From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from NAM01-BY2-obe.outbound.protection.outlook.com (mail-by2nam01on0056.outbound.protection.outlook.com [104.47.34.56]) by dpdk.org (Postfix) with ESMTP id 26E096DB2 for ; Thu, 4 May 2017 08:57:19 +0200 (CEST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=CAVIUMNETWORKS.onmicrosoft.com; s=selector1-cavium-com; h=From:Date:Subject:Message-ID:Content-Type:MIME-Version; bh=1D67vAyyQT/7pltU4p9BDaSIUgBlR+XJPDO+7sXLlIA=; b=DRxOvn6z+PopVz5o6HYsVEtHlT5jzMILYvzFncbIiLZAHBuJQdKoGaG+8SxwRHYheGk6p4seCQBaMu/L6cRBUKiZOenXkur0R/oG7veAOKzCRSBIG5UGYWX1Zs5xiqfZyjKolfU3qv8ERzhpUmPlFAEVDY/iS0yaCyUBu/f1Ftk= Authentication-Results: intel.com; dkim=none (message not signed) header.d=none;intel.com; dmarc=none action=none header.from=caviumnetworks.com; Received: from 1scrb-1.caveonetworks.com (50.233.148.156) by BY2PR07MB2421.namprd07.prod.outlook.com (10.166.115.13) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256_P256) id 15.1.1061.12; Thu, 4 May 2017 06:57:15 +0000 From: Ashwin Sekhar T K To: cristian.dumitrescu@intel.com, thomas@monjalon.net, jasvinder.singh@intel.com, viktorin@rehivetech.com, jerin.jacob@caviumnetworks.com, jianbo.liu@linaro.org Cc: dev@dpdk.org, Ashwin Sekhar T K Date: Wed, 3 May 2017 23:57:01 -0700 Message-Id: <20170504065702.22404-3-ashwin.sekhar@caviumnetworks.com> X-Mailer: git-send-email 2.13.0.rc1 In-Reply-To: <20170504065702.22404-1-ashwin.sekhar@caviumnetworks.com> References: <20170427140618.35897-1-ashwin.sekhar@caviumnetworks.com> <20170504065702.22404-1-ashwin.sekhar@caviumnetworks.com> MIME-Version: 1.0 Content-Type: text/plain X-Originating-IP: [50.233.148.156] X-ClientProxiedBy: CY1PR03CA0028.namprd03.prod.outlook.com (10.174.128.38) To BY2PR07MB2421.namprd07.prod.outlook.com (10.166.115.13) X-MS-PublicTrafficType: Email X-MS-Office365-Filtering-Correlation-Id: 023dc5a4-0a6d-4325-2ade-08d492baccc4 X-Microsoft-Antispam: UriScan:; BCL:0; PCL:0; RULEID:(22001)(201703131423075)(201703031133081); SRVR:BY2PR07MB2421; X-Microsoft-Exchange-Diagnostics: 1; BY2PR07MB2421; 3:NVY0Yr6XCRMXwzqRSBXJqi35kxuPZWDaIg9Q+WdVcveyJlXAPH2O3ydfhkq0LqFTwQWmgJdy6qNMQBxR4bJrzpPE07DDR6AYdsONZ9KZEsDQHyCtrKHjW6XpD0fDpoxI4his+sYkuQspk6q+4bWz30iivk2PnOctq+gGFcuxs6OChRiWlfYREXjse2B8hRBKHmnsiZEP1ipM26/mEIxaxVv8SYfftUqnY+A+n/PP0L54uJvGj1qCp4GQ5fz7VZMbXzthUSJH0Kja9bdffxuM9pWe2H1tfQt7nVNKO4/PwcUuJg1Yt4gvjv9f54XzHmGmXnomv3np8gE+mOdCQXvwBg==; 25:/D/B534q4aR7VOFgzMn4MCtFGPFLJmCLYZohT7v1xEWQgMxSfbaALasPLinxafQ4hPQfxSzfQFcmaHNzcsS1WUg7yZmOeRtL6D5vwcfcQZ8oLRy6N2HodRhyjGE0d+6zCqG9g3x7pLHRwb9xxWq7V5gHOMgN8l+mEy4I2RwwgY7YDXcoo33B3HJ6A0mg9R7jew4mexRJIyCYtQNEeJYvn9jFmeqt2YOSMn5zz/b83ePFWz7udcLXmw2h0gATVkhkqR7PdqWFopReAZuHz8U1i0/qz56rFf2Phuol2CWQSlMhj82fBzyrWabsxky3HOSlqQN3WE20jL02esn5NNHam/1TLB2VVtt8CZpbEarGppzx5or0+8+4l0Q8lwSZFGL4JmsMDKWOjSBwKx//hU1TD3sfXzKN2q6LZT754XAdp/yZeiWaXu8BCQ2zg0iSm1HEPxTkX3vsW8Ka5IJ69owj6w== X-Microsoft-Exchange-Diagnostics: 1; BY2PR07MB2421; 31:XXylDundXqRJ3YEOYXwlZ32pets5DNRC3/2xy4pwcLyYfm4F62mCjAco2sP5K7IZCLtf8GsEy3IQLDZdjRi9/HP+2U+YytJfeQs1Cn78sgZFWR/Ce5pJvrr1cZebK1BqrUEzhAdkGfOurI65U7r2oUdhUplv155hOGlf5FW/gksbveOqcGz9Wh63Gv0+/VYC2+HKb2OjIfdSjs4x8HI0OmVsTwZalTE9EAgNjczGrXo=; 20:vpWPzuyoOSqJhww4Hr50nEvfgXNE35y4kNHJMPqoEdBAX8gy/7aTL8zp1r3xRX/zjAILt2ZZhuzI/6L9GKx2xAHQdfqRjffg+aKppRE7Ut4+xb27PisoHeEQt9xprXgGU8PnPpf0fl9vtRfDkOBPoF7sZexKZTH6Yrp+GTYg+2/LCt59Mtmsjz9XEvGP7iJ+rKdjoaMOiSXDFb7ESNk3/ySVllepAzpCLqv2xkUue1PeBfKMu2tlVpjWrKGlbSKlxMfRVzrGjE9KYw277A70ikzdKffn8qDtLi32hsRl6gtfGmGzcVG9SXW1aUthfZj1nyZlWINlgelbjUOLoub28sOGmbz8NTBS4SyMfLhZ37slIhFuof0sRhyYjnle5VINzEXbwZdGdOH2XdLuEAZQcEKjgggaKzdzGy45pMo4uQHltJib1+GF4A1uY7woza4T69J/NrSAKRwMfOvOuP0C0ksZNM/avtnunE75674jcY76MT8kGrJM+kc2kGufK98SDpPWf0fpIbfll99Y0KSbSQvZ9ikIlH7LlFcxCE+N+SRuWw9eAWs+5gS0aDcBAUy6IiPZulMMhJCvy86VLx0woAQ/YdiW0uNBzjDJu/+ciE8= X-Microsoft-Antispam-PRVS: X-Exchange-Antispam-Report-Test: UriScan:; X-Exchange-Antispam-Report-CFA-Test: BCL:0; PCL:0; RULEID:(6040450)(601004)(2401047)(5005006)(8121501046)(10201501046)(3002001)(93006095)(6041248)(20161123555025)(20161123562025)(20161123558100)(20161123560025)(20161123564025)(201703131423075)(201702281528075)(201703061421075)(201703061406153)(6072148); SRVR:BY2PR07MB2421; BCL:0; PCL:0; RULEID:; SRVR:BY2PR07MB2421; X-Microsoft-Exchange-Diagnostics: 1; BY2PR07MB2421; 4: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 X-Forefront-PRVS: 02973C87BC X-Forefront-Antispam-Report: SFV:NSPM; SFS:(10009020)(4630300001)(979002)(6009001)(39850400002)(39400400002)(39840400002)(39410400002)(39450400003)(76176999)(8676002)(50986999)(5003940100001)(50226002)(6486002)(36756003)(48376002)(50466002)(3846002)(81166006)(6116002)(1076002)(66066001)(6506006)(47776003)(189998001)(25786009)(53416004)(42186005)(4326008)(2906002)(33646002)(53936002)(5660300001)(305945005)(478600001)(7736002)(110136004)(107886003)(2950100002)(6666003)(42882006)(6512007)(38730400002)(575784001)(969003)(989001)(999001)(1009001)(1019001); DIR:OUT; SFP:1101; SCL:1; SRVR:BY2PR07MB2421; H:1scrb-1.caveonetworks.com; FPR:; SPF:None; MLV:ovrnspm; PTR:InfoNoRecords; LANG:en; X-Microsoft-Exchange-Diagnostics: =?us-ascii?Q?1; BY2PR07MB2421; 23:hv5BSHWsek1+cu1C1LoQhBS0O0qnBP3lxMpD5Z477?= =?us-ascii?Q?PwPFa76nRpnZ+v5S7aS7f+urtXnP8y9ZU5btQsF+JFoyR05Hvu0eHo/iBoQX?= =?us-ascii?Q?Y/X8VuXMBdbtpttrwGsxAAMFZUrYmvfw4zkxb1puiQvONtEiUSZjJaYlH9r+?= =?us-ascii?Q?K5TUm9OxsuaOZuOqD3uz72vMZgvq2TjDwYvJuQxiEzjsCTTWI6nlWMir/J4r?= =?us-ascii?Q?wnYhfRrjxBUfShz2Yya4NgJj8HiIwip18PHTGznvYFjpxTUWxKCvGQILvGa4?= =?us-ascii?Q?gPv+szwMduzMU3EBrIlNrHNgeWriZpPzuv5iWq5LkdArUwg+MuGpyMIbNtmQ?= =?us-ascii?Q?TJad7h/cqZoxwtbx/HGInpJUr3feoIR2+CgiMr+T0UoRSUMkE/22hnPtCJrK?= =?us-ascii?Q?EJjx1kDGfHPvgQ+gzc2cKwNPabjzlqkPyXYiU8WPvMle2WneeIi/4TpGRw/V?= =?us-ascii?Q?DnbdsaUzBw6ZIEeRSmZJzklROj8MuGWHYHpR4+Hq3k9LEbfYk8VSjLHeD4m4?= =?us-ascii?Q?zGRQ/0V5WsW3JGYkFCHjZghTL03u67ns2YEkDDdqMNgChlIy1+/CpItX5m7F?= =?us-ascii?Q?++rw9uz7tEkShTdHsOOojQ1eGXa2qJzVUnXc4NMGNCoaawV1N7Vfn2Pfbi9I?= =?us-ascii?Q?dLcfaWULvupm323U4Ao2nSca4l3yUjFNmpj1hYHUfracLncqOwR/4jyW3zhz?= =?us-ascii?Q?K/iKnCsPlvOobb8b+NprLxqxC3p8/nBywIaYx7ruxLPvXIbDr82adE45drUV?= =?us-ascii?Q?hORDPUYtqhIl6RmeCGY/T7GnJ3tgl6ImlLhu9ScvwoPPAAbiu+9JD3SzfgaU?= =?us-ascii?Q?PbwXgcFOhmQmrDoAiITLoPJ3Tt/Tvhe6K4DqfnHAw1IX7pecs2fpbiD3k3bO?= =?us-ascii?Q?JbqrDI9MTu+sXFC0DFln+0CGLgQ90/mOPTBwucHzeebGJQesFe0psDj7ksQX?= =?us-ascii?Q?1xuYznx2fe0Jm8WAEW4FEH20U4hCRzrcxEuTIELxoNr9ypSaAFu8P92AcsL1?= =?us-ascii?Q?FiXfrE83yAn7JbYA+LFIwiLIvpVhpfOR7P+JnORAwpj7dyo9RGYMGBIa2e+y?= =?us-ascii?Q?edMJwZiOjJU13I2TqtPK1ulpqaaDBwgOy7f8toPO//pnUTHOt0h80TMyA0eb?= =?us-ascii?Q?jCKCQVgtwo/GxNfHOI9EiawcF0Dcc2m83EWzK/IZwphPzmg6kcKnn591g3Oz?= =?us-ascii?Q?/oGD1dydmdOXfk=3D?= X-Microsoft-Exchange-Diagnostics: 1; BY2PR07MB2421; 6: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; 5:1yGhA/0YF05IdeAj1aoCWQDL+XH+6aS1+i0HeJ8t25qIveEU+GhMnPAlWrilniOmmoD7SZHrPOZ6jwWpxIt0d4KZPYLMHgy9SxEFdSziVMPJ5H2naEwcnLwnls6P3+QaO0OSganYyyXRV1R2d3x5yA==; 24:v1iP2r5Mr6eR+ugaPGYpN6mrTvF4ZEDuRfh1rPx3IVfNmtEM7I4oAHrXpwHcurCqQiV+d+xu6FzyfLHcRMfAF4c2pomM0i4DspLfU5UGMIk= SpamDiagnosticOutput: 1:99 SpamDiagnosticMetadata: NSPM X-Microsoft-Exchange-Diagnostics: 1; BY2PR07MB2421; 7:oRGZEwF4J5R4WQxYx45Coi0v0sIeB0q02xW61oQ+4Zd9jmky3vdbBPA8cJ5vA3yyyoGDEOnG9VX4iddHJ13DaVoh2Zt+2X1qtqm5m5BDoq1RdTAgeICRIP5HrMRourwwlTxkqsS0KY/3VuhKwb0UNDIwDdS4coXpz/FmENG9rqoGsuvUNQ4kb0J9oP0A4BAA4qGRzDYwjaOnC34QtZopx01g3hv1nyCC+2v6sTZcUDCWv2TOLxKGn9IL40NKIqctLHLqbCTJG408Lok27bzHQtOjcGyKU7quqRVkgfBy+eQRENLgZZGvyc6/n4t7jZXCzG2gmS7DDzD8tLSPBPCW1g== X-OriginatorOrg: caviumnetworks.com X-MS-Exchange-CrossTenant-OriginalArrivalTime: 04 May 2017 06:57:15.7108 (UTC) X-MS-Exchange-CrossTenant-FromEntityHeader: Hosted X-MS-Exchange-Transport-CrossTenantHeadersStamped: BY2PR07MB2421 Subject: [dpdk-dev] [PATCH v3 3/4] net: add arm64 neon version of CRC compute APIs X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 04 May 2017 06:57:20 -0000 Added CRC compute APIs for arm64 utilizing the pmull capability Added new file net_crc_neon.h to hold the arm64 pmull CRC implementation Verified the changes with crc_autotest unit test case Signed-off-by: Ashwin Sekhar T K --- v2: * Fixed merge conflict in MAINTAINERS v3: * Moved feature detection changes and GCC_VERSION definition changes to separate commit * Replaced usage of assert() with RTE_ASSERT() * Made the comments in rte_vect.h more positive in sense MAINTAINERS | 1 + lib/librte_eal/common/include/arch/arm/rte_vect.h | 29 ++ lib/librte_net/net_crc_neon.h | 357 ++++++++++++++++++++++ lib/librte_net/rte_net_crc.c | 32 +- lib/librte_net/rte_net_crc.h | 2 + 5 files changed, 416 insertions(+), 5 deletions(-) create mode 100644 lib/librte_net/net_crc_neon.h diff --git a/MAINTAINERS b/MAINTAINERS index b6495d2..66d64c2 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -147,6 +147,7 @@ F: lib/librte_eal/common/include/arch/arm/*_64.h F: lib/librte_acl/acl_run_neon.* F: lib/librte_lpm/rte_lpm_neon.h F: lib/librte_hash/rte*_arm64.h +F: lib/librte_net/net_crc_neon.h F: drivers/net/ixgbe/ixgbe_rxtx_vec_neon.c F: drivers/net/i40e/i40e_rxtx_vec_neon.c F: drivers/net/virtio/virtio_rxtx_simple_neon.c diff --git a/lib/librte_eal/common/include/arch/arm/rte_vect.h b/lib/librte_eal/common/include/arch/arm/rte_vect.h index 4107c99..f28a19d 100644 --- a/lib/librte_eal/common/include/arch/arm/rte_vect.h +++ b/lib/librte_eal/common/include/arch/arm/rte_vect.h @@ -34,7 +34,9 @@ #define _RTE_VECT_ARM_H_ #include + #include "generic/rte_vect.h" +#include "rte_debug.h" #include "arm_neon.h" #ifdef __cplusplus @@ -78,6 +80,33 @@ vqtbl1q_u8(uint8x16_t a, uint8x16_t b) } #endif +#if (GCC_VERSION < 70000) +/* NEON intrinsic vreinterpretq_u64_p128() is supported since GCC version 7 */ +static inline uint64x2_t +vreinterpretq_u64_p128(poly128_t x) +{ + return (uint64x2_t)x; +} + +/* NEON intrinsic vreinterpretq_p64_u64() is supported since GCC version 7 */ +static inline poly64x2_t +vreinterpretq_p64_u64(uint64x2_t x) +{ + return (poly64x2_t)x; +} + +/* NEON intrinsic vgetq_lane_p64() is supported since GCC version 7 */ +static inline poly64_t +vgetq_lane_p64(poly64x2_t x, const int lane) +{ + RTE_ASSERT(lane >= 0 && lane <= 1); + + poly64_t *p = (poly64_t *)&x; + + return p[lane]; +} +#endif + #ifdef __cplusplus } #endif diff --git a/lib/librte_net/net_crc_neon.h b/lib/librte_net/net_crc_neon.h new file mode 100644 index 0000000..05120a7 --- /dev/null +++ b/lib/librte_net/net_crc_neon.h @@ -0,0 +1,357 @@ +/* + * BSD LICENSE + * + * Copyright (C) Cavium networks Ltd. 2017. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * + * * Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * * Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in + * the documentation and/or other materials provided with the + * distribution. + * * Neither the name of Cavium networks nor the names of its + * contributors may be used to endorse or promote products derived + * from this software without specific prior written permission. + * + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, + * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT + * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE + * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + */ + +#ifndef _NET_CRC_NEON_H_ +#define _NET_CRC_NEON_H_ + +#include +#include +#include +#include + +#ifdef __cplusplus +extern "C" { +#endif + +/** PMULL CRC computation context structure */ +struct crc_pmull_ctx { + uint64x2_t rk1_rk2; + uint64x2_t rk5_rk6; + uint64x2_t rk7_rk8; +}; + +struct crc_pmull_ctx crc32_eth_pmull __rte_aligned(16); +struct crc_pmull_ctx crc16_ccitt_pmull __rte_aligned(16); + +static inline uint8x16_t +extract_vector(uint8x16_t v0, uint8x16_t v1, const int n) +{ + switch (n) { + case 0: return vextq_u8(v0, v1, 0); + case 1: return vextq_u8(v0, v1, 1); + case 2: return vextq_u8(v0, v1, 2); + case 3: return vextq_u8(v0, v1, 3); + case 4: return vextq_u8(v0, v1, 4); + case 5: return vextq_u8(v0, v1, 5); + case 6: return vextq_u8(v0, v1, 6); + case 7: return vextq_u8(v0, v1, 7); + case 8: return vextq_u8(v0, v1, 8); + case 9: return vextq_u8(v0, v1, 9); + case 10: return vextq_u8(v0, v1, 10); + case 11: return vextq_u8(v0, v1, 11); + case 12: return vextq_u8(v0, v1, 12); + case 13: return vextq_u8(v0, v1, 13); + case 14: return vextq_u8(v0, v1, 14); + case 15: return vextq_u8(v0, v1, 15); + } + return v1; +} + +/** + * Shifts right 128 bit register by specified number of bytes + * + * @param reg 128 bit value + * @param num number of bytes to shift reg by (0-16) + * + * @return reg << (num * 8) + */ +static inline uint64x2_t +shift_bytes_right(uint64x2_t reg, const unsigned int num) +{ + /* Right Shift */ + return vreinterpretq_u64_u8(extract_vector( + vreinterpretq_u8_u64(reg), + vdupq_n_u8(0), + num)); +} + +/** + * Shifts left 128 bit register by specified number of bytes + * + * @param reg 128 bit value + * @param num number of bytes to shift reg by (0-16) + * + * @return reg << (num * 8) + */ +static inline uint64x2_t +shift_bytes_left(uint64x2_t reg, const unsigned int num) +{ + /* Left Shift */ + return vreinterpretq_u64_u8(extract_vector( + vdupq_n_u8(0), + vreinterpretq_u8_u64(reg), + 16 - num)); +} + +/** + * @brief Performs one folding round + * + * Logically function operates as follows: + * DATA = READ_NEXT_16BYTES(); + * F1 = LSB8(FOLD) + * F2 = MSB8(FOLD) + * T1 = CLMUL(F1, RK1) + * T2 = CLMUL(F2, RK2) + * FOLD = XOR(T1, T2, DATA) + * + * @param data_block 16 byte data block + * @param precomp precomputed rk1 constanst + * @param fold running 16 byte folded data + * + * @return New 16 byte folded data + */ +static inline uint64x2_t +crcr32_folding_round(uint64x2_t data_block, uint64x2_t precomp, + uint64x2_t fold) +{ + uint64x2_t tmp0 = vreinterpretq_u64_p128(vmull_p64( + vgetq_lane_p64(vreinterpretq_p64_u64(fold), 1), + vgetq_lane_p64(vreinterpretq_p64_u64(precomp), 0))); + + uint64x2_t tmp1 = vreinterpretq_u64_p128(vmull_p64( + vgetq_lane_p64(vreinterpretq_p64_u64(fold), 0), + vgetq_lane_p64(vreinterpretq_p64_u64(precomp), 1))); + + return veorq_u64(tmp1, veorq_u64(data_block, tmp0)); +} + +/** + * Performs reduction from 128 bits to 64 bits + * + * @param data128 128 bits data to be reduced + * @param precomp rk5 and rk6 precomputed constants + * + * @return data reduced to 64 bits + */ +static inline uint64x2_t +crcr32_reduce_128_to_64(uint64x2_t data128, + uint64x2_t precomp) +{ + uint64x2_t tmp0, tmp1, tmp2; + + /* 64b fold */ + tmp0 = vreinterpretq_u64_p128(vmull_p64( + vgetq_lane_p64(vreinterpretq_p64_u64(data128), 0), + vgetq_lane_p64(vreinterpretq_p64_u64(precomp), 0))); + tmp1 = shift_bytes_right(data128, 8); + tmp0 = veorq_u64(tmp0, tmp1); + + /* 32b fold */ + tmp2 = shift_bytes_left(tmp0, 4); + tmp1 = vreinterpretq_u64_p128(vmull_p64( + vgetq_lane_p64(vreinterpretq_p64_u64(tmp2), 0), + vgetq_lane_p64(vreinterpretq_p64_u64(precomp), 1))); + + return veorq_u64(tmp1, tmp0); +} + +/** + * Performs Barret's reduction from 64 bits to 32 bits + * + * @param data64 64 bits data to be reduced + * @param precomp rk7 precomputed constant + * + * @return data reduced to 32 bits + */ +static inline uint32_t +crcr32_reduce_64_to_32(uint64x2_t data64, + uint64x2_t precomp) +{ + static uint32_t mask1[4] __rte_aligned(16) = { + 0xffffffff, 0xffffffff, 0x00000000, 0x00000000 + }; + static uint32_t mask2[4] __rte_aligned(16) = { + 0x00000000, 0xffffffff, 0xffffffff, 0xffffffff + }; + uint64x2_t tmp0, tmp1, tmp2; + + tmp0 = vandq_u64(data64, vld1q_u64((uint64_t *)mask2)); + + tmp1 = vreinterpretq_u64_p128(vmull_p64( + vgetq_lane_p64(vreinterpretq_p64_u64(tmp0), 0), + vgetq_lane_p64(vreinterpretq_p64_u64(precomp), 0))); + tmp1 = veorq_u64(tmp1, tmp0); + tmp1 = vandq_u64(tmp1, vld1q_u64((uint64_t *)mask1)); + + tmp2 = vreinterpretq_u64_p128(vmull_p64( + vgetq_lane_p64(vreinterpretq_p64_u64(tmp1), 0), + vgetq_lane_p64(vreinterpretq_p64_u64(precomp), 1))); + tmp2 = veorq_u64(tmp2, tmp1); + tmp2 = veorq_u64(tmp2, tmp0); + + return vgetq_lane_u32(vreinterpretq_u32_u64(tmp2), 2); +} + +static inline uint32_t +crc32_eth_calc_pmull( + const uint8_t *data, + uint32_t data_len, + uint32_t crc, + const struct crc_pmull_ctx *params) +{ + uint64x2_t temp, fold, k; + uint32_t n; + + /* Get CRC init value */ + temp = vreinterpretq_u64_u32(vsetq_lane_u32(crc, vmovq_n_u32(0), 0)); + + /** + * Folding all data into single 16 byte data block + * Assumes: fold holds first 16 bytes of data + */ + if (unlikely(data_len < 32)) { + if (unlikely(data_len == 16)) { + /* 16 bytes */ + fold = vld1q_u64((const uint64_t *)data); + fold = veorq_u64(fold, temp); + goto reduction_128_64; + } + + if (unlikely(data_len < 16)) { + /* 0 to 15 bytes */ + uint8_t buffer[16] __rte_aligned(16); + + memset(buffer, 0, sizeof(buffer)); + memcpy(buffer, data, data_len); + + fold = vld1q_u64((uint64_t *)buffer); + fold = veorq_u64(fold, temp); + if (unlikely(data_len < 4)) { + fold = shift_bytes_left(fold, 8 - data_len); + goto barret_reduction; + } + fold = shift_bytes_left(fold, 16 - data_len); + goto reduction_128_64; + } + /* 17 to 31 bytes */ + fold = vld1q_u64((const uint64_t *)data); + fold = veorq_u64(fold, temp); + n = 16; + k = params->rk1_rk2; + goto partial_bytes; + } + + /** At least 32 bytes in the buffer */ + /** Apply CRC initial value */ + fold = vld1q_u64((const uint64_t *)data); + fold = veorq_u64(fold, temp); + + /** Main folding loop - the last 16 bytes is processed separately */ + k = params->rk1_rk2; + for (n = 16; (n + 16) <= data_len; n += 16) { + temp = vld1q_u64((const uint64_t *)&data[n]); + fold = crcr32_folding_round(temp, k, fold); + } + +partial_bytes: + if (likely(n < data_len)) { + uint64x2_t last16, a, b, mask; + uint32_t rem = data_len & 15; + + last16 = vld1q_u64((const uint64_t *)&data[data_len - 16]); + a = shift_bytes_left(fold, 16 - rem); + b = shift_bytes_right(fold, rem); + mask = shift_bytes_left(vdupq_n_u64(-1), 16 - rem); + b = vorrq_u64(b, vandq_u64(mask, last16)); + + /* k = rk1 & rk2 */ + temp = vreinterpretq_u64_p128(vmull_p64( + vgetq_lane_p64(vreinterpretq_p64_u64(a), 1), + vgetq_lane_p64(vreinterpretq_p64_u64(k), 0))); + fold = vreinterpretq_u64_p128(vmull_p64( + vgetq_lane_p64(vreinterpretq_p64_u64(a), 0), + vgetq_lane_p64(vreinterpretq_p64_u64(k), 1))); + fold = veorq_u64(fold, temp); + fold = veorq_u64(fold, b); + } + + /** Reduction 128 -> 32 Assumes: fold holds 128bit folded data */ +reduction_128_64: + k = params->rk5_rk6; + fold = crcr32_reduce_128_to_64(fold, k); + +barret_reduction: + k = params->rk7_rk8; + n = crcr32_reduce_64_to_32(fold, k); + + return n; +} + +static inline void +rte_net_crc_neon_init(void) +{ + /* Initialize CRC16 data */ + uint64_t ccitt_k1_k2[2] = {0x189aeLLU, 0x8e10LLU}; + uint64_t ccitt_k5_k6[2] = {0x189aeLLU, 0x114aaLLU}; + uint64_t ccitt_k7_k8[2] = {0x11c581910LLU, 0x10811LLU}; + + /* Initialize CRC32 data */ + uint64_t eth_k1_k2[2] = {0xccaa009eLLU, 0x1751997d0LLU}; + uint64_t eth_k5_k6[2] = {0xccaa009eLLU, 0x163cd6124LLU}; + uint64_t eth_k7_k8[2] = {0x1f7011640LLU, 0x1db710641LLU}; + + /** Save the params in context structure */ + crc16_ccitt_pmull.rk1_rk2 = vld1q_u64(ccitt_k1_k2); + crc16_ccitt_pmull.rk5_rk6 = vld1q_u64(ccitt_k5_k6); + crc16_ccitt_pmull.rk7_rk8 = vld1q_u64(ccitt_k7_k8); + + /** Save the params in context structure */ + crc32_eth_pmull.rk1_rk2 = vld1q_u64(eth_k1_k2); + crc32_eth_pmull.rk5_rk6 = vld1q_u64(eth_k5_k6); + crc32_eth_pmull.rk7_rk8 = vld1q_u64(eth_k7_k8); +} + +static inline uint32_t +rte_crc16_ccitt_neon_handler(const uint8_t *data, + uint32_t data_len) +{ + return (uint16_t)~crc32_eth_calc_pmull(data, + data_len, + 0xffff, + &crc16_ccitt_pmull); +} + +static inline uint32_t +rte_crc32_eth_neon_handler(const uint8_t *data, + uint32_t data_len) +{ + return ~crc32_eth_calc_pmull(data, + data_len, + 0xffffffffUL, + &crc32_eth_pmull); +} + +#ifdef __cplusplus +} +#endif + +#endif /* _NET_CRC_NEON_H_ */ diff --git a/lib/librte_net/rte_net_crc.c b/lib/librte_net/rte_net_crc.c index e8326fe..be65f34 100644 --- a/lib/librte_net/rte_net_crc.c +++ b/lib/librte_net/rte_net_crc.c @@ -43,10 +43,16 @@ && defined(RTE_MACHINE_CPUFLAG_SSE4_2) \ && defined(RTE_MACHINE_CPUFLAG_PCLMULQDQ) #define X86_64_SSE42_PCLMULQDQ 1 +#elif defined(RTE_ARCH_ARM64) +#if defined(RTE_MACHINE_CPUFLAG_PMULL) +#define ARM64_NEON_PMULL 1 +#endif #endif #ifdef X86_64_SSE42_PCLMULQDQ #include +#elif defined(ARM64_NEON_PMULL) +#include #endif /* crc tables */ @@ -74,6 +80,11 @@ static rte_net_crc_handler handlers_sse42[] = { [RTE_NET_CRC16_CCITT] = rte_crc16_ccitt_sse42_handler, [RTE_NET_CRC32_ETH] = rte_crc32_eth_sse42_handler, }; +#elif defined(ARM64_NEON_PMULL) +static rte_net_crc_handler handlers_neon[] = { + [RTE_NET_CRC16_CCITT] = rte_crc16_ccitt_neon_handler, + [RTE_NET_CRC32_ETH] = rte_crc32_eth_neon_handler, +}; #endif /** @@ -162,14 +173,20 @@ void rte_net_crc_set_alg(enum rte_net_crc_alg alg) { switch (alg) { - case RTE_NET_CRC_SSE42: #ifdef X86_64_SSE42_PCLMULQDQ + case RTE_NET_CRC_SSE42: handlers = handlers_sse42; -#else - alg = RTE_NET_CRC_SCALAR; break; +#elif defined(ARM64_NEON_PMULL) + case RTE_NET_CRC_NEON: + if (rte_cpu_get_flag_enabled(RTE_CPUFLAG_PMULL)) { + handlers = handlers_neon; + break; + } + //-fallthrough #endif case RTE_NET_CRC_SCALAR: + //-fallthrough default: handlers = handlers_scalar; break; @@ -199,8 +216,13 @@ rte_net_crc_init(void) rte_net_crc_scalar_init(); #ifdef X86_64_SSE42_PCLMULQDQ - alg = RTE_NET_CRC_SSE42; - rte_net_crc_sse42_init(); + alg = RTE_NET_CRC_SSE42; + rte_net_crc_sse42_init(); +#elif defined(ARM64_NEON_PMULL) + if (rte_cpu_get_flag_enabled(RTE_CPUFLAG_PMULL)) { + alg = RTE_NET_CRC_NEON; + rte_net_crc_neon_init(); + } #endif rte_net_crc_set_alg(alg); diff --git a/lib/librte_net/rte_net_crc.h b/lib/librte_net/rte_net_crc.h index d22286c..d01cf4b 100644 --- a/lib/librte_net/rte_net_crc.h +++ b/lib/librte_net/rte_net_crc.h @@ -57,6 +57,7 @@ enum rte_net_crc_type { enum rte_net_crc_alg { RTE_NET_CRC_SCALAR = 0, RTE_NET_CRC_SSE42, + RTE_NET_CRC_NEON, }; /** @@ -68,6 +69,7 @@ enum rte_net_crc_alg { * This parameter is used to select the CRC implementation version. * - RTE_NET_CRC_SCALAR * - RTE_NET_CRC_SSE42 (Use 64-bit SSE4.2 intrinsic) + * - RTE_NET_CRC_NEON (Use ARM Neon intrinsic) */ void rte_net_crc_set_alg(enum rte_net_crc_alg alg); -- 2.7.4