Re: [PATCH v2 2/3] arm64: dts: qcom: sc8280xp: Add GPU related nodes

From: Akhil P Oommen
Date: Thu Jun 01 2023 - 14:46:31 EST

Next message: Lee Jones: "Re: [PATCH v5 2/2] leds: add aw20xx driver"
Previous message: Boqun Feng: "Re: [PATCH v2 5/8] rust: workqueue: add helper for defining work_struct fields"
Next in thread: Akhil P Oommen: "Re: [PATCH v2 2/3] arm64: dts: qcom: sc8280xp: Add GPU related nodes"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On Mon, May 29, 2023 at 09:38:59AM +0200, Konrad Dybcio wrote:
>
>
>
> On 28.05.2023 19:07, Manivannan Sadhasivam wrote:
> > On Tue, May 23, 2023 at 09:59:53AM +0200, Konrad Dybcio wrote:
> >>
> >>
> >> On 23.05.2023 03:15, Bjorn Andersson wrote:
> >>> From: Bjorn Andersson <bjorn.andersson@xxxxxxxxxx>
> >>>
> >>> Add Adreno SMMU, GPU clock controller, GMU and GPU nodes for the
> >>> SC8280XP.
> >>>
> >>> Signed-off-by: Bjorn Andersson <bjorn.andersson@xxxxxxxxxx>
> >>> Signed-off-by: Bjorn Andersson <quic_bjorande@xxxxxxxxxxx>
> >>> ---
> >> It does not look like you tested the DTS against bindings. Please run
> >> `make dtbs_check` (see
> >> Documentation/devicetree/bindings/writing-schema.rst for instructions).
> >>
> >>>
> >>> Changes since v1:
> >>> - Dropped gmu_pdc_seq region from &gmu, as it shouldn't have been used.
> >>> - Added missing compatible to &adreno_smmu.
> >>> - Dropped aoss_qmp clock in &gmu and &adreno_smmu.
> >>>
> >>> arch/arm64/boot/dts/qcom/sc8280xp.dtsi | 169 +++++++++++++++++++++++++
> >>> 1 file changed, 169 insertions(+)
> >>>
> >>> diff --git a/arch/arm64/boot/dts/qcom/sc8280xp.dtsi b/arch/arm64/boot/dts/qcom/sc8280xp.dtsi
> >>> index d2a2224d138a..329ec2119ecf 100644
> >>> --- a/arch/arm64/boot/dts/qcom/sc8280xp.dtsi
> >>> +++ b/arch/arm64/boot/dts/qcom/sc8280xp.dtsi
> >>> @@ -6,6 +6,7 @@
> >>>
> >>> #include <dt-bindings/clock/qcom,dispcc-sc8280xp.h>
> >>> #include <dt-bindings/clock/qcom,gcc-sc8280xp.h>
> >>> +#include <dt-bindings/clock/qcom,gpucc-sc8280xp.h>
> >>> #include <dt-bindings/clock/qcom,rpmh.h>
> >>> #include <dt-bindings/interconnect/qcom,osm-l3.h>
> >>> #include <dt-bindings/interconnect/qcom,sc8280xp.h>
> >>> @@ -2331,6 +2332,174 @@ tcsr: syscon@1fc0000 {
> >>> reg = <0x0 0x01fc0000 0x0 0x30000>;
> >>> };
> >>>
> >>> + gpu: gpu@3d00000 {
> >>> + compatible = "qcom,adreno-690.0", "qcom,adreno";
> >>> +
> >>> + reg = <0 0x03d00000 0 0x40000>,
> >>> + <0 0x03d9e000 0 0x1000>,
> >>> + <0 0x03d61000 0 0x800>;
> >>> + reg-names = "kgsl_3d0_reg_memory",
> >>> + "cx_mem",
> >>> + "cx_dbgc";
> >>> + interrupts = <GIC_SPI 300 IRQ_TYPE_LEVEL_HIGH>;
> >>> + iommus = <&adreno_smmu 0 0xc00>, <&adreno_smmu 1 0xc00>;
> >>> + operating-points-v2 = <&gpu_opp_table>;
> >>> +
> >>> + qcom,gmu = <&gmu>;
> >>> + interconnects = <&gem_noc MASTER_GFX3D 0 &mc_virt SLAVE_EBI1 0>;
> >>> + interconnect-names = "gfx-mem";
> >>> + #cooling-cells = <2>;
> >>> +
> >>> + status = "disabled";
> >>> +
> >>> + gpu_opp_table: opp-table {
> >>> + compatible = "operating-points-v2";
> >>> +
> >>> + opp-270000000 {
> >>> + opp-hz = /bits/ 64 <270000000>;
> >>> + opp-level = <RPMH_REGULATOR_LEVEL_LOW_SVS>;
> >>> + opp-peak-kBps = <451000>;
> >>> + };
> >>> +
> >>> + opp-410000000 {
> >>> + opp-hz = /bits/ 64 <410000000>;
> >>> + opp-level = <RPMH_REGULATOR_LEVEL_SVS>;
> >>> + opp-peak-kBps = <1555000>;
> >>> + };
> >>> +
> >>> + opp-500000000 {
> >>> + opp-hz = /bits/ 64 <500000000>;
> >>> + opp-level = <RPMH_REGULATOR_LEVEL_SVS_L1>;
> >>> + opp-peak-kBps = <1555000>;
> >>> + };
> >>> +
> >>> + opp-547000000 {
> >>> + opp-hz = /bits/ 64 <547000000>;
> >>> + opp-level = <RPMH_REGULATOR_LEVEL_SVS_L2>;
> >>> + opp-peak-kBps = <1555000>;
> >>> + };
> >>> +
> >>> + opp-606000000 {
> >>> + opp-hz = /bits/ 64 <606000000>;
> >>> + opp-level = <RPMH_REGULATOR_LEVEL_NOM>;
> >>> + opp-peak-kBps = <2736000>;
> >>> + };
> >>> +
> >>> + opp-640000000 {
> >>> + opp-hz = /bits/ 64 <640000000>;
> >>> + opp-level = <RPMH_REGULATOR_LEVEL_NOM_L1>;
> >>> + opp-peak-kBps = <2736000>;
> >>> + };
> >>> +
> >>> + opp-690000000 {
> >>> + opp-hz = /bits/ 64 <690000000>;
> >>> + opp-level = <RPMH_REGULATOR_LEVEL_TURBO>;
> >>> + opp-peak-kBps = <2736000>;
> >>> + };
> >>> + };
> >>> + };
> >>> +
> >>> + gmu: gmu@3d6a000 {
> >>> + compatible = "qcom,adreno-gmu-690.0", "qcom,adreno-gmu";
> >>> + reg = <0 0x03d6a000 0 0x34000>,
> >>> + <0 0x03de0000 0 0x10000>,
> >>> + <0 0x0b290000 0 0x10000>;
> >>> + reg-names = "gmu", "rscc", "gmu_pdc";
> >>> + interrupts = <GIC_SPI 304 IRQ_TYPE_LEVEL_HIGH>,
> >>> + <GIC_SPI 305 IRQ_TYPE_LEVEL_HIGH>;
> >>> + interrupt-names = "hfi", "gmu";
> >>> + clocks = <&gpucc GPU_CC_CX_GMU_CLK>,
> >>> + <&gpucc GPU_CC_CXO_CLK>,
> >>> + <&gcc GCC_DDRSS_GPU_AXI_CLK>,
> >>> + <&gcc GCC_GPU_MEMNOC_GFX_CLK>,
> >>> + <&gpucc GPU_CC_AHB_CLK>,
> >>> + <&gpucc GPU_CC_HUB_CX_INT_CLK>,
> >>> + <&gpucc GPU_CC_HLOS1_VOTE_GPU_SMMU_CLK>;
> >>> + clock-names = "gmu",
> >>> + "cxo",
> >>> + "axi",
> >>> + "memnoc",
> >>> + "ahb",
> >>> + "hub",
> >>> + "smmu_vote";
> >>> + power-domains = <&gpucc GPU_CC_CX_GDSC>,
> >>> + <&gpucc GPU_CC_GX_GDSC>;
> >>> + power-domain-names = "cx",
> >>> + "gx";
> >>> + iommus = <&adreno_smmu 5 0xc00>;
> >>> + operating-points-v2 = <&gmu_opp_table>;
> >>> +
> >>> + status = "disabled";
> >> I've recently discovered that - and I am not 100% sure - all GMUs are
> >> cache-coherent. Could you please ask somebody at qc about this?

If supported at hw and necessary pte attributes are present, all GPU transactions are
cache-coherent. Since gmu is part of GPU, it is a fairly good assumption that it would
be too.

But current set of GMUs doesn't do enough frequent chatter with CPU to
get any meaningful benefit with coherency. So I feel it is better to
leave it similar to downstream.

-Akhil

> >>
> >
> > AFAIU, GMU's job is controlling the voltage and clock to the GPU.
> Not just that, it's only the limited functionality we've implemented
> upstream so far.
>
> It doesn't do
> > any data transactions on its own.
> Of course it does. AP communication is done through MMIO writes and
> the GMU talks to RPMh via the GPU RSC directly. Apart from that, some
> of the GPU registers (that nota bene don't have anything to do with
> the GMU M3 core itself) lay within the GMU address space.
>
>
> Bjorn noticed that this coherent mask setting downstream may be
> a bluff, but I guess we could poke Qualcomm about whether it's
> cache-coherent (Akhil, could you say anything about that?).
>
> Konrad
>
> So cache-coherent doesn't make sense to me.
> >
> > - Mani
> >
> >>> +
> >>> + gmu_opp_table: opp-table {
> >>> + compatible = "operating-points-v2";
> >>> +
> >>> + opp-200000000 {
> >>> + opp-hz = /bits/ 64 <200000000>;
> >>> + opp-level = <RPMH_REGULATOR_LEVEL_MIN_SVS>;
> >>> + };
> >> Missing 500MHz + RPMH_REGULATOR_LEVEL_SVS
> >>
> >> (that may be used in the future for hw scheduling)
> >>> + };
> >>> + };
> >>> +
> >>> + gpucc: clock-controller@3d90000 {
> >>> + compatible = "qcom,sc8280xp-gpucc";
> >>> + reg = <0 0x03d90000 0 0x9000>;
> >>> + clocks = <&rpmhcc RPMH_CXO_CLK>,
> >>> + <&gcc GCC_GPU_GPLL0_CLK_SRC>,
> >>> + <&gcc GCC_GPU_GPLL0_DIV_CLK_SRC>;
> >>> + clock-names = "bi_tcxo",
> >>> + "gcc_gpu_gpll0_clk_src",
> >>> + "gcc_gpu_gpll0_div_clk_src";
> >> FWIW the driver doesn't use clock-names, but the binding defines it,
> >> so I suppose it's fine
> >>
> >>> +
> >>> + power-domains = <&rpmhpd SC8280XP_GFX>;
> >>> + #clock-cells = <1>;
> >>> + #reset-cells = <1>;
> >>> + #power-domain-cells = <1>;
> >>> +
> >>> + status = "disabled";
> >>> + };
> >>> +
> >>> + adreno_smmu: iommu@3da0000 {
> >>> + compatible = "qcom,sc8280xp-smmu-500", "qcom,adreno-smmu",
> >>> + "qcom,smmu-500", "arm,mmu-500";
> >>> + reg = <0 0x03da0000 0 0x20000>;
> >>> + #iommu-cells = <2>;
> >>> + #global-interrupts = <2>;
> >>> + interrupts = <GIC_SPI 672 IRQ_TYPE_LEVEL_HIGH>,
> >>> + <GIC_SPI 673 IRQ_TYPE_LEVEL_HIGH>,
> >>> + <GIC_SPI 678 IRQ_TYPE_LEVEL_HIGH>,
> >>> + <GIC_SPI 679 IRQ_TYPE_LEVEL_HIGH>,
> >>> + <GIC_SPI 680 IRQ_TYPE_LEVEL_HIGH>,
> >>> + <GIC_SPI 681 IRQ_TYPE_LEVEL_HIGH>,
> >>> + <GIC_SPI 682 IRQ_TYPE_LEVEL_HIGH>,
> >>> + <GIC_SPI 683 IRQ_TYPE_LEVEL_HIGH>,
> >>> + <GIC_SPI 684 IRQ_TYPE_LEVEL_HIGH>,
> >>> + <GIC_SPI 685 IRQ_TYPE_LEVEL_HIGH>,
> >>> + <GIC_SPI 686 IRQ_TYPE_LEVEL_HIGH>,
> >>> + <GIC_SPI 687 IRQ_TYPE_LEVEL_HIGH>,
> >>> + <GIC_SPI 688 IRQ_TYPE_LEVEL_HIGH>,
> >>> + <GIC_SPI 689 IRQ_TYPE_LEVEL_HIGH>;
> >>> +
> >>> + clocks = <&gcc GCC_GPU_MEMNOC_GFX_CLK>,
> >>> + <&gcc GCC_GPU_SNOC_DVM_GFX_CLK>,
> >>> + <&gpucc GPU_CC_AHB_CLK>,
> >>> + <&gpucc GPU_CC_HLOS1_VOTE_GPU_SMMU_CLK>,
> >>> + <&gpucc GPU_CC_CX_GMU_CLK>,
> >>> + <&gpucc GPU_CC_HUB_CX_INT_CLK>,
> >>> + <&gpucc GPU_CC_HUB_AON_CLK>;
> >>> + clock-names = "gcc_gpu_memnoc_gfx_clk",
> >>> + "gcc_gpu_snoc_dvm_gfx_clk",
> >>> + "gpu_cc_ahb_clk",
> >>> + "gpu_cc_hlos1_vote_gpu_smmu_clk",
> >>> + "gpu_cc_cx_gmu_clk",
> >>> + "gpu_cc_hub_cx_int_clk",
> >>> + "gpu_cc_hub_aon_clk";
> >>> +
> >>> + power-domains = <&gpucc GPU_CC_CX_GDSC>;
> >>> +
> >>> + status = "disabled";
> >> This one should be dma-coherent (per downstream, plus 8350's mmu is for sure)
> >>
> >> Konrad
> >>> + };
> >>> +
> >>> usb_0_hsphy: phy@88e5000 {
> >>> compatible = "qcom,sc8280xp-usb-hs-phy",
> >>> "qcom,usb-snps-hs-5nm-phy";
> >

Next message: Lee Jones: "Re: [PATCH v5 2/2] leds: add aw20xx driver"
Previous message: Boqun Feng: "Re: [PATCH v2 5/8] rust: workqueue: add helper for defining work_struct fields"
Next in thread: Akhil P Oommen: "Re: [PATCH v2 2/3] arm64: dts: qcom: sc8280xp: Add GPU related nodes"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]