Ioannis Angelakopoulos (Boston University), Gianluca Stringhini (Boston University), Manuel Egele (Boston University)

Re-hosting Internet of Things (IoT) firmware can oftentimes be a tedious process, especially when analysts have to intervene with the analysis to ensure further progress. When it comes to Linux-based firmware, one crucial problem that current re-hosting systems face, is that the configuration of the custom kernels used by these systems, significantly deviates from the configuration of the IoT kernel modules used in firmware images. As a consequence, kernel artifacts, such as the memory layout of data structures might differ between the custom kernels and the IoT kernel modules. To analyze the IoT kernel modules within these kernels, the analyst often has to invest significant amount of engineering effort and time to align the offending data structures within the custom kernels. In this paper, we present FirmDiff, an automated binary diffing framework that enables analysts to effectively detect and align the misaligned data structures between the custom kernels produced by the FirmSolo re-hosting framework and the Linux kernel modules in IoT firmware. The goal of FirmDiff is to improve the configuration of FirmSolo’s kernels to closely approximate the configuration of the IoT kernels in the firmware images, such that the IoT kernel modules can be analyzed without errors. We evaluate FirmDiff on a dataset of 10 firmware images with 148 IoT kernel modules that crash during re-hosting with FirmSolo. Using FirmDiff’s findings, we identify 37 misaligned data structures in FirmSolo’s kernels for these images. After aligning the layout of 35 of these data structures, FirmSolo’s refined kernels successfully load 28 previously crashing kernel modules.

View More Papers

From Hardware Fingerprint to Access Token: Enhancing the Authentication...

Yue Xiao (Wuhan University), Yi He (Tsinghua University), Xiaoli Zhang (Zhejiang University of Technology), Qian Wang (Wuhan University), Renjie Xie (Tsinghua University), Kun Sun (George Mason University), Ke Xu (Tsinghua University), Qi Li (Tsinghua University)

Read More

ORL-AUDITOR: Dataset Auditing in Offline Deep Reinforcement Learning

Linkang Du (Zhejiang University), Min Chen (CISPA Helmholtz Center for Information Security), Mingyang Sun (Zhejiang University), Shouling Ji (Zhejiang University), Peng Cheng (Zhejiang University), Jiming Chen (Zhejiang University), Zhikun Zhang (CISPA Helmholtz Center for Information Security and Stanford University)

Read More

Towards Better CFG Layouts

Jack Royer (CentraleSupélec), Frédéric TRONEL (CentraleSupélec, Inria, CNRS, University of Rennes), Yaëlle Vinçont (Univ Rennes, Inria, CNRS, IRISA)

Read More

Efficient and Timely Revocation of V2X Credentials

Gianluca Scopelliti (Ericsson & KU Leuven), Christoph Baumann (Ericsson), Fritz Alder (KU Leuven), Eddy Truyen (KU Leuven), Jan Tobias Mühlberg (Université libre de Bruxelles & KU Leuven)

Read More