FFmpeg
vp9dsp_init_mips.c
Go to the documentation of this file.
1 /*
2  * Copyright (c) 2015 Shivraj Patil (Shivraj.Patil@imgtec.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 "libavutil/attributes.h"
22 #include "libavutil/mips/cpu.h"
23 #include "config.h"
24 #include "libavutil/common.h"
25 #include "libavcodec/vp9dsp.h"
26 #include "vp9dsp_mips.h"
27 
28 #if HAVE_MSA
29 static av_cold void vp9dsp_intrapred_init_msa(VP9DSPContext *dsp, int bpp)
30 {
31  if (bpp == 8) {
32 #define init_intra_pred_msa(tx, sz) \
33  dsp->intra_pred[tx][VERT_PRED] = ff_vert_##sz##_msa; \
34  dsp->intra_pred[tx][HOR_PRED] = ff_hor_##sz##_msa; \
35  dsp->intra_pred[tx][DC_PRED] = ff_dc_##sz##_msa; \
36  dsp->intra_pred[tx][LEFT_DC_PRED] = ff_dc_left_##sz##_msa; \
37  dsp->intra_pred[tx][TOP_DC_PRED] = ff_dc_top_##sz##_msa; \
38  dsp->intra_pred[tx][DC_128_PRED] = ff_dc_128_##sz##_msa; \
39  dsp->intra_pred[tx][DC_127_PRED] = ff_dc_127_##sz##_msa; \
40  dsp->intra_pred[tx][DC_129_PRED] = ff_dc_129_##sz##_msa; \
41  dsp->intra_pred[tx][TM_VP8_PRED] = ff_tm_##sz##_msa; \
42 
43  init_intra_pred_msa(TX_16X16, 16x16);
44  init_intra_pred_msa(TX_32X32, 32x32);
45 #undef init_intra_pred_msa
46 
47 #define init_intra_pred_msa(tx, sz) \
48  dsp->intra_pred[tx][DC_PRED] = ff_dc_##sz##_msa; \
49  dsp->intra_pred[tx][LEFT_DC_PRED] = ff_dc_left_##sz##_msa; \
50  dsp->intra_pred[tx][TOP_DC_PRED] = ff_dc_top_##sz##_msa; \
51  dsp->intra_pred[tx][TM_VP8_PRED] = ff_tm_##sz##_msa; \
52 
53  init_intra_pred_msa(TX_4X4, 4x4);
54  init_intra_pred_msa(TX_8X8, 8x8);
55 #undef init_intra_pred_msa
56  }
57 }
58 
59 static av_cold void vp9dsp_itxfm_init_msa(VP9DSPContext *dsp, int bpp)
60 {
61  if (bpp == 8) {
62 #define init_itxfm(tx, sz) \
63  dsp->itxfm_add[tx][DCT_DCT] = ff_idct_idct_##sz##_add_msa; \
64  dsp->itxfm_add[tx][DCT_ADST] = ff_iadst_idct_##sz##_add_msa; \
65  dsp->itxfm_add[tx][ADST_DCT] = ff_idct_iadst_##sz##_add_msa; \
66  dsp->itxfm_add[tx][ADST_ADST] = ff_iadst_iadst_##sz##_add_msa \
67 
68 #define init_idct(tx, nm) \
69  dsp->itxfm_add[tx][DCT_DCT] = \
70  dsp->itxfm_add[tx][ADST_DCT] = \
71  dsp->itxfm_add[tx][DCT_ADST] = \
72  dsp->itxfm_add[tx][ADST_ADST] = nm##_add_msa
73 
74  init_itxfm(TX_4X4, 4x4);
75  init_itxfm(TX_8X8, 8x8);
76  init_itxfm(TX_16X16, 16x16);
77  init_idct(TX_32X32, ff_idct_idct_32x32);
78 #undef init_itxfm
79 #undef init_idct
80  }
81 }
82 
83 static av_cold void vp9dsp_mc_init_msa(VP9DSPContext *dsp, int bpp)
84 {
85  if (bpp == 8) {
86 #define init_fpel(idx1, idx2, sz, type) \
87  dsp->mc[idx1][FILTER_8TAP_SMOOTH ][idx2][0][0] = ff_##type##sz##_msa; \
88  dsp->mc[idx1][FILTER_8TAP_REGULAR][idx2][0][0] = ff_##type##sz##_msa; \
89  dsp->mc[idx1][FILTER_8TAP_SHARP ][idx2][0][0] = ff_##type##sz##_msa; \
90  dsp->mc[idx1][FILTER_BILINEAR ][idx2][0][0] = ff_##type##sz##_msa
91 
92 #define init_copy_avg(idx, sz) \
93  init_fpel(idx, 0, sz, copy); \
94  init_fpel(idx, 1, sz, avg)
95 
96 #define init_avg(idx, sz) \
97  init_fpel(idx, 1, sz, avg)
98 
99  init_copy_avg(0, 64);
100  init_copy_avg(1, 32);
101  init_copy_avg(2, 16);
102  init_copy_avg(3, 8);
103  init_avg(4, 4);
104 
105 #undef init_copy_avg
106 #undef init_avg
107 #undef init_fpel
108 
109 #define init_subpel1(idx1, idx2, idxh, idxv, sz, dir, type) \
110  dsp->mc[idx1][FILTER_BILINEAR ][idx2][idxh][idxv] = \
111  ff_##type##_bilin_##sz##dir##_msa; \
112  dsp->mc[idx1][FILTER_8TAP_SMOOTH ][idx2][idxh][idxv] = \
113  ff_##type##_8tap_smooth_##sz##dir##_msa; \
114  dsp->mc[idx1][FILTER_8TAP_REGULAR][idx2][idxh][idxv] = \
115  ff_##type##_8tap_regular_##sz##dir##_msa; \
116  dsp->mc[idx1][FILTER_8TAP_SHARP ][idx2][idxh][idxv] = \
117  ff_##type##_8tap_sharp_##sz##dir##_msa;
118 
119 #define init_subpel2(idx, idxh, idxv, dir, type) \
120  init_subpel1(0, idx, idxh, idxv, 64, dir, type); \
121  init_subpel1(1, idx, idxh, idxv, 32, dir, type); \
122  init_subpel1(2, idx, idxh, idxv, 16, dir, type); \
123  init_subpel1(3, idx, idxh, idxv, 8, dir, type); \
124  init_subpel1(4, idx, idxh, idxv, 4, dir, type)
125 
126 #define init_subpel3(idx, type) \
127  init_subpel2(idx, 1, 1, hv, type); \
128  init_subpel2(idx, 0, 1, v, type); \
129  init_subpel2(idx, 1, 0, h, type)
130 
131  init_subpel3(0, put);
132  init_subpel3(1, avg);
133 
134 #undef init_subpel1
135 #undef init_subpel2
136 #undef init_subpel3
137  }
138 }
139 
140 static av_cold void vp9dsp_loopfilter_init_msa(VP9DSPContext *dsp, int bpp)
141 {
142  if (bpp == 8) {
149 
152 
161  }
162 }
163 
164 static av_cold void vp9dsp_init_msa(VP9DSPContext *dsp, int bpp)
165 {
166  vp9dsp_intrapred_init_msa(dsp, bpp);
167  vp9dsp_itxfm_init_msa(dsp, bpp);
168  vp9dsp_mc_init_msa(dsp, bpp);
169  vp9dsp_loopfilter_init_msa(dsp, bpp);
170 }
171 #endif // #if HAVE_MSA
172 
173 #if HAVE_MMI
174 static av_cold void vp9dsp_mc_init_mmi(VP9DSPContext *dsp)
175 {
176 #define init_subpel1(idx1, idx2, idxh, idxv, sz, dir, type) \
177  dsp->mc[idx1][FILTER_8TAP_SMOOTH ][idx2][idxh][idxv] = \
178  ff_##type##_8tap_smooth_##sz##dir##_mmi; \
179  dsp->mc[idx1][FILTER_8TAP_REGULAR][idx2][idxh][idxv] = \
180  ff_##type##_8tap_regular_##sz##dir##_mmi; \
181  dsp->mc[idx1][FILTER_8TAP_SHARP ][idx2][idxh][idxv] = \
182  ff_##type##_8tap_sharp_##sz##dir##_mmi;
183 
184 #define init_subpel2(idx, idxh, idxv, dir, type) \
185  init_subpel1(0, idx, idxh, idxv, 64, dir, type); \
186  init_subpel1(1, idx, idxh, idxv, 32, dir, type); \
187  init_subpel1(2, idx, idxh, idxv, 16, dir, type); \
188  init_subpel1(3, idx, idxh, idxv, 8, dir, type); \
189  init_subpel1(4, idx, idxh, idxv, 4, dir, type)
190 
191 #define init_subpel3(idx, type) \
192  init_subpel2(idx, 1, 1, hv, type); \
193  init_subpel2(idx, 0, 1, v, type); \
194  init_subpel2(idx, 1, 0, h, type)
195 
196  init_subpel3(0, put);
197  init_subpel3(1, avg);
198 
199 #undef init_subpel1
200 #undef init_subpel2
201 #undef init_subpel3
202 }
203 
204 static av_cold void vp9dsp_init_mmi(VP9DSPContext *dsp, int bpp)
205 {
206  if (bpp == 8) {
207  vp9dsp_mc_init_mmi(dsp);
208  }
209 }
210 #endif // #if HAVE_MMI
211 
213 {
214 #if HAVE_MSA || HAVE_MMI
215  int cpu_flags = av_get_cpu_flags();
216 #endif
217 
218 #if HAVE_MMI
219  if (have_mmi(cpu_flags))
220  vp9dsp_init_mmi(dsp, bpp);
221 #endif
222 
223 #if HAVE_MSA
224  if (have_msa(cpu_flags))
225  vp9dsp_init_msa(dsp, bpp);
226 #endif
227 }
ff_loop_filter_h_88_16_msa
void ff_loop_filter_h_88_16_msa(uint8_t *src, ptrdiff_t pitch, int32_t b_limit_ptr, int32_t limit_ptr, int32_t thresh_ptr)
Definition: vp9_lpf_msa.c:1349
ff_loop_filter_v_44_16_msa
void ff_loop_filter_v_44_16_msa(uint8_t *src, ptrdiff_t pitch, int32_t b_limit_ptr, int32_t limit_ptr, int32_t thresh_ptr)
Definition: vp9_lpf_msa.c:225
have_mmi
#define have_mmi(flags)
Definition: cpu.h:25
ff_loop_filter_h_8_8_msa
void ff_loop_filter_h_8_8_msa(uint8_t *src, ptrdiff_t pitch, int32_t b_limit_ptr, int32_t limit_ptr, int32_t thresh_ptr)
Definition: vp9_lpf_msa.c:1271
ff_loop_filter_v_4_8_msa
void ff_loop_filter_v_4_8_msa(uint8_t *src, ptrdiff_t pitch, int32_t b_limit_ptr, int32_t limit_ptr, int32_t thresh_ptr)
Definition: vp9_lpf_msa.c:196
ff_loop_filter_h_48_16_msa
void ff_loop_filter_h_48_16_msa(uint8_t *src, ptrdiff_t pitch, int32_t b_limit_ptr, int32_t limit_ptr, int32_t thresh_ptr)
Definition: vp9_lpf_msa.c:1565
VP9DSPContext::loop_filter_8
void(* loop_filter_8[3][2])(uint8_t *dst, ptrdiff_t stride, int mb_lim, int lim, int hev_thr)
Definition: vp9dsp.h:80
VP9DSPContext
Definition: vp9dsp.h:39
av_get_cpu_flags
int av_get_cpu_flags(void)
Return the flags which specify extensions supported by the CPU.
Definition: cpu.c:98
cpu_flags
static atomic_int cpu_flags
Definition: cpu.c:50
VP9DSPContext::loop_filter_mix2
void(* loop_filter_mix2[2][2][2])(uint8_t *dst, ptrdiff_t stride, int mb_lim, int lim, int hev_thr)
Definition: vp9dsp.h:102
av_cold
#define av_cold
Definition: attributes.h:90
have_msa
#define have_msa(flags)
Definition: cpu.h:26
init_avg
#define init_avg(idx, sz, suffix)
ff_loop_filter_v_48_16_msa
void ff_loop_filter_v_48_16_msa(uint8_t *src, ptrdiff_t pitch, int32_t b_limit_ptr, int32_t limit_ptr, int32_t thresh_ptr)
Definition: vp9_lpf_msa.c:479
vp9dsp_mips.h
ff_loop_filter_h_84_16_msa
void ff_loop_filter_h_84_16_msa(uint8_t *src, ptrdiff_t pitch, int32_t b_limit_ptr, int32_t limit_ptr, int32_t thresh_ptr)
Definition: vp9_lpf_msa.c:1462
TX_8X8
@ TX_8X8
Definition: vp9.h:29
ff_loop_filter_v_16_16_msa
void ff_loop_filter_v_16_16_msa(uint8_t *src, ptrdiff_t pitch, int32_t b_limit_ptr, int32_t limit_ptr, int32_t thresh_ptr)
Definition: vp9_lpf_msa.c:958
ff_loop_filter_h_4_8_msa
void ff_loop_filter_h_4_8_msa(uint8_t *src, ptrdiff_t pitch, int32_t b_limit_ptr, int32_t limit_ptr, int32_t thresh_ptr)
Definition: vp9_lpf_msa.c:1198
TX_16X16
@ TX_16X16
Definition: vp9.h:30
init_idct
#define init_idct(tx, nm, bpp)
vp9dsp.h
init_subpel3
#define init_subpel3(idx, type)
Definition: vp9dsp_init_loongarch.c:42
avg
#define avg(a, b, c, d)
Definition: colorspacedsp_template.c:28
TX_4X4
@ TX_4X4
Definition: vp9.h:28
ff_loop_filter_h_16_8_msa
void ff_loop_filter_h_16_8_msa(uint8_t *src, ptrdiff_t pitch, int32_t b_limit_ptr, int32_t limit_ptr, int32_t thresh_ptr)
Definition: vp9_lpf_msa.c:2066
attributes.h
ff_loop_filter_v_84_16_msa
void ff_loop_filter_v_84_16_msa(uint8_t *src, ptrdiff_t pitch, int32_t b_limit_ptr, int32_t limit_ptr, int32_t thresh_ptr)
Definition: vp9_lpf_msa.c:408
common.h
ff_loop_filter_v_16_8_msa
void ff_loop_filter_v_16_8_msa(uint8_t *src, ptrdiff_t pitch, int32_t b_limit_ptr, int32_t limit_ptr, int32_t thresh_ptr)
Definition: vp9_lpf_msa.c:974
ff_loop_filter_h_44_16_msa
void ff_loop_filter_h_44_16_msa(uint8_t *src, ptrdiff_t pitch, int32_t b_limit_ptr, int32_t limit_ptr, int32_t thresh_ptr)
Definition: vp9_lpf_msa.c:1225
init_itxfm
#define init_itxfm(tx, sz, bpp)
ff_vp9dsp_init_mips
av_cold void ff_vp9dsp_init_mips(VP9DSPContext *dsp, int bpp)
Definition: vp9dsp_init_mips.c:212
init_copy_avg
#define init_copy_avg(idx, sz1, sz2)
ff_loop_filter_v_8_8_msa
void ff_loop_filter_v_8_8_msa(uint8_t *src, ptrdiff_t pitch, int32_t b_limit_ptr, int32_t limit_ptr, int32_t thresh_ptr)
Definition: vp9_lpf_msa.c:255
ff_loop_filter_h_16_16_msa
void ff_loop_filter_h_16_16_msa(uint8_t *src, ptrdiff_t pitch, int32_t b_limit_ptr, int32_t limit_ptr, int32_t thresh_ptr)
Definition: vp9_lpf_msa.c:2502
TX_32X32
@ TX_32X32
Definition: vp9.h:31
ff_loop_filter_v_88_16_msa
void ff_loop_filter_v_88_16_msa(uint8_t *src, ptrdiff_t pitch, int32_t b_limit_ptr, int32_t limit_ptr, int32_t thresh_ptr)
Definition: vp9_lpf_msa.c:329
cpu.h
VP9DSPContext::loop_filter_16
void(* loop_filter_16[2])(uint8_t *dst, ptrdiff_t stride, int mb_lim, int lim, int hev_thr)
Definition: vp9dsp.h:88