UBIFS: ubifs_recover_leb() always execute fix_unclean_leb() even all the nodes are valid

[Zhihao Cheng]

Hi, 仁杰
> Sorry for the late reply. I saw the comment in the commit, but I don’t
> understand the meaning of unstable bits very well. Does it mean that
> unstable bits may appear in the area after the current offset and cause
> subsequent write operations to fail? But when I copy the original data
> to another PEB, there should be the same probability that unstable bits
> will appear?
I meant unstable bits may look good in first recovery process, but they 
might be changed after running for a while.
> I have a guess. Is it possible that the data is copied to another PEB
> because there may be more useless data in the current PEB in order
> to recover the entire PEB?
> 
Last written lebs are processed by ubifs_recover_leb(), some of them 
carry much usefull data, like GCHD and DATAHD(for creation nodes). In my 
opinion, fix_unclean_leb() is used to handle hardware error.