Linux creator Linus Torvalds is known for using AI to help with kernel debugging, and outside of the kernel tree he has also done some flavor coding with Google’s Antigravity. He has previously stated that he sees AI as a useful tool. Turns out he’s still using Claude Code to work on new Linux kernel features.
While waiting for the release of Linux kernel 7.3-rc7 and updating its torvalds/linux.git repository, I was surprised to see a new kernel branch appear. In general, Torvalds doesn’t do much work on branches of its main repository and usually just hosts the master bifurcate.
The new branch to appear is noref-filemap-read-with-llm and as the suffix “LLM” implies, there are new kernel improvements that he has been working on with the help of AI/LLM. Claude Code in particular. The branched code works to adapt the memory management code to read from the page cache without folio references.
The code aims to solve a Linux problem of small reads of a file from many threads conflicting over the folio’s reference counting. So the purpose of the branch/patch series is to read from the page cache without folio references.
For those who are passionate about AI, love it or hate it, in the main message of the patch, Linus Torvalds confirmed his use of AI in the creation of this job and associated testing and even the initial version of the cover letter:
“The code, test programs, review behind the list above and the first draft of this text were done with an LLM coding assistant (Claude Code) under my direction. The measurements were performed on my machine.”
And for the impact on this situation of solving the bottleneck of small reads of a file from multiple threads, the impact of the series is significant:
Quite a difference. The fixes are currently on Torvalds’ noref-filemap-read-with-llm Git branch. We will see if it merges its branch for the next Linux 7.4 merge window.
Gn tech
