HackerPapers

arXiv:2212.05614cs.CR

Generic Tagging for RISC-V Binaries

David Demicco, Matthew Cole, Gokturk Yuksek, Ravi Theja Gollapudi, Aravind Prakash, Kanad Ghose, Zerksis Umrigar

HN submission. Found in a submission link or body. Submitters do not count as commenters in the main ranking; a submission is not an endorsement.

Abstract

With the widespread popularity of RISC-V -- an open-source ISA -- custom hardware security solutions targeting specific defense needs are gaining popularity. These solutions often require specialized compilers that can insert metadata (called tags) into the generated binaries, and/or extend the RISC-V ISA with new instructions. Developing such compilers can be a tedious and time-consuming process. In this paper, we present COGENT, a generic instruction tag generator for RISC-V architecture. COGENT is capable of associating a tag of configurable and varying widths (1 to 20 bits) to each instruction. It is also capable of emitting labels that are central to the implementation of control-flow integrity (CFI) solutions. COGENT encodes all tags and labels as nop instructions thereby providing full backward compatibility. We evaluate COGENT on a subset of programs from the SPEC CPU2017 benchmark suite and report the binary size increase to be 29.3% and 18.27% for the lowest and highest tag coverage levels respectively. Additionally, we executed tagged programs on COTS RISC-V unmodified hardware and found the execution time overhead (with respect to backward compatibility) to be 13.4% and 5.72% for the lowest and highest coverage levels respectively. Finally, using a case study, we present possible use case scenarios where COGENT can be applied.

arXiv comments
Journal reference
DOI
Subjects
cs.CRCryptography and Security cs.ARHardware Architecture

On Hacker News

No comments linking this paper were found in the collected sources. Found in 1 HN submission from Dec 2022 to Dec 2022. Open submission discussions.

Comments linking this paper, per year.

Discussed in comments0 threads

No comments linking this paper were found in the collected sources. Open submission discussions.

Submitted to Hacker News

Connected through shared HN threads, not inferred similarity or a recommendation.