ACPI For 64-bit ARM Proposed For Linux 4.1

[phoronix]

Phoronix: ACPI For 64-bit ARM Proposed For Linux 4.1

Going back many months there's been work on adding ACPI support to ARM64/AArch64. That long journey may now be wrapping up with a pending pull request for landing full ACPI support for 64-bit ARM in Linux 4.1...

http://www.phoronix.com/scan.php?page=news_item&px=Linux-4.1-ARM64-ACPI

[plonoma]

How is ARM fragmentation being handled?
What is happening with ARM platform?
Why ARM are you not working on this?
When or even will we ever see a sort of standard platform for all ARM SoC's.
Not just one for servers.
http://www.anandtech.com/show/7721/a...tform-standard

kpb321 - Wednesday, January 29, 2014 - link
From what I understand this is a bigger deal than it would seem. Unlike x86 where pretty much everyone builds cpus/chipset with a set baseline that is needed to boot windows ARM SOCs aren't as standardized and might have different types of buses available, might handle the boot sequence differently etc which is a difficulty for the OS and part of the reason why most ARM SOCs can't just compile and run the standard linux kernel.

Interestingly, Microsoft did something similar for their Windows Phone and tablets to make things easier on them.

[vadix]

How is ARM fragmentation being handled?
What is happening with ARM platform?
Why ARM are you not working on this?
When or even will we ever see a sort of standard platform for all ARM SoC's.
Not just one for servers.
http://www.anandtech.com/show/7721/a...tform-standard

That would be nice, but ARM isn't a rich company; they only make a flat licensing fee to each ARM producer, so their resources and reach is limited. Companies should naturally start to use Coreboot and ACPI now that they are ported to ARM, and it might just become a standard without any intervention. It should be an incentive for them to use these systems because it makes the user's and the company's life easier, at least for consumer systems.

[GreatEmerald]

That would be nice, but ARM isn't a rich company; they only make a flat licensing fee to each ARM producer, so their resources and reach is limited. Companies should naturally start to use Coreboot and ACPI now that they are ported to ARM, and it might just become a standard without any intervention. It should be an incentive for them to use these systems because it makes the user's and the company's life easier, at least for consumer systems.

I'm not sure that using ACPI is a good thing... It's a major pain to deal with it. Like the whole situation of not being able to use lm_sensors due to ACPI...

As for Coreboot, yeah right... Not going to happen. Companies have their own bootloaders, and they don't see any value in using anything standard. God forbid the user tries to install something they weren't supposed to on their ARM devices...

[Adarion]

I'm not sure that using ACPI is a good thing... It's a major pain to deal with it. Like the whole situation of not being able to use lm_sensors due to ACPI...

As for Coreboot, yeah right... Not going to happen. Companies have their own bootloaders, and they don't see any value in using anything standard. God forbid the user tries to install something they weren't supposed to on their ARM devices...

Seconded. But I'd really support the idea of standardization and general use of coreboot.

> fragmentation

The fragmentation of ARM driver code in the kernel made iirc. Linus explode once. We had a nice talk about this mess by one of the maintainers (Heiko St?bner) recently here on a Linux Day in Germany. I don't remember if it was Linus or another maintainer / dev who said trying to deal with all the ARMs is like herding cats.
(There are audio recordings https://chemnitzer.linux-tage.de/2015/de/programm/plan but most of them are in German, though)

I am not sure if this hell called ACPI will improve things that much. Maybe if ACPI was reduced to a reasonable size and without any Microsoft interference.

[curaga]

That would be nice, but ARM isn't a rich company; they only make a flat licensing fee to each ARM producer, so their resources and reach is limited.

http://www.design-reuse.com/news/37165/arm-q1-2015-financial-results.html

They made 120 million dollars profit in Q1. Totally not a rich company.

[DrYak]

As for Coreboot, yeah right... Not going to happen. Companies have their own bootloaders, and they don't see any value in using anything standard.

Actually, several companies sees Coreboot as a way to get a cheap (mostly free) firmware.

And a few companies are actively developing/using coreboot.
- Google is notoriously using it a lot on their chromebooks (as a cheap source of BIOS/firmware).
- AMD provide a lot of code and spec (basically, if your platform is AMD, all the northbridge/etc. functionnality in the CPU package are 100% supported)
- motherboard manufacturer like MSI actively helps developping (and at some point Gigabyte and Tyan, not sure what exactly their current contributions are. Specs only? Code? Dev on payroll ?) mainly because coreboot gives them another firmware alternative with no license fee required.

And remember that:
- AMD has an AArch64 CPU in the pipeline (to be used interchangeably with the x86_64) (the whole K12 / Zen generation)
- MSI and Gigabyte are very active at producing motherboards for AMD platforms.
That means probably that in less than 1 year, you're almost guaranteed to be able to build a server that runs on AMD's ARM, and runs coreboot firmware.

So, although you won't probably find coreboot on Windows RT tablets and smartphones, coreboot on ARM is here to stay, and you'll probably see it on a whole other range of hardware (mostly servers).

They made 120 million dollars profit in Q1. Totally not a rich company.

Compared to Intel who made ~2 bilions and AMD who only made half of that (~1 billion)...
Yup, ARM only have 1/10 to 1/20 of the budget. It's normal that they don't have the same kind of ressources.

(Mainly due to the explanation given above: ARM only sell very cheap license to their architecture.
ARM CPUs are a huge market, but only when you total all the actor. The total revenues are big, but they are distributed over a lots of companies fabbing their own chips.)

[curaga]

Hiring 1 coreboot dev from that 120 million profit -> oh no, 119.9 million profit.