Re: [PATCH 2/2] clk: qcom: clk-rpmh: Make all VRMs optional by default

[Konrad Dybcio]

On 4/13/26 11:02 AM, Alexander Koskovich wrote:
> On Monday, April 13th, 2026 at 4:48 AM, Konrad Dybcio <konrad.dybcio@xxxxxxxxxxxxxxxx> wrote:
> 
>> On 4/12/26 5:30 PM, Alexander Koskovich wrote:
>>> Some VRMs aren't present on all boards, so mark them as optional. This
>>> prevents probe failures on boards where not all VRMs are present.
>>>
>>> This mirrors the downstream approach for this issue.
>>>
>>> Signed-off-by: Alexander Koskovich <akoskovich@xxxxx>
>>> ---
>>
>> [...]
>>
>>> @@ -976,6 +982,11 @@ static int clk_rpmh_probe(struct platform_device *pdev)
>>>  		rpmh_clk = to_clk_rpmh(hw_clks[i]);
>>>  		res_addr = cmd_db_read_addr(rpmh_clk->res_name);
>>>  		if (!res_addr) {
>>> +			hw_clks[i] = NULL;
>>> +
>>> +			if (rpmh_clk->optional)
>>> +				continue;
>>
>> if (rpmh_clk->offset == CLK_RPMH_VRM_EN_OFFSET)?
>>
>> That would let you drop the rest of the patch.
>>
>> This also needs some additional background.. I would assume this comes from
>> a slightly different HW (PMIC) configuration
>>
>> Could you please do `grep ^ /sys/kernel/debug/qcom_socinfo/*`, (perhaps
>> censoring your serial number)?
> 
> /sys/kernel/debug/qcom_socinfo/accessory_chip:0
> /sys/kernel/debug/qcom_socinfo/boot_cluster:0
> /sys/kernel/debug/qcom_socinfo/boot_core:0
> /sys/kernel/debug/qcom_socinfo/build_id:
> /sys/kernel/debug/qcom_socinfo/chip_family:0x0000009f
> /sys/kernel/debug/qcom_socinfo/chip_id:SM7750
> /sys/kernel/debug/qcom_socinfo/feature_code:2
> /sys/kernel/debug/qcom_socinfo/foundry_id:1
> /sys/kernel/debug/qcom_socinfo/hardware_platform:11
> /sys/kernel/debug/qcom_socinfo/hardware_platform_subtype:0
> /sys/kernel/debug/qcom_socinfo/info_fmt:0x00000016
> /sys/kernel/debug/qcom_socinfo/ncluster_array_offset:228
> /sys/kernel/debug/qcom_socinfo/nmodem_supported:4294967295
> /sys/kernel/debug/qcom_socinfo/nproduct_id:1135
> /sys/kernel/debug/qcom_socinfo/nsubset_parts_array_offset:232
> /sys/kernel/debug/qcom_socinfo/num_clusters:1
> /sys/kernel/debug/qcom_socinfo/num_func_clusters:1
> /sys/kernel/debug/qcom_socinfo/num_subset_parts:20
> /sys/kernel/debug/qcom_socinfo/oem_variant:0
> /sys/kernel/debug/qcom_socinfo/pcode:1
> /sys/kernel/debug/qcom_socinfo/platform_version:65536
> /sys/kernel/debug/qcom_socinfo/pmic_die_rev:2.1
> /sys/kernel/debug/qcom_socinfo/pmic_model:PMK8550
> /sys/kernel/debug/qcom_socinfo/pmic_model_array:PMK8550 2.1
> /sys/kernel/debug/qcom_socinfo/pmic_model_array:PM7550 1.1
> /sys/kernel/debug/qcom_socinfo/pmic_model_array:PM8550VS 2.0
> /sys/kernel/debug/qcom_socinfo/pmic_model_array:PM8550VS 2.0
> /sys/kernel/debug/qcom_socinfo/pmic_model_array:unknown (90)
> /sys/kernel/debug/qcom_socinfo/pmic_model_array:PMR735D 2.0

So you have:

your phone	vs	Abel's Eliza-MTP

PMK8550			PMK8550
PM7550			PM7550
PM8550VS		PM8550VS
PM8550VS		PM8550VS
PMIV0102 (90)		PM7550BA	[!]
PMR735D			PMR735D
			PM8010		[!]
			PM8010		[!]

And hence the difference, makes sense

Konrad