Re: [PATCH 1/7] dt-bindings: interrupt-controller: Add bcm2712 MSI-X DT bindings

[Stanimir Varbanov]

Hi Rob,

Thank you for the comments!

On 6/29/24 01:05, Rob Herring wrote:
> On Wed, Jun 26, 2024 at 01:45:38PM +0300, Stanimir Varbanov wrote:
>> Adds DT bindings for bcm2712 MSI-X interrupt peripheral controller.
>>
>> Signed-off-by: Stanimir Varbanov <svarbanov@xxxxxxx>
>> ---
>>  .../brcm,bcm2712-msix.yaml                    | 74 +++++++++++++++++++
>>  1 file changed, 74 insertions(+)
>>  create mode 100644 Documentation/devicetree/bindings/interrupt-controller/brcm,bcm2712-msix.yaml
>>
>> diff --git a/Documentation/devicetree/bindings/interrupt-controller/brcm,bcm2712-msix.yaml b/Documentation/devicetree/bindings/interrupt-controller/brcm,bcm2712-msix.yaml
>> new file mode 100644
>> index 000000000000..ca610e4467d9
>> --- /dev/null
>> +++ b/Documentation/devicetree/bindings/interrupt-controller/brcm,bcm2712-msix.yaml
>> @@ -0,0 +1,74 @@
>> +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
>> +%YAML 1.2
>> +---
>> +$id: http://devicetree.org/schemas/interrupt-controller/brcm,bcm2712-msix.yaml#
>> +$schema: http://devicetree.org/meta-schemas/core.yaml#
>> +
>> +title: Broadcom bcm2712 MSI-X Interrupt Peripheral support
>> +
>> +maintainers:
>> +  - Stanimir Varbanov <svarbanov@xxxxxxx>
>> +
>> +description: >
>> +  This interrupt controller is used to provide intterupt vectors to the
> 
> typo

Will fix it.

> 
>> +  generic interrupt controller (GIC) on bcm2712. It will be used as
>> +  external MSI-X controller for PCIe root complex.
>> +
>> +allOf:
>> +  - $ref: /schemas/interrupt-controller/msi-controller.yaml#
>> +
>> +properties:
>> +  compatible:
>> +    items:
>> +      - enum:
>> +          - "brcm,bcm2712-mip-intc"
> 
> Don't need quotes. Nor 'items'. And enum can be 'const' 

OK.

> 
>> +  reg:
>> +    maxItems: 1
>> +    description: >
>> +      Specifies the base physical address and size of the registers
> 
> drop. That's *every* reg property.

OK.

> 
>> +
>> +  interrupt-controller: true
>> +
>> +  "#interrupt-cells":
>> +    const: 2
>> +
>> +  msi-controller: true
> 
> Add #msi-cells. The default is 0, but that's legacy.

OK.

> 
>> +
>> +  brcm,msi-base-spi:
>> +    $ref: /schemas/types.yaml#/definitions/uint32
>> +    description: The SGI number that MSIs start.
>> +
>> +  brcm,msi-num-spis:
>> +    $ref: /schemas/types.yaml#/definitions/uint32
>> +    description: The number of SGIs for MSIs.
>> +
>> +  brcm,msi-offset:
>> +    $ref: /schemas/types.yaml#/definitions/uint32
>> +    description: Shift the allocated MSIs up by N.
> 
> If only we had a property that every MSI controller seems to need. Go 
> check msi-controller.yaml...

This exists because the second instance of MIP MSI-X
interrupt-controller (mip1) has some limitations.

Snippet from donwstream dtsi:

	brcm,msi-base-spi = <247>;
	/* Actually 20 total, but the others are
	 * both sparse and non-consecutive
	 */
	brcm,msi-num-spis = <8>;
	brcm,msi-offset = <8>;
	brcm,msi-pci-addr = <0xff 0xffffe000>;

I don't know how to model this except private property.

> 
>> +
>> +  brcm,msi-pci-addr:
>> +    $ref: /schemas/types.yaml#/definitions/uint64
>> +    description: MSI-X message address.
> 
> Why don't other platforms need something like this?

This is a destination address for MSI mem writes from PCIe endpoint
devices, i.e. msi_msg.address filled when composing msi_msg
(irq_chip::irq_compose_msi_msg).

~Stan

> 
>> +
>> +additionalProperties: false
>> +
>> +required:
>> +  - compatible
>> +  - reg
>> +  - interrupt-controller
>> +  - "#interrupt-cells"
>> +  - msi-controller
>> +
>> +examples:
>> +  - |
>> +    msi-controller@130000 {
>> +      compatible = "brcm,bcm2712-mip-intc";
>> +      reg = <0x00130000 0xc0>;
>> +      msi-controller;
>> +      interrupt-controller;
>> +      #interrupt-cells = <2>;
>> +      brcm,msi-base-spi = <128>;
>> +      brcm,msi-num-spis = <64>;
>> +      brcm,msi-offset = <0>;
>> +      brcm,msi-pci-addr = <0xff 0xfffff000>;
>> +    };
>> -- 
>> 2.43.0
>>