I'm fine with LIO the code. The fact that it's in upstream now makes many integration issues easier, and lets us all work on it to improve it.
I've been working on improving the out-of-box user experience for the LIO user tools as well. With my enhanced version of targetcli (in Fedora 17+, repos on github) configuring a scsi target can be pretty painless. I've done some Youtube screencasts on it, and added a manpage for targetcli.
At this point, I'm definitely not in favor of dumping LIO, but we need to get these license and maintainer potential conflict-of-interest issues straightened out.