RE: [PATCH 14/23] arm64: dts: fsd: Add initial device tree support

[Alim Akhtar]

Hi Arnd and Krzysztof

>-----Original Message-----
>From: Arnd Bergmann [mailto:arnd@xxxxxxxx]
>Sent: Thursday, January 13, 2022 7:54 PM
>To: Krzysztof Kozlowski <krzysztof.kozlowski@xxxxxxxxxxxxx>
>Cc: Alim Akhtar <alim.akhtar@xxxxxxxxxxx>; Linux ARM <linux-arm-
>kernel@xxxxxxxxxxxxxxxxxxx>; Linux Kernel Mailing List <linux-
>kernel@xxxxxxxxxxxxxxx>; SoC Team <soc@xxxxxxxxxx>; linux-clk <linux-
>clk@xxxxxxxxxxxxxxx>; DTML <devicetree@xxxxxxxxxxxxxxx>; Olof Johansson
><olof@xxxxxxxxx>; Linus Walleij <linus.walleij@xxxxxxxxxx>; Catalin Marinas
><catalin.marinas@xxxxxxx>; Rob Herring <robh+dt@xxxxxxxxxx>; Sylwester
>Nawrocki <s.nawrocki@xxxxxxxxxxx>; moderated list:ARM/SAMSUNG
>EXYNOS ARM ARCHITECTURES <linux-samsung-soc@xxxxxxxxxxxxxxx>; Pankaj
>Dubey <pankaj.dubey@xxxxxxxxxxx>; linux-fsd@xxxxxxxxx; Arjun K V
><arjun.kv@xxxxxxxxxxx>; Aswani Reddy <aswani.reddy@xxxxxxxxxxx>;
>Ajay Kumar <ajaykumar.rs@xxxxxxxxxxx>; Sriranjani P
><sriranjani.p@xxxxxxxxxxx>; Chandrasekar R <rcsekar@xxxxxxxxxxx>;
>Shashank Prashar <s.prashar@xxxxxxxxxxx>; Arnd Bergmann
><arnd@xxxxxxxx>
>Subject: Re: [PATCH 14/23] arm64: dts: fsd: Add initial device tree support
>
>On Thu, Jan 13, 2022 at 2:16 PM Krzysztof Kozlowski
><krzysztof.kozlowski@xxxxxxxxxxxxx> wrote:
>> >  ARM/TETON BGA MACHINE SUPPORT
>> >  M:   "Mark F. Brown" <mark.brown314@xxxxxxxxx>
>> >  L:   linux-arm-kernel@xxxxxxxxxxxxxxxxxxx (moderated for non-subscribers)
>> > diff --git a/arch/arm64/Kconfig.platforms
>> > b/arch/arm64/Kconfig.platforms index 54e3910e8b9b..bb8a047c2359
>> > 100644
>> > --- a/arch/arm64/Kconfig.platforms
>> > +++ b/arch/arm64/Kconfig.platforms
>> > @@ -267,6 +267,12 @@ config ARCH_TEGRA
>> >       help
>> >         This enables support for the NVIDIA Tegra SoC family.
>> >
>> > +config ARCH_TESLA_FSD
>> > +     bool "ARMv8 based Tesla platform"
>> > +     select ARCH_EXYNOS
>>
>> How similar it is? I think it is better to duplicate Exynos
>> selections/options here, instead of selecting entire ARCH. If this
>> would require "depends on ARCH_EXYNOS || ARCH_TESLA_FSD"
>everywhere in
>> the drivers, it's a hint that it is not a separate SoC but it is an
>> Exynos, so it might not need a new sub-architecture.
>
>Agreed, the SoC family options mainly exist so we can quickly enable or
>disable drivers based on what a kernel is built for. If most of the drivers for
>this SoC are shared with Exynos, I think having a single option is sufficient, but
>it may be worth pointing out both in the help text.
>
>If we want to have a separate option (mainly to help users find it), maybe a
>'depends on ARCH_EXYNOS' would be better. How many uses of
>ARCH_TESLA_FSD are there?
>
It is true that FSD shares few IPs with Exynos and it dose contains Telsa specific IPs/ PCIe/some of the PHYs/ GPUs (different then Exynos) etc. to name a few.
And drivers for those will be posted in upcoming phases by Samsung, Telsa etc.
AFA architecture is concerns both Exynos and FSD has completely different architecture (at least at HW level).
In my opinion, it is make sense to have a separate option for FSD.
And as Arnd suggested "'depends on ARCH_EXYNOS" may be the way forward.

>        Arnd