Giter Site home page Giter Site logo

sectrs-acai / acai Goto Github PK

View Code? Open in Web Editor NEW
18.0 3.0 1.0 11.33 MB

Protecting Accelerator Execution with Arm Confidential Computing Architecture (USENIX Security 2024)

Home Page: https://sectrs.ethz.ch/

License: Other

Makefile 3.18% Shell 16.49% Python 3.37% C 76.96%
armv9 confidential-computing armcca

acai's Introduction

ACAI

License: MIT License: MIT acai-artifact-eval Twitter

Abstract

Trusted execution environments in several existing and upcoming CPUs demonstrate the success of confidential computing, with the caveat that tenants cannot use accelerators such as GPUs and FPGAs. Even after hardware changes to enable TEEs on both sides and software changes to adopt existing code to leverage these features, it results in redundant data copies and hardware encryption at the bus-level and on the accelerator thus degrading the performance and defeating the purpose of using accelerators. In this paper, we reconsider the Arm Confidential Computing Architecture (CCA) design โ€” an upcoming TEE feature in Arm v9 โ€” to address this gap. We observe that CCA offers the right abstraction and mechanisms to allow confidential VMs to use accelerators as a first class abstraction, while relying on the hardware-based memory protection to preserve security. We build ACAI, a CCA-based solution, to demonstrate the feasibility of our approach while addressing several critical security gaps. Our experimental results on GPU and FPGA show that ACAI can achieve strong security guarantees while maintaining performance and compatibility.

@article{DBLP:journals/corr/abs-2305-15986,
  author       = {Supraja Sridhara and Andrin Bertschi and Benedict Schl{\"{u}}ter and Mark Kuhne and Fabio Aliberti and Shweta Shinde},
  title        = {{ACAI:} Extending Arm Confidential Computing Architecture Protection
                  from CPUs to Accelerators},
  journal      = {CoRR},
  volume       = {abs/2305.15986},
  year         = {2023},
  url          = {https://doi.org/10.48550/arXiv.2305.15986},
  doi          = {10.48550/ARXIV.2305.15986},
  eprinttype    = {arXiv},
  eprint       = {2305.15986},
  timestamp    = {Wed, 07 Jun 2023 14:31:13 +0200},
  biburl       = {https://dblp.org/rec/journals/corr/abs-2305-15986.bib},
  bibsource    = {dblp computer science bibliography, https://dblp.org}
}

https://arxiv.org/abs/2305.15986

Build the Project

Please follow Artifact Evaluation for instructions how to build and run the project. You find more tutorials and documentation here.

Files and Directories

/ext...................................: External project dependencies 
/buildconf.............................: Scripts and configuration to build artifacts
/scripts...............................: Helper scripts
/output................................: Artifact output
/output-distrobox......................: Artifact output built in container
/src...................................: Sources and source submodules
/scr/tfa...............................: TFA Monitor
/src/rmm...............................: RMM
/src/linux-cca-guest...................: CCA-enabled kernel
/src/fpga_driver.......................: PCIe bypass for FPGA
/src/fpga_driver/fh_host...............: ioctl library for PCIe bypass on x86
/src/fpga_driver/fpga_escape_libhook...: x86 FPGA Userspace Manager
/src/fpga_driver/xdma..................: x86 FPGA Host driver
/src/fpga_driver/xdma_stub.............: aarch64 FPGA guest driver
/src/fpga_driver/libhook...............: FVP memory alignment for DMA/ mmap
/src/fpga_driver/devmem_intercept .....: ACAI aarch64 kernel helper
/src/gpu_driver........................: PCIe bypass for GPU
/src/gpu_driver/gdev-guest.............: aarch64 GPU guest driver
/src/gpu_driver/gdev-host..............: x86 GPU host driver
/src/gpu_driver/gpu_gdev_usr_manager...: x86 GPU Userspace Manager
/src/gpu_driver/rodinia-bench..........: GPU Benchmarks CUDA Driver API
/src/benchmarking/fpga.................: FPGA Benchmarks
/src/linux-host........................: Faulthook host kernel
/src/encrypted-cuda....................: Encryption layer
/src/kvmtool...........................: Virtual Machine Manager
/src/tfa-tests.........................: TFA tests

acai's People

Contributors

abertschi avatar benschlueter avatar

Stargazers

 avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar

Watchers

 avatar  avatar  avatar

Forkers

chang-steve

acai's Issues

Error: unknown option `realm'

Hi experts,

I followed the instructions in artifact-evaluation.md, and got this error. lkvm run does not have the --realm argument:

# ./run_toy_realm.sh
Booting a realm...
Command:
+ nice -n -20 taskset 0x4 ./../assets/snapshots/lkvm run --realm --disable-sve -c 1 -m 800 -k ../assets/snapshots/Image-cca -i ../assets/snapshots/rootfs.realm.cpio --9p /,host0 -p 'fvp_escape_loop fvp_escape_off ip=off'
  Error: unknown option `realm'

What problem is this? and where is the source code of lkvm?

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.