We have two libchrome mirrors, one in ChromeOS and the other in Android.
Currently, Android's one is upstream, and ChromeOS's follows.

I'd propose to swap the upstream, when ChromeOS's libchrome catches up with Android's. Mainly because;
- ChromeOS release cycle is shorter than Android's. So fresher libchrome is more important for ChromeOS, than Android.
- ChromeOS's source code is closer to Chrome, than Android, so uprev'ing in ChromeOS should be relatively easier. However, currently Android's is upstream, so even if we could easily uprev in ChromeOS, it would be delayed by Android's.

Swapping the upstream resolves them. Also, then, we may be able to set up libchrome-PFQ-ish thing, to update almost automatically. (If there is a conflict or stable API changes, indeed, manual work is still needed, but according to some manual uprev experience, it would be not so frequently, I think).

Diverging two libraries would increase the total maintenance cost, I think, because anyways those need to be updated periodically.
E.g., ARC uses libchrome/libbase in Chrome/ChromeOS/Android, and managing multiple versions is just tough.