linux VS linux

Linux is a family of free and open source operating systems based on the Linux kernel.

You mean apart from 6.6 being the current latest longterm kernel?

https://kernel.org/

I don't like that, it's not good practice.

One should give links to original sources, i.e. https://kernel.org as far as Linux is concerned.

Even if git guarantees that the content is the same (if someone bothers to verify that the SHA-1 is the same and we exclude the possibility of a SHA-1 collision in git, which is yet to be demonstrated).

kernel.org existed before github.

- Modern Operating Systems, 5th Edition by Andrew Tanenbaum (of MINIX fame) and Herbert Bos (https://www.pearson.com/en-us/subject-catalog/p/modern-opera...) is the latest edition of a solid graduate-level textbook on operating system concepts.

It may also be beneficial studying the source code of existing operating systems. I recommend starting with smaller, simpler systems, such as MINIX and xv6 (https://github.com/mit-pdos/xv6-public), before moving on to larger, more complex systems such as the Linux kernel (https://kernel.org/) and its userland (e.g., GNU utilities, systemd, etc.).

Another cool thing is to study the designs of non-Unix operating systems, such as the classic Mac OS, VMS, IBM OS/400, Plan 9 (yes, this is "more Unix than Unix" in many ways, but it's quite a departure from Unix) and its successor Inferno, and Symbolics Genera. Bonus points for reading academic papers on OS concepts such as exokernels.

Good luck! It's a long but very interesting journey!

Those other flashy distros like mint and ubuntus are designed with rich people with very fresh machines in mind, they don't care if you have an AMDx4 or core2duo or even 32bit older machine. Even Mint and ubuntu people will tell you, if you have an old machine with little ram, use antiX. It still works very well with machines not even released yet, buy one in May 2024 and I "guaranty you" antiX will run fine. That's because kernels, even 6.1 have code for machines not yet released in the market, this is where manufacturers of new hardware send their prototypes to be tested, to kernel.org Some are even industrial machines we will never see in the market.

The memory_order_relaxed explanation on the kernel.org documentation heavily implies (never explicitly) that the direct memory load is implicit in the barrier(so by preventing it's reordering we are also forcing a LOAD from main), and that THIS specific barrier (relaxed) is what we NEED for these type of scenarios, so I am not entirely sure if a loadLoadFence() would prevent the hoisting... maybe it will prevent the reordering but not the hoisting/caching... in which case we would still have a visibility issue.

Are all versions of the kernel from kernel.org called mainline kernels or only 6.6-rc4 as shown in the picture?

Devuan is a fork of Debian GNU+Linux without systemd.

I reached to Arch's kernel repository and found that device.h exists.

bash $ pacman -Q --info linux-headers Name : linux-headers Version : 6.2.11.arch1-1 Description : Headers and scripts for building modules for the Linux kernel Architecture : x86_64 URL : https://github.com/archlinux/linux/commits/v6.2.11-arch1 Licenses : GPL2 Groups : None Provides : None Depends On : pahole Optional Deps : None Required By : None Optional For : dkms Conflicts With : None Replaces : None Installed Size : 158.24 MiB Packager : Jan Alexander Steffens (heftig) Build Date : Thu 13 Apr 2023 06:59:24 PM CEST Install Date : Wed 19 Apr 2023 09:32:40 AM CEST Install Reason : Explicitly installed Install Script : No Validated By : Signature

So, the "bleeding edge" of https://github.com/archlinux/linux? That's going to just get you a release candidate of the mainline kernel, with no Arch-specific patches. You can get that on NixOS by using linuxPackages_testing.

Thanks for the suggestion. I'm guessing I have to do it with the Arch Linux fork of the linux kernel?

I've now downloaded v6.0.2-arch1 and confirmed the patch isn't applied. Idk why but will check on archs forums.

I've confirmed that in v6.0.2-arch1 the patch isn't applied. But it is on the master branch.

$ pacman -Qii linux Name : linux Version : 6.0.arch1-1 Description : The Linux kernel and modules Architecture : x86_64 URL : https://github.com/archlinux/linux/commits/v6.0-arch1 Licenses : GPL2 Groups : None Provides : VIRTUALBOX-GUEST-MODULES WIREGUARD-MODULE KSMBD-MODULE Depends On : coreutils kmod initramfs Optional Deps : wireless-regdb: to set the correct wireless channels of your country linux-firmware: firmware images needed for some devices [installed] Required By : nvidia Optional For : base Conflicts With : None Replaces : virtualbox-guest-modules-arch wireguard-arch Installed Size : 163.52 MiB Packager : Jan Alexander Steffens (heftig) Build Date : Tue 04 Oct 2022 03:16:52 PM EDT Install Date : Wed 05 Oct 2022 04:35:19 AM EDT Install Reason : Explicitly installed Install Script : No Validated By : Signature Backup Files : (none)

Lately there appear to be changes in Arch as well, making it hard to keep up as well. Let's examine for example the kernel building. It was the tradition and commitment by arch to use direct source from upstream, and only modify what is necessary to make software comply with the entire distro. Now, with linux-5.19 appearing, and maybe with the last versions of 5.18, instead of using directly a tar ball from kernel.org Arch is using its own source library https://github.com/archlinux/linux/tags made by forking Torvalds' git. The traditional config file is gone, and all configuration and patching takes place at Arch's forked source.

It's not present in 5.15/16 and i noticed that lockdep_assert_preemption_disabled was added in the 5.17. As arch's kernel is by default PREEMPT_DYNAMIC, tried preempt=voluntary and the warnings went away.