[PATCH] acpi: move acpi_get_rsdp_addr() to acpi/acpi_table.h

[Simon Glass]

Hi Tom, Heinrich,

On Thu, 9 Nov 2023 at 13:42, Tom Rini <trini at konsulko.com> wrote:
>
> On Thu, Nov 09, 2023 at 12:28:26PM -0800, Heinrich Schuchardt wrote:
> > On 11/9/23 11:24, Tom Rini wrote:
> > > On Thu, Nov 09, 2023 at 09:23:02AM -0800, Heinrich Schuchardt wrote:
> > >
> > > > Function acpi_get_rsdp_addr() is needed on all architectures which
> > > > write ACPI tables. Move the definition from the x86 include to an
> > > > architecture independent one.
> > > >
> > > > Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt at canonical.com>
> > > > ---
> > > >   arch/x86/include/asm/acpi_table.h | 9 ---------
> > > >   drivers/misc/qfw.c                | 1 +
> > > >   include/acpi/acpi_table.h         | 9 +++++++++
> > > >   3 files changed, 10 insertions(+), 9 deletions(-)
> > >
> > > My question here is, does this work right on non-x86? I know you can
> > > have ACPI without UEFI on x86, but elsewhere doesn't the location have
> > > to be provided in some manner by UEFI? I know I'm thinking that's the
> > > case with SMBIOS (with a few exception).
> > >
> > > I'm assuming you're doing something here with qemu and its qfw interface
> > > and risc-v.
> > >
> >
> > Hello Tom,
> >
> > yes, I want to enable ACPI passthrough on QEMU RISC-V. There will be more
> > patches needed. But the current change seemed obvious looking at the code so
> > I did not want to pile up more patches first.
> >
> > Do you know what the status of ACPI passthrough is on ARM?
>
> On U-Boot? No. I know a while back I asked Simon to, and he did some
> hacking such that on a Raspberry Pi you could do UEFI boot and pass
> through some ACPI tables and it was somewhat functional (which was I
> think an issue with the ACPI tables as much as anything else). And I

Yes it worked OK, at least to a point. See [1]

> mentioned to some Linaro/EBBR people that it was because in part I don't
> object to U-Boot being able to do SystemReady ES level support either
> (and to be clear, I don't mean instead of, or preferred over IR, just
> that ES mandates ACPI and if we can pass it through .... and yes,
> there's more to it than that, I know).

This is what U-Boot does now when booted from coreboot. So it should
be fairly easy to arrange.

There is a blockage in Linux re booting with ACPI but without EFI,
though. I suppose it doesn't affect U-Boot right now since it can boot
with EFI, but it is causing problems with other projects.

>
> And yeah, I see now I had mis-interpreted the "rsdp" portion of the
> function name.
>
> Reviewed-by: Tom Rini <trini at konsulko.com>
>
> --
> Tom

[1] https://patchwork.ozlabs.org/project/uboot/list/?series=274675&state=*