| Message ID | 20230103010904.3201835-1-bryan.odonoghue@linaro.org |
|---|---|
| Headers | show |
| Series | Add MSM8939 SoC support with two devices | expand |
On 03/01/2023 09:14, Krzysztof Kozlowski wrote: > ../arch/arm64/boot/dts/qcom/msm8939.dtsi:1825.23-1842.5: Warning > (simple_bus_reg): /soc@0/mmc@7824000: simple-bus unit address format > error, expected "7824900 For the record the driver consuming this dt entry requires the host regs to come first followed by the core sdhc_1: mmc@7824000 { compatible = "qcom,msm8916-sdhci", "qcom,sdhci-msm-v4"; reg = <0x07824900 0x11c>, <0x07824000 0x800>; reg-names = "hc", "core"; } If I change this and the msm8916 to sdhc_1: mmc@7824000 { compatible = "qcom,msm8916-sdhci", "qcom,sdhci-msm-v4"; reg = <0x07824000 0x800>, <0x07824900 0x11c>; reg-names = "core", "hc"; } the warning goes away but then this code won't work as intended anymore. drivers/mmc/host/sdhci-msm.c: msm_host->core_mem = devm_platform_ioremap_resource(pdev, 1); Weirdly when I change the ordering of the registers in the dts the SD seems to work just fine, which is at variance to my reading of the code. Converting the sdhci driver to use the register name, instead of a static ordering of the regs feels like something to be done in a separate series. --- bod
On 10/01/2023 14:14, Bryan O'Donoghue wrote: > On 03/01/2023 09:14, Krzysztof Kozlowski wrote: >> ../arch/arm64/boot/dts/qcom/msm8939.dtsi:1825.23-1842.5: Warning >> (simple_bus_reg): /soc@0/mmc@7824000: simple-bus unit address format >> error, expected "7824900 > > For the record the driver consuming this dt entry requires the host regs > to come first followed by the core > > sdhc_1: mmc@7824000 { > compatible = "qcom,msm8916-sdhci", "qcom,sdhci-msm-v4"; > reg = <0x07824900 0x11c>, <0x07824000 0x800>; > reg-names = "hc", "core"; > } > > If I change this and the msm8916 to That's not the solution. The warning is saying that unit address does not match your reg. You need to correct unit address. Best regards, Krzysztof
On 10/01/2023 13:24, Krzysztof Kozlowski wrote: > On 10/01/2023 14:14, Bryan O'Donoghue wrote: >> On 03/01/2023 09:14, Krzysztof Kozlowski wrote: >>> ../arch/arm64/boot/dts/qcom/msm8939.dtsi:1825.23-1842.5: Warning >>> (simple_bus_reg): /soc@0/mmc@7824000: simple-bus unit address format >>> error, expected "7824900 >> >> For the record the driver consuming this dt entry requires the host regs >> to come first followed by the core >> >> sdhc_1: mmc@7824000 { >> compatible = "qcom,msm8916-sdhci", "qcom,sdhci-msm-v4"; >> reg = <0x07824900 0x11c>, <0x07824000 0x800>; >> reg-names = "hc", "core"; >> } >> >> If I change this and the msm8916 to > > That's not the solution. The warning is saying that unit address does > not match your reg. You need to correct unit address. > > Best regards, > Krzysztof > Is it not the case that the unit-address should match the first reg and that the first reg should also be the lowest address ? --- bod
On 10.01.2023 14:42, Bryan O'Donoghue wrote: > On 10/01/2023 13:24, Krzysztof Kozlowski wrote: >> On 10/01/2023 14:14, Bryan O'Donoghue wrote: >>> On 03/01/2023 09:14, Krzysztof Kozlowski wrote: >>>> ../arch/arm64/boot/dts/qcom/msm8939.dtsi:1825.23-1842.5: Warning >>>> (simple_bus_reg): /soc@0/mmc@7824000: simple-bus unit address format >>>> error, expected "7824900 >>> >>> For the record the driver consuming this dt entry requires the host regs >>> to come first followed by the core >>> >>> sdhc_1: mmc@7824000 { >>> compatible = "qcom,msm8916-sdhci", "qcom,sdhci-msm-v4"; >>> reg = <0x07824900 0x11c>, <0x07824000 0x800>; >>> reg-names = "hc", "core"; >>> } >>> >>> If I change this and the msm8916 to >> >> That's not the solution. The warning is saying that unit address does >> not match your reg. You need to correct unit address. >> >> Best regards, >> Krzysztof >> > > Is it not the case that the unit-address should match the first reg Yes and that the first reg should also be the lowest address ? Depends on what the first reg is, bindings usually dictate the order. Konrad > > --- > bod
On 10/01/2023 14:42, Bryan O'Donoghue wrote: > On 10/01/2023 13:24, Krzysztof Kozlowski wrote: >> On 10/01/2023 14:14, Bryan O'Donoghue wrote: >>> On 03/01/2023 09:14, Krzysztof Kozlowski wrote: >>>> ../arch/arm64/boot/dts/qcom/msm8939.dtsi:1825.23-1842.5: Warning >>>> (simple_bus_reg): /soc@0/mmc@7824000: simple-bus unit address format >>>> error, expected "7824900 >>> >>> For the record the driver consuming this dt entry requires the host regs >>> to come first followed by the core >>> >>> sdhc_1: mmc@7824000 { >>> compatible = "qcom,msm8916-sdhci", "qcom,sdhci-msm-v4"; >>> reg = <0x07824900 0x11c>, <0x07824000 0x800>; >>> reg-names = "hc", "core"; >>> } >>> >>> If I change this and the msm8916 to >> >> That's not the solution. The warning is saying that unit address does >> not match your reg. You need to correct unit address. >> >> Best regards, >> Krzysztof >> > > Is it not the case that the unit-address should match the first reg and As Konrad wrote, yes. > that the first reg should also be the lowest address ? I never heard such requirement, although it would be logical choice. In this case your solution is to use the higher address as unit address. apq8084, msm8226 and maybe more already do it. You can also look at this that HC was made as the main IO address of a device (if one can call something "main" here), thus it should be expressed in unit address. Best regards, Krzysztof
V2: - Sorts core dtsi node list by address followed by alpahbetical sorting within address sorted nodes - Bjorn - Drops use of 8916-pins - Bjorn - Adds msm8939-pm8916.dtsi - Stephan - Fixes every dts splat from previous submission minus non-converted .txt compat strings [1] and one yaml error in Bjorn's tree not in -next yet - I haven't applied Dmitry's change for tsens since that's not been picked up yet - Picks up a number of suggestions and fixes from Stephan Gerhold and Vincent Knecht - Depends on Applied: [PATCH v4 0/7] remoteproc: qcom_q6v5_mss: Add MSM8909 and MSM8953 https://lore.kernel.org/linux-arm-msm/167216232800.738877.17567287056128563074.b4-ty@kernel.org/ [PATCH v6 0/5] remoteproc: qcom: Add support for pronto-v3 https://lore.kernel.org/linux-arm-msm/167216232801.738877.15895916910585144737.b4-ty@kernel.org/ [PATCH v6 00/18] mdss-dsi-ctrl binding and dts fixes https://lore.kernel.org/linux-arm-msm/167233461766.1099840.17628700245792986354.b4-ty@kernel.org/ Awaiting application: https://lore.kernel.org/linux-arm-msm/20221228133058.213886-1-bryan.odonoghue@linaro.org/ - Previous https://lore.kernel.org/linux-arm-msm/20220419010903.3109514-1-bryan.odonoghue@linaro.org/ - Bootable tree https://git.linaro.org/people/bryan.odonoghue/kernel.git/log/?h=linux-next-23-01-03-msm8939-no-cpr - [1] DTC_CHK arch/arm64/boot/dts/qcom/apq8039-t2.dtb Documentation/devicetree/bindings/arm/msm/qcom,idle-state.txt qcom/apq8039-t2.dtb: idle-states: cpu-sleep-0:compatible:0: 'qcom,idle-state-spc' is not one of ['arm,idle-state', 'riscv,idle-state'] From schema: Documentation/devicetree/bindings/cpu/idle-states.yaml qcom/apq8039-t2.dtb: idle-states: cpu-sleep-0:compatible: ['qcom,idle-state-spc', 'arm,idle-state'] is too long From schema: Documentation/devicetree/bindings/cpu/idle-states.yaml arch/arm64/boot/dts/qcom/apq8039-t2.dtb:0:0: /cpus/idle-states/cpu-sleep-0: failed to match any schema with compatible: ['qcom,idle-state-spc', 'arm,idle-state'] Documentation/devicetree/bindings/iommu/qcom,iommu.txt arch/arm64/boot/dts/qcom/apq8039-t2.dtb:0:0: /soc@0/iommu@1ef0000: failed to match any schema with compatible: ['qcom,msm8916-iommu', 'qcom,msm-iommu-v1'] arch/arm64/boot/dts/qcom/apq8039-t2.dtb:0:0: /soc@0/iommu@1ef0000/iommu-ctx@4000: failed to match any schema with compatible: ['qcom,msm-iommu-v1-ns'] arch/arm64/boot/dts/qcom/apq8039-t2.dtb:0:0: /soc@0/iommu@1ef0000/iommu-ctx@5000: failed to match any schema with compatible: ['qcom,msm-iommu-v1-sec'] arch/arm64/boot/dts/qcom/apq8039-t2.dtb:0:0: /soc@0/iommu@1f08000: failed to match any schema with compatible: ['qcom,msm8916-iommu', 'qcom,msm-iommu-v1'] arch/arm64/boot/dts/qcom/apq8039-t2.dtb:0:0: /soc@0/iommu@1f08000/iommu-ctx@1000: failed to match any schema with compatible: ['qcom,msm-iommu-v1-ns'] arch/arm64/boot/dts/qcom/apq8039-t2.dtb:0:0: /soc@0/iommu@1f08000/iommu-ctx@2000: failed to match any schema with compatible: ['qcom,msm-iommu-v1-ns'] arch/arm64/boot/dts/qcom/pm8916.dtsi f5d7bca55425c8 qcom/apq8039-t2.dtb: pmic@0: 'extcon@1300' does not match any of the regexes: '(.*)?(wled|leds)@[0-9a-f]+$', '^adc-tm@[0-9a-f]+$', '^adc@[0-9a-f]+$', '^audio-codec@[0-9a-f]+$', '^charger@[0-9a-f]+$', '^mpps@[0-9a-f]+$', '^rtc@[0-9a-f]+$', '^temp-alarm@[0-9a-f]+$', '^usb-detect@[0-9a-f]+$', '^usb-vbus-regulator@[0-9a-f]+$', '^vibrator@[0-9a-f]+$', 'gpio@[0-9a-f]+$', 'pinctrl-[0-9]+', 'pon@[0-9a-f]+$' From schema: Documentation/devicetree/bindings/mfd/qcom,spmi-pmic.yaml Documentation/devicetree/bindings/sound/qcom,msm8916-wcd-analog.txt arch/arm64/boot/dts/qcom/apq8039-t2.dtb:0:0: /soc@0/spmi@200f000/pmic@1/audio-codec@f000: failed to match any schema with compatible: ['qcom,pm8916-wcd-analog-codec'] yaml documentation error not yet in -next arm64/boot/dts/qcom/apq8039-t2.dtb: remoteproc@4080000: qcom,halt-regs:0: [33] is too short From schema: Documentation/devicetree/bindings/remoteproc/qcom,msm8916-mss-pil.yaml Documentation/devicetree/bindings/usb/ci-hdrc-usb2.txt arch/arm64/boot/dts/qcom/apq8039-t2.dtb:0:0: /soc@0/usb@78d9000: failed to match any schema with compatible: ['qcom,ci-hdrc'] Documentation/devicetree/bindings/arm/msm/qcom,kpss-acc.txt: compatible = "qcom,kpss-acc-v2"; arch/arm64/boot/dts/qcom/apq8039-t2.dtb:0:0: /soc@0/clock-controller@b088000: failed to match any schema with compatible: ['qcom,kpss-acc-v2'] arch/arm64/boot/dts/qcom/apq8039-t2.dtb:0:0: /soc@0/clock-controller@b098000: failed to match any schema with compatible: ['qcom,kpss-acc-v2'] arch/arm64/boot/dts/qcom/apq8039-t2.dtb:0:0: /soc@0/clock-controller@b0a8000: failed to match any schema with compatible: ['qcom,kpss-acc-v2'] arch/arm64/boot/dts/qcom/apq8039-t2.dtb:0:0: /soc@0/clock-controller@b0b8000: failed to match any schema with compatible: ['qcom,kpss-acc-v2'] arch/arm64/boot/dts/qcom/apq8039-t2.dtb:0:0: /soc@0/clock-controller@b188000: failed to match any schema with compatible: ['qcom,kpss-acc-v2'] arch/arm64/boot/dts/qcom/apq8039-t2.dtb:0:0: /soc@0/clock-controller@b198000: failed to match any schema with compatible: ['qcom,kpss-acc-v2'] arch/arm64/boot/dts/qcom/apq8039-t2.dtb:0:0: /soc@0/clock-controller@b1a8000: failed to match any schema with compatible: ['qcom,kpss-acc-v2'] arch/arm64/boot/dts/qcom/apq8039-t2.dtb:0:0: /soc@0/clock-controller@b1b8000: failed to match any schema with compatible: ['qcom,kpss-acc-v2'] V1: This series adds in MSM8939 SoC support with two supported devices. - CPU MSM8939 is a non-PSCI compliant device. As such in the downstreaming shipped image custom code is used to bring non-boot cores out of reset. This drop specifies the boot-method as spin-table instead and is completely standard. To accomplish this, we rely on lk2nd. https://github.com/msm8916-mainline/lk2nd/pull/142 - Serial - i2c - USB - eMMC - MDP/DSI - WiFi - Bluetooth What's not included - CPR We have CPR working in a 4.19 kernel quite well but for now it feels like putting the cart before the horse to gate the SoC and boards on CPR. - Venus I've been told this works but I haven't tried it myself and right now consider it maybe working but probably not 100%. - Sound We have a copy-exactly from the 4.19 kernel here in the DTS. I haven't run the sound through any sort of reasonable test. Vincent Knecht has some PostmarketOS kernels which use a 5.17 version of this DTS to get sound up so, I think sound is in good shape. - CAMSS There are slight differences between msm8916 and msm8939 for CAMSS. It doesn't feel like tons of work but, right now it is work we haven't even started. - Devices I've booted on the Square device obviously and this is my regular hardware for upstream development. I've also booted on the Sony Xperia M4 Aqua including mutli-core bring-up, WiFi and ADB. Dependencies for this drop: qcom-cpufreq-nvmem: Add msm8939 with some fixups link: https://lore.kernel.org/linux-arm-msm/20220418162226.2983117-1-bryan.odonoghue@linaro.org/T/#t Fix apq8016 compat string link: https://lore.kernel.org/linux-arm-msm/20220418230956.3059563-1-bryan.odonoghue@linaro.org/T/#t dt-bindings: soc: qcom: smd-rpm: Fix missing MSM8936 compatible link: https://lore.kernel.org/linux-arm-msm/20220418231857.3061053-1-bryan.odonoghue@linaro.org/T/#u Bootable tree here: https://git.linaro.org/people/bryan.odonoghue/kernel.git/log/?h=v5.18-rc2%2bapq8039-without-cpr Bryan O'Donoghue (6): dt-bindings: interconnect: Move interconnect child node definition dt-bindings: arm: qcom: Document MSM8939 SoC binding dt-bindings: soc: qcom: smd-rpm: Exclude MSM8936 from glink-channels arm64: dts: qcom: Add msm8939 SoC arm64: dts: qcom: Add Square apq8039-t2 board arm64: dts: qcom: Add msm8939 Sony Xperia M4 Aqua Stephan Gerhold (1): arm64: dts: qcom: Add msm8939-pm8916.dtsi include .../devicetree/bindings/arm/qcom.yaml | 8 + .../bindings/interconnect/qcom,rpm.yaml | 62 +- .../bindings/soc/qcom/qcom,smd-rpm.yaml | 1 + arch/arm64/boot/dts/qcom/Makefile | 2 + arch/arm64/boot/dts/qcom/apq8039-t2.dts | 557 ++++ arch/arm64/boot/dts/qcom/msm8939-pm8916.dtsi | 82 + .../qcom/msm8939-sony-xperia-kanuti-tulip.dts | 472 ++++ arch/arm64/boot/dts/qcom/msm8939.dtsi | 2470 +++++++++++++++++ 8 files changed, 3623 insertions(+), 31 deletions(-) create mode 100644 arch/arm64/boot/dts/qcom/apq8039-t2.dts create mode 100644 arch/arm64/boot/dts/qcom/msm8939-pm8916.dtsi create mode 100644 arch/arm64/boot/dts/qcom/msm8939-sony-xperia-kanuti-tulip.dts create mode 100644 arch/arm64/boot/dts/qcom/msm8939.dtsi