[BUG] Infinite boot loop after charging control

scruch

Today , I encountered a wwird bug with GrapheneOS on my Pixel 6:
After leaving the phone plugged in for several days, GrapheneOS seems to have activated a battery protection feature that limits charging (presumably to extend battery health). All good so far.
However, after unplugging it and leaving it for a few days, the phone fully discharged. When I plugged it back in, it entered an infinite boot loop:

Bootloader starts the OS
A few seconds later the phone shuts down
Then it tries to reboot, and again shuts down...
After digging a bit, it looks like the Pixel firmware does charge the phone initially, but once GrapheneOS starts booting, the battery management kicks in and blocks charging, likely due to the previously activated charging protection.
This results in a loop where the phone never gets enough charge to complete boot and exit the protection mode. Also the risk of file corruption from power loss , especially sensible i think because encrypted.

⚠️ Recommendation (I'm not a dev): Consider resetting the charging protection flag before boot or only activating it when the battery is above a certain threshold (e.g. 10%). Otherwise, phones with not super new batteries like my Pixel 6 may never boot again without workarounds.

✅ Workaround that helped me: I let the phone sit plugged blocking the booting of Graphene.. eventually it got enough charge during those short firmware-powered moments to boot fully.
Just wanted to report in case it helps

de0u

scruch It is best in general to avoid letting the battery run flat, because it can damage the battery. But it can happen sometimes!

On the boot screen, when it suggests tapping the power button to pause booting, tap the power button and let the device charge for a while. Depending on the charger you might need only 10 minutes.

scruch

de0u thank you , yes this is what i have done. But if I was not manually doing it , the phone would have rebooted in loop for the next month.. until probably FS or HW got corrupted. (compared to this the damage of a complete drain battery is nothing) . Probably is a bug that need to be fixed .

de0u

scruch Thank you , yes this is what i have done. But if I was not manually doing it , the phone would have rebooted in loop for the next month.. until probably FS or HW got corrupted.

I think for that to play out these steps would be required:

The battery runs flat,
The user plugs in a charger,
3a. The user goes away for a month without seeing if the system boots, or
3b. The user sees the system rebooting in a loop but does not take advantage of the "Press power button to pause" message and also doesn't ask anybody for advice for a month.

scruch Probably is a bug that need to be fixed .

I don't think it would be easy for the bootloader to predict how much power is needed for the early stages of booting the OS. It would not be easy for the very early stages of the OS to enable charging. And presumably somebody decided it was unwise or infeasible to leave charging enabled when the bootloader hands control off to the OS.

I suspect that the GrapheneOS team would consider a code change if one were submitted, but I would be surprised if there were developer time available to work on this any time soon.

Does Google's stock OS avoid this phenomenon?

schweizer

Duplicate of https://discuss.grapheneos.org/d/13422-grapheneos-wont-start-after-dead-battery