diff options
Diffstat (limited to 'src/vppinfra/memcpy_sse3.h')
| -rw-r--r-- | src/vppinfra/memcpy_sse3.h | 355 |
1 files changed, 355 insertions, 0 deletions
diff --git a/src/vppinfra/memcpy_sse3.h b/src/vppinfra/memcpy_sse3.h new file mode 100644 index 00000000000..4fc48c86c8b --- /dev/null +++ b/src/vppinfra/memcpy_sse3.h @@ -0,0 +1,355 @@ +/* + * Copyright (c) 2016 Cisco and/or its affiliates. + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at: + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +/*- + * BSD LICENSE + * + * Copyright(c) 2010-2014 Intel Corporation. All rights reserved. + * All rights reserved. + * + * 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 Intel 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 included_clib_memcpy_sse3_h +#define included_clib_memcpy_sse3_h + +#include <stdint.h> +#include <x86intrin.h> + +typedef u8 u8x16u __attribute__ ((vector_size (16), aligned (1))); +typedef u8 u8x32u __attribute__ ((vector_size (32), aligned (1))); + +static inline void +clib_mov16 (u8 * dst, const u8 * src) +{ + *(u8x16u *) dst = *(u8x16u *) src; +} + +static inline void +clib_mov32 (u8 * dst, const u8 * src) +{ + *(u8x32u *) dst = *(u8x32u *) src; +} + +static inline void +clib_mov64 (u8 * dst, const u8 * src) +{ + clib_mov32 ((u8 *) dst + 0 * 32, (const u8 *) src + 0 * 32); + clib_mov32 ((u8 *) dst + 1 * 32, (const u8 *) src + 1 * 32); +} + +static inline void +clib_mov128 (u8 * dst, const u8 * src) +{ + clib_mov64 ((u8 *) dst + 0 * 64, (const u8 *) src + 0 * 64); + clib_mov64 ((u8 *) dst + 1 * 64, (const u8 *) src + 1 * 64); +} + +static inline void +clib_mov256 (u8 * dst, const u8 * src) +{ + clib_mov128 ((u8 *) dst + 0 * 128, (const u8 *) src + 0 * 128); + clib_mov128 ((u8 *) dst + 1 * 128, (const u8 *) src + 1 * 128); +} + +/** + * Macro for copying unaligned block from one location to another with constant load offset, + * 47 bytes leftover maximum, + * locations should not overlap. + * Requirements: + * - Store is aligned + * - Load offset is <offset>, which must be immediate value within [1, 15] + * - For <src>, make sure <offset> bit backwards & <16 - offset> bit forwards are available for loading + * - <dst>, <src>, <len> must be variables + * - __m128i <xmm0> ~ <xmm8> must be pre-defined + */ +#define CLIB_MVUNALIGN_LEFT47_IMM(dst, src, len, offset) \ +({ \ + int tmp; \ + while (len >= 128 + 16 - offset) { \ + xmm0 = _mm_loadu_si128((const __m128i *)((const u8 *)src - offset + 0 * 16)); \ + len -= 128; \ + xmm1 = _mm_loadu_si128((const __m128i *)((const u8 *)src - offset + 1 * 16)); \ + xmm2 = _mm_loadu_si128((const __m128i *)((const u8 *)src - offset + 2 * 16)); \ + xmm3 = _mm_loadu_si128((const __m128i *)((const u8 *)src - offset + 3 * 16)); \ + xmm4 = _mm_loadu_si128((const __m128i *)((const u8 *)src - offset + 4 * 16)); \ + xmm5 = _mm_loadu_si128((const __m128i *)((const u8 *)src - offset + 5 * 16)); \ + xmm6 = _mm_loadu_si128((const __m128i *)((const u8 *)src - offset + 6 * 16)); \ + xmm7 = _mm_loadu_si128((const __m128i *)((const u8 *)src - offset + 7 * 16)); \ + xmm8 = _mm_loadu_si128((const __m128i *)((const u8 *)src - offset + 8 * 16)); \ + src = (const u8 *)src + 128; \ + _mm_storeu_si128((__m128i *)((u8 *)dst + 0 * 16), _mm_alignr_epi8(xmm1, xmm0, offset)); \ + _mm_storeu_si128((__m128i *)((u8 *)dst + 1 * 16), _mm_alignr_epi8(xmm2, xmm1, offset)); \ + _mm_storeu_si128((__m128i *)((u8 *)dst + 2 * 16), _mm_alignr_epi8(xmm3, xmm2, offset)); \ + _mm_storeu_si128((__m128i *)((u8 *)dst + 3 * 16), _mm_alignr_epi8(xmm4, xmm3, offset)); \ + _mm_storeu_si128((__m128i *)((u8 *)dst + 4 * 16), _mm_alignr_epi8(xmm5, xmm4, offset)); \ + _mm_storeu_si128((__m128i *)((u8 *)dst + 5 * 16), _mm_alignr_epi8(xmm6, xmm5, offset)); \ + _mm_storeu_si128((__m128i *)((u8 *)dst + 6 * 16), _mm_alignr_epi8(xmm7, xmm6, offset)); \ + _mm_storeu_si128((__m128i *)((u8 *)dst + 7 * 16), _mm_alignr_epi8(xmm8, xmm7, offset)); \ + dst = (u8 *)dst + 128; \ + } \ + tmp = len; \ + len = ((len - 16 + offset) & 127) + 16 - offset; \ + tmp -= len; \ + src = (const u8 *)src + tmp; \ + dst = (u8 *)dst + tmp; \ + if (len >= 32 + 16 - offset) { \ + while (len >= 32 + 16 - offset) { \ + xmm0 = _mm_loadu_si128((const __m128i *)((const u8 *)src - offset + 0 * 16)); \ + len -= 32; \ + xmm1 = _mm_loadu_si128((const __m128i *)((const u8 *)src - offset + 1 * 16)); \ + xmm2 = _mm_loadu_si128((const __m128i *)((const u8 *)src - offset + 2 * 16)); \ + src = (const u8 *)src + 32; \ + _mm_storeu_si128((__m128i *)((u8 *)dst + 0 * 16), _mm_alignr_epi8(xmm1, xmm0, offset)); \ + _mm_storeu_si128((__m128i *)((u8 *)dst + 1 * 16), _mm_alignr_epi8(xmm2, xmm1, offset)); \ + dst = (u8 *)dst + 32; \ + } \ + tmp = len; \ + len = ((len - 16 + offset) & 31) + 16 - offset; \ + tmp -= len; \ + src = (const u8 *)src + tmp; \ + dst = (u8 *)dst + tmp; \ + } \ +}) + +/** + * Macro for copying unaligned block from one location to another, + * 47 bytes leftover maximum, + * locations should not overlap. + * Use switch here because the aligning instruction requires immediate value for shift count. + * Requirements: + * - Store is aligned + * - Load offset is <offset>, which must be within [1, 15] + * - For <src>, make sure <offset> bit backwards & <16 - offset> bit forwards are available for loading + * - <dst>, <src>, <len> must be variables + * - __m128i <xmm0> ~ <xmm8> used in CLIB_MVUNALIGN_LEFT47_IMM must be pre-defined + */ +#define CLIB_MVUNALIGN_LEFT47(dst, src, len, offset) \ +({ \ + switch (offset) { \ + case 0x01: CLIB_MVUNALIGN_LEFT47_IMM(dst, src, n, 0x01); break; \ + case 0x02: CLIB_MVUNALIGN_LEFT47_IMM(dst, src, n, 0x02); break; \ + case 0x03: CLIB_MVUNALIGN_LEFT47_IMM(dst, src, n, 0x03); break; \ + case 0x04: CLIB_MVUNALIGN_LEFT47_IMM(dst, src, n, 0x04); break; \ + case 0x05: CLIB_MVUNALIGN_LEFT47_IMM(dst, src, n, 0x05); break; \ + case 0x06: CLIB_MVUNALIGN_LEFT47_IMM(dst, src, n, 0x06); break; \ + case 0x07: CLIB_MVUNALIGN_LEFT47_IMM(dst, src, n, 0x07); break; \ + case 0x08: CLIB_MVUNALIGN_LEFT47_IMM(dst, src, n, 0x08); break; \ + case 0x09: CLIB_MVUNALIGN_LEFT47_IMM(dst, src, n, 0x09); break; \ + case 0x0A: CLIB_MVUNALIGN_LEFT47_IMM(dst, src, n, 0x0A); break; \ + case 0x0B: CLIB_MVUNALIGN_LEFT47_IMM(dst, src, n, 0x0B); break; \ + case 0x0C: CLIB_MVUNALIGN_LEFT47_IMM(dst, src, n, 0x0C); break; \ + case 0x0D: CLIB_MVUNALIGN_LEFT47_IMM(dst, src, n, 0x0D); break; \ + case 0x0E: CLIB_MVUNALIGN_LEFT47_IMM(dst, src, n, 0x0E); break; \ + case 0x0F: CLIB_MVUNALIGN_LEFT47_IMM(dst, src, n, 0x0F); break; \ + default:; \ + } \ +}) + +static inline void * +clib_memcpy (void *dst, const void *src, size_t n) +{ + __m128i xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7, xmm8; + uword dstu = (uword) dst; + uword srcu = (uword) src; + void *ret = dst; + size_t dstofss; + size_t srcofs; + + /** + * Copy less than 16 bytes + */ + if (n < 16) + { + if (n & 0x01) + { + *(u8 *) dstu = *(const u8 *) srcu; + srcu = (uword) ((const u8 *) srcu + 1); + dstu = (uword) ((u8 *) dstu + 1); + } + if (n & 0x02) + { + *(u16 *) dstu = *(const u16 *) srcu; + srcu = (uword) ((const u16 *) srcu + 1); + dstu = (uword) ((u16 *) dstu + 1); + } + if (n & 0x04) + { + *(u32 *) dstu = *(const u32 *) srcu; + srcu = (uword) ((const u32 *) srcu + 1); + dstu = (uword) ((u32 *) dstu + 1); + } + if (n & 0x08) + { + *(u64 *) dstu = *(const u64 *) srcu; + } + return ret; + } + + /** + * Fast way when copy size doesn't exceed 512 bytes + */ + if (n <= 32) + { + clib_mov16 ((u8 *) dst, (const u8 *) src); + clib_mov16 ((u8 *) dst - 16 + n, (const u8 *) src - 16 + n); + return ret; + } + if (n <= 48) + { + clib_mov32 ((u8 *) dst, (const u8 *) src); + clib_mov16 ((u8 *) dst - 16 + n, (const u8 *) src - 16 + n); + return ret; + } + if (n <= 64) + { + clib_mov32 ((u8 *) dst, (const u8 *) src); + clib_mov16 ((u8 *) dst + 32, (const u8 *) src + 32); + clib_mov16 ((u8 *) dst - 16 + n, (const u8 *) src - 16 + n); + return ret; + } + if (n <= 128) + { + goto COPY_BLOCK_128_BACK15; + } + if (n <= 512) + { + if (n >= 256) + { + n -= 256; + clib_mov128 ((u8 *) dst, (const u8 *) src); + clib_mov128 ((u8 *) dst + 128, (const u8 *) src + 128); + src = (const u8 *) src + 256; + dst = (u8 *) dst + 256; + } + COPY_BLOCK_255_BACK15: + if (n >= 128) + { + n -= 128; + clib_mov128 ((u8 *) dst, (const u8 *) src); + src = (const u8 *) src + 128; + dst = (u8 *) dst + 128; + } + COPY_BLOCK_128_BACK15: + if (n >= 64) + { + n -= 64; + clib_mov64 ((u8 *) dst, (const u8 *) src); + src = (const u8 *) src + 64; + dst = (u8 *) dst + 64; + } + COPY_BLOCK_64_BACK15: + if (n >= 32) + { + n -= 32; + clib_mov32 ((u8 *) dst, (const u8 *) src); + src = (const u8 *) src + 32; + dst = (u8 *) dst + 32; + } + if (n > 16) + { + clib_mov16 ((u8 *) dst, (const u8 *) src); + clib_mov16 ((u8 *) dst - 16 + n, (const u8 *) src - 16 + n); + return ret; + } + if (n > 0) + { + clib_mov16 ((u8 *) dst - 16 + n, (const u8 *) src - 16 + n); + } + return ret; + } + + /** + * Make store aligned when copy size exceeds 512 bytes, + * and make sure the first 15 bytes are copied, because + * unaligned copy functions require up to 15 bytes + * backwards access. + */ + dstofss = 16 - ((uword) dst & 0x0F) + 16; + n -= dstofss; + clib_mov32 ((u8 *) dst, (const u8 *) src); + src = (const u8 *) src + dstofss; + dst = (u8 *) dst + dstofss; + srcofs = ((uword) src & 0x0F); + + /** + * For aligned copy + */ + if (srcofs == 0) + { + /** + * Copy 256-byte blocks + */ + for (; n >= 256; n -= 256) + { + clib_mov256 ((u8 *) dst, (const u8 *) src); + dst = (u8 *) dst + 256; + src = (const u8 *) src + 256; + } + + /** + * Copy whatever left + */ + goto COPY_BLOCK_255_BACK15; + } + + /** + * For copy with unaligned load + */ + CLIB_MVUNALIGN_LEFT47 (dst, src, n, srcofs); + + /** + * Copy whatever left + */ + goto COPY_BLOCK_64_BACK15; +} + + +#undef CLIB_MVUNALIGN_LEFT47_IMM +#undef CLIB_MVUNALIGN_LEFT47 + +#endif /* included_clib_memcpy_sse3_h */ + + +/* + * fd.io coding-style-patch-verification: ON + * + * Local Variables: + * eval: (c-set-style "gnu") + * End: + */ |