Category: os

On Sunday’s release of Linux 6.0 release candidate version 1 (rc-1), he explained his reasoning behind choosing a new major version number and its purpose for developers. Again, it’s about avoiding confusion rather than signaling that the release has major new features. His threshold for changing the lead version number was .20 because it is difficult to remember incremental version numbers beyond that. “Despite the major number change, there’s nothing fundamentally different about this release – I’ve long eschewed the notion that major numbers are meaningful, and the only reason for a ‘hierarchical; numbering system is to make the numbers easier to remember and distinguish,” said Torvalds. Torvalds lamented some Rust-enabling code didn’t make it into the release. The Register adds: “I actually was hoping that we’d get some of the first rust infrastructure, and the multi-gen LRU VM, but neither of them happened this time around,” he mused, before observing “There’s always more releases. This is one of those releases where you should not look at the diffstat too closely, because more than half of it is yet another AMD GPU register dump,” he added, noting that Intel’s Gaudi2 Ai processors are also likely to produce plenty of similar kernel additions. “The CPU people also show up in the JSON files that describe the perf events, but they look absolutely tiny compared to the ‘asic_reg’ auto-generated GPU and AI hardware definitions,” he added.

NetBSD supports eight “tier I” architectures: 32-bit and 64-bit x86 and Arm, plus MIPS, PowerPC, Sun UltraSPARC, and the Xen hypervisor. Alongside those, there are no less than 49 “tier II” supported architectures, which are not as complete and not everything works — although almost all of them are on version 9.3 except for the version for original Acorn computers with 32-bit Arm CPUs, which is still only on NetBSD 8.1. There’s also a “tier III” for ports which are on “life support” so there may be a risk Archimedes support could drop to that. This is an OS that can run on 680×0 hardware, DEC VAX minicomputers and workstations, and Sun 2, 3, and 32-bit SPARC boxes. In other words, it reaches back as far as some 1970s hardware. Let this govern your expectations. For instance, in VirtualBox, if you tell it you want to create a NetBSD guest, it disables SMP support.

Development work on this latest Linux-on-iDevices effort is still in its early days. The photos that the developers shared both show a basic boot process that fails because it can’t mount a filesystem, and Dybcio notes that basic things like USB and Bluetooth support aren’t working. Getting networking, audio, and graphics acceleration all working properly will also be a tall order. But being able to boot Linux at all could draw the attention of other developers who want to help the project.

Compared to modern hardware with an Apple M1 chip, A7 and A8-powered devices wouldn’t be great as general-purpose Linux machines. While impressive at the time, their CPUs and GPUs are considerably slower than modern Apple devices, and they all shipped with either 1GB or 2GB of RAM. But their performance still stacks up well next to the slow processors in devices like the Raspberry Pi 4, and most (though not all) A7 and A8 hardware has stopped getting new iOS and iPadOS updates from Apple at this point; Linux support could give some of these devices a second life as retro game consoles, simple home servers, or other things that low-power Arm hardware is good for. Further reading: Linux For M1 Macs? First Alpha Release Announced for Asahi Linux

The daemon will work with the kernel’s CFS scheduler to give priority to components that System76 deems important for its distro. Out of the box, the scheduler will assign priority to the X.Org Server and desktop window managers/compositors, while pushing compilers and other background tasks lower. However, the scheduler will be configurable via Rusty Object Notation (RON) files found in /etc/system76-scheduler/assignments/ and /usr/share/system76-scheduler/assignments/.

Over on the GitHub page for the project, the team indicates that they are indeed making a trade-off from the default CFS to benefit Desktop configurations over the typical load a server might see.