[PATCH 3/7] dt-bindings: pci: Add document for ASPEED PCIe RC
Jacky Chou
jacky_chou at aspeedtech.com
Fri Jun 20 01:36:02 PDT 2025
> > +++ b/Documentation/devicetree/bindings/pci/aspeed-pcie.yaml
>
> Same comments.
>
> > @@ -0,0 +1,159 @@
> > +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) %YAML 1.2
> > +---
> > +$id: http://devicetree.org/schemas/pci/aspeed-pcie.yaml#
> > +$schema: http://devicetree.org/meta-schemas/core.yaml#
> > +
> > +title: ASPEED PCIe Root Complex Controller
> > +
> > +maintainers:
> > + - Jacky Chou <jacky_chou at aspeedtech.com>
> > +
> > +description: |
>
> Do not need '|' unless you need to preserve formatting.
>
Agreed.
>
> > + Device tree binding for the ASPEED PCIe Root Complex controller.
>
> No, describe the hardware. Your current description is 100% redundant.
> It is never useful to say in DT binding description that it is a DT binding. It
> cannot be anything else, can it?
>
Agreed.
> > +
> > +properties:
> > + compatible:
> > + enum:
> > + - aspeed,ast2600-pcie
> > + - aspeed,ast2700-pcie
> > +
> > + device_type:
> > + const: pci
>
> You need to include proper pci schema and drop all redundant properties.
>
> Look at other schemas.
Agreed.
>
> > +
> > + reg:
> > + maxItems: 1
> > +
> > + ranges:
> > + minItems: 2
> > + maxItems: 2
> > +
> > + interrupts:
> > + description: IntX and MSI interrupt
>
> Need to list the items. Look at other schemas.
Agreed.
>
> > +
> > + resets:
> > + items:
> > + - description: Module reset
> > + - description: PCIe PERST
> > +
> > + reset-names:
> > + items:
> > + - const: h2x
> > + - const: perst
> > +
> > + msi-parent: true
> > +
> > + msi_address:
>
> Where is this property defined? I do not see in kernel nor in dtschema.
> Drop and use existing properties. I am not even talking about coding style...
Agreed.
>
> > + $ref: /schemas/types.yaml#/definitions/uint32
> > + description: MSI address
> > +
> > + aspeed,ahbc:
> > + $ref: /schemas/types.yaml#/definitions/phandle
> > + description: Phandle to ASPEED AHBC syscon.
>
> For what purpose?
>
> > +
> > + aspeed,pciecfg:
> > + $ref: /schemas/types.yaml#/definitions/phandle
> > + description: Phandle to ASPEED PCIe configuration syscon.
>
> For what purpose?
>
> > +
> > + aspeed,pciephy:
> > + $ref: /schemas/types.yaml#/definitions/phandle
> > + description: Phandle to ASPEED PCIe PHY syscon.
>
> For what purpose?
>
I will add more description in next version.
> > +
> > + clocks:
> > + description: PCIe BUS clock
>
> Missing constraints.
>
> Just open any other binding and do not implement things diferently.
>
Agreed.
> > +
> > + interrupt-controller:
> > + description: Interrupt controller node for handling legacy PCI
> interrupts.
> > + type: object
> > + properties:
> > + '#address-cells':
> > + const: 0
> > + '#interrupt-cells':
> > + const: 1
> > + interrupt-controller: true
> > +
> > + required:
> > + - '#address-cells'
> > + - '#interrupt-cells'
> > + - interrupt-controller
> > +
> > + additionalProperties: false
> > +
> > +allOf:
> > + - $ref: /schemas/pci/pci-bus.yaml#
> > + - $ref: /schemas/interrupt-controller/msi-controller.yaml#
> > + - if:
> > + properties:
> > + compatible:
> > + contains:
> > + const: aspeed,ast2600-pcie
> > + then:
> > + required:
> > + - aspeed,ahbc
>
> else: make it false
>
Agreed.
> > +
> > +required:
> > + - interrupts
> > + - bus-range
> > + - ranges
> > + - resets
> > + - reset-names
> > + - msi-parent
> > + - msi-controller
> > + - aspeed,pciephy
> > + - aspeed,pciecfg
> > + - interrupt-map-mask
> > + - interrupt-map
> > + - interrupt-controller
>
> Messed order, missing properties. Open other bindings...
>
Agreed.
> > +
> > +unevaluatedProperties: false
> > +
> > +examples:
> > + - |
> > + #include <dt-bindings/interrupt-controller/arm-gic.h>
> > + #include <dt-bindings/clock/ast2600-clock.h>
> > +
> > + apb {
> > + #address-cells = <1>;
> > + #size-cells = <1>;
> > +
> > + pcie0: pcie at 1e7700C0 {
> > + compatible = "aspeed,ast2600-pcie";
> > + device_type = "pci";
> > + reg = <0x1e7700C0 0x40>;
>
> Lower case hex. Please follow carefully DTS coding style.
>
Agreed.
Thanks,
Jacky
More information about the linux-arm-kernel
mailing list