1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
|
/*
* Copyright (C) 2022 Loongson Technology Corporation Limited
* Contributed by Hao Chen(chenhao@loongson.cn)
*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "swscale_loongarch.h"
#include "libavutil/loongarch/loongson_intrinsics.h"
#define YUV2RGB_LOAD_COE \
/* Load x_offset */ \
__m256i y_offset = __lasx_xvreplgr2vr_d(c->yOffset); \
__m256i u_offset = __lasx_xvreplgr2vr_d(c->uOffset); \
__m256i v_offset = __lasx_xvreplgr2vr_d(c->vOffset); \
/* Load x_coeff */ \
__m256i ug_coeff = __lasx_xvreplgr2vr_d(c->ugCoeff); \
__m256i vg_coeff = __lasx_xvreplgr2vr_d(c->vgCoeff); \
__m256i y_coeff = __lasx_xvreplgr2vr_d(c->yCoeff); \
__m256i ub_coeff = __lasx_xvreplgr2vr_d(c->ubCoeff); \
__m256i vr_coeff = __lasx_xvreplgr2vr_d(c->vrCoeff); \
#define LOAD_YUV_16 \
m_y1 = __lasx_xvld(py_1, 0); \
m_y2 = __lasx_xvld(py_2, 0); \
m_u = __lasx_xvldrepl_d(pu, 0); \
m_v = __lasx_xvldrepl_d(pv, 0); \
m_u = __lasx_xvilvl_b(m_u, m_u); \
m_v = __lasx_xvilvl_b(m_v, m_v); \
DUP4_ARG1(__lasx_vext2xv_hu_bu, m_y1, m_y2, m_u, m_v, \
m_y1, m_y2, m_u, m_v); \
/* YUV2RGB method
* The conversion method is as follows:
* R = Y' * y_coeff + V' * vr_coeff
* G = Y' * y_coeff + V' * vg_coeff + U' * ug_coeff
* B = Y' * y_coeff + U' * ub_coeff
*
* where X' = X * 8 - x_offset
*
*/
#define YUV2RGB \
m_y1 = __lasx_xvslli_h(m_y1, 3); \
m_y2 = __lasx_xvslli_h(m_y2, 3); \
m_u = __lasx_xvslli_h(m_u, 3); \
m_v = __lasx_xvslli_h(m_v, 3); \
m_y1 = __lasx_xvsub_h(m_y1, y_offset); \
m_y2 = __lasx_xvsub_h(m_y2, y_offset); \
m_u = __lasx_xvsub_h(m_u, u_offset); \
m_v = __lasx_xvsub_h(m_v, v_offset); \
y_1 = __lasx_xvmuh_h(m_y1, y_coeff); \
y_2 = __lasx_xvmuh_h(m_y2, y_coeff); \
u2g = __lasx_xvmuh_h(m_u, ug_coeff); \
u2b = __lasx_xvmuh_h(m_u, ub_coeff); \
v2r = __lasx_xvmuh_h(m_v, vr_coeff); \
v2g = __lasx_xvmuh_h(m_v, vg_coeff); \
r1 = __lasx_xvsadd_h(y_1, v2r); \
v2g = __lasx_xvsadd_h(v2g, u2g); \
g1 = __lasx_xvsadd_h(y_1, v2g); \
b1 = __lasx_xvsadd_h(y_1, u2b); \
r2 = __lasx_xvsadd_h(y_2, v2r); \
g2 = __lasx_xvsadd_h(y_2, v2g); \
b2 = __lasx_xvsadd_h(y_2, u2b); \
DUP4_ARG1(__lasx_xvclip255_h, r1, g1, b1, r2, r1, g1, b1, r2); \
DUP2_ARG1(__lasx_xvclip255_h, g2, b2, g2, b2); \
#define YUV2RGB_RES \
m_y1 = __lasx_xvldrepl_d(py_1, 0); \
m_y2 = __lasx_xvldrepl_d(py_2, 0); \
m_u = __lasx_xvldrepl_w(pu, 0); \
m_v = __lasx_xvldrepl_w(pv, 0); \
m_y1 = __lasx_xvilvl_d(m_y2, m_y1); \
m_u = __lasx_xvilvl_b(m_u, m_u); \
m_v = __lasx_xvilvl_b(m_v, m_v); \
m_y1 = __lasx_vext2xv_hu_bu(m_y1); \
m_u = __lasx_vext2xv_hu_bu(m_u); \
m_v = __lasx_vext2xv_hu_bu(m_v); \
m_y1 = __lasx_xvslli_h(m_y1, 3); \
m_u = __lasx_xvslli_h(m_u, 3); \
m_v = __lasx_xvslli_h(m_v, 3); \
m_y1 = __lasx_xvsub_h(m_y1, y_offset); \
m_u = __lasx_xvsub_h(m_u, u_offset); \
m_v = __lasx_xvsub_h(m_v, v_offset); \
y_1 = __lasx_xvmuh_h(m_y1, y_coeff); \
u2g = __lasx_xvmuh_h(m_u, ug_coeff); \
u2b = __lasx_xvmuh_h(m_u, ub_coeff); \
v2r = __lasx_xvmuh_h(m_v, vr_coeff); \
v2g = __lasx_xvmuh_h(m_v, vg_coeff); \
r1 = __lasx_xvsadd_h(y_1, v2r); \
v2g = __lasx_xvsadd_h(v2g, u2g); \
g1 = __lasx_xvsadd_h(y_1, v2g); \
b1 = __lasx_xvsadd_h(y_1, u2b); \
r1 = __lasx_xvclip255_h(r1); \
g1 = __lasx_xvclip255_h(g1); \
b1 = __lasx_xvclip255_h(b1); \
#define RGB_PACK(r, g, b, rgb_l, rgb_h) \
{ \
__m256i rg; \
rg = __lasx_xvpackev_b(g, r); \
DUP2_ARG3(__lasx_xvshuf_b, b, rg, shuf2, b, rg, shuf3, rgb_l, rgb_h); \
}
#define RGB32_PACK(a, r, g, b, rgb_l, rgb_h) \
{ \
__m256i ra, bg, tmp0, tmp1; \
ra = __lasx_xvpackev_b(r, a); \
bg = __lasx_xvpackev_b(b, g); \
tmp0 = __lasx_xvilvl_h(bg, ra); \
tmp1 = __lasx_xvilvh_h(bg, ra); \
rgb_l = __lasx_xvpermi_q(tmp1, tmp0, 0x20); \
rgb_h = __lasx_xvpermi_q(tmp1, tmp0, 0x31); \
}
#define RGB_STORE_RES(rgb_l, rgb_h, image_1, image_2) \
{ \
__lasx_xvstelm_d(rgb_l, image_1, 0, 0); \
__lasx_xvstelm_d(rgb_l, image_1, 8, 1); \
__lasx_xvstelm_d(rgb_h, image_1, 16, 0); \
__lasx_xvstelm_d(rgb_l, image_2, 0, 2); \
__lasx_xvstelm_d(rgb_l, image_2, 8, 3); \
__lasx_xvstelm_d(rgb_h, image_2, 16, 2); \
}
#define RGB_STORE(rgb_l, rgb_h, image) \
{ \
__lasx_xvstelm_d(rgb_l, image, 0, 0); \
__lasx_xvstelm_d(rgb_l, image, 8, 1); \
__lasx_xvstelm_d(rgb_h, image, 16, 0); \
__lasx_xvstelm_d(rgb_l, image, 24, 2); \
__lasx_xvstelm_d(rgb_l, image, 32, 3); \
__lasx_xvstelm_d(rgb_h, image, 40, 2); \
}
#define RGB32_STORE(rgb_l, rgb_h, image) \
{ \
__lasx_xvst(rgb_l, image, 0); \
__lasx_xvst(rgb_h, image, 32); \
}
#define RGB32_STORE_RES(rgb_l, rgb_h, image_1, image_2) \
{ \
__lasx_xvst(rgb_l, image_1, 0); \
__lasx_xvst(rgb_h, image_2, 0); \
}
#define YUV2RGBFUNC(func_name, dst_type, alpha) \
int func_name(SwsInternal *c, const uint8_t *const src[], \
const int srcStride[], int srcSliceY, int srcSliceH, \
uint8_t *const dst[], const int dstStride[]) \
{ \
int x, y, h_size, vshift, res; \
__m256i m_y1, m_y2, m_u, m_v; \
__m256i y_1, y_2, u2g, v2g, u2b, v2r, rgb1_l, rgb1_h; \
__m256i rgb2_l, rgb2_h, r1, g1, b1, r2, g2, b2; \
__m256i shuf2 = {0x0504120302100100, 0x0A18090816070614, \
0x0504120302100100, 0x0A18090816070614}; \
__m256i shuf3 = {0x1E0F0E1C0D0C1A0B, 0x0101010101010101, \
0x1E0F0E1C0D0C1A0B, 0x0101010101010101}; \
YUV2RGB_LOAD_COE \
y = (c->opts.dst_w + 7) & ~7; \
h_size = y >> 4; \
res = y & 15; \
\
vshift = c->opts.src_format != AV_PIX_FMT_YUV422P; \
for (y = 0; y < srcSliceH; y += 2) { \
dst_type *image1 = (dst_type *)(dst[0] + (y + srcSliceY) * dstStride[0]);\
dst_type *image2 = (dst_type *)(image1 + dstStride[0]);\
const uint8_t *py_1 = src[0] + y * srcStride[0]; \
const uint8_t *py_2 = py_1 + srcStride[0]; \
const uint8_t *pu = src[1] + (y >> vshift) * srcStride[1]; \
const uint8_t *pv = src[2] + (y >> vshift) * srcStride[2]; \
for(x = 0; x < h_size; x++) { \
#define YUV2RGBFUNC32(func_name, dst_type, alpha) \
int func_name(SwsInternal *c, const uint8_t *const src[], \
const int srcStride[], int srcSliceY, int srcSliceH, \
uint8_t *const dst[], const int dstStride[]) \
{ \
int x, y, h_size, vshift, res; \
__m256i m_y1, m_y2, m_u, m_v; \
__m256i y_1, y_2, u2g, v2g, u2b, v2r, rgb1_l, rgb1_h; \
__m256i rgb2_l, rgb2_h, r1, g1, b1, r2, g2, b2; \
__m256i a = __lasx_xvldi(0xFF); \
\
YUV2RGB_LOAD_COE \
y = (c->opts.dst_w + 7) & ~7; \
h_size = y >> 4; \
res = y & 15; \
\
vshift = c->opts.src_format != AV_PIX_FMT_YUV422P; \
for (y = 0; y < srcSliceH; y += 2) { \
int yd = y + srcSliceY; \
dst_type av_unused *r, *g, *b; \
dst_type *image1 = (dst_type *)(dst[0] + (yd) * dstStride[0]); \
dst_type *image2 = (dst_type *)(dst[0] + (yd + 1) * dstStride[0]); \
const uint8_t *py_1 = src[0] + y * srcStride[0]; \
const uint8_t *py_2 = py_1 + srcStride[0]; \
const uint8_t *pu = src[1] + (y >> vshift) * srcStride[1]; \
const uint8_t *pv = src[2] + (y >> vshift) * srcStride[2]; \
for(x = 0; x < h_size; x++) { \
#define DEALYUV2RGBREMAIN \
py_1 += 16; \
py_2 += 16; \
pu += 8; \
pv += 8; \
image1 += 48; \
image2 += 48; \
} \
if (res) { \
#define DEALYUV2RGBREMAIN32 \
py_1 += 16; \
py_2 += 16; \
pu += 8; \
pv += 8; \
image1 += 16; \
image2 += 16; \
} \
if (res) { \
#define END_FUNC() \
} \
} \
return srcSliceH; \
}
YUV2RGBFUNC(yuv420_rgb24_lasx, uint8_t, 0)
LOAD_YUV_16
YUV2RGB
RGB_PACK(r1, g1, b1, rgb1_l, rgb1_h);
RGB_PACK(r2, g2, b2, rgb2_l, rgb2_h);
RGB_STORE(rgb1_l, rgb1_h, image1);
RGB_STORE(rgb2_l, rgb2_h, image2);
DEALYUV2RGBREMAIN
YUV2RGB_RES
RGB_PACK(r1, g1, b1, rgb1_l, rgb1_h);
RGB_STORE_RES(rgb1_l, rgb1_h, image1, image2);
END_FUNC()
YUV2RGBFUNC(yuv420_bgr24_lasx, uint8_t, 0)
LOAD_YUV_16
YUV2RGB
RGB_PACK(b1, g1, r1, rgb1_l, rgb1_h);
RGB_PACK(b2, g2, r2, rgb2_l, rgb2_h);
RGB_STORE(rgb1_l, rgb1_h, image1);
RGB_STORE(rgb2_l, rgb2_h, image2);
DEALYUV2RGBREMAIN
YUV2RGB_RES
RGB_PACK(b1, g1, r1, rgb1_l, rgb1_h);
RGB_STORE_RES(rgb1_l, rgb1_h, image1, image2);
END_FUNC()
YUV2RGBFUNC32(yuv420_rgba32_lasx, uint32_t, 0)
LOAD_YUV_16
YUV2RGB
RGB32_PACK(r1, g1, b1, a, rgb1_l, rgb1_h);
RGB32_PACK(r2, g2, b2, a, rgb2_l, rgb2_h);
RGB32_STORE(rgb1_l, rgb1_h, image1);
RGB32_STORE(rgb2_l, rgb2_h, image2);
DEALYUV2RGBREMAIN32
YUV2RGB_RES
RGB32_PACK(r1, g1, b1, a, rgb1_l, rgb1_h);
RGB32_STORE_RES(rgb1_l, rgb1_h, image1, image2);
END_FUNC()
YUV2RGBFUNC32(yuv420_bgra32_lasx, uint32_t, 0)
LOAD_YUV_16
YUV2RGB
RGB32_PACK(b1, g1, r1, a, rgb1_l, rgb1_h);
RGB32_PACK(b2, g2, r2, a, rgb2_l, rgb2_h);
RGB32_STORE(rgb1_l, rgb1_h, image1);
RGB32_STORE(rgb2_l, rgb2_h, image2);
DEALYUV2RGBREMAIN32
YUV2RGB_RES
RGB32_PACK(b1, g1, r1, a, rgb1_l, rgb1_h);
RGB32_STORE_RES(rgb1_l, rgb1_h, image1, image2);
END_FUNC()
YUV2RGBFUNC32(yuv420_argb32_lasx, uint32_t, 0)
LOAD_YUV_16
YUV2RGB
RGB32_PACK(a, r1, g1, b1, rgb1_l, rgb1_h);
RGB32_PACK(a, r2, g2, b2, rgb2_l, rgb2_h);
RGB32_STORE(rgb1_l, rgb1_h, image1);
RGB32_STORE(rgb2_l, rgb2_h, image2);
DEALYUV2RGBREMAIN32
YUV2RGB_RES
RGB32_PACK(a, r1, g1, b1, rgb1_l, rgb1_h);
RGB32_STORE_RES(rgb1_l, rgb1_h, image1, image2);
END_FUNC()
YUV2RGBFUNC32(yuv420_abgr32_lasx, uint32_t, 0)
LOAD_YUV_16
YUV2RGB
RGB32_PACK(a, b1, g1, r1, rgb1_l, rgb1_h);
RGB32_PACK(a, b2, g2, r2, rgb2_l, rgb2_h);
RGB32_STORE(rgb1_l, rgb1_h, image1);
RGB32_STORE(rgb2_l, rgb2_h, image2);
DEALYUV2RGBREMAIN32
YUV2RGB_RES
RGB32_PACK(a, b1, g1, r1, rgb1_l, rgb1_h);
RGB32_STORE_RES(rgb1_l, rgb1_h, image1, image2);
END_FUNC()
|
