[PATCH 2/2] dt-bindings: memory-controllers: Add mediatek common-dramc dt-bindings

Krzysztof Kozlowski krzk at kernel.org
Thu Dec 12 02:58:00 PST 2024


On 12/12/2024 09:59, Crystal Guo wrote:
> Add devicetree binding for mediatek common-dramc driver.
> 
> The DRAM controller of MediaTek SoC provides an interface to
> get the current data rate of DRAM.

Bindings are before users.



A nit, subject: drop second/last, redundant "dt-bindings". The
"dt-bindings" prefix is already stating that these are bindings.
See also:
https://elixir.bootlin.com/linux/v6.7-rc8/source/Documentation/devicetree/bindings/submitting-patches.rst#L18

> 
> Signed-off-by: Crystal Guo <crystal.guo at mediatek.com>
> ---
>  .../mediatek,common-dramc.yaml                | 129 ++++++++++++++++++
>  1 file changed, 129 insertions(+)
>  create mode 100644 Documentation/devicetree/bindings/memory-controllers/mediatek,common-dramc.yaml
> 
> diff --git a/Documentation/devicetree/bindings/memory-controllers/mediatek,common-dramc.yaml b/Documentation/devicetree/bindings/memory-controllers/mediatek,common-dramc.yaml
> new file mode 100644
> index 000000000000..c9e608c7f183
> --- /dev/null
> +++ b/Documentation/devicetree/bindings/memory-controllers/mediatek,common-dramc.yaml
> @@ -0,0 +1,129 @@
> +# SPDX-License-Identifier: GPL-2.0-only OR BSD-2-Clause
> +# Copyright (c) 2024 MediaTek Inc.
> +%YAML 1.2
> +---
> +$id: http://devicetree.org/schemas/memory-controllers/mediatek,common-dramc.yaml#
> +$schema: http://devicetree.org/meta-schemas/core.yaml#
> +
> +title: MediaTek Common DRAMC (DRAM Controller)

Common? Is this a real thing? Please describe the hardware.

> +
> +maintainers:
> +  - Crystal Guo <crystal.guo at mediatek.com>
> +
> +description: |

Do not need '|' unless you need to preserve formatting.

> +  The DRAM controller of MediaTek SoC provides an interface to
> +  get the current data rate of DRAM.

So not common here?

> +
> +properties:
> +  compatible:
> +    const: mediatek,common-dramc

This has to be SoC.

> +
> +  reg:
> +    minItems: 9

Why this is flexible?

> +    items:
> +      - description: DRAMC_AO_CHA_BASE
> +      - description: DRAMC_AO_CHB_BASE
> +      - description: DRAMC_AO_CHC_BASE
> +      - description: DRAMC_AO_CHD_BASE
> +      - description: DRAMC_NAO_CHA_BASE
> +      - description: DRAMC_NAO_CHB_BASE
> +      - description: DRAMC_NAO_CHC_BASE
> +      - description: DRAMC_NAO_CHD_BASE
> +      - description: DDRPHY_AO_CHA_BASE
> +      - description: DDRPHY_AO_CHB_BASE
> +      - description: DDRPHY_AO_CHC_BASE
> +      - description: DDRPHY_AO_CHD_BASE
> +      - description: DDRPHY_NAO_CHA_BASE
> +      - description: DDRPHY_NAO_CHB_BASE
> +      - description: DDRPHY_NAO_CHC_BASE
> +      - description: DDRPHY_NAO_CHD_BASE
> +      - description: SLEEP_BASE

Don't use some defines. Look at other bindings how they describe items.

> +
> +  support-ch-cnt:

Nope

> +    maxItems: 1
> +
> +  fmeter-version:
> +    maxItems: 1
> +    description:
> +      Fmeter version for calculating dram data rate
> +
> +  crystal-freq:
> +    maxItems: 1
> +    description:
> +      Reference clock rate in MHz
> +
> +  shu-of:
> +    maxItems: 1
> +
> +  pll-id: true
> +  shu-lv: true
> +  sdmpcw: true
> +  posdiv: true
> +  fbksel: true
> +  dqsopen: true
> +  async-ca: true
> +  dq-ser-mode: true


This binding is terrible. Was not tested and does not follow any
guidelines. Please read example schema and writing bindings document.
You can also read slides from my talks...


> +
> +required:
> +  - compatible
> +  - reg
> +  - support-ch-cnt
> +  - fmeter-version
> +  - crystal-freq
> +  - pll-id
> +  - shu-lv
> +  - shu-of
> +  - sdmpcw
> +  - posdiv
> +  - fbksel
> +  - dqsopen
> +  - async-ca
> +  - dq-ser-mode
> +
> +additionalProperties: false
> +
> +examples:
> +  - |
> +    soc {
> +        #address-cells = <2>;
> +        #size-cells = <2>;
> +
> +        dramc: dramc at 10230000 {

memory-controller@
and drop unused label.

Best regards,
Krzysztof



More information about the Linux-mediatek mailing list