13 ms·

Asahi Linux for M1 Macs Progress October-November 2021

367 points by svenpeter 5y ago

SamuelAdams 5y ago
> Amusingly, while implementing support for this in Linux, we ran into a bug in Linux’s ARM SMMU support that had been there ever since 52-bit address support was introduced. This was breaking systems with more than 256 TiB of RAM - I wonder why nobody noticed? Either way, Linux now correctly supports standard ARM systems with up to 4 PiB of RAM ;-). Interesting how working on new technologies helps improve existing things for all users. Sort of reminds me how NASA invests in R&D and those investments eventually come back to military and consumers.
- fragmede 5y ago
  Especially handy for SpaceX, which employs a number of ex-NASA folk. SpaceX wouldn't be where they are today if it weren't for the investment in NASA.
- Leherenn 5y ago
  > This was breaking systems with more than 256 TiB of RAM - I wonder why nobody noticed? Are there even ARM systems with that amount of RAM? I assume it's something in the order of what super computers use, but I don't think any of them are ARM. That doesn't sound that surprising to me that no one noticed if no one uses that much RAM.
  - pthariensflame 5y ago
    There’s at least one Armv8 supercomputer out there today, and it’s a beast too: https://en.m.wikipedia.org/wiki/Fugaku_(supercomputer) https://en.m.wikipedia.org/wiki/Fugaku_(supercomputer)
yosito 5y ago
As a Linux user coming from Mac, I'm super excited about this, since Apple's hardware is some of the best. As a web application developer and Linux newbie, I don't feel like I have the expertise to be able to contribute to this, but I wish I did.
- bartvk 5y ago
  Me neither, so I donate to Marcan on Patreon: https://www.patreon.com/marcan/posts https://www.patreon.com/marcan/posts (I'm not saying you should, I just thought your point is good and I wanted to point out the possibility of sponsoring to everyone here).
  - e40 5y ago
    I do, as well, and I don't even know if I'll ever use the result of this project. However, I know it's important work.
- kreetx 5y ago
  I don't think being even a linux "old-timer" would help with the low-level stuff that they are doing. (As the sibling) Supporting on Patreon is what hopefully gets us what we want quicker :)
captn3m0 5y ago
Very excited about this work, planning to move to a M1 Air once it gets a stable relase.
RMPR 5y ago
This is great and might actually prompt me to get a Mac, as much as I like ThinkPads, this battery life is simply on another level.
- ur-whale 5y ago
  > This is great and might actually prompt me to get a Mac Before you jump in, be aware that actually installing a Linux distro on one of these is not exactly trivial (and many "guides" out there are outdated). [EDIT]: as a matter of fact, I've been browsing the Asahi site for a "howto install asahi", and there's basically no such thing.
  - marcan_42 5y ago
    There's no HOWTO because the experience isn't up to end user standards; I haven't even gotten around to settling on the precise U-Boot config/boot chain yet, and I don't really encourage non-kernel-developers to try to install things at this point (though we have some folks doing it anyway). Right now the focus is very much on kernel developers doing tethered boots via USB. That will change rather soon, as enough core drivers are getting merged or usable to make this actually useful for people without very high standards (e.g. we just merged in proper keyboard/touchpad support), so now it's worth spending more time on the installer, to get it to the point where the stand-alone boot flow is where it needs to be. (If you're a developer and you want to do tethered boot and you know what you are doing, the guide is simple: make sure you have macOS 12.0.1 or newer, use diskutil to resize macOS to leave at least 3GB of unpartitioned space, boot holding down power / options / terminal, curl -L mrcn.st/alxsh | sh and follow the prompts, once you're booted into m1n1 you can use m1n1.git/proxyclient/tools/linux.py from another machine connected via USB to run a kernel / initramfs).
- LeonM 5y ago
  Don't get your expectations up too high. My ThinkPad X1E has horrible battery life under Linux, and has a boatload of issues with thunderbolt, the external dock, HDMI port, audio and WiFi. For example: when my Thinkpad goes standby with an external monitor attached via TB, some ACPI interrupt goes insane and starts burning 100% CPU resources. USB ports regularly don't work after stand-by. BMC support seems problematic as well, my battery status is often 'unknown' with the Lenovo ACPI kernel modules. And this is even with 'official' support from Lenovo for Linux. I can only imagine how bad the experience will be for running Linux on Apple products for the coming years.
chewyland 5y ago
I have been running Ubuntu on my 14 inch Mac for a little while now and it works great. This project here is also really exciting. Also, Windows ARM with the recent updates also works well in the new base 14 inch. I've been using it daily, no issues. I have nothing but praise and swiftly switching between 3 OSs sure is neat. This computer is sooo fast. I can see myself using this Mac for many years ahead.
- mlindner 5y ago
  How are you getting Windows ARM on a mac? I thought it wasn't available yet because of some kind of exclusivity agreement.
  - zamadatix 5y ago
    It'll be VM, use Parallels as it has better Windows 11 ARM64 support. You do not want to use Windows 10, it doesn't have x86-64 app compatibility like Windows 11. GPU acceleration is decent but no DX12 support yet. I've been using Windows ARM on M1 as my primary work PC device since ~January when I got a 13" M1 MacBook Pro. I now have a 16" MacBook Pro M1 Max. Can share particulars of my Parallels config or answer any questions.
- sys_64738 5y ago
  What is Linux graphics performance like? Can you comment on battery life also compared to the Mac OS running?
andy_ppp 5y ago
I will probably never install Linux on this MacBook but I'm so impressed with the technical quality of this work that I sponsor them a coffee per month through GitHub Sponsors, you can find out more here: https://asahilinux.org/support/ https://asahilinux.org/support/
hbbio 5y ago
The progress on all fronts is really impressive, and even the write-up is top notch. The pace at which the team advances is incredible, as it seems Apple support is minimal... Maybe the core tech users is negligible to Apple, but it shouldn't be: Winning over the power users generally precedes larger adoption and I wouldn't be surprised if Macs continue to gain market share in the years to come with their CPU lead.
- peoplefromibiza 5y ago
  power users come and go. general population don't need macs, a $ 200 smartphone is more than enough for the majority who would spend 1.500 euros (in Europe) for a laptop that has a limit of 16gb of RAM? only people with lots of money to waste.
  - hvgk 5y ago
    A $200 smartphone isn’t enough for most people. They’ve been sold that ideology but all I see is people struggling with them constantly assuming that’s the best of the future in their hands.
  - saagarjha 5y ago
    There's a new set of Apple silicon Macs that go all the way up to 64 GB of RAM, although you'll pay a premium for it.
    - peoplefromibiza 5y ago
      premium = an arm and a leg. My 7 years old Lenovo laptop mounts 64gb of RAM and 2 1TB ssds. It costed less than a Mac M1 Pro 7 years later.
      - saagarjha 5y ago
        I suspect that your workload that uses 64 GB of RAM and 2 1 TB SSDs could benefit from an extremely fast processor.
retskrad 5y ago
Alder Lake, Intel's flagship dekstop class CPU, is only 10-15% faster than the M1 Pro/M1 Max in CPU performance but it requires an insane amount of cooling and the power consumption is frankly ridiculous. Seems like Apple Silion will be 5 years ahead of the competition - similar to how long it took for others to catch up with the first iPhone.
- sys_64738 5y ago
  Performance per Watt only matters to Intel when they are leading. Otherwise they brute force to be "fastest". I think that the ARM in Macs is on a different trajectory than Intel. The battery life along in an ARM based Mac is second to none. That's more important to most folks, IMO.
  - specialist 5y ago
    > ...the ARM in Macs is on a different trajectory than Intel Agreed. Yes and: I'd like more analysis and punditry (predictions) comparing SoC offerings from Apple, Intel, AMD, etc. Especially wrt Apple's anticipated consumer features and markets, like AR/VR, video chat, integrated Apple Pay, whatever. For instance, I want to hear more about Apple Silicon's codecs. More about biometrics UX (Face ID, Touch ID). I think a server optimized SoC from Intel is right and proper. Ditto for Apple Silicon's strategy for mobile and media. And I don't anticipate Apple caring so much about server use cases. Maybe in-house stuff, like server farms for Siri. And I mean "market focus", vs "technical focus". While Apple Silicon is probably fine for server, I don't anticipate Apple caring, leaving those segments to existing cloud providers, eg Amazon's Graviton. Now that I think about it, I'd like to see an Apple Silicon vs Graviton matchup. Just to get a sense of the landscape. For instance, I'd guess Apple Silicon's TPU (Neural Engine) is optimized for their software stack, whereas Amazon's doing something more general purpose, as befits their respective markets. Sorry for stream of conscious, blathering for too long. I'm just writing out loud.
    - smoldesu 5y ago
      > While Apple Silicon is probably fine for server, I don't anticipate Apple caring, leaving those segments to existing cloud providers, eg Amazon's Graviton. It mostly comes down to the supply chain, from what I can see. Apple is already pushing their luck with the 5nm node, their current strategy is pretty much just keeping everyone else off the cutting edge technology so they can monopolize TSMC's manufacturing chops. They could pivot to servers, but that would cannibalize their plans for the Mac and probably force them back onto the 7nm node, which would probably throw away the majority of their bragging points right now. Plus, if history has shown us anything, people simply don't want Macs in the datacenter. xServe gave people the option, but that quickly went the way of Itanium after a year or two. Apple's history of "it's not a bug, it's a feature" and "you're holding it wrong" doesn't exactly appeal to corporate buyers who want to set-and-forget a system and have a number to call when it breaks.
Aissen 5y ago
Anyone has a pointer to what "Blocked on PSCI discussion" means ? I'd be interested to read those (mailing list?) discussions.
- marcan_42 5y ago
  It hasn't really started yet, other than some discussions on IRC, but the TL;DR is on every other ARM64 system those features are handled by either a hypervisor or a supervisor, via special PSCI calls. These machines don't have the option for a supervisor (and no nested virtualization, so an under-OS hypervisor would preclude VMs). Rather than introducing a special snowflake driver for these machines, the current thinking is we can introduce some kind of third PSCI mechanism to make these calls into code set up by the bootloader (a la EFI runtime services), without switching to a higher privilege level. That means we need to define how all that would work first.
  - phh 5y ago
    Implementing a PSCI mechanism in m1n1 sounds like a nice workaround (I could almost consider it an actual fix, even if it's not totally standard), but I'm a bit surprised: - Doesn't ARMv8 architecture require supervisor mode? You should be able to just pop a good old Arm Trusted Framework sample, implement PSCI in there just like other standard ARMv8? - Linux kernel already supports various PSCI-less architectures (rpi4 notably for armv8 if I'm not mistaken), so I don't think upstream would mind about that at all? I guess your proposed approach makes sense when the target is more than just Linux though, do you have hopes that m1n1 could be eventually suitable to boot Windows?
    - svenpeter 5y ago
      I don't think ARMv8 requires EL3 but even if it did there is no EL3 on the M1. Booting Windows natively would require Microsoft's support since some rather invasive changes to the kernel would be required (FIQ support, DART instead of SMMU) on top of implementing drivers for everything.
    - marcan_42 5y ago
terafo 5y ago
While it's obvious that within two years every GPU manufacturer is going to release MCM offering to the consumer market I'm surprised by a fact that Apple, of all companies, actually might deliver first consumer(actually more prosumer with their pricing) MCM GPU to the market. I think two M1 Max dies, if overclocked, actually can beat 3090 in most tasks and take performance crown if they release it this spring. Nvidia won't have a new GPU generation for nearly a year, same for AMD and Intel will probably shoot for performance crown only with Battlemage, which will probably be released in mid-2023.
- frant-hartm 5y ago
  MCM = multi chip module
- vbezhenar 5y ago
  Can it mine like Nvidia?
  - terafo 5y ago
    It definitely can, but it will not be price efficient since you can buy 3090 for 3k$ and that's the price of macbook pro with single m1 max chip.
    - vbezhenar 5y ago
      Hypothetical mac minis with m1 max might be of interest in that regard.
      - terafo 5y ago
        I think even mac mini will likely cost 5k given two chip config.
minimaul 5y ago
It's still not really ready for prime-time, but I've got a M1 mac mini running debian as an arm64 build bot (and a few other small things) now, and I can say that the bits that work, work really well. Storage is fast, CPU performance is solid, PCIe for ethernet works just fine. Now if I could just get one with 64G of RAM and 2TB of SSD... ;) edit: the setup process is still far from polished but that's not unexpected this early, and it's still in a 'build your own kernel' state. But it was easy enough to get going in an hour or two with a pointer or two from IRC and from reading the wiki.
- smoldesu 5y ago
  > Now if I could just get one with 64G of RAM and 2TB of SSD... ;) Careful what you wish for. With the way Apple marks up memory and storage upgrades, you might be looking at a $5,000 machine there...
  - vbezhenar 5y ago
    According to MBP 16" price, Apple charges $800 for 64 GB RAM and $800 for 2 TB SSD. IMO that's a reasonable price.
    - smoldesu 5y ago
      2 TB of NVME storage is little more than $200 off-the-shelf. 64 gigs of DDR4, high-bandwidth laptop memory costs ~$250. Even assuming Apple is springing for high-quality, high-speed RAM, $800 borders on insanity when other laptops offer similar configuration options at less than half the price. Apple's price gouging in this department is well-documented, I don't think I need to argue with HN users about that.
      - vbezhenar 5y ago
        I just checked out some random Dell Precision laptop. They want $800 for 64GB RAM and $720 for SSD. Seems pretty comparable to me. Samsung 970 Pro 1TB is $270 on newegg, so it makes $540 for 2TB. Cheaper, but it's only PCI-E 3.
noisy_boy 5y ago
Though I'm ok with my Thinkpad X1E Gen 2 (not super happy but ok), I had made up my mind about Framework being my next machine. Now with Asahi Linux's progress, build quality of the Macbook combined with flexibility of Linux makes it a serious option. Very interested to see how things turn out on both these fronts in next few years.
- juancampa 5y ago
  I'm curious, what don't you like about the X1E? I have a Gen 3 and it's super wonky with an external monitor, nvidia + linux seems to still be a mistake.
  - noisy_boy 5y ago
    Battery life isn't great (I occasionally take it out for about 2-2.5 hours which it does manage but a longer duration would probably kill it). Once froze and refused to resume when I tried to put it on standby and I had to do multiple hard resets to somehow make it going again (could be related to S3 config in BIOS) - scared me off to ever try to standby again. Also, I think you are 100% correct on Nvidia + Linux point; I did an upgrade to Nvidia driver version 460 (I think) and it basically made everything so slow that it was unusable. When I tried to downgrade, apt showed that uninstalling the driver would uninstall entire pop-desktop so had to do some hacks to downgrade back to version 450. Thankfully everything works well with that version including multiple external displays so I have locked the version via Synaptic. But yeah, not a great overall experience; someone less experienced with Linux might have (justifiably) given in to frustration and switched to Windows or Mac.
    - brokenmachine 5y ago
      Nvidia and Linux, a frustrating duo. Nvidia! Fuck you!
sohrob 5y ago
Does the method whereby they circumvent the T2 security enclave work on Intel-based Macs? I have a 2018 Mac mini which I would love to repurpose with Linux once the end-of-life support stage is reached.
- marcan_42 5y ago
  You want to look at t2linux.org. There is nothing to circumvent, it's just that some of the drivers aren't upstream yet. NVMe has been upstream for years, so you won't have any issues doing a vanilla Linux install with an up to date distro, especially on a Mac Mini (the bigger issue is no upstream keyboard support on the T2 laptops). I'm working on Wi-Fi for M1 Macs this week and that particular patch will be useful for T2 Macs too, so hopefully that will be upstream in the near future.
- unix_fan 5y ago
  There is no circumvention going on. The problem with the t2 chip is more of the lack of proper driver support, because it’s Frankenstein hardware even for Apple. There should be a driver to support apple NVME implementation, and that should get you to boot on a t2 mac.

367 points by svenpeter 5y ago

> Amusingly, while implementing support for this in Linux, we ran into a bug in Linux’s ARM SMMU support that had been there ever since 52-bit address support was introduced. This was breaking systems with more than 256 TiB of RAM - I wonder why nobody noticed? Either way, Linux now correctly supports standard ARM systems with up to 4 PiB of RAM ;-). Interesting how working on new technologies helps improve existing things for all users. Sort of reminds me how NASA invests in R&D and those investments eventually come back to military and consumers.

fragmede 5y ago
Especially handy for SpaceX, which employs a number of ex-NASA folk. SpaceX wouldn't be where they are today if it weren't for the investment in NASA.
Leherenn 5y ago
> This was breaking systems with more than 256 TiB of RAM - I wonder why nobody noticed? Are there even ARM systems with that amount of RAM? I assume it's something in the order of what super computers use, but I don't think any of them are ARM. That doesn't sound that surprising to me that no one noticed if no one uses that much RAM.
- pthariensflame 5y ago
  There’s at least one Armv8 supercomputer out there today, and it’s a beast too: https://en.m.wikipedia.org/wiki/Fugaku_(supercomputer) https://en.m.wikipedia.org/wiki/Fugaku_(supercomputer)

yosito 5y ago

As a Linux user coming from Mac, I'm super excited about this, since Apple's hardware is some of the best. As a web application developer and Linux newbie, I don't feel like I have the expertise to be able to contribute to this, but I wish I did.

bartvk 5y ago
Me neither, so I donate to Marcan on Patreon: https://www.patreon.com/marcan/posts https://www.patreon.com/marcan/posts (I'm not saying you should, I just thought your point is good and I wanted to point out the possibility of sponsoring to everyone here).
- e40 5y ago
  I do, as well, and I don't even know if I'll ever use the result of this project. However, I know it's important work.
kreetx 5y ago
I don't think being even a linux "old-timer" would help with the low-level stuff that they are doing. (As the sibling) Supporting on Patreon is what hopefully gets us what we want quicker :)

captn3m0 5y ago

Very excited about this work, planning to move to a M1 Air once it gets a stable relase.

RMPR 5y ago

This is great and might actually prompt me to get a Mac, as much as I like ThinkPads, this battery life is simply on another level.

ur-whale 5y ago
> This is great and might actually prompt me to get a Mac Before you jump in, be aware that actually installing a Linux distro on one of these is not exactly trivial (and many "guides" out there are outdated). [EDIT]: as a matter of fact, I've been browsing the Asahi site for a "howto install asahi", and there's basically no such thing.
- marcan_42 5y ago
  There's no HOWTO because the experience isn't up to end user standards; I haven't even gotten around to settling on the precise U-Boot config/boot chain yet, and I don't really encourage non-kernel-developers to try to install things at this point (though we have some folks doing it anyway). Right now the focus is very much on kernel developers doing tethered boots via USB. That will change rather soon, as enough core drivers are getting merged or usable to make this actually useful for people without very high standards (e.g. we just merged in proper keyboard/touchpad support), so now it's worth spending more time on the installer, to get it to the point where the stand-alone boot flow is where it needs to be. (If you're a developer and you want to do tethered boot and you know what you are doing, the guide is simple: make sure you have macOS 12.0.1 or newer, use diskutil to resize macOS to leave at least 3GB of unpartitioned space, boot holding down power / options / terminal, curl -L mrcn.st/alxsh | sh and follow the prompts, once you're booted into m1n1 you can use m1n1.git/proxyclient/tools/linux.py from another machine connected via USB to run a kernel / initramfs).
LeonM 5y ago
Don't get your expectations up too high. My ThinkPad X1E has horrible battery life under Linux, and has a boatload of issues with thunderbolt, the external dock, HDMI port, audio and WiFi. For example: when my Thinkpad goes standby with an external monitor attached via TB, some ACPI interrupt goes insane and starts burning 100% CPU resources. USB ports regularly don't work after stand-by. BMC support seems problematic as well, my battery status is often 'unknown' with the Lenovo ACPI kernel modules. And this is even with 'official' support from Lenovo for Linux. I can only imagine how bad the experience will be for running Linux on Apple products for the coming years.

chewyland 5y ago

I have been running Ubuntu on my 14 inch Mac for a little while now and it works great. This project here is also really exciting. Also, Windows ARM with the recent updates also works well in the new base 14 inch. I've been using it daily, no issues. I have nothing but praise and swiftly switching between 3 OSs sure is neat. This computer is sooo fast. I can see myself using this Mac for many years ahead.

mlindner 5y ago
How are you getting Windows ARM on a mac? I thought it wasn't available yet because of some kind of exclusivity agreement.
- zamadatix 5y ago
  It'll be VM, use Parallels as it has better Windows 11 ARM64 support. You do not want to use Windows 10, it doesn't have x86-64 app compatibility like Windows 11. GPU acceleration is decent but no DX12 support yet. I've been using Windows ARM on M1 as my primary work PC device since ~January when I got a 13" M1 MacBook Pro. I now have a 16" MacBook Pro M1 Max. Can share particulars of my Parallels config or answer any questions.
sys_64738 5y ago
What is Linux graphics performance like? Can you comment on battery life also compared to the Mac OS running?

andy_ppp 5y ago

I will probably never install Linux on this MacBook but I'm so impressed with the technical quality of this work that I sponsor them a coffee per month through GitHub Sponsors, you can find out more here: https://asahilinux.org/support/ https://asahilinux.org/support/

hbbio 5y ago

The progress on all fronts is really impressive, and even the write-up is top notch. The pace at which the team advances is incredible, as it seems Apple support is minimal... Maybe the core tech users is negligible to Apple, but it shouldn't be: Winning over the power users generally precedes larger adoption and I wouldn't be surprised if Macs continue to gain market share in the years to come with their CPU lead.

peoplefromibiza 5y ago
power users come and go. general population don't need macs, a $ 200 smartphone is more than enough for the majority who would spend 1.500 euros (in Europe) for a laptop that has a limit of 16gb of RAM? only people with lots of money to waste.
- hvgk 5y ago
  A $200 smartphone isn’t enough for most people. They’ve been sold that ideology but all I see is people struggling with them constantly assuming that’s the best of the future in their hands.
- saagarjha 5y ago
  There's a new set of Apple silicon Macs that go all the way up to 64 GB of RAM, although you'll pay a premium for it.
  - peoplefromibiza 5y ago
    premium = an arm and a leg. My 7 years old Lenovo laptop mounts 64gb of RAM and 2 1TB ssds. It costed less than a Mac M1 Pro 7 years later.
    - saagarjha 5y ago
      I suspect that your workload that uses 64 GB of RAM and 2 1 TB SSDs could benefit from an extremely fast processor.

retskrad 5y ago

Alder Lake, Intel's flagship dekstop class CPU, is only 10-15% faster than the M1 Pro/M1 Max in CPU performance but it requires an insane amount of cooling and the power consumption is frankly ridiculous. Seems like Apple Silion will be 5 years ahead of the competition - similar to how long it took for others to catch up with the first iPhone.

sys_64738 5y ago
Performance per Watt only matters to Intel when they are leading. Otherwise they brute force to be "fastest". I think that the ARM in Macs is on a different trajectory than Intel. The battery life along in an ARM based Mac is second to none. That's more important to most folks, IMO.
- specialist 5y ago
  > ...the ARM in Macs is on a different trajectory than Intel Agreed. Yes and: I'd like more analysis and punditry (predictions) comparing SoC offerings from Apple, Intel, AMD, etc. Especially wrt Apple's anticipated consumer features and markets, like AR/VR, video chat, integrated Apple Pay, whatever. For instance, I want to hear more about Apple Silicon's codecs. More about biometrics UX (Face ID, Touch ID). I think a server optimized SoC from Intel is right and proper. Ditto for Apple Silicon's strategy for mobile and media. And I don't anticipate Apple caring so much about server use cases. Maybe in-house stuff, like server farms for Siri. And I mean "market focus", vs "technical focus". While Apple Silicon is probably fine for server, I don't anticipate Apple caring, leaving those segments to existing cloud providers, eg Amazon's Graviton. Now that I think about it, I'd like to see an Apple Silicon vs Graviton matchup. Just to get a sense of the landscape. For instance, I'd guess Apple Silicon's TPU (Neural Engine) is optimized for their software stack, whereas Amazon's doing something more general purpose, as befits their respective markets. Sorry for stream of conscious, blathering for too long. I'm just writing out loud.
  - smoldesu 5y ago
    > While Apple Silicon is probably fine for server, I don't anticipate Apple caring, leaving those segments to existing cloud providers, eg Amazon's Graviton. It mostly comes down to the supply chain, from what I can see. Apple is already pushing their luck with the 5nm node, their current strategy is pretty much just keeping everyone else off the cutting edge technology so they can monopolize TSMC's manufacturing chops. They could pivot to servers, but that would cannibalize their plans for the Mac and probably force them back onto the 7nm node, which would probably throw away the majority of their bragging points right now. Plus, if history has shown us anything, people simply don't want Macs in the datacenter. xServe gave people the option, but that quickly went the way of Itanium after a year or two. Apple's history of "it's not a bug, it's a feature" and "you're holding it wrong" doesn't exactly appeal to corporate buyers who want to set-and-forget a system and have a number to call when it breaks.

Aissen 5y ago

Anyone has a pointer to what "Blocked on PSCI discussion" means ? I'd be interested to read those (mailing list?) discussions.

marcan_42 5y ago
It hasn't really started yet, other than some discussions on IRC, but the TL;DR is on every other ARM64 system those features are handled by either a hypervisor or a supervisor, via special PSCI calls. These machines don't have the option for a supervisor (and no nested virtualization, so an under-OS hypervisor would preclude VMs). Rather than introducing a special snowflake driver for these machines, the current thinking is we can introduce some kind of third PSCI mechanism to make these calls into code set up by the bootloader (a la EFI runtime services), without switching to a higher privilege level. That means we need to define how all that would work first.
- phh 5y ago
  Implementing a PSCI mechanism in m1n1 sounds like a nice workaround (I could almost consider it an actual fix, even if it's not totally standard), but I'm a bit surprised: - Doesn't ARMv8 architecture require supervisor mode? You should be able to just pop a good old Arm Trusted Framework sample, implement PSCI in there just like other standard ARMv8? - Linux kernel already supports various PSCI-less architectures (rpi4 notably for armv8 if I'm not mistaken), so I don't think upstream would mind about that at all? I guess your proposed approach makes sense when the target is more than just Linux though, do you have hopes that m1n1 could be eventually suitable to boot Windows?
  - svenpeter 5y ago
    I don't think ARMv8 requires EL3 but even if it did there is no EL3 on the M1. Booting Windows natively would require Microsoft's support since some rather invasive changes to the kernel would be required (FIQ support, DART instead of SMMU) on top of implementing drivers for everything.
  - marcan_42 5y ago

terafo 5y ago

While it's obvious that within two years every GPU manufacturer is going to release MCM offering to the consumer market I'm surprised by a fact that Apple, of all companies, actually might deliver first consumer(actually more prosumer with their pricing) MCM GPU to the market. I think two M1 Max dies, if overclocked, actually can beat 3090 in most tasks and take performance crown if they release it this spring. Nvidia won't have a new GPU generation for nearly a year, same for AMD and Intel will probably shoot for performance crown only with Battlemage, which will probably be released in mid-2023.

frant-hartm 5y ago
MCM = multi chip module
vbezhenar 5y ago
Can it mine like Nvidia?
- terafo 5y ago
  It definitely can, but it will not be price efficient since you can buy 3090 for 3k$ and that's the price of macbook pro with single m1 max chip.
  - vbezhenar 5y ago
    Hypothetical mac minis with m1 max might be of interest in that regard.
    - terafo 5y ago
      I think even mac mini will likely cost 5k given two chip config.

minimaul 5y ago

It's still not really ready for prime-time, but I've got a M1 mac mini running debian as an arm64 build bot (and a few other small things) now, and I can say that the bits that work, work really well. Storage is fast, CPU performance is solid, PCIe for ethernet works just fine. Now if I could just get one with 64G of RAM and 2TB of SSD... ;) edit: the setup process is still far from polished but that's not unexpected this early, and it's still in a 'build your own kernel' state. But it was easy enough to get going in an hour or two with a pointer or two from IRC and from reading the wiki.

smoldesu 5y ago
> Now if I could just get one with 64G of RAM and 2TB of SSD... ;) Careful what you wish for. With the way Apple marks up memory and storage upgrades, you might be looking at a $5,000 machine there...
- vbezhenar 5y ago
  According to MBP 16" price, Apple charges $800 for 64 GB RAM and $800 for 2 TB SSD. IMO that's a reasonable price.
  - smoldesu 5y ago
    2 TB of NVME storage is little more than $200 off-the-shelf. 64 gigs of DDR4, high-bandwidth laptop memory costs ~$250. Even assuming Apple is springing for high-quality, high-speed RAM, $800 borders on insanity when other laptops offer similar configuration options at less than half the price. Apple's price gouging in this department is well-documented, I don't think I need to argue with HN users about that.
    - vbezhenar 5y ago
      I just checked out some random Dell Precision laptop. They want $800 for 64GB RAM and $720 for SSD. Seems pretty comparable to me. Samsung 970 Pro 1TB is $270 on newegg, so it makes $540 for 2TB. Cheaper, but it's only PCI-E 3.

noisy_boy 5y ago

Though I'm ok with my Thinkpad X1E Gen 2 (not super happy but ok), I had made up my mind about Framework being my next machine. Now with Asahi Linux's progress, build quality of the Macbook combined with flexibility of Linux makes it a serious option. Very interested to see how things turn out on both these fronts in next few years.

juancampa 5y ago
I'm curious, what don't you like about the X1E? I have a Gen 3 and it's super wonky with an external monitor, nvidia + linux seems to still be a mistake.
- noisy_boy 5y ago
  Battery life isn't great (I occasionally take it out for about 2-2.5 hours which it does manage but a longer duration would probably kill it). Once froze and refused to resume when I tried to put it on standby and I had to do multiple hard resets to somehow make it going again (could be related to S3 config in BIOS) - scared me off to ever try to standby again. Also, I think you are 100% correct on Nvidia + Linux point; I did an upgrade to Nvidia driver version 460 (I think) and it basically made everything so slow that it was unusable. When I tried to downgrade, apt showed that uninstalling the driver would uninstall entire pop-desktop so had to do some hacks to downgrade back to version 450. Thankfully everything works well with that version including multiple external displays so I have locked the version via Synaptic. But yeah, not a great overall experience; someone less experienced with Linux might have (justifiably) given in to frustration and switched to Windows or Mac.
  - brokenmachine 5y ago
    Nvidia and Linux, a frustrating duo. Nvidia! Fuck you!

sohrob 5y ago

Does the method whereby they circumvent the T2 security enclave work on Intel-based Macs? I have a 2018 Mac mini which I would love to repurpose with Linux once the end-of-life support stage is reached.

marcan_42 5y ago
You want to look at t2linux.org. There is nothing to circumvent, it's just that some of the drivers aren't upstream yet. NVMe has been upstream for years, so you won't have any issues doing a vanilla Linux install with an up to date distro, especially on a Mac Mini (the bigger issue is no upstream keyboard support on the T2 laptops). I'm working on Wi-Fi for M1 Macs this week and that particular patch will be useful for T2 Macs too, so hopefully that will be upstream in the near future.
unix_fan 5y ago
There is no circumvention going on. The problem with the t2 chip is more of the lack of proper driver support, because it’s Frankenstein hardware even for Apple. There should be a driver to support apple NVME implementation, and that should get you to boot on a t2 mac.