From: Evgeniy Polyakov Subject: Re: [0/11] Add IV generators and givcrypt Date: Thu, 22 Nov 2007 14:25:59 +0300 Message-ID: <20071122112558.GE2444@2ka.mipt.ru> References: <20071122084758.GA7536@gondor.apana.org.au> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Linux Crypto Mailing List To: Herbert Xu Return-path: Received: from relay.2ka.mipt.ru ([194.85.82.65]:49443 "EHLO 2ka.mipt.ru" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751905AbXKVL0O (ORCPT ); Thu, 22 Nov 2007 06:26:14 -0500 Content-Disposition: inline In-Reply-To: <20071122084758.GA7536@gondor.apana.org.au> Sender: linux-crypto-owner@vger.kernel.org List-Id: linux-crypto.vger.kernel.org Hi Herbert. On Thu, Nov 22, 2007 at 04:47:58PM +0800, Herbert Xu (herbert@gondor.apana.org.au) wrote: > This series of patches introduces IV generators which can be > accessed through the givcrypt operation on block ciphers (other > than crypto_blkcipher which should not be used for new code). > > The givcrypt operation is identical to encrypt except that the > IV is generated by the algorithm instead of given by the user. > > Algorithms can either provide their own IV generator or (as > most will do) use the default IV generators provided by the > system. If no IV generator is specified by the system, then > chainiv will be used for synchronous algorithms and eseqiv will > be used for asynchronous algorithms. > > Counter mode algorithms (which include some stream ciphers) must > choose something other than the default to ensure uniqueness for > the IV. Idea and implementation look very good, I have couple of comments on patches and one generic comment here: you absolutely have to write at least bits of documentation for this new interfaces, how they behave and who and how should use it :) -- Evgeniy Polyakov