Compilomorphic Fuzzing: Turning a Compiler Against Itself
TL;DR
Compilomorphic fuzzing is introduced, a new validation approach that turns a compiler into its own test oracle via structure-preserving transformations over programs and structured inputs, establishing a practical, model-free foundation for validating compilers in critical DSL domains.
Abstract
Compilers for structured data pipelines, often specified using domain-specific languages such as P4, are notoriously difficult to validate. We introduce compilomorphic fuzzing, a new validation approach that turns a compiler into its own test oracle via structure-preserving transformations over programs and structured inputs. The core idea is to generate pairs of semantically equivalent but structurally distinct inputs (e.g., packets) and programs (e.g., pipeline specifications), and then assert that the compiler's observable behaviour remains invariant. Any deviation exposes a deep semantic bug. We present the formal basis for this approach, compilomorphism, and a prototype fuzzing framework that operationalises it, establishing a practical, model-free foundation for validating compilers in critical DSL domains.