Exploring Semantic Stability Across Reviews in the Linux Kernel

• Lucas Ciziks USP
• Paulo Meirelles USP
• Marco Aurélio Gerosa NAU

DOI: https://doi.org/10.5753/vem.2026.33015

Resumo

Code review is credited with substantially changing a patch’s code between its first submission and the version that eventually lands. However, prior work typically studied only the final merged patch without comparing it to the first submission.We present a functionlevel measurement that tracks 10,117 trajectories (each function followed across the numbered revisions of one patch series) through the patch history of the Linux IIO subsystem, comparing similarity scores against unrelated function pairs as a baseline. A naive reading yields near-total similarity, but this is largely an artifact of composition: 75.3% of tracked trajectories are never textually modified between versions, contributing a trivial 100% similarity that inflates the headline. Restricting to the trajectories with a real edit, semantic purpose is still largely preserved (mean similarity 0.990 vs. a 0.909 baseline), but drift appears to concentrate in the first review round mainly because later rounds contain more functions that nobody touched, not because edits become more conservative over time. After controlling for it, a statistically detectable but small residual effect remains. This points to an open question: whether near-ceiling similarity reflects preserved purpose or a measurement tool that cannot detect the significance of small, localized edits. We present this work as a first look and outline next steps.

Palavras-chave: Code review, Code embeddings, Semantic similarity, Linux kernel

Referências

Luyao Bai, Kenan Alghythee, Hang Zhang, and Xiaoguang Wang. 2026. Beyond Crash-to-Patch: Patch Evolution for Linux Kernel Repair. arXiv preprint arXiv:2604.03851 (2026).

Edward E. Cureton. 1956. Rank-biserial correlation. Psychometrika 21, 3 (1956), 287–290. DOI: 10.1007/BF02289138

Kawin Ethayarajh. 2019. How Contextual are Contextualized Word Representations? Comparing the Geometry of BERT, ELMo, and GPT-2 Embeddings. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP). Association for Computational Linguistics, Hong Kong, China, 55–65. DOI: 10.18653/v1/D19-1006

Mohammad Farhad and Shuvalaxmi Dass. 2026. Residual Risk Analysis in Benign Code: How Far Are We? A Multi-Model Semantic and Structural Similarity Approach. arXiv preprint arXiv:2604.21051.

Zhangyin Feng, Daya Guo, Duyu Tang, Nan Duan, Xiaocheng Feng, Ming Gong, Linjun Shou, Bing Qin, Ting Liu, Daxin Jiang, and Ming Zhou. 2020. CodeBERT: A Pre-Trained Model for Programming and Natural Languages. In Findings of the Association for Computational Linguistics: EMNLP 2020. Association for Computational Linguistics, Online, 1536–1547. DOI: 10.18653/v1/2020.findings-emnlp.139

Daya Guo, Shuai Lu, Nan Duan, Yanlin Wang, Ming Zhou, and Jian Yin. 2022. UniXcoder: Unified Cross-Modal Pre-training for Code Representation. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). Association for Computational Linguistics, Dublin, Ireland, 7212–7225. DOI: 10.18653/v1/2022.acl-long.499

Thong Hoang, Hong Jin Kang, David Lo, and Julia Lawall. 2020. CC2Vec: Distributed Representations of Code Changes. In Proceedings of the ACM/IEEE 42nd International Conference on Software Engineering (ICSE ’20). ACM, New York, NY, USA, 518–529. DOI: 10.1145/3377811.3380361

Thong Hoang, Julia Lawall, Richard J. Oentaryo, Yuan Tian, and David Lo. 2019. PatchNet: A Tool for Deep Patch Classification. In 2019 IEEE/ACM 41st International Conference on Software Engineering: Companion Proceedings (ICSE-Companion). IEEE, 83–86. DOI: 10.1109/ICSE-Companion.2019.00044

Syful Islam and Stefano Zacchiroli. 2026. On the Use of Commit Messages for Corrective Software Maintenance: A Systematic Mapping Study. In Proceedings of the International Conference on Evaluation and Assessment in Software Engineering (EASE). Preprint: arXiv:2604.16404.

H. B. Mann and D. R. Whitney. 1947. On a Test of Whether one of Two Random Variables is Stochastically Larger than the Other. The Annals of Mathematical Statistics 18, 1 (1947), 50–60. DOI: 10.1214/aoms/1177730491

Huu Hung Nguyen, Duc Manh Tran, Yiran Cheng, Thanh Le-Cong, Hong Jin Kang, Ratnadira Widyasari, Shar Lwin Khin, Ouh Eng Lieh, Ting Zhang, and David Lo. 2025. Mapping NVD Records to Their Vulnerability-Fixing Commits: How Hard Is It? arXiv preprint arXiv:2506.09702 (2025). Preprint: arXiv:2506.09702.

Serge Lionel Nikiema, Albérick Euraste Djiré, Abdoul Aziz Bonkoungou, Micheline BénédicteMoumoula, Jordan Samhi, Abdoul Kader Kaboré, Jacques Klein, and Tegawendé F. Bissyandé. 2025. How Small Transformation Expose the Weakness of Semantic Similarity Measures. arXiv preprint arXiv:2509.09714 (2025).

Rafael Passos, Arthur Pilone, David Tadokoro, Laura Arakaki, and PauloMeirelles. 2026. LKML5Ws: The What, When, Who, Where, and Why in the Linux Kernel Mailing Lists. In 2026 IEEE International Conference on Software Maintenance and Evolution (ICSME).

Rafael Passos, Arthur Pilone, David Tadokoro, and Paulo Meirelles. 2025. Streamlining Analyses on the Linux Kernel with DUKS . In 2025 IEEEWorking Conference on Software Visualization (VISSOFT). IEEE Computer Society, Los Alamitos, CA, USA, 125–128. DOI: 10.1109/VISSOFT67405.2025.00025

Ralf Ramsauer, Daniel Lohmann, and Wolfgang Mauerer. 2019. The List is the Process: Reliable Pre-Integration Tracking of Commits on Mailing Lists. In Proceedings of the 41st International Conference on Software Engineering (ICSE). IEEE, 807–818. DOI: 10.1109/ICSE.2019.00088

Nils Reimers and Iryna Gurevych. 2019. Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks. In Proc. EMNLP 2019. DOI: 10.18653/v1/D19-1410

The Linux Kernel Organization. 2024. Submitting patches: the essential guide to getting your code into the kernel. [link]. Accessed July 2026.