Linux Kernel sucks for Android Devices

Context

This project is my try on making a old phone a Motorola G9 Plus have the latest Android (16 because 17 is not yet released by LineageOS). In the past had bootloop problems, phone wipes problems, almost bricking it a few thousand times.

Semi-Bricking fix

Opus 5, fixed this problem really easily it realized that the Motorola Bootloader Menu (MBM) didn’t like LineageOS standart bootctrl, so it built one by itself!

Another bootloop problem!

After the fix, we tried the OS for the first time and it was AWESOME! It restarted the phone, the Bootloader Unlocked Warning appered (good sign that the OS exists), them I got stuck on the Motorola Logo for ~20-30 seconds and it bootlooped…

That wasn’t a good thing but it meant that the “Semi-Bricking” was fixed and now we had some Kernel problem again…

Bootloop fix (8x tries…)

I just smashed Opus 5, GPT-5.6-Sol, and Kimi K3 (which Weights became avaliable today!) at the problem and they all got stuck trying to make the Logger work.

In so many words and tokens they tried ALL kinds of loggers, and after ~2 days of trying to make some kind of logger work, they found that the problem…

Culprit

It was the “kernel 4.14 vs Android 16 BPF”, which means that the Android 16 requires a newer kernel in order to have a more up-to-date BPF (Berkeley Packet Filter, aka a Subsystem that allows User code to be ran at the Kernel Level) to work.

Possible paths forward

1. Try to use OpenEL Assosiation (openela) Kernel LTS, which has a longer support for the 4.14.x kernel (which had the official End-Of-Life at January 2024, with an extension to December 2024)
2. Vibecode the shit out of the kernel to try to update to the 5.10.x LTS (which has End-Of-Life December of 2026) or a newer version like 6.1 LTS, 6.6 LTS, 6.12 LTS, 6.18 LTS.
3. Give up…

Well, what I am going to do is likely the 2nd option. But I need to find out if is going to be 5.10, or one of the 6.x versions. I am going to start with what other developers at LineageOS did. Its likely the best option