Back to search

Virginia Tech

RALLY: Retargetable Assembly Lifter to LLVM IR

Abstract

dc:description.abstract

Binary lifting to compiler-level intermediate representations enables powerful analysis, re-optimization, and cross-architecture porting for legacy software. However, traditional lifters operating directly on raw binaries suffer from irreversible information loss, forcing reliance on fragile heuristics that misclassify code, break position-independent code (PIC) semantics, and produce unsound control-flow graphs. In this paper, we present RALLY, a retargetable lifter that translates symbolized x86-64 NASM assembly into LLVM IR. By consuming symbolized assembly, such as the output of formally verified decompilation tools, RALLY circumvents the error-prone disassembly and code/data separation phases that plague binary-first approaches. Instead, RALLY treats assembly as a first-class input, using a deterministic grammar to structurally preserve PIC constructs and symbolic references from the outset. To bridge the semantic gap between assembly's implicit machine state and LLVM's explicit value semantics, RALLY introduces an integrated pipeline that recovers control flow, reconstructs ABI-compliant function signatures, and explicitly models architectural state. This approach produces semantically rich, optimizable LLVM IR without the heuristic compromises inherent in raw-binary lifting, enabling robust static analysis and retargetable compilation for modern position-independent software. To understand RALLY's effectiveness for recompilation and retargetability, we conducted experimental studies using NASM suite test cases, algorithmic C programs, microbenchmarks, and real-world applications targeting x86-64 and ARM Linux. Our evaluations reveal that RALLY achieves high semantic correctness, broad instruction-family coverage, and produces lifted IR that faithfully reproduces native execution behavior while successfully enabling cross-architecture recompilation.

Degree

thesis:*
Name thesis:degree_name
Master of Science
Level thesis:degree_level
masters
Discipline thesis:degree_discipline
Computer Engineering
Department dc:contributor.department
Electrical and Computer Engineering
Grantor dc:publisher
Virginia Tech
Year dc:date.issued
2026

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Goel, Kartik
Chair dc:contributor.committeechair
  • Ravindran, Binoy
Committee members dc:contributor.committeemember
  • Verbeek, Freek
  • Giles, Kendall Everett

Subjects

dc:subject × 5

Rights

dc:rights
Statement dc:rights
  • Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International
Language dc:language.iso
en

Identifiers

dc:identifier.*
Dc Identifier Other
vt_gsexam:47185

Chain of custody

source
Harvested from
Virginia Tech
Base URL
vtechworks.lib.vt.edu/oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Goel, Kartik. RALLY: Retargetable Assembly Lifter to LLVM IR. masters thesis, Virginia Tech, 2026. https://hdl.handle.net/10919/143477