Re: [PATCH 4/5] arm64: dts: qcom: sm6350: Add GPU nodes
From: Konrad Dybcio
Date: Mon Apr 03 2023 - 14:52:28 EST
On 3.04.2023 20:50, Dmitry Baryshkov wrote:
> On 18/03/2023 15:45, Konrad Dybcio wrote:
>>
>>
>> On 17.03.2023 09:56, Luca Weiss wrote:
>>> On Thu Mar 16, 2023 at 12:16 PM CET, Konrad Dybcio wrote:
>>>> From: Konrad Dybcio <konrad.dybcio@xxxxxxxxxxxxxx>
>>>>
>>>> Add Adreno, GPU SMMU and GMU nodes to hook up everything that
>>>> the A619 needs to function properly.
>>>>
>>>> Co-developed-by: Luca Weiss <luca.weiss@xxxxxxxxxxxxx>
>>>> Signed-off-by: Konrad Dybcio <konrad.dybcio@xxxxxxxxxxxxxx>
>>>> Signed-off-by: Konrad Dybcio <konrad.dybcio@xxxxxxxxxx>
>>>> ---
>> [...]
>>
>>> What about adding interconnect already? I also have opp-peak-kBps
>>> additions in the opp table for that. I'll attach the diff I have at the
>>> end of the email.
>> I believe the GMU takes care of it internally (or at least should)
>> with the bandwidth tables we send in a6xx_hfi.c : a6xx_hfi_send_bw_table()
>
> We should still declare the interconnects. If at some point we attempt to fill these tables in a proper way, the interconnects will be required to get addresses of the nodes.
A619 has all the "proper" data filled in. This should arguably
be switched to per-SoC and not per-GPU btw.
The interconnect endpoints should be looked up through the cmd_db
function like Bjorn did in the A690 patchset.
Konrad
>
>>
>>
>> [...]
>>
>>>> + adreno_smmu: iommu@3d40000 {
>>>
>>> This and gmu should be above gpucc @3d90000?
>> Absolutely.
>>
>> Konrad
>>>
>>>> + compatible = "qcom,sm6350-smmu-v2", "qcom,adreno-smmu", "qcom,smmu-v2";
>>>> + reg = <0 0x03d40000 0 0x10000>;
>>>> + #iommu-cells = <1>;
>>>> + #global-interrupts = <2>;
>>>> + interrupts = <GIC_SPI 229 IRQ_TYPE_LEVEL_HIGH>,
>>>> + <GIC_SPI 231 IRQ_TYPE_LEVEL_HIGH>,
>>>> + <GIC_SPI 364 IRQ_TYPE_LEVEL_HIGH>,
>>>> + <GIC_SPI 365 IRQ_TYPE_LEVEL_HIGH>,
>>>> + <GIC_SPI 366 IRQ_TYPE_LEVEL_HIGH>,
>>>> + <GIC_SPI 367 IRQ_TYPE_LEVEL_HIGH>,
>>>> + <GIC_SPI 368 IRQ_TYPE_LEVEL_HIGH>,
>>>> + <GIC_SPI 369 IRQ_TYPE_LEVEL_HIGH>,
>>>> + <GIC_SPI 370 IRQ_TYPE_LEVEL_HIGH>,
>>>> + <GIC_SPI 371 IRQ_TYPE_LEVEL_HIGH>;
>>>> +
>>>> + clocks = <&gpucc GPU_CC_AHB_CLK>,
>>>> + <&gcc GCC_GPU_MEMNOC_GFX_CLK>,
>>>> + <&gcc GCC_GPU_SNOC_DVM_GFX_CLK>;
>>>> + clock-names = "ahb",
>>>> + "bus",
>>>> + "iface";
>>>> +
>>>> + power-domains = <&gpucc GPU_CX_GDSC>;
>>>> + };
>>>> +
>>>> + gmu: gmu@3d6a000 {
>>>> + compatible = "qcom,adreno-gmu-619.0", "qcom,adreno-gmu";
>>>> + reg = <0 0x03d6a000 0 0x31000>,
>>>> + <0 0x0b290000 0 0x10000>,
>>>> + <0 0x0b490000 0 0x10000>;
>>>> + reg-names = "gmu",
>>>> + "gmu_pdc",
>>>> + "gmu_pdc_seq";
>>>> +
>>>> + interrupts = <GIC_SPI 304 IRQ_TYPE_LEVEL_HIGH>,
>>>> + <GIC_SPI 305 IRQ_TYPE_LEVEL_HIGH>;
>>>> + interrupt-names = "hfi",
>>>> + "gmu";
>>>> +
>>>> + clocks = <&gpucc GPU_CC_AHB_CLK>,
>>>> + <&gpucc GPU_CC_CX_GMU_CLK>,
>>>> + <&gpucc GPU_CC_CXO_CLK>,
>>>> + <&gcc GCC_DDRSS_GPU_AXI_CLK>,
>>>> + <&gcc GCC_GPU_MEMNOC_GFX_CLK>;
>>>> + clock-names = "ahb",
>>>> + "gmu",
>>>> + "cxo",
>>>> + "axi",
>>>> + "memnoc";
>>>> +
>>>> + power-domains = <&gpucc GPU_CX_GDSC>,
>>>> + <&gpucc GPU_GX_GDSC>;
>>>> + power-domain-names = "cx",
>>>> + "gx";
>>>> +
>>>> + iommus = <&adreno_smmu 5>;
>>>> +
>>>> + operating-points-v2 = <&gmu_opp_table>;
>>>> +
>>>> + status = "disabled";
>>>> +
>>>> + gmu_opp_table: opp-table {
>>>> + compatible = "operating-points-v2";
>>>> +
>>>> + opp-200000000 {
>>>> + opp-hz = /bits/ 64 <200000000>;
>>>> + opp-level = <RPMH_REGULATOR_LEVEL_MIN_SVS>;
>>>> + };
>>>> + };
>>>> + };
>>>> +
>>>> mpss: remoteproc@4080000 {
>>>> compatible = "qcom,sm6350-mpss-pas";
>>>> reg = <0x0 0x04080000 0x0 0x4040>;
>>>>
>>>> --
>>>> 2.39.2
>>>
>>> Here's the diff I have for interconnect on top of this:
>>>
>>> diff --git a/arch/arm64/boot/dts/qcom/sm6350.dtsi b/arch/arm64/boot/dts/qcom/sm6350.dtsi
>>> index 4954cbc2c0fc..51c5ac679a32 100644
>>> --- a/arch/arm64/boot/dts/qcom/sm6350.dtsi
>>> +++ b/arch/arm64/boot/dts/qcom/sm6350.dtsi
>>> @@ -1142,6 +1142,8 @@ gpu: gpu@3d00000 {
>>> iommus = <&adreno_smmu 0>;
>>> operating-points-v2 = <&gpu_opp_table>;
>>> qcom,gmu = <&gmu>;
>>> + interconnects = <&gem_noc MASTER_GRAPHICS_3D 0 &clk_virt SLAVE_EBI_CH0 0>;
>>> + interconnect-names = "gfx-mem";
>>> nvmem-cells = <&gpu_speed_bin>;
>>> nvmem-cell-names = "speed_bin";
>>> @@ -1157,42 +1159,49 @@ gpu_opp_table: opp-table {
>>> opp-850000000 {
>>> opp-hz = /bits/ 64 <850000000>;
>>> opp-level = <RPMH_REGULATOR_LEVEL_TURBO_L1>;
>>> + opp-peak-kBps = <8371200>;
>>> opp-supported-hw = <0x02>;
>>> };
>>> opp-800000000 {
>>> opp-hz = /bits/ 64 <800000000>;
>>> opp-level = <RPMH_REGULATOR_LEVEL_TURBO>;
>>> + opp-peak-kBps = <8371200>;
>>> opp-supported-hw = <0x04>;
>>> };
>>> opp-650000000 {
>>> opp-hz = /bits/ 64 <650000000>;
>>> opp-level = <RPMH_REGULATOR_LEVEL_NOM_L1>;
>>> + opp-peak-kBps = <6220000>;
>>> opp-supported-hw = <0x08>;
>>> };
>>> opp-565000000 {
>>> opp-hz = /bits/ 64 <565000000>;
>>> opp-level = <RPMH_REGULATOR_LEVEL_NOM>;
>>> + opp-peak-kBps = <5412000>;
>>> opp-supported-hw = <0x10>;
>>> };
>>> opp-430000000 {
>>> opp-hz = /bits/ 64 <430000000>;
>>> opp-level = <RPMH_REGULATOR_LEVEL_SVS_L1>;
>>> + opp-peak-kBps = <4068000>;
>>> opp-supported-hw = <0xff>;
>>> };
>>> opp-355000000 {
>>> opp-hz = /bits/ 64 <355000000>;
>>> opp-level = <RPMH_REGULATOR_LEVEL_SVS>;
>>> + opp-peak-kBps = <3072000>;
>>> opp-supported-hw = <0xff>;
>>> };
>>> opp-253000000 {
>>> opp-hz = /bits/ 64 <253000000>;
>>> opp-level = <RPMH_REGULATOR_LEVEL_LOW_SVS>;
>>> + opp-peak-kBps = <2188000>;
>>> opp-supported-hw = <0xff>;
>>> };
>>> };
>>>
>