OpenBSD laptops | Hacker News Reader

OpenBSD laptops

tedunangst.com

177 pointsfcambus95 comments

Can confirm that snapshots from December run almost flawlessly already on an X250. Great work OpenBSD team and roll on 5.9.

what is the "almost"?

The devs advise that there are still some problems with suspend / resume while 3D acceleration is in use. I imagine this should be fixed before 5.9 is released.

I also didn't run in to this bug.

Edit: Also didn't test Bluetooth or WWAN.

AFAIK there is no Bluetooth support since 5.6, they dropped the entire subsystem. Found that out when I wanted to use a wireless mouse with my Macbook Air... other than that and the embedded WiFi it works fine, they've done a great job.

I loved running OpenBSD on my laptop. The man-pages were insanely good and the *ctl commands to configure stuff were refreshingly well organized - no more mucking around with a 100 different config files. My major gripe last time was that Firefox was unusably slow. I was told that this was because multi-threading support in OpenBSD was poor. Also, while chromium was tolerable, playing something on youtube would make it feel like the system was under a lot of load. I finally had to move away from it because of this. That said, I plan to move to Vultr and run an OpenBSD box when my current Digital Ocean account runs out of balance.

Almost all laptops these days come with SSDs and last I checked, OpenBSD doesn't support TRIM. Has this been fixed? Is lack of TRIM a non-issue?

I thought SSDs were integrating TRIM into the drive's firmware/controller nowadays.

That's impossible (AFAIK at least) - the drives can't know what parts of the filesystem are in use, only the OS can.

A quick Googling shows it exists only in high-end SSDs, but the information is from 2013.

Just anecdotally, my Nexus 9 does not even have the fstrim command anymore, nor do I see any discard options in the fstab. My Nexus 7 2012 had lots of interest about fstrim improving performance, but not my Nexus 9. I assumed the TRIM issue had disappeared.

Uhh... The nexus 7 and 9? You realize those have SD and MMC cards soldered to the motherboards and work /entirely/ unlike an SSD, right?

Also, the OP is right: TRIM works by telling the SSD which blocks that were previously used are now unused, so that the SSD, at a lower level, can eventually erase the larger flash sector. There is literally no way for the SSD to know this intrinsically because SSDs do not expose a filesystem interface.

Regardless of the differences, the Nexus 7 used trim (introduced in Android 4.3) on it's flash-based (SSD also uses flash) solid-state memory and my Nexus 9 does not. The trim "bug" is well known for some Android devices, but now it has died off. I was attempting to guess why.

Another quick Googling seems to say improved garbage collection (in the SSD itself) has lessened the need for fstrim.

Hopefully someone who is knowledgeable will chime in.

trim is implemented in the controller in that a controller implementing trim is capable of receiving, understanding and acting upon the trim command. It's actually issued higher up in the stack; I've only ever seen it mentioned in the context of an OS issuing the command to the drive if a file on a filesystem on the drive is deleted, or if a utility intended to issue trim against all empty blocks was run.

You're correct in that trim is less necessary if a drive has good garbage collection. I don't know anything about the implementation of either, but I would expect trim to be easier, cheaper and better than GC. It seems like GC would require work to discover/remember things provided as input w/ trim. I've assumed that GC was only implemented by vendors in their controllers because trim wasn't available in the popular OSs yet, and/or because their controllers were already doing most of the work required by GC anyway so it was cheap to implement.

Finally, you can't equate SSDs with the SD or eMMC used in some of the phones/tablets. Yes it's all flash, but an SSD is typically multi-channel with many flash modules, a DRAM cache and a much more sophisticated controller implementing a standard storage interface (AHCI/SATA or SAS or NVMe). eMMC and SD are much simpler, cheaper and slower.

My point was that they are equal from the fstrim perspective.

Not saying that you're wrong, but SD cards for example assume that they're going to be running FAT and make assumptions/optimizations based on that.

Been using an SSD in my laptop for 6 months now... UFS / FFS doesnt really seem to fragment that much

# fsck / | grep frag 6395 files, 106651 used, 409612 free (300 frags, 51164 blocks, 0.1% fragmentation)

# fsck /usr | grep frag 396741 files, 4832173 used, 19820289 free (4953 frags, 2476917 blocks, 0.0% fragmentation)

# fsck /home | grep frag 1192199 files, 27780639 used, 3176062 free (216374 frags, 369961 blocks, 0.7% fragmentation)

The effect of fragmentation on file systems on SSDs is minimal, and fragmentation is not related to TRIM. http://askubuntu.com/a/537211/1624

Due to TRIM bugs I just follow an advise of never using it. Rather I set aside like 5% of disk that is never touched. This works with both old and new SSD without affecting long-term performance on laptops. As an added bonus it makes full-disk encryption more sound.

http://www.tedunangst.com/flak/post/lessons-learned-about-TR... I do believe it supports trim, since 2011. Need to check the code if and how it's implemented today. I'm using 2 SSDs in a RAID-0 with OpenBSD for 1 year now. No problems and lightning fast.

> UEFI boot support means systems that lack BIOS compat will work now.

> ASUS X205TA is an example of a machine with reduced hardware ACPI. Also has Broadcom wifi. I don’t know of a particularly good choice in the netbook category.

The ASUS X205TA is a machine that is hostile to anything not Windows.

It doesn't have bios. It has 64 bit architecture, but a 32 bit EFI. That makes it harder ot install for example some Linuxes.

Here's a 51 page thread in the Ubuntu forums: http://ubuntuforums.org/showthread.php?t=2254322&page=51

Here's someone trying to write a driver for sound: http://comments.gmane.org/gmane.linux.alsa.devel/138822

Here's the guide to getting debian on it: https://wiki.debian.org/InstallingDebianOn/Asus/X205TA

Here's the (as usual, excellent) guide to getting Arch on it: https://wiki.archlinux.org/index.php/Asus_x205ta

I guess people using BSD are happy with installs that are more complex than "burn an image to a USB stick; boot from that USB stick; click a button". But here's an example of someone installing Arch: http://ifranali.blogspot.co.uk/2015/04/installing-arch-linux...

It's a shame it's so hostile because, considering the price, it's a nice enough machine. (The screen is lousy, and the sound is odd. If you liked the EEE PC 701, well, this isn't as good.)

OpenBSD doesn't use GRUB. OpenBSD's EFI loader is a port of the existing second-stage loader, boot(8). It can load both i386 and amd64 kernels, and there are EFI binaries for 64-bit and 32-bit firmware.

"I guess people using BSD... burn an image to a USB stick..."

I always thought "burn" referred to optical storage like CD-ROM?

Not sure about others, but when I do BSD installs I do not have to click any buttons. I do not need a mouse.

I can lazily use something as small as an 8M stick for the install image. It couldn't be much easier.

With the size of today's RAM (500M+), using the install image I can make the custom image with my personal kernel configs, utilities, and settings entirely in memory without needing to write anything to a HDD. Then I just dd the image to the removable media. That is the slowest step.

My custom images are quite small too, under 16M, and need no access to a HDD. The system runs just fine in RAM, no swap.

I guess I could do the same setup with "Arch" but I imagine it would take considerable work undoing pre-configurations based on assumptions about how I would want my system configured.

Easier perhaps to just compile my own custom Linux kernel and "initrd". Build from the ground up.

I like the flexibility and control with BSD. Not to say Linux does not have flexibility and control, too. It certainly does. I just rarely see ordinary Linux users making use of it when installing. Instead many if not all decisions have already been made, and everything is pre-configured. "One click..."

Kudos to the OpenBSD folks for the network drivers.

Can I PXE boot the OpenBSD image? I know when I looked it up for FreeBSD the information was quite sparse.

http://ftp.openbsd.org/pub/OpenBSD/5.8/i386/INSTALL.i386

Also, if you learn to use installboot(8) you can make your own PXE bootable images.

I would read the Makefiles in the OpenBSD source tree to see how they make their install images.

Note I'm not an OpenBSD user and have not installed it in many years so I'm not the best person to ask. I use an earlier distribution, the one that spawned OpenBSD. Similar but not the same.

PXE booting OpenBSD is covered in the FAQ: http://www.openbsd.org/faq/faq6.html#PXE

Yes, all you need is `pxeboot` and `bsd.rd`. Put them in the root of your TFTP server's directory, have your DHCP server hand out `pxeboot` as the filename (and rename `bsd.rd` to `bsd` on your TFTP server) and point it towards your TFTP server, boot via PXE, and install.

You can mirror the install sets locally (on a HTTP server) for performance reasons, if you like; otherwise, you can just download them from an HTTP mirror during the installation.

I've been working on automating the whole process (fully-automated installations for bringing up new VMs) and it's so much simpler than with other operating systems.

>Lots of fixes to USB and and xhci mean USB 3 devices should attach much more reliably.

Does anyone know how much code FreeBSD and OpenBSD share? FreeNAS uses FreeBSD 9 and the forum is allergic to any mention of xhci. I've got it working on a Atom C2758 system but it runs at USB 2 speed.

IIRC, they both forked the USB stack from NetBSD a long time ago and haven't really remained in sync since. The XHCI drivers seem to be independent implementations.

Damnit. Looks like I'll be trialling FreeBSD 10 and OpenBSD 5. If I had a good alternative to ZFS I'd jump back to Linux before loading too much data on to this NAS. FreeNAS 9 is like going back to the 90s.

Speaking about NetBSD.

I never see any news about NetBSD and hardly any about FreeBSD. Though quite a few about OpenBSD. Why?

We borrow a lot, however Freebsd tends to like to do things as library layers. e.g. libgraph, and other things we don't do.

Does anyone else just run Linux (or OS X if you're on a MacBook) and then just run OpenBSD in a VM? Run it fullscreen and can't even tell. You get all the advantages of OpenBSD plus all the advantages of using the regular host OS.

I'd love to run OpenBSD on my laptop computers, but last time I tried (5.7), the fan on my Dell Latitude 6410 was permanently on, while on FreeBSD with performance_cx_lowest="Cmax" the fan barely runs at all.

As much as I love OpenBSD, this is a knock-out criterium to me. Has the situation improved significantly since 5.7?

Did you run apmd with the conservative option?

CPU power management support has improved in 5.8. Better ACPI support and the drivers now use more lower-power C states.

"acpicpu(4) uses ACPI C-state information to reduce power consumption of idle CPUs. "

http://www.openbsd.org/58.html

That's great to hear! I'll definitely give it a try!

What C states does it currently support? My FreeBSD box spends most of its time in C8.

As I understand previously only C1 was used. The range has been expanded now though you'd have to read the code to see exactly what it's doing.

It seems to produce acceptable power draw figures on a ThinkPad X250 in any case.

> There’s also some very preliminary support for HID over IIC, for systems that attach keyboards and touchpads over an I2C bus.

Keyboards & trackpads over I2C? I liked the trend of moving everything to USB, even for internal devices.

HID over I2C specification was created by Microsoft for external devices connectable to their Surface tablets. IIRC they claimed that it saved power because of usage of simpler controllers.

Regardless of what it was intended for, my Skylake laptop has its touchpad set up using HID over I2C. It works on legacy systems using some kind of PS/2 emulation, but it works much better using I2C.

Everything old is new again. One of the original intents of I2C was just that -- a universal bus for semi-low speed devices, including keyboards. USB just happened to come around at the same time and defined a physical spec along with an protocol, and the battle was done, with I2C relegated to embedded devices and things like querying monitors for supported features/modes.

Why? USB adds tons of complexity/overhead relative to what HID devices actually need -- especially hardwired devices.

USB - not Universal, not Serial and not a Bus

USB isn't the right choice for a lot of interconnects, or even most interconnects, especially things like HID for mobile devices. One of the big advantages you can have with I2C is that because it's multi-master, both sides of the HID chain can become a lot simpler. The host side doesn't need to send occasional messages asking if the various HID devices have data, and the device side doesn't need the more complicated logic to handle USB HID events. Instead, the HID device can send HID events as they happen, and the host only needs to send the much less common host communications -- for commands like turning on and off lights, etc -- when those happens. This means a bit deeper sleep, and as such, less power consumption.

> There are a few minor quirks, but generally it works well

> As ever, support for Broadcom wifi and Nvidia graphics is nonexistent.

This is exactly the big issue with adoption of any *nix system on laptop or desktop. No user wants to have to live with quirks. Users just want their system to work.

- This mid-2013 Mac Pro wakes up every few hours at night, even though I have disabled every "wake from sleep" setting that exists in OS X

- My mid-2013 MBP has, several times this week, turned on while closed, and continued as if it was opened (playing YouTube videos or whatever else it be doing), only shutting back off after about 120 seconds or so

- Every time I turn this Mac Pro on, it doesn't recognize the wired Apple keyboard that's plugged into the Apple Cinema Display, and says it's looking for a bluetooth keyboard, until I unplug and replug the keyboard in at least 3 or 4 times

- My MBP once emitted a (very) loud buzzing sound from its speakers, for absolutely no reason, that lasted about 3 or 4 seconds, startling everyone nearby, when no sound-based programs were running

- My wife's Surface Pro 3 continually needs to be restarted because touch-screen controls simply stop being able to open programs on the home/grid page.

- If our family ASUS $800 computer is left on and falls asleep, it won't wake up, and will need to be manually shut down and restarted

These are just off the top of my head. Point is, Microsoft/Apple reliability is a complete myth. I'm tired of it and I'm tired of throwing extra money away for the sake of "higher reliability" and "fewer quirks".

THANK you!

I'm tired of the cult of perfection surrounding frankly middling proprietary operating systems.

- Every mac I've ever used has had Launchpad UI bugs requiring me to try several times to launch things.

- OS X has woeful graphics drivers, to the point where Mesa is doing better.

- My Air got erratically unusably laggy until I finally gave up on it.

- My MBP will occasionally forget that it has any integrated input devices, requiring me to either find a USB keyboard or reboot.

- Microsoft still hasn't resolved their graphics driver stability situation; blue screens are far from a relic of the past in my experience.

These are anecdotes, yes. Sure, some (rare) people might have a perfect experience with OS X or Windows. But the point is, free operating systems don't have a monopoly on driver and spitshine issues.

Windows 10's UI is pretty half-baked, as well. I mean, the UI concepts aren't bad, but they're only applied to half of the OS, with everything just below the surface still using the old-fashioned dialogs that've been around since Windows 98.

They're converting stuff to Windows Runtime. There's a lot of it. It will take a while....

Your anecdotal experience runs so completely contrary to mine that I don't even know what to say.

My MacBook works. It just works. No tinkering, compiling, fidgeting, anything -- and this is precisely why I like it and will continue to use it. I don't want my hardware getting in the way.

And reliability is certainly not a myth. In my days I've supported all three platforms and by far it was MPB that gave me the least grief. This is doubly true with hardware. Linux is close behind Mac but I think most engineers brave enough to venture it would competent to deal with problems on their own.

When a MPB has a problem: take it to a Mac store and get a replacement. When a Lenovo had a problem? Ship it to Virginia on your dime and wait 6 weeks.

I seriously suggest you take your MPB to a mac store and speak to them about your problem.

What you in effect just did: "Your experience is anecdotal, so now I tell you that Reliability is not a myth based on my anecdotal experience."

The point he wanted to make is most likely not that all Macs suck all the time, but that Mac does not automatically mean reliable or more reliable than any other OS/system. So while it might not be worse than other systems (I don't want to make a judgement based on the few Macs I had to fix for friends without ever having used them myself), it certainly is no super-reliable saint either.

And my limited first-hand experience agrees with him.

Additional interesting anecdote: my brother and I bought identical macbook pros at the same time. I generally didn't have much of a problem with it. Him? Well, let's say it was barely more reliable than the windows laptop he had before. Powering down wrong, failing to identify displays, color gamut screwups, power issues, boot problems. Mostly transient, but very real, failures in both hardware and software.

> When a Lenovo had a problem? Ship it to Virginia on your dime and wait 6 weeks.

When I've had problems with US purchased Dell's, I had someone there within 48 hours with a replacement piece, despite being in 1) Austria and 2) Italy.

Lenovo support has been great to me as well. With my second-hand machine, a year out of warranty, they paid to have my laptop shipped to them for a keyboard and possible motherboard replacement for free. This is after I'd damaged the keyboard by intentionally testing, "How well DOES the water drainage system work?" by pouring water into the keyboard. I was frank with them about what happened, and it did not matter.

Care to elaborate on what happened when you poured water onto your laptop? Which model was it? Do you know why it failed?

I had a run-in with Lenovo support last year. Three-year warranty, two and a half years old, fried motherboard caused by the plastic in the power socket coming apart. Not sure how, don't care - it's wear-and-tear. On initial calls they refused to honour the warranty. I went through the complaints procedure and eventually they relented. It took six weeks but they covered all costs and completely fixed it.

In the meanwhile I'd got a macbook so I could continue working. The mac was driving me crazy. The keyboard feel, the trackpad feel, the lack of nipple mouse, the every-changing UI that never does quite what I want it to.

When the Lenovo came back I tried it out, fell in love again, and moved back to it. At some point the SSD will die and I'll get another one. There's nothing close to a Thinkpad. To me, they feel like they built as tools in a way that other devices lack.

I once had a 1st gen Chromebook wake up in its bag near my bed and resume playing a YouTube video at 2am. I forget what the video was but it was rather creepy at first.

Is your apple display an actual Cinema display or one of the newer Thunderbolt displays?

I've noticed some issues with the USB hub that is built into the Thunderbolt display whereby for some reason it doesn't enumerate correctly every 40 - 50 plug/unplug cycles while using my MBPr at work. It only happens when the Apple keyboard is plugged into the display, with alternate keyboards it doesn't happen.

So I personally feel like it is an issue with the Apple keyboard.

I don't know, but it shouldn't matter. Apple products are known for how they Just Work™ but in this case it's not true.

On the other hand, my desktop that's been dirty-upgraded from Win 7 to Win 8 to Win 10, and is hand-assembled from components I didn't necessarily check too closely for compatibility has uptimes regularly measured in weeks, despite running multiple instances of Visual Studio, a full SQL Server instance, piles of long-lived Chrome tabs, and whatever crazy stuff Steam and my Steam Link are doing.

I'm not sure I'll ever buy another OEM PC. So much crapware comes pre-installed, and it's so difficult to get rid of it, that I usually wipe the disks and install Windows from an MSDN iso. Of course, there's still the drivers... And more importantly, the tens and hundreds of MB of shitware that comes bundled with a tiny device driver. No, I do not want the ASUS Smart Keyboard Whoop-dee-doo application launching at startup. I just want my keyboard to work like every other keyboard has worked for the last 30 years.

Did you disable just wake-from-sleep or standby mode using pmset utility? In my case only after I disabled standby completely both MacBook Air 2014 and MacBook Retina 2015 stopped waking up under the last 3 versions of OSX. I suspect the bug depends on the software one uses that is why only some users are affected.

This has nothing to do with operating systems, and everything to do with hardware makers. If you design hardware for a particular OS, it will work very well with that OS and poorly with anything else.

So for instance: Every Chromebook out there is running a Linux kernel, and they all have exemplary hardware support -- power management is great, trackpads are phenomenal, there are no issues with sound or wifi or GPU. BUT if you try to install Windows on them, it's super-awkward and hacky if you can get it to work at all.

For historical reasons, most people who try to install BSD or Linux are doing so on hardware that was designed for Windows, with components that were selected on the basis of how well they work with Windows, and driver work that is done for Windows. This does not lead to great outcomes, any more than buying a Chromebook and using it as a Windows machine would.

So in a sense "hardware makers aren't making hardware designed for Linux (other than Chromebooks) or BSD" is a problem for adoption of Linux/BSD, but it's a problem that can only be fixed by hardware makers.

There's a social component as well. If a hypothetical user buys a device with a proprietary OS (eg MS, Apple, or stock Android), the blame for any quirks falls squarely on the manufacturer. The user then copes by blaming themselves for buying that manufacturer, not spending enough money, etc. The solution is to simply deal with the quirks until purchasing a replacement thing can be justified. This phenomenon should be readily apparent to anybody who's ever inherited perfectly good hardware from someone who was sick of their computer and just wanted to get a new one.

But installing something that the manufacturer didn't supply is a deliberate choice. Any quirks get blamed directly on the changed software, even if there are actually fewer of them. The change becomes the issue.

> But installing something that the manufacturer didn't supply is a deliberate choice. Any quirks get blamed directly on the changed software, even if there are actually fewer of them. The change becomes the issue.

This rings true for me very much this week. Even more fun when there are multiple changes, but rather than examine the situation and determine objectively which change caused a problem, whichever change is the biggest magnet for FUD gets blamed by default.

This is so true. If you plan to run Linux or BSD, you should better cherry-pick hardware---especially laptops.

E.g., Linus himself was using a MacBook Air 11 late 2012 for some time. It works fantastically well on Linux. Newer MacBooks are sometimes not so well supported. Only thing lacking is no recognition of ACPI events to flag each time the battery discharges by 1% or below a threshold. But very few laptops have this.

I also don't trust a position that considers nonfunctional WiFi a quirk.

WiFi not working at all is a different matter, but not working with a specific wireless card is a quirk.

Mini PCI Express is a standard (as was Mini PCI before it). Swapping out a laptop wireless card you don't like isn't that hard. I've done it several times (to get dual band support, for better Linux drivers, etc).

Spending around $100 (including installation if you don't want to do it yourself) may be annoying, but it shouldn't be a deal breaker.

Some laptops check wifi minipci cards against a PCI ID device whitelist in the BIOS. For instance, all Thinkpads do this. It's seriously annoying and means that the OS must support a very specific set of devices or wifi won't work. For some machines a hacked BIOS can be flashed to disable this check but that's not always available.

Apart from reverse-engineering yet more hardware and writing more drivers, there is not much the OpenBSD project can do about this.

Fixing the root cause would be best. Vendors who do such things must be convinced to stop doing them.

If vendors also released freely available documentation for the hardware (read: freely accessible manuals for driver developers, no NDA required to read) then end users of free software could live in a world where wireless not working would be the exception.

You can buy supported hardware. Dell makes decent Ubuntu laptops for instance.

I would guess a $30 USB wifi adapter fixes it for most people.

I don't think there's an issue with OpenBSD adoption.

OpenBSD attracts the type of users they want to support I imagine. Leave it to someone else to push the "Year of *nix desktop" agenda.

Don't tar all \*nix with the same brush. Broadcom wifi and Nvidia graphics work fine on FreeBSD. OpenBSD makes particular tradeoffs; if you like them use it, if you don't, don't.

Unfortunately FreeBSD doesn't even have working sleep support, and hardly even a nod to power saving, whereas OpenBSD supports this automatically on all the laptops I've tested it with.

Makes no sense that the more niche, security focussed OpenBSD has the much better laptop support, until you consider that OpenBSD devs actually dogfood it on their laptops.

Many FreeBSD devs seem to run Macbooks for development, using FreeBSD only in a VM or via SSH.

> No user wants to have to live with quirks. Users just want their system to work.

I agree. Which is why I'm so frustrated at the moment, as I've had issues with my 1 year old rMBP, my girlfriend had issues with her Windows laptop, and my Ubuntu install on my HP 11" runs into the weirdest bugs.

It's 2016, and nothing works. I don't understand it.

Everything is Broken[1]

[1]: https://medium.com/message/everything-is-broken-81e5f33a24e1

I've never tried OpenBSD, but it seems to have better laptop/desktop support. Maybe I'll fire up an instance in a VM tonight and play around a bit.

Is there anything significantly different right off the bat from someone coming from the FreeBSD world?

Only binary patches for base provided by a third party (M:Tier, where a number of OpenBSD developers work).

I haven't run FBSD in years so I can't comment but I highly recommend OBSD for laptops (obviously with supported hardware).

I set up an OBSD laptop, an ancient Dell Inspiron 8000. Based on the experience, I'd like to get a lightweight, cheap, relatively modern laptop. Any recommendations? Cheapest thinkpad I can find?

Did you read the article? Ted specifically mentioned a few laptops by name.

Thinkpad x200's are perfect in my experience and dirt cheap

X201 - swap the HD for an SSD and enjoy.

Unless you want to enjoy Bluetooth of course :)

Do they still exclude all 3rd party binaries?

What 3rd party binaries? You mean the ~9000 packages?

blobs? Yes.

Yes. This is why no support for Nvidia or Broadcom.

Don't they ship the free Nvidia Mesa driver either?

From the base system, they exclude any proprietary code & binary blobs.

There are many 3rd party packages in ports.

Up and running on a 366Mhz Pentium II based Dell CPi. Using w3m-image through FVWM for browsing.

5.9 is going to be amazing. I've been running current and with the patches from Bitrig's port of Netbsd features, we now have a working dtrace/zfs. No more fsck! Dtrace lacks the pledge(2) probes, but hope that gets fixed soon. Otherwise it's nice to have alternatives to desktop linux. Only miss Spotify. oh well.

NetBSD ZFS has rather severe bugs, was this addressed by bitrig? They have remained unfixed in NetBSD for quite some years now.

    Known Bugs
    
        vnode reclaiming deadlocks
        vnode fsync bug
        kmem_cache_alloc/ aka pool_cache_get panics

https://wiki.netbsd.org/users/haad/porting_zfs/

Should also be noted that dtrace and zfs can never be actually integrated into OpenBSD due to them refusing to import CDDL licensed code.

Last edited 3 years and 9 months ago That sounds authoritative. The CDDL is an anti forking license. It requires spooning.

Bitrig does not have support for ZFS. Must be something different from Bitrig the poster is talking about.

Is there any reason not to run -current?

Why sit around six months out of date? Packages stop getting updated. With current that's not an issue.

Which patches from Bitrig are you exactly talking about?

Reply on news.ycombinator.com