一句话：提出内核辅助的共享库细粒度去膨胀与布局随机化方法，在复用前提下缓解代码膨胀。
代码复用与“一刀切”软件开发实践导致共享库等组件严重代码膨胀。现有去膨胀方案多依赖删除冗余代码，但无论离线还是运行时，这种思路都与可复用组件的设计哲学相冲突：同一库可能被不同程序以不同方式使用。为此，论文提出内核辅助的共享库细粒度去膨胀与布局随机化方法。其核心是利用内核支持，在不破坏共享库可复用性的前提下，对共享库进行细粒度代码裁剪，并辅以布局随机化，以降低冗余代码带来的攻击面与安全风险。摘要未报告具体实验设置与量化结果，因此其实际去膨胀效果、随机化强度及性能开销尚无法从摘要判断。
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.
@inproceedings{paperbot11529,
title = {One Size Does Fit All: Kernel-Assisted Fine-Grained Debloating and Layout Randomization for Shared Libraries},
author = {Ziyi Lin and Haoyu Ma and Kexin Liu and Bowen Li and Jinku Li and Jianfeng Ma},
booktitle = {ICSE},
year = {2026},
doi = {10.1145/3744916.3773152}
}