Interesting. I use FreeBSD on my desktop too but it's really a desktop so I don't have to bother with WiFi or bluetooth. I generally dislike laptops for ergonomic reasons, and I never bring my computers anywhere anyway so I just buy NUCs. Not having to buy for a display, keyboard, trackpad, battery helps keep the price down.
I like it for several reasons. It's a holistic system which means it's much easier to understand, not a collection of random parts thrown together. There is only really one (big) distro so documentation is easy to come by and consistent. I love the way the updates of the system are uncoupled from the userland software so you can have rolling packages but a stable OS.
Also the ports collection is great (being able to manually compile every package with different flags where needed). And jails. And ZFS first-class citizen. Also I like the attitude. Less involvement from big tech, less strive to change for change's sake. It feels a lot more stable, every new version there's only a few things changed. It's not that with every major update I have to learn everything anew again because someone wanted to include their new init system (like systemd), configuration tools (like ifconfig -> ip), packaging system (like snap) etc. Things that work fine are just left alone.
It has some really good ideas also, like boot environments. But it's not linux. It's not meant to be.
But yeah if you want everything all figured out for you, don't use FreeBSD. Just take a commercial linux like ubuntu. You'll need to tinker a bit, which I like because it helps me understand my system. FreeBSD is a bit like Linux was in the early 2000s, it mostly works but you often have to dive into a shell for some magic. The good thing is having ZFS snapshots as a safety net though. Never really get caught out that way.
If bun and node.js run I'm sure the agents will, you might have to fight tool calls since the system utils differ from GNU core utils a tiny bit here and there, but you could toss the agent and whatever tools into a jail and have a nice package, use zfs snapshots between prompts so you can disect it later.
I don't get what they have been ironically saying. Direct reading suggests it's that Linux is like Windows in 2000 but it doesn't make sense to me, I never heard a comparison like that.
Love the Thinkpad line. It's kinda the hackers laptop of choice.
> X220/X230
These are pretty solid with dual batteries. Also popular for OpenBSD and 9front, the latter of which I run on an x230 (it stopped charging the removable battery :-/) You can get about 2 hours off the internal and 6~8 with a big fat removable battery, maybe more if the OS and drivers can properly throttle hardware power settings.
I have an X1 carbon 5th gen and it's quite light but not useful for 9front without some Ethernet driver tweaking (likely some phy bits need twiddling.) Instead I tossed Debian on it and run 9front in a VM if I need a local CPU. So far it just seems to work including Ethernet (via a dongle) and WiFi.
I just went through T14 Gen 1 hell. I tried to replace a broken LCD and the system just wouldn't boot. I fussed with it so much, the flimsy case cracked and broke in several places. I gave up and decided to use it as a clamshell, but had to keep the broken screen connected to boot.
But I couldn't get it to see the external monitor. I fought this for hours. I finally reached my patience threshold and tore the fucking LCD off the case. It hurt the hands a bit but was cathartic. Amazingly, it booted and saw the monitor, magically. I was astonished. With or without the broken screen, it wouldn't boot. But going psycho on it fixed it. Proof in the wholesomeness of violence, maybe not. But...
it's not running BSD. It's now my mom's desktop, and it's running Void. Works great, and I love the Ryzen 4700, with 16gb ram and atheros wifi chip. Delicate, but capable of some extremes :)
From the link: "Note: The inbuilt WiFi chip is not natively supported by FreeBSD, so you will need to (temporarily) use a USB WiFi or Ethernet dongle, or (as I will explain) copy some files from a different system to the Macbook. You could also just transplant a different chip into the system."
You say "works perfectly". I do not think it means what you think it means.
To be fair, Linux also has trouble with the Broadcom chip, the driver needs to be installed as a separate step on most distros.
| Works Perfectly | Mostly Works | Has Lots Of Bugs
-------------------+-----------------+--------------+-----------------
Default Install | | |
-------------------+-----------------+--------------+-----------------
With Add-Ons | X | |
-------------------+-----------------+--------------+-----------------
Major Config Work | | |
i.e. Declare its working quality after the install is done. The install may take multiple steps. (In this case, copying some files over, apparently.)
Broadcom (and to a lesser extent, Realtek) devices had always been anywhere between hit-or-miss and completely unworkable on Linux, LONG before Raspberry Pi came around.
It's MIT licensed now, which isn't particularly useful when it comes to Pi (there's some Broadcom crap in that boot loader so it won't be open sourced) but otherwise is kind of interesting.
I always saw Broadcom as evil, and saw Raspberry Pi as just reusing cheap parts from set top boxes or similar, with all the proprietary stuff that that comes with.
By that logic, every piece of software ever made can be said to work perfectly in every situation, because there is always some amount of additional work which could be done to make up for its native deficiencies.
That's quite the leap. The work is already done, they just can't/won't ship the driver in base, right? Isn't it comparable to installing Debian and needing to load in non-free drivers separately?
That is cool in ways, but many manufactures change the internals without changing the model number and so I'm not sure how much I can trust it. There is a recycled computers place near me that will sell me some of those cheap, but how can I be sure the one I'm buying is the same as the one tested (if indeed I can find any of those model numbers at all - which is a factor of what companies near me are recycling this month)
That kind of seems crazy to me, considering OpenBSD has worked perfectly fine with every wifi capable device I've tested it on. Granted, most of them were older machines.
Is this just an artifact of FreeBSD primarily focusing on server hardware rather than consumer/end-user stuff?
Basically no one supports Broadcom SoftMAC WiFi cards very well, but OpenBSD just doesn't. I have a 2015 MacBook Air with a BCM4350 where the recommended fix is to go buy a FullMAC card from a similar vintage MacBook Pro and just cope with the fact that the card barely fits in the case and can't be secured properly.
But then you’re stuck running Linux as your primary OS, which is a huge drawback.
FWIW, I’ve done both, FreeBSD with a virtualised OpenBSD for WiFi, and currently I run Arch so I can do gaming, with FreeBSD virtualised for development. I’m kind of looking to go back to my previous setup.
Why? Nothing wrong with running your network interface in a VM. There are reasons for doing so even if drivers aren't an issue. Qubes OS does this, for instance, for security reasons.
Windows also does. Almost everything is a VM in windows these days.
It's just how things work these days. If you'd say "I run my VPN client in a docker container" it would raise a lot less eyebrows. Yet it's not very different, really.
Though conceptually I'd frown at having to run Linux. I'd prefer upgrading the hardware to a supported chip.
Definitely not independent kernels but my guess is he's referring to Virtualization-Based Security (VBS) - it gets turned on by default if your Win10/11 system has virtualization enabled.
One such example security solution is memory integrity, which protects and hardens Windows by running kernel mode code integrity within the isolated virtual environment of VBS. Kernel mode code integrity is the Windows process that checks all kernel mode drivers and binaries before they're started, and prevents unsigned or untrusted drivers or system files from being loaded into system memory.
Why not? FreeBSD has never been intended as a batteries included, everything "just works" out of the box OS. It's meant to have a bare minimum install and let the user choose how things are set up. You can disagree with that philosophy, but that's not an indictment of FreeBSD. Just go use something that aligns better with your preferences.
I agree that 9/10 is a bit of a strange score there, but it's not all that bad: You can get a $15 wifi dongle and use that instead. It occupies a USB port and looks a bit ugly, but it's still a fairly easy workaround.
I'd like to know more about what it takes to turn on PCI pass through for laptop hardware. On desktops and servers it's typically the IOMMU setting in the BIOS. Is that also commonly available on laptops?
Some years ago, I was workig with FreeBSD on an old laptop. The laptop had a wireless adapter that ostensible should be supported, but was not.
After some digging, I realized the driver was just missing some PCIe device identifiers. I added them to driver and bam my WiFi is working without issue.
I tried to submit a bug report and patch, and it got positive feedback at first any changes even got committed. But then I learned why it’s better to not even try.
Apparently this was a known issue, but only in the heads of the FreeBSD wireless developers. They had their reasons for not adding the device, but the reasons did not appear to be documented in mailing lists or docs until my thread. At that point I realized it’s not worth it to try and contribute to such large projects as I just lack the decades of institutional knowledge of the system.
Anyway, I’m not sure it ever got released. I believe there’s an umbrella bug somewhere left after the version my patch supported went out of support.
Maybe one day! One of the things that has changed in the past 20 years with FreeBSD is the addition of bhyve. Virtualization on FreeBSD was not great prior to bhyve. Because FreeBSD waited to implement it, they ended up with an arguably more modern hypervisor versus what Linux had at the time. Maybe history will repeat when it comes to application containers / docker images. Podman came 5 years after Docker and the result was better. It could happen with FreeBSD too.
Wow, searched this thread for docker hoping to realize I was behind the times and thinking surely it must work by now. It's so crazy that the BSDs have a reputation for stable, implying sysadmins and servers, then it won't support the one thing that every server needs. VMs running linux is the preferred path of least resistance then? Honestly you'd have to work hard to find a better way to alienate your target audience.
It's crazy how much negativity there is in comment threads like this. I would get it if FreeBSD was a product you paid for, or someone was evangelizing about how you're missing out if you don't get the FreeBSD laptop experience, or something.
As someone who liked FreeBSD in the past and curious to check it out again, I'm glad to have this handy list.
What is sad is that even though Linux now has hardware support that is miles ahead of Windows, we've exchanged one problem with another, because nowadays most of the hardware I see is only supported on Linux and nothing else.
Even on PCs, latest generation AMD graphics cards (already >1yr old) are not supported in _anything_ other than Linux (and Windows). This is just sad.
FreeBSD uses a compatibility layer to run the Linux graphics drivers, though it lags behind Linux. So if FreeBSD currently does not support the graphics cards, it will soon. It looks like they are currently porting over 6.11: https://github.com/FreeBSDFoundation/proj-laptop/issues/41
It is the _only_ OSS operating system that supports AMD cards from this decade, and it does so by having to emulate the Linux kernel API, and yet _still_ it lags years behind Linux itself. I've chosen this example for a reason -- this is exactly what I'm sad about.
Not to sound like goal post mover, but once you tried 802.11be @ 6GHz, you never go back (assuming what your AP connected to can handle it).
That's my problem with FreeBSD on non-servers - eventually it's supported, usually via Linux shim, but it's too late. By the time FreeBSD started to support (on CURRENT) GPU that forced me to switch, I already upgraded twice.
eh I've got 802.11be @ 6ghz at home (u7 pro AP and QCNCM865 client) and it is truly impressive, but I only notice the difference once every 1-2 years when I'm transferring a full-disk image over the network for backup. We've long passed the point where the speed improvements matter for daily usage (browsing the web, streaming video, remote desktop, installing updates). For those use cases, you wouldn't notice a slowdown on 802.11ac and I'd argue even 802.11n would be fine.
The one exception I can think of would be video content creators since they end up with large amounts of raw video that would benefit from transferring at much-faster-than-streaming speeds.
And I guess steam downloads if you don't plan ahead at all, but if I'm planning to play a game later, I'll tell steam to install it hours or days in advance.
I often have to transfer large files in LAN. 802.11ac and 802.11n definitely enough for most of the time. I often have to transfer large files between machines in LAN, with 802.11ac makes you remember that it's wireless, while 802.11be makes you forget.
I have the same setup for a framework main board next to the AP, and it's reliably faster than using their usb-c ethernet extension card.
I wouldn’t mind faster wifi speed, but reverse engineering stuff has always been slow (i.e Asahi Linux, and they only have a few device to investigate)
Windows is cleaning up a lot of legacy drivers. A bunch of printers (+ scanners) that predate updates to the printer driver framework in recent versions of windows just don't have functioning drivers anymore, despite being perfectly functional.
All these devices work out of the box on linux, more or less.
Most devices that you can buy for under $400 now run on ARM chips (frequently Mediatek). We're talking tablets (with keyboards), convertibles, even outright laptops (i.e. "netbooks"). These things qualify as computers. They are replacing traditional laptops, just as those replaced desktops.
If we're looking at sub-$400 computers, especially on ARM, it seems like we have to include the large segment of ChromeOS devices that only run Linux out of the box (or at all, generally).
Referring to Intel Chromebooks (i.e. laptops), that segment is now dwindling in size much as its predecessor (Intel Windows netbooks) did a few years ago. Most low-end ChromeOS devices now run on ARM. And Android is nipping at their heels.
Netbooks were originally Linux. MSFT created a special licensing class just to try to undercut it. It wasn't great, but because Windows and Microsoft licensing, it quickly took off. People realized Windows on netbooks sucks, thought that meant that netbooks sucked, and eventually netbooks died. Until, arguably, ChromeOS arrived.
Sure. And all of those devices run Linux. Some of them even run other Linux OSs decently; one of my daily drivers is an ARM Chromebook running postmarketos.
It is not trivial to get FOSS Linux onto a write-protected Intel Chromebook, compared to a Windows netbook of yore. It is harder still to get it onto an ARM Chromebook or Android tablet. PostmarketOS is a bit simpler (or at least better documented) but it is not a full Linux distro.
Installing a fully-fledged FOSS OS on low-end general-purpose computing hardware is getting harder. Certainly for the non-techies who have to be part of FOSS if it is to survive.
I think it's better and worse in slightly different ways. On the one hand, yeah a Chromebook won't let you touch the default OS without switching to developer mode, and won't let you install a new OS without disabling the write-protect screw or firmware option. On the other hand, every ChromeOS device allows you to do exactly that, and then you can run whatever you want and you should have at least some support for upstream Linux because ChromeOS upstreams their drivers. (I will happily agree that the Android device situation is awful.)
> PostmarketOS is a bit simpler (or at least better documented) but it is not a full Linux distro.
By what definition is PMOS not a "full" distro? It's Alpine plus some extra stuff, including device tweaks and out-of-the-box desktop environments.
Most devices in that class I see run some vendor flavor of Android or ChromeOS and not Windows, so definitionally speaking they do run Linux out of the box.
Yes but it's a bit academic. The problem is that getting a FOSS distro of Linux onto low-end general-purpose computing hardware is harder now than it was a decade ago. I speak from bitter recent experience.
Oh, I know perfectly well what you mean. The move to the SoC paradigm has serious implications for the future of computing freedom. I can't imagine how we might be able to fight this crap, realistically.
There is only one area where windows excels: recent PC-based hardware. For everything else, primarily including anything that is not PC, and anything that is from more than a decade and a half or so ago, Linux is miles ahead, and there's no discussion possible.
Whether this is any helpful to us is another story.
I think of it more in the reverse, the choice being removed is the hardware you can use. It has been the case from the dawn of computing that you start from a usecase (which correlates to software, which maps to an operating system) and then look at your options for hardware. The more specific your usecase, the more specific your software, which correlates to a specific choice of hardware. There is no, and can be no, "have it all". It's a fundamental principle of mathematics, the postulates you choose radically change the set of proofs you have access to, and the set of proofs you choose entail the axioms and structures you can take.
Now it can be better or worse, and right now it's never been better. There was a time when your language, your shell and your operating system were specific to the exact model of computer (not just a patched kernel, everything was fully bespoke) and you have a very limited set of peripherals. That we suffer from more esoteric operating systems lagging behind the bleeding edge of extremely complicated peripherals is a very good place to be in. That there's always room for improvement shouldn't be cause for sadness.
> Now it can be better or worse, and right now it's never been better. There was a time when your language, your shell and your operating system were specific to the exact model of comput
No, it is not. There was a small period of time between the 90s and the 2010s where you could grab almost every 386 OS and have your hardware mostly decently run for it, and if not, drivers would be easily written from manufacturer specifications. That time was definitely better then than what we have today, or what we had before then. I am writing this as someone who was written serial port controller drivers for the BeOS.
> That we suffer from more esoteric operating systems lagging behind the bleeding edge of extremely complicated peripherals is a very good place to be in.
This is the wrong logic, because operating systems become esoteric since they can't support the hardware, and hardware becomes complicated and under-specified because there's basically only one operating system to take care of. You may _think_ you have no reason to be sad if you're a user of Windows or Linux, but you have plenty anyway.
> There was a small period of time between the 90s and the 2010s where you could grab almost every 386 OS and have your hardware mostly decently run for it
And prior to that, you could grab every OS running on IBM clones and not have to worry about graphics drivers at all, because graphics acceleration wasn't a thing. The era you refer to had already introduced software contingency on hardware within x86. This disparity was further compounded in the mid-2010s as GPUs exploded in complexity and their drivers screamed into tens of millions of lines of code, eclipsing kernels themselves. This is not distinguishable from the introduction of graphics drivers in any generalized manner. They were driven by the same process.
An important thing I want to point out as well; you're doing a lot of heavy lifting by limiting the pool to x86 computers, which is already giving up and admitting to a very strong restriction of hardware choice. Don't take that as pedantry, it's a very well hidden assumption that you've accidentally overlooked, or in the case that you think it's irrelevant, I'm letting you know that I don't consider it irrelevant in the slightest. When I think of computers, I'm not just thinking of x86 PCs. In the 90s I'm thinking of SGI workstations, Acorns, Amiga, Macs. I'm thinking of mainframes and supercomputers and everything else.
> This is the wrong logic, because operating systems become esoteric since they can't support the hardware
On the contrary, I assure you that this logic rests on faulty premises. As a general principle it's clearly false since most operating systems (which are long forgotten) predate it by decades, and in the specific context of Linux winning over FreeBSD, it's still not applicable as that happened smack dab in this era you describe.
> You may _think_ you have no reason to be sad if you're a user of Windows or Linux, but you have plenty anyway.
I'm a user of Linux, FreeBSD and 9Front. I just don't (and never have) bought hardware at random. You can reason your way into sadness any which way, but rationalization isn't always meaningfully justified. I just don't find it sad that my second desktop can't have an RX 9000 whatever in it. Where's the cut off line for that? Why not be sad that I can't jam a Fujitsu ARM processor into a PCIE slot as another type of satellite processor? The incompatibility is of the same effect, but I don't see you lamenting or even considering the latter, as though mounting a processor to a PCB is somehow fundamentally less possible than writing a modern graphics driver.
> And prior to that, you could grab every OS running on IBM clones and not have to worry about graphics drivers at all, because graphics acceleration wasn't a thing. The era you refer to had already introduced software contingency on hardware within x86. This disparity was further compounded in the mid-2010s as GPUs exploded in complexity and their drivers screamed into tens of millions of lines of code, eclipsing kernels themselves.
Not at all; I excluded this early era because you could _not_ be sure to find an OS that would support your graphics card at all, other than maybe what the BIOS supported. I am talking about the 90s because GPUs already had plenty of non-BIOS-supported features, like multiple CRTCs, weird fixed acceleration pipelines, weird coprocessors with random ISAs, and yet you could still find operating systems with 3rd party drivers supporting them.
It is a _perfectly_ distinguishable era. See how many OSes support 3D cards from the era like i9xx. Heck, FreeBSD itself qualifies, but also BeOS and many others.
In addition, I am talking about the _kernel_ part, which by any logic should be ridiculously simple. E.g. this is not a compiler to a random ISA or anything like that. It is what in Linux you would call a DRM driver, and the only reason they are complex and millions of LoCs is that they are under-specified, by AMD and the rest. Most of lines of AMD driver code in Linux are the register indices for each and every card submodel (yes, really, header files!), when it is clearly that before they would just have standarized on one set and abstracted from it. Compare AtomBIOS support in cards from a decade ago and cards from today. It is literally easier today for a 3rd party to implement support for the more complicated parts of the GPU (e.g. running compute code!), which AMD more or less documents, than it is to support basic modesetting support as it was in the 00s. This has happened!
Hardware may be more complicated, but interfaces needn't be more complicated. This, I believe, is a symptom, not the cause.
> I just don't find it sad that my second desktop can't have an RX 9000 whatever in it. Where's the cut off line for that? Why not be sad that I can't jam a Fujitsu ARM processor into a PCIE slot as another type of satellite processor?
You do not find it sad that there is no longer any operating system other than Linux supporting any amount of hardware, simple or not ?
Also, you call every non-Linux OS as "esoteric" as a counter-argument to my point , yet you try to use support for definitely esoteric hardware (which would be even hard to acquire!) as an argument for your point, whatever it is ? When I'm complaining that I can no longer rely on FreeBSD, literally the 2nd open OS with most hardware support, on supporting basic hardware (!) from this decade, when on the past I could more or less rely on _all_ BSDs supporting it, as well as a myriad other OSes , the argument that "oh well it never supported hardware that it is impossible to find in stores anyway, so I don't care" sounds pretty hollow.
Certainly even slightly deviating from the popular hardware has always resulted in diminishing returns, but today it is much worse, _except_ for Linux.
A few years back I attended a talk by one of the Netflix developers who contributed kTLS. It was very informative as he pointed the reasons why Netflix went with FreeBSD. The first question afterwards was “why not Linux”… Didn’t you listen to the talk, man…
It's not as polished as linux obviously, especially for desktop usage but the maintainers are very much on the ball (and they do a lot of work to get things to compile and work, there's a lot of linuxisms they have to work around).
> this simplicity also comes specifically because there's less contributions.
Not entirely. A rather large amount of Linux's mess stems from the fact that it was a hobbyist project in its foundational years. It was never clean or well designed, at any point in its life. Go look at Linux 1.2.0 vs FreeBSD 2.0
Even when Linux began to get traction, it had already developed an ingrained culture that didn't particularly care about "nice" code or architectural solutions. The BSDs inherited their culture where such things were prioritized. You're right that things get messier as they get larger, but the gap between the two is much, much larger than can possibly be accounted for. Things like Linux not respecting NICE values have very little to do with surface-level problems like stylistic inconsistencies in the source code.
>It's crazy how much negativity there is in comment threads like this
I think it's because this chart continues a trend I've noticed with BSD zealots. Namely, there's some sort of reality distortion effect at play.
Consider that there are obvious bullshit scores on TFA, like giving a laptop 9/10 when the fucking wifi doesn't work. In reality, this should be 5/10 or arguably 0/10. After all, what use is a laptop without wifi? If my laptop's wifi didn't work I wouldn't just buy a usb-ethernet adapter and never bring it anywhere; I would get a new laptop because a laptop without WiFi is useless.
On top of that there was a while here where every BSD thread had:
- a comment about how BSD powers the PlayStation, Netflix, and other FAANGs, except those corps don't contribute enough back because of the license so won't you please subsidize these giant corps by donating to BSD?
- people who argue BSD is superior because it's "more cohesive" and "feels cleaner" or similar
- OpenBSD zealots claiming it's 110% secure because trust me bro
Mostly I'm just tired of people claiming BSD is this amazing new thing with no flaws, when reality is that it has got some niche use cases, I suspect lots of its developers don't even dogfood it, and is otherwise superceded by Linux in nearly every meaningful way.
I have no problem with BSD, and I have two boxes in my basement running freeBSD right now, but I'm not delusional about BSD's limitations.
> Mostly I'm just tired of people claiming BSD is this amazing new thing
I don't think I've heard anybody claim BSD is new.
> Netflix, and other FAANGs, except those corps don't contribute enough back because of the license
I believe Netflix has upstreamed a lot to FreeBSD. They don't do it because the license compels them, they do it because upstreaming your changes makes maintenance easier.
> If my laptop's wifi didn't work I wouldn't just buy a usb-ethernet adapter and never bring it anywhere
I'm going to guess with this rant that you weren't using Linux in the olden days, because that's what it was like. The workaround isn't using wired ethernet by the way..you can get a USB wifi adapter or you can buy an m.2 wifi card. On on one of my machines I got a cheap m.2 Intel ax200 (just checked, about $15 on eBay) because it runs faster on FreeBSD than the one that shipped with my laptop.
>I'm going to guess with this rant that you weren't using Linux in the olden days, because that's what it was like.
I've been using Linux and BSD in one form or another since 2003, and I definitely used wpa_supplicant on the command line to connect my Thinkpad to WiFi. And you're right, it did suck. It was not a 9/10 experience by a long shot.
FreeBSD actually has a similar thing, you can run Linux wifi drivers inside a VM and pass through the adapter. There's a port called wifibox that does this.
You can even forward the Unix domain socket for wpa-supplicant from the guest to host, so all the normal tools that talk to wifi cards via that socket work transparently.
Regarding your wifi example. I did have to replace it with an intel one on my Lenovo because wifi would not work with something connected to Bluetooth (might have been USB . I don't recall). This is on Windows by the way. I just replaced it instead of fighting it. Same reason people prefer AMD on linux but this is changing with better Nvidia support.
If you look the table you will realize 9/10 means 9 of 10 included HW devices run. Is not a scale from 0 to 10. Is not a "out of ten" in usability scale. Just count the devices that work, vs. the ones included in the HW.
> I would get a new laptop because a laptop without WiFi is useless.
You can run Linux in a VM and PCI passthrough your WiFi Adapter. Linux drivers will be able to connect to your wifi card and you can then supply internet to FreeBSD.
Doing this manually is complicated but the whole process has been automated on FreeBSD by "Wifibox"
is there a similar thing for GPUs? I want to build a workstation and have it work on freebsd but would prefer to use an intel arc card which has no information about freebsd compatibility online
Fun fact: My old Lenovo Y50 only supports like 3 specific WiFi cards else it doesn't even POST. And I think none of them work with upstream Linux drivers (I think, have only 2 different ones and neither worked ages ago and I changed laptops a while ago and haven't retested). Actually I think one didn't have bluetooth work (the non-standard one) and the other needed the broadcom-wl package.
Paradoxically, given their otherwise positive standing, Lenovo has keept allowlists on their BIOS for specific devices on specific ports. For example, I have a T460 that has an m2 slot that only works with two specific WWAN modules.
I remember seeing something in that direction when I was looking but never did look deeper into it.
The post made me actually take out the laptop again and maybe use it as a server or something like that in the future and for that I'd use ethernet anyway.
Replacing the wifi card isn't necessarily easy. I'd rather not buy and use a USB dongle for it if I can just get it out of the box.
I remember doing those kind of things nearly two decades ago now, I don't expect to have to do that in 2026. If people want to, that's fine, but the parent comment is right here: giving it 9/10 without working wifi is ridiculous.
Seriously. I'd rip the wifi hardware out of the laptop with a spoon if it somehow got me a laptop that handles sleep mode properly. I can't even imagine what that would be like with a Unix (aside from a Mac).
I think what you're seeing is partly a consequence of how capable Linux has become. Linux is in a weird phase where it can still be enjoyed by hobbyists/enthusiasts/eccentric types, which were arguably its original audience, but now you can also Zoom and do work and install Steam on it, which gives it less appeal from the niche/hobby angle. The software ecosystem in Linux is also increasingly homogenizing, which helps with the "practicality" aspect, but also diminishes the niche appeal. BSDs are in a position to snap up that audience that appreciates engineering elegance/design and uses the computer as an end unto itself (not just as a means to an end). This audience isn't necessarily bothered by wonky laptop WiFi, and may even enjoy tinkering with it as a hobby project. Just my take.
If you don’t care about administrating your computer and just want to use some software on some hardware, the BSDs are not that great. But if you do, the experience is better on the BSD land because cohesiveness reduces cognitive debt.
Also I wouldn’t make hardware support an OS quality metric. Linux get by with NDA and with direct contributions from the vendors. Which is something the BSDs don’t want/don’t benefit from.
>If you don’t care about administrating your computer and just want to use some software on some hardware, the BSDs are not that great.
Yes this is my opinion also. BSD seems more suited to people for whom fiddling with the OS itself is the point, rather than the OS being a tool to get other things done.
I fall firmly into the latter camp. I'd rather chew glass than manually set flags in rc.conf
I like the word tune rather than fiddle. The BSD are very stable. You adjust some configuration, and then updates without having to change your tools or your config with every release. The config are not provided out of the box but the manuals can be very informative.
A lot of current GNU/Linux complexity have no benefits for most users and may be an hindrance when they want to slightly alter their use cases.
The first column may have valid use cases, but I strongly doubt those cases include casual usage. Simple tools that work well is better than complex tools that solves everything.
* Openbsd does not like containers or being a vm host
I would argue that much of the mentioned zealotry is a sort of kneejerk response to cult-like behavior from some Linux adherents. It’s mostly defensive; these people want continued variety in the FOSS desktop space and feel that’s threatened by Linux.
"Only two remote code execution vulns in the default install" isn't saying much, because the default install has essentially no functionality. Similarly, RCE is not the only kind of vuln.
Let's just say it is not the mainstream consensus that OpenBSD is meaningfully more secure than an up-to-date linux. This may have been true in 1995, but it's generally acknowledged by people who know what they're talking about that OpenBSD's reputation for security is overstated.
> because the default install has essentially no functionality
I dunno, it's got a built in HTTP/S web server and everything needed to be VPN or router.
> Let's just say it is not the mainstream consensus that OpenBSD is meaningfully more secure than an up-to-date linux. This may have been true in 1995, but it's generally acknowledged by people who know what they're talking about that OpenBSD's reputation for security is overstated.
Yes, I've read plenty of vague aspersions that it's totally not as secure as claimed. Since those claims never come with evidence, I'm going with the traditional response: PoC||GTFO.
In my opinion pre alder lake intel is the sweet spot for FreeBSD. Not sure about AMD but anything before 2020 should work just fine. Just avoid CPUs with heterogenous core configurations for now.
I'd say Juana Manso laptops are usable with FreeBSD. sure, you lose brightness control, you can't see how much battery remains, (I didn't try wifi but the 9650AC chip seems to be supported), but it is usable. audio works, USB works, video works when you load the Intel drivers.
Yeah. The reason is that a lot (almost all?) consumer hardware is broken, but in ways that either minimally impact Windows or which are worked around in drivers.
So can anyone give me a short explanation on why someone would use freeBSD over linux? I do run it technically, on my router (OPNSense), but that's not a personal computer, like a desktop or laptop. What are the advantages to running FreeBSD?
For the average joe there it's just an option and personal taste, and it comes with its own tradeoffs and learning curve. Well integrated ZFS would probably be the main thing for average joe.
For developers, it is interesting to think of as a self contained toolkit. If you are building firmware, platform images for bare metal or cloud, it creates a much better demarcation than any attempts Linux can put forth. This is related to why you might like OPNSense. But if you are just a consumer it only indirectly matters to you.. consistency of build and product, quality of subset of network drivers and subsystems like pf to support your mission, ability go in and quickly and correctly fix the right problem at the right level etc.
1. I subjectively just like it better. Things like dtrace, jails, the init system, just click for me.
2. I think it's good to not support a Linux mono-culture. Yes, there is Windows and macOS, but in terms of open source OS's, I think it's good to have more than one choice and so for any rough edges in FreeBSD, I'm willing to deal with them to support that goal.
3. I don't think you'll find any actual, hard, technical reason to want to prefer FreeBSD over Linux on a desktop. Anything you can do in FreeBSD you can do in Linux. Heck, FreeBSD is probably even running the Linux version (for example video drivers).
But really, which Linux do you mean? Nix? Gentoo? Red Hat?
I use arch, but I don't see the distro as mattering that much. For the most part it just feels like picking your poison on which init system, what package manager you want to use, etc, but the end experience isn't too different once everything is set up. I can run steam on each one, they all have gcc, they all have support for the same set of desktop environments, they all have chrome/ff/whatever browser you want. At this point I'm just using what I have set up because the setup is what takes most of the effort, once that's done, it's smooth sailing.
Yes; linux is generally supported better than freebsd. CUDA and Docker work out of the box on linux. Linux has better graphics drivers and steam support. Opensource software (libraries, tools) are much more likely to be tested & work properly on linux. I've also run into several rust crates which don't build on freebsd - particularly crates which depend on C code.
But the comment you're replying to said there weren't many good technical reasons to prefer freebsd over linux. I think that's broadly true.
I still really like freebsd though. Unlike linux, one community is responsible for the kernel and userspace. That makes the whole OS feel much more cohesive. You don't have to worry about supporting 18 different distributions, which all do their own thing.
FreeBSD's development philosophy, it's aversion to design decisions like - we must allow systemd everywhere, stability, zfs and jails, consistent configuration (for decades) are all technical reasons I prefer it over Linux.
How about Ubuntu and snaps? License needed for certain security updates, etc.
It is an old-school UNIX experience, not great for desktops but excellent for long-lived “pet servers” where long-term stability over decades of service is valued. I treasure it for running small Web servers and shell hosts, instead of Debian/Ubuntu.
Same. I've been running it on a "pet" server since the mid 90's, for shell, web, email, etc. I started on FreeBSD 2.x and has been through many upgrades and migrations! I also worked at an early ISP and FreeBSD was our go-to for email, NNTP, and DNS.
Some may call me a devil's advocate (I love FreeBSD as my first alternate OS, it's more sensible as a whole, and less chaotic then linux) it provided the network stack for Microsoft and a base for darwin (OSX) Maybe it's just the low brow art, maybe it's the root of evil and the label is on the tin, how cute!
It's different than Linux, and mainstream enough to be actively developed and used in production. For a hobbyist, that is useful in itself.
It's a pretty rock-solid system, from my memory of the 2000s.
BSDs in general are tightly integrated between kernel and userland tools. FreeBSD has a lot of modern concepts built in that Linux also has, such as Jails and bhyve VM hosting.
FreeBSD has ZFS has a first-class citizen and in my sysadmin opinion, ZFS is one of the best filesystems ever created. While others dunk on BSD for "catching up to Linux" on certain features, BSD equivalents seem to be really well architected. ZFS is one place where Linux (btrfs) is only beginning to catch up. I just learned about bectl (Boot Environment Ctl) that makes snapshotting and rolling back the system partition of installs really easy, and ZFS-on-root is critical to that tool.
For many, having the simple RC init system is a boon over systemd. Services that need started up at boot are defined in /etc/rc.conf, as well as networking and other core services. Editing rc.conf can be done manually or with the sysrc tool.
Ooh! I have lne of those T490 laptops. Except my wife had 1 liter of liquid detergent into it after putting a single bagged carton of it into my suitcase to bring on vacay. Great fun. The screen flickers.
I have the latitude 7490 and it worked great with Linux, FreeBSD and OpenBSD. The only issue is some hardware design issue where lifting it with one hand will cause it to freeze (possibly some stress causing a shock or a displacement).
This happened exactly to me also, I suspect some flexing in the motherboard or other component; right now it is complaining about the RAM and reseating hasn’t fixed it. Great laptop otherwise however!
Consider balling up some electrical tape underneath the Ram stick. This solved this very specific issue with my laptop that was flexing too much and crashing.
Yeah you run into this head on trying to use BSD. It’s too much glue and compat work. By the end of it you no longer have a coherent system, you’re back to Linux.
I use FBSD on an old-ish Lenovo W540 without too many hiccups. No, it’s not for everyone and never was. I wouldn’t suggest to anyone to run a BSD as a daily driver, or at all, unless they have a good reason to. Once you cross that line you need to know what and why.
> Once you cross that line you need to know what and why.
This is counterbalanced by the fact there is often one straightforward solution to every problem you run into, and those have been abundantly discussed online. Written as someone who just gave it a try.
It’s also that so many real-world use cases of BSDs entirely avoid these issues. If they were dominant concerns they would be addressed in a very professional manner, like the rest of the system. But the target market for these things just doesn’t overlap. Maybe there is a market for a BSD-esque approach to solving these things, but honestly? Meh.
So you use FreeBSD on your laptop. You think people with reason should run it. You agree with others who have said you need to know the system to use it, and you do. So why does your response sound like you don't like it? I think you do cause you do.
One of the guys on the FreeBSD forums said, "FreeBSD is for professionals and serious computer enthusiasts." and I don't see anything wrong with that.
I am using it on a 12th Gen Intel (so a few years old), and it works great. I'm sure I have naturally acclimated and subconsciously work around things that I don't realize but suspend/resume works (I rebooted for the first time in like 3 months today). All the apps, yadda. For me, it's great.
Years ago, there was a project combining Debian with the kernel from FreeBSD. That never made sense to me and the project seems to have died meanwhile. More sensible, IMHO, might be to bolt the FreeBSD user space unto the Linux kernel. That way one would get fairly broad and current hardware support and could still enjoy a classic Unix look&feel and stable ABI.
> More sensible, IMHO, might be to bolt the FreeBSD user space unto the Linux kernel.
Chimera Linux is doing something like this, although without aiming for the full BSD experience, just utilising its core userland [1]. The project is in an alpha stage.
IMHO the biggest advantage that Debian/kFreeBSD would have had would be first-class ZFS support. You can use ZFS with Debian today, but the license problem means it only gets supported through DKMS, which is a pain; a FreeBSD-based Debian could ship binary packages for ZFS that just worked out of the box.
Considering the amount of Linux distros based on Debian, it is truly a shame that Debian/kFreeBSD stagnated. We could have had Proxmox/kFreeBSD with support for ZFS Boot environments! The evolution of FreeNAS/TrueNAS demonstrates things have been going in the opposite direction.
Moreover, many laptops working on Linux perfectly, are not Ubuntu certified. Lenovo Legion series generally works well, but it is not in the Ubuntu list. Id we'd make a list of all 8/10 or more compatible laptops, it would be huge.
> More sensible, IMHO, might be to bolt the FreeBSD user space unto the Linux kernel.
A lot of BSD utilities that are not POSIX has really close interaction with the kernel. OpenBSD’s *ctl binaries are often the user-facing part of some OS subsystem. Linux subsystem often expose a very complex internal that you need to use some other project to tame down
Now to be fair, in a few ways I think it is ahead. Now if you said "catch up to Linux in hardware support" I would fully agree.
Last I heard, its VM (swap/memory) processes is still better, but seems many Linux people avoid swap space these days. FWIW, I always have swap on any system that allows it.
And Jails, IMO nothing on Linux comes close to how good FreeBSD Jails is.
Incus is pretty damn good to be fair. You can mix and match VMs and containers, the terraform provider "just works", the setup is fast and easy, it plays well with ZFS. Now I wouldn't be surprised if it still lags jails (or Illumos Zones) in robustness or some capabilities but I'm a happy user of them now.
I use IncusOS when I need the Linux kernel for some specific hardware like the latest graphics or some AI accelerator. FreeBSD for when none of those things are necessary.
I personally feel like the race to support a vast array of hardware is very costly for such a small team and might be a waste of their precious resources.
Of course I love FreeBSD and want it to be supported on my desktop or laptop but at what cost?
Here is the question I have always wanted to ask:
Why not make the ultimate compromise and say: you will be able to run FreeBSD on almost all laptops but it is gonna be through let say an Alpine Linux hypervisor and we are gonna ship it with all the glue you need to have a great experience.
About every CPU has great visualization capabilities nowadays and the perf are amazing.
Now some might start screaming at the idea but you already run your favorite operating system through a stack of software you do not trust or control: UEFI, CPU microcode, etc.
I believe we need OS diversity and if so much of the energy of project is spent on working on an infinite hardware support, how much is left for the real innovation?
I agree. Linux has a wealth of hardware drivers and the time would be better spent on a translation layer or do it via running a VM or even using LLMs to port the drivers over to FreeBSD en masse. That way BSD team can focus on their unique strengths.
I like it for several reasons. It's a holistic system which means it's much easier to understand, not a collection of random parts thrown together. There is only really one (big) distro so documentation is easy to come by and consistent. I love the way the updates of the system are uncoupled from the userland software so you can have rolling packages but a stable OS.
Also the ports collection is great (being able to manually compile every package with different flags where needed). And jails. And ZFS first-class citizen. Also I like the attitude. Less involvement from big tech, less strive to change for change's sake. It feels a lot more stable, every new version there's only a few things changed. It's not that with every major update I have to learn everything anew again because someone wanted to include their new init system (like systemd), configuration tools (like ifconfig -> ip), packaging system (like snap) etc. Things that work fine are just left alone.
It has some really good ideas also, like boot environments. But it's not linux. It's not meant to be.
But yeah if you want everything all figured out for you, don't use FreeBSD. Just take a commercial linux like ubuntu. You'll need to tinker a bit, which I like because it helps me understand my system. FreeBSD is a bit like Linux was in the early 2000s, it mostly works but you often have to dive into a shell for some magic. The good thing is having ZFS snapshots as a safety net though. Never really get caught out that way.
reply