One Size Does Fit All: Kernel-Assisted Fine-Grained Debloating and Layout Randomization for Shared Libraries

Prevalent software engineering practices, such as code reuse and the “one-size-fits-all” methodology, have significantly contributed to code bloat. Code debloating, which targets the reduction of extraneous code, emerges as a promising solution to combat this issue. However, current methods that depend on redundant code deletion are at odds with the foundational design philosophy of reusable code components, regardless of whether they are applied offline or at runtime.