ACPI vs DT at runtime

[Olof Johansson]

On Sun, Nov 17, 2013 at 10:10 AM, Arnd Bergmann <arnd at arndb.de> wrote:
> On Sunday 17 November 2013 17:18:03 Stefano Stabellini wrote:
>> On Fri, 15 Nov 2013, Olof Johansson wrote:
>> > On Fri, Nov 15, 2013 at 12:58 PM, Arnd Bergmann <arnd at arndb.de> wrote:
>> > > On Friday 15 November 2013, Olof Johansson wrote:
>> > >> So, I'm strongly urging that whatever the server guys try to do, it
>> > >> will in the end result in the ACPI data being translated into DT
>> > >> equivalents, such that the kernel only needs to handle data via DT.
>> > >
>> > > I don't think that a translation layer is the answer, I see the problem
>> > > more in things that cannot be translated automatically. The parts that
>> > > are similar enough to allow translation could also just be handled by
>> > > a thin abstraction layer in the kernel, which I think we will see
>> > > on embedded x86 with DT-in-ACPI-syntax.
>> >
>> > I'm not so sure that converting everything yet again to another
>> > abstraction layer is a good solution. We've got one level of
>> > abstraction today -- DT. And we've got the un-abstracted layer of
>> > platform devices. Churning all drivers yet again seems like the wrong
>> > way to do this. For things that we can pre-populate instead of adding
>> > runtime abstraction, we should.
>
> My point was not that everything would be good if we change the kernel
> this way, it clearly wouldn't. I think the problem is more the parts
> for which neither an automated translation nor a unified driver level
> interface would work well.

Do you have predictions for what those areas would be? Chances are
that at least the initial ones can be handled with the methods we
already have today such as quirks and the like.

Once we've seen enough of them, we'll have more knowledge of what a
new solution really needs to cover. Speculating prematurely will
likely give us the wrong abstractions (and/or at too high a level,
causing extra work and churn for everybody). I think we agree that
holding off is the right answer, just maybe not (yet) what the actual
solution might be when we start to need it. :)

>> Simply using DT would help avoiding the awkward situation where a driver
>> of a device only works with one of the two description formats and not
>> the other.
>
> Yes, but remember that Intel still have the problem on their embedded
> systems, and will want to solve them.
>
> For ARM Linux I agree that we're better off not having to change all
> the drivers again for this, but we will have a growing number of
> drivers that are shared with ACPI based x86 SoCs. At least there are
> only one or two vendors of those chips ;-)

Well, you menion DT-in-ACPI above. If they're really exploring that
path then a lot will sort itself out from there. Making a decision to
hold off with ACPI for ARM will help get everybody on the same page if
they're already considering a solution along those lines.

>> > > I think we can still treat ACPI on ARM64 as a beginner's mistake and
>> > > provide hand-written DT blobs for the few systems that start shipping
>> > > with that.
>> >
>> > I think we can do better -- we can ask those vendors to not ship ACPI
>> > at all, and ship DT themselves (together with us for review, etc).

Yep, and together for review is a critical part. I'm not saying that
the ideal solution is a flashed DTB, but it's better than ACPI. A
flashed DTB that's _easy to update_ as part of an OS install would
probably be one of the best solutions we could ask for though.

>> They can ship with ACPI if they want to, what is important is that they
>> ship with device tree too.
>> Encouraging them to do that is definitely the right thing to do today,
>> regardless of the medium to long term ACPI strategy for the Linux
>> kernel.
>
> I'm skeptical about getting people to ship both, except if they
> want to support multiple operating systems. For those vendors
> we are talking to, we are possibly able to convince them to drop
> ACPI for Linux based servers.

+1. Especially if their initial focus is just on linux and not on
enabling RT on the same set of machines (with the same firmware).

> The bigger problem is system vendors
> we don't talk to. They are going to to do any number of crazy things
> in their private kernels:
>
> * Board files
> * Hardcoded platform devices with no multiplatform support
> * Custom hypervisor (EL2 or EL3) interfaces for probing
> * FEX
> * ACPI
> * Some other firmware ported over from their MIPS products
> * Incompatible Open Firmware
> * Incompatible DT extensions
> * ...

Given that we have a documented boot protocol today, it shouldn't
matter much if someone ports over some MIPS firmware (we did add OF
client interface to CFE at PA Semi, and that worked well for us :-).
Most of the others listed above can hopefully be avoided by us having
appropriate solutions in-kernel today that they can reuse. I.e. FEX
happened because DT didn't exist at the time.

ACPI is a big problem here in that sense: Vendors have been told to
use it, but there's no infrastructure, and even if it would be there
it would be immature and we don't really know what we want from it.

Board files and hardcoded devices are things we've dealt with already
and know what to do with, so I'm less worried about that. And finally,
custom firmware interfaces should hopefully be rare given the
reference PSCI implementation from ARM, but we can deal with it if we
have to.

> The only thing we really have a handle on is stuff that gets submitted
> for inclusion into Linux. One would hope that this would include
> any server platform, but I think the people trying get into that
> market are still learning about how to do these things.

Yes, we're likely to see some false starts, or even successful starts
done wrong. Hopefully we can catch them early and get them onto a
better track quickly. Especially once they realize they really do
want/need upstream support for their platforms.

I do hope we'll start to see more from the vendors that are working on
silicon soon, or things run the risk of getting really messy. :(

>> > Especially since doing a broken ACPI implementation today _just for
>> > us_ will just distract them. If they need one for RT, fine. As I
>> > mentioned elsewhere, shipping a flashed DTB is no different from
>> > shipping ACPI from a future-proof point of view; we'll end up
>> > overriding either at some point once we figure out things better.
>>
>> Well said.
>
> +1


-Olof