Re: [PATCH 4/5] ARM: Add basic support for EcoNet EN7523 SoC

From: Ard Biesheuvel
Date: Mon Aug 09 2021 - 08:49:45 EST


On Mon, 9 Aug 2021 at 14:46, Geert Uytterhoeven <geert@xxxxxxxxxxxxxx> wrote:
>
> Hoi Bert,
>
> On Wed, Aug 4, 2021 at 6:43 PM Bert Vermeulen <bert@xxxxxxxx> wrote:
> > On 7/30/21 4:48 PM, Arnd Bergmann wrote:
> > > On Fri, Jul 30, 2021 at 3:45 PM Bert Vermeulen <bert@xxxxxxxx> wrote:
> > >> diff --git a/arch/arm/Makefile b/arch/arm/Makefile
> > >> index 173da685a52e..1bff0aa29c07 100644
> > >> --- a/arch/arm/Makefile
> > >> +++ b/arch/arm/Makefile
> > >> @@ -152,6 +152,7 @@ textofs-$(CONFIG_ARCH_MSM8X60) := 0x00208000
> > >> textofs-$(CONFIG_ARCH_MSM8960) := 0x00208000
> > >> textofs-$(CONFIG_ARCH_MESON) := 0x00208000
> > >> textofs-$(CONFIG_ARCH_AXXIA) := 0x00308000
> > >> +textofs-$(CONFIG_ARCH_ECONET) := 0x00088000
> > >
> > > Why is this needed?
> > >
> > > Note also the comment directly above it exlaining
> > > # Text offset. This list is sorted numerically by address in order to
> > > # provide a means to avoid/resolve conflicts in multi-arch kernels.
> >
> > I didn't make that patch, but it turns out it's needed to get PSCI working;
> > detection hangs without it. That makes no sense to me, but I'll examine further.
>
> Probably PSCI relies on the memory contents at the start of RAM not
> being overwritten?
> Does it help if you remove the first 512 KiB from the /memory node

I /think/ we rely on the first memblock being mappable using section
mappings, so it might be better to remove the first 1 MB (or 2 MB in
case the platform is LPAE capable). Note that this memory is discarded
in any case, so this change is not as costly as it may seem.


> (which should be declared in en7523-evb.dts instead of en7523.dtsi
> BTW)?
>
> Gr{oetje,eeting}s,
>
> Geert
>
> --
> Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@xxxxxxxxxxxxxx
>
> In personal conversations with technical people, I call myself a hacker. But
> when I'm talking to journalists I just say "programmer" or something like that.
> -- Linus Torvalds