Re: [PATCH 3/5] dt-bindings: PCI: kirin: Add support for Kirin970

[Rob Herring <robh@kernel.org>]

On Thu, Jul 29, 2021 at 5:56 AM Mauro Carvalho Chehab
<mchehab+huawei@kernel.org> wrote:
>
> Add a new compatible, plus the new bindings needed by
> HiKey970 board.
>
> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
> ---
>  .../bindings/pci/hisilicon,kirin-pcie.yaml    | 61 ++++++++++++++++++-
>  1 file changed, 60 insertions(+), 1 deletion(-)
>
> diff --git a/Documentation/devicetree/bindings/pci/hisilicon,kirin-pcie.yaml b/Documentation/devicetree/bindings/pci/hisilicon,kirin-pcie.yaml
> index 90cab09e8d4b..bb0c3a081d68 100644
> --- a/Documentation/devicetree/bindings/pci/hisilicon,kirin-pcie.yaml
> +++ b/Documentation/devicetree/bindings/pci/hisilicon,kirin-pcie.yaml
> @@ -24,11 +24,13 @@ properties:
>      contains:
>        enum:
>          - hisilicon,kirin960-pcie
> +        - hisilicon,kirin970-pcie
>
>    reg:
>      description: |
>        Should contain dbi, apb, config registers location and length.
> -      For HiKey960, it should also contain phy.
> +      For HiKey960, it should also contain phy. All other devices
> +      should use a separate phy driver.
>      minItems: 3
>      maxItems: 4
>
> @@ -47,6 +49,7 @@ examples:
>    - |
>      #include <dt-bindings/interrupt-controller/arm-gic.h>
>      #include <dt-bindings/clock/hi3660-clock.h>
> +    #include <dt-bindings/clock/hi3670-clock.h>
>
>      soc {
>        #address-cells = <2>;
> @@ -83,4 +86,60 @@ examples:
>          clock-names = "pcie_phy_ref", "pcie_aux", "pcie_apb_phy",
>                        "pcie_apb_sys", "pcie_aclk";
>        };
> +
> +      pcie@f5000000 {
> +        compatible = "hisilicon,kirin970-pcie";
> +        reg = <0x0 0xf4000000 0x0 0x1000000>,
> +              <0x0 0xfc180000 0x0 0x1000>,
> +              <0x0 0xf5000000 0x0 0x2000>;
> +        reg-names = "dbi", "apb", "config";
> +        bus-range = <0x0  0x1>;
> +        msi-parent = <&its_pcie>;
> +        #address-cells = <3>;
> +        #size-cells = <2>;
> +        device_type = "pci";
> +        phys = <&pcie_phy>;
> +        ranges = <0x02000000 0x0 0x00000000
> +                  0x0 0xf6000000
> +                  0x0 0x02000000>;
> +        num-lanes = <1>;
> +        #interrupt-cells = <1>;
> +        interrupts = <GIC_SPI 283 IRQ_TYPE_LEVEL_HIGH>;
> +        interrupt-names = "msi";
> +        interrupt-map-mask = <0 0 0 7>;
> +        interrupt-map = <0x0 0 0 1 &gic GIC_SPI 282 IRQ_TYPE_LEVEL_HIGH>,
> +                        <0x0 0 0 2 &gic GIC_SPI 283 IRQ_TYPE_LEVEL_HIGH>,
> +                        <0x0 0 0 3 &gic GIC_SPI 284 IRQ_TYPE_LEVEL_HIGH>,
> +                        <0x0 0 0 4 &gic GIC_SPI 285 IRQ_TYPE_LEVEL_HIGH>;
> +        pcie@4,0 { // Lane 4: M.2
> +          reg = <0 0 0 0 0>;
> +          compatible = "pciclass,0604";
> +          device_type = "pci";
> +          reset-gpios = <&gpio7 1 0>;
> +          clkreq-gpios = <&gpio27 3 0 >;

Looking at the schematics some more, this is not right. CLKREQ# is an
input from the device, and they are not connected to any GPIO (just
pulled high) on hikey970. These GPIOs are simply clock enables and
very much specific to hikey. So I'd call this 'hisilicon,clken-gpios'
and you can just stick them in the host bridge node.

I think the way the board should have been designed is the CLKREQ#
signals to the clock driver chip OE signals. Then there'd be no s/w
control needed.

Rob