AOSP E2E: Internals, Security, Optimization - Building ROM & Automotive
In this talk, I’ll take you on a journey through the depths of the Android Open Source Project (AOSP) — from creating a custom AOSP extension for automotive devices, to reviving an old smartphone far beyond its official support.
I’ll start with how I built a custom AOSP repo that can turn virtually any device into an Android Automotive system with just a few additions to the build. This modular setup makes it easy for developers and hobbyists to experiment with automotive-grade AOSP without needing expensive setups.
Next, I’ll dive into my journey with the Xiaomi Mi A2 (SDM660), a device launched in 2018 and abandoned in 2020. I’ll share how I kept it alive through Android 15 (and potentially 16), improving its performance through deep debugging, optimization, and smart usage of Android features like Android Go configurations. This part includes real-world debugging stories, performance tuning tips, and practical AOSP development strategies.
We’ll also explore general AOSP security features, why they're often overlooked, and how enabling them — alongside lesser-known optimizations — can result in smoother, more secure Android experiences, even on older hardware.
How to use a custom AOSP repository to enable Android Automotive functionality on almost any device.
The process of extending the life of an end-of-support device like the Mi A2 to Android 15+.
Performance tuning and debugging techniques learned from working on real-world legacy hardware.
Hidden AOSP features and flags (like Android Go configs) that can improve efficiency on low-end devices.
Insights into contributing to custom ROMs and the AOSP community.
Practical security enhancements you can apply in AOSP builds.
Which track are you applying for?
Immanuel Raj
The talk presents a custom Android Open Source Project (AOSP) contributed by the author. It is an important scenario considering the lack of support of the older mobile devices. This issue helps reduce the electronic waste significantly by reusing the devices that have lost support. While it will be interesting to understand how does the device security function, or the device will be exclusively used for custom purposes instead of normal cellular functions. It is an important talk considering a consideration for the environment, but lack of security hardening may create issues later. The author may address these points as well.
As we’re reviewing the program, we’ve noticed that several proposals cover similar foundational topics, such as AOSP porting/bring-up and Custom ROM basics. So we have asked proposer if he would be opento the idea of presenting his session as a 15min (10+5) lightning talk rather than a full-length 25min (20+5) session.
Rather than talking about AOSP or custom ROMs in general, we have suggested him to focus on the technical details of recycling old/neglected phones as Android Auto systems. Highlight its modular framework or what makes his project stand out in the landscape of Custom ROMs, and more importantly the value it adds to the ecosystem.
Approved as a lightning talk.
We requested the submitter if they could convert to a lightning talk (10+5 minutes), focussing on efforts around phones recycled / enabled as Android Auto systems that could also serve as dev platforms.
We're excited to accept it as a lightning talk.