On Wed, Jul 11, 2018 at 06:46:50AM +0200, Theo Buehler wrote:
> Our current implementations of BN_{,u}{add,sub}() are plain disgusting.
> OpenSSL cleaned this up quite a bit recently -- still under
> '(the "License")' -- and now you can actually follow what is going on.
> There should be no performance impact from this change, the code is
> doing essentially the same thing, just in a more orderly way.
>
> For review, note that are two pairs of functions, BN_{add,sub}() and
> BN_{uadd,usub}() which have similar changes. I recommend just applying
> the diff and looking at the resulting file, the new functions are
> pretty obvious.
I wrapped by head around these in two stages,
your bit fiddling looks good to me.
OK kn
