Return-Path: Received: from mail.kernel.org ([198.145.29.99]:44492 "EHLO mail.kernel.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726889AbeJWHAk (ORCPT ); Tue, 23 Oct 2018 03:00:40 -0400 Date: Mon, 22 Oct 2018 15:40:10 -0700 From: Eric Biggers To: Ard Biesheuvel Cc: "open list:HARDWARE RANDOM NUMBER GENERATOR CORE" , linux-fscrypt@vger.kernel.org, linux-arm-kernel , Linux Kernel Mailing List , Herbert Xu , Paul Crowley , Greg Kaiser , Michael Halcrow , "Jason A . Donenfeld" , Samuel Neves , Tomer Ashur Subject: Re: [RFC PATCH v2 09/12] crypto: nhpoly1305 - add NHPoly1305 support Message-ID: <20181022224008.GB59695@gmail.com> References: <20181015175424.97147-1-ebiggers@kernel.org> <20181015175424.97147-10-ebiggers@kernel.org> <20181020053834.GC876@sol.localdomain> <20181022184236.GA59695@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: Sender: linux-crypto-owner@vger.kernel.org List-ID: Hi Ard, On Mon, Oct 22, 2018 at 07:25:27PM -0300, Ard Biesheuvel wrote: > > > > Hmm, I'm actually leaning towards the following instead. Unrolling multiple > > strides to try to reduce loads of the keys doesn't seem worthwhile in the C > > implementation; for one, it bloats the code size a lot > > (412 => 2332 bytes on arm32). > > > > static void nh_generic(const u32 *key, const u8 *message, size_t message_len, > > __le64 hash[NH_NUM_PASSES]) > > { > > u64 sums[4] = { 0, 0, 0, 0 }; > > > > BUILD_BUG_ON(NH_PAIR_STRIDE != 2); > > BUILD_BUG_ON(NH_NUM_PASSES != 4); > > > > while (message_len) { > > u32 m0 = get_unaligned_le32(message + 0); > > u32 m1 = get_unaligned_le32(message + 4); > > u32 m2 = get_unaligned_le32(message + 8); > > u32 m3 = get_unaligned_le32(message + 12); > > > > sums[0] += (u64)(u32)(m0 + key[ 0]) * (u32)(m2 + key[ 2]); > > sums[1] += (u64)(u32)(m0 + key[ 4]) * (u32)(m2 + key[ 6]); > > sums[2] += (u64)(u32)(m0 + key[ 8]) * (u32)(m2 + key[10]); > > sums[3] += (u64)(u32)(m0 + key[12]) * (u32)(m2 + key[14]); > > sums[0] += (u64)(u32)(m1 + key[ 1]) * (u32)(m3 + key[ 3]); > > sums[1] += (u64)(u32)(m1 + key[ 5]) * (u32)(m3 + key[ 7]); > > sums[2] += (u64)(u32)(m1 + key[ 9]) * (u32)(m3 + key[11]); > > sums[3] += (u64)(u32)(m1 + key[13]) * (u32)(m3 + key[15]); > > Are these (u32) casts really necessary? All the addends are u32 types, > so I'd expect each (x + y) subexpression to have a u32 type already as > well. Or am I missing something? > The (u32) casts are only necessary when sizeof(int) > sizeof(u32), as then the addends will be promoted to 'int'. Of course, that's never the case for the Linux kernel. But I prefer it to be as robust and well-defined as possible, since people might use this as a reference when coding other implementations, which could end up finding their way into unusual and/or future platforms. - Eric