Re: [PATCH v3 2/3] arm64: dts: qcom: talos: Add Talos Lyra EVK SoM platform

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] arm64: dts: qcom: talos-lyra-evk-som: wrong vqmmc-supply on sdhc_2 breaks 3.3V/UHS-I
--

commit 9f59ab084ac5176f461f2df3b23f3bc182845992
Author: Avaneesh Kumar Dwivedi <avaneesh.dwivedi@oss.qualcomm.com>

arm64: dts: qcom: talos: Add Talos Lyra EVK SoM platform

This commit introduces the initial device tree for the QCS615-based Talos Lyra
EVK SoM platform, including definitions for CPU, memory, PMIC regulators, UFS
and SD/MMC storage, QUPv3, remoteprocs, and GPU.

> diff --git a/arch/arm64/boot/dts/qcom/talos-lyra-evk-som.dtsi b/arch/arm64/boot/dts/qcom/talos-lyra-evk-som.dtsi
> new file mode 100644
> index 0000000000000..7d9a084719067
> --- /dev/null
> +++ b/arch/arm64/boot/dts/qcom/talos-lyra-evk-som.dtsi
> @@ -0,0 +1,281 @@

[ ... ]

> +&apps_rsc {
> +	regulators-0 {
> +		compatible = "qcom,pm8150-rpmh-regulators";
> +		qcom,pmic-id = "a";
> +
> +		vreg_s4a: smps4 {
> +			regulator-name = "vreg_s4a";
> +			regulator-min-microvolt = <1800000>;
> +			regulator-max-microvolt = <1829000>;
> +			regulator-initial-mode = <RPMH_REGULATOR_MODE_HPM>;
> +		};

[ ... ]

> +		vreg_l2a: ldo2 {
> +			regulator-name = "vreg_l2a";
> +			regulator-min-microvolt = <1650000>;
> +			regulator-max-microvolt = <3100000>;
> +			regulator-initial-mode = <RPMH_REGULATOR_MODE_HPM>;
> +			regulator-allow-set-load;
> +			regulator-allowed-modes = <RPMH_REGULATOR_MODE_LPM
> +						   RPMH_REGULATOR_MODE_HPM>;
> +		};

[ ... ]

> +&sdhc_2 {
> +	pinctrl-0 = <&sdc2_state_on>;
> +	pinctrl-1 = <&sdc2_state_off>;
> +	pinctrl-names = "default", "sleep";
> +	bus-width = <4>;
> +	cd-gpios = <&tlmm 99 GPIO_ACTIVE_LOW>;
> +	vmmc-supply = <&vreg_l10a>;
> +	vqmmc-supply = <&vreg_s4a>;

[Severity: Medium]
Should sdhc_2 use vreg_l2a instead of vreg_s4a for vqmmc-supply?

Here, vreg_s4a is constrained to 1.8V-1.829V, while sdhc_2 is the removable SD
card controller. When sdhci_msm_set_regulator_caps() in
drivers/mmc/host/sdhci-msm.c and sdhci_setup_host() in
drivers/mmc/host/sdhci.c check vqmmc, they see that 2.7V-3.6V is unsupported:

drivers/mmc/host/sdhci-msm.c:sdhci_msm_set_regulator_caps()
	if (!IS_ERR(mmc->supply.vqmmc)) {
		if (regulator_is_supported_voltage(supply, 1700000, 1950000))
			caps |= CORE_1_8V_SUPPORT;
		if (regulator_is_supported_voltage(supply, 2700000, 3600000))
			caps |= CORE_3_0V_SUPPORT;

drivers/mmc/host/sdhci.c:sdhci_setup_host()
		if (!regulator_is_supported_voltage(mmc->supply.vqmmc, 2700000,
						    3600000))
			host->flags &= ~SDHCI_SIGNALING_330;

Because CORE_3_0V_SUPPORT is not set and SDHCI_SIGNALING_330 is cleared,
initial 3.3V SD signaling in mmc_power_up() ->
mmc_set_initial_signal_voltage() -> sdhci_msm_start_signal_voltage_switch()
fails with -EINVAL, breaking 3.3V initial I/O signaling and UHS-I voltage
switching.

Meanwhile, vreg_l2a (ldo2, 1.65V-3.1V) is defined in this file with no
consumers, similar to the issue on QCS615 Ride fixed in commit 4614628bea86
("arm64: dts: qcom: qcs615-ride: fix sdhc_2 vqmmc-supply for UHS-I mode").

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261027-talos-lyra-evk-v2-v3-0-658fd5a7ec99@oss.qualcomm.com?part=2