What would you like to see over the next few FreeBSD versions?

That is not in doubt. But nothing in the computer world is so good that it cannot be improved further. I guess that otherwise we would still monkey around on 32kB CP/M boxes. ;)
Indeed.
Something can be improved and also over improved in terms that stabilization would be rather hard to achieve.
In terms of improvements I still prefer *BSDs approach more than linux approach. :P
 
Indeed.
Something can be improved and also over improved in terms that stabilization would be rather hard to achieve.
In terms of improvements I still prefer *BSDs approach more than linux approach. :P
I beg to differ. Linux often introduces change for the sake of change and calls it improvement (systemd, ALSA, Pulseaudio etc.). Such changes are often not a real improvement and not really the topic here.
 
I know, I was kinda nervous about it spilling off into "Linux is xyz because of steam xyz xyz xyz".

How would something like SteamVR work today (if possible)?

I'm not sure on Linuxlator's overhead, but it sounds like Proton stuff on FreeBSD is hooked up through that. I'm curious how SteamVR virtual desktop and audio output stuff would work FreeBSD vs native Linux, and headset drivers/their tech (OG Vive/Index native, OpenHMD CV1, ALVR Quest with video encoding, etc).

Windows SteamVR going through Wine doesn't sound impossible (probably easier with ALVR/generic video vs headset-specific drivers), but I never heard of anyone doing that on Linux and assume it won't work.

FreeBSD has the low-latency and performance, and that'd be cool to see how it handles Beat Saber!
 
Linuxulator does not cost any measureable performance penalty, in fact it is even rumored to run binaries faster than some real distros! It also opens you up to the very real, very big threat of Linux malware, so take what you will.
Do you think linuxulator can run proton-ge and provide the same compatibility with windows games like a linux distro can ?
If yes, it would be worth a try.
 
quarter increments without pkgs vanishing en masse for months. net-im/signal-desktop to name one.
editors/vscode being another one. Sure, Node and Electron are dumpster fires with frequent vulnerabilities, but reliably having the latest safe version in quarterly would be nice.

Two other fanciful wishes:
  1. MediaTek MT972X wireless driver (effort seems to have stalled due to LinuxKPI/VM issues): for users of the many newer laptops with this family of cards, the only current options are dongle and net/wifibox (or own similar solution)
  2. Bluetooth HID stack redesign, moving more things into userspace and integrating with the iichid layer
Otherwise all of my needs continue to be more than sufficiently met by FreeBSD!
 
As long as Linuxulator has support for the same Linux Kernel version that proton-ge requires, then yes.
That would be a great improvement. Half the games I have in steam work with Mizuma, or over half, but linux-steam-utils works with like 12. :/ It would be nice to have a better proton version. The last time I tested, 1 year ago, GloriousEggroll's NobaraOS in wayland actually got less FPS in the same titles on the same machine when running Wayland on FreeBSD with Mizuma. So it would be really cool if we could get the GE version of proton going for the linux-steam-utils package. That would allow for more titles to work well and with better performance on FreeBSD. It would also be nice to see the GE version of Proton for Mizuma as well. Mizuma, in my experience, has a significantly higher compatibility with Steam titles. I do, of course, realize that this is a package request and not a base system request. :D
 
Me neither, but the handbook quotes version 3.17, last edit 2025.
If I am not wrong the binary compatibility of linuxulator should be around linux kernel version 5, and that is still somewhat behind of the version proton-ge requires.
 
I would suggest to make the language supporting variants local to the packages instead of having them as a program package like branch in the ports tree.
 
I think FreeBSD is on a good trajectory, but we still have a large way to go.
  • pkgbase. We've been talking about it in current@ and stable@ for a decade now, it'll debut with 15.0-RELEASE
  • improvements to bhyve
  • GPU virtualization support in bhyve
  • better documentation for FIBs
  • better support for alderlake p-core/e-core architecture
  • system level support for DNS-over-HTTPS and DNS-over-TLS
  • LoongArch support
 
better support for alderlake p-core/e-core architecture
Maybe review D45387 (already landed with multiple related commits on main) would be the first step.

For virtualization support, it would be better embedded into UEFI firmware as runtime services, means, all OS runs virtualized, all hardware resources to be dynamically reallocatable (i.e., CPU cores, GPU cores, memories, network bandwidth,...) like in mainframes.
 
Back
Top