Con Kolivas has warned of his intention to cease development of his projects for the Linux kernel, aimed at enhancing the responsiveness and interactivity of user tasks. This includes stopping the development of the MuQSS (Multiple Queue Skiplist Scheduler, previously known as BFS) task scheduler and halting the adaptation of the '-ck' patch set for new kernel releases.
The reason mentioned is a loss of interest in Linux kernel development after 20 years of such activity and an inability to regain former motivation after returning to medical work during the Covid19 pandemic (by education, Con is an anesthesiologist and during the pandemic he led a project for developing new designs of artificial ventilation devices and using 3D printing to create accompanying parts).
Notably, in 2007, Con Kolivas already halted the development of the '-ck' patches due to the inability to promote his fixes to the mainline Linux kernel, but later returned to their development. If Con Kolivas is unable to find motivation to continue this time, the release of the 5.12-ck1 patches will be the last.
The '-ck' patches, in addition to the MuQSS scheduler, which continues the BFS project, include various changes affecting memory management, priority handling, timer interrupt generation, and kernel tuning. The key goal of the patches is to improve the responsiveness of desktop applications. Since the proposed changes may negatively impact the performance of server systems, computers with a large number of CPU cores, and operation in conditions of high concurrency, many of Con Kolivas's changes were not accepted into the mainline kernel, forcing him to maintain them as a separate patch set adapted to each new kernel release.
The last update of the '-ck' branch was an adaptation for the 5.12 kernel release. The release of the '-ck' patches for the 5.13 kernel was skipped, and following the release of the 5.14 kernel, an intention to cease porting for new kernel versions was announced. The Liquorix and Xanmod projects may possibly take over the maintenance of the patches, as they already utilize developments from the '-ck' set in their versions of the Linux kernel.
Con Kolivas is ready to hand over the maintenance of patches to others, but he doesn't believe this will be a successful solution, as all previous attempts at creating forks have resulted in problems he tried to avoid. For users who wish to maximize the benefits of using the main Linux kernel without porting the MuQSS scheduler to it, Con Kolivas believes the simplest and most effective approach would be to transfer patches to increase the timer interrupt frequency (HZ) to 1000 Hz.
Source: opennet.ru
