[PATCH 15/25] EAX: provide an implementation of EAX

[Ian Jackson]

Mark Wooding writes ("Re: [PATCH 15/25] EAX: provide an implementation of EAX"):
> Ian Jackson <ijackson at chiark.greenend.org.uk> wrote:
> > Then for completeness we also provide a set of EAX-Serpent test
> > vectors and the corresponding test code.  The EAX-Serpent test vectors
> > were generated by this very code, so aren't independently verified.
> 
> I've understood how my Serpent implementation differs from Secnet's, and
> have reproduced your test vectors.
> [stuff]

Thanks for the research.  This is now a bit of a mess, which I will
deal with in a bit.

> > +eax-%-test.confirm: eax-%-test eax-%-test.vectors
> > +	./$< <eax-$*-test.vectors >$@.new
> > +	mv -f $@.new $@
> 
> This should read
> 	./$< <$(srcdir)/eax-$*-test.vectors >$@.new
> otherwise VPATH builds fail.

Fixed.

> Annoyingly, `xor' is a reserved word in C++, and a macro defined in
> <iso646.h>.  I'd recommend a different name, like `block_xor' or
> something.

Fixed.

> > +static void alg_omac_t_k(INFO, uint8_t *mac_out, uint8_t t,
> > +			 const uint8_t *m, size_t m_len)
...
> This can be made less fiddly, I think, by splitting `cbc_iter' in two
> and delaying the `BLOCK_ENCRYPT' application.  Something like this
> (tested).  This subsumes the `cbc_init' and `cbc_iter' functions, so I
> removed them.

Thanks, I like your version a lot better.  I've incorporated it.
(I've kept your S-o-b from your previous contribution to this patch.)

> Also, this file has lines containing trailing whitespace and a mix of
> tabs and spaces for indentation, sometimes on the same line.  (My Emacs
> goes into `angry-fruit-salad-mode'.)

I don't think I would normally care abuut this in secnet.  But this
particular file is probably useful elsewhere, so to make that slightly
less awkward I've untabified it and removed the trailing spaces.

Thanks
Ian.