ZGC: Introduce ZConcurrentRootsIterator for scanning a subset of strong IN_NATIVE roots concurrently

We now have enough load barriers to support scanning of CLDs and JNI handles concurrently. Also, when ZWeakRoots is false, the vm weaks, string table and jni weaks can also be scanned concurrently. I propose to do that and move these root sets out from ZRootsIterator, and hence the GC pauses.