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

This is my big request. I can do some work using SAMBA clients such as Thunar, but my interaction with our network shares really requires scripting and mounting. Our organization is 100% Microsoft on the server side, so all SMB shares. As far as I can tell, there is no technical or licensing reason to not update, it is just a manpower issue. I don't know how difficult it would be to write the code, but Linux CIFS utils was updated rather quickly, so it can't be an insurmountable challenge. Too bad I'm just a technician/low level admin, and never learned any C programming.
FYI: Ongoing plans including SMB/CIFS.
Not sure about actual progresses.
 
You can use stat for that. find .... -print0 | xargs -0 stat -f "..."

Thanks. I had forgotten that stat(2) takes more than one file argument.

However, this solution exposes drastic incompatibilities between the BSD and Linux versions of stat(2). Maybe I should drag around one or the other in my personal tools.
 
Developer Requests

I'd really like to see fcntl() with F_KINFO work reliably for all FSes and not just if the kernel feels like it because it finds something in a cache. fcntl() with F_GETPATH works on macOS and Linux and Solaris can just look in /proc.

Documentation for ioctls.

Use only one guardpage for thread stacks.

libc tunables like GNU libc (and in particular some way to turn off the pthread stack cache)

A libc freeres function

Desktop (KDE) User Requests

Working Discover

Fewer (or no) kwin/drkonqi coredumps

My old system intermittenly gets stuck during login and I have to use ssh to kill kwin. That system is probably going to switch to CURRENT so we'll see if that problem persists.

Platforms

Better Raspberry Pi 5 support

That's all that I can think of for now.
 
0. pkg working reliably

1. fix the bug that crashes graphics hardware acceleration after a VT switch:-

2. waterfox web browser package (www.waterfox.net)
- when I install the www/waterfox package, I appear to get firefox Nightly??!

3. something like the debian 'ufw' simple firewall setup tool to simplify firewalls

4. working bluetooth stack (reliably)

5. go compiler in packages

6. do all the laptop support stuff that this page talks about https://github.com/FreeBSDFoundation/proj-laptop

7. a standardised and simplified way to set up encrypted dns over tls using local unbound without having to read and work my way thorough 5 or 6 different web howtos all of which describe different ways to do it

8. kde6 / plasma6 working with wayland

9. it would be nice if there was a concise and simple way (eg, one line!) to tell the system to configure a machine's ethernet and wireless lan interfaces for lagg failover without having to add all this rigmarole to rc.conf :-

# failover between wifi and ethernet
ifconfig_em0="up"
wlans_iwn0="wlan0"
ifconfig_wlan0="ether xx:xx:xx:xx:xx:xx WPA"
cloned_interfaces="lagg0"
ifconfig_lagg0="up laggproto failover laggport em0 laggport wlan0 DHCP"
 
Even though my main use of BSD is on my server and desktops, I would love better laptop support, so I could finally use it on my main laptop! Finally, the last trace of Windows banished from my life!
Windows is not bad for what it is.
It just shouldn't be run on the same storage device where your FreeBSD boot loader is. 😅
But virtualization could encapsulate windows, and let you use the pros of both worlds.
 
Aoki-san kansha itashimasu.
Yes, I know that the FreeBSD foundation works with AMD to improve the amdgpu driver support, but I need to wait a little bit more to get RDNA 3.5 GPU support for my iGPU.
The problem would be the driver codes are ported from Linux drivers, which are behind Windoze drivers, and if the drivers require additional KPI/KBI level of kernel support, FreeBSD project needs to implement them in LinuxKPI, thus, larger delays.

What is wanted is that GPU vendors directly provide FreeBSD-native drivers under BSD-compatible licenses. Maybe some parts of the drivers would be provided as pre-compiled binaries like nvidia does in those cases.
 
What is wanted is that GPU vendors directly provide FreeBSD-native drivers under BSD-compatible licenses. Maybe some parts of the drivers would be provided as pre-compiled binaries like nvidia does in those cases.
I don't support every move NVidia takes, but they did a good choice providing atleast binary drivers for FreeBSD.
Now I am excited to see what steps amd will take to provide better support for FreeBSD.
Sure they could provide binary-drivers as well.
FreeBSD is not at fault at all, contrary it is very good, and works good, but RNDA 3.5 support is just missing.
I wish I could do that part, but I am just to busy, too.
 
Windows is not bad for what it is.
It just shouldn't be run on the same storage device where your FreeBSD boot loader is. 😅
But virtualization could encapsulate windows, and let you use the pros of both worlds.
I don't run BSD on that laptop at all. I don't dual-boot unless there is a seperate physical drive for it. That is why most of my desktop computers have 3+ drives! You know you are weird when you kit your desktop out with as many drives as your server.
 
You know you are weird when you kit your desktop out with as many drives as your server.
Yes, it was the firstl lesson of don'ts in my cybersecurity course back then.

I don't run BSD on that laptop at all. I don't dual-boot unless there is a seperate physical drive for it. T
Yes, you are doing it the recommended way, not my recommended way, though, but I have some trust in professionals.

My server is my desktop anyway, with a little bit more HDD storage...
I mean it serves me, anyway, uncoditonally 😂

I have a puri librem 14 laptop with binary harming blobs from intel removed.
I might to try to install FreeBSD 14 on it. :D
 
I would like to see 3 major outstanding bugs fixed or at least proper release errata documented for them. First is a race condition with jails and ip6 where if a jail is stopped and it's resources freed while a IP6 connection was left open, the whole hypervisor's kernel with kpanic with a page fault. This particular bug has quite a few duplicates in the tracker, but the PFsense project was able to narrow it down and issue a temporary workaround until a proper solution is found which for some reason still hasn't made it into the tree.
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=279653
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=238326

This is quite an important bug because without it the jail subsystem isn't ready for production use. The reason it hasn't been discovered before is many people are using legacy-IP in their VNETs.

Application mixing broke with a change to the audio subsystem. Since the major change most audio mixing programs such as dsbmixer report 'no audio apps found' where per-application controls used to be. This is a problem in particular with some sound cards that have really low output and need amplification to heard properly.

X11 Acceleration broken on VT switching

This has been an outstanding issue for the last couple major releases. A system in which this occurs is a terrible experience to use with stuttering and excessive heat unless your on the latest and greatest cheapsets which can afford to do all rendering in software. It also has weird interactions with some accelerated userspace software such as MPV media player causing a single invocation while acceleration has been broken to hijack the entire screen and put it into an unusable state. If no solution to this can be found soon it should be documented as release errata. On some shared systems VT switching is the primary way in which users are able to share the computer.
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=267915

Better wifi support

While wifi drivers for the Intel chipsets are getting some love, progress seems slow. This also leaves other chipset families which used to be flagship in a sorry state such as Atheros. https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=154567 Atheros at the time only being able to use 802.11n, which most manufacturers now are only adding in for legacy compatibility as an afterthought. Ounce the new WiFi standard that is able to segment off OFDM carriers to individial clients for better efficiency in high density areas becomes more mainstream this is going to be even more of an issue.
 
1. bhyve fbuf w,h, expanded beyond 1920x1080. Lotta 4k monitors out there these days.
2. bhyve nvidia passthrough first class support.
3. Brave browser available in pkg/ports without Linuxulator (not great in jails).
4. One or two browsers with the latest sandboxing and security features implemented in the FreeBSD port
 
While wifi drivers for the Intel chipsets are getting some love, progress seems slow. This also leaves other chipset families which used to be flagship in a sorry state such as Atheros. https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=154567 Atheros at the time only being able to use 802.11n, which most manufacturers now are only adding in for legacy compatibility as an afterthought. Ounce the new WiFi standard that is able to segment off OFDM carriers to individial clients for better efficiency in high density areas becomes more mainstream this is going to be even more of an issue.
That's interesting about Atheros. Years ago I bought a ath9k card for a Lenovo T500 specifically because it was open-source friendly! (iirc it was also against Lenovo's BIOS whitelist, which didn't matter cuz Libreboot :p)
 
  1. I remember reading (20 yrs ago) how the BSD network stack was far superior (speed, structure, kernel integration) than the linux kernels, but recently that Linux has caught up if not surpassed the BSDs. I have absolutely no idea if this development is true, but if it were, I would like to see the BSDs regain that superior position.
  2. Native full speed wifi without resorting to wifibox.
  3. The package managers for the language specific implementations (e.g. cargo for rust, npm for node, etc...) provide first class support for BSDs. No clue how to solve that one.
  4. Documentation has really improved. Would like to see more documentation around building custom implementations with embedded apps such as PI. Installing direct is one thing, but building your own is quite another.
 
What the title says. What features (specifically ones that have not been announced yet) do you want to see in 15.x, 16.x, and beyond?

Better Laptop support, especially finger print support on Thinkpads, automatic decryption of boot drives.
Installer support for encrypted UFS.

Automatic switching between Wifi and ethernet.

better Linuxulator that supports Docker.
 
Back
Top