Re: [PATCH v2 2/3] arm64: dts: qcom: pm8916: Add BMS and charger

[Konrad Dybcio]

On 27.10.2023 07:44, Nikita Travkin wrote:
> Konrad Dybcio писал(а) 27.10.2023 01:03:
>> On 10/26/23 21:17, Stephan Gerhold wrote:
>>> On Thu, Oct 26, 2023 at 08:54:00PM +0200, Konrad Dybcio wrote:
>>>> On 10/24/23 11:29, Nikita Travkin wrote:
>>>>> Konrad Dybcio писал(а) 24.10.2023 13:34:
>>>>>> On 10/23/23 08:20, Nikita Travkin wrote:
>>>>>>> pm8916 contains some hardware blocks for battery powered devices:
>>>>>>>
>>>>>>> - VM-BMS: Battery voltage monitoring block.
>>>>>>> - LBC: Linear battery charger.
>>>>>>>
>>>>>>> Add them to the pmic dtsi so the devices that make use of those blocks
>>>>>>> can enable them.
>>>>>>>
>>>>>>> Signed-off-by: Nikita Travkin <nikita@xxxxxxx>
>>>>>>> ---
>>>>>>>     arch/arm64/boot/dts/qcom/pm8916.dtsi | 48 ++++++++++++++++++++++++++++++++++++
>>>>>>>     1 file changed, 48 insertions(+)
>>>>>>>
>>>>>>> diff --git a/arch/arm64/boot/dts/qcom/pm8916.dtsi b/arch/arm64/boot/dts/qcom/pm8916.dtsi
>>>>>>> index f4de86787743..4b2e8fb47d2d 100644
>>>>>>> --- a/arch/arm64/boot/dts/qcom/pm8916.dtsi
>>>>>>> +++ b/arch/arm64/boot/dts/qcom/pm8916.dtsi
>>>>>>> @@ -41,6 +41,35 @@ watchdog {
>>>>>>>     			};
>>>>>>>     		};
>>>>>>>     +		pm8916_charger: charger@1000 {
>>>>>>> +			compatible = "qcom,pm8916-lbc";
>>>>>>> +			reg = <0x1000>, <0x1200>, <0x1300>, <0x1600>;
>>>>>>> +			reg-names = "chgr", "bat_if", "usb", "misc";
>>>>>>> +
>>>>>>> +			interrupts = <0x0 0x10 0 IRQ_TYPE_EDGE_BOTH>,
>>>>>>> +				     <0x0 0x10 5 IRQ_TYPE_EDGE_BOTH>,
>>>>>>> +				     <0x0 0x10 6 IRQ_TYPE_EDGE_BOTH>,
>>>>>>> +				     <0x0 0x10 7 IRQ_TYPE_EDGE_BOTH>,
>>>>>>> +				     <0x0 0x12 0 IRQ_TYPE_EDGE_BOTH>,
>>>>>>> +				     <0x0 0x12 1 IRQ_TYPE_EDGE_BOTH>,
>>>>>>> +				     <0x0 0x13 0 IRQ_TYPE_EDGE_BOTH>,
>>>>>>> +				     <0x0 0x13 1 IRQ_TYPE_EDGE_BOTH>,
>>>>>>> +				     <0x0 0x13 2 IRQ_TYPE_EDGE_BOTH>,
>>>>>>> +				     <0x0 0x13 4 IRQ_TYPE_EDGE_BOTH>;
>>>>>>> +			interrupt-names = "vbat_det",
>>>>>>> +					  "fast_chg",
>>>>>>> +					  "chg_fail",
>>>>>>> +					  "chg_done",
>>>>>>> +					  "bat_pres",
>>>>>>> +					  "temp_ok",
>>>>>>> +					  "coarse_det",
>>>>>>> +					  "usb_vbus",
>>>>>> So, both the charger and the USBIN driver use the same irq? :/
>>>>>>
>>>>>
>>>>> AFAIU the usbin extcon driver pretty much just tracks the state
>>>>> of the IRQ to report extcon. It happens to assume the same part
>>>>> of the pmic though, yes, which also means there will be no user
>>>>> that would enable both charger and vbus extcon, since charger
>>>>> driver provides this functionality as well.
>>>> So, should USBIN be removed from PM8916 dt since it's essentially
>>>> a part of the charger block?
>>>>
>>>
>>> The "USB_IN" pad of the PM8916 seems to be connected on pretty much all
>>> devices, even if they are using external chargers and the charging
>>> functionality of PM8916 is completely disabled. For those devices, the
>>> &pm8916_usbin device provides a convenient way to detect the USB state,
>>> even without a working charger driver.
>>>
>>> While we could modify the PM8916 charger driver and DT node to have some
>>> special mode where charging and battery monitoring is completely
>>> disabled and only the USBIN extcon is provided, I'm not sure if that
>>> would provide a significant advantage compared to just keeping the
>>> simple &pm8916_usbin node with the existing driver.
>> Hmm okay I see..
>>
>> Generally it's rather "no bueno" to have two DT nodes consuming the
>> same register space.. What happens when you enable BMS on a device
>> with a non-PM8916 charger? Does it correctly recognize "no battery"
>> etc.?
>>
> 
> The _charger and _bms are separate and communicate in a generic
> manner via power-supplies and supply core (see 3/3) so giving
> a different charger to _bms can work.
> 
> If an external charger is present in the device, qcom mandates
> "external charger" optional line of the pmic to be tied, and
> _charger is then disabled. The driver bails out in this case,
> but _usbin could still be used.
Meh..

I guess I'll reluctantly let it slide, unless Bjorn has some objections

Konrad