Re: [PATCH v4 1/1] dt-bindings: remoteproc: mtk,scp: Allow multiple memory regions for MT8188

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] The dt-bindings are updated to allow MT8188 to provide L1TCM via a second `memory-region`, but the `mtk_scp` driver does not support parsing L1TCM from `memory-region`.
--

commit a5d9a931f3a31fccb0fec319f48f7f29be5209e8
Author: Arnab Layek <arnab.layek@mediatek.com>

dt-bindings: remoteproc: mtk,scp: Allow multiple memory regions for MT8188

This commit updates the MT8188 SCP binding to allow an optional second
memory region for Level 1 Tightly Coupled Memory. The base schema is
updated to allow up to two memory regions, keeping other devices
restricted to one region for backward compatibility.

> diff --git a/Documentation/devicetree/bindings/remoteproc/mtk,scp.yaml b/Documentation/devicetree/bindings/remoteproc/mtk,scp.yaml
> index bdbb12118da43..fca9b0675eae4 100644
> --- a/Documentation/devicetree/bindings/remoteproc/mtk,scp.yaml
> +++ b/Documentation/devicetree/bindings/remoteproc/mtk,scp.yaml

[ ... ]

> @@ -205,6 +207,45 @@ allOf:

[ ... ]

> +  - if:
> +      properties:
> +        compatible:
> +          enum:
> +            - mediatek,mt8188-scp
> +            - mediatek,mt8188-scp-dual
> +    then:
> +      properties:
> +        memory-region:
> +          minItems: 1
> +          items:
> +            - description: Main SCP SRAM memory region
> +            - description: Optional SCP L1TCM memory region

[Severity: Medium]
Are there regressions when the L1TCM region is specified as a memory-region
instead of a reg property?

Looking at the mtk_scp driver implementation, it appears to fetch L1TCM
exclusively by looking up an IORESOURCE_MEM resource named l1tcm, which
corresponds to the reg property rather than memory-region:

drivers/remoteproc/mtk_scp.c:scp_probe() {
    ...
    res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "l1tcm");
    if (res) {
        scp->l1tcm_base = devm_ioremap_resource(dev, res);
    ...
}

Since of_reserved_mem_device_init() typically only maps the first memory
region, and the driver checks the reg properties for l1tcm, could this
cause a regression where the second memory-region is silently ignored at
runtime, leaving the L1TCM performance optimizations non-functional?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260520112629.3420612-2-arnab.layek@mediatek.com?part=1