Re: [PATCH v5 1/3] dt-bindings: phy: qcom: Add Glymur QMP PCIe multiple link-mode PHY

From: Krzysztof Kozlowski

Date: Sun Jul 19 2026 - 02:50:25 EST


On Fri, Jul 17, 2026 at 02:58:33AM -0700, Qiang Yu wrote:
> diff --git a/Documentation/devicetree/bindings/phy/qcom,glymur-qmp-pcie-multiphy.yaml b/Documentation/devicetree/bindings/phy/qcom,glymur-qmp-pcie-multiphy.yaml
> new file mode 100644
> index 000000000000..73550f77e9b6
> --- /dev/null
> +++ b/Documentation/devicetree/bindings/phy/qcom,glymur-qmp-pcie-multiphy.yaml

Filename: qcom,glymur-qmp-gen5x8-pcie-phy.yaml

> @@ -0,0 +1,176 @@
> +# SPDX-License-Identifier: (GPL-2.0 OR BSD-2-Clause)
> +%YAML 1.2
> +---
> +$id: http://devicetree.org/schemas/phy/qcom,glymur-qmp-pcie-multiphy.yaml#
> +$schema: http://devicetree.org/meta-schemas/core.yaml#
> +
> +title: Qualcomm QMP PCIe multiple link-mode PHY controller (PCIe, Glymur)
> +
> +maintainers:
> + - Qiang Yu <qiang.yu@xxxxxxxxxxxxxxxx>
> + - Manivannan Sadhasivam <manivannan.sadhasivam@xxxxxxxxxxxxxxxx>
> +
> +description: |
> + The Glymur SoC uses a single PCIe Gen5 PHY hardware block for the
> + PCIe3a/PCIe3b controllers. This block supports two link modes, selected
> + at runtime via a TCSR syscon register:
> +
> + 1. x8 - a single 8-lane PHY instance is exposed (PCIe3a only)
> + 2. x4+x4 - two independent 4-lane PHY instances are exposed (PCIe3a and
> + PCIe3b)
> +
> + The node always describes both PHY instances ("port a" and "port b"),
> + regardless of which link mode is active on the board.
> +
> +properties:
> + compatible:
> + enum:
> + - qcom,glymur-qmp-gen5x8-pcie-phy
> +
> + reg:
> + minItems: 2

Dop

> + maxItems: 2
> +
> + reg-names:
> + items:
> + - const: port_a
> + - const: port_b
> +
> + clocks:
> + minItems: 10

Drop

> + maxItems: 10
> +
> + clock-names:
> + items:
> + - const: aux
> + - const: cfg_ahb
> + - const: ref
> + - const: rchng
> + - const: pipe
> + - const: phy_b_aux
> + - const: cfg_ahb_b
> + - const: rchng_b
> + - const: pipe_b
> + - const: pipediv2_b
> +
> + power-domains:
> + minItems: 2

Drop

> + maxItems: 2
> +
> + power-domain-names:
> + items:
> + - const: phy_a_gdsc

port_a

> + - const: phy_b_gdsc

port_b

> +
> + resets:
> + minItems: 4

Drop

> + maxItems: 4
> +
> + reset-names:
> + items:
> + - const: phy_a

port_a

> + - const: phy_a_nocsr

port_a_nocsr

etc.

> + - const: phy_b
> + - const: phy_b_nocsr
> +
> + vdda-phy-supply: true
> +
> + vdda-pll-supply: true
> +
> + vdda-refgen0p9-supply: true
> +
> + vdda-refgen1p2-supply: true
> +
> + qcom,link-mode:
> + description:
> + Reference to a register in the TCSR syscon that reports the link
> + mode the PCIe PHY is currently configured for, either a single x8
> + link or two independent x4 links. The link mode is programmed by
> + firmware before Linux boots; this property is only used to read
> + the active link mode, specified as a phandle to the syscon and
> + the register offset.
> + $ref: /schemas/types.yaml#/definitions/phandle-array
> + items:
> + - items:
> + - description: phandle of TCSR syscon
> + - description: offset of link mode register
> +
> + "#clock-cells":
> + const: 1
> +
> + clock-output-names:
> + items:
> + - description: Name of the PHY A pipe clock output.
> + - description: Name of the PHY B pipe clock output.
> +
> + "#phy-cells":
> + const: 1

You should explain in description the meaning of cell value. Or provide
filename of header with constants.

> +
> +required:
> + - compatible
> + - reg
> + - reg-names
> + - clocks
> + - clock-names
> + - power-domains
> + - power-domain-names
> + - resets
> + - reset-names
> + - vdda-phy-supply
> + - vdda-pll-supply
> + - vdda-refgen0p9-supply
> + - vdda-refgen1p2-supply
> + - qcom,link-mode
> + - "#clock-cells"
> + - clock-output-names
> + - "#phy-cells"
> +
> +additionalProperties: false
> +
> +examples:
> + - |
> + #include <dt-bindings/clock/qcom,glymur-gcc.h>
> + #include <dt-bindings/clock/qcom,glymur-tcsr.h>
> +
> + pcie3_phy: phy@f00000 {

Drop unused label.

The old 5x4 glymur compatible could be deprecated as well, but fine for
me without doing so.

Best regards,
Krzysztof