build your project upon, we are recommending to use The public version of Nyx instead. It is a mainted, improved version of Nyx with all of its latest and greatest features. Other than that, we are not planning to maintain this prototype implementation anymore. So please do yourself a favor and use the real release of Nyx instead.

Setup:

Install dependencies:

The following commands will install all required dependencies (assuming an Ubuntu 20.04):

sudo apt-get install gcc-8 libisoburn1 libgcc-8-dev-i386-cross libx32gcc-8-dev-i386-cross build-essential curl qemu-utils python3-pip python-is-python3
pip install msgpack
sudo apt-get remove libcapstone3

# install rust
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh

# install capstone-v4
git clone https://github.com/aquynh/capstone.git && \
cd capstone && \
git checkout v4 && \
make && \
sudo make install && \
cd -

# install all QEMU build dependencies (requires deb-src URIs in your /etc/apt/sources.list file)
sudo apt-get build-dep qemu

KVM-PT

One of Nyx's main components is KVM-PT. This standalone LKM is based on KVM (Kernel 4.20-rc4), but implements multiple features on top of KVM required to build a fast coverage-guided ring -1 fuzzer. KVM-PT provides an Intel-PT tracing module for vCPUs (instead of Intel-PT tracing per logical CPU), nested hypercalls (which are fully transparent to L1), an in-kernel dirty-memory tracker and many more other hacks.

  1. Install the following kernel and reboot your system. Ensure that you are using the right kernel by using uname:

    cd /tmp/ && \
    wget -c https://kernel.ubuntu.com/~kernel-ppa/mainline/v4.19/linux-headers-4.19.0-041900_4.19.0-041900.201810221809_all.deb && \
    wget -c https://kernel.ubuntu.com/~kernel-ppa/mainline/v4.19/linux-headers-4.19.0-041900-generic_4.19.0-041900.201810221809_amd64.deb && \
    wget -c https://kernel.ubuntu.com/~kernel-ppa/mainline/v4.19/linux-image-unsigned-4.19.0-041900-generic_4.19.0-041900.201810221809_amd64.deb && \
    wget -c https://kernel.ubuntu.com/~kernel-ppa/mainline/v4.19/linux-modules-4.19.0-041900-generic_4.19.0-041900.201810221809_amd64.deb && \
    sudo dpkg -i *.deb && \
    sudo reboot
    
    
  2. After the system reboot verify that your system is using kernel 4.19 (by using uname -a). Otherwise, you can manually set kernel 4.19 in your grub configuration by adding the following line to /etc/default/grub and by generating a new GRUB config via update-grub:

GRUB_DEFAULT="Advanced options for Ubuntu>Ubuntu, with Linux 4.19.0-041900-generic"
  1. Execute load.sh in KVM-PT/ to compile and load KVM-PT.
  2. Check if the /dev/kvm-pt device node has been created.

QEMU-PT:

QEMU-PT is the user-space counterpart of KVM-PT. It is based on QEMU 4.2.50 and implements userspace handlers for hypercalls, fast snapshots and a custom Intel-PT decoder.

  1. Run sh compile_qemu_pt.h in QEMU-PT/ to build QEMU-PT. `

Build GRUB:

  1. Move to nyx_fuzzer/hypertrash_spec/build/hypertrash_os/grub/ and unpack grub.zip (unzip grub.zip). In case you want to compile GRUB by yourself instead of using our precompiled binaries, you can use install.sh to do so (gcc-6 is required).

Nyx-Fuzzer:

  1. Move to nyx_fuzzer/rust_fuzzer and run cargo build --release to build the fuzzer.

Getting Started:

bhyve (AHCI) Fuzzing:

  1. Move to the Targets/bhyve/VM folder.
  2. Execute prepare.sh. This script will download an ISO image of FreeBSD 11-3, prepare a virtual disk and spawn a QEMU instance. Connect via VNC (localhost:5900) and install FreeBSD. Install the following packages: base-dbg, kernel-dbg, ib32, ports. Create the follwing users: root / password: root, user / password: userand add user to wheelgroup.
  3. Once the OS installation has finished, reboot the VM and execute copy_files.shon the host (in another shell). Connect via SSH (localhost:2222) and run install.sh as root. Once the script has finished and the OS has shutdown, you can terminate QEMU via CTRL+c.
  4. Run create_pre_snapshot.sh and connect via VNC. Login as rootand run /home/user/loader. QEMU will create a pre-snapshot image and terminate.
  5. Create a config file by executing cp Targets/bhyve/config_template.ron Targets/bhyve/config.ron and adjust all paths in Targets/bhyve/config.ron (use absolute paths).
  6. Run make_sharedir.sh in nyx_fuzzer/hypertrash_spec/.
  7. Move to nyx-fuzzer/rust_fuzzer and execute cargo run --release -- -c ../../Targets/bhyve/config.ron to start the fuzzer.
  8. Use the following command to show all findings: cat /tmp/workdir_bhyve/corpus/crash/*.log | sort | uniq -c

qemu (legacy XHCI) Fuzzing:

  1. Move to the Targets/qemu/VM folder.
  2. Execute prepare.sh. This script will download an ISO image of Ubuntu 18-04 LTS, prepare a virtual disk and spawn a QEMU instance. Connect via VNC (localhost:5900) and install Ubuntu. Create the following users: user: user / password: userand add install openssh-server.
  3. Once the OS installation has finished, reboot the VM and execute copy_install_files.shon the host (in another shell). Connect via SSH (localhost:2223) and run ./install.sh as root. Once the install script is done and the OS has shutdown, you can terminate QEMU via CTRL+c.
  4. Run create_pre_snapshot.sh and connect via VNC. Login as rootand run /home/user/loader. QEMU will create a pre-snapshot image and exit.
  5. Create a config file by executing cp Targets/qemu/config_template.ron Targets/qemu/config.ron and adjust all paths in Targets/qemu/config.ron (use absolute paths).
  6. Run make_sharedir.sh in nyx_fuzzer/hypertrash_spec/.
  7. Move to Targets/qemu/agent and run build_sharedir.sh.
  8. Move to nyx-fuzzer/rust_fuzzer and execute cargo run --release -- -c ../../Targets/qemu/config.ron to start the fuzzer.
  9. Use the following command to show all findings: cat /tmp/workdir_qemu/corpus/crash/*.log | sort | uniq -c

License

Nyx is provided under GPLv2 license.

Free Software Hell Yeah!

Proudly provided by:

nyx's People

Contributors

rub-syssec avatar eqv 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.