Message ID | 20200605185916.318494-3-stephan@gerhold.net (mailing list archive) |
---|---|
State | Accepted |
Commit | e2ee9edc282961783d519c760bbaa20fed4dec38 |
Headers | show |
Series | [1/4] arm64: dts: qcom: msm8916: Set #address-cells for lpass | expand |
diff --git a/arch/arm64/boot/dts/qcom/msm8916-pins.dtsi b/arch/arm64/boot/dts/qcom/msm8916-pins.dtsi index 5785bf0a807c..591f48a57535 100644 --- a/arch/arm64/boot/dts/qcom/msm8916-pins.dtsi +++ b/arch/arm64/boot/dts/qcom/msm8916-pins.dtsi @@ -569,7 +569,7 @@ pinconf { pins = "gpio63", "gpio64", "gpio65", "gpio66", "gpio67", "gpio68"; drive-strength = <2>; - bias-disable; + bias-pull-down; }; }; };
The original qcom kernel changed the PDM GPIOs to be pull-down during sleep at some point. Reportedly this was done because there was some "leakage at PDM outputs during sleep": https://source.codeaurora.org/quic/la/kernel/msm-3.10/commit/?id=0f87e08c1cd3e6484a6f7fb3e74e37340bdcdee0 I cannot say how effective this is, but everything seems to work fine with this change so let's apply the same to mainline just to be sure. Cc: Srinivas Kandagatla <srinivas.kandagatla@linaro.org> Signed-off-by: Stephan Gerhold <stephan@gerhold.net> --- I found this at some point while trying to make audio work (searched for differences in mainline compared to downstream). This was not a difference that prevented it to work, but I thought I would just submit it in case this could improve power efficiency a bit. Feel free to ignore this if this is stupid. --- arch/arm64/boot/dts/qcom/msm8916-pins.dtsi | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)