You won’t find nearly as many online resources about setting up OpenBSD, because honestly, you really don’t need any. Unlike much of Linux and FreeBSD, the included manuals are high quality, coherent, and filled with practical examples. You also need very little third party software to do basic tasks – almost everything you need is well-integrated into the base system.
You’ll notice that many features that require toil to achieve on FreeBSD, such as suspend on lid close, working volume buttons, and decent battery life, work out of the box on OpenBSD. You can tell the developers actually use this thing on their personal devices.
And while the official OpenBSD FAQ has all you need to get an installation up and running, it takes a bit of grinding to massage the base installation into a seamless laptop experience. So, I wrote this guide to give you a jump start. Things should just work as long as you have a non-bleeding-edge, semi-mainstream laptop, but ThinkPads are your best bet. Enjoy!
Unfortunately AMD graphics support is lagging quite a bit behind Intel on OpenBSD perhaps that will change now that you can buy decent AMD laptops with integrated graphics again.
Not keeping my hopes up unfortunately.
The most support it will come from them it will be for their latest line of products.
Even on Linux and Windows AMD support hasn’t been stellar for older products from their lineup (including their APU lineup).
No luck for AMD-based Thinkpads.
OpenBSD also doesn’t have TRIM support which is a must-have on many newer laptops (and desktops) these days for their SSD. There’s also no power management support (yet? unless that changed in 6.4) for Ryzen systems. This keeps it off my main desktop (Ryzen 5 1600 w/ SSD) and my main laptop (has an SSD).
kills the deal.
Theo De Raadt.. is a well known ass. But man he writes some good clean code. I don’t really have any qualms about saying OpenBSD is one of the cleanest well documented Unix OSes in existence. It’s a pleasure to install and setup as well.
You realize that the only reason this clean OS exists because of a clash some 23+ years ago with the NetBSD project of which he was one of 3 founders… again over personal conduct, if they had gotten along instead of ousting him NetBSD would probably be even better than it is. However that isn’t to say that the current situation is bad either as the projects actually do share code and drivers, maintaining a sort of amicable separation in development however.
That leads me to conclude that CoC are useless… for several reasons, 1. They wont change people, 2. they will be abused just as any political power is, 3. Just like licenses they are a distraction from actual development, 4. most CoC ban violator in such a way that future cooperation is impossible 5. society itself is already an effective CoC… basically mob justice but if you act like a jerk for no reason or bad reasons people will call you out on it, but if you act like a jerk for reasons they can relate to… perhaps you have a point. CoC set in stone and administered by the biased are nothing but censorship.
Edited 2018-11-04 05:44 UTC
Why?
I fixed it for you
For me: not being distracted by “Social Justice” is a strength for an open source project.
Martin Luther King said something along the lines that people should be judged by their character and not their skin color. IMHO open source should be produced on the basis of the quality of the code rather than any snapshot of someone’s private political perspective.
Open source is intrinsically equal opportunity and most projects (AFAIR) are immensely grateful for contributions from anyone. In this context good manners are essential; and of course every BDFL will make decisions that are unpopular decisions from time to time.
Being color blind in practice means ignoring the causes that lead to the injustice that still persist. Its like saying the only rule of a race should be the one that crossess the finish line first.
Seems fair right?
But the *only* rule? The guy strapped into a dodge daemon, will probably beat the guy that has ankle chains on every time. What would happen if they rules were modified to say that they had to use their own leg power and without any leg chains on?
I think those are the rules that everyone wants and they think that’s what they are saying by ” color blind”. But it usually means they are prohibiting looking for leg chains or cars.
Frankly, while I’m all for diversity, this CoC thing really does seem like it’s aimed at getting a clique of incompetent coders into good/powerful positions within influential open source projects. If I was Theo (or Linus) and headed a project that’s essentially my life’s work, why would I hand over power to some group of people who can arbitrarily say what kind of behaviour is accepted or not?
Edited 2018-11-04 14:32 UTC
“COC”? Code of Conduct?
Interesting. But the lack of trim support is a deal breaker IMHO.
Meh, doesn’t bother me as I’m A) using a modern SSD with built-in wear levelling B) mount the ffs filesystem with noatime, softdep
Also, here is an older blogpost from one of the devs:
https://www.tedunangst.com/flak/post/lessons-learned-about-TRIM
As fffx said, wear leveling doesn’t replace TRIM. They don’t address the same physical condition. They’re meant to work in tandem on modern SSD drives.
Wear leveling makes sure all cells wear at similar rates depending on storage patterns. It lowers the chances of data being stored in a dead or dying cell and therefore lost. This is a somewhat similar idea to legacy HDD firmware transparently moving data out of physical sectors starting to go bad and marking them unusable.
What TRIM does, however, has no mechanical drive equivalent. On NAND flash devices, TRIM marks the cells available for reuse as they can’t be used again for new data until they are marked physically free. Early SSD drives lacked TRIM support and their lifetimes were short. TRIM extends the lifetimes of modern SSDs considerably. Since OpenBSD doesn’t support TRIM operations, the lifetime of such devices is much shorter than it would be otherwise. It’s true that wear leveling (assuming the filesystem doesn’t use the entire device) can partly compensate for lack of TRIM, but you’re still artificially limiting the drive’s lifespan.
There’s also a security aspect to TRIM that should appeal to the truly paranoid (oddly enough… that SHOULD be OpenBSD’s central audience!): on some (many? -depends on firmware and how GC is implemented) SSDs the TRIM command flushes all cells that have invalid (deleted/erased) data making it entirely unrecoverable even by aggressive forensic methods.
For a quick rundown on TRIM and why it’s needed, use considerations, and performance impact on SSDs you can just take a look at the wikipedia article.
https://en.wikipedia.org/wiki/TRIM#Flash_drive_specific_issues
Any idea as to what’s keeping it out of OpenBSD? TRIM has been around in the wild for quite awhile now, hasn’t it?
They made two attempts, failed, and decided it was too difficult.
The devs stated that TRIM interferes with their disk encryption and/or they just don’t seem to care?
I tried using OpenBSD but they have issues with nvidia cards. Linux worked much better for me.
this is all well and good, but since ZFS is not even planned because of some bollocks about licensing, the whole thing is a lot less appealing, however clean and well-structured it is.