[PATCH] arm64: dts: qcom: hamoa/x1: fix idle exit latency

[PATCH] arm64: dts: qcom: hamoa/x1: fix idle exit latency

[Daniel J Blueman]

Designs based on the Qualcomm X1 Hamoa reference platform report:
driver: Idle state 1 target residency too low

This is because the declared X1 idle entry plus exit latency of 680us
exceeds the declared minimum 600us residency time:
  entry-latency-us = <180>;
  exit-latency-us = <500>;
  min-residency-us = <600>;

As 500us idle exit latency is overly conservative, fix this to be a more
realistic 200us, in range of similar silicon. Tested on a Lenovo Yoga Slim
7x with Qualcomm X1E-80-100.

Signed-off-by: Daniel J Blueman <daniel@quora.org>
---
 arch/arm64/boot/dts/qcom/hamoa.dtsi | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/arm64/boot/dts/qcom/hamoa.dtsi b/arch/arm64/boot/dts/qcom/hamoa.dtsi
index a17900eacb20..b870f4dc9c42 100644
--- a/arch/arm64/boot/dts/qcom/hamoa.dtsi
+++ b/arch/arm64/boot/dts/qcom/hamoa.dtsi
@@ -269,7 +269,7 @@ cluster_c4: cpu-sleep-0 {
 				idle-state-name = "ret";
 				arm,psci-suspend-param = <0x00000004>;
 				entry-latency-us = <180>;
-				exit-latency-us = <500>;
+				exit-latency-us = <200>;
 				min-residency-us = <600>;
 			};
 		};
-- 
2.51.0

Re: [PATCH] arm64: dts: qcom: hamoa/x1: fix idle exit latency

[Konrad Dybcio]

On 2/20/26 8:02 AM, Daniel J Blueman wrote:
> Designs based on the Qualcomm X1 Hamoa reference platform report:
> driver: Idle state 1 target residency too low
> 
> This is because the declared X1 idle entry plus exit latency of 680us
> exceeds the declared minimum 600us residency time:
>   entry-latency-us = <180>;
>   exit-latency-us = <500>;
>   min-residency-us = <600>;
> 
> As 500us idle exit latency is overly conservative, fix this to be a more
> realistic 200us, in range of similar silicon. Tested on a Lenovo Yoga Slim
> 7x with Qualcomm X1E-80-100.
> 
> Signed-off-by: Daniel J Blueman <daniel@quora.org>
> ---
>  arch/arm64/boot/dts/qcom/hamoa.dtsi | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/arch/arm64/boot/dts/qcom/hamoa.dtsi b/arch/arm64/boot/dts/qcom/hamoa.dtsi
> index a17900eacb20..b870f4dc9c42 100644
> --- a/arch/arm64/boot/dts/qcom/hamoa.dtsi
> +++ b/arch/arm64/boot/dts/qcom/hamoa.dtsi
> @@ -269,7 +269,7 @@ cluster_c4: cpu-sleep-0 {
>  				idle-state-name = "ret";
>  				arm,psci-suspend-param = <0x00000004>;
>  				entry-latency-us = <180>;
> -				exit-latency-us = <500>;
> +				exit-latency-us = <200>;
>  				min-residency-us = <600>;

+Maulik?

Konrad

Re: [PATCH] arm64: dts: qcom: hamoa/x1: fix idle exit latency

[Maulik Shah (mkshah)]

On 2/20/2026 3:11 PM, Konrad Dybcio wrote:
> On 2/20/26 8:02 AM, Daniel J Blueman wrote:
>> Designs based on the Qualcomm X1 Hamoa reference platform report:
>> driver: Idle state 1 target residency too low
>>
>> This is because the declared X1 idle entry plus exit latency of 680us
>> exceeds the declared minimum 600us residency time:
>>   entry-latency-us = <180>;
>>   exit-latency-us = <500>;
>>   min-residency-us = <600>;
>>
>> As 500us idle exit latency is overly conservative, fix this to be a more
>> realistic 200us, in range of similar silicon. Tested on a Lenovo Yoga Slim
>> 7x with Qualcomm X1E-80-100.
>>
>> Signed-off-by: Daniel J Blueman <daniel@quora.org>
>> ---
>>  arch/arm64/boot/dts/qcom/hamoa.dtsi | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/arch/arm64/boot/dts/qcom/hamoa.dtsi b/arch/arm64/boot/dts/qcom/hamoa.dtsi
>> index a17900eacb20..b870f4dc9c42 100644
>> --- a/arch/arm64/boot/dts/qcom/hamoa.dtsi
>> +++ b/arch/arm64/boot/dts/qcom/hamoa.dtsi
>> @@ -269,7 +269,7 @@ cluster_c4: cpu-sleep-0 {
>>  				idle-state-name = "ret";
>>  				arm,psci-suspend-param = <0x00000004>;
>>  				entry-latency-us = <180>;
>> -				exit-latency-us = <500>;
>> +				exit-latency-us = <200>;
>>  				min-residency-us = <600>;
> 
> +Maulik?
> 
> Konrad

Linux dt idle driver sums up entry & exit latency and displays final number as total exit latency at
/sys/devices/system/cpu/cpu*/cpuidle/state1/latency

Can you please keep 320 usec as exit-latency-us to match with downstream 500usec total exit latency number from cluster_c4 state.

Thanks,
Maulik