Sashiko: automatic Linux kernel patch review with a language model

According to the machine translation of the LWN feed, patch review has long been one of the constraints limiting the Linux kernel project, as in most others, because there are not enough people to properly review all code submitted for inclusion. Sashiko is a system that uses a large language model (LLM) to generate patch reviews automatically, offering some relief for this bottleneck. According to the text, it has already become an important part of the kernel development process. At the 2026 edition of Kernel Recipes, Roman Gushchin, Sashiko's maintainer, gave an overview of how the system works and what is being done to improve it. To check the original, use the source link given in the LWN feed item, which carries the full article, and confirm the publication date before citing details. The item describes AI use in open source software engineering; it does not concern personal data and does not report accuracy metrics, which should be sought in the project's official documentation.