[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20251029-adventurous-russet-jackal-64e3e5@kuoka>
Date: Wed, 29 Oct 2025 07:50:02 +0100
From: Krzysztof Kozlowski <krzk@...nel.org>
To: niravkumarlaxmidas.rabara@...era.com
Cc: dinguyen@...nel.org, matthew.gerlach@...era.com, robh@...nel.org,
krzk+dt@...nel.org, conor+dt@...nel.org, bp@...en8.de, tony.luck@...el.com,
linux-edac@...r.kernel.org, devicetree@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 1/6] dt-bindings: edac: altera: Document additional ECC
instances
On Tue, Oct 28, 2025 at 05:22:27PM +0800, niravkumarlaxmidas.rabara@...era.com wrote:
> From: Niravkumar L Rabara <niravkumarlaxmidas.rabara@...era.com>
>
> Add support for Secure Device Manager(SDM) QSPI ECC, IO96B memory
> controller ECC and Configuration RAM(CRAM) Single Event Upset(SEU).
>
> Add interrupt-names property and increase interrupts maxItems from 2 to 7
> to accommodate additional interrupts.
>
> Signed-off-by: Niravkumar L Rabara <niravkumarlaxmidas.rabara@...era.com>
> ---
> .../edac/altr,socfpga-ecc-manager.yaml | 77 ++++++++++++++++++-
> 1 file changed, 76 insertions(+), 1 deletion(-)
>
> diff --git a/Documentation/devicetree/bindings/edac/altr,socfpga-ecc-manager.yaml b/Documentation/devicetree/bindings/edac/altr,socfpga-ecc-manager.yaml
> index 3d787dea0f14..5e0c08a15ab9 100644
> --- a/Documentation/devicetree/bindings/edac/altr,socfpga-ecc-manager.yaml
> +++ b/Documentation/devicetree/bindings/edac/altr,socfpga-ecc-manager.yaml
> @@ -33,7 +33,13 @@ properties:
>
> interrupts:
> minItems: 1
> - maxItems: 2
> + maxItems: 7
No, list the interrupts instead. Your commit msg must clearly explain
why exception of not-fixed length/entries is justified.
See writing bindings.
> +
> + interrupt-names:
> + items:
> + enum: [global_sbe, global_dbe, io96b0, io96b1, sdm_qspi_sbe, sdm_qspi_dbe, sdm_seu]
Nope, list the items instead. Please do not come up with some custom
syntax.
> + minItems: 1
> + maxItems: 7
>
> interrupt-controller: true
>
> @@ -70,6 +76,41 @@ properties:
> - interrupts
> - altr,sdr-syscon
>
> + cram-seu:
Missing description, so difficult to say what is here.
Node names should be generic. See also an explanation and list of
examples (not exhaustive) in DT specification:
https://devicetree-specification.readthedocs.io/en/latest/chapter2-devicetree-basics.html#generic-names-recommendation
If you cannot find a name matching your device, please check in kernel
sources for similar cases or you can grow the spec (via pull request to
DT spec repo).
> + type: object
> + additionalProperties: false
> +
> + properties:
> + compatible:
> + items:
> + - const: altr,socfpga-cram-seu
Why do you need compatible?
> +
> + reg:
> + maxItems: 1
So you created child node only for reg? No, fold it into parent.
You also forgot to update the example.
> +
> + altr,seu-safe-inject-ce-msb:
> + $ref: /schemas/types.yaml#/definitions/uint32
> + description: MSB of error injection command for Correctable Error
> +
> + altr,seu-safe-inject-ce-lsb:
> + $ref: /schemas/types.yaml#/definitions/uint32
> + description: LSB of error injection command for Correctable Error
> +
> + altr,seu-safe-inject-ue-msb:
> + $ref: /schemas/types.yaml#/definitions/uint32
> + description: MSB of error injection command for Uncorrectable Error
> +
> + altr,seu-safe-inject-ue-lsb:
> + $ref: /schemas/types.yaml#/definitions/uint32
> + description: LSB of error injection command for Uncorrectable Error
How are these board-level properties?
> +
> + required:
> + - compatible
> + - altr,seu-safe-inject-ce-msb
> + - altr,seu-safe-inject-ce-lsb
> + - altr,seu-safe-inject-ue-msb
> + - altr,seu-safe-inject-ue-lsb
> +
> patternProperties:
> "^ocram-ecc@[a-f0-9]+$":
> type: object
> @@ -191,6 +232,40 @@ patternProperties:
> - interrupts
> - altr,ecc-parent
>
> + "^sdm-qspi-ecc@[a-f0-9]+$":
> + type: object
> + additionalProperties: false
> +
> + properties:
> + compatible:
> + items:
> + - const: altr,socfpga-sdm-qspi-ecc
No, drop.
> +
> + reg:
> + maxItems: 1
> +
> + required:
> + - compatible
> + - reg
No point for empty children. One reg is not justification for having a
child.
> +
> + "^io96b[0-9]-ecc@[a-f0-9]+$":
You need to stop coming with random node names. Nothing explains why you
need children, why these are not part of parent node.
> + type: object
> + additionalProperties: false
> +
> + properties:
> + compatible:
> + items:
> + - enum:
> + - altr,socfpga-io96b0-ecc
> + - altr,socfpga-io96b1-ecc
Plus all your compatibles have WRONG format. See writing bindings and
numerouse presentations - you always must use SoC specific compatible.
Best regards,
Krzysztof
Powered by blists - more mailing lists