Re: [PATCH v5 15/15] devicetree: bindings: Document qcom,pvs

From: Rob Herring
Date: Wed Dec 27 2017 - 16:59:05 EST


On Wed, Dec 27, 2017 at 4:20 AM, Sricharan R <sricharan@xxxxxxxxxxxxxx> wrote:
> Hi Rob,
>
> On 12/26/2017 11:06 PM, Rob Herring wrote:
>> On Thu, Dec 21, 2017 at 5:53 AM, Sricharan R <sricharan@xxxxxxxxxxxxxx> wrote:
>>> Hi Rob,
>>>
>>> On 12/21/2017 2:48 AM, Rob Herring wrote:
>>>> On Wed, Dec 20, 2017 at 11:55:33AM +0530, Sricharan R wrote:
>>>>> Hi Viresh,
>>>>>
>>>>> On 12/20/2017 8:56 AM, Viresh Kumar wrote:
>>>>>> On 19-12-17, 21:25, Sricharan R wrote:
>>>>>>> + cpu@0 {
>>>>>>> + compatible = "qcom,krait";
>>>>>>> + enable-method = "qcom,kpss-acc-v1";
>>>>>>> + device_type = "cpu";
>>>>>>> + reg = <0>;
>>>>>>> + qcom,acc = <&acc0>;
>>>>>>> + qcom,saw = <&saw0>;
>>>>>>> + clocks = <&kraitcc 0>;
>>>>>>> + clock-names = "cpu";
>>>>>>> + cpu-supply = <&smb208_s2a>;
>>>>>>> + operating-points-v2 = <&cpu_opp_table>;
>>>>>>> + };
>>>>>>> +
>>>>>>> + qcom,pvs {
>>>>>>> + qcom,pvs-format-a;
>>>>>>> + };
>>>>>>
>>>>>> Not sure what Rob is going to say on that :)
>>>>>>
>>>>>
>>>>> Yes. Would be good to know the best way.
>>>>
>>>> Seems like this should be a property of an efuse node either implied by
>>>> the compatible or a separate property. What determines format A vs. B?
>>>>
>>>
>>> Yes, this efuse registers are part of the eeprom (qfprom) tied to the soc.
>>> So this property (details like bitfields and register offsets that it represents)
>>> can be put soc specific and nvmem apis can be used to read
>>> the registers. Does something like below look ok ?
>>>
>>> qcom,pvs {
>>> compatible = "qcom,pvs-ipq8064";
>>> nvmem-cells = <&pvs_efuse>;
>>> }
>>
>> Why do you need this node? It doesn't look like it corresponds to a
>> h/w block. It looks like you are just creating it to instantiate a
>> driver.
>>
>>> qfprom: qfprom@700000 {
>>> compatible = "qcom,qfprom";
>>
>> Either this or...
>>
>>> reg = <0x00700000 0x1000>;
>>> #address-cells = <1>;
>>> #size-cells = <1>;
>>> ranges;
>>> pvs_efuse: pvs {
>>
>> a compatible here should be specific enough so the OS can know what
>> the bits are.
>
> Infact the above "qcom,pvs" node is required mainly to act as a consumer
> for the nvmem data provider ("qcom,qfprom") (using nvmem-cells = <&pvs_efuse>)
> Then "qfprom" can be made to contain a "format_a" or "format_b" specific cell.
>
> So all that is needed is, nvmem-cells = <&pvs_efuse_phandle> needs to be available
> somewhere. The requirement is similar what is now done by "operating-points-v2-ti-cpu"
> and the ti-cpufreq.c. There "operating-points-v2-ti-cpu" node, contains the syscon
> register to read the efuse values. Similarly does defining a new
> "operating-points-v2-krait-cpu" which would contain the nvmem-cells property look ok ?
> This would avoid defining a new qcom,pvs node.

Yes, this seems reasonable.

>
> cpu@0 {
> compatible = "qcom,krait";
> enable-method = "qcom,kpss-acc-v1";
> device_type = "cpu";
> reg = <0>;
> qcom,acc = <&acc0>;
> qcom,saw = <&saw0>;
> clocks = <&kraitcc 0>;
> clock-names = "cpu";
> cpu-supply = <&smb208_s2a>;
> operating-points-v2 = <&cpu_opp_table>;
> };
>
> cpu_opp_table: opp_table {
> compatible = "operating-points-v2-krait-cpu";
>
> nvmem-cells = <&pvs_efuse_format_a>;
> /*
> * Missing opp-shared property means CPUs switch DVFS states
> * independently.
> */
>
> opp-1400000000 {
> opp-hz = /bits/ 64 <1400000000>;
> opp-microvolt-speed0-pvs0-v0 = <1250000>;
> opp-microvolt-speed0-pvs1-v0 = <1175000>;
> opp-microvolt-speed0-pvs2-v0 = <1125000>;
> opp-microvolt-speed0-pvs3-v0 = <1050000>;
>
> };
> ...
> }
>
> qfprom: qfprom@700000 {
> compatible = "qcom,qfprom";
> reg = <0x00700000 0x1000>;
> #address-cells = <1>;
> #size-cells = <1>;
> ranges;
> pvs_efuse_format_a: pvs {
> reg = <0xc0 0x8>;
> };
> }
>
> Regards,
> Sricharan
>
> --
> "QUALCOMM INDIA, on behalf of Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation