FFmpeg
hevcdsp_init_aarch64.c
Go to the documentation of this file.
1 /*
2  * Copyright (c) 2020 Reimar Döffinger
3  * Copyright (c) 2023 xu fulong <839789740@qq.com>
4  *
5  * This file is part of FFmpeg.
6  *
7  * FFmpeg is free software; you can redistribute it and/or
8  * modify it under the terms of the GNU Lesser General Public
9  * License as published by the Free Software Foundation; either
10  * version 2.1 of the License, or (at your option) any later version.
11  *
12  * FFmpeg is distributed in the hope that it will be useful,
13  * but WITHOUT ANY WARRANTY; without even the implied warranty of
14  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15  * Lesser General Public License for more details.
16  *
17  * You should have received a copy of the GNU Lesser General Public
18  * License along with FFmpeg; if not, write to the Free Software
19  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
20  */
21 
22 #include <stdint.h>
23 
24 #include "libavutil/attributes.h"
25 #include "libavutil/cpu.h"
26 #include "libavutil/aarch64/cpu.h"
27 #include "libavcodec/hevc/dsp.h"
28 
29 void ff_hevc_v_loop_filter_chroma_8_neon(uint8_t *_pix, ptrdiff_t _stride,
30  const int *_tc, const uint8_t *_no_p, const uint8_t *_no_q);
31 void ff_hevc_v_loop_filter_chroma_10_neon(uint8_t *_pix, ptrdiff_t _stride,
32  const int *_tc, const uint8_t *_no_p, const uint8_t *_no_q);
33 void ff_hevc_v_loop_filter_chroma_12_neon(uint8_t *_pix, ptrdiff_t _stride,
34  const int *_tc, const uint8_t *_no_p, const uint8_t *_no_q);
35 void ff_hevc_h_loop_filter_chroma_8_neon(uint8_t *_pix, ptrdiff_t _stride,
36  const int *_tc, const uint8_t *_no_p, const uint8_t *_no_q);
37 void ff_hevc_h_loop_filter_chroma_10_neon(uint8_t *_pix, ptrdiff_t _stride,
38  const int *_tc, const uint8_t *_no_p, const uint8_t *_no_q);
39 void ff_hevc_h_loop_filter_chroma_12_neon(uint8_t *_pix, ptrdiff_t _stride,
40  const int *_tc, const uint8_t *_no_p, const uint8_t *_no_q);
41 void ff_hevc_v_loop_filter_luma_8_neon(uint8_t *_pix, ptrdiff_t _stride, int beta,
42  const int *_tc, const uint8_t *_no_p, const uint8_t *_no_q);
43 void ff_hevc_v_loop_filter_luma_10_neon(uint8_t *_pix, ptrdiff_t _stride, int beta,
44  const int *_tc, const uint8_t *_no_p, const uint8_t *_no_q);
45 void ff_hevc_v_loop_filter_luma_12_neon(uint8_t *_pix, ptrdiff_t _stride, int beta,
46  const int *_tc, const uint8_t *_no_p, const uint8_t *_no_q);
47 void ff_hevc_h_loop_filter_luma_8_neon(uint8_t *_pix, ptrdiff_t _stride, int beta,
48  const int *_tc, const uint8_t *_no_p, const uint8_t *_no_q);
49 void ff_hevc_h_loop_filter_luma_10_neon(uint8_t *_pix, ptrdiff_t _stride, int beta,
50  const int *_tc, const uint8_t *_no_p, const uint8_t *_no_q);
51 void ff_hevc_h_loop_filter_luma_12_neon(uint8_t *_pix, ptrdiff_t _stride, int beta,
52  const int *_tc, const uint8_t *_no_p, const uint8_t *_no_q);
53 void ff_hevc_add_residual_4x4_8_neon(uint8_t *_dst, const int16_t *coeffs,
54  ptrdiff_t stride);
55 void ff_hevc_add_residual_4x4_10_neon(uint8_t *_dst, const int16_t *coeffs,
56  ptrdiff_t stride);
57 void ff_hevc_add_residual_4x4_12_neon(uint8_t *_dst, const int16_t *coeffs,
58  ptrdiff_t stride);
59 void ff_hevc_add_residual_8x8_8_neon(uint8_t *_dst, const int16_t *coeffs,
60  ptrdiff_t stride);
61 void ff_hevc_add_residual_8x8_10_neon(uint8_t *_dst, const int16_t *coeffs,
62  ptrdiff_t stride);
63 void ff_hevc_add_residual_8x8_12_neon(uint8_t *_dst, const int16_t *coeffs,
64  ptrdiff_t stride);
65 void ff_hevc_add_residual_16x16_8_neon(uint8_t *_dst, const int16_t *coeffs,
66  ptrdiff_t stride);
67 void ff_hevc_add_residual_16x16_10_neon(uint8_t *_dst, const int16_t *coeffs,
68  ptrdiff_t stride);
69 void ff_hevc_add_residual_16x16_12_neon(uint8_t *_dst, const int16_t *coeffs,
70  ptrdiff_t stride);
71 void ff_hevc_add_residual_32x32_8_neon(uint8_t *_dst, const int16_t *coeffs,
72  ptrdiff_t stride);
73 void ff_hevc_add_residual_32x32_10_neon(uint8_t *_dst, const int16_t *coeffs,
74  ptrdiff_t stride);
75 void ff_hevc_add_residual_32x32_12_neon(uint8_t *_dst, const int16_t *coeffs,
76  ptrdiff_t stride);
77 void ff_hevc_idct_4x4_8_neon(int16_t *coeffs, int col_limit);
78 void ff_hevc_idct_4x4_10_neon(int16_t *coeffs, int col_limit);
79 void ff_hevc_idct_8x8_8_neon(int16_t *coeffs, int col_limit);
80 void ff_hevc_idct_8x8_10_neon(int16_t *coeffs, int col_limit);
81 void ff_hevc_idct_16x16_8_neon(int16_t *coeffs, int col_limit);
82 void ff_hevc_idct_16x16_10_neon(int16_t *coeffs, int col_limit);
83 void ff_hevc_idct_32x32_8_neon(int16_t *coeffs, int col_limit);
84 void ff_hevc_idct_32x32_10_neon(int16_t *coeffs, int col_limit);
85 void ff_hevc_idct_4x4_dc_8_neon(int16_t *coeffs);
86 void ff_hevc_idct_8x8_dc_8_neon(int16_t *coeffs);
87 void ff_hevc_idct_16x16_dc_8_neon(int16_t *coeffs);
88 void ff_hevc_idct_32x32_dc_8_neon(int16_t *coeffs);
89 void ff_hevc_idct_4x4_dc_10_neon(int16_t *coeffs);
90 void ff_hevc_idct_8x8_dc_10_neon(int16_t *coeffs);
91 void ff_hevc_idct_16x16_dc_10_neon(int16_t *coeffs);
92 void ff_hevc_idct_32x32_dc_10_neon(int16_t *coeffs);
93 void ff_hevc_transform_luma_4x4_neon_8(int16_t *coeffs);
94 void ff_hevc_sao_band_filter_8x8_8_neon(uint8_t *_dst, const uint8_t *_src,
95  ptrdiff_t stride_dst, ptrdiff_t stride_src,
96  const int16_t *sao_offset_val, int sao_left_class,
97  int width, int height);
98 void ff_hevc_sao_edge_filter_16x16_8_neon(uint8_t *dst, const uint8_t *src, ptrdiff_t stride_dst,
99  const int16_t *sao_offset_val, int eo, int width, int height);
100 void ff_hevc_sao_edge_filter_8x8_8_neon(uint8_t *dst, const uint8_t *src, ptrdiff_t stride_dst,
101  const int16_t *sao_offset_val, int eo, int width, int height);
102 void ff_hevc_put_hevc_qpel_h4_8_neon(int16_t *dst, const uint8_t *_src, ptrdiff_t _srcstride, int height,
103  intptr_t mx, intptr_t my, int width);
104 void ff_hevc_put_hevc_qpel_h6_8_neon(int16_t *dst, const uint8_t *_src, ptrdiff_t _srcstride, int height,
105  intptr_t mx, intptr_t my, int width);
106 void ff_hevc_put_hevc_qpel_h8_8_neon(int16_t *dst, const uint8_t *_src, ptrdiff_t _srcstride, int height,
107  intptr_t mx, intptr_t my, int width);
108 void ff_hevc_put_hevc_qpel_h12_8_neon(int16_t *dst, const uint8_t *_src, ptrdiff_t _srcstride, int height,
109  intptr_t mx, intptr_t my, int width);
110 void ff_hevc_put_hevc_qpel_h16_8_neon(int16_t *dst, const uint8_t *_src, ptrdiff_t _srcstride, int height,
111  intptr_t mx, intptr_t my, int width);
112 void ff_hevc_put_hevc_qpel_h32_8_neon(int16_t *dst, const uint8_t *_src, ptrdiff_t _srcstride, int height,
113  intptr_t mx, intptr_t my, int width);
114 void ff_hevc_put_hevc_qpel_uni_h4_8_neon(uint8_t *_dst, ptrdiff_t _dststride, const uint8_t *_src,
115  ptrdiff_t _srcstride, int height, intptr_t mx, intptr_t my,
116  int width);
117 void ff_hevc_put_hevc_qpel_uni_h6_8_neon(uint8_t *_dst, ptrdiff_t _dststride, const uint8_t *_src,
118  ptrdiff_t _srcstride, int height, intptr_t mx, intptr_t my,
119  int width);
120 void ff_hevc_put_hevc_qpel_uni_h8_8_neon(uint8_t *_dst, ptrdiff_t _dststride, const uint8_t *_src,
121  ptrdiff_t _srcstride, int height, intptr_t mx, intptr_t my,
122  int width);
123 void ff_hevc_put_hevc_qpel_uni_h12_8_neon(uint8_t *_dst, ptrdiff_t _dststride, const uint8_t *_src,
124  ptrdiff_t _srcstride, int height, intptr_t mx, intptr_t
125  my, int width);
126 void ff_hevc_put_hevc_qpel_uni_h16_8_neon(uint8_t *_dst, ptrdiff_t _dststride, const uint8_t *_src,
127  ptrdiff_t _srcstride, int height, intptr_t mx, intptr_t
128  my, int width);
129 void ff_hevc_put_hevc_qpel_uni_h32_8_neon(uint8_t *_dst, ptrdiff_t _dststride, const uint8_t *_src,
130  ptrdiff_t _srcstride, int height, intptr_t mx, intptr_t
131  my, int width);
132 void ff_hevc_put_hevc_qpel_bi_h4_8_neon(uint8_t *_dst, ptrdiff_t _dststride, const uint8_t *_src,
133  ptrdiff_t _srcstride, const int16_t *src2, int height, intptr_t
134  mx, intptr_t my, int width);
135 void ff_hevc_put_hevc_qpel_bi_h6_8_neon(uint8_t *_dst, ptrdiff_t _dststride, const uint8_t *_src,
136  ptrdiff_t _srcstride, const int16_t *src2, int height, intptr_t
137  mx, intptr_t my, int width);
138 void ff_hevc_put_hevc_qpel_bi_h8_8_neon(uint8_t *_dst, ptrdiff_t _dststride, const uint8_t *_src,
139  ptrdiff_t _srcstride, const int16_t *src2, int height, intptr_t
140  mx, intptr_t my, int width);
141 void ff_hevc_put_hevc_qpel_bi_h12_8_neon(uint8_t *_dst, ptrdiff_t _dststride, const uint8_t *_src,
142  ptrdiff_t _srcstride, const int16_t *src2, int height, intptr_t
143  mx, intptr_t my, int width);
144 void ff_hevc_put_hevc_qpel_bi_h16_8_neon(uint8_t *_dst, ptrdiff_t _dststride, const uint8_t *_src,
145  ptrdiff_t _srcstride, const int16_t *src2, int height, intptr_t
146  mx, intptr_t my, int width);
147 void ff_hevc_put_hevc_qpel_bi_h32_8_neon(uint8_t *_dst, ptrdiff_t _dststride, const uint8_t *_src,
148  ptrdiff_t _srcstride, const int16_t *src2, int height, intptr_t
149  mx, intptr_t my, int width);
150 
151 #define NEON8_FNPROTO(fn, args, ext) \
152  void ff_hevc_put_hevc_##fn##4_8_neon##ext args; \
153  void ff_hevc_put_hevc_##fn##6_8_neon##ext args; \
154  void ff_hevc_put_hevc_##fn##8_8_neon##ext args; \
155  void ff_hevc_put_hevc_##fn##12_8_neon##ext args; \
156  void ff_hevc_put_hevc_##fn##16_8_neon##ext args; \
157  void ff_hevc_put_hevc_##fn##24_8_neon##ext args; \
158  void ff_hevc_put_hevc_##fn##32_8_neon##ext args; \
159  void ff_hevc_put_hevc_##fn##48_8_neon##ext args; \
160  void ff_hevc_put_hevc_##fn##64_8_neon##ext args
161 
162 #define NEON8_FNPROTO_PARTIAL_4(fn, args, ext) \
163  void ff_hevc_put_hevc_##fn##4_8_neon##ext args; \
164  void ff_hevc_put_hevc_##fn##8_8_neon##ext args; \
165  void ff_hevc_put_hevc_##fn##16_8_neon##ext args; \
166  void ff_hevc_put_hevc_##fn##64_8_neon##ext args
167 
168 #define NEON8_FNPROTO_PARTIAL_5(fn, args, ext) \
169  void ff_hevc_put_hevc_##fn##4_8_neon##ext args; \
170  void ff_hevc_put_hevc_##fn##8_8_neon##ext args; \
171  void ff_hevc_put_hevc_##fn##16_8_neon##ext args; \
172  void ff_hevc_put_hevc_##fn##32_8_neon##ext args; \
173  void ff_hevc_put_hevc_##fn##64_8_neon##ext args
174 
175 NEON8_FNPROTO(pel_pixels, (int16_t *dst,
176  const uint8_t *src, ptrdiff_t srcstride,
177  int height, intptr_t mx, intptr_t my, int width),);
178 
179 NEON8_FNPROTO(pel_bi_pixels, (uint8_t *dst, ptrdiff_t dststride,
180  const uint8_t *_src, ptrdiff_t _srcstride, const int16_t *src2,
181  int height, intptr_t mx, intptr_t my, int width),);
182 
183 NEON8_FNPROTO(epel_bi_h, (uint8_t *dst, ptrdiff_t dststride,
184  const uint8_t *src, ptrdiff_t srcstride, const int16_t *src2,
185  int height, intptr_t mx, intptr_t my, int width),);
186 
187 NEON8_FNPROTO(epel_bi_v, (uint8_t *dst, ptrdiff_t dststride,
188  const uint8_t *src, ptrdiff_t srcstride, const int16_t *src2,
189  int height, intptr_t mx, intptr_t my, int width),);
190 
191 NEON8_FNPROTO(epel_bi_hv, (uint8_t *dst, ptrdiff_t dststride,
192  const uint8_t *src, ptrdiff_t srcstride, const int16_t *src2,
193  int height, intptr_t mx, intptr_t my, int width),);
194 
195 NEON8_FNPROTO(epel_bi_hv, (uint8_t *dst, ptrdiff_t dststride,
196  const uint8_t *src, ptrdiff_t srcstride, const int16_t *src2,
197  int height, intptr_t mx, intptr_t my, int width), _i8mm);
198 
199 NEON8_FNPROTO(epel_v, (int16_t *dst,
200  const uint8_t *src, ptrdiff_t srcstride,
201  int height, intptr_t mx, intptr_t my, int width),);
202 
203 NEON8_FNPROTO(pel_uni_pixels, (uint8_t *_dst, ptrdiff_t _dststride,
204  const uint8_t *_src, ptrdiff_t _srcstride,
205  int height, intptr_t mx, intptr_t my, int width),);
206 
207 NEON8_FNPROTO(pel_uni_w_pixels, (uint8_t *_dst, ptrdiff_t _dststride,
208  const uint8_t *_src, ptrdiff_t _srcstride,
209  int height, int denom, int wx, int ox,
210  intptr_t mx, intptr_t my, int width),);
211 
212 NEON8_FNPROTO(epel_uni_v, (uint8_t *dst, ptrdiff_t dststride,
213  const uint8_t *src, ptrdiff_t srcstride,
214  int height, intptr_t mx, intptr_t my, int width),);
215 
216 NEON8_FNPROTO(epel_uni_hv, (uint8_t *dst, ptrdiff_t _dststride,
217  const uint8_t *src, ptrdiff_t srcstride,
218  int height, intptr_t mx, intptr_t my, int width),);
219 
220 NEON8_FNPROTO(epel_uni_hv, (uint8_t *dst, ptrdiff_t _dststride,
221  const uint8_t *src, ptrdiff_t srcstride,
222  int height, intptr_t mx, intptr_t my, int width), _i8mm);
223 
224 NEON8_FNPROTO(epel_uni_w_v, (uint8_t *_dst, ptrdiff_t _dststride,
225  const uint8_t *_src, ptrdiff_t _srcstride,
226  int height, int denom, int wx, int ox,
227  intptr_t mx, intptr_t my, int width),);
228 
229 NEON8_FNPROTO_PARTIAL_4(qpel_uni_w_v, (uint8_t *_dst, ptrdiff_t _dststride,
230  const uint8_t *_src, ptrdiff_t _srcstride,
231  int height, int denom, int wx, int ox,
232  intptr_t mx, intptr_t my, int width),);
233 
234 NEON8_FNPROTO(epel_h, (int16_t *dst,
235  const uint8_t *_src, ptrdiff_t _srcstride,
236  int height, intptr_t mx, intptr_t my, int width),);
237 
238 NEON8_FNPROTO(epel_hv, (int16_t *dst,
239  const uint8_t *src, ptrdiff_t srcstride,
240  int height, intptr_t mx, intptr_t my, int width), );
241 
242 NEON8_FNPROTO(epel_h, (int16_t *dst,
243  const uint8_t *_src, ptrdiff_t _srcstride,
244  int height, intptr_t mx, intptr_t my, int width), _i8mm);
245 
246 NEON8_FNPROTO(epel_hv, (int16_t *dst,
247  const uint8_t *src, ptrdiff_t srcstride,
248  int height, intptr_t mx, intptr_t my, int width), _i8mm);
249 
250 NEON8_FNPROTO(epel_uni_w_h, (uint8_t *_dst, ptrdiff_t _dststride,
251  const uint8_t *_src, ptrdiff_t _srcstride,
252  int height, int denom, int wx, int ox,
253  intptr_t mx, intptr_t my, int width),);
254 
255 NEON8_FNPROTO(epel_uni_w_h, (uint8_t *_dst, ptrdiff_t _dststride,
256  const uint8_t *_src, ptrdiff_t _srcstride,
257  int height, int denom, int wx, int ox,
258  intptr_t mx, intptr_t my, int width), _i8mm);
259 
260 NEON8_FNPROTO(qpel_h, (int16_t *dst,
261  const uint8_t *_src, ptrdiff_t _srcstride,
262  int height, intptr_t mx, intptr_t my, int width), _i8mm);
263 
264 NEON8_FNPROTO(qpel_v, (int16_t *dst,
265  const uint8_t *src, ptrdiff_t srcstride,
266  int height, intptr_t mx, intptr_t my, int width),);
267 
268 NEON8_FNPROTO(qpel_hv, (int16_t *dst,
269  const uint8_t *src, ptrdiff_t srcstride,
270  int height, intptr_t mx, intptr_t my, int width),);
271 
272 NEON8_FNPROTO(qpel_hv, (int16_t *dst,
273  const uint8_t *src, ptrdiff_t srcstride,
274  int height, intptr_t mx, intptr_t my, int width), _i8mm);
275 
276 NEON8_FNPROTO(qpel_uni_v, (uint8_t *dst, ptrdiff_t dststride,
277  const uint8_t *src, ptrdiff_t srcstride,
278  int height, intptr_t mx, intptr_t my, int width),);
279 
280 NEON8_FNPROTO(qpel_uni_hv, (uint8_t *dst, ptrdiff_t dststride,
281  const uint8_t *src, ptrdiff_t srcstride,
282  int height, intptr_t mx, intptr_t my, int width),);
283 
284 NEON8_FNPROTO(qpel_uni_hv, (uint8_t *dst, ptrdiff_t dststride,
285  const uint8_t *src, ptrdiff_t srcstride,
286  int height, intptr_t mx, intptr_t my, int width), _i8mm);
287 
288 NEON8_FNPROTO(qpel_uni_w_h, (uint8_t *_dst, ptrdiff_t _dststride,
289  const uint8_t *_src, ptrdiff_t _srcstride,
290  int height, int denom, int wx, int ox,
291  intptr_t mx, intptr_t my, int width),);
292 
293 NEON8_FNPROTO(qpel_uni_w_h, (uint8_t *_dst, ptrdiff_t _dststride,
294  const uint8_t *_src, ptrdiff_t _srcstride,
295  int height, int denom, int wx, int ox,
296  intptr_t mx, intptr_t my, int width), _i8mm);
297 
298 NEON8_FNPROTO(epel_uni_w_hv, (uint8_t *_dst, ptrdiff_t _dststride,
299  const uint8_t *_src, ptrdiff_t _srcstride,
300  int height, int denom, int wx, int ox,
301  intptr_t mx, intptr_t my, int width),);
302 
303 NEON8_FNPROTO(epel_uni_w_hv, (uint8_t *_dst, ptrdiff_t _dststride,
304  const uint8_t *_src, ptrdiff_t _srcstride,
305  int height, int denom, int wx, int ox,
306  intptr_t mx, intptr_t my, int width), _i8mm);
307 
308 NEON8_FNPROTO_PARTIAL_5(qpel_uni_w_hv, (uint8_t *_dst, ptrdiff_t _dststride,
309  const uint8_t *_src, ptrdiff_t _srcstride,
310  int height, int denom, int wx, int ox,
311  intptr_t mx, intptr_t my, int width),);
312 
313 NEON8_FNPROTO_PARTIAL_5(qpel_uni_w_hv, (uint8_t *_dst, ptrdiff_t _dststride,
314  const uint8_t *_src, ptrdiff_t _srcstride,
315  int height, int denom, int wx, int ox,
316  intptr_t mx, intptr_t my, int width), _i8mm);
317 
318 NEON8_FNPROTO(qpel_bi_v, (uint8_t *dst, ptrdiff_t dststride,
319  const uint8_t *src, ptrdiff_t srcstride, const int16_t *src2,
320  int height, intptr_t mx, intptr_t my, int width),);
321 
322 NEON8_FNPROTO(qpel_bi_hv, (uint8_t *dst, ptrdiff_t dststride,
323  const uint8_t *src, ptrdiff_t srcstride, const int16_t *src2,
324  int height, intptr_t mx, intptr_t my, int width),);
325 
326 NEON8_FNPROTO(qpel_bi_hv, (uint8_t *dst, ptrdiff_t dststride,
327  const uint8_t *src, ptrdiff_t srcstride, const int16_t *src2,
328  int height, intptr_t mx, intptr_t my, int width), _i8mm);
329 
330 #define NEON8_FNASSIGN(member, v, h, fn, ext) \
331  member[1][v][h] = ff_hevc_put_hevc_##fn##4_8_neon##ext; \
332  member[2][v][h] = ff_hevc_put_hevc_##fn##6_8_neon##ext; \
333  member[3][v][h] = ff_hevc_put_hevc_##fn##8_8_neon##ext; \
334  member[4][v][h] = ff_hevc_put_hevc_##fn##12_8_neon##ext; \
335  member[5][v][h] = ff_hevc_put_hevc_##fn##16_8_neon##ext; \
336  member[6][v][h] = ff_hevc_put_hevc_##fn##24_8_neon##ext; \
337  member[7][v][h] = ff_hevc_put_hevc_##fn##32_8_neon##ext; \
338  member[8][v][h] = ff_hevc_put_hevc_##fn##48_8_neon##ext; \
339  member[9][v][h] = ff_hevc_put_hevc_##fn##64_8_neon##ext;
340 
341 #define NEON8_FNASSIGN_SHARED_32(member, v, h, fn, ext) \
342  member[1][v][h] = ff_hevc_put_hevc_##fn##4_8_neon##ext; \
343  member[2][v][h] = ff_hevc_put_hevc_##fn##6_8_neon##ext; \
344  member[3][v][h] = ff_hevc_put_hevc_##fn##8_8_neon##ext; \
345  member[4][v][h] = ff_hevc_put_hevc_##fn##12_8_neon##ext; \
346  member[5][v][h] = ff_hevc_put_hevc_##fn##16_8_neon##ext; \
347  member[6][v][h] = ff_hevc_put_hevc_##fn##24_8_neon##ext; \
348  member[7][v][h] = \
349  member[8][v][h] = \
350  member[9][v][h] = ff_hevc_put_hevc_##fn##32_8_neon##ext;
351 
352 #define NEON8_FNASSIGN_PARTIAL_4(member, v, h, fn, ext) \
353  member[1][v][h] = ff_hevc_put_hevc_##fn##4_8_neon##ext; \
354  member[3][v][h] = ff_hevc_put_hevc_##fn##8_8_neon##ext; \
355  member[5][v][h] = ff_hevc_put_hevc_##fn##16_8_neon##ext; \
356  member[7][v][h] = ff_hevc_put_hevc_##fn##64_8_neon##ext; \
357  member[8][v][h] = ff_hevc_put_hevc_##fn##64_8_neon##ext; \
358  member[9][v][h] = ff_hevc_put_hevc_##fn##64_8_neon##ext;
359 
360 #define NEON8_FNASSIGN_PARTIAL_5(member, v, h, fn, ext) \
361  member[1][v][h] = ff_hevc_put_hevc_##fn##4_8_neon##ext; \
362  member[3][v][h] = ff_hevc_put_hevc_##fn##8_8_neon##ext; \
363  member[5][v][h] = ff_hevc_put_hevc_##fn##16_8_neon##ext; \
364  member[7][v][h] = ff_hevc_put_hevc_##fn##32_8_neon##ext; \
365  member[9][v][h] = ff_hevc_put_hevc_##fn##64_8_neon##ext;
366 
368 {
369  int cpu_flags = av_get_cpu_flags();
370  if (!have_neon(cpu_flags)) return;
371 
372  if (bit_depth == 8) {
373  c->hevc_h_loop_filter_luma = ff_hevc_h_loop_filter_luma_8_neon;
374  c->hevc_v_loop_filter_luma = ff_hevc_v_loop_filter_luma_8_neon;
375  c->hevc_h_loop_filter_chroma = ff_hevc_h_loop_filter_chroma_8_neon;
376  c->hevc_v_loop_filter_chroma = ff_hevc_v_loop_filter_chroma_8_neon;
377  c->add_residual[0] = ff_hevc_add_residual_4x4_8_neon;
378  c->add_residual[1] = ff_hevc_add_residual_8x8_8_neon;
379  c->add_residual[2] = ff_hevc_add_residual_16x16_8_neon;
380  c->add_residual[3] = ff_hevc_add_residual_32x32_8_neon;
381  c->idct[0] = ff_hevc_idct_4x4_8_neon;
382  c->idct[1] = ff_hevc_idct_8x8_8_neon;
383  c->idct[2] = ff_hevc_idct_16x16_8_neon;
384  c->idct[3] = ff_hevc_idct_32x32_8_neon;
385  c->idct_dc[0] = ff_hevc_idct_4x4_dc_8_neon;
386  c->idct_dc[1] = ff_hevc_idct_8x8_dc_8_neon;
387  c->idct_dc[2] = ff_hevc_idct_16x16_dc_8_neon;
388  c->idct_dc[3] = ff_hevc_idct_32x32_dc_8_neon;
389  c->transform_4x4_luma = ff_hevc_transform_luma_4x4_neon_8;
390  c->sao_band_filter[0] =
391  c->sao_band_filter[1] =
392  c->sao_band_filter[2] =
393  c->sao_band_filter[3] =
394  c->sao_band_filter[4] = ff_hevc_sao_band_filter_8x8_8_neon;
395  c->sao_edge_filter[0] = ff_hevc_sao_edge_filter_8x8_8_neon;
396  c->sao_edge_filter[1] =
397  c->sao_edge_filter[2] =
398  c->sao_edge_filter[3] =
399  c->sao_edge_filter[4] = ff_hevc_sao_edge_filter_16x16_8_neon;
400  c->put_hevc_qpel[1][0][1] = ff_hevc_put_hevc_qpel_h4_8_neon;
401  c->put_hevc_qpel[2][0][1] = ff_hevc_put_hevc_qpel_h6_8_neon;
402  c->put_hevc_qpel[3][0][1] = ff_hevc_put_hevc_qpel_h8_8_neon;
403  c->put_hevc_qpel[4][0][1] =
404  c->put_hevc_qpel[6][0][1] = ff_hevc_put_hevc_qpel_h12_8_neon;
405  c->put_hevc_qpel[5][0][1] = ff_hevc_put_hevc_qpel_h16_8_neon;
406  c->put_hevc_qpel[7][0][1] =
407  c->put_hevc_qpel[8][0][1] =
408  c->put_hevc_qpel[9][0][1] = ff_hevc_put_hevc_qpel_h32_8_neon;
409  c->put_hevc_qpel_uni[1][0][1] = ff_hevc_put_hevc_qpel_uni_h4_8_neon;
410  c->put_hevc_qpel_uni[2][0][1] = ff_hevc_put_hevc_qpel_uni_h6_8_neon;
411  c->put_hevc_qpel_uni[3][0][1] = ff_hevc_put_hevc_qpel_uni_h8_8_neon;
412  c->put_hevc_qpel_uni[4][0][1] =
413  c->put_hevc_qpel_uni[6][0][1] = ff_hevc_put_hevc_qpel_uni_h12_8_neon;
414  c->put_hevc_qpel_uni[5][0][1] = ff_hevc_put_hevc_qpel_uni_h16_8_neon;
415  c->put_hevc_qpel_uni[7][0][1] =
416  c->put_hevc_qpel_uni[8][0][1] =
417  c->put_hevc_qpel_uni[9][0][1] = ff_hevc_put_hevc_qpel_uni_h32_8_neon;
418  c->put_hevc_qpel_bi[1][0][1] = ff_hevc_put_hevc_qpel_bi_h4_8_neon;
419  c->put_hevc_qpel_bi[2][0][1] = ff_hevc_put_hevc_qpel_bi_h6_8_neon;
420  c->put_hevc_qpel_bi[3][0][1] = ff_hevc_put_hevc_qpel_bi_h8_8_neon;
421  c->put_hevc_qpel_bi[4][0][1] =
422  c->put_hevc_qpel_bi[6][0][1] = ff_hevc_put_hevc_qpel_bi_h12_8_neon;
423  c->put_hevc_qpel_bi[5][0][1] = ff_hevc_put_hevc_qpel_bi_h16_8_neon;
424  c->put_hevc_qpel_bi[7][0][1] =
425  c->put_hevc_qpel_bi[8][0][1] =
426  c->put_hevc_qpel_bi[9][0][1] = ff_hevc_put_hevc_qpel_bi_h32_8_neon;
427 
428  NEON8_FNASSIGN(c->put_hevc_epel, 0, 0, pel_pixels,);
429  NEON8_FNASSIGN(c->put_hevc_epel, 1, 0, epel_v,);
430  NEON8_FNASSIGN(c->put_hevc_qpel, 0, 0, pel_pixels,);
431  NEON8_FNASSIGN(c->put_hevc_qpel, 1, 0, qpel_v,);
432  NEON8_FNASSIGN(c->put_hevc_epel_bi, 0, 0, pel_bi_pixels,);
433  NEON8_FNASSIGN(c->put_hevc_epel_bi, 0, 1, epel_bi_h,);
434  NEON8_FNASSIGN(c->put_hevc_epel_bi, 1, 0, epel_bi_v,);
435  NEON8_FNASSIGN(c->put_hevc_qpel_bi, 0, 0, pel_bi_pixels,);
436  NEON8_FNASSIGN(c->put_hevc_qpel_bi, 1, 0, qpel_bi_v,);
437  NEON8_FNASSIGN(c->put_hevc_epel_uni, 0, 0, pel_uni_pixels,);
438  NEON8_FNASSIGN(c->put_hevc_epel_uni, 1, 0, epel_uni_v,);
439  NEON8_FNASSIGN(c->put_hevc_qpel_uni, 0, 0, pel_uni_pixels,);
440  NEON8_FNASSIGN(c->put_hevc_qpel_uni, 1, 0, qpel_uni_v,);
441  NEON8_FNASSIGN(c->put_hevc_epel_uni_w, 0, 0, pel_uni_w_pixels,);
442  NEON8_FNASSIGN(c->put_hevc_qpel_uni_w, 0, 0, pel_uni_w_pixels,);
443  NEON8_FNASSIGN(c->put_hevc_epel_uni_w, 1, 0, epel_uni_w_v,);
444  NEON8_FNASSIGN_PARTIAL_4(c->put_hevc_qpel_uni_w, 1, 0, qpel_uni_w_v,);
445 
446  NEON8_FNASSIGN_SHARED_32(c->put_hevc_epel, 0, 1, epel_h,);
447  NEON8_FNASSIGN_SHARED_32(c->put_hevc_epel_uni_w, 0, 1, epel_uni_w_h,);
448 
449  NEON8_FNASSIGN(c->put_hevc_epel, 1, 1, epel_hv,);
450  NEON8_FNASSIGN(c->put_hevc_epel_uni, 1, 1, epel_uni_hv,);
451  NEON8_FNASSIGN(c->put_hevc_epel_uni_w, 1, 1, epel_uni_w_hv,);
452  NEON8_FNASSIGN(c->put_hevc_epel_bi, 1, 1, epel_bi_hv,);
453 
454  NEON8_FNASSIGN_SHARED_32(c->put_hevc_qpel_uni_w, 0, 1, qpel_uni_w_h,);
455 
456  NEON8_FNASSIGN(c->put_hevc_qpel, 1, 1, qpel_hv,);
457  NEON8_FNASSIGN(c->put_hevc_qpel_uni, 1, 1, qpel_uni_hv,);
458  NEON8_FNASSIGN_PARTIAL_5(c->put_hevc_qpel_uni_w, 1, 1, qpel_uni_w_hv,);
459  NEON8_FNASSIGN(c->put_hevc_qpel_bi, 1, 1, qpel_bi_hv,);
460 
461  if (have_i8mm(cpu_flags)) {
462  NEON8_FNASSIGN(c->put_hevc_epel, 0, 1, epel_h, _i8mm);
463  NEON8_FNASSIGN(c->put_hevc_epel, 1, 1, epel_hv, _i8mm);
464  NEON8_FNASSIGN(c->put_hevc_epel_uni, 1, 1, epel_uni_hv, _i8mm);
465  NEON8_FNASSIGN(c->put_hevc_epel_uni_w, 0, 1, epel_uni_w_h ,_i8mm);
466  NEON8_FNASSIGN(c->put_hevc_epel_uni_w, 1, 1, epel_uni_w_hv, _i8mm);
467  NEON8_FNASSIGN(c->put_hevc_epel_bi, 1, 1, epel_bi_hv, _i8mm);
468  NEON8_FNASSIGN(c->put_hevc_qpel, 0, 1, qpel_h, _i8mm);
469  NEON8_FNASSIGN(c->put_hevc_qpel, 1, 1, qpel_hv, _i8mm);
470  NEON8_FNASSIGN(c->put_hevc_qpel_uni, 1, 1, qpel_uni_hv, _i8mm);
471  NEON8_FNASSIGN(c->put_hevc_qpel_uni_w, 0, 1, qpel_uni_w_h, _i8mm);
472  NEON8_FNASSIGN_PARTIAL_5(c->put_hevc_qpel_uni_w, 1, 1, qpel_uni_w_hv, _i8mm);
473  NEON8_FNASSIGN(c->put_hevc_qpel_bi, 1, 1, qpel_bi_hv, _i8mm);
474  }
475 
476  }
477  if (bit_depth == 10) {
478  c->hevc_h_loop_filter_luma = ff_hevc_h_loop_filter_luma_10_neon;
479  c->hevc_v_loop_filter_luma = ff_hevc_v_loop_filter_luma_10_neon;
480  c->hevc_h_loop_filter_chroma = ff_hevc_h_loop_filter_chroma_10_neon;
481  c->hevc_v_loop_filter_chroma = ff_hevc_v_loop_filter_chroma_10_neon;
482  c->add_residual[0] = ff_hevc_add_residual_4x4_10_neon;
483  c->add_residual[1] = ff_hevc_add_residual_8x8_10_neon;
484  c->add_residual[2] = ff_hevc_add_residual_16x16_10_neon;
485  c->add_residual[3] = ff_hevc_add_residual_32x32_10_neon;
486  c->idct[0] = ff_hevc_idct_4x4_10_neon;
487  c->idct[1] = ff_hevc_idct_8x8_10_neon;
488  c->idct[2] = ff_hevc_idct_16x16_10_neon;
489  c->idct[3] = ff_hevc_idct_32x32_10_neon;
490  c->idct_dc[0] = ff_hevc_idct_4x4_dc_10_neon;
491  c->idct_dc[1] = ff_hevc_idct_8x8_dc_10_neon;
492  c->idct_dc[2] = ff_hevc_idct_16x16_dc_10_neon;
493  c->idct_dc[3] = ff_hevc_idct_32x32_dc_10_neon;
494  }
495  if (bit_depth == 12) {
496  c->hevc_h_loop_filter_luma = ff_hevc_h_loop_filter_luma_12_neon;
497  c->hevc_v_loop_filter_luma = ff_hevc_v_loop_filter_luma_12_neon;
498  c->hevc_h_loop_filter_chroma = ff_hevc_h_loop_filter_chroma_12_neon;
499  c->hevc_v_loop_filter_chroma = ff_hevc_v_loop_filter_chroma_12_neon;
500  c->add_residual[0] = ff_hevc_add_residual_4x4_12_neon;
501  c->add_residual[1] = ff_hevc_add_residual_8x8_12_neon;
502  c->add_residual[2] = ff_hevc_add_residual_16x16_12_neon;
503  c->add_residual[3] = ff_hevc_add_residual_32x32_12_neon;
504  }
505 }
NEON8_FNPROTO_PARTIAL_5
#define NEON8_FNPROTO_PARTIAL_5(fn, args, ext)
Definition: hevcdsp_init_aarch64.c:168
ff_hevc_v_loop_filter_chroma_10_neon
void ff_hevc_v_loop_filter_chroma_10_neon(uint8_t *_pix, ptrdiff_t _stride, const int *_tc, const uint8_t *_no_p, const uint8_t *_no_q)
NEON8_FNASSIGN
#define NEON8_FNASSIGN(member, v, h, fn, ext)
Definition: hevcdsp_init_aarch64.c:330
ff_hevc_idct_4x4_dc_8_neon
void ff_hevc_idct_4x4_dc_8_neon(int16_t *coeffs)
ff_hevc_put_hevc_qpel_h4_8_neon
void ff_hevc_put_hevc_qpel_h4_8_neon(int16_t *dst, const uint8_t *_src, ptrdiff_t _srcstride, int height, intptr_t mx, intptr_t my, int width)
ff_hevc_put_hevc_qpel_bi_h6_8_neon
void ff_hevc_put_hevc_qpel_bi_h6_8_neon(uint8_t *_dst, ptrdiff_t _dststride, const uint8_t *_src, ptrdiff_t _srcstride, const int16_t *src2, int height, intptr_t mx, intptr_t my, int width)
ff_hevc_transform_luma_4x4_neon_8
void ff_hevc_transform_luma_4x4_neon_8(int16_t *coeffs)
ff_hevc_h_loop_filter_luma_10_neon
void ff_hevc_h_loop_filter_luma_10_neon(uint8_t *_pix, ptrdiff_t _stride, int beta, const int *_tc, const uint8_t *_no_p, const uint8_t *_no_q)
ff_hevc_sao_edge_filter_8x8_8_neon
void ff_hevc_sao_edge_filter_8x8_8_neon(uint8_t *dst, const uint8_t *src, ptrdiff_t stride_dst, const int16_t *sao_offset_val, int eo, int width, int height)
_stride
ptrdiff_t _stride
Definition: h264pred_template.c:411
ff_hevc_idct_16x16_8_neon
void ff_hevc_idct_16x16_8_neon(int16_t *coeffs, int col_limit)
ff_hevc_put_hevc_qpel_bi_h12_8_neon
void ff_hevc_put_hevc_qpel_bi_h12_8_neon(uint8_t *_dst, ptrdiff_t _dststride, const uint8_t *_src, ptrdiff_t _srcstride, const int16_t *src2, int height, intptr_t mx, intptr_t my, int width)
ff_hevc_put_hevc_qpel_h8_8_neon
void ff_hevc_put_hevc_qpel_h8_8_neon(int16_t *dst, const uint8_t *_src, ptrdiff_t _srcstride, int height, intptr_t mx, intptr_t my, int width)
av_get_cpu_flags
int av_get_cpu_flags(void)
Return the flags which specify extensions supported by the CPU.
Definition: cpu.c:103
ff_hevc_v_loop_filter_luma_8_neon
void ff_hevc_v_loop_filter_luma_8_neon(uint8_t *_pix, ptrdiff_t _stride, int beta, const int *_tc, const uint8_t *_no_p, const uint8_t *_no_q)
bit_depth
static void bit_depth(AudioStatsContext *s, const uint64_t *const mask, uint8_t *depth)
Definition: af_astats.c:246
cpu_flags
static atomic_int cpu_flags
Definition: cpu.c:52
ff_hevc_add_residual_16x16_10_neon
void ff_hevc_add_residual_16x16_10_neon(uint8_t *_dst, const int16_t *coeffs, ptrdiff_t stride)
ff_hevc_idct_32x32_8_neon
void ff_hevc_idct_32x32_8_neon(int16_t *coeffs, int col_limit)
ff_hevc_add_residual_4x4_10_neon
void ff_hevc_add_residual_4x4_10_neon(uint8_t *_dst, const int16_t *coeffs, ptrdiff_t stride)
ff_hevc_idct_8x8_8_neon
void ff_hevc_idct_8x8_8_neon(int16_t *coeffs, int col_limit)
ff_hevc_idct_16x16_dc_10_neon
void ff_hevc_idct_16x16_dc_10_neon(int16_t *coeffs)
ff_hevc_h_loop_filter_luma_8_neon
void ff_hevc_h_loop_filter_luma_8_neon(uint8_t *_pix, ptrdiff_t _stride, int beta, const int *_tc, const uint8_t *_no_p, const uint8_t *_no_q)
av_cold
#define av_cold
Definition: attributes.h:90
ff_hevc_h_loop_filter_luma_12_neon
void ff_hevc_h_loop_filter_luma_12_neon(uint8_t *_pix, ptrdiff_t _stride, int beta, const int *_tc, const uint8_t *_no_p, const uint8_t *_no_q)
width
#define width
ff_hevc_idct_32x32_10_neon
void ff_hevc_idct_32x32_10_neon(int16_t *coeffs, int col_limit)
ff_hevc_add_residual_32x32_12_neon
void ff_hevc_add_residual_32x32_12_neon(uint8_t *_dst, const int16_t *coeffs, ptrdiff_t stride)
ff_hevc_put_hevc_qpel_uni_h8_8_neon
void ff_hevc_put_hevc_qpel_uni_h8_8_neon(uint8_t *_dst, ptrdiff_t _dststride, const uint8_t *_src, ptrdiff_t _srcstride, int height, intptr_t mx, intptr_t my, int width)
ff_hevc_v_loop_filter_luma_10_neon
void ff_hevc_v_loop_filter_luma_10_neon(uint8_t *_pix, ptrdiff_t _stride, int beta, const int *_tc, const uint8_t *_no_p, const uint8_t *_no_q)
ff_hevc_put_hevc_qpel_h32_8_neon
void ff_hevc_put_hevc_qpel_h32_8_neon(int16_t *dst, const uint8_t *_src, ptrdiff_t _srcstride, int height, intptr_t mx, intptr_t my, int width)
ff_hevc_put_hevc_qpel_uni_h12_8_neon
void ff_hevc_put_hevc_qpel_uni_h12_8_neon(uint8_t *_dst, ptrdiff_t _dststride, const uint8_t *_src, ptrdiff_t _srcstride, int height, intptr_t mx, intptr_t my, int width)
ff_hevc_put_hevc_qpel_bi_h8_8_neon
void ff_hevc_put_hevc_qpel_bi_h8_8_neon(uint8_t *_dst, ptrdiff_t _dststride, const uint8_t *_src, ptrdiff_t _srcstride, const int16_t *src2, int height, intptr_t mx, intptr_t my, int width)
NEON8_FNASSIGN_PARTIAL_5
#define NEON8_FNASSIGN_PARTIAL_5(member, v, h, fn, ext)
Definition: hevcdsp_init_aarch64.c:360
ff_hevc_put_hevc_qpel_bi_h4_8_neon
void ff_hevc_put_hevc_qpel_bi_h4_8_neon(uint8_t *_dst, ptrdiff_t _dststride, const uint8_t *_src, ptrdiff_t _srcstride, const int16_t *src2, int height, intptr_t mx, intptr_t my, int width)
ff_hevc_idct_16x16_dc_8_neon
void ff_hevc_idct_16x16_dc_8_neon(int16_t *coeffs)
ff_hevc_put_hevc_qpel_bi_h32_8_neon
void ff_hevc_put_hevc_qpel_bi_h32_8_neon(uint8_t *_dst, ptrdiff_t _dststride, const uint8_t *_src, ptrdiff_t _srcstride, const int16_t *src2, int height, intptr_t mx, intptr_t my, int width)
ff_hevc_idct_32x32_dc_10_neon
void ff_hevc_idct_32x32_dc_10_neon(int16_t *coeffs)
c
Undefined Behavior In the C some operations are like signed integer dereferencing freed accessing outside allocated Undefined Behavior must not occur in a C it is not safe even if the output of undefined operations is unused The unsafety may seem nit picking but Optimizing compilers have in fact optimized code on the assumption that no undefined Behavior occurs Optimizing code based on wrong assumptions can and has in some cases lead to effects beyond the output of computations The signed integer overflow problem in speed critical code Code which is highly optimized and works with signed integers sometimes has the problem that often the output of the computation does not c
Definition: undefined.txt:32
have_i8mm
#define have_i8mm(flags)
Definition: cpu.h:29
ff_hevc_put_hevc_qpel_uni_h4_8_neon
void ff_hevc_put_hevc_qpel_uni_h4_8_neon(uint8_t *_dst, ptrdiff_t _dststride, const uint8_t *_src, ptrdiff_t _srcstride, int height, intptr_t mx, intptr_t my, int width)
ff_hevc_put_hevc_qpel_uni_h32_8_neon
void ff_hevc_put_hevc_qpel_uni_h32_8_neon(uint8_t *_dst, ptrdiff_t _dststride, const uint8_t *_src, ptrdiff_t _srcstride, int height, intptr_t mx, intptr_t my, int width)
ff_hevc_sao_band_filter_8x8_8_neon
void ff_hevc_sao_band_filter_8x8_8_neon(uint8_t *_dst, const uint8_t *_src, ptrdiff_t stride_dst, ptrdiff_t stride_src, const int16_t *sao_offset_val, int sao_left_class, int width, int height)
dsp.h
ff_hevc_add_residual_8x8_8_neon
void ff_hevc_add_residual_8x8_8_neon(uint8_t *_dst, const int16_t *coeffs, ptrdiff_t stride)
ff_hevc_v_loop_filter_chroma_12_neon
void ff_hevc_v_loop_filter_chroma_12_neon(uint8_t *_pix, ptrdiff_t _stride, const int *_tc, const uint8_t *_no_p, const uint8_t *_no_q)
cpu.h
ff_hevc_add_residual_16x16_12_neon
void ff_hevc_add_residual_16x16_12_neon(uint8_t *_dst, const int16_t *coeffs, ptrdiff_t stride)
ff_hevc_idct_4x4_10_neon
void ff_hevc_idct_4x4_10_neon(int16_t *coeffs, int col_limit)
ff_hevc_add_residual_8x8_12_neon
void ff_hevc_add_residual_8x8_12_neon(uint8_t *_dst, const int16_t *coeffs, ptrdiff_t stride)
ff_hevc_put_hevc_qpel_h6_8_neon
void ff_hevc_put_hevc_qpel_h6_8_neon(int16_t *dst, const uint8_t *_src, ptrdiff_t _srcstride, int height, intptr_t mx, intptr_t my, int width)
ff_hevc_idct_4x4_dc_10_neon
void ff_hevc_idct_4x4_dc_10_neon(int16_t *coeffs)
have_neon
#define have_neon(flags)
Definition: cpu.h:26
ff_hevc_h_loop_filter_chroma_12_neon
void ff_hevc_h_loop_filter_chroma_12_neon(uint8_t *_pix, ptrdiff_t _stride, const int *_tc, const uint8_t *_no_p, const uint8_t *_no_q)
NEON8_FNPROTO
#define NEON8_FNPROTO(fn, args, ext)
Definition: hevcdsp_init_aarch64.c:151
height
#define height
ff_hevc_idct_8x8_10_neon
void ff_hevc_idct_8x8_10_neon(int16_t *coeffs, int col_limit)
HEVCDSPContext
Definition: dsp.h:47
attributes.h
ff_hevc_idct_4x4_8_neon
void ff_hevc_idct_4x4_8_neon(int16_t *coeffs, int col_limit)
ff_hevc_idct_32x32_dc_8_neon
void ff_hevc_idct_32x32_dc_8_neon(int16_t *coeffs)
src2
const pixel * src2
Definition: h264pred_template.c:422
ff_hevc_idct_8x8_dc_8_neon
void ff_hevc_idct_8x8_dc_8_neon(int16_t *coeffs)
ff_hevc_put_hevc_qpel_uni_h6_8_neon
void ff_hevc_put_hevc_qpel_uni_h6_8_neon(uint8_t *_dst, ptrdiff_t _dststride, const uint8_t *_src, ptrdiff_t _srcstride, int height, intptr_t mx, intptr_t my, int width)
ff_hevc_put_hevc_qpel_h12_8_neon
void ff_hevc_put_hevc_qpel_h12_8_neon(int16_t *dst, const uint8_t *_src, ptrdiff_t _srcstride, int height, intptr_t mx, intptr_t my, int width)
ff_hevc_put_hevc_qpel_uni_h16_8_neon
void ff_hevc_put_hevc_qpel_uni_h16_8_neon(uint8_t *_dst, ptrdiff_t _dststride, const uint8_t *_src, ptrdiff_t _srcstride, int height, intptr_t mx, intptr_t my, int width)
ff_hevc_idct_8x8_dc_10_neon
void ff_hevc_idct_8x8_dc_10_neon(int16_t *coeffs)
ff_hevc_dsp_init_aarch64
av_cold void ff_hevc_dsp_init_aarch64(HEVCDSPContext *c, const int bit_depth)
Definition: hevcdsp_init_aarch64.c:367
stride
#define stride
Definition: h264pred_template.c:537
ff_hevc_sao_edge_filter_16x16_8_neon
void ff_hevc_sao_edge_filter_16x16_8_neon(uint8_t *dst, const uint8_t *src, ptrdiff_t stride_dst, const int16_t *sao_offset_val, int eo, int width, int height)
NEON8_FNASSIGN_SHARED_32
#define NEON8_FNASSIGN_SHARED_32(member, v, h, fn, ext)
Definition: hevcdsp_init_aarch64.c:341
ff_hevc_add_residual_32x32_8_neon
void ff_hevc_add_residual_32x32_8_neon(uint8_t *_dst, const int16_t *coeffs, ptrdiff_t stride)
ff_hevc_v_loop_filter_luma_12_neon
void ff_hevc_v_loop_filter_luma_12_neon(uint8_t *_pix, ptrdiff_t _stride, int beta, const int *_tc, const uint8_t *_no_p, const uint8_t *_no_q)
ff_hevc_v_loop_filter_chroma_8_neon
void ff_hevc_v_loop_filter_chroma_8_neon(uint8_t *_pix, ptrdiff_t _stride, const int *_tc, const uint8_t *_no_p, const uint8_t *_no_q)
ff_hevc_put_hevc_qpel_h16_8_neon
void ff_hevc_put_hevc_qpel_h16_8_neon(int16_t *dst, const uint8_t *_src, ptrdiff_t _srcstride, int height, intptr_t mx, intptr_t my, int width)
ff_hevc_idct_16x16_10_neon
void ff_hevc_idct_16x16_10_neon(int16_t *coeffs, int col_limit)
src
INIT_CLIP pixel * src
Definition: h264pred_template.c:418
ff_hevc_h_loop_filter_chroma_10_neon
void ff_hevc_h_loop_filter_chroma_10_neon(uint8_t *_pix, ptrdiff_t _stride, const int *_tc, const uint8_t *_no_p, const uint8_t *_no_q)
ff_hevc_add_residual_4x4_8_neon
void ff_hevc_add_residual_4x4_8_neon(uint8_t *_dst, const int16_t *coeffs, ptrdiff_t stride)
cpu.h
ff_hevc_add_residual_32x32_10_neon
void ff_hevc_add_residual_32x32_10_neon(uint8_t *_dst, const int16_t *coeffs, ptrdiff_t stride)
NEON8_FNPROTO_PARTIAL_4
#define NEON8_FNPROTO_PARTIAL_4(fn, args, ext)
Definition: hevcdsp_init_aarch64.c:162
ff_hevc_add_residual_8x8_10_neon
void ff_hevc_add_residual_8x8_10_neon(uint8_t *_dst, const int16_t *coeffs, ptrdiff_t stride)
ff_hevc_h_loop_filter_chroma_8_neon
void ff_hevc_h_loop_filter_chroma_8_neon(uint8_t *_pix, ptrdiff_t _stride, const int *_tc, const uint8_t *_no_p, const uint8_t *_no_q)
ff_hevc_add_residual_16x16_8_neon
void ff_hevc_add_residual_16x16_8_neon(uint8_t *_dst, const int16_t *coeffs, ptrdiff_t stride)
NEON8_FNASSIGN_PARTIAL_4
#define NEON8_FNASSIGN_PARTIAL_4(member, v, h, fn, ext)
Definition: hevcdsp_init_aarch64.c:352
ff_hevc_put_hevc_qpel_bi_h16_8_neon
void ff_hevc_put_hevc_qpel_bi_h16_8_neon(uint8_t *_dst, ptrdiff_t _dststride, const uint8_t *_src, ptrdiff_t _srcstride, const int16_t *src2, int height, intptr_t mx, intptr_t my, int width)
ff_hevc_add_residual_4x4_12_neon
void ff_hevc_add_residual_4x4_12_neon(uint8_t *_dst, const int16_t *coeffs, ptrdiff_t stride)