From patchwork Sun Aug 11 22:59:10 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Ard Biesheuvel X-Patchwork-Id: 11089191 X-Patchwork-Delegate: herbert@gondor.apana.org.au Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id DCA381399 for ; Sun, 11 Aug 2019 22:59:29 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id CB60D27F2B for ; Sun, 11 Aug 2019 22:59:29 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id BC14D27F82; Sun, 11 Aug 2019 22:59:29 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-8.0 required=2.0 tests=BAYES_00,DKIM_SIGNED, DKIM_VALID,DKIM_VALID_AU,MAILING_LIST_MULTI,RCVD_IN_DNSWL_HI autolearn=ham version=3.3.1 Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 80B1B27F2B for ; Sun, 11 Aug 2019 22:59:28 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1726014AbfHKW72 (ORCPT ); Sun, 11 Aug 2019 18:59:28 -0400 Received: from mail-wr1-f66.google.com ([209.85.221.66]:39835 "EHLO mail-wr1-f66.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726144AbfHKW71 (ORCPT ); Sun, 11 Aug 2019 18:59:27 -0400 Received: by mail-wr1-f66.google.com with SMTP id t16so12925074wra.6 for ; Sun, 11 Aug 2019 15:59:25 -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:in-reply-to:references; bh=WBpXCKjdfq1yPU+OJBvQfW6gPuB8RZ7Tc8kFP902ZIM=; b=kCofGDQRD+qKmOAJWzOnSnR6Wwz3Y9bYdcIfvJWQa1nk/Nnm8BgcQvgttireKoC9i4 ZhRJBbwXTXAil6kuWzTLxLgxUBrB4DqBxWjDgKCnT0jNVaRnvTUkm/pa9qLwCMx3EQA+ slZ+LKH/Z3EDpse+Qr0PhDq5mMt10BPSx2FSaGL2OfjZFMdCiZWaZPQSHrHpoYaqcm2d 4Mv5K/NUqtSUGqJslAsSTRtrSrK4FKGWKxXrngRBx147X5CwzS1m71p3gu4zXxHeuBLt t47ZnwSxcebtwTrcRbjsboIR8EHe3FYQKtFEv+pRVMNlPVkM18VWZ/mqpUE6enpUFZy1 gbdA== 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:in-reply-to :references; bh=WBpXCKjdfq1yPU+OJBvQfW6gPuB8RZ7Tc8kFP902ZIM=; b=STMNTzDavXuJCiAb3bPdb2bDn9NJDoat1SuHjlalRtEAfZvE49mphA4dJqaRjNgMYF 87vStBem2m//qqMd7JFjMTXtUA6OhUHmdL1QIXMeO0VI95PD74lzbcwpWUqCClwGGxTj 7FWuR2fdG0Mh3p5QbhuNZ+LSGhmapiQSgaXelbJW2HUZc/Y68KHKcQsFe4/uZPSaWefr hhBrcp++3TTxlllrs3OiafSWf9toWmAUHLWIBLudjiGWiNJr+EWuQmneX0Khrg4SbZQ5 u2tT32E5kaVdI9D0/dZnjgV8joljP/xcUqpe70V923cwshkMKp0pii1RRT3molXg5Mal 5IKQ== X-Gm-Message-State: APjAAAWO1+9rODBJO23Ou1dZp9WHPV5h9R5nXqAxtBOeIF4CHbUAGfRV KMhkFLYOKePlmztcZIX1Z2Obykiq6WXZug== X-Google-Smtp-Source: APXvYqwNeOvn8OGLFcFwne3f/5KXZFeXbvl/XTwD3j+KLcZuFebcUFtpjmrkX1JUouUphdHFACnecA== X-Received: by 2002:adf:9224:: with SMTP id 33mr23351134wrj.238.1565564364796; Sun, 11 Aug 2019 15:59:24 -0700 (PDT) Received: from localhost.localdomain ([2a02:587:a407:da00:5df1:4fcc:7bd1:4860]) by smtp.gmail.com with ESMTPSA id a17sm5930888wmm.47.2019.08.11.15.59.23 (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Sun, 11 Aug 2019 15:59:24 -0700 (PDT) From: Ard Biesheuvel To: linux-crypto@vger.kernel.org Cc: herbert@gondor.apana.org.au, ebiggers@kernel.org, Ard Biesheuvel Subject: [PATCH v2 1/3] crypto: aegis128 - add support for SIMD acceleration Date: Mon, 12 Aug 2019 01:59:10 +0300 Message-Id: <20190811225912.19412-2-ard.biesheuvel@linaro.org> X-Mailer: git-send-email 2.17.1 In-Reply-To: <20190811225912.19412-1-ard.biesheuvel@linaro.org> References: <20190811225912.19412-1-ard.biesheuvel@linaro.org> Sender: linux-crypto-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-crypto@vger.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP Add some plumbing to allow the AEGIS128 code to be built with SIMD routines for acceleration. Reviewed-by: Ondrej Mosnacek Signed-off-by: Ard Biesheuvel Signed-off-by: Herbert Xu --- crypto/Makefile | 1 + crypto/{aegis128.c => aegis128-core.c} | 52 ++++++++++++++++++-- 2 files changed, 49 insertions(+), 4 deletions(-) diff --git a/crypto/Makefile b/crypto/Makefile index cfcc954e59f9..92e985714ff6 100644 --- a/crypto/Makefile +++ b/crypto/Makefile @@ -90,6 +90,7 @@ obj-$(CONFIG_CRYPTO_GCM) += gcm.o obj-$(CONFIG_CRYPTO_CCM) += ccm.o obj-$(CONFIG_CRYPTO_CHACHA20POLY1305) += chacha20poly1305.o obj-$(CONFIG_CRYPTO_AEGIS128) += aegis128.o +aegis128-y := aegis128-core.o obj-$(CONFIG_CRYPTO_PCRYPT) += pcrypt.o obj-$(CONFIG_CRYPTO_CRYPTD) += cryptd.o obj-$(CONFIG_CRYPTO_DES) += des_generic.o diff --git a/crypto/aegis128.c b/crypto/aegis128-core.c similarity index 89% rename from crypto/aegis128.c rename to crypto/aegis128-core.c index 32840d5e7f65..fa69e99968e2 100644 --- a/crypto/aegis128.c +++ b/crypto/aegis128-core.c @@ -8,6 +8,7 @@ #include #include +#include #include #include #include @@ -16,6 +17,8 @@ #include #include +#include + #include "aegis.h" #define AEGIS128_NONCE_SIZE 16 @@ -40,6 +43,24 @@ struct aegis128_ops { const u8 *src, unsigned int size); }; +static bool have_simd; + +static bool aegis128_do_simd(void) +{ +#ifdef CONFIG_CRYPTO_AEGIS128_SIMD + if (have_simd) + return crypto_simd_usable(); +#endif + return false; +} + +bool crypto_aegis128_have_simd(void); +void crypto_aegis128_update_simd(struct aegis_state *state, const void *msg); +void crypto_aegis128_encrypt_chunk_simd(struct aegis_state *state, u8 *dst, + const u8 *src, unsigned int size); +void crypto_aegis128_decrypt_chunk_simd(struct aegis_state *state, u8 *dst, + const u8 *src, unsigned int size); + static void crypto_aegis128_update(struct aegis_state *state) { union aegis_block tmp; @@ -55,12 +76,22 @@ static void crypto_aegis128_update(struct aegis_state *state) static void crypto_aegis128_update_a(struct aegis_state *state, const union aegis_block *msg) { + if (aegis128_do_simd()) { + crypto_aegis128_update_simd(state, msg); + return; + } + crypto_aegis128_update(state); crypto_aegis_block_xor(&state->blocks[0], msg); } static void crypto_aegis128_update_u(struct aegis_state *state, const void *msg) { + if (aegis128_do_simd()) { + crypto_aegis128_update_simd(state, msg); + return; + } + crypto_aegis128_update(state); crypto_xor(state->blocks[0].bytes, msg, AEGIS_BLOCK_SIZE); } @@ -365,7 +396,7 @@ static void crypto_aegis128_crypt(struct aead_request *req, static int crypto_aegis128_encrypt(struct aead_request *req) { - static const struct aegis128_ops ops = { + const struct aegis128_ops *ops = &(struct aegis128_ops){ .skcipher_walk_init = skcipher_walk_aead_encrypt, .crypt_chunk = crypto_aegis128_encrypt_chunk, }; @@ -375,7 +406,12 @@ static int crypto_aegis128_encrypt(struct aead_request *req) unsigned int authsize = crypto_aead_authsize(tfm); unsigned int cryptlen = req->cryptlen; - crypto_aegis128_crypt(req, &tag, cryptlen, &ops); + if (aegis128_do_simd()) + ops = &(struct aegis128_ops){ + .skcipher_walk_init = skcipher_walk_aead_encrypt, + .crypt_chunk = crypto_aegis128_encrypt_chunk_simd }; + + crypto_aegis128_crypt(req, &tag, cryptlen, ops); scatterwalk_map_and_copy(tag.bytes, req->dst, req->assoclen + cryptlen, authsize, 1); @@ -384,7 +420,7 @@ static int crypto_aegis128_encrypt(struct aead_request *req) static int crypto_aegis128_decrypt(struct aead_request *req) { - static const struct aegis128_ops ops = { + const struct aegis128_ops *ops = &(struct aegis128_ops){ .skcipher_walk_init = skcipher_walk_aead_decrypt, .crypt_chunk = crypto_aegis128_decrypt_chunk, }; @@ -398,7 +434,12 @@ static int crypto_aegis128_decrypt(struct aead_request *req) scatterwalk_map_and_copy(tag.bytes, req->src, req->assoclen + cryptlen, authsize, 0); - crypto_aegis128_crypt(req, &tag, cryptlen, &ops); + if (aegis128_do_simd()) + ops = &(struct aegis128_ops){ + .skcipher_walk_init = skcipher_walk_aead_decrypt, + .crypt_chunk = crypto_aegis128_decrypt_chunk_simd }; + + crypto_aegis128_crypt(req, &tag, cryptlen, ops); return crypto_memneq(tag.bytes, zeros, authsize) ? -EBADMSG : 0; } @@ -429,6 +470,9 @@ static struct aead_alg crypto_aegis128_alg = { static int __init crypto_aegis128_module_init(void) { + if (IS_ENABLED(CONFIG_CRYPTO_AEGIS128_SIMD)) + have_simd = crypto_aegis128_have_simd(); + return crypto_register_aead(&crypto_aegis128_alg); } From patchwork Sun Aug 11 22:59:11 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Ard Biesheuvel X-Patchwork-Id: 11089193 X-Patchwork-Delegate: herbert@gondor.apana.org.au Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 5E6921399 for ; Sun, 11 Aug 2019 22:59:32 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 50D8027F2B for ; Sun, 11 Aug 2019 22:59:32 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 4530B27F82; Sun, 11 Aug 2019 22:59:32 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-8.0 required=2.0 tests=BAYES_00,DKIM_SIGNED, DKIM_VALID,DKIM_VALID_AU,MAILING_LIST_MULTI,RCVD_IN_DNSWL_HI autolearn=ham version=3.3.1 Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 9018727F2B for ; Sun, 11 Aug 2019 22:59:31 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1726179AbfHKW7b (ORCPT ); Sun, 11 Aug 2019 18:59:31 -0400 Received: from mail-wm1-f67.google.com ([209.85.128.67]:52706 "EHLO mail-wm1-f67.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726144AbfHKW7b (ORCPT ); Sun, 11 Aug 2019 18:59:31 -0400 Received: by mail-wm1-f67.google.com with SMTP id s3so10539608wms.2 for ; Sun, 11 Aug 2019 15:59:28 -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:in-reply-to:references; bh=NdlkfYolXkiJlDnDSe4O3QjW41knJBAaFb7TIELDbHQ=; b=pbNLyV8bnXHnNatF6b/yKRL+M/BxVcZIk7gICdle+gopoNDWbwOojNw/cJirgcUHGw +GDnRMsor26jbxCA1P/zlrX2nU1hg0I6ShjytVlVQQvTVugZ6wpV5nedwJwtTwb4MtNH QiL09d8sz4JuJJ5JP6lFiYtJ9519/vwJtXzhf7DYBJYoQ0GiXMS902B5zQVupLRBvT4h gRtIv4M3+k/j9xmOXAvuB7T0hKP3S+h1wM8j5xy+2pQkEnd7oe8uVWy5+VwhswpTT88o Yxdh3yc7nHyYQhCmzyco7psQLFlu5s6Bk20RbV3bLHgCGjElKkcsu8C61IFIgYXA6NGV 6+bQ== 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:in-reply-to :references; bh=NdlkfYolXkiJlDnDSe4O3QjW41knJBAaFb7TIELDbHQ=; b=uiNWF+p6Ej5g3a4V0liaAIESpZPlyf3OrAS4sOWD8t8xs+7/62dsBflYgrjNbfv1qg JCToXnRposF3vhEynoYubX4fNl8983TmP5GObnzIGp88RCG91D3JVurL/HgbltOlldpc aj0Hiz5GHRPKn6H7WCw/ACipxh/X5eu6Trg4ttjmyAyjmZ2+AOw0nH0r8/gbrwELcT08 A6yGdU64e5H1td8lPKk+WATtpYmTzIdE0nsW9zNzy3UmBMIrvw/EckwAG3g6XnQNGCkC HNPHKD/qLzBiQUjWBG+8iI6rHqpke0NHpTNJjEF1AFBPC3WoRCDwUmNhRAU8OGgCZkWL zUjw== X-Gm-Message-State: APjAAAU8d60yXfi27SGtsM/JmFUpEii4n7C0vj6bDouX+8DLUZr3jLVK MWRNVfR7O167PvA4Xqbqkf1J0hkRSD3Njg== X-Google-Smtp-Source: APXvYqzMdKP+BBMI8AJdZ1zJhqQteKiCJ27Gje1RUHj9eTJ41ZwfVGEpRCvwS7m8ZUhiIYGnJAagmw== X-Received: by 2002:a1c:968c:: with SMTP id y134mr23989984wmd.75.1565564367293; Sun, 11 Aug 2019 15:59:27 -0700 (PDT) Received: from localhost.localdomain ([2a02:587:a407:da00:5df1:4fcc:7bd1:4860]) by smtp.gmail.com with ESMTPSA id a17sm5930888wmm.47.2019.08.11.15.59.24 (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Sun, 11 Aug 2019 15:59:26 -0700 (PDT) From: Ard Biesheuvel To: linux-crypto@vger.kernel.org Cc: herbert@gondor.apana.org.au, ebiggers@kernel.org, Ard Biesheuvel Subject: [PATCH v2 2/3] crypto: aegis128 - provide a SIMD implementation based on NEON intrinsics Date: Mon, 12 Aug 2019 01:59:11 +0300 Message-Id: <20190811225912.19412-3-ard.biesheuvel@linaro.org> X-Mailer: git-send-email 2.17.1 In-Reply-To: <20190811225912.19412-1-ard.biesheuvel@linaro.org> References: <20190811225912.19412-1-ard.biesheuvel@linaro.org> Sender: linux-crypto-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-crypto@vger.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP Provide an accelerated implementation of aegis128 by wiring up the SIMD hooks in the generic driver to an implementation based on NEON intrinsics, which can be compiled to both ARM and arm64 code. This results in a performance of 2.2 cycles per byte on Cortex-A53, which is a performance increase of ~11x compared to the generic code. Reviewed-by: Ondrej Mosnacek Signed-off-by: Ard Biesheuvel Signed-off-by: Herbert Xu --- crypto/Kconfig | 5 + crypto/Makefile | 12 ++ crypto/aegis128-neon-inner.c | 147 ++++++++++++++++++++ crypto/aegis128-neon.c | 43 ++++++ 4 files changed, 207 insertions(+) diff --git a/crypto/Kconfig b/crypto/Kconfig index 8880c1fc51d8..455a3354e291 100644 --- a/crypto/Kconfig +++ b/crypto/Kconfig @@ -306,6 +306,11 @@ config CRYPTO_AEGIS128 help Support for the AEGIS-128 dedicated AEAD algorithm. +config CRYPTO_AEGIS128_SIMD + bool "Support SIMD acceleration for AEGIS-128" + depends on CRYPTO_AEGIS128 && ((ARM || ARM64) && KERNEL_MODE_NEON) + default y + config CRYPTO_AEGIS128_AESNI_SSE2 tristate "AEGIS-128 AEAD algorithm (x86_64 AESNI+SSE2 implementation)" depends on X86 && 64BIT diff --git a/crypto/Makefile b/crypto/Makefile index 92e985714ff6..99a9fa9087d1 100644 --- a/crypto/Makefile +++ b/crypto/Makefile @@ -91,6 +91,18 @@ obj-$(CONFIG_CRYPTO_CCM) += ccm.o obj-$(CONFIG_CRYPTO_CHACHA20POLY1305) += chacha20poly1305.o obj-$(CONFIG_CRYPTO_AEGIS128) += aegis128.o aegis128-y := aegis128-core.o + +ifeq ($(ARCH),arm) +CFLAGS_aegis128-neon-inner.o += -ffreestanding -march=armv7-a -mfloat-abi=softfp +CFLAGS_aegis128-neon-inner.o += -mfpu=crypto-neon-fp-armv8 +aegis128-$(CONFIG_CRYPTO_AEGIS128_SIMD) += aegis128-neon.o aegis128-neon-inner.o +endif +ifeq ($(ARCH),arm64) +CFLAGS_aegis128-neon-inner.o += -ffreestanding -mcpu=generic+crypto +CFLAGS_REMOVE_aegis128-neon-inner.o += -mgeneral-regs-only +aegis128-$(CONFIG_CRYPTO_AEGIS128_SIMD) += aegis128-neon.o aegis128-neon-inner.o +endif + obj-$(CONFIG_CRYPTO_PCRYPT) += pcrypt.o obj-$(CONFIG_CRYPTO_CRYPTD) += cryptd.o obj-$(CONFIG_CRYPTO_DES) += des_generic.o diff --git a/crypto/aegis128-neon-inner.c b/crypto/aegis128-neon-inner.c new file mode 100644 index 000000000000..3d8043c4832b --- /dev/null +++ b/crypto/aegis128-neon-inner.c @@ -0,0 +1,147 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright (C) 2019 Linaro, Ltd. + */ + +#ifdef CONFIG_ARM64 +#include + +#define AES_ROUND "aese %0.16b, %1.16b \n\t aesmc %0.16b, %0.16b" +#else +#include + +#define AES_ROUND "aese.8 %q0, %q1 \n\t aesmc.8 %q0, %q0" +#endif + +#define AEGIS_BLOCK_SIZE 16 + +#include + +void *memcpy(void *dest, const void *src, size_t n); +void *memset(void *s, int c, size_t n); + +struct aegis128_state { + uint8x16_t v[5]; +}; + +static struct aegis128_state aegis128_load_state_neon(const void *state) +{ + return (struct aegis128_state){ { + vld1q_u8(state), + vld1q_u8(state + 16), + vld1q_u8(state + 32), + vld1q_u8(state + 48), + vld1q_u8(state + 64) + } }; +} + +static void aegis128_save_state_neon(struct aegis128_state st, void *state) +{ + vst1q_u8(state, st.v[0]); + vst1q_u8(state + 16, st.v[1]); + vst1q_u8(state + 32, st.v[2]); + vst1q_u8(state + 48, st.v[3]); + vst1q_u8(state + 64, st.v[4]); +} + +static inline __attribute__((always_inline)) +uint8x16_t aegis_aes_round(uint8x16_t w) +{ + uint8x16_t z = {}; + + /* + * We use inline asm here instead of the vaeseq_u8/vaesmcq_u8 intrinsics + * to force the compiler to issue the aese/aesmc instructions in pairs. + * This is much faster on many cores, where the instruction pair can + * execute in a single cycle. + */ + asm(AES_ROUND : "+w"(w) : "w"(z)); + return w; +} + +static inline __attribute__((always_inline)) +struct aegis128_state aegis128_update_neon(struct aegis128_state st, + uint8x16_t m) +{ + m ^= aegis_aes_round(st.v[4]); + st.v[4] ^= aegis_aes_round(st.v[3]); + st.v[3] ^= aegis_aes_round(st.v[2]); + st.v[2] ^= aegis_aes_round(st.v[1]); + st.v[1] ^= aegis_aes_round(st.v[0]); + st.v[0] ^= m; + + return st; +} + +void crypto_aegis128_update_neon(void *state, const void *msg) +{ + struct aegis128_state st = aegis128_load_state_neon(state); + + st = aegis128_update_neon(st, vld1q_u8(msg)); + + aegis128_save_state_neon(st, state); +} + +void crypto_aegis128_encrypt_chunk_neon(void *state, void *dst, const void *src, + unsigned int size) +{ + struct aegis128_state st = aegis128_load_state_neon(state); + uint8x16_t msg; + + while (size >= AEGIS_BLOCK_SIZE) { + uint8x16_t s = st.v[1] ^ (st.v[2] & st.v[3]) ^ st.v[4]; + + msg = vld1q_u8(src); + st = aegis128_update_neon(st, msg); + vst1q_u8(dst, msg ^ s); + + size -= AEGIS_BLOCK_SIZE; + src += AEGIS_BLOCK_SIZE; + dst += AEGIS_BLOCK_SIZE; + } + + if (size > 0) { + uint8x16_t s = st.v[1] ^ (st.v[2] & st.v[3]) ^ st.v[4]; + uint8_t buf[AEGIS_BLOCK_SIZE] = {}; + + memcpy(buf, src, size); + msg = vld1q_u8(buf); + st = aegis128_update_neon(st, msg); + vst1q_u8(buf, msg ^ s); + memcpy(dst, buf, size); + } + + aegis128_save_state_neon(st, state); +} + +void crypto_aegis128_decrypt_chunk_neon(void *state, void *dst, const void *src, + unsigned int size) +{ + struct aegis128_state st = aegis128_load_state_neon(state); + uint8x16_t msg; + + while (size >= AEGIS_BLOCK_SIZE) { + msg = vld1q_u8(src) ^ st.v[1] ^ (st.v[2] & st.v[3]) ^ st.v[4]; + st = aegis128_update_neon(st, msg); + vst1q_u8(dst, msg); + + size -= AEGIS_BLOCK_SIZE; + src += AEGIS_BLOCK_SIZE; + dst += AEGIS_BLOCK_SIZE; + } + + if (size > 0) { + uint8x16_t s = st.v[1] ^ (st.v[2] & st.v[3]) ^ st.v[4]; + uint8_t buf[AEGIS_BLOCK_SIZE]; + + vst1q_u8(buf, s); + memcpy(buf, src, size); + msg = vld1q_u8(buf) ^ s; + vst1q_u8(buf, msg); + memcpy(dst, buf, size); + + st = aegis128_update_neon(st, msg); + } + + aegis128_save_state_neon(st, state); +} diff --git a/crypto/aegis128-neon.c b/crypto/aegis128-neon.c new file mode 100644 index 000000000000..c1c0a1686f67 --- /dev/null +++ b/crypto/aegis128-neon.c @@ -0,0 +1,43 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright (C) 2019 Linaro Ltd + */ + +#include +#include + +#include "aegis.h" + +void crypto_aegis128_update_neon(void *state, const void *msg); +void crypto_aegis128_encrypt_chunk_neon(void *state, void *dst, const void *src, + unsigned int size); +void crypto_aegis128_decrypt_chunk_neon(void *state, void *dst, const void *src, + unsigned int size); + +bool crypto_aegis128_have_simd(void) +{ + return cpu_have_feature(cpu_feature(AES)); +} + +void crypto_aegis128_update_simd(union aegis_block *state, const void *msg) +{ + kernel_neon_begin(); + crypto_aegis128_update_neon(state, msg); + kernel_neon_end(); +} + +void crypto_aegis128_encrypt_chunk_simd(union aegis_block *state, u8 *dst, + const u8 *src, unsigned int size) +{ + kernel_neon_begin(); + crypto_aegis128_encrypt_chunk_neon(state, dst, src, size); + kernel_neon_end(); +} + +void crypto_aegis128_decrypt_chunk_simd(union aegis_block *state, u8 *dst, + const u8 *src, unsigned int size) +{ + kernel_neon_begin(); + crypto_aegis128_decrypt_chunk_neon(state, dst, src, size); + kernel_neon_end(); +} From patchwork Sun Aug 11 22:59:12 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Ard Biesheuvel X-Patchwork-Id: 11089195 X-Patchwork-Delegate: herbert@gondor.apana.org.au Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 169601399 for ; Sun, 11 Aug 2019 22:59:34 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 065BF27F2B for ; Sun, 11 Aug 2019 22:59:34 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id ED5F627F82; Sun, 11 Aug 2019 22:59:33 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-8.0 required=2.0 tests=BAYES_00,DKIM_SIGNED, DKIM_VALID,DKIM_VALID_AU,MAILING_LIST_MULTI,RCVD_IN_DNSWL_HI autolearn=ham version=3.3.1 Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 5600727F2B for ; Sun, 11 Aug 2019 22:59:33 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1726144AbfHKW7c (ORCPT ); Sun, 11 Aug 2019 18:59:32 -0400 Received: from mail-wm1-f67.google.com ([209.85.128.67]:38097 "EHLO mail-wm1-f67.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726164AbfHKW7c (ORCPT ); Sun, 11 Aug 2019 18:59:32 -0400 Received: by mail-wm1-f67.google.com with SMTP id m125so6123494wmm.3 for ; Sun, 11 Aug 2019 15:59:30 -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:in-reply-to:references; bh=AWsY2soOMxAD12H34LvTvFzff/WUbKiZ2Byzr98+9CE=; b=wy/cf3sMMgiSvzTXFMbmS7h0Mhvph+S/NHLixX1zQdYlr683eAAoVKqRmwqhsqquR9 12euTzNqgruvc8jguJyQStdac7TcEHZxSpfSRERPeUDdMGfRVhLdnaKYAyQzb3I0uaZY rNVTQgiQslhz9pmrfrRbKI9ToZk4stlk+NxEQCueNd9OY+B/dc7+8jGb406/AC8yeQK5 mY9+2IsXhzPeAqbO7EdLcGbCuQHvqn13LMzIvxuGGuAlT2Xv8pN6gPQKCk+ozzoHoQmU llw449Gne0ww0PTYYyltJ4nA6jdCvuLWKM9CfBdC+H3hIeM9jMzDnOlisIHnxZzzafXS ahaA== 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:in-reply-to :references; bh=AWsY2soOMxAD12H34LvTvFzff/WUbKiZ2Byzr98+9CE=; b=hWeXwMmNn2ksZP2NXPUhVz0RcCnyWxyJDUXMHtR++vQbgTD8WxD8I5CnhcutB/YhHx 4CtATJ1+RXtcHP1NOQFLmufhKvfRnN5OQ1LdMjtwUGD7fFtSC3CDzR4V2J0PmMwuehhE fkqmNX39J220JMeMPfEFkGckTlfzoS0b1A5jKFhuz28d10tQV5syeWFHhoj0NbCjgcdU CxODzAH6l7DroIohlfG2mMlO91hU5CU1MBHCUTZWJoUPIkTXuSIwTTdGzxj9d1gYr/nF NhR0vIQoNfsZXqvFSOLsxR9D7KM8vGjmdxE5RdgKcPnLTBnjuRxyuEnCYZwuzN+/q691 P4Dg== X-Gm-Message-State: APjAAAV/N8Jd9/QRz69LsPVOnebTJHg/YO7jYEzXj7CN5ViwR7gx69n+ HzhxCyO14vM780Hyk6JB2chciAAmj3tTMA== X-Google-Smtp-Source: APXvYqyYdi02jzLVcvQq3H4NR8X0UNZPHvmjYB7akbU+vHk6HkTmdgRkrGuJvqTjupR2yYdJAFZ/Ew== X-Received: by 2002:a1c:61d4:: with SMTP id v203mr23504532wmb.164.1565564369426; Sun, 11 Aug 2019 15:59:29 -0700 (PDT) Received: from localhost.localdomain ([2a02:587:a407:da00:5df1:4fcc:7bd1:4860]) by smtp.gmail.com with ESMTPSA id a17sm5930888wmm.47.2019.08.11.15.59.27 (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Sun, 11 Aug 2019 15:59:28 -0700 (PDT) From: Ard Biesheuvel To: linux-crypto@vger.kernel.org Cc: herbert@gondor.apana.org.au, ebiggers@kernel.org, Ard Biesheuvel , Nick Desaulniers Subject: [PATCH v2 3/3] crypto: arm64/aegis128 - implement plain NEON version Date: Mon, 12 Aug 2019 01:59:12 +0300 Message-Id: <20190811225912.19412-4-ard.biesheuvel@linaro.org> X-Mailer: git-send-email 2.17.1 In-Reply-To: <20190811225912.19412-1-ard.biesheuvel@linaro.org> References: <20190811225912.19412-1-ard.biesheuvel@linaro.org> Sender: linux-crypto-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-crypto@vger.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP Provide a version of the core AES transform to the aegis128 SIMD code that does not rely on the special AES instructions, but uses plain NEON instructions instead. This allows the SIMD version of the aegis128 driver to be used on arm64 systems that do not implement those instructions (which are not mandatory in the architecture), such as the Raspberry Pi 3. Since GCC makes a mess of this when using the tbl/tbx intrinsics to perform the sbox substitution, preload the Sbox into v16..v31 in this case and use inline asm to emit the tbl/tbx instructions. Clang does not support this approach, nor does it require it, since it does a much better job at code generation, so there we use the intrinsics as usual. Cc: Nick Desaulniers Signed-off-by: Ard Biesheuvel Acked-by: Nick Desaulniers --- crypto/Makefile | 9 ++- crypto/aegis128-neon-inner.c | 65 ++++++++++++++++++++ crypto/aegis128-neon.c | 8 ++- 3 files changed, 80 insertions(+), 2 deletions(-) diff --git a/crypto/Makefile b/crypto/Makefile index 99a9fa9087d1..0d2cdd523fd9 100644 --- a/crypto/Makefile +++ b/crypto/Makefile @@ -98,7 +98,14 @@ CFLAGS_aegis128-neon-inner.o += -mfpu=crypto-neon-fp-armv8 aegis128-$(CONFIG_CRYPTO_AEGIS128_SIMD) += aegis128-neon.o aegis128-neon-inner.o endif ifeq ($(ARCH),arm64) -CFLAGS_aegis128-neon-inner.o += -ffreestanding -mcpu=generic+crypto +aegis128-cflags-y := -ffreestanding -mcpu=generic+crypto +aegis128-cflags-$(CONFIG_CC_IS_GCC) += -ffixed-q16 -ffixed-q17 -ffixed-q18 \ + -ffixed-q19 -ffixed-q20 -ffixed-q21 \ + -ffixed-q22 -ffixed-q23 -ffixed-q24 \ + -ffixed-q25 -ffixed-q26 -ffixed-q27 \ + -ffixed-q28 -ffixed-q29 -ffixed-q30 \ + -ffixed-q31 +CFLAGS_aegis128-neon-inner.o += $(aegis128-cflags-y) CFLAGS_REMOVE_aegis128-neon-inner.o += -mgeneral-regs-only aegis128-$(CONFIG_CRYPTO_AEGIS128_SIMD) += aegis128-neon.o aegis128-neon-inner.o endif diff --git a/crypto/aegis128-neon-inner.c b/crypto/aegis128-neon-inner.c index 3d8043c4832b..ed55568afd1b 100644 --- a/crypto/aegis128-neon-inner.c +++ b/crypto/aegis128-neon-inner.c @@ -17,6 +17,8 @@ #include +extern int aegis128_have_aes_insn; + void *memcpy(void *dest, const void *src, size_t n); void *memset(void *s, int c, size_t n); @@ -24,6 +26,8 @@ struct aegis128_state { uint8x16_t v[5]; }; +extern const uint8x16x4_t crypto_aes_sbox[]; + static struct aegis128_state aegis128_load_state_neon(const void *state) { return (struct aegis128_state){ { @@ -49,6 +53,46 @@ uint8x16_t aegis_aes_round(uint8x16_t w) { uint8x16_t z = {}; +#ifdef CONFIG_ARM64 + if (!__builtin_expect(aegis128_have_aes_insn, 1)) { + static const uint8x16_t shift_rows = { + 0x0, 0x5, 0xa, 0xf, 0x4, 0x9, 0xe, 0x3, + 0x8, 0xd, 0x2, 0x7, 0xc, 0x1, 0x6, 0xb, + }; + static const uint8x16_t ror32by8 = { + 0x1, 0x2, 0x3, 0x0, 0x5, 0x6, 0x7, 0x4, + 0x9, 0xa, 0xb, 0x8, 0xd, 0xe, 0xf, 0xc, + }; + uint8x16_t v; + + // shift rows + w = vqtbl1q_u8(w, shift_rows); + + // sub bytes + if (!IS_ENABLED(CONFIG_CC_IS_GCC)) { + v = vqtbl4q_u8(crypto_aes_sbox[0], w); + v = vqtbx4q_u8(v, crypto_aes_sbox[1], w - 0x40); + v = vqtbx4q_u8(v, crypto_aes_sbox[2], w - 0x80); + v = vqtbx4q_u8(v, crypto_aes_sbox[3], w - 0xc0); + } else { + asm("tbl %0.16b, {v16.16b-v19.16b}, %1.16b" : "=w"(v) : "w"(w)); + w -= 0x40; + asm("tbx %0.16b, {v20.16b-v23.16b}, %1.16b" : "+w"(v) : "w"(w)); + w -= 0x40; + asm("tbx %0.16b, {v24.16b-v27.16b}, %1.16b" : "+w"(v) : "w"(w)); + w -= 0x40; + asm("tbx %0.16b, {v28.16b-v31.16b}, %1.16b" : "+w"(v) : "w"(w)); + } + + // mix columns + w = (v << 1) ^ (uint8x16_t)(((int8x16_t)v >> 7) & 0x1b); + w ^= (uint8x16_t)vrev32q_u16((uint16x8_t)v); + w ^= vqtbl1q_u8(v ^ w, ror32by8); + + return w; + } +#endif + /* * We use inline asm here instead of the vaeseq_u8/vaesmcq_u8 intrinsics * to force the compiler to issue the aese/aesmc instructions in pairs. @@ -73,10 +117,27 @@ struct aegis128_state aegis128_update_neon(struct aegis128_state st, return st; } +static inline __attribute__((always_inline)) +void preload_sbox(void) +{ + if (!IS_ENABLED(CONFIG_ARM64) || + !IS_ENABLED(CONFIG_CC_IS_GCC) || + __builtin_expect(aegis128_have_aes_insn, 1)) + return; + + asm("ld1 {v16.16b-v19.16b}, [%0], #64 \n\t" + "ld1 {v20.16b-v23.16b}, [%0], #64 \n\t" + "ld1 {v24.16b-v27.16b}, [%0], #64 \n\t" + "ld1 {v28.16b-v31.16b}, [%0] \n\t" + :: "r"(crypto_aes_sbox)); +} + void crypto_aegis128_update_neon(void *state, const void *msg) { struct aegis128_state st = aegis128_load_state_neon(state); + preload_sbox(); + st = aegis128_update_neon(st, vld1q_u8(msg)); aegis128_save_state_neon(st, state); @@ -88,6 +149,8 @@ void crypto_aegis128_encrypt_chunk_neon(void *state, void *dst, const void *src, struct aegis128_state st = aegis128_load_state_neon(state); uint8x16_t msg; + preload_sbox(); + while (size >= AEGIS_BLOCK_SIZE) { uint8x16_t s = st.v[1] ^ (st.v[2] & st.v[3]) ^ st.v[4]; @@ -120,6 +183,8 @@ void crypto_aegis128_decrypt_chunk_neon(void *state, void *dst, const void *src, struct aegis128_state st = aegis128_load_state_neon(state); uint8x16_t msg; + preload_sbox(); + while (size >= AEGIS_BLOCK_SIZE) { msg = vld1q_u8(src) ^ st.v[1] ^ (st.v[2] & st.v[3]) ^ st.v[4]; st = aegis128_update_neon(st, msg); diff --git a/crypto/aegis128-neon.c b/crypto/aegis128-neon.c index c1c0a1686f67..751f9c195aa4 100644 --- a/crypto/aegis128-neon.c +++ b/crypto/aegis128-neon.c @@ -14,9 +14,15 @@ void crypto_aegis128_encrypt_chunk_neon(void *state, void *dst, const void *src, void crypto_aegis128_decrypt_chunk_neon(void *state, void *dst, const void *src, unsigned int size); +int aegis128_have_aes_insn __ro_after_init; + bool crypto_aegis128_have_simd(void) { - return cpu_have_feature(cpu_feature(AES)); + if (cpu_have_feature(cpu_feature(AES))) { + aegis128_have_aes_insn = 1; + return true; + } + return IS_ENABLED(CONFIG_ARM64); } void crypto_aegis128_update_simd(union aegis_block *state, const void *msg)