From mboxrd@z Thu Jan 1 00:00:00 1970 From: Rob Herring Subject: Re: [PATCH 09/38] dt-bindings: gpu: tegra: Convert to json-schema Date: Wed, 17 Jun 2020 20:29:38 -0600 Message-ID: <20200618022938.GA3333151@bogus> References: <20200612141903.2391044-1-thierry.reding@gmail.com> <20200612141903.2391044-10-thierry.reding@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <20200612141903.2391044-10-thierry.reding-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> Sender: devicetree-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: Thierry Reding Cc: devicetree-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, linux-tegra-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-Id: linux-tegra@vger.kernel.org On Fri, Jun 12, 2020 at 04:18:34PM +0200, Thierry Reding wrote: > From: Thierry Reding > > Convert the NVIDIA Tegra GPU bindings from the previous free-form text > format to json-schema. > > Signed-off-by: Thierry Reding > --- > .../devicetree/bindings/gpu/nvidia,gk20a.txt | 90 --------- > .../devicetree/bindings/gpu/nvidia,gk20a.yaml | 189 ++++++++++++++++++ > 2 files changed, 189 insertions(+), 90 deletions(-) > delete mode 100644 Documentation/devicetree/bindings/gpu/nvidia,gk20a.txt > create mode 100644 Documentation/devicetree/bindings/gpu/nvidia,gk20a.yaml > diff --git a/Documentation/devicetree/bindings/gpu/nvidia,gk20a.yaml b/Documentation/devicetree/bindings/gpu/nvidia,gk20a.yaml > new file mode 100644 > index 000000000000..726ef39b2221 > --- /dev/null > +++ b/Documentation/devicetree/bindings/gpu/nvidia,gk20a.yaml > @@ -0,0 +1,189 @@ > +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) > +%YAML 1.2 > +--- > +$id: http://devicetree.org/schemas/gpu/nvidia,gk20a.yaml# > +$schema: http://devicetree.org/meta-schemas/core.yaml# > + > +title: NVIDIA Tegra Graphics Processing Units > + > +maintainers: > + - Thierry Reding > + - Jon Hunter > + > +properties: > + compatible: > + items: > + - enum: > + - nvidia,gk20a > + - nvidia,gm20b > + - nvidia,gp10b > + > + reg: > + items: > + - description: base address and size for BAR0 > + - description: base address and size for BAR1 > + > + interrupts: > + items: > + - description: stalling interrupt > + - description: non-stalling interrupt > + > + interrupt-names: > + items: > + - const: stall > + - const: nonstall > + > + clocks: > + minItems: 2 > + maxItems: 3 > + items: > + - description: GPU clock > + - description: PWR clock > + - description: REF clock > + > + clock-names: > + minItems: 2 > + maxItems: 3 > + items: > + - const: gpu > + - const: pwr > + - const: ref > + > + resets: > + items: > + - description: GPU reset > + > + reset-names: > + minItems: 1 > + maxItems: 1 You can drop minItems/maxItems. The tooling will add them based on 'items' length if not specified. > + items: > + - const: gpu > + > + power-domains: > + description: GPUs that make use of power domains can define this property > + instead of vdd-supply. Currently "nvidia,gp10b" makes use of this. How many? > + > + vdd-supply: > + description: GPU power supply. > + > + iommus: > + description: The phandle of the IOMMU and the IOMMU specifier. See > + ../iommu/iommu.txt for details. > + $ref: "/schemas/types.yaml#/definitions/phandle-array" Again, just need to define how many. > + > +required: > + - compatible > + - reg > + - interrupts > + - interrupt-names > + - clocks > + - clock-names > + - resets > + - reset-names > + > +allOf: > + - if: > + properties: > + compatible: > + contains: > + const: nvidia,gm20b > + then: > + required: > + - vdd-supply > + > + - if: > + properties: > + compatible: > + contains: > + const: nvidia,gm20b > + then: > + properties: > + clock-names: > + minItems: 3 > + > + required: > + - vdd-supply > + > + - if: > + properties: > + compatible: > + contains: > + const: nvidia,gp10b > + then: > + required: > + - power-domains > + > +unevaluatedProperties: false > + > +examples: > + - | > + #include > + #include > + #include > + > + gpu@57000000 { > + compatible = "nvidia,gk20a"; > + reg = <0x0 0x57000000 0x0 0x01000000>, > + <0x0 0x58000000 0x0 0x01000000>; The default size is 1 cell each for examples. So either use that or add a parent node defining 2 cells. > + interrupts = , > + ; > + interrupt-names = "stall", "nonstall"; > + vdd-supply = <&vdd_gpu>; > + clocks = <&tegra_car TEGRA124_CLK_GPU>, > + <&tegra_car TEGRA124_CLK_PLL_P_OUT5>; > + clock-names = "gpu", "pwr"; > + resets = <&tegra_car 184>; > + reset-names = "gpu"; > + iommus = <&mc TEGRA_SWGROUP_GPU>; > + }; > + > + - | > + #include > + #include > + /* > + * XXX can't include this because it conflicts with the tegra124-mc.h > + * include from the Tegra124 example above. > + */ > + /* > + #include > + */ > + > + gpu@57000000 { > + compatible = "nvidia,gm20b"; > + reg = <0x0 0x57000000 0x0 0x01000000>, > + <0x0 0x58000000 0x0 0x01000000>; > + interrupts = , > + ; > + interrupt-names = "stall", "nonstall"; > + vdd-supply = <&vdd_gpu>; > + clocks = <&tegra_car TEGRA210_CLK_GPU>, > + <&tegra_car TEGRA210_CLK_PLL_P_OUT5>, > + <&tegra_car TEGRA210_CLK_PLL_G_REF>; > + clock-names = "gpu", "pwr", "ref"; > + resets = <&tegra_car 184>; > + reset-names = "gpu"; > + iommus = <&mc TEGRA_SWGROUP_GPU>; > + }; > + > + - | > + #include > + #include > + #include > + #include > + #include > + > + gpu@17000000 { > + compatible = "nvidia,gp10b"; > + reg = <0x0 0x17000000 0x0 0x1000000>, > + <0x0 0x18000000 0x0 0x1000000>; > + interrupts = , > + ; > + interrupt-names = "stall", "nonstall"; > + clocks = <&bpmp TEGRA186_CLK_GPCCLK>, > + <&bpmp TEGRA186_CLK_GPU>; > + clock-names = "gpu", "pwr"; > + resets = <&bpmp TEGRA186_RESET_GPU>; > + reset-names = "gpu"; > + power-domains = <&bpmp TEGRA186_POWER_DOMAIN_GPU>; > + iommus = <&smmu TEGRA186_SID_GPU>; > + }; > -- > 2.24.1 >