Follow Slashdot stories on Twitter

 



Forgot your password?
typodupeerror
×
Operating Systems Linux

Linux 6.2: The First Mainstream Linux Kernel with Upstream Support for Apple M1 Chips Arrives (twitter.com) 65

Steven Vaughan-Nichols, writing for ZDNet: Linux 6.2 was released yesterday, and Linus Torvalds described the latest Linux kernel release as, "Maybe it's not a sexy LTS release like 6.1 ended up being, but all those regular pedestrian kernels want some test love too." For once, I disagree with Torvalds. By adding upstream support for the Apple M1 Pro, M1 Max, and M1 Ultra chips, newer Mac owners can look forward to running Linux on their M1-powered machines. And, for techies, that's sexy. Getting Linux to run on the M1 family wasn't easy.

When these high-powered ARM chips first arrived, Torvalds told me in an exclusive interview that he'd like to run Linux on these next-generation Macs. But, while he'd been "waiting for an ARM laptop that can run Linux for a long time," he worried, saying, "The main problem with the M1 for me is the GPU and other devices around it because that's likely what would hold me off using it because it wouldn't have any Linux support unless Apple opens up."

UPDATE (2/26/2023): Asahi Linux called ZDNet's report "misleading and borderline false," posting on Twitter that "You will not be able to run Ubuntu nor any other standard distro with 6.2 on any M1 Mac. Please don't get your hopes up." We are continuously upstreaming kernel features, and 6.2 notably adds device trees and basic boot support for M1 Pro/Max/Ultra machines. However, there is still a long road before upstream kernels are usable on laptops. There is no trackpad/keyboard support upstream yet.

While you can boot an upstream 6.2 kernel on desktops (M1 Mac Mini, M1 Max/Ultra Mac Studio) and do useful things with it, that is only the case for 16K page size kernel builds. No generic ARM64 distro ships 16K kernels today, to our knowledge.

Our goal is to upstream everything, but that doesn't mean distros instantly get Apple Silicon support. As with many other platforms, there is some integration work required. Distros need to package our userspace tooling and, at this time, offer 16K kernels. In the future, once 4K kernel builds are somewhat usable, you can expect zero-integration distros to somewhat work on these machines (i.e. some hardware will work, but not all, or only partially). This should be sufficient to add a third-party repo with the integration packages.

But for out-of-the-box hardware support, distros will need to work with us to get everything right. We are already working with some, and we expect to announce official Apple Silicon support for a mainstream distro in the near future. Just not quite yet!

This discussion has been archived. No new comments can be posted.

Linux 6.2: The First Mainstream Linux Kernel with Upstream Support for Apple M1 Chips Arrives

Comments Filter:
  • understanding that sentence

    wouldn't have any Linux support unless Apple opens up."

    • Re:I have difficulty (Score:4, Informative)

      by franzrogar ( 3986783 ) on Tuesday February 21, 2023 @09:16AM (#63311115)

      Apple won't release any documentation as how to properly support them because that would mean they would release design documents, trade secrets, etc. that would help competitors to create better chips.

      Think of it like nVIDIA and how the opensource codec had to hack-and-slash until it got something workable without support from the company.

      • That was sarcasm ..
      • by dskoll ( 99328 ) on Tuesday February 21, 2023 @09:33AM (#63311169) Homepage

        If Apple released enough documentation for people to write open-source drivers, it would not help competitors one iota.

        Competitors already keep a close eye on one another, and you can bet that major competitors have acquired Apple chips and reverse-engineered them to see what's up.

        No, this is just control on the part of large companies. They can't stand the idea of open-source software somehow reducing their control over their products. It's a bullshit control-freak mentality... nothing to do with preventing competitors from gaining knowledge.

        • by samkass ( 174571 )

          They can't stand the idea of open-source software somehow reducing their control over their products. It's a bullshit control-freak mentality... nothing to do with preventing competitors from gaining knowledge.

          I doubt this. After all, you can download all the source code to the MacOS XNU kernel too, edit it, recompile it, and install it on your Mac. I think it's just that Apple has a lot better things to do with their engineer's time, especially when it comes to making money, than trying to get this to work and getting the documentation to widespread-distribution levels of quality.

        • Yeah.... I'm not so sure? On one hand, we know Apple is a control-freak type company in general. That attitude has served them well over the decades, and was pretty much demanded from the top down by Steve Jobs.

          But with the M series processors? What we seem to have going on is a CPU/GPU combo that Microsoft has no interest in trying to port Windows onto. (Their official stance is to buy a copy of Parallels desktop to run it via emulation, or to pay for them to serve an instance of it to you via the cloud.)

          • Yeah.... I'm not so sure? On one hand, we know Apple is a control-freak type company in general. That attitude has served them well over the decades, and was pretty much demanded from the top down by Steve Jobs.

            But with the M series processors? What we seem to have going on is a CPU/GPU combo that Microsoft has no interest in trying to port Windows onto. (Their official stance is to buy a copy of Parallels desktop to run it via emulation, or to pay for them to serve an instance of it to you via the cloud.) Every time MS has tried to shoehorn Windows onto ARM CPUs, it wound up performing poorly and had almost no third party software support. And from what I've read from others here? Apparently, there are some logistical problems trying to move the x86 type code over to ARM having to do with sequences/order of operations. It's much easier to start from scratch like Apple did, writing an OS specifically for the chipset (and note how they completely dumped 32-bit app support before ever attempting it, too).

            So .... it *would* be potentially compelling for Apple to cooperate with providing tech details the open source community could use to make Linux run better on it. Apple would still make the money on the hardware sales that it makes now, and since MacOS comes WITH the machines and not sold separately? They even sell a MacOS license with each one -- even if the user plans to wipe it and make it a Linux box. This could be done without aiding Microsoft in getting people more comfortable with the Windows platform and software applications.

            I don't think you know the difference between Emulation and Virtualization.

            Even running x86 code isn't done under Emulation, unless
            there is no other way; like if code is dynamically-generated during execution. It is Translated from x86 to ARM before running. Not even close to the same thing.

            Rosetta2 and MS' version of the same thing is not "SoftPC".

            Apple did dump 32 bit support, because the ARM architecture was SO much more efficient at 64 bit code that it would be silly to do anything else; and bec

        • If Apple released enough documentation for people to write open-source drivers, it would not help competitors one iota.

          Err that is demonstrably false given the history we have of performance improvements that have occurred due to code leaks. Driver level optimisations are a real thing and complete documentation of a device provides a lot of insight into how a device does what it does.

          Sure you could likely eventually figure it out through reverse engineering the existing delivered product and code, but that takes a long time.

        • >"If Apple released enough documentation for people to write open-source drivers, it would not help competitors one iota. "

          Correct

          >this is just control on the part of large companies. They can't stand the idea of open-source software somehow reducing their control over their products."

          Not correct. It is the part of APPLE, and MICROSOFT (and perhaps some other large companies), but many other large manufacturers, Lenovo for example, have no hostility at all to Open Source/Linux. In some ways, they ha

      • Re:I have difficulty (Score:4, Informative)

        by AmiMoJo ( 196126 ) on Tuesday February 21, 2023 @09:50AM (#63311209) Homepage Journal

        I don't think releasing a description of the registers and the rendering pipeline would make much difference to competitors. Apple's GPU isn't magic, it's the same as many other mobile GPUs. Performance is okay, not spectacular.

        Same with their ARM CPUs really. Nothing revolutionary there, just massive caches and all the usual pipeline tricks. There are a apparently a few enhancements to make x86 emulation faster, but they aren't needed for Linux and there doesn't seem to be any public compiler support.

        • We live in a world where major vendors fall over themselves providing optimised driver updates that deliver actual step changes in performance on their hardware. And sadly in this same world Linux either resorts to binary blobs or a significant performance hit from open source drivers.

          There is absolutely IP to be discovered in the complete documentation of a GPU.

      • by Burdell ( 228580 )

        It's not that releasing docs would give competitors any significant advantage, it's that Apple is not a "hardware vendor", they're a "package supplier". They are not in the business of selling hardware for you to run whatever you want on it, they are selling a packaged thing (hardware, software, and support) - they want to control as much as they can what you do with their packaged thing. You're just supposed to look good using it and advertise it for them.

      • by Anonymous Coward

        I preferred Apple when it was pre-post-Woz. Remember when they used to release schematics for their mainboards? Remember when they used to provide commented assembly code for their mainboard and I/O board firmware? Remember when they used to publish thick dead tree volumes through Addison-Wesley about their hardware, firmware and toolbox routines on Apple II and Macintosh computers?

        They were pretty much Open Source before that was even a term. But the money grabbers have taken over since then and now it's a

      • Yeah, no. Even Nvidia has open sourced their drivers [github.com] (kind of) at this point, because nothing that is important in the chip design is exposed by the driver code.

        If even Nvidia has come around on that, then Apple is VERY late to the party. And it's not like Nvidia, AMD, Intel, etc. can't just go buy a Macbook and de-lid the SoC and have a look in a scanning electron microscope anyway...

    • by vbdasc ( 146051 )

      So, forgive my ignorance, but did Apple open up?

  • by Hydrian ( 183536 ) on Tuesday February 21, 2023 @09:41AM (#63311183) Homepage
    I think that M1 support in Linux 6.2 is just a stepping stone. There will still be bugs and external devices that will need to be figured out. It is going to be far from usable on Mac M* based hardware.
    • by caseih ( 160668 ) on Tuesday February 21, 2023 @10:25AM (#63311321)

      But yet Asahi Linux is now good enough that Linus uses an Apple M1 as his daily driver. This kernel release came from his Apple M1. So it's usable now to a large degree. Yes there are bugs and some things are not working that need to be, but people are using the M1s every day as their main laptops.

    • by Holi ( 250190 )

      It's not the CPU support that made this newsworthy, they got the GPU working with OpenGL.

  • for closed hardware. Funny that Linus gives NVIDIA the finger but buys Macs.

    • The dude seems a bit bipolar.

    • for closed hardware. Funny that Linus gives NVIDIA the finger but buys Macs.

      Heck even Apple gave the finger to NVIDIA, so the enemy of an enemy is a friend?

      Given his work history and the fact he describes himself as non-technical, outside of the kernel space [youtube.com], maybe the Apple hardware is a place where Linus can look good on a laptop, if it works?

    • Some people enjoy the path of more resistance. I've got a friend who could easily afford a new car, but he'd rather tinker away on some clunkers.

      • Some people enjoy the path of more resistance. I've got a friend who could easily afford a new car, but he'd rather tinker away on some clunkers.

        I thought enjoying the path of most resistance is the watchword of all Linux masochists.

        Seriously. Ya buy an Mx Laptop with a perfectly good Unix (with the revered CLI) installed, updated and maintained; on which you can install and run basically any 64 bit x86 and ARM software.

        But you'd rather exhaust yourself swimming upstream, just to. . . to what, exactly?

    • for closed hardware. Funny that Linus gives NVIDIA the finger but buys Macs.

      Maybe he knows something that you don't.

      Note that Apple gave the finger to nVidia, too. . .

    • by bn-7bc ( 909819 )
      That is a bit different, Nvidia designs and manufactures (via partners) a component that goes into an otherwise linux compatible open-ish platform (ie pcs), Apple makes its own SOC which they but into their own laptops/tablets/desktops. The difference is with macs /after the M-series transition) there are no expectations of compatibility, but for a repurposed pc it is annoying that the drivers break on every kernel update (I'm talking about nvidias binary drivers, at lest several years ago when I last tried
      • this is a bit of a stretch. the lack of Linux drivers for hardware in your PC is a small matter of closeness. And entire hardware platform that is closed is a large matter. As much as I cringe when RMS is trotted out, he's right about these closed platforms being a serious problem. And I'll go a step further than RMS and say this isn't just a philosophical or political soapbox, but if you can't control hardware you bought with your money then did you even buy it? If I can't operate something as I see fit, a

Experiments must be reproducible; they should all fail in the same way.

Working...