bin2cet artifacts

# This README presents the workflow required to reproduce the experimental results and validation tests described in the article *Retrofitting Intel CET's Indirect Branch Tracking into Legacy Binaries through Static Binary Rewriting*. # It provides instructions to prepare the artifact environment, execute the available demonstration, rebuild the experimental data, and extract the metrics used to evaluate the proposed approach. ## Abstract ** This artifact reproduces the experiments conducted in Section 5. **** A Docker container with scripts to recreate the paper results can be found in https://doi.org/10.5281/zenodo.20420793. **** Additionally, the artifact contains a demonstration script with a step-by-step guide to use BIN2CET to make the program in Figure 1 CET-compatible. ** ## Checklist * **Program:** BIN2CET, LIEF 0.17.6, E9Patch, PIN 4.2, Ghidra 11.0.3, perf, ROPgadget, and CoreUtils 9.11.* **Compilation:** gcc-15.* **Run-time environment:** Linux on x86-64 architecture inside a Docker container.* **Hardware:** Intel processor with hardware counters, required for perf, and Docker support.* **Metrics:** Effectiveness (ß5.1), runtime overhead (ß5.2), and attack surface reduction (ß5.3).* **Experiments:** Rebuild the paper results by producing data and extracting information into CSV files.* **Required disk space:** The Docker image requires approximately 4.6 GB uncompressed, and the dataset requires approximately 2.4 GB uncompressed.* **Workflow preparation time:** Approximately 1 hour to build the Docker image.* **Experiment execution time:** Approximately 36 hours.* **Availability:** BIN2CET is publicly available at ***REMOVED***, and the artifact is available at https://doi.org/10.5281/zenodo.20420793. ## Installation ### 1. Download and decompress the artifact ```tar -jxvf bin2cet-artifacts.tar.bz2``` ### 2. Prepare the dataset Use the dataset provided with the artifact: ```sudo tar --same-owner -xvf dataset.tar``` Alternatively, create a directory to build the dataset from scratch: ```mkdir datasetchmod 777 dataset``` ### 3. Prepare the Docker image Build the image: ```docker build --platform=linux/amd64 -t bin2cet_image .``` Alternatively, load the prebuilt image: ```bunzip2 bin2cet_image.tar.bz2docker load -i bin2cet_image.tar``` ### 4. Run the container ```docker run --rm -it --privileged --name bin2cet_$$ \ -v ./dataset:/work/dataset bin2cet_image``` ## Experiment Workflow From within the Docker container terminal, use one of the following commands: * `demo`: performs a step-by-step demonstration of BIN2CET.* `sim`: executes scripts to simulate an optimization level.* `extract`: extracts analyses for the three research questions.* `everything`: simulates and extracts all results.# 本自述文件介绍了复现论文《通过静态二进制重写为遗留二进制文件适配Intel CET间接分支跟踪》中所述实验结果与验证测试所需的工作流。本文档提供了准备实验工件环境、运行可用演示程序、重建实验数据以及提取用于评估所提方法的评估指标的操作指南。 ## 摘要 **本工件可复现论文第5节所述的实验。** 可复现该论文结果的Docker容器脚本可在https://doi.org/10.5281/zenodo.20420793获取。** 此外，本工件包含一份演示脚本，附带分步指南，用于使用BIN2CET使图1中的程序兼容CET。 ## 核对清单 * **所用程序：** BIN2CET、LIEF 0.17.6、E9Patch、PIN 4.2、Ghidra 11.0.3、perf、ROPgadget 以及 CoreUtils 9.11。 * **编译工具：** gcc-15。 * **运行时环境：** Docker容器内的x86-64架构Linux系统。 * **硬件要求：** 支持硬件计数器（供perf工具使用）且兼容Docker的Intel处理器。 * **评估指标：** 有效性（对应第5.1节）、运行时开销（对应第5.2节）以及攻击面缩减效果（对应第5.3节）。 * **实验内容：** 通过生成数据并将信息导出至CSV文件来重建论文结果。 * **所需磁盘空间：** Docker镜像未压缩大小约为4.6 GB，数据集未压缩大小约为2.4 GB。 * **工作流准备时长：** 构建Docker镜像约需1小时。 * **实验执行时长：** 约需36小时。 * **获取渠道：** BIN2CET已公开发布于***REMOVED***，本工件可在https://doi.org/10.5281/zenodo.20420793获取。 ## 安装 ### 1. 下载并解压工件 tar -jxvf bin2cet-artifacts.tar.bz2 ### 2. 准备数据集 使用工件附带的数据集： sudo tar --same-owner -xvf dataset.tar 或者创建目录以从零构建数据集： mkdir dataset && chmod 777 dataset ### 3. 准备Docker镜像 构建镜像： docker build --platform=linux/amd64 -t bin2cet_image . 亦可加载预构建镜像： bunzip2 bin2cet_image.tar.bz2 && docker load -i bin2cet_image.tar ### 4. 运行容器 docker run --rm -it --privileged --name bin2cet_$$ -v ./dataset:/work/dataset bin2cet_image ## 实验工作流 在Docker容器终端中，可使用以下任一命令： * `demo`：执行BIN2CET的分步演示。 * `sim`：运行脚本以模拟指定优化级别。 * `extract`：针对三个研究问题提取分析结果。 * `everything`：运行所有模拟并提取全部结果。