What features are you excited about for FreeBSD 12?

There was no incentive for it because it worked. But Microsoft has completely disabled SMBv1 since then and turning SMBv1 back on is really bad advice. So now there's an actual need to support SMBv2 and above.
 
tankist02 You need to choose more wisely. Call it luck but I've never had hardware issues except for one Broadcom network card in almost 15 years of using FreeBSD. My current high end workstation, I chose parts I knew worked with FreeBSD. Off the shelf components but name brand parts easily found and widely used like Gigabyte, nVidia, etc. It was easy. I bought them all from Newegg cause it was cheaper than the local Microcenter store which also carried the same things.

Personally, I hate having to "cherry pick" parts just to ensure compatibility. When it comes to compatibility, I don't shop for parts that work with my OS; I search for an OS that is comparable with whatever hardware I want to use. :beer:
 
With FreeBSD we definitely have to shop for the hardware to fit the OS, and, let's be honest, you'd better not buy the latest and the greatest, and even with that, there's a risk of not running your gear to its full potential—I'm still locked to 802.11g. I wonder if instead of chasing to support everything FreeBSD would focus more on selected hardware "lineages," e.g. some series of ThinkPad laptops, workstations, and so on - whatever is popular. Call me a heretic but it would be useful to have a BSD OS that is guaranteed to support specific hardware platforms. Perhaps, hardware vendors should be involved. And yes, I'm yearning for NeXT-like application bundles, a feature that isn't even on a horizon. It would be great to not having to build your own or download the entire repo to run a desktop with thousand of packages without special effort or care.

Sorry for off-topic. I'm also looking for better hardware support.
 
I hate having to "cherry pick" parts just to ensure compatibility.
With FreeBSD we definitely have to shop for the hardware to fit the OS, and, let's be honest, you'd better not buy the latest and the greatest, and even with that, there's a risk of not running your gear to its full potential
It's not hard to do. While other OSes may be compatible with 10,000 devices, FreeBSD is probably compatible with 9734 of them. You only need one and, most likely, the one you have in mind. I did not have to compromise with the off-the-shelf, leading edge hardware I chose when I built my workstation at all while going for performance.

If you notice, almost every complaint about hardware--maybe even every complaint--is only due to Broadcom networking cards or a graphics card. Little else. And, even then, isn't the graphics complaint about non-nVidia or AMD?
 
Mostly the complains come when the cheapest thing ended up on your buy list. The shop where I got my server hardware has a policy allowing to return stuff if it would not work, or using a customer supplied USB stick to boot it, or install from when they build the system from your wish list. And they don't sell the cheapest stuff.

Back to topic, I'd want better power saving & ACPI.
 
It's not hard to do. While other OSes may be compatible with 10,000 devices, FreeBSD is probably compatible with 9734 of them. You only need one and, most likely, the one you have in mind. I did not have to compromise with the off-the-shelf, leading edge hardware I chose when I built my workstation at all while going for performance.

If you notice, almost every complaint about hardware--maybe even every complaint--is only due to Broadcom networking cards or a graphics card. Little else. And, even then, isn't the graphics complaint about non-nVidia or AMD?

We have very different experiences with FreeBSD and hardware compatibility. I've tried FreeBSD with dozens of workstations and laptops over the past ten years and never had the OS work 100% with any of them. With over half FreeBSD won't even boot. This covers a range of OEMs, different video cards, different network cards (none of them were Broadcom), different CPU brands. Either I have unusually bad luck, or you have had exceptionally good luck.

With servers I tend to have pretty good luck, at least with older or virtualized platforms. But consumer hardware is always a no-go for me with FreeBSD.
 
I don't shop for parts that work with my OS; I search for an OS that is comparable with whatever hardware I want to use. :beer:

Really? I am actually the opposite. I am more of a software guy so I suppose that might be why. But when I sit down at a computer, the operating system is the important part for me, I generally don't give a damn about if it is running an Intel, AMD, or... (nope, that is about it in 2018 unfortunately) processor.

So I am quite happy to pick the compatible hardware up to an extent. I.e I will be upset one day when Laptops supporting FreeBSD are non-existent.

I guess it is slightly similar to a macOS user expecting to have to buy Apple hardware in order to use macOS. The only difference is that other than the price, It is extremely easy to find resellers for this type of "compatible" hardware for macOS.
 
I am happy that FreeBSD is still a portable piece of C software and the prime example for that is the ARM port. Really, great work.
I am not missing much in 11.2 though. It works great, is really stable and has many useful features. Does anyone know about bhyve improvements?
 
I don't see support for Optimus Technology but top has gotten an overhaul. Yay.
 
Last edited by a moderator:
POWER9, PowerPC e6500 (for PowerPC Notebook project), Pine64 and RISC-V support, GELI-encrypted ZFS boot, jailed bhyve, bectl, lua loader
 
There's some work on virual filesystem passthrough that looks very very interesting for bhyve. I like this better than loopback to the host over nfs (which is itself a neat trick). No idea if this will be part of 12.
 
I just tested 12 RC1 with OpenSim 0.9.0.1 (latest) with 4 OpenSim servers and 1 Robust Grid server and 1 Percona 5.7 server setup and it all works - including custom builds using packages gcc, mono and more. Normally consists of a set of virtual machines on vmware but I also run it on a set of jails and the jail setup outperforms any other setup, including doing the setup on OpenSUSE. I'll be upgrading these systems to FreeBSD 12.0 when it's released.
 
I don't see support for Optimus Technology but top has gotten an overhaul. Yay.

What a fool I've been...

I just installed FreeBSD 11.2-RELEASE on my Thinkpad W520 with support for Nvidia Optimus Technology out of the box.
 
Haven't looked into it since it was first talked about but I believe it's a bit hacky and redirects blocks that were on the removed device to a new location. I still think you're better off avoiding removing devices unless you get really stuck,



Yeah I saw encryption mentioned on the mailing list today, which at the moment is just on GitHub. I'm not really a fan of the timed releases for .0 versions. It ends up with major new versions often containing just incremental changes, rather than reviewing possible major features, such as zfs encryption, and planning to get them in for the release. Apart from general improvements and a few new drivers I don't think there's much to get excited about in 12, and a major feature like zfs encryption may randomly appear in a minor point release.

Yeah this is a concern for me as well. e.g. the version of clang got bumped up some major revisions between minor point 11 releases. In the past the compiler version was static for an entire branch.

Also that ports get forcefully disabled now when the world is EOL, and there is effectively no LTS branch of FreeBSD anymore as the extended release's have been scrapped.

But for sure I feel a release should be released when ready not to a time schedule.

What happed to the "whats cooking" pages for Major FreeBSD versions, they provided good detailed information, now they seem to have been replaced with sparsely populated wiki pages.

Compare this

http://www.ivoras.net/freebsd/freebsd9.html

to this

https://wiki.freebsd.org/WhatsNew/FreeBSD12
 
The Clang change isn't a problem unless you can prove that it breaks a userland API/ABI (go ahead and prove me wrong, I doubt you can), that's what the stability in FreeBSD is all about and not about whether the tools used to build the programs are this version or that version.
 
Back
Top