14:59 <calebccff> Hi, does anyone know the current state of battery fuel gauge / charger support on C630?

15:17 <bamse> calebccff: some-battery is just one of the features exposed by the EC, another one is plug-detect for usb-c/displayport...so what remains is to figure out how the type-c mux really works and then polish up the whole thing

15:18 <bamse> calebccff: and come up with a name for it

15:20 <bamse> calebccff: i made an attempt to teach drm/msm/dp about type-c hpd a while ago, which was shot down...and have been reworking that...so once we have that side in place it should be trivial to test the ec-driver-side of the puzzle

15:21 <bamse> calebccff: the one thing that worries me is that in the GPU node in acpi there's a reference to an i2c bus...which could imply that there's something like an SBU mux that we need to drive as well (per how other boards are designed)

16:29 <calebccff> bamse: huh interesting, so it doesn't use the smb2/fuel gauge exposed by pmi8998?

16:30 <bamse> calebccff: no, battery management is entirely hidden by the EC

16:33 <bamse> calebccff: i suspect that pmi8998 doesn't support the type of batteries used in the laptop, because in flex 5g they use a pair of smb pmics again, but at least Lenovo follows the qualcomm reference design there

17:45 <calebccff> huh interesting, thanks.

19:57 <steev> calebccff: you gonna work on it? :D

20:06 <calebccff> steev: I'm working on pmi8998 fuel gauge and smb2 charger support for the oneplus 6 and some other devices, unfortunately i don't own a c630

20:06 <steev> ah, yeah i saw those patches the other day

20:06 <calebccff> i was hopeful that we'd have some common goals and that would maybe make my life a bit easier aha

20:07 <calebccff> oh? they're a bit off a mess right now as I'm desperately trying to figure out why the op6 won't draw more than 500ma no matter what where the other devices that have been tested will happily draw much more

20:09 <calebccff> the downstream regmaps are HUUUUGE and zero docs make it quite tough to figure out what's what. especially fun when the driver doesn't use a bunch of the registers

20:20 <steev> just because i saw them doesn't mean i understand or can help, sadly

20:20 <steev> i'd had the same thought re c630 battery

20:22 <macc24> calebccff: what's the source for vendor's battery driver?

20:22 <bamse> calebccff: i would expect that there's some logic in the downstream kernel to adjust the current limit in the charger based on usb state

20:23 <bamse> calebccff: analog to the SMBB_USB_IMAX register in smbb

20:36 <calebccff> macc24: https://github.com/LineageOS/android_kernel_oneplus_sdm845/tree/lineage-18.1/drivers/power/supply/qcom files "qpnp-smb2.c", "qpnp-fg-gen3.c" and "smb-lib.c" are most relevant, this fork also contains a bunch of oneplus specific hacks on top

20:38 <calebccff> bamse: yeah there are *several* current limit registers, which seem to behave correctly on some other devices but simply don't on the oneplus 6... It's a special case as oneplus made a lot of changes to support their proprietary fast charge implementation

20:39 <calebccff> they have a whole STM8 which may or may not have a say in it somehow, as well as a bq27xxx fuel gauge which seems to live on the battery iteslf https://wiki.postmarketos.org/wiki/OnePlus_6_(oneplus-enchilada)/Hacking#UART_Muxing

20:39 <macc24> calebccff: i've seen worse.

20:40 <calebccff> macc24: honestly i don't want to know XD

20:42 <calebccff> The MCU is hooked up to VBAT_AD, VBUS_ADC, V_ADC (?), and 3 temperature probes, it talks to the SoC via some bitbang SPI for some reason, I'm wary to turn it on

20:43 <macc24> calebccff: check this out https://github.com/LineageOS/android_kernel_oneplus_sdm845/blob/lineage-18.1/drivers/power/supply/qcom/smb-lib.c#L3292

20:44 <macc24> and this https://github.com/LineageOS/android_kernel_oneplus_sdm845/blob/lineage-18.1/drivers/power/supply/qcom/smb-lib.c#L3149

20:44 <macc24> as i don't have the device myself, that's all i can do. you're on your own. good luck :)

20:45 <calebccff> cheers