Added proper failhandling to nand.cpp . Amap gets too long packet, probably because of different clock (artix 50Mhz, Nexys 60Mhz)?

Fixed interesting condition where setTargetInode was called on an Inode that does not exist in Tree anymore.
This would lead to a lazy-load of this Inode, which may be illegal during journal replay if the cache is full.

Fixed le 'ol bug (findNextBestArea did not state that area is empty if it found an area with valid contents before