Link-time compaction and optimization of ARM executables

Abstract

The overhead in terms of code size, power consumption, and execution time caused by the use of precompiled libraries and separate compilation is often unacceptable in the embedded world, where real-time constraints, battery life-time, and production costs are of critical importance. In this paper, we present our link-time optimizer for the ARM architecture. We discuss how we can deal with the peculiarities of the ARM architecture related to its visible program counter and how the introduced overhead can to a large extent be eliminated. Our link-time optimizer is evaluated with four tool chains, two proprietary ones from ARM and two open ones based on GNU GCC. When used with proprietary tool chains from ARM Ltd., our link-time optimizer achieved average code size reductions of 16.0 and 18.5%, while the programs have become 12.8 and 12.3% faster, and 10.7 to 10.1% more energy efficient. Finally, we show how the incorporation of link-time optimization in tool chains may influence library interface design.

Link-time smart card code hardening

“…Regarding the wide range of transformations that can be applied at link-time, as necessary to insert the necessary redundancy, we point to the existing applications of the open-source link-time binary rewriting framework Diablo 1 [48]. These applications include optimization and compaction [19], obfuscation [4,34] and deobfuscation [32], anti-tampering [51], formal verification of binary code [54], instrumentation [16], GUI binary code editing [49], and operating system customization [12].…”

“…C, C++, assembly and Fortran; binary code generated with different compiler generations covering more than a decade of proprietary as well as open-source compilers (incl. ARM ADS, ARM RVCT, and gcc) [19]; system libraries that contain significant 1 http://diablo.elis.ugent.be amounts of manually-written assembly, incl. newlibc 2 , glibc 3 , and uClibc 4 [19]; a range of architectures, incl.…”

“…ARM ADS, ARM RVCT, and gcc) [19]; system libraries that contain significant 1 http://diablo.elis.ugent.be amounts of manually-written assembly, incl. newlibc 2 , glibc 3 , and uClibc 4 [19]; a range of architectures, incl. SH [14], PowerPC [10], MIPS [33], IA64 [5], x86 [12], and ARM incl.…”

“…SH [14], PowerPC [10], MIPS [33], IA64 [5], x86 [12], and ARM incl. Thumb [12,19]; the Linux kernel [12,11], which features artifacts such as mixed code for physical and virtual address spaces, privileged instructions, manually written assembly not adhering to the conventions as specified in application binary interfaces (ABIs), and a complicated, non-standard build process.…”

Towards Tamper Resistant Code Encryption: Practice and Experience

“…Subsequently, we apply our on-demand encryption scheme where we protect a maximal number of functions, such that our scheme can offer tamper-resistance according to the properties mentioned in Section 4. To insert the guard code we used Diablo [7], a link-time binary rewriter, allowing us to patch binary code, insert extra encryption functionality, and perform dominator analysis on either the control flow graph or the function call graph. As we are generating self-modifying code, we mark all code segments to be readable and writable.…”