TITLE:
Program Counter Encoding for ARM® Architecture
AUTHORS:
Seho Park, Yongsuk Lee, Gyungho Lee
KEYWORDS:
Compiler, Control-Flow Integrity, PC-Encoding, Software Security, ARM®
JOURNAL NAME:
Journal of Information Security,
Vol.8 No.1,
January
10,
2017
ABSTRACT: ARM® is the prevalent processor architecture for embedded and mobile applications. For the smartphones, it is the processor for which software applications are running, whether the platform is with Apple’s iOS or Google’s Android. Software operations under these platforms are prone to semantic gap, which refers to potential difference between intended operations described in software and actual operations done by processor. Attacks that compromise program control flows, which result in these mantic gaps, are a major attack type in modern software attacks. Many recent software protection schemes for servers and desktops focus on protecting program control flows, but there are little protection tools available for protecting program control flows of mobile applications for ARM processor architecture. This paper uses a program counter (PC) encoding technique (PC-Encoding) to harden program control flows under ARM processor architecture. The PC-Encoding directly encodes control flow target addresses that will load into the PC. It is simple and intuitive to implement and incur little overhead. Encoding the control flow target addresses can minimize the semantic gap by preventing potential compromises of the control flows. This paper describes our efforts of implementing PC-Encoding to harden portable binaries in ELF (Executable and Linkable Format).