Re: [PATCH 1/5] arm: apple: Add initial support for Apple's M1 SoC

[Simon Glass <sjg@chromium.org>]

Hi Mark,

On Mon, 20 Sept 2021 at 02:49, Mark Kettenis <mark.kettenis@xs4all.nl> wrote:
>
> > From: Simon Glass <sjg@chromium.org>
> > Date: Sun, 19 Sep 2021 21:15:57 -0600
> >
> > Hi Mark,
> >
> > On Sat, 18 Sept 2021 at 07:55, Mark Kettenis <kettenis@openbsd.org> wrote:
> > >
> > > Add support for Apple's M1 SoC that is used in "Apple Silicon"
> > > Macs.  This builds a basic U-Boot that can be used as a payload
> > > for the m1n1 boot loader being developed by the Asahi Linux
> > > project.
> > >
> > > Signed-off-by: Mark Kettenis <kettenis@openbsd.org>
> > > ---
> > >  arch/arm/Kconfig                    |  22 ++++
> > >  arch/arm/Makefile                   |   1 +
> > >  arch/arm/mach-apple/Kconfig         |  18 ++++
> > >  arch/arm/mach-apple/Makefile        |   4 +
> > >  arch/arm/mach-apple/board.c         | 158 ++++++++++++++++++++++++++++
> > >  arch/arm/mach-apple/lowlevel_init.S |  16 +++
> > >  configs/apple_m1_defconfig          |  14 +++
> > >  include/configs/apple.h             |  38 +++++++
> > >  8 files changed, 271 insertions(+)
> > >  create mode 100644 arch/arm/mach-apple/Kconfig
> > >  create mode 100644 arch/arm/mach-apple/Makefile
> > >  create mode 100644 arch/arm/mach-apple/board.c
> > >  create mode 100644 arch/arm/mach-apple/lowlevel_init.S
> > >  create mode 100644 configs/apple_m1_defconfig
> > >  create mode 100644 include/configs/apple.h
> > >
> > > diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig
> > > index b5bd3284cd..7cdea1f615 100644
> > > --- a/arch/arm/Kconfig
> > > +++ b/arch/arm/Kconfig
> > > @@ -895,6 +895,26 @@ config ARCH_NEXELL
> > >         select DM
> > >         select GPIO_EXTRA_HEADER
> > >
> > > +config ARCH_APPLE
> > > +       bool "Apple SoCs"
> > > +       select ARM64
> > > +       select LINUX_KERNEL_IMAGE_HEADER
> > > +       select POSITION_INDEPENDENT
> > > +       select BLK
> > > +       select DM
> > > +       select DM_KEYBOARD
> > > +       select DM_SERIAL
> > > +       select DM_USB
> > > +       select DM_VIDEO
> > > +       select CMD_USB
> > > +       select MISC
> > > +       select OF_CONTROL
> > > +       select OF_BOARD
> > > +       select USB
> > > +       imply CMD_DM
> > > +       imply CMD_GPT
> > > +       imply DISTRO_DEFAULTS
> >
> > Suggest sorting these
>
> As in sort all the selects alphabetically and sort all the implies
> alphabetically seperately?

I suppose so.

>
> Does my use of impy here even make sense?

It allows people to disable it, whereas select does not. It looks OK to me.

>
> This whole Kconfig stuff is a bit alien to me and I must say that it
> isn't obvious what "best-practice" is in this area...

We try to avoid select so only use it if it really cannot build/run
without that feature. From what I can tell you have done that.

I know this is a different project, but there are some ideas here:

https://docs.zephyrproject.org/1.14.0/guides/kconfig/index.html

[..]

> > > +int dram_init(void)
> > > +{
> > > +       int index, node, ret;
> > > +       fdt_addr_t base;
> > > +       fdt_size_t size;
> > > +
> > > +       ret = fdtdec_setup_mem_size_base();
> > > +       if (ret)
> > > +               return ret;
> > > +
> > > +       /* Update RAM mapping. */
> >
> > Do you want the period at the end of that?
>
> Given that Bin brought up the same thing, I gather that not having a
> full stop at the end of single-line comments is preferred?

Yes, the periods just clutter things and people start adding them only
when it is a valid sentence, etc. Best to just leave them out.

>
> > > +       index = ARRAY_SIZE(apple_mem_map) - 3;
> > > +       apple_mem_map[index].virt = gd->ram_base;
> > > +       apple_mem_map[index].phys = gd->ram_base;
> > > +       apple_mem_map[index].size = gd->ram_size;
> > > +
> > > +       node = fdt_path_offset(gd->fdt_blob, "/chosen/framebuffer");
> >
> > Can you use the ofnode interface here and below?
>
> I can try...

Something like:

ofnode node;

node = ofnode_path("/chosen/framebuffer");

then use ofnode_get_addr_size()

>
> > > +       if (node < 0)
> > > +               return 0;
> > > +
> > > +       base = fdtdec_get_addr_size(gd->fdt_blob, node, "reg", &size);
> > > +       if (base == FDT_ADDR_T_NONE)
> > > +               return 0;
> > > +
> > > +       /* Add framebuffer mapping. */
> > > +       index = ARRAY_SIZE(apple_mem_map) - 2;
> > > +       apple_mem_map[index].virt = base;
> > > +       apple_mem_map[index].phys = base;
> > > +       apple_mem_map[index].size = size;
> > > +       apple_mem_map[index].attrs = PTE_BLOCK_MEMTYPE(MT_NORMAL_NC) |
> > > +               PTE_BLOCK_INNER_SHARE | PTE_BLOCK_PXN | PTE_BLOCK_UXN;
> > > +
> > > +       return 0;
> > > +}

[..]

> > > diff --git a/arch/arm/mach-apple/lowlevel_init.S b/arch/arm/mach-apple/lowlevel_init.S
> > > new file mode 100644
> > > index 0000000000..0f5313163e
> > > --- /dev/null
> > > +++ b/arch/arm/mach-apple/lowlevel_init.S
> > > @@ -0,0 +1,16 @@
> > > +/* SPDX-License-Identifier: GPL-2.0+ */
> > > +/*
> > > + * (C) Copyright 2021 Mark Kettenis <kettenis@openbsd.org>
> > > + */
> > > +
> > > +.align 8
> > > +.global fw_dtb_pointer
> > > +fw_dtb_pointer:
> > > +       .quad   0
> > > +
> > > +.global save_boot_params
> > > +save_boot_params:
> > > +       adr     x1, fw_dtb_pointer
> >
> > could use a comment as to where this is set up. Previous-stage
> > firmware, I suppose?
>
> Yes.  I'm basically making U-Boot look like a Linux kernel, and m1n1
> passes the DT in x1 just like the Linux kernel expects.
>
> I suspect using CONFIG_OF_PRIOR_STAGE would make it more obvious what
> is happening here.  But (a) I didn't know that existed and (b) we
> discussed removing that option in the near future.

That sounds right to me, but a comment would help.

Regards,
Simon