FreeBSD Adding More AMD64 SIMD, Continues Improving Linux Compatibility Layer

[phoronix]

Phoronix: FreeBSD Adding More AMD64 SIMD, Continues Improving Linux Compatibility Layer

FreeBSD developers have been busy preparing for the release of FreeBSD 14 as well as making a variety of enhancements to this leading BSD operating system...

https://www.phoronix.com/news/FreeBSD-In-Q3-2023

[chromer]

The SquashFS driver for the FreeBSD kernel has been finished.

This is a great feature, now the image files can be compressed and custom ISO files is easier to handle.

[darkoverlordofdata]

Improvements to Linux compatibility would be very welcome. One of the biggest challenges while running FreeBSD as a desktop is finding desktop software that works. Wine seems more successful at this than the ‘Linuxulator’.

It’s a shame that unix-like does not translate to unix compatible. I often wonder if Linux chose incompatibility with Unix for the same reasons Microsoft did - Market Lock.

[Dharc]

I often wonder if Linux chose incompatibility with Unix for the same reasons Microsoft did - Market Lock.

You have any doubts? But it is normal different OS having their own idiosyncrasies. I believe BSDs should follow linux LTS compatibility only.

[skeevy420]

Pot is a jail management tool that also supports orchestration through Nomad.

Pot is a jail management tool in a lot of the world.

I doubt that Linux or FreeBSD or whatever all intentionally tried to make themselves incompatible with each other when it's all the more likely that they started in an era where there wasn't mass communications or lots of shared open source drivers so they all did their own thing in ways that made sense to them. From what I can tell, lately they've been trying to make themselves as compatible as possible by sharing drivers and code where possible and when licensing allows. They're both open source operating systems that run a lot of the same open source software so it's in their best interests to share and work together. They're not corporate operating systems where a hostile walled garden approach can be used to lock consumers in to maximize profits.

[sophisticles]

The FreeBSD Foundation was helped along by significant donations from NetApp, Netflix, and ARM. So far their fundraising total for the year is $375,000 USD out of their $2.23M budget.

FreeBSD developers have been working to fill gaps within the FreeBSD desktop experience. This includes installation improvements and various new ports to enhance the desktop usability

FreeBSD's Linux compatibility layer for running Linux binaries has added support for additional system calls. It's now possible to use Linux rsync and debootstrap Ubuntu 23.04 on FreeBSD.

The question that pops into my mind is if there is a industry wide movement to replace Linux use with FreeBSD use, due to the silly GPL restrictions that plague Linux based OSes.

Think about it, there was a short-lived project called Ubuntu BSD, the FreeBSD Foundation is raising significant amount of cash from companies that use Linux and they are trying to achieve binary compatibility with Linux.

All this tells me that there is substantial interest in an open source alternative that is not hampered by the GPL.

Who knows, maybe in 10 years Linux will be a footnote computing history.

[ClosedSource]

It’s a shame that unix-like does not translate to unix compatible. I often wonder if Linux chose incompatibility with Unix for the same reasons Microsoft did - Market Lock.

Linux is a free kernel. It never had the chance to choose anything. It's a collaborative project where companies push their code (which may be well tested but can be badly tested too) into Linux. They share maintenance of different parts of the kernel. These companies are only interested in their own products and not in UNIX compatibility. There is no basement hermit contributing to the Linux kernel. It's all very professional developers. It's much cheaper to contribute to a few parts of a kernel than write your own and so greed (otherwise known as cutting costs) is part of it. This whole ecosystem is somewhat unfair at times but people are doing their parts and at least it lets us still have Linux so it's not bad .

[dragon321]

I often wonder if Linux chose incompatibility with Unix for the same reasons Microsoft did - Market Lock.

What do you mean by "incompatibility with Unix"? First of all Linux was never intended to be "real Unix". It's started as free and hobby OS inspired by Unix. Second - different Unix operating systems were not compatible with each other. Sure there was POSIX so you could write app that would compile and run on AIX, SunOS, HP-UX etc. but aside from that they weren't compatible with each other. AIX wouldn't run SunOS binaries (and vice versa), kernels were also completely different often providing different features. Even source compatibility wasn't perfect because standards implementation wasn't always perfect in different Unix operating systems.

So basically Linux couldn't choose "compatibility with Unix" because there is no such thing. As for the limited source compatibility, Linux is pretty good at it as well - most Linux software can be easily compiled and work in FreeBSD/OpenBSD/NetBSD/etc.

[Kjell]

The SquashFS driver for the FreeBSD kernel has been finished

Perfect
Let's add OSTree

[Dharc]

Think about it, there was a short-lived project called Ubuntu BSD, the FreeBSD Foundation is raising significant amount of cash from companies that use Linux and they are trying to achieve binary compatibility with Linux.

All this tells me that there is substantial interest in an open source alternative that is not hampered by the GPL.

Yes, there are interests in non-GPL OS. But the ones available are full OS. Linux is cheaper to build a project on top of it because it is just a kernel. Creating a customized Linux distro for servers, desktop or embedded is easier than customizing a whole OS. From BSD side I believe NetBSD is the only one worth it. Google is creating the non-GPL Fuschia OS. GPL is outdated nowadays.