FFmpeg coverage

Directory: ../../../ffmpeg/
File: src/libavcodec/x86/cavsdsp.c
Date: 2025-10-21 20:24:59
Exec Total Coverage
Lines: 34 34 100.0%
Functions: 12 12 100.0%
Branches: 4 4 100.0%
Line Branch Exec Source
1 /*
2 * Chinese AVS video (AVS1-P2, JiZhun profile) decoder.
3 * Copyright (c) 2006 Stefan Gehrer <stefan.gehrer@gmx.de>
4 *
5 * MMX-optimized DSP functions, based on H.264 optimizations by
6 * Michael Niedermayer and Loren Merritt
7 *
8 * This file is part of FFmpeg.
9 *
10 * FFmpeg is free software; you can redistribute it and/or
11 * modify it under the terms of the GNU Lesser General Public
12 * License as published by the Free Software Foundation; either
13 * version 2.1 of the License, or (at your option) any later version.
14 *
15 * FFmpeg is distributed in the hope that it will be useful,
16 * but WITHOUT ANY WARRANTY; without even the implied warranty of
17 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
18 * Lesser General Public License for more details.
19 *
20 * You should have received a copy of the GNU Lesser General Public
21 * License along with FFmpeg; if not, write to the Free Software
22 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
23 */
24
25 #include "libavutil/attributes.h"
26 #include "libavutil/cpu.h"
27 #include "libavutil/mem_internal.h"
28 #include "libavutil/x86/cpu.h"
29 #include "libavcodec/cavsdsp.h"
30 #include "libavcodec/idctdsp.h"
31 #include "fpel.h"
32 #include "idctdsp.h"
33 #include "config.h"
34
35
36 #if HAVE_SSE2_EXTERNAL
37
38 void ff_cavs_idct8_sse2(int16_t *out, const int16_t *in);
39
40 4921 static void cavs_idct8_add_sse2(uint8_t *dst, int16_t *block, ptrdiff_t stride)
41 {
42 4921 LOCAL_ALIGNED(16, int16_t, b2, [64]);
43 4921 ff_cavs_idct8_sse2(b2, block);
44 4921 ff_add_pixels_clamped_sse2(b2, dst, stride);
45 4921 }
46
47 #endif /* HAVE_SSE2_EXTERNAL */
48
49 #if HAVE_SSE2_EXTERNAL
50 #define DEF_QPEL(OPNAME) \
51 void ff_ ## OPNAME ## _cavs_qpel8_mc20_sse2(uint8_t *dst, const uint8_t *src, ptrdiff_t stride); \
52 void ff_ ## OPNAME ## _cavs_qpel8_mc02_sse2(uint8_t *dst, const uint8_t *src, ptrdiff_t stride); \
53 void ff_ ## OPNAME ## _cavs_qpel8_mc03_sse2(uint8_t *dst, const uint8_t *src, ptrdiff_t stride); \
54 void ff_ ## OPNAME ## _cavs_qpel8_h_sse2(uint8_t *dst, const uint8_t *src, ptrdiff_t stride, int h); \
55 void ff_ ## OPNAME ## _cavs_qpel8_v2_sse2(uint8_t *dst, const uint8_t *src, ptrdiff_t stride, int h);\
56 void ff_ ## OPNAME ## _cavs_qpel8_v3_sse2(uint8_t *dst, const uint8_t *src, ptrdiff_t stride, int h);\
57
58 DEF_QPEL(put)
59 DEF_QPEL(avg)
60
61 #define QPEL_CAVS_XMM(OPNAME, XMM) \
62 static void OPNAME ## _cavs_qpel16_mc02_ ## XMM(uint8_t *dst, const uint8_t *src, ptrdiff_t stride) \
63 { \
64 ff_ ## OPNAME ## _cavs_qpel8_v2_ ## XMM(dst, src, stride, 16); \
65 ff_ ## OPNAME ## _cavs_qpel8_v2_ ## XMM(dst + 8, src + 8, stride, 16); \
66 } \
67 static void OPNAME ## _cavs_qpel16_mc03_ ## XMM(uint8_t *dst, const uint8_t *src, ptrdiff_t stride) \
68 { \
69 ff_ ## OPNAME ## _cavs_qpel8_v3_ ## XMM(dst, src, stride, 16); \
70 ff_ ## OPNAME ## _cavs_qpel8_v3_ ## XMM(dst + 8, src + 8, stride, 16); \
71 } \
72 static void OPNAME ## _cavs_qpel8_mc01_ ## XMM(uint8_t *dst, const uint8_t *src, ptrdiff_t stride) \
73 { \
74 ff_ ## OPNAME ## _cavs_qpel8_mc03_ ## XMM(dst + 7 * stride, src + 8 * stride, -stride); \
75 } \
76 static void OPNAME ## _cavs_qpel16_mc01_ ## XMM(uint8_t *dst, const uint8_t *src, ptrdiff_t stride) \
77 { \
78 OPNAME ## _cavs_qpel16_mc03_ ## XMM(dst + 15 * stride, src + 16 * stride, -stride); \
79 } \
80 static void OPNAME ## _cavs_qpel16_mc20_ ## XMM(uint8_t *dst, const uint8_t *src, ptrdiff_t stride) \
81 { \
82 ff_ ## OPNAME ## _cavs_qpel8_h_ ## XMM(dst, src, stride, 16); \
83 ff_ ## OPNAME ## _cavs_qpel8_h_ ## XMM(dst + 8, src + 8, stride, 16); \
84 }
85
86 12 QPEL_CAVS_XMM(put, sse2)
87 12 QPEL_CAVS_XMM(avg, sse2)
88 #endif
89
90 19 av_cold void ff_cavsdsp_init_x86(CAVSDSPContext *c)
91 {
92 19 av_unused int cpu_flags = av_get_cpu_flags();
93
94 #if HAVE_MMX_EXTERNAL
95
2/2
✓ Branch 0 taken 13 times.
✓ Branch 1 taken 6 times.
 19 if (EXTERNAL_MMXEXT(cpu_flags)) {
96 13 c->avg_cavs_qpel_pixels_tab[1][0] = ff_avg_pixels8x8_mmxext;
97 }
98 #endif
99 #if HAVE_SSE2_EXTERNAL
100
2/2
✓ Branch 0 taken 11 times.
✓ Branch 1 taken 8 times.
 19 if (EXTERNAL_SSE2(cpu_flags)) {
101 11 c->put_cavs_qpel_pixels_tab[0][ 0] = ff_put_pixels16x16_sse2;
102 11 c->put_cavs_qpel_pixels_tab[0][ 2] = put_cavs_qpel16_mc20_sse2;
103 11 c->put_cavs_qpel_pixels_tab[0][ 4] = put_cavs_qpel16_mc01_sse2;
104 11 c->put_cavs_qpel_pixels_tab[0][ 8] = put_cavs_qpel16_mc02_sse2;
105 11 c->put_cavs_qpel_pixels_tab[0][12] = put_cavs_qpel16_mc03_sse2;
106 11 c->put_cavs_qpel_pixels_tab[1][ 0] = ff_put_pixels8x8_sse2;
107 11 c->put_cavs_qpel_pixels_tab[1][ 2] = ff_put_cavs_qpel8_mc20_sse2;
108 11 c->put_cavs_qpel_pixels_tab[1][ 4] = put_cavs_qpel8_mc01_sse2;
109 11 c->put_cavs_qpel_pixels_tab[1][ 8] = ff_put_cavs_qpel8_mc02_sse2;
110 11 c->put_cavs_qpel_pixels_tab[1][12] = ff_put_cavs_qpel8_mc03_sse2;
111
112 11 c->avg_cavs_qpel_pixels_tab[0][ 0] = ff_avg_pixels16x16_sse2;
113 11 c->avg_cavs_qpel_pixels_tab[0][ 2] = avg_cavs_qpel16_mc20_sse2;
114 11 c->avg_cavs_qpel_pixels_tab[0][ 4] = avg_cavs_qpel16_mc01_sse2;
115 11 c->avg_cavs_qpel_pixels_tab[0][ 8] = avg_cavs_qpel16_mc02_sse2;
116 11 c->avg_cavs_qpel_pixels_tab[0][12] = avg_cavs_qpel16_mc03_sse2;
117 11 c->avg_cavs_qpel_pixels_tab[1][ 2] = ff_avg_cavs_qpel8_mc20_sse2;
118 11 c->avg_cavs_qpel_pixels_tab[1][ 4] = avg_cavs_qpel8_mc01_sse2;
119 11 c->avg_cavs_qpel_pixels_tab[1][ 8] = ff_avg_cavs_qpel8_mc02_sse2;
120 11 c->avg_cavs_qpel_pixels_tab[1][12] = ff_avg_cavs_qpel8_mc03_sse2;
121
122 11 c->cavs_idct8_add = cavs_idct8_add_sse2;
123 11 c->idct_perm = FF_IDCT_PERM_TRANSPOSE;
124 }
125 #endif
126 19 }
127