The on and off values of the charging switch cannot contain spaces about acc HOT 16 CLOSED

Wow that's so cool! Manufacturers sure know how to make things interesting 😂

from acc.

Just to make things more fun I found out that if current_cmd is toggled on while the device is not plugged in, the battery will report "charging". Yes really -_-

from acc.

A while ago, I came across /proc/mtk_battery_cmd/current_cmd "0 0" "0 1". Is that your switch as well?

ACC is undergoing major changes. As part of these, several config variables are now arrays (e.g., chargingSwitch=(/proc/mtk_battery_cmd/current_cmd "0 0" "0 1")).

If you want to try experimental builds, message me on https://t.me/vr25xda (preferred), https://fb.me/vr25xda or https://m.me/vr25xda .

Thanks!

from acc.

Yep, that's the switch (btw it's a Blackview 9600 Pro).
One unrelated thing I've noticed is that toggling it stops charging but keeps the battery at that level, while to make it actually discharge I need to toggle en_power_path as well.
As for trying the experimental builds, thanks a lot for the offer, but unfortunately I cannot tinker much with this device as it is the primary phone of a friend :(

from acc.

I can't get enough of Mediatek's "nice" surprises. Holly nuts!

Could you elaborate more on en_power_path? I can make acc toggle it.

from acc.

So /proc/mtk_battery_cmd/en_power_path is normally 1, when set to 0 it stops charging and also begins discharging but there is a catch: the battery status always remains "charging" , so acc doesn't work.
To get the complete behaviour (battery stops charging, starts discharging and reports the correct status) one has to toggle both en_power_path and current_cmd.

from acc.

Interesting and Intriguing!
This will give birth to a new feature: toggling multiple switches at once.

from acc.

Terrible Design!

Lets say the device is plugged in and charging. After charging is disabled (i.e., through toggling the 2 switches), is it enabled automatically on replug, without toggling the switches back?

Were you able to find any file that reports the correct external power supply status (i.e., plugged/unplugged)?
Still on this question, does su -c acpi -a do that?

from acc.

Sorry for the delay, I managed to get some time with the phone to run the tests. The results are braindead haha

• Once one touches current_cmd it becomes fully manual, ie the phone won't charge when plugged in unless the switch is toggled
• acpi -a reports two adapters, "Adapter 0" seems to reliably indicate whether the phone is plugged in while "Adapter 1" is always "Off-line"

from acc.

Finally, at least one pleasant surprise... hopefully! 😁

If acpi -a really reports the correct external power supply status (plugged/unplugged), regardless of whether switches were toggled, we have a winner.

From acpi -a and battery/status's data combination, we have three basic battery statuses:

• Plugged in and charging (at least one adapter online AND battery/status == "charging")
• Plugged in, but not charging (at least one adapter online AND battery/status == "discharging|not charging")
• Unplugged and discharging (all adapters offline)

I just finished implementing and testing the "multiple switches toggling" feature. It's ready for third-party testing. I'm in touch with another MTK user - who will be testing the real deal.

from acc.

That's great! So to ensure that the braindead status "unplugged but charging" never happens I guess acc will have to toggle the switches when it detects an unplug haha

from acc.

Thou Shalt Go Mad.
MediaTek Inc.

from acc.

I implemented everything we discussed so far.
Even though this is a pre-release, the daemon is pretty much stable at this point.

https://www.dropbox.com/s/cdfok5fek2g5z5f/acc-202001180.zip?dl=0

Notes

• Don't use acc -s|--set for editing the config. It's not ready yet. Use acc -c|--config EDITOR OPTS instead (e.g., acc -c vim). If you run just acc -c, the fallback editors are nano -l > vim > vi. You can always edit /data/adb/acc-data/config.txt with any other means.

• Apparently, MTK requires a [slightly] longer delay after toggling the switches - for changes to take effect. ACC has the config variable switchDelay (default: 1 second). You may want to increase that before performing any testing - and play with it along the way. This is because, in auto mode (chargingSwitch unset), switch_cycle() is in charge. After switchDelay, it resets switches which don't work. That's where the default delay can be an issue. switchDelay=10 is a good starting point.

• If you don't want to play with switchDelay right now, set chargingSwitch=(/proc/mtk_battery_cmd/current_cmd "0 0" "0 1" /proc/mtk_battery_cmd/en_power_path 1 0) and test whether acc -d disables charging - and acc -e re-enables it. This is the manual/daemon-less mode.

• The documentation is not fully updated yet. At this point, besides switchDelay and chargingSwitch[@], the only other values you have to care about are capacity[2] (resume, default: 75) and capacity[3] (stop, default: 80).

from acc.

Wow that's great! I can't wait to try it, as soon as I see my friend I'll try and report back

from acc.

So I finally got the chance to test (sorry for the delay!), but I'm running into a weird problem: I flashed the zip with magisk but apparently it gets installed into /sbin/.magisk/modules/acc/ but not into /sbin/.acc/, and this prevents it from starting... if i try manually I get like

Blackview:/ # /sbin/.magisk/modules/acc/acc.sh                                                                                                                   

/sbin/.magisk/modules/acc/acc.sh[466]: .: /sbin/.acc/acc/setup-busybox.sh: No such file or directory

I'm probably missing something obvious sorry

from acc.

@kbios, the repo is up-to-date.
Dowload the latest version: https://GitHub.com/VR-25/ACC/archive/dev.zip .
Extract the zip and run the following to install:
$ su
# sh /path/to/extracted/install-current.sh

from acc.