linux-kernel - Re: [PATCH 1/2] dt-bindings: misc: xlnx,sd-fec: convert bindings to yaml

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

Date: Tue, 16 Jan 2024 16:44:30 +0100
From: Krzysztof Kozlowski <krzysztof.kozlowski@...aro.org>
To: Michal Simek <michal.simek@....com>, Dragan Cvetic
 <dragan.cvetic@....com>, arnd@...db.de, gregkh@...uxfoundation.org,
 michal.simek@...inx.com, linux-arm-kernel@...ts.infradead.org,
 robh+dt@...nel.org, mark.rutland@....com, devicetree@...r.kernel.org
Cc: linux-kernel@...r.kernel.org, git@....com
Subject: Re: [PATCH 1/2] dt-bindings: misc: xlnx,sd-fec: convert bindings to
 yaml

On 16/01/2024 16:36, Michal Simek wrote:
>>> +  clock-names:
>>> +    additionalItems: true
>>
>> Nope
>>
>>> +    minItems: 2
>>> +    maxItems: 8
>>
>> Nope
>>
>>> +    items:
>>> +      - const: core_clk
>>> +      - const: s_axi_aclk
>>> +      - enum:
>>> +          - s_axis_ctrl_aclk
>>> +          - s_axis_din_aclk
>>> +          - m_axis_status_aclk
>>> +          - m_axis_dout_aclk
>>> +          - s_axis_din_words_aclk
>>> +          - m_axis_dout_words_aclk
>>
>> Why order is not enforced?
> 
> Let me comment this one. Based on my discussion with Dragan IP itself is 
> configurable and only the first two clocks are in all combinations. But based on 
> his description that last 6 clocks can be present in some of them.
> It means order is not really fixed and any combination is possible.
> That's why I have suggested him to use this description because I didn't find 
> any better one.
> I actually tested this schema here but didn't get a feedback on it yet.
> https://lore.kernel.org/r/3e86244a840a45c970289ba6d2fa700a74f5b259.1705051222.git.michal.simek@amd.com
> 
> It means not sure about not defining maxItems but when I don't do it it is not 
> passing dtbs_check.


This would explain why you want additionalItems:true, but it should be
also explained in commit msg. Old code did not have such relaxed
statement, at least not explicitly written, and commit msg explicitly
says it is 1-to-1 conversion.

Anyway, current solution won't work, because additional items can be
anything. Try it. Put as fourth clock "yellow_duck" and see what happens.

I don't find such names as useful and maybe the drivers should just get
by index. Especially that Linux driver does not care. It would be a ABI
change, though, so up to you.

If you want to keep the names, then:
1. Look at snps,dwmac.yaml
2. or just list 6 enums with all possibilities.


Best regards,
Krzysztof