linux-kernel - Re: [PATCH 2/5] dt-bindings: soc: canaan: Add top syscon for Canaan K230 SoC

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <572407e8-bac7-4277-bfbd-ed42327b0ff4@kernel.org>
Date: Tue, 30 Dec 2025 15:00:26 +0100
From: Krzysztof Kozlowski <krzk@...nel.org>
To: Jiayu Du <jiayu.riscv@...c.iscas.ac.cn>
Cc: conor@...nel.org, vkoul@...nel.org, gregkh@...uxfoundation.org,
 pjw@...nel.org, palmer@...belt.com, aou@...s.berkeley.edu, alex@...ti.fr,
 neil.armstrong@...aro.org, krzk+dt@...nel.org,
 linux-riscv@...ts.infradead.org, devicetree@...r.kernel.org,
 linux-kernel@...r.kernel.org, linux-phy@...ts.infradead.org,
 linux-usb@...r.kernel.org
Subject: Re: [PATCH 2/5] dt-bindings: soc: canaan: Add top syscon for Canaan
 K230 SoC

On 30/12/2025 14:14, Jiayu Du wrote:
> On Tue, Dec 30, 2025 at 08:39:19AM +0100, Krzysztof Kozlowski wrote:
>> On Tue, Dec 30, 2025 at 10:37:21AM +0800, Jiayu Du wrote:
>>> The Canaan K230 SoC top system controller provides register access
>>> to configure related modules. It includes a USB2 PHY and eMMC/SDIO PHY.
>>>
>>> Signed-off-by: Jiayu Du <jiayu.riscv@...c.iscas.ac.cn>
> ...
>>> +
>>> +  "#size-cells":
>>> +    const: 1
>>> +
>>> +  usb-phy@70:
>>> +    $ref: schemas/phy/canaan,k230-usb-phy.yaml#
>>
>> So that's why you did not have example there? But where did you explain
>> merging strategy/constraints/dependencies? How maintainers can now they
>> can apply this or not?
> 
> Sorry, I will update in v2.
> 
>>
>>
>>> +    unevaluatedProperties: false
>>> +
>>> +  usb-phy@90:
>>> +    $ref: schemas/phy/canaan,k230-usb-phy.yaml#
>>> +    unevaluatedProperties: false
>>
>> Anyway, these are not really real children. Defining child per phy,
>> where each such phy is just few registers, is way too granular. Instead
>> define one phy with phy-cells=2.

Just a note: phy-cells=1, I made mistake before.

>>
>> You also MUST make this device - hisys - binding complete. If you do
>> not, then my review is: fold the children here, because you do not have
>> any other resources for the parent.
> 
> This hisys memory area not only includes the usbphy registers,
> but also contains the registers of sd/mmc phy. Therefore, the
> hisys node is necessary and cannot be folded.

Can be. There is absolutely nothing stopping it.

Anyway, define all nodes.

> 
> 
> If what I said above is accepted by you, do I still need to
> merge the two usb phy nodes by defining one phy with phy-cells=2?

You should read your datasheet, not exactly rely on me guessing. In
current form of the binding, you must fold the child into the parent.

Best regards,
Krzysztof