multiboot2 protocol on ARM with ACPI

multiboot2 protocol on ARM with ACPI

[Daniel Kiper]

Hi folks,

I heard that ARM server platforms will be provided with ACPI.
Is it true? If yes then how multiboot2 protocol for ARM
will work on such hardware? As I know it was designed for
device tree and it is tightly linked with it. So?

Daniel

Re: multiboot2 protocol on ARM with ACPI

[Fu Wei]

On 09/24/2014 09:11 PM, Daniel Kiper wrote:
> Hi folks,
> 
> I heard that ARM server platforms will be provided with ACPI.
> Is it true? If yes then how multiboot2 protocol for ARM
> will work on such hardware? 

Hi Daniel Kiper,

For now, we are *NOT* using multiboot2 in GRUB for multiboot(xen)
Please check here:
http://lists.linaro.org/pipermail/linaro-uefi/2014-September/000521.html

Let me know if you need more info about this  :-)

FYI,https://wiki.linaro.org/LEG/Engineering/Grub2/Xen_booting_on_FVP_Base_AEMv8A

> As I know it was designed for
> device tree and it is tightly linked with it. So?
> 
> Daniel
> 


-- 
Best regards,

Fu Wei
Enterprise Server Engineer From Red Hat
LEG Team
Linaro.org | Open source software for ARM SoCs
Ph: +86 186 2020 4684 (mobile)
IRC: fuwei
Skype: tekkamanninja
Room 1512, Regus One Corporate Avenue,Level 15,
One Corporate Avenue,222 Hubin Road,Huangpu District,
Shanghai,China 200021 

Re: multiboot2 protocol on ARM with ACPI

[Ian Campbell]

On Wed, 2014-09-24 at 15:11 +0200, Daniel Kiper wrote:
> Hi folks,
> 
> I heard that ARM server platforms will be provided with ACPI.
> Is it true?

Yes.

>  If yes then how multiboot2 protocol for ARM
> will work on such hardware? As I know it was designed for
> device tree and it is tightly linked with it. So?

The thing (I think) you are speaking about is not strictly "multiboot2"
as defined by the grub project (or whoever). Rather it is the thing
which we happen to call multiboot which is described at:
http://wiki.xen.org/wiki/Xen_ARM_with_Virtualization_Extensions/Multiboot

Anyway, with that out of the way...

ARM systems with ACPI still use DT for certain things, including the
initrd and boot modules (stuff which in speak is in the "/chosen node").
It's the platform/hardware side of things (i.e. all the stuff
outside /chosen on a DT system) which is described via ACPI.

Ian.

Re: multiboot2 protocol on ARM with ACPI

[Daniel Kiper]

On Wed, Sep 24, 2014 at 02:43:02PM +0100, Ian Campbell wrote:
> On Wed, 2014-09-24 at 15:11 +0200, Daniel Kiper wrote:
> > Hi folks,
> >
> > I heard that ARM server platforms will be provided with ACPI.
> > Is it true?
>
> Yes.
>
> >  If yes then how multiboot2 protocol for ARM
> > will work on such hardware? As I know it was designed for
> > device tree and it is tightly linked with it. So?
>
> The thing (I think) you are speaking about is not strictly "multiboot2"
> as defined by the grub project (or whoever). Rather it is the thing

Yep, I know.

> which we happen to call multiboot which is described at:
> http://wiki.xen.org/wiki/Xen_ARM_with_Virtualization_Extensions/Multiboot

I thought exactly about that. Anyway, name is a bit unfortunate because
it clashes with name of legacy boot multiboot protocol for x86, MIPS and
maybe others which is defined in completely different way. Once I even heard
that "The thing" is called multiboot2 for ARM. That is why I put multiboot2
in this email, ehh...

> Anyway, with that out of the way...
>
> ARM systems with ACPI still use DT for certain things, including the
> initrd and boot modules (stuff which in speak is in the "/chosen node").
> It's the platform/hardware side of things (i.e. all the stuff
> outside /chosen on a DT system) which is described via ACPI.

Thank you for explanation. I thought that this could be done more or
less in that way but I wanted to be sure.

Daniel

Re: multiboot2 protocol on ARM with ACPI

[Ian Campbell]

On Wed, 2014-09-24 at 20:20 +0200, Daniel Kiper wrote:
> > which we happen to call multiboot which is described at:
> > http://wiki.xen.org/wiki/Xen_ARM_with_Virtualization_Extensions/Multiboot
> 
> I thought exactly about that. Anyway, name is a bit unfortunate because
> it clashes with name of legacy boot multiboot protocol for x86, MIPS and
> maybe others

This was partially deliberate since we wanted the existing grub.cfg
multiboot syntax ("multiboot <foo>; kernel <bar>; etc") people were used
to (and which the tools generate already) to Just Work.

AFAIK multiboot(1) is mostly x86 specific, or at least will never be
implemented on on ARM. multiboot2 is something else which is newer and
more cross platform[0] and it uses a different grub.cfg syntax
(multiboot2 <foo> I think).

Which is why we felt it would be ok to reuse the multiboot1 grub syntax
for this ARM thing we invented.

Ian.

[0] i.e. supports MIPS and maybe others, possibly including ARM,
although I don't know of anyone who is working on that. Xen on ARM
doesn't really find multiboot2 useful/interesting, although if it became
widely used we would of course think about implementing support for it.

>  which is defined in completely different way. Once I even heard
> that "The thing" is called multiboot2 for ARM. That is why I put multiboot2
> in this email, ehh...
> 
> > Anyway, with that out of the way...
> >
> > ARM systems with ACPI still use DT for certain things, including the
> > initrd and boot modules (stuff which in speak is in the "/chosen node").
> > It's the platform/hardware side of things (i.e. all the stuff
> > outside /chosen on a DT system) which is described via ACPI.
> 
> Thank you for explanation. I thought that this could be done more or
> less in that way but I wanted to be sure.
> 
> Daniel