Side channel silent karatsuba / mpn_addmul_2 karatsuba

[Torbjörn Granlund]

"Marco Bodrato" <bodrato at mail.dm.unipi.it> writes:

  Yes, in that case ((g >> 32) + e1h)>>32 will be non-zero, so that the
  expression ((g >> 32) + e1h + (e1hc << 32)) will give at most a 33 bit
  result. Right?

I don't understand my code that well.  :-)

  >   ulong e1ha =-((a + b) >> 32)          & ((c + d) & 0xfffffffful);
  >   ulong e1hb = ((a + b) & 0xfffffffful) &-((c + d) >> 32);
  >   ulong e1hc = ((a + b) >> 32)          & ((c + d) >> 32);
  >   ulong e1h = e1ha + e1hb;

  when moing from half- to full-word, the carry of e1h will be added to e1hc

Yes, they will be \beta apart.

  and here e1hc will get also the carry from (g >> 32) + e1h ... after that
  the value of e1hc will not be "negative" any more :-)

I got stuck with that code since there are so many ways of summing
things.

-- 
Torbjörn
Please encrypt, key id 0xC8601622