1/*
2 * Copyright © 2013 Intel Corporation
3 *
4 * Permission is hereby granted, free of charge, to any person obtaining a
5 * copy of this software and associated documentation files (the "Software"),
6 * to deal in the Software without restriction, including without limitation
7 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
8 * and/or sell copies of the Software, and to permit persons to whom the
9 * Software is furnished to do so, subject to the following conditions:
10 *
11 * The above copyright notice and this permission notice (including the next
12 * paragraph) shall be included in all copies or substantial portions of the
13 * Software.
14 *
15 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
18 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
20 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
21 * IN THE SOFTWARE.
22 *
23 * Authors:
24 *    Eric Anholt <eric@anholt.net>
25 *    Matt Turner <mattst88@gmail.com>
26 *
27 */
28
29#include "main/macros.h"
30#include "util/streaming-load-memcpy.h"
31#include "x86/common_x86_asm.h"
32#ifdef USE_SSE41
33#include <smmintrin.h>
34#endif
35
36/* Copies memory from src to dst, using SSE 4.1's MOVNTDQA to get streaming
37 * read performance from uncached memory.
38 */
39void
40util_streaming_load_memcpy(void *restrict dst, void *restrict src, size_t len)
41{
42   char *restrict d = dst;
43   char *restrict s = src;
44
45#ifdef USE_SSE41
46   /* If dst and src are not co-aligned, or if SSE4.1 is not present, fallback to memcpy(). */
47   if (((uintptr_t)d & 15) != ((uintptr_t)s & 15) || !cpu_has_sse4_1) {
48      memcpy(d, s, len);
49      return;
50   }
51
52   /* memcpy() the misaligned header. At the end of this if block, <d> and <s>
53    * are aligned to a 16-byte boundary or <len> == 0.
54    */
55   if ((uintptr_t)d & 15) {
56      uintptr_t bytes_before_alignment_boundary = 16 - ((uintptr_t)d & 15);
57      assert(bytes_before_alignment_boundary < 16);
58
59      memcpy(d, s, MIN2(bytes_before_alignment_boundary, len));
60
61      d = (char *)ALIGN((uintptr_t)d, 16);
62      s = (char *)ALIGN((uintptr_t)s, 16);
63      len -= MIN2(bytes_before_alignment_boundary, len);
64   }
65
66   if (len >= 64)
67      _mm_mfence();
68
69   while (len >= 64) {
70      __m128i *dst_cacheline = (__m128i *)d;
71      __m128i *src_cacheline = (__m128i *)s;
72
73      __m128i temp1 = _mm_stream_load_si128(src_cacheline + 0);
74      __m128i temp2 = _mm_stream_load_si128(src_cacheline + 1);
75      __m128i temp3 = _mm_stream_load_si128(src_cacheline + 2);
76      __m128i temp4 = _mm_stream_load_si128(src_cacheline + 3);
77
78      _mm_store_si128(dst_cacheline + 0, temp1);
79      _mm_store_si128(dst_cacheline + 1, temp2);
80      _mm_store_si128(dst_cacheline + 2, temp3);
81      _mm_store_si128(dst_cacheline + 3, temp4);
82
83      d += 64;
84      s += 64;
85      len -= 64;
86   }
87#endif
88   /* memcpy() the tail. */
89   if (len) {
90      memcpy(d, s, len);
91   }
92}
93