On Tue, 2015-01-27 at 19:58 +0100, Nikos Mavrogiannopoulos wrote:
I guess we need to get into release mode soon (I'll send another message to try to sort out loose ends), but it might be possible to to ocb. I've had a quick look at RFC7253. According to wikipedia (https://en.wikipedia.org/wiki/OCB_mode), that is "OCB2", is that the relevant version?
Has the ietf discussion clarified the patent issues?
I'm going to mail fsf lawyers about the patent license (http://web.cs.ucdavis.edu/~rogaway/ocb/license1.pdf), I suspect they're not compatible with the LGPL.
It is that one: https://tools.ietf.org/html/rfc7253
Let me know if you get some reply from FSF. In that case I'd recommend against standardizing OCB in the IETF TLS WG.
In the case it is not compatible with LGPL.
About the release... Since you added the fat, would it include AESNI +PCLMUL? If yes that would reduce significantly the assembly shipped in gnutls (only the padlock functions would remain).
regards, Nikos