linux-kernel - Re: [PATCH v2] dt-bindings: mailbox: add Versal IPI bindings

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: <1027ec80-bed6-4b51-b346-60bc19aa04e6@amd.com>
Date:   Wed, 13 Dec 2023 10:35:57 -0600
From:   Tanmay Shah <tanmay.shah@....com>
To:     Krzysztof Kozlowski <krzysztof.kozlowski@...aro.org>,
        jassisinghbrar@...il.com, robh+dt@...nel.org,
        krzysztof.kozlowski+dt@...aro.org, conor+dt@...nel.org,
        michal.simek@....com, shubhrajyoti.datta@....com
Cc:     linux-kernel@...r.kernel.org, devicetree@...r.kernel.org,
        linux-arm-kernel@...ts.infradead.org,
        "Levinsky, Ben" <ben.levinsky@....com>
Subject: Re: [PATCH v2] dt-bindings: mailbox: add Versal IPI bindings


On 12/13/23 10:21 AM, Krzysztof Kozlowski wrote:
> On 13/12/2023 17:18, Tanmay Shah wrote:
>
>
> >>> +    minItems: 1
> >>> +    maxItems: 2
> >>
> >> I don't understand why this change is here. Previously you did not have
> >> MMIO address space? If yes, then where do you restrict the old device to
> >> disallow these?
> > 
> > 
> > Hardware description is different between ZynqMP and Versal. And so, we have to design
> > 
> > new bindings for Versal. reg and reg-names for parent node, is only available for new devices.
> > 
> > The new device is checked with compatible string after "required" section.
>
> Hm? What does that mean?


Following [1]. I think You answered what needs to be done. I will have reg: false for old devices.

Thanks.


>
> > 
> > Should I add "unevaluatedProperties: false" after "additionalProperties: false" for parent node below [1] ?
>
> No, you should disallow reg/reg-names for older device.
>
> ...
>
> > 
> >>> +        then:
> >>> +          properties:
> >>> +            reg:
> >>> +              items:
> >>> +                - description: Host agent request message buffer
> >>> +                - description: Host agent response message buffer
> >>> +                - description: Remote agent request message buffer
> >>> +                - description: Remote agent response message buffer
> >>> +
> >>> +            reg-names:
> >>> +              items:
> >>> +                - const: local_request_region
> >>> +                - const: local_response_region
> >>> +                - const: remote_request_region
> >>> +                - const: remote_response_region
> >>> +        else:
> >>> +          properties:
> >>> +            reg:
> >>> +              minItems: 1
> >>> +              items:
> >>> +                - description: Remote IPI agent control register
> >>> +                - description: Remote IPI agent optional message buffer
> >>
> >> Were these described in old binding? If not, it's a separate change.
> > 
> > Okay, so I will split this in two patches:
> > 
> > 1) Clean up current bindings (like remove redundant descriptino, sort "required" property order etc..)
>
> If you want to combine cleanup and functional changes, then this would
> be two patches.
>
> > 
> > 2) Add versal platforms bindings doc. (This will add if else cases and Versal platform support)

In that case, can I just post single patch that adds new device support ?

Won't touch anything that is not needed for new device support.


>
> > 
> > 
> >>> +
> >>> +            reg-names:
> >>> +              minItems: 1
> >>> +              items:
> >>> +                - const: ctrl
> >>> +                - const: msg
> >>
> >> Blank line
> >>
> >>>  required:
> >>>    - compatible
> >>> -  - interrupts
> >>>    - '#address-cells'
> >>>    - '#size-cells'
> >>> +  - interrupts
> >>
> >> Separate change with its own rationale. Trivial cleanups can be
> >> organized in one patch, but should not be mixed with adding new devices.
> > 
> > Ack
> > 
> > 
> >>>    - xlnx,ipi-id
> >>>  
> >>> +allOf:
> >>> +  - if:
> >>> +      properties:
> >>> +        compatible:
> >>> +          contains:
> >>> +            enum:
> >>> +              - xlnx,versal-ipi-mailbox
> >>> +    then:
> >>> +      properties:
> >>> +        reg:
> >>> +          items:
> >>> +            - description: Host IPI agent control registers
> >>> +            - description: Host IPI agent optional message buffers
> >>> +
> >>> +        reg-names:
> >>> +          items:
> >>> +            - const: ctrl
> >>> +            - const: msg
> >>> +
> >>> +      required:
> >>> +        - reg
> >>> +        - reg-names
> >>> +
> >>> +additionalProperties: false
> > 
> > [1] Here, if I add unevaluatedProperties: false then is it enough for old device to disallow
>
> No, you need reg:false
>
> Best regards,
> Krzysztof
>