Virtualization can increase CPU utilization by allowing a large number of virtual CPUs to share a smaller number of physical CPUs. The amount of CPU time that is actually available does not change, though, so heavy activity on too many virtual CPUs can lead to contention and significant performance loss. The steal governor patch series from Shrikanth Hegde is an attempt to address that problem with a mechanism that allows virtual machines to voluntarily reduce the number of virtual CPUs they use when contention is high.

For people who would like to experiment with the multi-kernel Linux concept, Cong Wang has announced the release of mklinux v7.0-mk2.

mklinux lets one machine run several independent Linux kernels at the same time on bare metal, without a hypervisor. A host kernel owns a pool of CPUs, memory and PCI devices, carves that pool into instances, and boots a spawn kernel into each instance through kexec_file_load(). Every spawn kernel runs natively on its own CPUs, its own physical memory and its own devices. Nothing is emulated and nothing is trapped; the only thing shared is what you choose to share.

Note that this is not the old MkLinux, which was a port to PowerPC Macintosh systems.

The 7.1.10, 6.18.46, 6.12.105, 6.6.153, 6.1.184, 5.15.217, and 5.10.266 stable kernels have all been released; each contains another set of important fixes.

From Jeremy Allison we have the sad news of the passing of Steve French. He was the maintainer of the kernel's SMB filesystem code for many years, having only dropped that role due to health issues in the last week. "I've known Steve for over 20 years. He was a legend in the community, and a really good friend. He will be greatly missed. Farewell Steve." He will indeed be missed.

The open-source world has been struggling for a few years now to understand how to approach large language models (LLMs) and the licensing applied to them. What constitutes "freedom" with respect to a black box filled with numerical weights? The process taken by the Open Source Initiative (OSI) in the development of its Open Source AI Definition was controversial at best, as was its output. Now, the Linux Foundation's Mike Dolan has brought a new license to the OSI for approval. It is called the OpenMDW ("Open Model, Data, and Weights"), and it aims to clarify licensing for the distribution of LLMs and related materials, but consensus is proving hard to find for this license as well.

The Rust blog reports on a malicious crate, called proc-macro1, that was uploaded to the crates.io repository.

Furthermore, we discovered that the popular arrayref crate had recently been republished and made to depend on this crate, with the most recent versions yanked. We have removed the malicious version and unyanked the maliciously-yanked versions. Other crates by that author (internment, append-only-vec) were also affected so we have done the same for those, and locked the account as a precaution. We do not believe the author of arrayref to be acting maliciously, but their computer or credentials are likely compromised, and we are attempting to contact them.

As of this writing, 2,346 non-merge changesets have been pulled into the mainline repository for the 7.3 kernel release. That, clearly, is a mere down payment on the flood that is to come. Even so, those early pulls brought in some noteworthy changes, including (but not limited to) a significant reworking of how group scheduling works on multiprocessor systems.