10cf1993335981f7d1f9c76d72a8c8ba56add658
[ffmpeg.git] / libavcodec / aarch64 / h264dsp_init_aarch64.c
1 /*
2  * Copyright (c) 2010 Mans Rullgard <mans@mansr.com>
3  *
4  * This file is part of FFmpeg.
5  *
6  * FFmpeg is free software; you can redistribute it and/or
7  * modify it under the terms of the GNU Lesser General Public
8  * License as published by the Free Software Foundation; either
9  * version 2.1 of the License, or (at your option) any later version.
10  *
11  * FFmpeg is distributed in the hope that it will be useful,
12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
14  * Lesser General Public License for more details.
15  *
16  * You should have received a copy of the GNU Lesser General Public
17  * License along with FFmpeg; if not, write to the Free Software
18  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
19  */
20
21 #include <stdint.h>
22
23 #include "libavutil/attributes.h"
24 #include "libavutil/cpu.h"
25 #include "libavutil/aarch64/cpu.h"
26 #include "libavcodec/h264dsp.h"
27
28 void ff_h264_v_loop_filter_luma_neon(uint8_t *pix, int stride, int alpha,
29                                      int beta, int8_t *tc0);
30 void ff_h264_h_loop_filter_luma_neon(uint8_t *pix, int stride, int alpha,
31                                      int beta, int8_t *tc0);
32 void ff_h264_v_loop_filter_luma_intra_neon(uint8_t *pix, int stride, int alpha,
33                                            int beta);
34 void ff_h264_h_loop_filter_luma_intra_neon(uint8_t *pix, int stride, int alpha,
35                                            int beta);
36 void ff_h264_v_loop_filter_chroma_neon(uint8_t *pix, int stride, int alpha,
37                                        int beta, int8_t *tc0);
38 void ff_h264_h_loop_filter_chroma_neon(uint8_t *pix, int stride, int alpha,
39                                        int beta, int8_t *tc0);
40 void ff_h264_v_loop_filter_chroma_intra_neon(uint8_t *pix, int stride,
41                                              int alpha, int beta);
42 void ff_h264_h_loop_filter_chroma_intra_neon(uint8_t *pix, int stride,
43                                              int alpha, int beta);
44 void ff_h264_h_loop_filter_chroma_mbaff_intra_neon(uint8_t *pix, int stride,
45                                                    int alpha, int beta);
46
47 void ff_weight_h264_pixels_16_neon(uint8_t *dst, ptrdiff_t stride, int height,
48                                    int log2_den, int weight, int offset);
49 void ff_weight_h264_pixels_8_neon(uint8_t *dst, ptrdiff_t stride, int height,
50                                   int log2_den, int weight, int offset);
51 void ff_weight_h264_pixels_4_neon(uint8_t *dst, ptrdiff_t stride, int height,
52                                   int log2_den, int weight, int offset);
53
54 void ff_biweight_h264_pixels_16_neon(uint8_t *dst, uint8_t *src, ptrdiff_t stride,
55                                      int height, int log2_den, int weightd,
56                                      int weights, int offset);
57 void ff_biweight_h264_pixels_8_neon(uint8_t *dst, uint8_t *src, ptrdiff_t stride,
58                                     int height, int log2_den, int weightd,
59                                     int weights, int offset);
60 void ff_biweight_h264_pixels_4_neon(uint8_t *dst, uint8_t *src, ptrdiff_t stride,
61                                     int height, int log2_den, int weightd,
62                                     int weights, int offset);
63
64 void ff_h264_idct_add_neon(uint8_t *dst, int16_t *block, int stride);
65 void ff_h264_idct_dc_add_neon(uint8_t *dst, int16_t *block, int stride);
66 void ff_h264_idct_add16_neon(uint8_t *dst, const int *block_offset,
67                              int16_t *block, int stride,
68                              const uint8_t nnzc[6*8]);
69 void ff_h264_idct_add16intra_neon(uint8_t *dst, const int *block_offset,
70                                   int16_t *block, int stride,
71                                   const uint8_t nnzc[6*8]);
72 void ff_h264_idct_add8_neon(uint8_t **dest, const int *block_offset,
73                             int16_t *block, int stride,
74                             const uint8_t nnzc[6*8]);
75
76 void ff_h264_idct8_add_neon(uint8_t *dst, int16_t *block, int stride);
77 void ff_h264_idct8_dc_add_neon(uint8_t *dst, int16_t *block, int stride);
78 void ff_h264_idct8_add4_neon(uint8_t *dst, const int *block_offset,
79                              int16_t *block, int stride,
80                              const uint8_t nnzc[6*8]);
81
82 av_cold void ff_h264dsp_init_aarch64(H264DSPContext *c, const int bit_depth,
83                                      const int chroma_format_idc)
84 {
85     int cpu_flags = av_get_cpu_flags();
86
87     if (have_neon(cpu_flags) && bit_depth == 8) {
88         c->h264_v_loop_filter_luma   = ff_h264_v_loop_filter_luma_neon;
89         c->h264_h_loop_filter_luma   = ff_h264_h_loop_filter_luma_neon;
90         c->h264_v_loop_filter_luma_intra= ff_h264_v_loop_filter_luma_intra_neon;
91         c->h264_h_loop_filter_luma_intra= ff_h264_h_loop_filter_luma_intra_neon;
92
93         c->h264_v_loop_filter_chroma = ff_h264_v_loop_filter_chroma_neon;
94         if (chroma_format_idc <= 1)
95         c->h264_h_loop_filter_chroma = ff_h264_h_loop_filter_chroma_neon;
96         c->h264_v_loop_filter_chroma_intra = ff_h264_v_loop_filter_chroma_intra_neon;
97         c->h264_h_loop_filter_chroma_intra = ff_h264_h_loop_filter_chroma_intra_neon;
98         c->h264_h_loop_filter_chroma_mbaff_intra = ff_h264_h_loop_filter_chroma_mbaff_intra_neon;
99
100         c->weight_h264_pixels_tab[0] = ff_weight_h264_pixels_16_neon;
101         c->weight_h264_pixels_tab[1] = ff_weight_h264_pixels_8_neon;
102         c->weight_h264_pixels_tab[2] = ff_weight_h264_pixels_4_neon;
103
104         c->biweight_h264_pixels_tab[0] = ff_biweight_h264_pixels_16_neon;
105         c->biweight_h264_pixels_tab[1] = ff_biweight_h264_pixels_8_neon;
106         c->biweight_h264_pixels_tab[2] = ff_biweight_h264_pixels_4_neon;
107
108         c->h264_idct_add        = ff_h264_idct_add_neon;
109         c->h264_idct_dc_add     = ff_h264_idct_dc_add_neon;
110         c->h264_idct_add16      = ff_h264_idct_add16_neon;
111         c->h264_idct_add16intra = ff_h264_idct_add16intra_neon;
112         if (chroma_format_idc <= 1)
113             c->h264_idct_add8   = ff_h264_idct_add8_neon;
114         c->h264_idct8_add       = ff_h264_idct8_add_neon;
115         c->h264_idct8_dc_add    = ff_h264_idct8_dc_add_neon;
116         c->h264_idct8_add4      = ff_h264_idct8_add4_neon;
117     }
118 }