Last month, Punya described his stance on why he feels updating Android kernels to the latest minor Linux version (upstreaming) is important and asserted that, upstreaming is important and while you could merge individual patches, it takes significantly less effort and time to merge the entire branch.
Initially, I felt the same way: that upstreaming was absolutely necessary to keep the kernel up-to-date with the latest patches, so I ended up doing so on several based kernels I maintained at the time. However, now when I look back, I feel it was not worth the time.
If your kernel has a maintained upstream repository in addition to Linux and the Android Common Kernel (such as Code Aurora Forum for devices with Qualcomm Snapdragon processors), then upstreaming from Linux can cause issues and will also cause merge conflicts, duplicate code, and break drivers when pulling from the other upstream. In addition, it will increase the delta from the vendor upstream code (such as CAF) to your tree, which can lead to merge conflicts and loss of functionality and stability.
Often, many of the updates in minor Linux kernel upgrades are specific to x86 and desktop environments, so the only difference in user experience is the larger number in Kernel Version under About Phone in the best case scenario, while in the worst case scenario, they will encounter issues.
Rather than simply merging Linux kernel updates in order to have a newer version number for marketing purposes, it is better to be more thoughtful about what you change in your kernel. You can still manually add any patches from Linux, Google, Qualcomm, or any other source that you feel will improve your device, which minimizes risk and makes it easier to evaluate and understand the additions to your kernel.
Concluding, let us know your stance when it comes to Andorid Kernel upstreaming through the comments!
Want to have your story published? All you have to do is head over to the Submit A Post page !