linux-kernel - Re: [PATCH v1 2/2] arm64: dts: qcom: Enable TSENS support for QCS615 SoC

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <2b889254-2847-4c6b-a01d-3626332dcb0a@oss.qualcomm.com>
Date: Mon, 14 Apr 2025 11:53:42 +0200
From: Konrad Dybcio <konrad.dybcio@....qualcomm.com>
To: Gaurav Kohli <quic_gkohli@...cinc.com>, amitk@...nel.org,
        rafael@...nel.org, daniel.lezcano@...aro.org, rui.zhang@...el.com,
        lukasz.luba@....com, robh@...nel.org, krzk+dt@...nel.org,
        andersson@...nel.org, konradybcio@...nel.org
Cc: devicetree@...r.kernel.org, linux-kernel@...r.kernel.org,
        linux-arm-msm@...r.kernel.org, quic_manafm@...cinc.com
Subject: Re: [PATCH v1 2/2] arm64: dts: qcom: Enable TSENS support for QCS615
 SoC

On 4/14/25 10:28 AM, Gaurav Kohli wrote:
> thanks for review!
> 
> On 4/12/2025 5:13 AM, Konrad Dybcio wrote:
>> On 4/10/25 4:00 PM, Gaurav Kohli wrote:
>>> Add TSENS and thermal devicetree node for QCS615 SoC.
>>>
>>> Signed-off-by: Gaurav Kohli <quic_gkohli@...cinc.com>
>>> ---
>>
>> subject: "arm64: dts: qcom: qcs615: ..">  arch/arm64/boot/dts/qcom/qcs615.dtsi | 281 +++++++++++++++++++++++++++
>>>   1 file changed, 281 insertions(+)
>>>
> will fix this.
>>> diff --git a/arch/arm64/boot/dts/qcom/qcs615.dtsi b/arch/arm64/boot/dts/qcom/qcs615.dtsi
>>> index edfb796d8dd3..f0d8aed7da29 100644
>>> --- a/arch/arm64/boot/dts/qcom/qcs615.dtsi
>>> +++ b/arch/arm64/boot/dts/qcom/qcs615.dtsi
>>> @@ -3668,6 +3668,17 @@ usb_2_dwc3: usb@...0000 {
>>>                   maximum-speed = "high-speed";
>>>               };
>>>           };
>>> +
>>> +        tsens0: tsens@...2000 {
>>> +            compatible = "qcom,qcs615-tsens", "qcom,tsens-v2";
>>> +            reg = <0x0 0xc263000 0x0 0x1ff>,
>>> +                <0x0 0xc222000 0x0 0x8>;
>> Pad the address part to 8 hex digits with leading zeroes> +            interrupts = <GIC_SPI 506 IRQ_TYPE_LEVEL_HIGH>,
>>
>> &pdc 26
>>
>>> +                    <GIC_SPI 508 IRQ_TYPE_LEVEL_HIGH>;
>>
>> &pdc 28
> we don't want to mark this as wake up capable, so using this approach.

Why not?

>>> +
>>> +        cpuss-0-thermal {
>>> +            thermal-sensors = <&tsens0 1>;
>>> +
>>> +            trips {
>>> +
>>> +                trip-point0 {
>>> +                    temperature = <115000>;
>>> +                    hysteresis = <5000>;
>>> +                    type = "passive";
>>> +                };
>>> +
>>> +                trip-point1 {
>>> +                    temperature = <118000>;
>>> +                    hysteresis = <5000>;
>>> +                    type = "passive";
>>> +                };
>>
>> Please drop the passive trip point for the *CPU* tzones, see
>>
> 
> we are using trip-point 0 for cpu idle injection mitigation which i will add in subsequent patches, if you are fine i will add cpu idle injection cooling map in this series only ?

The folks working on qcs9xxx have made this point too, but I'm lukewarm
on duplicating meaningless dt description everywhere. I've asked them to
conduct some measurements on whether random default settings (that would
be preset in the driver and require no additional dt fluff) show any
significant difference - if not, we can save up on boilerplate.

So let's wait to hear back from them on this.

>> commit 06eadce936971dd11279e53b6dfb151804137836
>> ("arm64: dts: qcom: x1e80100: Drop unused passive thermal trip points for CPU")
>>
>> and add a single critical point instead, see
>>
> As critical shutdown is already supported by hardware, so we are not defining here.

The hardware critical shutdown will literally pull the plug out with the OS
having no chance to sync the filesystem etc.

Please define one that's like 5 degC below the hardware limit, so that the
operating system can try to take some steps to avoid data loss

Konrad