1. Same IP, different hashes;

Yes, the same IP could emit events that get handled by different instances of the now-parallel processor.  This means that events with the same IP could have different hashes.
 
2. Different IPs, same hash;

This was always possible in the old system due to hash collisions.  But, like, very very unlikely.  And this hasn't been influenced by the new bug in any way I'm aware of.  Though maybe some number theoretician somewhere just did a telepathic spit take.