FreeBSD Foundation Buying Newer Laptops To Help Improve Hardware Support

[kpedersen]

This 2018 mini has outstayed its welcome and has to go. I need to recover the money I spent on it so I can do something more meaningful with it.

Haha, OK, I understand completely!

Also agree with the user who mentioned the Pinebook, would be great to see OpenBSD or FreeBSD working on it.

In principle I agree but if you try to actually buy a Pinebook, you'll see what I mean. It is basically vapourware. It doesn't exit. They are unavailable to purchase (and probably always will be). Just read all the red bollocks here: https://store.pine64.org/?product=11-6-pinebook
I probably sound harsh. I really do wish it existed!

<broken fix, needs reverting>

The latest in Linux news is here: https://www.phoronix.com/forums/foru...the-sgi-octane
(Loongson and SGI).

https://www.openbsd.org/loongson.html

BSD has had this support since 2010

https://www.openbsd.org/sgi.html

BSD has had this support since 2009

You are right. I take it back; BSD often gets support for hardware long before Linux.

[starshipeleven]

BSD has had this support since 2010 (longsoon)

So you are telling me that BSD supported the Longsoon 3A4000, a CPU that was released this year, since 2010? Is time travel a thing for BSD developers?

May I remind you that for Longsoon the OS offered by the vendors is Linux? What you see in that article is same-year mainline support of modern hardware for Linux.

BSD has had this support since 2009 (SGI octane)

That hardware was long obsolete by then, the Octane II was 6 years old in 2009, which is A LOT for hardware of its era. Other SGI workstations were even older than that.

[aht0]

Why the hell thinkpads? Thinkpads are the best laptops for freebsd since ever. Why the hel the dont buy an Asus, Dell, HP, Acer?

I've had problems with nearly all of these, excepting Asus, problems on Linux as well.
Dell - one 17" model (Inspiron 17R) I have cannot shutdown/reboot/suspend properly even on Linux 5.x kernels. "Reboot" performs actual shutdown and that's the only way to close the machine when it runs Linux. Works as expected on Free-/DragonflyBSD's.
Acer - one I have has weird UEFI implementation that does not like neither Linux nor BSD.
HP - device whitelisting (sometimes neither Linux or BSD cannot recognize branded cellular modems), shitty sort of build quality often.
Asus (3 models so far I've owned) have had the best compatibility (even UEFI-wise) but RMA may be problematic when you bought yours from abroad or store you bought it from happened to go under meanwhile. Asus does not deal with clients RMA-requests direct.

Good Linux compatibility among laptops is largely a myth. It's better but in no way comparable to Windows hardware support. Sometimes it's fixable (swap out cellular modem/WiFi card against supported model, manually edit some config files disabling/reconfiguring stuff by hand), sometimes it's not (goofy shit with ACPI, UEFI etc)

[aht0]

So you are telling me that BSD supported the Longsoon 3A4000, a CPU that was released this year, since 2010? Is time travel a thing for BSD developers?

May I remind you that for Longsoon the OS offered by the vendors is Linux? What you see in that article is same-year mainline support of modern hardware for Linux.

That hardware was long obsolete by then, the Octane II was 6 years old in 2009, which is A LOT for hardware of its era. Other SGI workstations were even older than that.

You are just plainly trolling here. 3A4000 is anything but theoretical only when you can somehow get your hands on it. I couldn't see it even on Loongsoon's Chinese site. General support for BSD is there, if some particular model works or not - and if it can be made if it didn't - gets clear after you somehow get one. Little-endian is no issue in itself.

[k1e0x]

I had to buy a five-year-old HP desktop machine in order for FreeBSD to work well.

Ya right.. all I can say is you're doing it wrong.

FreeBSD runs on most higher end Intel stuff just fine. Pretty much any common U1 (like supermicro) you find it will run on. I haven't seen an Intel server it *didn't* run on. (and most cloud providers provide images for FreeBSD)

For workstations you have graphics cards to deal with. Most Nvidia cards work with the proprietary driver support for FreeBSD is provided by Nvidia. For AMD Polaris works great, Vega and Navi I'm not sure about YMMV. (Linux has trouble with some of these too). for AMD CPU's Ryzen works including the power states.

So.. in the end here if you have widely available mid to upper tier hardware you're fine, but the more enthusiast exotic stuff you might run into problems. (your l33t water cooler or rgb fan controller isn't going to work.)

For desktops occasionally you'll run into oddities.. tho usually there is a kernel parameter you can tweak to work around it. (FreeBSD has a ton of kernel parameters for system tuning)

For Laptops.. well.. That is where you really start to run into problems. But the same is true for Linux, tho there are many great Linux laptops and even laptop vendors. OEM's make tweaks to the hardware and it's often undocumented. I have two Laptops, they run MacOS and Windows because I don't want to be bothered trying to deal with getting Linux running right on them.. I've been down that road before. Even if I did want to do it and hardware compatibility wasn't an issue.. I think I'd prefer OpenBSD on a portable over FreeBSD or Linux anyhow. FreeBSD is just to me.. a server/workstation OS and it's very good at that. Being extremely good in one role is where I think the project should focus.. instead of splitting it's limited effort to support every use case.. but that isn't up to me.

As such I'd like to see FreeBSD focus on becoming a really great server for storage and other OS containers. Focusing on OpenRC, pkgbase, bhyve, jails, linuxulator (to run docker on metal in a jail) and improving performance, ZFS and PF. It's positioned pretty well for that role already and this is the role enterprises tend to use FreeBSD for. A cheep Netapp or modern Solaris replacement for a stand alone server usually.

[willbprog177]

Ya right.. all I can say is you're doing it wrong.

I wouldn't be the first time!

In all seriousness, I wanted to make sure FreeBSD would work well without a lot of drama. Plus the machine wasn't too expensive and has 'just right' performance for me, even if it's compiling large projects or doing CPU-intensive stuff.

There is still a three or four-year-old AMD video card just sitting in my drawer because FreeBSD won't boot all the way with it in. It's frustrating because it's not some card that just came out last year, it's three to four years old and you'd think the devs would have had adequate time to get it working. So I'll just use the integrated Intel 4600 HD adapter until whenever the newer AMD card is supported.

I'm not bashing FreeBSD, just disappointed in some areas. It's my day-to-day OS and works pretty much fine for most things (but won't run Minecraft :'( ). I know the FreeBSD devs have more work than they can handle, so I'm not raging and bothering them with dozens of bug reports for the things that don't work. If I was a better dev, I'd certainly join their team to help out.

[aht0]

There is still a three or four-year-old AMD video card just sitting in my drawer because FreeBSD won't boot all the way with it in. It's frustrating because it's not some card that just came out last year, it's three to four years old and you'd think the devs would have had adequate time to get it working. So I'll just use the integrated Intel 4600 HD adapter until whenever the newer AMD card is supported.

You may meet same situation on Linux as well, at least with Polaris-based cards. Should be supported but really card refuses to work. For whatever reason.

[starshipeleven]

You are just plainly trolling here.

A dumb ass is comparing a patch where someone contributed support for Longsoon 3A4000 RIGHT NOW in Linux with generic support for (a possibly older) Longsoon arch in BSD, and claiming BSD has better support for this hardware.
And I'm the one trolling if I point out that he is out of his mind.

3A4000 is anything but theoretical only when you can somehow get your hands on it.

The hardware vendor itself is writing the driver, you tool. The vendor's hardware samples and test boards are not theoretical.
Intel and AMD hardware usually merge hardware support before the product is even released, for example.
And it makes sense to have Linux support out asap if your main OS is Linux. Longsoon systems are sold with Linux on them.

Comparing this level of support with something that was added by someone in 2010 is not fair in the slightest. This is first class OS support by the fucking hardware vendor, for crying out loud.

[aht0]

The hardware vendor itself is writing the driver, you tool. The vendor's hardware samples and test boards are not theoretical.
Intel and AMD hardware usually merge hardware support before the product is even released, for example.
And it makes sense to have Linux support out asap if your main OS is Linux. Longsoon systems are sold with Linux on them.

Comparing this level of support with something that was added by someone in 2010 is not fair in the slightest. This is first class OS support by the fucking hardware vendor, for crying out loud.

When you have no way of getting this piece of hardware, then all the hardware support existence or not-existence remains effectively theoretical. It's like some truebeliever bragging about the miracles his God's prophet supposedly did 1500-2000 years a go. It has no plausible value, except bragging rights.

[starshipeleven]

When you have no way of getting this piece of hardware, then all the hardware support existence or not-existence remains effectively theoretical

I'm not the one that pulled that Longsoon to use as example to disprove my point, so yeah, the more you break his statement the more you are helping me telling him he is a moron and that I was actually factually right.

That said I personally find questionable your nitpicking, there are no indications that this is a paper launch.