From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from e28smtp06.in.ibm.com (e28smtp06.in.ibm.com [122.248.162.6]) by dpdk.org (Postfix) with ESMTP id A1A537F60 for ; Tue, 25 Nov 2014 11:05:37 +0100 (CET) Received: from /spool/local by e28smtp06.in.ibm.com with IBM ESMTP SMTP Gateway: Authorized Use Only! Violators will be prosecuted for from ; Tue, 25 Nov 2014 15:46:27 +0530 Received: from d28dlp03.in.ibm.com (9.184.220.128) by e28smtp06.in.ibm.com (192.168.1.136) with IBM ESMTP SMTP Gateway: Authorized Use Only! Violators will be prosecuted; Tue, 25 Nov 2014 15:46:24 +0530 Received: from d28relay03.in.ibm.com (d28relay03.in.ibm.com [9.184.220.60]) by d28dlp03.in.ibm.com (Postfix) with ESMTP id C712A1258023 for ; Tue, 25 Nov 2014 15:46:36 +0530 (IST) Received: from d28av05.in.ibm.com (d28av05.in.ibm.com [9.184.220.67]) by d28relay03.in.ibm.com (8.14.9/8.14.9/NCO v10.0) with ESMTP id sAPAIXeP2163108 for ; Tue, 25 Nov 2014 15:48:33 +0530 Received: from d28av05.in.ibm.com (localhost [127.0.0.1]) by d28av05.in.ibm.com (8.14.4/8.14.4/NCO v10.0 AVout) with ESMTP id sAPAGEca021271 for ; Tue, 25 Nov 2014 15:46:15 +0530 Received: from os_controller.crl.ibm.com ([9.186.57.58]) by d28av05.in.ibm.com (8.14.4/8.14.4/NCO v10.0 AVin) with ESMTP id sAPAG3kq020283 for ; Tue, 25 Nov 2014 15:46:13 +0530 From: Chao Zhu To: dev@dpdk.org Date: Tue, 25 Nov 2014 17:17:10 -0500 Message-Id: <1416953837-15894-8-git-send-email-chaozhu@linux.vnet.ibm.com> X-Mailer: git-send-email 1.7.1 In-Reply-To: <1416953837-15894-1-git-send-email-chaozhu@linux.vnet.ibm.com> References: <1416953837-15894-1-git-send-email-chaozhu@linux.vnet.ibm.com> X-TM-AS-MML: disable X-Content-Scanned: Fidelis XPS MAILER x-cbid: 14112510-0021-0000-0000-0000026EBF8B Subject: [dpdk-dev] [PATCH v5 07/14] Add vector memcpy for IBM Power architecture X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: patches and discussions about DPDK List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 25 Nov 2014 10:05:39 -0000 The SSE based memory copy in DPDK only support x86. This patch adds altivec based memory copy functions for IBM Power architecture. This patch includes altivec.h which requires GCC version>= 4.8. Signed-off-by: Chao Zhu --- .../common/include/arch/ppc_64/rte_memcpy.h | 226 ++++++++++++++++++++ 1 files changed, 226 insertions(+), 0 deletions(-) create mode 100644 lib/librte_eal/common/include/arch/ppc_64/rte_memcpy.h diff --git a/lib/librte_eal/common/include/arch/ppc_64/rte_memcpy.h b/lib/librte_eal/common/include/arch/ppc_64/rte_memcpy.h new file mode 100644 index 0000000..9f3b062 --- /dev/null +++ b/lib/librte_eal/common/include/arch/ppc_64/rte_memcpy.h @@ -0,0 +1,226 @@ +/* + * BSD LICENSE + * + * Copyright (C) IBM Corporation 2014. + * + * 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 IBM Corporation 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 _RTE_MEMCPY_PPC_64_H_ +#define _RTE_MEMCPY_PPC_64_H_ + +#include +#include +/*To include altivec.h, GCC version must >= 4.8 */ +#include + +#ifdef __cplusplus +extern "C" { +#endif + +#include "generic/rte_memcpy.h" + +static inline void +rte_mov16(uint8_t *dst, const uint8_t *src) +{ + vec_vsx_st(vec_vsx_ld(0, src), 0, dst); +} + +static inline void +rte_mov32(uint8_t *dst, const uint8_t *src) +{ + vec_vsx_st(vec_vsx_ld(0, src), 0, dst); + vec_vsx_st(vec_vsx_ld(16, src), 16, dst); +} + +static inline void +rte_mov48(uint8_t *dst, const uint8_t *src) +{ + vec_vsx_st(vec_vsx_ld(0, src), 0, dst); + vec_vsx_st(vec_vsx_ld(16, src), 16, dst); + vec_vsx_st(vec_vsx_ld(32, src), 32, dst); +} + +static inline void +rte_mov64(uint8_t *dst, const uint8_t *src) +{ + vec_vsx_st(vec_vsx_ld(0, src), 0, dst); + vec_vsx_st(vec_vsx_ld(16, src), 16, dst); + vec_vsx_st(vec_vsx_ld(32, src), 32, dst); + vec_vsx_st(vec_vsx_ld(48, src), 48, dst); +} + +static inline void +rte_mov128(uint8_t *dst, const uint8_t *src) +{ + vec_vsx_st(vec_vsx_ld(0, src), 0, dst); + vec_vsx_st(vec_vsx_ld(16, src), 16, dst); + vec_vsx_st(vec_vsx_ld(32, src), 32, dst); + vec_vsx_st(vec_vsx_ld(48, src), 48, dst); + vec_vsx_st(vec_vsx_ld(64, src), 64, dst); + vec_vsx_st(vec_vsx_ld(80, src), 80, dst); + vec_vsx_st(vec_vsx_ld(96, src), 96, dst); + vec_vsx_st(vec_vsx_ld(112, src), 112, dst); +} + +static inline void +rte_mov256(uint8_t *dst, const uint8_t *src) +{ + rte_mov128(dst, src); + rte_mov128(dst + 128, src + 128); +} + +#define rte_memcpy(dst, src, n) \ + ((__builtin_constant_p(n)) ? \ + memcpy((dst), (src), (n)) : \ + rte_memcpy_func((dst), (src), (n))) + +static inline void * +rte_memcpy_func(void *dst, const void *src, size_t n) +{ + void *ret = dst; + + /* We can't copy < 16 bytes using XMM registers so do it manually. */ + if (n < 16) { + if (n & 0x01) { + *(uint8_t *)dst = *(const uint8_t *)src; + dst = (uint8_t *)dst + 1; + src = (const uint8_t *)src + 1; + } + if (n & 0x02) { + *(uint16_t *)dst = *(const uint16_t *)src; + dst = (uint16_t *)dst + 1; + src = (const uint16_t *)src + 1; + } + if (n & 0x04) { + *(uint32_t *)dst = *(const uint32_t *)src; + dst = (uint32_t *)dst + 1; + src = (const uint32_t *)src + 1; + } + if (n & 0x08) + *(uint64_t *)dst = *(const uint64_t *)src; + return ret; + } + + /* Special fast cases for <= 128 bytes */ + if (n <= 32) { + rte_mov16((uint8_t *)dst, (const uint8_t *)src); + rte_mov16((uint8_t *)dst - 16 + n, + (const uint8_t *)src - 16 + n); + return ret; + } + + if (n <= 64) { + rte_mov32((uint8_t *)dst, (const uint8_t *)src); + rte_mov32((uint8_t *)dst - 32 + n, + (const uint8_t *)src - 32 + n); + return ret; + } + + if (n <= 128) { + rte_mov64((uint8_t *)dst, (const uint8_t *)src); + rte_mov64((uint8_t *)dst - 64 + n, + (const uint8_t *)src - 64 + n); + return ret; + } + + /* + * For large copies > 128 bytes. This combination of 256, 64 and 16 byte + * copies was found to be faster than doing 128 and 32 byte copies as + * well. + */ + for ( ; n >= 256; n -= 256) { + rte_mov256((uint8_t *)dst, (const uint8_t *)src); + dst = (uint8_t *)dst + 256; + src = (const uint8_t *)src + 256; + } + + /* + * We split the remaining bytes (which will be less than 256) into + * 64byte (2^6) chunks. + * Using incrementing integers in the case labels of a switch statement + * enourages the compiler to use a jump table. To get incrementing + * integers, we shift the 2 relevant bits to the LSB position to first + * get decrementing integers, and then subtract. + */ + switch (3 - (n >> 6)) { + case 0x00: + rte_mov64((uint8_t *)dst, (const uint8_t *)src); + n -= 64; + dst = (uint8_t *)dst + 64; + src = (const uint8_t *)src + 64; /* fallthrough */ + case 0x01: + rte_mov64((uint8_t *)dst, (const uint8_t *)src); + n -= 64; + dst = (uint8_t *)dst + 64; + src = (const uint8_t *)src + 64; /* fallthrough */ + case 0x02: + rte_mov64((uint8_t *)dst, (const uint8_t *)src); + n -= 64; + dst = (uint8_t *)dst + 64; + src = (const uint8_t *)src + 64; /* fallthrough */ + default: + ; + } + + /* + * We split the remaining bytes (which will be less than 64) into + * 16byte (2^4) chunks, using the same switch structure as above. + */ + switch (3 - (n >> 4)) { + case 0x00: + rte_mov16((uint8_t *)dst, (const uint8_t *)src); + n -= 16; + dst = (uint8_t *)dst + 16; + src = (const uint8_t *)src + 16; /* fallthrough */ + case 0x01: + rte_mov16((uint8_t *)dst, (const uint8_t *)src); + n -= 16; + dst = (uint8_t *)dst + 16; + src = (const uint8_t *)src + 16; /* fallthrough */ + case 0x02: + rte_mov16((uint8_t *)dst, (const uint8_t *)src); + n -= 16; + dst = (uint8_t *)dst + 16; + src = (const uint8_t *)src + 16; /* fallthrough */ + default: + ; + } + + /* Copy any remaining bytes, without going beyond end of buffers */ + if (n != 0) + rte_mov16((uint8_t *)dst - 16 + n, + (const uint8_t *)src - 16 + n); + return ret; +} + +#ifdef __cplusplus +} +#endif + +#endif /* _RTE_MEMCPY_PPC_64_H_ */ + -- 1.7.1