Coreboot 4.22 Released: Initial AMD OpenSIL Code Added, 17 New Motherboards

[phoronix]

Phoronix: Coreboot 4.22 Released: Initial AMD OpenSIL Code Added, 17 New Motherboards

A new release of Coreboot is available today as the increasingly popular open-source system firmware solution that's used by Chromebooks, increasing hyperscaler / data center industry interest due to increased code transparency and security, System76 laptops, and more. Coreboot 4.22 is the new release and brings initial AMD OpenSIL code integration, 17 new motherboard ports, and more. Coreboot 4.22 will be succeeded next year by Coreboot 24.02...

https://www.phoronix.com/news/Coreboot-4.22-Released

[Retramon]

I understand the priorities and enterprise investments, realities of open-source struggles, manufacturer cooperation, but seriously Desktop and HEDT support when !!!! This project needs a big boost. Donation, programmer help?

Usually it's the overclockers who want to tweak all kinds of knobs and a more flexible boot system wouldn't hurt, but ofcourse it's a form of entertainment ... otherwise there's still a lot of unrecognized, err rather I would say overshadowed serious work done on Desktop and HEDT, but in the mind of a lot of companies it feels like it always get's the third-class just-a-gamer treatment.

I'm being a bit ranty in general with the tone here, it's late, like first thing on my mind, obviously it's a more complicated topic, and eventually all of this will come to the Desktop (...rrrrright?)
I love the news but, where's the support for stuff that matters*!

[ayumu]

I understand the priorities and enterprise investments, realities of open-source struggles, manufacturer cooperation, but seriously Desktop and HEDT support when !!!! This project needs a big boost. Donation, programmer help?

Usually it's the overclockers who want to tweak all kinds of knobs and a more flexible boot system wouldn't hurt, but ofcourse it's a form of entertainment ... otherwise there's still a lot of unrecognized, err rather I would say overshadowed serious work done on Desktop and HEDT, but in the mind of a lot of companies it feels like it always get's the third-class just-a-gamer treatment.

I'm being a bit ranty in general with the tone here, it's late, like first thing on my mind, obviously it's a more complicated topic, and eventually all of this will come to the Desktop (...rrrrright?)
I love the news but, where's the support for stuff that matters*!

I tried coreboot on a supported board. It's meh.

I have since given up on x86 for open boot process. RISC-V is where it's at.

[Kjell]

AMD can have my wallet once open source/OpenSIL BIOS reach feature parity with AGESA

[RejectModernity]

Trying to make coreboot with tianocore work on my ASUS P8H61-M Pro was pain in the ass. No matter what I tried, I could not make my videocard work in the boot process. I even ended up patching GPUs firmware to support GOP driver but it didn't help. Sadly documentation lacks details. Very unpleasant experience when you try to compile dozens of firmware with different settings, but nothing helps. And no one could help, the community looks small and asking questions in the irc is like screaming into abyss.

[Adarion]

> There are 17 new motherboards supported by Coreboot 4.22
Oh. Well. And most of it is google-Chrome-stuff.
In the past that were more, right?
Coreboot goes nowhere unless there is decent support for real boards. The wiki was great, once. Showed supported chips of all kind. And yes, there were more boards. Now it is "useless" Google-stuff (their Chromebooks come with Coreboot by definition, which is good, but that is of no use to normal people w. normal boards) and once in a while an unobtainable developer board and if you're "lucky" one or two server boards which you don't need either as a normal person.

[pgeorgi]

> There are 17 new motherboards supported by Coreboot 4.22
Oh. Well. And most of it is google-Chrome-stuff.
In the past that were more, right?

The cadence was moved to quarterly releases recently, from bi-annual before. Of course the list of additions is shorter when it only covers half the time span.

The wiki was great, once. Showed supported chips of all kind.

You're probably referring to https://coreboot.org/Supported_Motherboards, which is a forward to a statically created HTML page these days. That page is created by the very same tooling, using the very same data set that updated the wiki page before (case in point: it under-reports the board count because it wasn't updated to board variants yet. It's really the _same_ code.)

And yes, there were more boards. Now it is "useless" Google-stuff (their Chromebooks come with Coreboot by definition, which is good, but that is of no use to normal people w. normal boards) and once in a while an unobtainable developer board and if you're "lucky" one or two server boards which you don't need either as a normal person.

Be the change you want to see. Chromebooks have the advantage of paid developers bringing up the support (and they have full hardware docs to get there, which helps).

There are a couple of other vendors (listed on https://coreboot.org/users.html) that mostly follow the chipset support pioneered by the Chromebook projects, so you can have devices with different specs and default installs. They'll be happy about every sale that they can make, and that'll feed back into further development of coreboot-enabled hardware.

The other option, of course, is to port coreboot to your favorite board yourself. That's a big time commitment, but surely you're not asking everybody else to do what you're not willing to do yourself, right?

Coreboot goes nowhere unless there is decent support for real boards.

coreboot would be dead for ~10 years by this point if it wasn't for those "unreal" Chromebook boards keeping the development pipeline alive.

(Full disclosure: coreboot old timer, coreboot.org admin, former ChromeOS Firmware developer, still working on ChromeOS)

[zir_blazer]

Coreboot goes nowhere unless there is decent support for real boards.

Aren't the MSI Z690-A and Z790-P real enough? You didn't used to have actual boards you can purchase worldwide in almost any eRetailer 2 years ago.

[Retramon]

I tried coreboot on a supported board. It's meh.

I have since given up on x86 for open boot process. RISC-V is where it's at.

That's a big thing I forgot, the feature parity, as the previous post mentioned, which is a bummer since this kind of stuff is suppose to be better and one would expect the reason for doing it was because traditionally BIOSes didn't get a serious treatment and being specific to manufacturer of the motherboard's good will support.

Well it turns out Coreboot doesn't even have the goals I thought it did ...

... a lightweight firmware designed to perform only the minimum number of tasks necessary to load and run a modern 32-bit or 64-bit operating system

From Wikipedia

Oh now I see on the about page they then transfer control to the payload, so they don't directly develop any payloads? I guess this project's not what I had in mind, meh

At some point it's just best for companies, if they're looking to go open-open source, to dump their work and effort over to one of the open-source projects and then it would make something like Coreboot (and payloads) viable but also practical, it's what AMD seems to be doing so it makes sense for them support something like Coreboot, but not the minimalistic approach that the project seems to be aiming at.

Advanced features and powerful customization does not need to be burdensome, heavy and unoptimized. I get irritated when people equate complexity, customizability and depth with inefficiency, instability and impracticality, not necessairly so at all. Reading up on Coreboot's philosophy I guess they didn't exactly have this kind of reasoning which is good, but then the payloads are where the big things happen I suppose, that is the place where it not only has to be parity but exceed in features of traditional BIOS, and I'm not saying in any kind of fancy graphical prowess or convenience feature such as networking, or even browsing the web (oh please), I'm talking about low-level and low hanging fruit that makes things usable and covers niches, that kind of stuff that motherboard manufacturers don't take seriously enough because there's not enough market demand and they get little in return. An open-source community would be able to fill in all those gaps as they come and work with passion not with corporate bottom line.

[pgeorgi]

An open-source community would be able to fill in all those gaps as they come and work with passion not with corporate bottom line.

The community generally gets done the things it's interested in (well, except for "it's too hard to figure out some stuff without documentation" type of issues).
Your pet theories about how coreboot development should work mean nothing if you're not willing to put in the work to make them happen.