Outrunning LLM Cutoffs: A Live Kernel Crash Resolution Benchmark for All
Chenxi Huang, Alex Mathai, Feiyang Yu, Aleksandr Nogikh, Petros Maniatis, Franjo Ivančić, Eugene Wu, Kostis Kaffes, Junfeng Yang, Baishakhi Ray
TL;DR
This work tackles the challenge of evaluating and improving Linux kernel crash repair by LLM-based agents in a rapidly evolving kernel. It introduces kEnv, an agent-agnostic crash-resolution environment, and Live-kBench, a self-evolving benchmark that continuously ingests fresh kernel bugs from Syzbot and evaluates agents under identical conditions. The inaugural dataset Live-kBench-2512 contains 534 bugs and reveals that agents fix crashes on the first attempt for about 74% of cases, but only around 20% patches are equivalent to developer fixes; pre-cutoff fixes yield up to ~25% higher equivalent patch rates. Crash-resolution feedback (CRF) improves crash-resolution rate by about 29%, and the framework provides a public dashboard to study distribution shift, data contamination, and cross-agent performance, enabling time-aware, reproducible kernel crash-resolution research.
Abstract
Repairing system crashes discovered by kernel fuzzers like Syzkaller is a critical yet underexplored challenge in software engineering. While recent works have introduced Large Language Model (LLM) based agents for Linux kernel crash-resolution, their evaluation benchmarks are usually static and thus, do not capture the evolving nature of the Linux kernel, and suffer from potential data contamination due to LLM knowledge cutoffs. To address the above problem, we present (i) Live-kBench, an evaluation framework for self-evolving benchmarks that continuously scrapes and evaluates agents on freshly discovered kernel bugs, and (ii) kEnv, an agent-agnostic standardized crash-resolution environment for kernel compilation, execution, and feedback. This design decouples agent workflows from heavy-weight execution, enabling fair and scalable comparison across diverse agent frameworks under identical conditions. To this end, we curate an inaugural dataset of 534 Linux kernel bugs and empirically demonstrate a significant performance gap, with agents achieving up to 25% higher equivalent patch rate on bugs fixed before the LLM knowledge cutoff. Using kEnv, we benchmark three state-of-the-art agents, showing that they resolve 74% of crashes on the first attempt (plausible patches); however only ~20% of generated patches closely match developer fixes. Additionally, exposing crash resolution feedback improves crash resolution rate by 29%. Live-kBench provides the community with an evaluation infrastructure for self-evolving benchmarks that is both time and attribute sensitive; complete with a public dashboard to track agent progress on Linux kernel bugs.