I'm not sure there's actually a limitation in the current approach - I think it should work at least as well as the current OpenISCSI target code. But there's some code messiness in the handling of the iSCSI target IP; happily the future direction of the SAN code is likely to evolve to fix this anyway. Not sure if we should do a bug/blueprint in this case (or what would really go in it if we did!)
I'm not sure there's actually a limitation in the current approach - I think it should work at least as well as the current OpenISCSI target code. But there's some code messiness in the handling of the iSCSI target IP; happily the future direction of the SAN code is likely to evolve to fix this anyway. Not sure if we should do a bug/blueprint in this case (or what would really go in it if we did!)