1/**
2 * @file
3 * SSE2-optimized functions for the VP6 decoder
4 *
5 * Copyright (C) 2009  Zuxy Meng <zuxy.meng@gmail.com>
6 *
7 * This file is part of FFmpeg.
8 *
9 * FFmpeg is free software; you can redistribute it and/or
10 * modify it under the terms of the GNU Lesser General Public
11 * License as published by the Free Software Foundation; either
12 * version 2.1 of the License, or (at your option) any later version.
13 *
14 * FFmpeg is distributed in the hope that it will be useful,
15 * but WITHOUT ANY WARRANTY; without even the implied warranty of
16 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
17 * Lesser General Public License for more details.
18 *
19 * You should have received a copy of the GNU Lesser General Public
20 * License along with FFmpeg; if not, write to the Free Software
21 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
22 */
23
24#include "libavutil/x86_cpu.h"
25#include "libavcodec/dsputil.h"
26#include "dsputil_mmx.h"
27#include "vp6dsp_sse2.h"
28
29#define DIAG4_SSE2(in1,in2,in3,in4)                                           \
30    "movq  "#in1"(%0), %%xmm0            \n\t"                                \
31    "movq  "#in2"(%0), %%xmm1            \n\t"                                \
32    "punpcklbw %%xmm7, %%xmm0            \n\t"                                \
33    "punpcklbw %%xmm7, %%xmm1            \n\t"                                \
34    "pmullw    %%xmm4, %%xmm0            \n\t" /* src[x-8 ] * biweight [0] */ \
35    "pmullw    %%xmm5, %%xmm1            \n\t" /* src[x   ] * biweight [1] */ \
36    "paddw     %%xmm1, %%xmm0            \n\t"                                \
37    "movq  "#in3"(%0), %%xmm1            \n\t"                                \
38    "movq  "#in4"(%0), %%xmm2            \n\t"                                \
39    "punpcklbw %%xmm7, %%xmm1            \n\t"                                \
40    "punpcklbw %%xmm7, %%xmm2            \n\t"                                \
41    "pmullw    %%xmm6, %%xmm1            \n\t" /* src[x+8 ] * biweight [2] */ \
42    "pmullw    %%xmm3, %%xmm2            \n\t" /* src[x+16] * biweight [3] */ \
43    "paddw     %%xmm2, %%xmm1            \n\t"                                \
44    "paddsw     %%xmm1, %%xmm0           \n\t"                                \
45    "paddsw "MANGLE(ff_pw_64)", %%xmm0   \n\t" /* Add 64 */                   \
46    "psraw         $7, %%xmm0            \n\t"                                \
47    "packuswb  %%xmm0, %%xmm0            \n\t"                                \
48    "movq      %%xmm0,   (%1)            \n\t"                                \
49
50void ff_vp6_filter_diag4_sse2(uint8_t *dst, uint8_t *src, int stride,
51                              const int16_t *h_weights,const int16_t *v_weights)
52{
53    uint8_t tmp[8*11], *t = tmp;
54    src -= stride;
55
56    __asm__ volatile(
57    "pxor           %%xmm7, %%xmm7       \n\t"
58    "movq               %4, %%xmm3       \n\t"
59    "pshuflw    $0, %%xmm3, %%xmm4       \n\t"
60    "punpcklqdq     %%xmm4, %%xmm4       \n\t"
61    "pshuflw   $85, %%xmm3, %%xmm5       \n\t"
62    "punpcklqdq     %%xmm5, %%xmm5       \n\t"
63    "pshuflw  $170, %%xmm3, %%xmm6       \n\t"
64    "punpcklqdq     %%xmm6, %%xmm6       \n\t"
65    "pshuflw  $255, %%xmm3, %%xmm3       \n\t"
66    "punpcklqdq     %%xmm3, %%xmm3       \n\t"
67    "1:                                  \n\t"
68    DIAG4_SSE2(-1,0,1,2)
69    "add  $8, %1                         \n\t"
70    "add  %2, %0                         \n\t"
71    "decl %3                             \n\t"
72    "jnz 1b                              \n\t"
73    : "+r"(src), "+r"(t)
74    : "g"((x86_reg)stride), "r"(11), "m"(*(const int64_t*)h_weights)
75    : "memory");
76
77    t = tmp + 8;
78
79    __asm__ volatile(
80    "movq               %4, %%xmm3       \n\t"
81    "pshuflw    $0, %%xmm3, %%xmm4       \n\t"
82    "punpcklqdq     %%xmm4, %%xmm4       \n\t"
83    "pshuflw   $85, %%xmm3, %%xmm5       \n\t"
84    "punpcklqdq     %%xmm5, %%xmm5       \n\t"
85    "pshuflw  $170, %%xmm3, %%xmm6       \n\t"
86    "punpcklqdq     %%xmm6, %%xmm6       \n\t"
87    "pshuflw  $255, %%xmm3, %%xmm3       \n\t"
88    "punpcklqdq     %%xmm3, %%xmm3       \n\t"
89    "1:                                  \n\t"
90    DIAG4_SSE2(-8,0,8,16)
91    "add  $8, %0                         \n\t"
92    "add  %2, %1                         \n\t"
93    "decl %3                             \n\t"
94    "jnz 1b                              \n\t"
95    : "+r"(t), "+r"(dst)
96    : "g"((x86_reg)stride), "r"(8), "m"(*(const int64_t*)v_weights)
97    : "memory");
98}
99