From: Eric Biggers Subject: Re: [PATCH v5 0/4] gf128mul refactoring Date: Tue, 4 Apr 2017 21:13:51 -0700 Message-ID: <20170405041351.GA4761@zzz> References: <20170402191916.9309-1-omosnacek@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Herbert Xu , "David S. Miller" , linux-crypto@vger.kernel.org, Eric Biggers , Jeffrey Walton , Milan Broz To: Ondrej Mosnacek Return-path: Received: from mail-pf0-f195.google.com ([209.85.192.195]:33202 "EHLO mail-pf0-f195.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750712AbdDEENy (ORCPT ); Wed, 5 Apr 2017 00:13:54 -0400 Received: by mail-pf0-f195.google.com with SMTP id 197so335721pfv.0 for ; Tue, 04 Apr 2017 21:13:54 -0700 (PDT) Content-Disposition: inline In-Reply-To: <20170402191916.9309-1-omosnacek@gmail.com> Sender: linux-crypto-owner@vger.kernel.org List-ID: Hi Ondrej, On Sun, Apr 02, 2017 at 09:19:12PM +0200, Ondrej Mosnacek wrote: > This patchset contains the following gf128mul-related changes: > 1. The gf128mul_x_* functions are moved to gf128mul.h for performance reasons. > 2. The gf128mul_x_ble function is fixed to use the correct block type. > 3. The le128_gf128mul_x_ble function from glue_helper.h is removed and its > usages replaced with gf128mul_x_ble calls. > 4. The now obsolete dependency of CRYPTO_XTS on CRYPTO_GF128MUL is removed. > > v4 -> v5: added the other three patches > v3 -> v4: a faster version of gf128mul_x_lle > v2 -> v3: constant-time implementation > v1 -> v2: move all _x_ functions to the header, not just gf128mul_x_ble > > Ondrej Mosnacek (4): > crypto: gf128mul - define gf128mul_x_* in gf128mul.h > crypto: gf128mul - switch gf128mul_x_ble to le128 > crypto: glue_helper - remove the le128_gf128mul_x_ble function > crypto: xts - drop gf128mul dependency These all look good to me, and you can add Reviewed-by: Eric Biggers to the patches. I think the change to le128 is an improvement, but what I've actually wanted to do is have the GF(2^128) elements strongly typed, with ble128, bbe128, and lle128 types, so that the types would reflect both the bit order and the byte order. But that's harder to do and I ran into some issues when I tried it last. - Eric