Linux Running on Apple M1
twitter.com
twitter.com
A Mac with subpar support for a webcam, energy saving, suspend/resume, the trackpad, brightness/volume controls, etc is not a laptop, it's an expensive paperweight.
In my opinion, other than doing it for the sake of learning and the challenge of it, this will lead nowhere in the long term unless you get any kind of commitment or support from Apple. And I hope I'm wrong but that's very, very unlikely to happen any time soon.
I grab a machine and install Linux, and it works more or less out of the box. Maybe a few fixable quirks. And I don't use Thinkpads.
I try to use Windows or macos, and it's a coin toss. Windows handling USB like hot garbage (hub balancing/buffer sizes leading to devices unable to activate, webcams glitching, input lagging at random intervals), display issues (macos doesn't support displayport MST), dock connectivity problems (macos freezing or crashing when connected except when it suddenly works for a day, Windows playing disconnect/reconnect sounds in a loop when the machine goes idle), and more, including just today out inexplicable crashing.
Install Linux or plug all the "made for windows/macOS" hardware into a Linux box, and all the problems went away.
My conclusion, supported by having been a proprietary kernel driver developer for Windows, FreeBSD and Linux, is that any hardware and driver combo tends to be a coin toss, irrespective of platform.
But with Linux, drivers that would otherwise be abandoned after a project was shipped, get a chance at being fixed and improved that their proprietary counterparts can never even dream of.
First, non-desktop uses of Linux significantly outnumber uses of Windows and macOS in general, and a large majority of drivers and functionality is shared there.
Second, drivers not made by the usual giants are, as I mentioned earlier and experienced first hand, written and shipped once, maybe with a few updates for obvious issues, but this shit is harder than you think, with obscure uncaught bugs littered all over the place. Drivers need pretty much perpetual maintenance for all but the simplest blinkenlights.
Third, even "desktop" drivers are run through their places in server environments, through e.g. workloads like Stadia.
Fourth, proprietary drivers are not subject to any scrutiny whatsoever, whereas you won't get past the gatekeepers if you don't pull your shit together for the open source kernels.
Popularity does not fix unmaintained code. Giving even just one annoyed and skilled person the ability to change the code does. And sometimes, external companies like Collabora will decide to overhaul things, which they again can't do for a proprietary driver.
Go to your favorite computer store. Randomly pick 5 laptops, regardless of the price. So no cherry picking.
Use these laptops with Windows installed on them, for a reasonably long period, for example 1 month, as your daily driver. Perform varied tasks on them, such as printing, connecting to external displays, to projectors, other peripherals, playing modern 3D games on them, etc.
Then do the same with Linux.
I'm willing to bet $100 that on average Windows will run better on them, have a longer lasting battery life, better network connectivity, etc.
And if you're trying to tell me that on average Linux runs better than MacOS on Apple laptops and desktops, then this discussion is not worth continuing, we both probably have better things to do with our time.
You're very right that there is no value in continuing a discussion if that's the level of debate you're presenting. I'm off.
Linux as a personal operating system got WAY better over the last years, sure, but you can't seriously argue that it got Windows AND MacOS beat.
Now, a mac on its own tends to work quite well out of the box, but this does not hold for peripherals, and I feel like the machines always end up developing... quirks.
I always felt that the path you picked just gives you the sum of all problems with few of the benefits. I could maybe do the other way around for compat, but the laptop wouldn't be able to stay on this side of the balcony railing for long if I did it the way suggested... :|
Apples, oranges.
But I guess that's kinda in-line with this thread, that Macs are a pain to run Linux on.
And actually, Windows 10 is also approaching this state, but it is not there yet.
I last installed drivers on Windows late last year.
So your example seems to be completely unrelated to the question.
Based on...? Apple is pretty straightforward: they support hardware until they don't. And they're quite explicit about ending support and it's almost always a major release. You might be able to hack support after that, but I've literally never had it be a "toss-up" about when Apple was or wasn't supporting their own hardware.
As for Windows... I've got a 10 year old 2600k based desktop that runs the latest version of windows flawlessly. I guess if you go back 20? 30? 40? years you might find something that can't run the latest version of windows, but you're going to be down a really, REALLY obscure rabbit hole. I can't say I ever recall it being a coin toss, it was about 10 seconds on google of finding or not finding a driver.
Linux on the other hand... the support of hardware is awesome, but determining if something is or isn't supported is generally an afternoon of reading mailing lists.
My favourite example is one that the OP mentioned - not supporting multiple stream transport on Displayport. For those who aren't familiar, Displayport MST is a feature that allows multiple streams over one Displayport cable. Some monitors support this directly, meaning you can have
Macbook -> Display1 -> Display2
Macbook -> Display1
Macbook -> Display2
Except they don't support it for this.
What they support MST for is to allow vastly-higher-resolution displays on macOS, such as 5K displays, by splitting the display image over multiple stream transports to bypass the limit on resolutions and refresh rates that Displayport provides (or provided at the time).
For some reason, they just haven't bothered to implement MST to allow for multiple displays; it exists and is supported, but only for high-resolution displays. This is great if you're googling around and see that macOS supports MST, then you buy monitors which support MST and hey surprise it doesn't work and there's literally no indication why.
Dell makes some Displayport monitors which support daisy chaining.
Linux, however, is another story. Bad sleep support, forget about printing, scanning, GPGPU computing, etc. We use it when we have to.
I upgraded from thinkpad x240 on which sleep worked perfectly too.
What phone are you having issues with? Every android phone I’ve used will communicate with adb. iPhone has never used USB mass storage, and support for that has nothing to do with MacOS.
Can’t comment on the daisy chain issue, I just learned that was a thing.
Screen will remain black. Or, screen will power on, display nothing. Or, screen will power on, mouse cursor will appear, but no password prompt.
No such problem with Linux.
Yeah there are things Linux doesn't support but just don't buy them. The things that it does support generally don't require any driver installation.
macOS does support MST! It actually does! But it only supports it for splitting one display image over multiple streams, to overcome bandwidth limits on Displayport streams.
For providing a signal to very-high-pixel-count displays, macOS uses MST.
For providing multiple displayport signals to multiple displays? Nope, not implemented.
Imagine my frustration after a day of googling and finding out that macOS supports the feature we wanted, but not the use that we wanted.
MST is supported specifically for early 4k displays that had scalers that couldn't handle 4k60, but could handle half the resolution, so they sent two streams. But that wasn't a bandwidth limitation; old MST 4k displays and modern SST 4k displays both used a single HBR2 link.
MSTs primary usecase today is multiple displays, by "daisy-chaining" through built-in MST bridges, dual DP dongles or docks.
Support for hacky displays is less interesting, and hopefully not relevant today.
The screens light up from the same DP stream, which is also why you can have, say, 2x 4k@60hz monitors in this configuration where true MST would need to drop it to at least 30Hz, or lower res.
May I ask how you learned how to do this? I'd like to learn too! I had a great experience developing a Linux user space driver for my own laptop's LEDs. Couldn't figure out how to control the fans though.
2. Play around with embedded. You can use an arduino, but get rid of the Arduino IDE. Once you've ridded yourself from their weird environment and code in C, you're pretty close to what kernel programming is: Direct hardware control, debugging over a serial console, and if you mess up you don't get saved by a segfault.
You can upgrade to playing with ARM boards later if you want. Things like a Raspberry Pi can also be useful to boot random kernels you've built later on for HW stuff, otherwise you can use VMs. QEMU can boot a kernel file directly, which makes debugging easier.
3. Look at one of the tutorial for writing hello-world kernel modules. There's also usually smaller cleanup tasks you can do to get started submitting work. Looking at Linux and FreeBSD both can be useful, and things like Plan9 have very small kernels that can be used as reference. Linux and FreeBSD are not that different. (Windows is a pain with really weird interfaces, but it can be made to work.)
4. Find something you want to do with the kernel or fix in it.
Kernel developers aren't that common, so I imagine a lot of places are willing to train people. The first job I had doing kernel work was pretty open, and just threw minor stuff to begin with at me, e.g. "things stopped working after kernel X.Y, figure out what happened". Bisecting, testing in VMs, printk'ing a lot to compare state, stuff like that. I later ended up being the owner of the kernel drivers of all our platforms, so I guess I did okay. :)
But not having the right documentation was a challenge. MSDN is okay, but the weird mechanics of MDL chains don't really get discussed on Stack Overflow.
How do people overcome this? I managed to reverse my laptop's LED commands: they were implemented via USB so I used wireshark to intercept and analyze the data sent by the proprietary vendor software. What if it's some ACPI thing though? Or some memory mapped I/O chip? How do people figure out how it works?
It isn't like the lack of support is Linux developers' fault. Apple doesn't provide datasheets for their hardware, and they don't cooperate with developers writing drivers for their custom hardware.
There are hundreds, if not thousands, of ARM SoCs that "run Linux", but that doesn't mean much because they're actually running Linux forks, and someone needs to maintain those forks, and build and release custom images for each SoC. I don't see M1 Macs diverting from that fate without significant support from Apple themselves.
Many of these Arm SoCs are running linux forks because there are some terrible baked in drivers, and no spec associated with them. Still, that is a step ahead of the M1 as there are at least drivers that don't have to be decompiled and reverse engineered.
The thing is, that these SoCs have no PCIe or other enumerable bus. You (=your kernel) must know what hardware it is running on and which drivers to load without being able to ask the hardware, what is really present. One wrong POKE and the entire system can hang.
And that is on top of the problem, how to boot in the first place. There is not such thing as UEFI on these SoCs. Every single one is a special snowflake with its own special way to boot.
Hence, kernels built for specific systems, even if it is just Device Tree.
Most device vendors don't really have a need or incentive to use this, though. A generic kernel is much larger and wastes flash space, and most vendors don't care about running anything but their OS image on it, so they of course just build a kernel specific to their hardware.
But in theory it should be possible at some point to use a generic ARM Debian installer to install your own OS on a random ARM device, assuming the relevant drivers and device tree have been upstreamed.
Again, the problem isn't that the mainline kernel _couldnt_ support these boards, is that no one wants to put in the work to bring their implementations up to the standards of the mainline kernel.
Eh, the majority of embedded crap comes with U-Boot (sometimes built without EFI support but always with Linux image support, at least).
It appears what should be being asked for is Apple to commit resources to natively support things like: Docker, Kubernetes, Virtualization and Linux drivers on their M1 Macs going forward.
> Apple designed their own interrupt controller, the Apple Interrupt Controller (AIC), not compatible with either of the major ARM GIC standards. And not only that: the timer interrupts - normally connected to a regular per-CPU interrupt on ARM - are instead routed to the FIQ, an abstruse architectural feature, seen more frequently in the old 32-bit ARM days. Naturally, Linux kernel did not support delivering any interrupts via the FIQ path, so we had to add that.
and many others
https://corellium.com/blog/linux-m1
It is on top of HN now https://news.ycombinator.com/item?id=25859907
You can see a list of differences between Macs and standard PCs here: https://en.wikipedia.org/wiki/Apple%E2%80%93Intel_architectu...
However, I’m optimistic that these will be mostly usable on Linux before they’re too obsolete or outdated. The platform is so popular and iconic that it’s drawing a lot of attention from Linux devs and reverse engineering crowds.
That's sound advice.
As someone who doesn't use Linux as their daily driver, can I expect Linux to run perfectly in a VM on an M1 Mac?
Sure the M1 is the new shiny, and people will be attracted to it in the short term, which might boost reverse-engineering efforts. But I expect that to die down as people get frustrated, and we'll have the same (or worse) situation as we do running Linux on Intel Macs.
Ethernet, waking up from sleep mode, brightness adjustment... take your pick.
A reminder that not only software bitrots, but hardware too. Make sure not to buy a Stinkpad (or any hardware) with non-replaceable parts!
(and I just bought an M1 Macbook Pro...)
Software will catch up with the M1, I hope!
Can't wait for [Your least favourite Electron App] to use even more CPU cycles!
Correct - that is why I bought a Thinkpad X1 Extreme - both SSD and RAM are user replaceable. Granted its not as thin as X1 Carbon but I don't need to carry it around much so that is not a big issue for me.
Can work in the meantime, before you have a new machine.
But I don't have a SATA reader either, so I'd need one of those :-)
Though, if you want a serviceable system, I'd stay away from Apple - the RAM in your new M1 mac is actually included in the SoC, not sure how much you'd like to get in there!
Sure, documentation helps, but it is overrated--people get lots of things working by reverse engineering.
https://shop.solid-run.com/product-category/embedded-compute...
You get upgradable ram, a PCIe slot for your favorite GPU/etc, and actual mainline linux support sufficient to boot most random linux distros (ok the onboard 10G+ nic might not work without patches).
Its not the fastest machine around (A72s) but there are 16 of them, so it does a decent job of building ARM software, and running VMs.
In fact I made the FreeBSD port of amdgpu work on my Macchiatobin :) Absolutely smooth experience btw, video output works even in UEFI (it actually runs the GOP driver from the card in QEMU), amdgpu works perfectly (played vkquake, supertuxkart, openmw, etc.)
[1]: https://www.jeffgeerling.com/blog/2020/external-gpus-and-ras...
So, yes you have to bring some storage, ram, ps and case. But one can cheap out and probably get all three for <$200, or go crazy and put 64G of ECC, and a few TB of storage in it. The cheaper off the shelf RAM/storage makes it a lot cheaper than a loaded mini with similar specifications.
And as another commenter points out, yes fairly common GPUs tend to work in it. Its not perfect, for that you have to spend more, but the time you will spend fighting with the mac mini+linux is going to be worth the difference.
Perhaps you're meaning to say Apple will iterate the hardware faster than people can add Linux support? That's plausible. It also was the status quo of Linux on PCs for many years. (Still the case with some hardware I guess.) That hasn't killed Linux yet.
That driver has been upstreamed, but I hear the 2017 and onward touchbar MBPs still require and out-of-tree driver, and a similar level of things don't work.
In 2019 I finally gave up and got a Dell XPS13. It's... basically perfect? The keyboard doesn't suck like recent MacBooks, though the touchpad isn't quite as nice (though really, it's perfectly fine). The only thing that doesn't work is the fingerprint reader, but I knew that going in and didn't care. And subjectively the hardware isn't as pretty, but... oh well. On the plus side, I really do like the "soft touch" palm rest better than the harsh aluminum of the MacBook.
I'm definitely interested in the Linux-on-M1 project's progress from an intellectual curiosity standpoint, but I'm expecting long-term as a user it'll be just as -- if not more -- frustrating as running Linux on any Intel MacBook from the last 5 years, which means I won't bother.
My personal laptop is an HP Envy X360 with a Ryzen 5 2500U. That was rough when it came out. Windows update updated the bios on the laptop which installed a new GPU firmware that wasn't supported by the Linux driver for a few months. It's been solid since kernel 5.0 came out.
IMHO, the Linux experience on the AMD APUs is better than the Windows experience due to how rough the AMD windows drivers can be for the APUs. The fans are always shrieking at me under Windows but at idle they are off in Linux and the battery lasts significantly longer.
The computer never woke and needed to be hard-rebooted if it was ever allowed to suspend? I ran into this after installing Linux on a MacBook, downgrading the kernel to an old LTS release eventually worked.
I think right now the new M1 machines are actually priced pretty reasonable. If they'd run Linux, I would buy them in an instant.
I mean, try to match the Air with a Thinkpad, especially if you consider the screen. I think there is not even one recent AMD Thinkpad with a > FHD screen. And who wants a 2020 Intel machine?
Edit: To make this clear: No Linux, no Mac for me (as I don't believe in MacOS's future at all). But even considering the horrible keyboard, I think the M1 macs are totally worth it just for the hardware. That's a first. And I am the kind of person who feels all warm and fuzzy over ejecting an ultrabay hard drive, or changing RAM/display in < 5 minutes in a laptop.
There are Mac desktops as well. That said, I'd say the list of mandatory components on a laptop is:
- Decent GPU Drivers
- Workable power management with sleep/ awake
- Trackpad (this should be straight forward)
The webcam, and brightness controls are stretch goals. I can't imaging the trackpad is vastly different from currently shipping ones. The Webcam might cause some significant headaches since Apple secures that fairly tightly.
The onboard wifi modem is likely worthless, from what I recall, many of them are even in Windows laptops due to proprietary drivers.
Hopefully with a little bit of money and some enthusiasm from the team, they'll be able to make more progress.
A lot of people bitching about this being a waste of effort, but all the things they do here have follow on effects. If they improve the Broadcom drivers for the M1, much of that work will directly benefit everyone else on Broadcom.
Honestly this sounds exactly like my experience with linux on any laptop, including three thinkpads.
Linux is a great desktop and server OS but the UX of a laptop actually demands a certain level of polish/fit-and-finish that linux desktop environments just don't have, otherwise it's constantly getting in your way.
Apple clearly doesn't seem interested. The M1 is just a scaled-up version of the iDevice SoCs. From the reverse engineering we've seen so far, it is extremely clear that not even a single step towards any standardization was taken. It's very much the ad-hoc hodgepodge of embedded crap companies build when they only care about their own complete product and don't allocate any budget towards any "refactoring" that doesn't directly benefit the end product.
They use some ancient Samsung UART for debug (because the first iPhones used a Samsung SoC?), old P.A. Semi I2C controller, the Synopsys DesignWare USB 3 controller (just like some random cheap Allwinner/Rockchip/etc), and here's the fucking kicker, a custom Apple interrupt controller and a custom IOMMU too. These probably predate GICv2/3 at least. But there was no reason for them to switch to the standard Arm GIC and SMMU so they didn't >_<
If you want something working, I'd also not recommend using hardware from someone like Apple.
Yes, it will have to be RE'd. Yes, that takes time.
Essentially, while this project may be quicker to get visible results, they may not be able to release all the code kosher for merging upstream. It will be interesting to see the next steps.
Winner takes all.
When teams of people compete intensely , the collaboration within the teams themselves must be just as intense.
I find it slightly offensive that someone would accuse me of discounting collaboration when 1. I never discounted it, 2. It's obvious that society is full of people who collaborate. In what universe are the words "zero sum" apt for my response? None. I never described a zero sum game. Obviously, the replier added the description with his biased imagination.
I'm just saying spiciness and intense rivalry and competition can lead to results beneficial to society. There are tons of examples of intense competition and rivalry leading to great results in science. The decoding of the human genome for one.
You wrote, "Winner takes all" which seems pretty zero sum to me, hence my comment.
As for “winner takes all” why don’t you look up the definition of a “zero sum game”. A zero sum game usually applies to simplistic games like chess or an island with limited resources aka things that have a measurable gains and losses. Complex situations like the one described are rarely zero sum.
When I say winner takes all its more of an “expression” symbolizing the intensity of competition. I think it’s quite obvious that the situation here is not some contest setup so that a single winner takes everything. There’s no need to make your self sound smart and use the words “zero sum game” redundantly. Only certain types of people use the words “zero sum game” colloquially for the purposes of sounding smart even though the majority of situations in nature aren’t actually artificial games setup to be zero sum.
The word is also used negatively as if zero sum games can’t ever exist. Like it’s obviously wrong if your describing a zero sum game. It’s rare but zero sum games do exist so stating that something is a zero sum game doesn’t move the conversation forward. Like so what? Yeah I could be describing a zero sum game, it doesn’t make me wrong, what’s your point?
Case in point, the credit for the First person or team who can get Linux running on the m1 IS a zero sum game and there already is a winner for that “game.”
That line made you sound all-in on zero sum competition. But you’ve clarified now and softened it to say it can lead to beneficial results.
“ We all like to think that collaboration, harmony and love drives innovation but much of innovation is built around intense rivalry and competition as well.”
Keyword here is “as well”. If you feel the need to respond or vote someone down please read the post carefully rather then respond or vote baselessly.
More clarification is necessary. Competition is the driver of natural selection. Your entire biological form exists as an evolution of winning traits because your ancestors out competed and defeated others who fought to reproduce so someone else could take your place.
Competition is therefore a primary driver of your existence while collaboration is secondary. It’s not that competition can lead to benefits, the phenomenon that occurs is that collaboration can actually work but only as a tertiary driver behind competition.
See communism if you want to know the results of a society formed with collaboration over competition as the primary driver.
These things can escalate surprisingly quickly, so you really need to be careful. I've watched it happen.
All of society and the development of capitalism to evolutionary biology is founded on competition. Competition is, in fact, the primary success story and collaboration is the side story. Citing the failure of one community discounts the view of the entire world. Competition works, and it works better than collaboration. See communism if you want an example about a community founded on collaboration as the primary driver.
Either way you're citing singular examples and calling it "seemingly unhealthy." It's a weak argument against my example of the entire modern world as a competitive arena.
I wouldn't dare considering it the most likely scenario, but it's surely the most lovely.
https://lore.kernel.org/linux-arm-kernel/20210120132717.3958...
marcan is CCed on that set of patches as a courtesy, and he can help with figuring out better approaches before it's merged. Because it's set in stone forever after that.
- someone
I like to view competition as a form of collaboration, for example:
When companies compete, they compete in a collaboration we call the market.
When sports teams compete, whether as groups or individuals, they compete in a collaboration we call the game, or even a tournament.
When the other side is beating you, you might look at their strengths and weaknesses, their knowledge and ability, and work out if there's anything you can learn, how to adapt your behaviour to strengthen your position.
Employees, or players, can move from one company to another, through hiring or acquisitions, bringing skills and experience with them. Here we have companies competing for employees, and then those employees collaborating anew.
Or a company might license certain IP from another, this is a form of collaboration too.
In this sense, collaboration and competition aren't necessarily opposing paradigms, they can be tools to apply to the situation.
WTF.
Still going. Don't think even stopped to eat, drink, or bio break. I'm also surprised his computer hasn't crashed or something! Really fun to watch him work.
I really think this is getting unhealthy hyperfixation and I really hope this is unusual for marcan, as this comes at a cost you can never make up for again.
Btw. if you enjoy these extended coding streams, you may like the Scanlime livestreams: https://www.youtube.com/c/scanlimeinprogress/videos. She's also a fascinating person to observe in her natural environment. The videos are also low key trippy and artsy.
Main channel here: https://www.youtube.com/c/scanlime/videos
I think opening up just enough to enable this effort also serves as some good nerd marketing for Apple. This race keeps the m1 in the news regularly, giving hope to those who want a low-power and performant ARM machine that wouldn't otherwise consider an apple machine.
(Dealing with the GPU is going to be the majority of the work, I'd think.)
...without the documentation that would help you.
When Broadcom act like this they're considered villains and we're recommended to stay away from their hardware. But when Apple do it, they're being benevolent?
And that's ignoring the fact that I could actually get Broadcom documentation in exchange for dollars and NDA.
It does work out for Apple in the end. Their current standard 10 years' support will look quite short now Moore's law is dead and their hardware has barely any moving parts. But they'll shush some complaints if up to date third party OSes are available in 2030.
Well, it's not really worse than their usual documentation on products they officially support: https://www.caseyliss.com/2020/11/10/on-apples-pisspoor-docu...
Apple makes money selling hardware, and Linux support will sell more hardware. Perhaps not much more, but for a commensurately low amount of effort. What does Apple gain by forcing Marcan to reverse engineer everything?
I realize that Marcan isn't Microsoft—but he's not quite "a random guy with a Patreon" either. He's a professional freelancer, and I'm sure he has an LLC† and a set of professional references he can point to.
Put another way—on a scale between "Marcan" and "Microsoft", where is the threshold in which Apple would be helpful? I don't personally see a huge difference between a one-person LLC and a 100,000-person company in this regard. If anything, the 100,000 person company offers more opportunities for things to get leaked.
---
† Or something similar.
There are some interesting situations though. Like in the big GPU world it's very common to document the ISA. Even nvidia does. I suppose that's because it benefits the vendor when games and GPGPU compute programs optimize for their GPUs, down to the assembly level. It's sad that Apple's approach is "just use Metal" rather than fully enabling developers to get to the actual… well, "metal"
Apple‘s competitors have sufficient cash flows to pay whatever they need to get the best people. They just don’t want to.
It’s a good sign that the latest betas (11.2 IIRC) officially support multiboot in the UI. That’s a good indication Apple sees the level of interest in Linux on M1 that they intend to at least let it happen.
I’d say it’s still up in the air whether they’ll go for full first party support with drivers or an open spec, but it’s definitely not out of the question. And they may even have direct interest in it, as I’m sure they’d like to get the benefits of their hardware in their data centers.
Maybe the people doing this for free are just interested in benefitting from the result? As many people who work for free on open source do.
As far as I’m aware, there is no free (as in beer) hardware that runs Linux. Someone has put the effort into running Linux on every single for-profit/for-pay platform it runs on.
Are you under the mistaken impression I was suggesting that Apple waits for a community solution to be developed then packages that as a product? As far as I’m aware they didn’t do that with Boot Camp, but instead offered in-house drivers and blessed boot loaders and proprietary UI/UX for accessing both.
They're hoping to get increased for themselves, increased adoption of their internal tools outside of the company (easier recruiting plus purely internal tools are notorious for rotting quickly) and... free labor.
Somebody wants to copy the hardware over and sell it for half price? Yeah, good luck reproducing the M1.
Opening the boot loader to allow for Linux is quite an opposite of a middle finger, tbh. I don't know if they will divert some guys from working on MacOS towards Linux support, but this is already looking much better than before.
They don’t want you installing their OS on other hardware, not the reverse.
Perhaps for some of the peripheral stuff (such as the touchbar), but the GPU ought not to need it.
I wouldn't bet on it. Apple are very insular on matters like this, it's their toy, end of.
If the latest is true that could prove to be another prestige point for the M1.
They may have in order of 100 registers to talk to them and they're horribly proprietary with pretty much no standardisation.
Reverse engineering that is hellish at best - you can see projects like nouveau which barely manages to get nVidia cores up and working without help from the manufacturer. And that's after years and years of development.
That's pretty much game over for Noveau, and it is not due to difficulties in figuring out registers and NV ISA.
But that's a bit besides the point :)
I can understand preventing FW patching, maybe. But redistributing the manufacturer's genuine FW, why not?
Multiple reasons:
1) GPU manufacturers are notorious for not publishing documentation out of IP/patent concerns. Worst offender is NVIDIA here.
2) For embedded GPUs there isn't much interest in open source drivers... the big customers (think Samsung and the likes) have direct support from the chip design vendor and get drop-in drivers as part of the board support package (BSP, basically a conglomerate of bootloader, kernel+modules+initrd, firmware blobs for components such as wifi) so they don't need OSS drivers
3) The mobile GPU space is... splintered. With desktops you got the three major players AMD/ATI, NVIDIA and Intel's built-in Iris, in the GPU space there are more.
I think easily Apple takes the cake from nVidia - they don't even provide drivers for anything but their platforms (that is for their proprietary GPU core). The GPU core that's actually in the M1.
Making any modern GPU work is a lot of work because of how complicated they are. That's even with the full documentation.
In the Apple M1 case, the GPU will have to be reverse engineered to understand how it works, then a driver will need to be written for Linux that supports it.
Getting GPU acceleration is now the real challenge.
"Oh."
I think I need to recalibrate my idea of a 10x developer...
(The standard would seem to have ratcheted up somewhat as the years have gone by!)
You might have to play shenanigans like copying a complete filesystem into a ramdisk from the bootloader if your kernel doesn't have support for any IO/networking/storage devices. But you'll still be able to get this screenshot!
Having said that, they have USB working, which is quite an effort, although I'd guess it's an IPCore that a driver already existed for, so it was a simple matter of figuring out memory mappings etc. With USB working, you can make a very usable system, because pretty much any peripheral will work over USB.
Apple is finally filling in one of the Arm gaps with a mid-tier chip that can handle a desktop load efficiently while an efficient GPU can handle graphics without thermally clobbering each other.
You can argue that their chips in this segment are not very good, but Qualcomm is actually specifically addressing the laptop market, and they're not "no one".
https://www.anandtech.com/show/15210/qualcomm-expands-lineup...
I have a OrangePi Zero w/ 512MB RAM at home and it handles a lot of stuff (Syncthing, DNSMasq, rsync based backups and more) without a glitch.
The only thing it doesn't like to handle SFTP encryption at high speeds. Processor gets visibly strained and overheats after a 4MB/s or so.
The GPU is make or break for Linux. No GPU means it's just another server in a different form factor.
Let that thing have more standard IO, and run Linux, and it'd be a heck of a fanless machine.
"macOS Big Sur 11.2 beta 2 is out with full custom kernel support" ... "The OS now finally includes the firmware and bootloaders and tools necessary to replace Big Sur with not-Big-Sur. That was previously not possible."
Even if Apple does do some open source support, Even if that hardware isn’t quite as nice, I prefer to buy/support hardware and vendors that support Linux at this point.
Perhaps the M1 is sufficiently compelling to muster the engineering resources needed to get every piece working nicely, we’ll see!
But I have pushed the decision back to see how Linux will be running on the M1. From first reports, it is simply tailor-made for a laptop computer that doesn't get burning hot or dog slow because of thermal throttling.
I hope intel and to a lesser extend AMD are looking at this thing red-faced. But they will need some time to play catch-up.
I wish Apple would help Linux on their hardware as it won't take away a lot from their software business and they do have the resources, but yeah, they are just a big company like others in that regard.
I think it raised their eyebrows a bit, but don't agree that they need some time to play catch-up. Bottom line is the vast majority of Intel and AMD chips are running Windows. You can't even buy Windows 10 for an ARM processor. Microsoft flirted with the idea at one point, but has pretty much abandoned it. So why would Intel or AMD market ARM chips at all when Microsoft isn't supporting them at all?
If you are talking about servers that is a different story (since lots of servers run Linux). AMD already made a server with an ARM processor, the latest "Opteron" series was ARM https://www.amd.com/en/amd-opteron-a1100 These came out in 2016 and don't think they sold nearly as well as Epyc has. Taking a second look at offering ARM for the datacenter in the fture might not be a bad idea for AMD though.
The M1 isn't a Xeon in disguise, it has its limitations. But for the sub-notebook form factor, it's in its own class and freaking intel can't match it. And they've been treading water since about 6y now.
Also: me personally, I don't care about Windows. I want to run Linux. I understand that's a niche market, but we are writing in a Linux thread, so...
And this was a distro geared towards PowerPC hardware.
This both reads a little overly optimistic to me knowing how proprietary things are - but linux and foss software have history of running on all manner of platforms. Arguments based on: "you can only get a client machine with at max 16 gig of ram" or "you can't plunk an M1 into a server chassis with all the hardware you need" and so on may be various amounts of preliminary (we only have 1 chip from them yet and it's clearly a 'first out the door' type near-beta) to actually relevant.
Is the CPU/SoC not the thing on which then operating system is running?
If the tweet was claiming that it was ready to go for general use and that everyone should install it and use it as their daily driver, then I could see that the external boot disk caveat would be more significant... but it seems kind of irrelevant in the context of what they've achieved so far.
The most common AArch64 distro usage is on RaspberryPies
I was able to get qemu to boot an x86 Linux after applying a huge pile of patches I found on a forum.
If you are expecting to get up and running instantly with a VM to put all of your comfy old software in, not so fast!
Happy to see frame buffer, excited for great linux experience on m1.
It will take years for the open-source community to reverse-engineer the ISA and write kernel code that optimally utilizes it. There's no shortcut here. Until then, it's not worth buying a Mac just to run Linux on it.
1. Instructions to change the memory ordering mode, which allows Rosetta2 to skip some expensive memory fences when translating x86 binaries to ARM.
2. Some instructions for matrix math.
I wouldn't expect either of these to affect typical usage or battery life much, unless your workflow involves running things that do a ton of matrix math.
The second is an extension that talks to a coprocessor that's quite application specific. And that has an unstable ISA that changes every year. Probably not the biggest priority for anyone.
"The main problem with the M1 for me is the GPU and other devices around it"
That's a separate issue than the ISA extensions, which apply to the ARM CPU.
Who's right now?
Plus, what do Apple fanboys know about hardware?