Re: [PATCH 4/4] arm64: renesas: Add R8A78000 Ironhide board code

[Krzysztof Kozlowski <krzk@kernel.org>]

On 10/09/2025 14:45, Geert Uytterhoeven wrote:
> On Wed, 10 Sept 2025 at 09:11, Krzysztof Kozlowski <krzk@kernel.org> wrote:
>> On 10/09/2025 01:39, Kuninori Morimoto wrote:
>>>>> +   chosen {
>>>>> +           bootargs = "ignore_loglevel rw maxcpus=1";
>>>>
>>>> Drop bootargs. Not suitable for mainline. I really do not understand why
>>>> all users in mainline and other projects (not your development) must be
>>>> limited to cpus=1.
>>>
>>> This is initial patch, and only 1 CPU is available for now.
>>> The limitaion will be removed in the future, but not yet for now.
>>
>> You do not understand the problem - DTS describes the hardware, not your
>> current stage of Linux support. These bootargs are wrong, I have been
>> commenting on this since years.
> 
> Well, we can't post the full DTS describing the full hardware yet, as
> that hasn't been written yet.  E.g. lots of DT bindings for new or

No one suggests that, but you tied DTS to your particular setup without
any explanation. Nothing in commit msg, nothing in comment explains why
you need "rw". Nothing explains why maxcpus=1.

I am sorry, but we talk about complete basics here. All odd and weird
code - and maxcpus=1 is the weirdest - MUST HAVE explanation.

> changed components are still missing.  So we have to fall back to the
> customary incremental DTS writing...
> 
> If currently only one CPU can/must be used, the right way to handle
> that is to drop all but the first CPU node from the .dtsi, and only
> add the other CPU nodes later.  This is what we've been doing for
> other SoCs before, too.

No, because on my imaginary system, these bootargs are not helping and
remaining CPUs work fine.

What's more, it breaks my imaginary system root, because I am running it
from readonly NFS root.


Best regards,
Krzysztof