[PATCH v3 13/13] arm64: defconfig: Enable Renesas RZ/T2H SoC option

[Krzysztof Kozlowski]

On 27/02/2025 10:24, Geert Uytterhoeven wrote:
> Hi Krzysztof,
> 
> On Wed, 26 Feb 2025 at 15:40, Krzysztof Kozlowski <krzk at kernel.org> wrote:
>> On 26/02/2025 15:32, Thierry Bultel wrote:
>>>> -----Original Message-----
>>>> From: Krzysztof Kozlowski <krzk at kernel.org>
>>>> Sent: mercredi 26 février 2025 15:22
>>>> To: Thierry Bultel <thierry.bultel.yh at bp.renesas.com>; thierry.bultel at linatsea.fr
>>>> Cc: linux-renesas-soc at vger.kernel.org; geert at linux-m68k.org; Paul Barker <paul.barker.ct at bp.renesas.com>; linux-arm->kernel at lists.infradead.org; linux-kernel at vger.kernel.org
>>>> Subject: Re: [PATCH v3 13/13] arm64: defconfig: Enable Renesas RZ/T2H SoC option
>>>>
>>>> On 26/02/2025 14:09, Thierry Bultel wrote:
>>>>> Selects support for RZ/T2H (aka r9a09g077), and SCI (serial) specific
>>>>> code for it.
>>>>>
>>>>> Signed-off-by: Thierry Bultel <thierry.bultel.yh at bp.renesas.com>
>>>>> Reviewed-by: Paul Barker <paul.barker.ct at bp.renesas.com>
>>>>> ---
>>>>
>>>> You never responded to my comments at v1. So I asked at v2. Still no answer.
>>>>
>>>> That's v3 and still silence from your side.
>>>
>>> Yes, I understand your position and have added a paragraph
>>> at the end of the cover letter about this point.
>>
>> We do no read cover letters, unless look for dependencies, so if you
>> disagree with someone you ought to respond to the email directly. Not
>> silently discard.
>>
>> You keep adding more and more symbols, so your "out of scope of this
>> patchset" is no true. Otherwise every contributor will use exactly the
>> same arguments - "not my problem".
>>
>> So again NAK because it is something ought to be finally fixed (and is
>> not even tricky to, so I don't ask for impossible).
> 
> Adding RAM to existing systems is usually quite hard ;-)
> 
> Not all Renesas SoCs are used in systems with multi-GiBs of RAM, so
> IMHO it is still valuable to have fine control over which SoCs are
> supported by your kernel (and e.g. which large pin control tables are
> included in your kernel image).


This is fine, I am not against fine-grained SoC-enable options. However
all your SoCs should be enabled by default (default y if ARCH_RENESAS or
any other option which works for you) thus you won't be growing
user-selectable choices.

Strictly speaking this still will be choice, because you need to trim
config, but all people and all distros will just ignore it  and don't
see it in defconfig.

> 
> Gr{oetje,eeting}s,
> 
>                         Geert
> 


Best regards,
Krzysztof