On Fri, 2012-03-16 at 14:34 +0100, Joel Reardon wrote:
> I actually just copied this comment from the comment following padding1
> and padding2; should they all just be omitted?

Ah, ok, ignore this comment then please.

> As for the ubifs being mounted with the old, it may be best to increase
> the version format number. The old version won't be able to 'read' (i.e.,
> decrypt) the data, while the new version has a switch to enable both
> modes. If new data is written by the old version then the new version will
> also have trouble to read it (unless we set crypto_lookup==0 to mean no
> key). But its probably for the best to just let older version mount the
> security enhanced one as read-only using the version format as the data
> will be anyhow unreadable. Non-security-enhanced ubifs (but
> aware) partitions can set the version format to the older value as they
> will be compatible.

OK. Then I guess the version increment patch should also be separate. In
general, try to make logically independent things separately and keep
patches small.

-- 
Best Regards,
Artem Bityutskiy