On 11/22/25 12:07 PM, Maud Spierings wrote:
> On 11/20/25 16:36, Konrad Dybcio wrote:
>> On 11/17/25 4:44 PM, Maud Spierings wrote:
>>> On 11/17/25 16:35, Konrad Dybcio wrote:
>>>> On 11/17/25 3:13 PM, Maud Spierings wrote:
>>>>> Hi Konrad,
>>>>>
>>>>> On 11/17/25 13:59, Konrad Dybcio wrote:
>>>>>> On 11/16/25 11:52 AM, Maud Spierings via B4 Relay wrote:
>>>>>>> From: Maud Spierings <[email protected]>
>>>>>>>
>>>>>>> Add nvmem cells for getting charge control thresholds if they have
>>>>>>> been set previously.
>>>>>>>
>>>>>>> Signed-off-by: Maud Spierings <[email protected]>
>>>>>>> ---
>>>>>> Have you verified that e.g.
>>>>>>
>>>>>> connecting the charger
>>>>>> setting the charge threshold
>>>>>> rebooting to windows
>>>>>> rebooting to windows once more for good measure
>>>>>> rebooting to linux
>>>>>>
>>>>>> still has the settings persist?
>>>>> Hmm I have tried several things but I can't seem to get the values to
>>>>> stick. I the spmi-sdam driver is compiled in, I am not quite sure if I
>>>>> might be missing something.
>>>> Hm, I wonder if Windows/UEFI overwrites these values or whether they're
>>>> used by something else..
>>>>
>>>> Can you set a threshold in windows and see if Linux can read back that
>>>> data?
>>> the values in /sys/class/power_supply/jada-jada/ are zero when rebooting
>>> from Windows into Linux after enabling charge limitting in the Asus
>>> application.
>>>
>>> I remember my old vivobook (x86) also forgot its settings each boot, but
>>> given the nvmem cells that should not be happing here I guess. It is odd
>>> that there seems to be no collision between Windows and Linux. Maybe the
>>> Windows mechanism is doing the old trick of writing it in there every boot?
>> Odd indeed.. Does it work if you reboot from Linux to Linux?
> It seems not, I seem to remember testing it quite some time ago, but I cannot
> get it to remember any way, at least it is not popping up in sysfs, always
> back to 0
It seems like the driver currently only populates the sysfs start/stop
values if the "enable" bit is set
Could you check this (hacky and wrong) diff and give it another try?
diff --git a/drivers/power/supply/qcom_battmgr.c
b/drivers/power/supply/qcom_battmgr.c
index c8028606bba0..9ebd8adfb8eb 100644
--- a/drivers/power/supply/qcom_battmgr.c
+++ b/drivers/power/supply/qcom_battmgr.c
@@ -733,7 +733,7 @@ static int
qcom_battmgr_charge_control_thresholds_init(struct qcom_battmgr *batt
u8 en, end_soc, start_soc, delta_soc;
ret = nvmem_cell_read_u8(battmgr->dev->parent, "charge_limit_en", &en);
- if (!ret && en != 0) {
+ if (!ret) {
ret = nvmem_cell_read_u8(battmgr->dev->parent,
"charge_limit_end", &end_soc);
if (ret < 0)
return ret;
Konrad
