[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <625b06a0-f7d6-51e4-2930-cfaaf74a9204@linaro.org>
Date: Thu, 18 Aug 2022 11:42:29 +0300
From: Krzysztof Kozlowski <krzysztof.kozlowski@...aro.org>
To: Geert Uytterhoeven <geert@...ux-m68k.org>,
Rob Herring <robh@...nel.org>
Cc: Lad Prabhakar <prabhakar.mahadev-lad.rj@...renesas.com>,
Magnus Damm <magnus.damm@...il.com>,
Krzysztof Kozlowski <krzysztof.kozlowski+dt@...aro.org>,
Linux-Renesas <linux-renesas-soc@...r.kernel.org>,
"open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS"
<devicetree@...r.kernel.org>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
Prabhakar <prabhakar.csengg@...il.com>,
Biju Das <biju.das.jz@...renesas.com>
Subject: Re: [PATCH] dt-bindings: soc: renesas: Move renesas.yaml from arm to
soc
On 17/08/2022 12:30, Geert Uytterhoeven wrote:
> Hi Rob,
>
> On Tue, Aug 16, 2022 at 7:41 PM Rob Herring <robh@...nel.org> wrote:
>> On Mon, Aug 15, 2022 at 12:17:08PM +0100, Lad Prabhakar wrote:
>>> renesas.yaml lists out all the Renesas SoC's and the platforms/EVK's which
>>> is either ARM32/ARM64. It would rather make sense if we move renesas.yaml
>>> to the soc/renesas folder instead. This is in preparation for adding a new
>>> SoC (RZ/Five) from Renesas which is based on RISC-V.
>>
>> Please post this as part of the above.
>>
>> bindings/soc/ is just a dumping ground for stuff that doesn't fit
>> anywhere. We've mostly cleaned bindings/arm/ of that, so I don't really
>
> Note that the target of this move is not .../bindings/soc/, but
bindings/soc/ means bindings/soc/vendor/, so it means bindings/soc/renesas/.
There are no files in bindings/soc/ directly.
> .../bindings/soc/renesas/, so it's a bit less of a dumping ground.
Therefore it is still dumping ground.
> Perhaps this is also a good opportunity to split renesas.yaml per
> family or product group
> (renesas,{rmobile,rcar-gen[1234],rza,rzg,rzn,...}.yaml?
> A fine-grained split may cause headaches with RZ/G2UL and RZ/Five
> sharing the same SoC Base, but a coarse-grained split keeping all RZ/G
> (after all RZ/Five is part of RZ/G) or even all RZ series together should work.
>
>> want to start that again. I would propose bindings/board/ instead if we
>> move in this direction.
>
> .../bindings/board has the issue with the same boards used with
> multiple pin-compatible SoCs, SiPs, and SoMs.
Best regards,
Krzysztof
Powered by blists - more mailing lists