Re: [PATCH 4/5] dt-bindings: net: document loongson.pci-gmac
From: Simon Horman <hidden>
Date: 2019-10-31 20:42:10
Also in:
linux-devicetree, linux-ide, linux-mips, linux-pci
On Thu, Oct 31, 2019 at 06:57:16PM +0800, Jiaxun Yang wrote:
在 2019/10/31 下午4:35, Simon Horman 写道:quoted
Hi Jiaxun, thanks for your patch. On Wed, Oct 30, 2019 at 09:53:46PM +0800, Jiaxun Yang wrote:quoted
This binding will provide extra information for PCI enabled device. Signed-off-by: Jiaxun Yang <jiaxun.yang@flygoat.com>Please verify the bindings using dtbs_check as described in Documentation/devicetree/writing-schema.rstquoted
--- .../net/wireless/loongson,pci-gmac.yaml | 71 +++++++++++++++++++ 1 file changed, 71 insertions(+) create mode 100644 Documentation/devicetree/bindings/net/wireless/loongson,pci-gmac.yamldiff --git a/Documentation/devicetree/bindings/net/wireless/loongson,pci-gmac.yaml b/Documentation/devicetree/bindings/net/wireless/loongson,pci-gmac.yaml new file mode 100644 index 000000000000..5f764bd46735 --- /dev/null +++ b/Documentation/devicetree/bindings/net/wireless/loongson,pci-gmac.yaml@@ -0,0 +1,71 @@ +# SPDX-License-Identifier: GPL-2.0 +%YAML 1.2 +--- +$id: http://devicetree.org/schemas/net/allwinner,sun7i-a20-gmac.yaml#The id does not match the filename of the schema. i.e. the above should be: $id: http://devicetree.org/schemas/net/wireless/loongson,pci-gmac.yaml#quoted
+$schema: http://devicetree.org/meta-schemas/core.yaml# + +title: Loongson PCI GMAC Device Tree Bindings + +allOf: + - $ref: "snps,dwmac.yaml#"snps,dwmac.yaml# is in the parent directory relative to loongson,pci-gmac.yaml. So I think the above needs to be: $ref: "../snps,dwmac.yaml#"quoted
+ +maintainers: + - Jiaxun Yang [off-list ref] + +properties: + compatible: + const: loongson,pci-gmac + + reg: + maxItems: 1 + + interrupts: + minItems: 1 + maxItems: 3 + items: + - description: Combined signal for various interrupt events + - description: The interrupt to manage the remote wake-up packet detection + - description: The interrupt that occurs when Rx exits the LPI state + + interrupt-names: + minItems: 1 + maxItems: 3 + items: + - const: macirq + - const: eth_wake_irq + - const: eth_lpi + + clocks: + items: + - description: GMAC main clock + + clock-names: + items: + - const: stmmaceth + +required: + - compatible + - reg + - interrupts + - interrupt-names + - clocks + - clock-names + - phy-mode + +examples: + - | + gmac: ethernet@ {I would have expected a bus address here, f.e.: gmac: ethernet@0x00001800quoted
+ compatible = "loongson,pci-irq"; + reg = <0x00001800 0 0 0 0>;I think there is one to many cell in the above, perhaps it should be. reg = <0x00001800 0 0 0>; Also, I would expect the registers to be wider than 0, i.e. no registers.Hi Simon, Thanks for your suggestions above, will fix in v1. Here, the reg domain is a standard 5-cell representing a PCI device, See: Documentation/devicetree/bindings/pci/pci.txt and IEEE Std 1275-1994,<https://github.com/torvalds/linux/blob/master/Documentation/devicetree/bindings/pci/pci.txt> Should I add some description?
Thanks, sorry for missing that.
As that is the case I think you need something like the following
as an example that compiles.
examples:
- |
pcie@0 {
reg = <0 0 0 0>;
#size-cells = <2>;
#address-cells = <3>;
ranges = <0 0 0 0 0 0>;
device_type = "pci";
gmac: ethernet@1800 {
compatible = "loongson,pci-irq";
reg = <0x00001800 0 0 0 0>;
interrupts = <12>, <13>;
interrupt-names = "macirq", "eth_lpi";
clocks = <&clk_pch_gmac>;
clock-names = "stmmaceth";
phy-mode = "rgmii";
};
};