From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from NAM01-BN3-obe.outbound.protection.outlook.com (mail-bn3nam01on0055.outbound.protection.outlook.com [104.47.33.55]) by dpdk.org (Postfix) with ESMTP id CFE85A2F for ; Thu, 27 Apr 2017 16:06:38 +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=DsxygULx/C1wgHPJ2algWwoO+fgoM0tCwnPuObNRmvQ=; b=AjIoelvfTALdZC7Xz8P/XCQLCulwqlwDzeTEBWSd6KY7wRqoOYugv51TAnqJor+jymK08u5rrpiafvqmXgbejvaY5JbC762A+bbWqtHvkaWDN+PAg9sfW6y0wmclIKe4MlwuvVjFMhWRjLwcDhSfmz/mcqFCk1zBSPJ5IQl1aS0= Authentication-Results: monjalon.net; dkim=none (message not signed) header.d=none;monjalon.net; dmarc=none action=none header.from=caviumnetworks.com; Received: from 1scrb-1.caveonetworks.com (50.233.148.156) by BY2PR07MB2423.namprd07.prod.outlook.com (10.166.115.15) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256_P256) id 15.1.1061.12; Thu, 27 Apr 2017 14:06:30 +0000 From: Ashwin Sekhar T K To: 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: Thu, 27 Apr 2017 07:06:17 -0700 Message-Id: <20170427140618.35897-1-ashwin.sekhar@caviumnetworks.com> X-Mailer: git-send-email 2.13.0.rc1 MIME-Version: 1.0 Content-Type: text/plain X-Originating-IP: [50.233.148.156] X-ClientProxiedBy: MWHPR01CA0024.prod.exchangelabs.com (10.168.201.162) To BY2PR07MB2423.namprd07.prod.outlook.com (10.166.115.15) X-MS-PublicTrafficType: Email X-MS-Office365-Filtering-Correlation-Id: efba9cb4-1109-49ce-9d74-08d48d769b15 X-Microsoft-Antispam: UriScan:; BCL:0; PCL:0; RULEID:(22001)(201703131423075)(201703031133081); SRVR:BY2PR07MB2423; X-Microsoft-Exchange-Diagnostics: 1; BY2PR07MB2423; 3:8dneoO5QjtDou9xyenQF3cPrCeW6ekyjBqF6H/BF+PERaTpmtxnW8XnQ+H8Puztl6V7U58arpXNO0V/Y17Jb6ScuFGrvX3Zx2tVGa6K2NYMnE9UdpM7s78Jq1udMzgqCUYrOObtqYEvdSVIrLPMWuFVrvDc7FAF9NV29VvBhIF8OCzrI03qNuxDRAV7FlbucrOHuxjuVTUuHXc7I4iND/uG6hLVn0alBqKOBWf7bCsjXl79MgtDLF8sl1b57TcLAgvUFuI6nDA04CBLQUFE2rIE0oGajT9sKPV5LI92GDM+R0ofiibwo1r3G3UUUZKxz/2tMTG65zH9x1FD1S0Bmfg==; 25:+B5DK/S0apsWB0bEb2AFyQweru1NGFXoeoIGg0XKYeoRVXqcmWqpMo3Jo9p4yLi0+6bUTmK3ES9tLE7P5oPf7J9kalnVCuTct8wOIHPBW6UhUxlhDcsKvjUvdwIra+7a3Xvs0fsandekS2gu0YdfB0Kw3MO4pxxMM8LlIGEeEb0Tc6ZvqrdUmR2Z923CZMrTJj0WuiuUf1a4Gng31ndMWhy3qTQljrMCeP2ZornKgimBg5wgd5roIUIXsFxvWijGGcRapzuopG7PDiHGvu5MprVP1x7UkWqiD6OCP79dzU9XZwgVNJpU2himRZwjPIjo6nfBLBkCSkdaibOa1V3GmsGovqrXMLQr/5EuMOI+YvDq7Bzn4SJJQS4YXglQ26Bm38aJhJWR8W0fhTLQNKq2fJsaXAnPX0TIIspeNwpqn4zIsYQCnHYUsWyvCiWOL6afJqET+hp7cPCYM8K+kvvK1w== X-Microsoft-Exchange-Diagnostics: 1; BY2PR07MB2423; 31:fO2lJe2U8qJi5DuKUD33DthwqtB/pd1oK76dvwPcyEms/xuhzRw7OmRRZwW7PlFgZdNhQ03F+mYvz7Cd6DYiDpNsLrxIDt6kDNAHBxwNzpIcL4GkA3miRb3WrKvGtNTNvgoS77yXExpn0XJpEntlVQ7loj98q4V+8buupdWfZcnQgsbb+kwDonFVWFFFvkOS+sNo+pnFAfsh3WXpC83Y8mEjrX3Tp9SzBrW8uEqlHzY=; 20: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 X-Microsoft-Antispam-PRVS: X-Exchange-Antispam-Report-Test: UriScan:; X-Exchange-Antispam-Report-CFA-Test: BCL:0; PCL:0; RULEID:(6040450)(601004)(2401047)(8121501046)(5005006)(93006095)(10201501046)(3002001)(6041248)(201703131423075)(201702281528075)(201703061421075)(201703061406153)(20161123555025)(20161123564025)(20161123558100)(20161123562025)(20161123560025)(6072148); SRVR:BY2PR07MB2423; BCL:0; PCL:0; RULEID:; SRVR:BY2PR07MB2423; X-Microsoft-Exchange-Diagnostics: 1; BY2PR07MB2423; 4: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 X-Forefront-PRVS: 029097202E X-Forefront-Antispam-Report: SFV:NSPM; SFS:(10009020)(4630300001)(6009001)(39450400003)(39840400002)(39850400002)(39400400002)(39410400002)(33646002)(6506006)(8676002)(2906002)(53416004)(42882006)(6486002)(6512007)(575784001)(6666003)(81166006)(53936002)(66066001)(50226002)(47776003)(36756003)(50466002)(42186005)(5660300001)(305945005)(3846002)(6116002)(1076002)(107886003)(38730400002)(110136004)(48376002)(25786009)(4326008)(50986999)(5003940100001)(189998001); DIR:OUT; SFP:1101; SCL:1; SRVR:BY2PR07MB2423; H:1scrb-1.caveonetworks.com; FPR:; SPF:None; MLV:sfv; LANG:en; X-Microsoft-Exchange-Diagnostics: =?us-ascii?Q?1; BY2PR07MB2423; 23:kgH7ivjCWRQu4T+QYlU6yBZOP2TiGkZUNyN95C943?= =?us-ascii?Q?ptkPGgvMVCqJYZuD9dAibL8re6S36Bar1ZkCrPwq4VPgt+x2hjj70/SfMDxZ?= =?us-ascii?Q?+Vu7g4RNGXeRnFwTA9rXe99fYtRjGZNRR7JVuSigcnWtdtln8GlMAVCQpZiN?= =?us-ascii?Q?UXkOQLPY9Zu+l7/sHj4ecHl4Rgic+fIhRRHyWLRJBAbgyNaKCes9t+lqN1Ji?= =?us-ascii?Q?fOyeUg6gomglDqFsYramHz4pg+hBj1Ovsyj73ZrUKBdwOYYJkmQ2UDB+j3Hm?= =?us-ascii?Q?eLI4hXuSkDlTS/ORvN036qIItkDnIDlSYPuzVJ19mqSWCKek9dvAp0oC/vXe?= =?us-ascii?Q?axMm+vMXG3K+t91uTtIqOqTmAIIQNEHq/G3F1bvBGHUt4QIUpI8jWFARpRgb?= =?us-ascii?Q?ddZWpzRLKsV2JqaDN7dagjjzir9asdUJSEvooS8BTRTWolTd0MpBOmE1IoTK?= =?us-ascii?Q?og7p4Goq2VUm4XJTXoFkDHKn+VJv5T+wi1YLNjvrmB+JrPW184Dyjt4ZiUMZ?= =?us-ascii?Q?xtVhNqBzsfLzO3tXT6Lik1hbRYCZo8YPmVR2Y30OhQJpUL9N9EH7tNZQXzWV?= =?us-ascii?Q?PZjLrNcvHnXR2fDwy+f8IyYNAKqy0Xa7LWaB0wPD2TC2STmlJVBxb9JSmQjk?= =?us-ascii?Q?xP0tOOuvtMkTHG9bWUbHBIvLJSHgiQp4Xw/D8OnXs1hMKsDnaRnN0/aAgHIP?= =?us-ascii?Q?A9v90YjoP++UPXpqgD6m3N4t/Zp+ozyVOUMDmyAN87hKusP6BqukY5caTo5U?= =?us-ascii?Q?fH28PzWjBXDikGXH1YyQgPWypp7AYJvo8lE70V54DmoaSxM1cr+eIMkoF1/U?= =?us-ascii?Q?GqaJoaed/nhTVmuFddZhx7qE/DYVk5Lj28zi7AreapZXZZtTofbf1/sYvCC9?= =?us-ascii?Q?5vtD5sMaJnKiVt7HKUVn13WMGMjtKHHSg6dzxmEM7lIZGeTwe1cGQpoqPGRV?= =?us-ascii?Q?gzZmYk415V+rn5JRwkSVbx1UtwcUgpUO0WTVdp2ubtOaSFQPjM+Fq0hJPkZj?= =?us-ascii?Q?xfafrkaJUwc3r0l/ygSAGpn?= X-Microsoft-Exchange-Diagnostics: 1; BY2PR07MB2423; 6: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; 5:uPweKzBpVy3DTRNrD+XpI2A8vUKccvYQiiCWaRmPKwTGdyf2TWHJ4XYnnCrIlVHDjPy2yOeza3a2+G6e5CR5ljfEvKE3poqLcu6Ktq/1nrL4OOdWOygRB8rYluES8ZIz2D8Morrmp7oebpeW3Pmv/Q==; 24:YhFr9VnjYxicJPMfSamCMm0lQamvu/VeSbKfHIAcpfCXfLbJHkkcqZH0NUQDPMivao72Klqn3rIeKOpr3xIhjDGl0V31rBGt/euwZMCSpGA= SpamDiagnosticOutput: 1:99 SpamDiagnosticMetadata: NSPM X-Microsoft-Exchange-Diagnostics: 1; BY2PR07MB2423; 7:C17QVgoE94xRVWCS/QazMXJFFRa4g9xZt02/WH5StiXP3OSBrK+4sSHNi4R/2q295KEMppI68RIyi+XhXcZIsKvSxck/e0NTeoXWuJfUd/O7SAP43O2UdQltjfGhaO06dw7lphP49+0qTROuMvvT5MB2TemcDFqq4jCSORx0AjLbf5+JTeOPnGsEyDo9CIs1GhTKw/UtKc/ZpKDbXL8y93cJz9qS4zpYGjks933jKPYDoBOToxdfv7QCHvK0GLIXwhyk1+MyrAwCVSj6nSHepc9o823bi0fNLkyhWUjN/wEu5qCszfEgMexsQnw7YjGtegP3WMQEa6P2EoY0H8PHDA== X-OriginatorOrg: caviumnetworks.com X-MS-Exchange-CrossTenant-OriginalArrivalTime: 27 Apr 2017 14:06:30.9249 (UTC) X-MS-Exchange-CrossTenant-FromEntityHeader: Hosted X-MS-Exchange-Transport-CrossTenantHeadersStamped: BY2PR07MB2423 Subject: [dpdk-dev] [PATCH 1/2] 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, 27 Apr 2017 14:06:39 -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 * Added crypto capability in compilation of generic armv8 and thunderx targets * pmull CRC version is used only after checking the pmull capability at runtime * Verified the changes with crc_autotest unit test case Signed-off-by: Ashwin Sekhar T K --- MAINTAINERS | 1 + lib/librte_eal/common/include/arch/arm/rte_vect.h | 45 +++ lib/librte_net/net_crc_neon.h | 357 ++++++++++++++++++++++ lib/librte_net/rte_net_crc.c | 32 +- lib/librte_net/rte_net_crc.h | 2 + mk/machine/armv8a/rte.vars.mk | 2 +- mk/machine/thunderx/rte.vars.mk | 2 +- mk/rte.cpuflags.mk | 3 + mk/toolchain/gcc/rte.toolchain-compat.mk | 1 + 9 files changed, 438 insertions(+), 7 deletions(-) create mode 100644 lib/librte_net/net_crc_neon.h diff --git a/MAINTAINERS b/MAINTAINERS index 576d60a..283743e 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -149,6 +149,7 @@ F: lib/librte_lpm/rte_lpm_neon.h F: lib/librte_hash/rte*_arm64.h F: lib/librte_efd/rte*_arm64.h F: lib/librte_table/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..9a3dfdf 100644 --- a/lib/librte_eal/common/include/arch/arm/rte_vect.h +++ b/lib/librte_eal/common/include/arch/arm/rte_vect.h @@ -34,9 +34,18 @@ #define _RTE_VECT_ARM_H_ #include +#include + #include "generic/rte_vect.h" #include "arm_neon.h" +#ifdef GCC_VERSION +#undef GCC_VERSION +#endif + +#define GCC_VERSION (__GNUC__ * 10000 + __GNUC_MINOR__ * 100 \ + + __GNUC_PATCHLEVEL__) + #ifdef __cplusplus extern "C" { #endif @@ -78,6 +87,42 @@ vqtbl1q_u8(uint8x16_t a, uint8x16_t b) } #endif +#if (GCC_VERSION < 70000) +/* + * NEON intrinsic vreinterpretq_u64_p128() is not supported + * in GCC versions < 7 + */ +static inline uint64x2_t +vreinterpretq_u64_p128(poly128_t x) +{ + return (uint64x2_t)x; +} + +/* + * NEON intrinsic vreinterpretq_p64_u64() is not supported + * in GCC versions < 7 + */ +static inline poly64x2_t +vreinterpretq_p64_u64(uint64x2_t x) +{ + return (poly64x2_t)x; +} + +/* + * NEON intrinsic vgetq_lane_p64() is not supported + * in GCC versions < 7 + */ +static inline poly64_t +vgetq_lane_p64(poly64x2_t x, const int lane) +{ + 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 76fd129..1daed30 100644 --- a/lib/librte_net/rte_net_crc.h +++ b/lib/librte_net/rte_net_crc.h @@ -55,6 +55,7 @@ enum rte_net_crc_type { enum rte_net_crc_alg { RTE_NET_CRC_SCALAR = 0, RTE_NET_CRC_SSE42, + RTE_NET_CRC_NEON, }; /** @@ -66,6 +67,7 @@ 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); diff --git a/mk/machine/armv8a/rte.vars.mk b/mk/machine/armv8a/rte.vars.mk index d5049e1..51966a5 100644 --- a/mk/machine/armv8a/rte.vars.mk +++ b/mk/machine/armv8a/rte.vars.mk @@ -55,4 +55,4 @@ # CPU_LDFLAGS = # CPU_ASFLAGS = -MACHINE_CFLAGS += -march=armv8-a+crc +MACHINE_CFLAGS += -march=armv8-a+crc+crypto diff --git a/mk/machine/thunderx/rte.vars.mk b/mk/machine/thunderx/rte.vars.mk index ad5a379..6784105 100644 --- a/mk/machine/thunderx/rte.vars.mk +++ b/mk/machine/thunderx/rte.vars.mk @@ -55,4 +55,4 @@ # CPU_LDFLAGS = # CPU_ASFLAGS = -MACHINE_CFLAGS += -march=armv8-a+crc -mcpu=thunderx +MACHINE_CFLAGS += -march=armv8-a+crc+crypto -mcpu=thunderx diff --git a/mk/rte.cpuflags.mk b/mk/rte.cpuflags.mk index e634abc..6bbd742 100644 --- a/mk/rte.cpuflags.mk +++ b/mk/rte.cpuflags.mk @@ -119,6 +119,9 @@ ifneq ($(filter $(AUTO_CPUFLAGS),__ARM_FEATURE_CRC32),) CPUFLAGS += CRC32 endif +ifneq ($(filter $(AUTO_CPUFLAGS),__ARM_FEATURE_CRYPTO),) +CPUFLAGS += PMULL +endif MACHINE_CFLAGS += $(addprefix -DRTE_MACHINE_CPUFLAG_,$(CPUFLAGS)) diff --git a/mk/toolchain/gcc/rte.toolchain-compat.mk b/mk/toolchain/gcc/rte.toolchain-compat.mk index 280dde2..01ac7e2 100644 --- a/mk/toolchain/gcc/rte.toolchain-compat.mk +++ b/mk/toolchain/gcc/rte.toolchain-compat.mk @@ -60,6 +60,7 @@ else # ifeq ($(shell test $(GCC_VERSION) -le 49 && echo 1), 1) MACHINE_CFLAGS := $(patsubst -march=armv8-a+crc,-march=armv8-a+crc -D__ARM_FEATURE_CRC32=1,$(MACHINE_CFLAGS)) + MACHINE_CFLAGS := $(patsubst -march=armv8-a+crc+crypto,-march=armv8-a+crc+crypto -D__ARM_FEATURE_CRC32=1,$(MACHINE_CFLAGS)) endif ifeq ($(shell test $(GCC_VERSION) -le 47 && echo 1), 1) MACHINE_CFLAGS := $(patsubst -march=core-avx-i,-march=corei7-avx,$(MACHINE_CFLAGS)) -- 2.7.4