Phoronix’s Q2 2026 kernel recap is a solid primary hub for both threads of that story:
Source: Linux Kernel Q2 2026 recap — Phoronix
i486 support is not “retro computing’s funeral”
Removing i486-class support from mainline mostly acknowledges reality:
- No major distro still targets 486 as a first-class citizen
- Maintainer cost grows when ancient paths interact with modern security and build tooling
- Hobby 486 boxes already live on older LTS kernels, custom trees, or non-Linux systems
The “potato” scheduling work is the other half of the plot
There is ongoing scheduler work aimed at aging hardware that is still in actual use — think early 2010s desktops that people still game on. That is a different constituency from museum 486s: machines that are slow by 2026 standards but still have multi-core CPUs, GPUs, and users.
Mosfetti take for old PC keepers
- Separate preservation machines (frozen OS images) from daily drivers (updated kernels)
- For very old x86, document the last known-good kernel and config
- For “potato but playable” boxes, track scheduler and desktop responsiveness patches — they can matter more than another 5 FPS fantasy
Where do you draw the line between “keep it on modern Linux” and “freeze a classic install forever” for older PCs?