From patchwork Mon Jul 3 10:29:19 2017 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Ard Biesheuvel X-Patchwork-Id: 106892 Delivered-To: patch@linaro.org Received: by 10.140.101.44 with SMTP id t41csp5606006qge; Mon, 3 Jul 2017 03:29:30 -0700 (PDT) X-Received: by 10.99.154.18 with SMTP id o18mr9495335pge.251.1499077770755; Mon, 03 Jul 2017 03:29:30 -0700 (PDT) ARC-Seal: i=1; a=rsa-sha256; t=1499077770; cv=none; d=google.com; s=arc-20160816; b=nBedahU4aX6PC6Hpv/K5mJ7htIi8Hk5tHrwxpmTgWAoBUH2RPY98WuTLIY+EWNKOkf 1rXEIYid/S2cBkrEro+Il7k08FC7/6LtaFMGhmcNgmdS+6oOovV6fmzwAfi0zpEPk2fb JCUAif6DT04JPFIM3XXxS8wvThvcv8zazG5Sh+LiLEPex07T0kZl5YrHBwI9oIWWGbQ2 abUEFGzJj+7cWMj0dsbXi4VQ9xleLcJxeUjiRrGFaa9zgU6Ht8cRUg8uBgkPKmD+9/LR HKrJ0KrezWY93U0cVANri05PE4LD6d1KRZiCJ0vZSDp+vHQlP88n65LayHfFL80HCWs+ hGng== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=arc-20160816; h=list-id:precedence:sender:message-id:date:subject:cc:to:from :dkim-signature:arc-authentication-results; bh=/0aXztWQJ9j79itOf+OzNf/ZpzoJAI4B632OsMyj3gM=; b=IVM0+094oJvRwVTD57ONHlInb7OK6DRIF6rdmXo0Tp31xIYL82SvBoQJ87NdaoRxzI Db1mIoxlCQpHJORLnzicroLchx2SBD6nL7VN9oANSkzDaYzvhCdiXhdq6C/W9k+fhiSD Ff5Uks8PSOBoIY6a5QhpouC0aTnEPV4pcjf5AhWn9gm1Gw03Ze2xyvqUY5YrU1xGPhT8 gooUbaqPqXm+EOwL4qwPdAeTfwgbPV1oS7ZUCoxCXMVbSKKqrpbNFe/qXX96exthAco2 k3jrNaftUgjel0cET8MpVDIRBLYAEfRSTDO9bSPJUAQ8gEAv74NFAxVsyn03fMXR3ERY ajvQ== ARC-Authentication-Results: i=1; mx.google.com; dkim=pass header.i=@linaro.org header.b=KJwTugIE; spf=pass (google.com: best guess record for domain of linux-crypto-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) smtp.mailfrom=linux-crypto-owner@vger.kernel.org; dmarc=pass (p=NONE sp=NONE dis=NONE) header.from=linaro.org Return-Path: Received: from vger.kernel.org (vger.kernel.org. [209.132.180.67]) by mx.google.com with ESMTP id r3si12711686plb.108.2017.07.03.03.29.30; Mon, 03 Jul 2017 03:29:30 -0700 (PDT) Received-SPF: pass (google.com: best guess record for domain of linux-crypto-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67; Authentication-Results: mx.google.com; dkim=pass header.i=@linaro.org header.b=KJwTugIE; spf=pass (google.com: best guess record for domain of linux-crypto-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) smtp.mailfrom=linux-crypto-owner@vger.kernel.org; dmarc=pass (p=NONE sp=NONE dis=NONE) header.from=linaro.org Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753319AbdGCK33 (ORCPT + 1 other); Mon, 3 Jul 2017 06:29:29 -0400 Received: from mail-wm0-f47.google.com ([74.125.82.47]:36981 "EHLO mail-wm0-f47.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753018AbdGCK32 (ORCPT ); Mon, 3 Jul 2017 06:29:28 -0400 Received: by mail-wm0-f47.google.com with SMTP id i127so103606668wma.0 for ; Mon, 03 Jul 2017 03:29:27 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linaro.org; s=google; h=from:to:cc:subject:date:message-id; bh=/0aXztWQJ9j79itOf+OzNf/ZpzoJAI4B632OsMyj3gM=; b=KJwTugIEb1/j09pQtmqDVUf4O/P++cu8WM8uQP98Hj5XKSwLCrb8yoArT0rp9d3Y6a nM0Bc4gredz4Fck58qHg7iXuJMZtCZFht4xagvNLAziDFsZIEccUs1wT+5k8w24RymOo Kte7mAdOSdW06wbH0nT6w3Hb+Xvi+qPuaPxIQ= X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:date:message-id; bh=/0aXztWQJ9j79itOf+OzNf/ZpzoJAI4B632OsMyj3gM=; b=mDtfD61w25S8IwHccgEvOH7wDVeg5GIIX+4w5yA9YXwdPLTFUM0RU8qbhoC7XUsN3G iJXQhNDqJehC2ezXxf6sXh9V1GMe+SC8LyHl3FMAxJ+ZIXWuQOl2uG3ANU14GdL2tOW1 IXFLpqlg9cCGvffM9nylEvVl9GZ9wR8W3UZFKUKoWG2nLO+EColhfaBPh8KdZlTcSK4+ vv6brjOst3RjYCCWGJzX0bNAU5lJmo/jeVVURiulxEdWVbxQmOdrq2hLfkot68aUOEnq KgMvH9EjTkf5rrNNpAyRUjddDxZ7Q7aMoudobJsW7w7piXu1llnK9WZaoKQ3IA1pqabc tXvg== X-Gm-Message-State: AKS2vOyX9WgwITJ4KxMc5dTj66MK+jKnwTEVDbxuZiThH/ymJLuLjb3z jwqOWkCjmoYL2Br7b0jOZA== X-Received: by 10.28.167.205 with SMTP id q196mr24036284wme.54.1499077766613; Mon, 03 Jul 2017 03:29:26 -0700 (PDT) Received: from localhost.localdomain ([154.151.196.199]) by smtp.gmail.com with ESMTPSA id v102sm2600484wrc.53.2017.07.03.03.29.24 (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Mon, 03 Jul 2017 03:29:25 -0700 (PDT) From: Ard Biesheuvel To: linux-crypto@vger.kernel.org, linux-arm-kernel@lists.infradead.org Cc: herbert@gondor.apana.org.au, will.deacon@arm.com, catalin.marinas@arm.com, Ard Biesheuvel Subject: [RFT PATCH] crypto: arm64/ghash - add NEON accelerated fallback for 64-bit PMULL Date: Mon, 3 Jul 2017 11:29:19 +0100 Message-Id: <20170703102919.21714-1-ard.biesheuvel@linaro.org> X-Mailer: git-send-email 2.9.3 Sender: linux-crypto-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-crypto@vger.kernel.org Implement a NEON fallback for systems that do support NEON but have no support for the optional 64x64->128 polynomial multiplication instruction that is part of the ARMv8 Crypto Extensions. It is based on the paper "Fast Software Polynomial Multiplication on ARM Processors Using the NEON Engine" by Danilo Camara, Conrado Gouvea, Julio Lopez and Ricardo Dahab (https://hal.inria.fr/hal-01506572) On a low-end core such as the Cortex-A53 found in the Raspberry Pi3, the NEON based implementation is ~2.8x faster than the table based one, and is time invariant as well, making it less vulnerable to timing attacks. When combined with the bit-sliced NEON implementation of AES-CTR, the AES-GCM performance increases by 75% (from 58 to 33 cycles per byte). Signed-off-by: Ard Biesheuvel --- Note that this is the arm64 counterpart of the patch "crypto: arm/ghash - add NEON accelerated fallback for vmull.p64" Raw numbers for a 1.2 Ghz Cortex-A53 (Raspberry Pi3) after the patch. This patch applies onto the patch "crypto: arm64/gcm - implement native driver using v8 Crypto Extensions" which can be found here: http://www.mail-archive.com/linux-crypto@vger.kernel.org/msg26385.html arch/arm64/crypto/ghash-ce-core.S | 161 +++++++++++++++++--- arch/arm64/crypto/ghash-ce-glue.c | 36 ++++- 2 files changed, 170 insertions(+), 27 deletions(-) -- 2.9.3 testing speed of async ghash-generic (ghash-generic) 0 ( 1 x 16 = 16 bytes): 1032792 opers/sec, 16524672 bytes/sec 1 ( 4 x 16 = 64 bytes): 303065 opers/sec, 19396160 bytes/sec 2 ( 1 x 64 = 64 bytes): 398480 opers/sec, 25502720 bytes/sec 3 ( 16 x 16 = 256 bytes): 79072 opers/sec, 20242432 bytes/sec 4 ( 4 x 64 = 256 bytes): 105639 opers/sec, 27043584 bytes/sec 5 ( 1 x 256 = 256 bytes): 115866 opers/sec, 29661696 bytes/sec 6 ( 64 x 16 = 1024 bytes): 20000 opers/sec, 20480000 bytes/sec 7 ( 4 x 256 = 1024 bytes): 29416 opers/sec, 30121984 bytes/sec 8 ( 1 x 1024 = 1024 bytes): 30202 opers/sec, 30926848 bytes/sec 9 (128 x 16 = 2048 bytes): 10021 opers/sec, 20523008 bytes/sec 10 ( 8 x 256 = 2048 bytes): 14749 opers/sec, 30205952 bytes/sec 11 ( 2 x 1024 = 2048 bytes): 15124 opers/sec, 30973952 bytes/sec 12 ( 1 x 2048 = 2048 bytes): 15204 opers/sec, 31137792 bytes/sec 13 (256 x 16 = 4096 bytes): 5010 opers/sec, 20520960 bytes/sec 14 ( 16 x 256 = 4096 bytes): 7391 opers/sec, 30273536 bytes/sec 15 ( 4 x 1024 = 4096 bytes): 7575 opers/sec, 31027200 bytes/sec 16 ( 1 x 4096 = 4096 bytes): 7620 opers/sec, 31211520 bytes/sec 17 (512 x 16 = 8192 bytes): 2507 opers/sec, 20537344 bytes/sec 18 ( 32 x 256 = 8192 bytes): 3698 opers/sec, 30294016 bytes/sec 19 ( 8 x 1024 = 8192 bytes): 3791 opers/sec, 31055872 bytes/sec 20 ( 2 x 4096 = 8192 bytes): 3815 opers/sec, 31252480 bytes/sec 21 ( 1 x 8192 = 8192 bytes): 3813 opers/sec, 31236096 bytes/sec testing speed of async ghash (ghash-ce) 0 ( 1 x 16 = 16 bytes): 1262369 opers/sec, 20197904 bytes/sec 1 ( 4 x 16 = 64 bytes): 374038 opers/sec, 23938432 bytes/sec 2 ( 1 x 64 = 64 bytes): 750298 opers/sec, 48019072 bytes/sec 3 ( 16 x 16 = 256 bytes): 98520 opers/sec, 25221120 bytes/sec 4 ( 4 x 64 = 256 bytes): 206875 opers/sec, 52960000 bytes/sec 5 ( 1 x 256 = 256 bytes): 285419 opers/sec, 73067264 bytes/sec 6 ( 64 x 16 = 1024 bytes): 24942 opers/sec, 25540608 bytes/sec 7 ( 4 x 256 = 1024 bytes): 73911 opers/sec, 75684864 bytes/sec 8 ( 1 x 1024 = 1024 bytes): 82371 opers/sec, 84347904 bytes/sec 9 (128 x 16 = 2048 bytes): 12490 opers/sec, 25579520 bytes/sec 10 ( 8 x 256 = 2048 bytes): 37233 opers/sec, 76253184 bytes/sec 11 ( 2 x 1024 = 2048 bytes): 41424 opers/sec, 84836352 bytes/sec 12 ( 1 x 2048 = 2048 bytes): 42277 opers/sec, 86583296 bytes/sec 13 (256 x 16 = 4096 bytes): 6255 opers/sec, 25620480 bytes/sec 14 ( 16 x 256 = 4096 bytes): 18676 opers/sec, 76496896 bytes/sec 15 ( 4 x 1024 = 4096 bytes): 20785 opers/sec, 85135360 bytes/sec 16 ( 1 x 4096 = 4096 bytes): 21369 opers/sec, 87527424 bytes/sec 17 (512 x 16 = 8192 bytes): 3132 opers/sec, 25657344 bytes/sec 18 ( 32 x 256 = 8192 bytes): 9356 opers/sec, 76644352 bytes/sec 19 ( 8 x 1024 = 8192 bytes): 10394 opers/sec, 85147648 bytes/sec 20 ( 2 x 4096 = 8192 bytes): 10701 opers/sec, 87662592 bytes/sec 21 ( 1 x 8192 = 8192 bytes): 10702 opers/sec, 87670784 bytes/sec testing speed of gcm(aes) (gcm_base(ctr-aes-neonbs,ghash-generic)) encryption 0 (128 bit key, 16 byte blocks): 129339 opers/sec, 2069424 bytes/sec 1 (128 bit key, 64 byte blocks): 106580 opers/sec, 6821120 bytes/sec 2 (128 bit key, 256 byte blocks): 50794 opers/sec, 13003264 bytes/sec 3 (128 bit key, 512 byte blocks): 31399 opers/sec, 16076288 bytes/sec 4 (128 bit key, 1024 byte blocks): 17835 opers/sec, 18263040 bytes/sec 5 (128 bit key, 2048 byte blocks): 9565 opers/sec, 19589120 bytes/sec 6 (128 bit key, 4096 byte blocks): 4973 opers/sec, 20369408 bytes/sec 7 (128 bit key, 8192 byte blocks): 2521 opers/sec, 20652032 bytes/sec 8 (192 bit key, 16 byte blocks): 123632 opers/sec, 1978112 bytes/sec 9 (192 bit key, 64 byte blocks): 102969 opers/sec, 6590016 bytes/sec 10 (192 bit key, 256 byte blocks): 48204 opers/sec, 12340224 bytes/sec 11 (192 bit key, 512 byte blocks): 29747 opers/sec, 15230464 bytes/sec 12 (192 bit key, 1024 byte blocks): 16873 opers/sec, 17277952 bytes/sec 13 (192 bit key, 2048 byte blocks): 9041 opers/sec, 18515968 bytes/sec 14 (192 bit key, 4096 byte blocks): 4700 opers/sec, 19251200 bytes/sec 15 (192 bit key, 8192 byte blocks): 2382 opers/sec, 19513344 bytes/sec 16 (256 bit key, 16 byte blocks): 118382 opers/sec, 1894112 bytes/sec 17 (256 bit key, 64 byte blocks): 98995 opers/sec, 6335680 bytes/sec 18 (256 bit key, 256 byte blocks): 45832 opers/sec, 11732992 bytes/sec 19 (256 bit key, 512 byte blocks): 28262 opers/sec, 14470144 bytes/sec 20 (256 bit key, 1024 byte blocks): 16006 opers/sec, 16390144 bytes/sec 21 (256 bit key, 2048 byte blocks): 8567 opers/sec, 17545216 bytes/sec 22 (256 bit key, 4096 byte blocks): 4447 opers/sec, 18214912 bytes/sec 23 (256 bit key, 8192 byte blocks): 2259 opers/sec, 18505728 bytes/sec testing speed of gcm(aes) (gcm_base(ctr-aes-neonbs,ghash-ce)) encryption 0 (128 bit key, 16 byte blocks): 139252 opers/sec, 2228032 bytes/sec 1 (128 bit key, 64 byte blocks): 128819 opers/sec, 8244416 bytes/sec 2 (128 bit key, 256 byte blocks): 71131 opers/sec, 18209536 bytes/sec 3 (128 bit key, 512 byte blocks): 47844 opers/sec, 24496128 bytes/sec 4 (128 bit key, 1024 byte blocks): 28988 opers/sec, 29683712 bytes/sec 5 (128 bit key, 2048 byte blocks): 16196 opers/sec, 33169408 bytes/sec 6 (128 bit key, 4096 byte blocks): 8613 opers/sec, 35278848 bytes/sec 7 (128 bit key, 8192 byte blocks): 4401 opers/sec, 36052992 bytes/sec 8 (192 bit key, 16 byte blocks): 132723 opers/sec, 2123568 bytes/sec 9 (192 bit key, 64 byte blocks): 123025 opers/sec, 7873600 bytes/sec 10 (192 bit key, 256 byte blocks): 66083 opers/sec, 16917248 bytes/sec 11 (192 bit key, 512 byte blocks): 44115 opers/sec, 22586880 bytes/sec 12 (192 bit key, 1024 byte blocks): 26518 opers/sec, 27154432 bytes/sec 13 (192 bit key, 2048 byte blocks): 14753 opers/sec, 30214144 bytes/sec 14 (192 bit key, 4096 byte blocks): 7825 opers/sec, 32051200 bytes/sec 15 (192 bit key, 8192 byte blocks): 3996 opers/sec, 32735232 bytes/sec 16 (256 bit key, 16 byte blocks): 126708 opers/sec, 2027328 bytes/sec 17 (256 bit key, 64 byte blocks): 117968 opers/sec, 7549952 bytes/sec 18 (256 bit key, 256 byte blocks): 61776 opers/sec, 15814656 bytes/sec 19 (256 bit key, 512 byte blocks): 40926 opers/sec, 20954112 bytes/sec 20 (256 bit key, 1024 byte blocks): 24459 opers/sec, 25046016 bytes/sec 21 (256 bit key, 2048 byte blocks): 13541 opers/sec, 27731968 bytes/sec 22 (256 bit key, 4096 byte blocks): 7154 opers/sec, 29302784 bytes/sec 23 (256 bit key, 8192 byte blocks): 3659 opers/sec, 29974528 bytes/sec diff --git a/arch/arm64/crypto/ghash-ce-core.S b/arch/arm64/crypto/ghash-ce-core.S index cb22459eba85..8a789f6154fc 100644 --- a/arch/arm64/crypto/ghash-ce-core.S +++ b/arch/arm64/crypto/ghash-ce-core.S @@ -1,7 +1,7 @@ /* * Accelerated GHASH implementation with ARMv8 PMULL instructions. * - * Copyright (C) 2014 Linaro Ltd. + * Copyright (C) 2014 - 2017 Linaro Ltd. * * This program is free software; you can redistribute it and/or modify it * under the terms of the GNU General Public License version 2 as published @@ -11,24 +11,119 @@ #include #include - SHASH .req v0 - SHASH2 .req v1 - T1 .req v2 - T2 .req v3 - MASK .req v4 - XL .req v5 - XM .req v6 - XH .req v7 - IN1 .req v7 + SHASH .req v0 + SHASH2 .req v1 + T1 .req v2 + T2 .req v3 + MASK .req v4 + XL .req v5 + XM .req v6 + XH .req v7 + IN1 .req v7 + + k00_16 .req v8 + k32_48 .req v9 + + t3 .req v10 + t4 .req v11 + t5 .req v12 + t6 .req v13 + t7 .req v14 + t8 .req v15 + t9 .req v16 + + perm1 .req v17 + perm2 .req v18 + perm3 .req v19 + perm4 .req v20 .text .arch armv8-a+crypto - /* - * void pmull_ghash_update(int blocks, u64 dg[], const char *src, - * struct ghash_key const *k, const char *head) - */ -ENTRY(pmull_ghash_update) + .macro __pmull_p64, rd, rn, rm, i + .ifb \i + pmull \rd\().1q, \rn\().1d, \rm\().1d + .else + pmull2 \rd\().1q, \rn\().2d, \rm\().2d + .endif + .endm + + .macro __pmull_p8, rq, ad, bd, i + .ifb \i + ext t4.8b, \ad\().8b, \ad\().8b, #1 // A1 + ext t8.8b, \bd\().8b, \bd\().8b, #1 // B1 + ext t5.8b, \ad\().8b, \ad\().8b, #2 // A2 + ext t7.8b, \bd\().8b, \bd\().8b, #2 // B2 + ext t6.8b, \ad\().8b, \ad\().8b, #3 // A3 + ext t9.8b, \bd\().8b, \bd\().8b, #3 // B3 + ext t3.8b, \bd\().8b, \bd\().8b, #4 // B4 + + pmull t4.8h, t4.8b, \bd\().8b // F = A1*B + pmull t8.8h, \ad\().8b, t8.8b // E = A*B1 + pmull t5.8h, t5.8b, \bd\().8b // H = A2*B + pmull t7.8h, \ad\().8b, t7.8b // G = A*B2 + pmull t6.8h, t6.8b, \bd\().8b // J = A3*B + pmull t9.8h, \ad\().8b, t9.8b // I = A*B3 + pmull t3.8h, \ad\().8b, t3.8b // K = A*B4 + pmull \rq\().8h, \ad\().8b, \bd\().8b // D = A*B + .else + tbl t4.16b, {\ad\().16b}, perm1.16b // A1 + tbl t8.16b, {\bd\().16b}, perm1.16b // B1 + tbl t5.16b, {\ad\().16b}, perm2.16b // A2 + tbl t7.16b, {\bd\().16b}, perm2.16b // B2 + tbl t6.16b, {\ad\().16b}, perm3.16b // A3 + tbl t9.16b, {\bd\().16b}, perm3.16b // B3 + tbl t3.16b, {\bd\().16b}, perm4.16b // B4 + + pmull2 t4.8h, t4.16b, \bd\().16b // F = A1*B + pmull2 t8.8h, \ad\().16b, t8.16b // E = A*B1 + pmull2 t5.8h, t5.16b, \bd\().16b // H = A2*B + pmull2 t7.8h, \ad\().16b, t7.16b // G = A*B2 + pmull2 t6.8h, t6.16b, \bd\().16b // J = A3*B + pmull2 t9.8h, \ad\().16b, t9.16b // I = A*B3 + pmull2 t3.8h, \ad\().16b, t3.16b // K = A*B4 + pmull2 \rq\().8h, \ad\().16b, \bd\().16b // D = A*B + .endif + + eor t4.16b, t4.16b, t8.16b // L = E + F + eor t5.16b, t5.16b, t7.16b // M = G + H + eor t6.16b, t6.16b, t9.16b // N = I + J + + uzp1 t8.2d, t4.2d, t5.2d + uzp2 t4.2d, t4.2d, t5.2d + uzp1 t7.2d, t6.2d, t3.2d + uzp2 t6.2d, t6.2d, t3.2d + + // t4 = (L) (P0 + P1) << 8 + // t5 = (M) (P2 + P3) << 16 + eor t8.16b, t8.16b, t4.16b + and t4.16b, t4.16b, k32_48.16b + + // t6 = (N) (P4 + P5) << 24 + // t7 = (K) (P6 + P7) << 32 + eor t7.16b, t7.16b, t6.16b + and t6.16b, t6.16b, k00_16.16b + + eor t8.16b, t8.16b, t4.16b + eor t7.16b, t7.16b, t6.16b + + zip2 t5.2d, t8.2d, t4.2d + zip1 t4.2d, t8.2d, t4.2d + zip2 t3.2d, t7.2d, t6.2d + zip1 t6.2d, t7.2d, t6.2d + + ext t4.16b, t4.16b, t4.16b, #15 + ext t5.16b, t5.16b, t5.16b, #14 + ext t6.16b, t6.16b, t6.16b, #13 + ext t3.16b, t3.16b, t3.16b, #12 + + eor t4.16b, t4.16b, t5.16b + eor t6.16b, t6.16b, t3.16b + eor \rq\().16b, \rq\().16b, t4.16b + eor \rq\().16b, \rq\().16b, t6.16b + .endm + + .macro __pmull_ghash, pm ld1 {SHASH.2d}, [x3] ld1 {XL.2d}, [x1] movi MASK.16b, #0xe1 @@ -52,23 +147,23 @@ CPU_LE( rev64 T1.16b, T1.16b ) eor T1.16b, T1.16b, T2.16b eor XL.16b, XL.16b, IN1.16b - pmull2 XH.1q, SHASH.2d, XL.2d // a1 * b1 + \pm XH, SHASH, XL, 2 // a1 * b1 eor T1.16b, T1.16b, XL.16b - pmull XL.1q, SHASH.1d, XL.1d // a0 * b0 - pmull XM.1q, SHASH2.1d, T1.1d // (a1 + a0)(b1 + b0) + \pm XL, SHASH, XL // a0 * b0 + \pm XM, SHASH2, T1 // (a1 + a0)(b1 + b0) ext T1.16b, XL.16b, XH.16b, #8 eor T2.16b, XL.16b, XH.16b eor XM.16b, XM.16b, T1.16b eor XM.16b, XM.16b, T2.16b - pmull T2.1q, XL.1d, MASK.1d + \pm T2, XL, MASK mov XH.d[0], XM.d[1] mov XM.d[1], XL.d[0] eor XL.16b, XM.16b, T2.16b ext T2.16b, XL.16b, XL.16b, #8 - pmull XL.1q, XL.1d, MASK.1d + \pm XL, XL, MASK eor T2.16b, T2.16b, XH.16b eor XL.16b, XL.16b, T2.16b @@ -76,7 +171,31 @@ CPU_LE( rev64 T1.16b, T1.16b ) st1 {XL.2d}, [x1] ret -ENDPROC(pmull_ghash_update) + .endm + + /* + * void pmull_ghash_update(int blocks, u64 dg[], const char *src, + * struct ghash_key const *k, const char *head) + */ +ENTRY(pmull_ghash_update_p64) + __pmull_ghash __pmull_p64 +ENDPROC(pmull_ghash_update_p64) + +ENTRY(pmull_ghash_update_p8) + // k00_16 := 0x0000000000000000_000000000000ffff + // k32_48 := 0x00000000ffffffff_0000ffffffffffff + movi k32_48.2d, #0xffffffff + mov k32_48.h[2], k32_48.h[0] + ushr k00_16.2d, k32_48.2d, #32 + + mov_q x5, 0x080f0e0d0c0b0a09 + dup perm1.2d, x5 + ext perm2.16b, perm1.16b, perm1.16b, #1 + ext perm3.16b, perm1.16b, perm1.16b, #2 + ext perm4.16b, perm1.16b, perm1.16b, #3 + + __pmull_ghash __pmull_p8 +ENDPROC(pmull_ghash_update_p8) KS .req v8 CTR .req v9 diff --git a/arch/arm64/crypto/ghash-ce-glue.c b/arch/arm64/crypto/ghash-ce-glue.c index 524dd5a5aca1..6bf08e4d84fe 100644 --- a/arch/arm64/crypto/ghash-ce-glue.c +++ b/arch/arm64/crypto/ghash-ce-glue.c @@ -26,6 +26,7 @@ MODULE_DESCRIPTION("GHASH and AES-GCM using ARMv8 Crypto Extensions"); MODULE_AUTHOR("Ard Biesheuvel "); MODULE_LICENSE("GPL v2"); +MODULE_ALIAS_CRYPTO("ghash"); #define GHASH_BLOCK_SIZE 16 #define GHASH_DIGEST_SIZE 16 @@ -48,8 +49,17 @@ struct gcm_aes_ctx { struct ghash_key ghash_key; }; -asmlinkage void pmull_ghash_update(int blocks, u64 dg[], const char *src, - struct ghash_key const *k, const char *head); +asmlinkage void pmull_ghash_update_p64(int blocks, u64 dg[], const char *src, + struct ghash_key const *k, + const char *head); + +asmlinkage void pmull_ghash_update_p8(int blocks, u64 dg[], const char *src, + struct ghash_key const *k, + const char *head); + +static void (*pmull_ghash_update)(int blocks, u64 dg[], const char *src, + struct ghash_key const *k, + const char *head); asmlinkage void pmull_gcm_encrypt(int blocks, u64 dg[], u8 dst[], const u8 src[], struct ghash_key const *k, @@ -554,9 +564,18 @@ static int __init ghash_ce_mod_init(void) { int ret; - ret = crypto_register_aead(&gcm_aes_alg); - if (ret) - return ret; + if (!(elf_hwcap & HWCAP_ASIMD)) + return -ENODEV; + + if (elf_hwcap & HWCAP_PMULL) { + pmull_ghash_update = pmull_ghash_update_p64; + + ret = crypto_register_aead(&gcm_aes_alg); + if (ret) + return ret; + } else { + pmull_ghash_update = pmull_ghash_update_p8; + } ret = crypto_register_shash(&ghash_alg); if (ret) @@ -570,5 +589,10 @@ static void __exit ghash_ce_mod_exit(void) crypto_unregister_aead(&gcm_aes_alg); } -module_cpu_feature_match(PMULL, ghash_ce_mod_init); +static const struct cpu_feature ghash_cpu_feature[] = { + { cpu_feature(PMULL) }, { } +}; +MODULE_DEVICE_TABLE(cpu, ghash_cpu_feature); + +module_init(ghash_ce_mod_init); module_exit(ghash_ce_mod_exit);