Linus Torvalds posted his first LLM-assisted patch to the Linux kernel development branch, addressing a cache contention issue that occurs when multiple CPUs read the same file. The patch targets scenarios encountered in shared libraries, databases, and web servers. Torvalds stated he decided to "try to actually solve this thing using an LLM" after drafting an initial solution in early 2024 and using it on his own laptop. The move follows his clear pro-AI directives for Linux kernel development, where he has stated willingness to "absolutely put his foot down" on AI critics.
Torvalds Reports 23x Performance Improvement in Synthetic Benchmark
Torvalds posted the solution in a development branch and reported a 23x performance improvement in a dedicated synthetic benchmark. He noted that actual real-world impact will be far less significant, stating "no real load" will witness those figures. The patch addresses an issue that several developers had attempted to fix over the years, all with varied results and none providing an ideal all-encompassing solution.
Torvalds drafted his initial approach in early 2024 and has used it on his own laptop since that time. He recently decided to employ an LLM to finalize the solution.
Linux Kernel Drops Mandatory AI Tool Disclosure Requirement
The Linux kernel team is changing its policy to no longer require patch authors to disclose they used an AI tool when generating their code. Torvalds took an assertive stance on the matter, reiterating that users should treat LLMs as yet another tool, likening it to compilers or editors. He stated he expects "people do their best and understand what they are doing."
Last July, Torvalds stated he's "willing to absolutely put his foot down" on AI haters. He pointed out that an automated code review tool the kernel team uses outperforms humans and caught basic programming mistakes that shouldn't have been submitted in the first place.
AI Bug Hunters Drive Linux CVE Count to 2,000 Per Release
The Linux security mailing list became "unmanageable" earlier in the year due to AI agents recognizing common security flaw patterns. Linux has been fixing CVEs at a rate of around 2,000 per release. The team's hand is being forced into dropping older hardware drivers, as they're often the target of numerous security reports despite their negligible install base.
FAQ
What did Linus Torvalds post to the Linux kernel development branch?
Linus Torvalds posted his first LLM-assisted patch to the Linux kernel development branch, addressing a cache contention issue when multiple CPUs read the same file. He reported a 23x performance improvement in a synthetic benchmark, though he noted actual real-world impact will be far less significant.
Why did the Linux kernel team change its AI disclosure policy?
The Linux kernel team changed its policy to no longer require patch authors to disclose they used an AI tool when generating their code. Torvalds stated he expects "people do their best and understand what they are doing," treating LLMs as tools comparable to compilers or editors.
How has AI affected Linux kernel security reporting?
AI agents recognizing common security flaw patterns made the Linux security mailing list "unmanageable" earlier in the year. Linux has been fixing CVEs at a rate of around 2,000 per release, forcing the team to drop older hardware drivers that are often targets of numerous security reports.
