FFmpeg coverage

Directory: ../../../ffmpeg/
File: src/libavcodec/x86/cavsdsp.c
Date: 2026-07-19 01:11:25
Exec Total Coverage
Lines: 28 33 84.8%
Functions: 11 12 91.7%
Branches: 2 2 100.0%
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 ✗ static void cavs_idct8_add_sse2(uint8_t *dst, int16_t *block, ptrdiff_t stride)
41 {
42 ✗ LOCAL_ALIGNED(16, int16_t, b2, [64]);
43 ✗ ff_cavs_idct8_sse2(b2, block);
44 ✗ ff_add_pixels_clamped_sse2(b2, dst, stride);
45 ✗ }
46
47 #define DEF_QPEL(OPNAME) \
48 void ff_ ## OPNAME ## _cavs_qpel8_mc20_sse2(uint8_t *dst, const uint8_t *src, ptrdiff_t stride); \
49 void ff_ ## OPNAME ## _cavs_qpel8_mc02_sse2(uint8_t *dst, const uint8_t *src, ptrdiff_t stride); \
50 void ff_ ## OPNAME ## _cavs_qpel8_mc03_sse2(uint8_t *dst, const uint8_t *src, ptrdiff_t stride); \
51 void ff_ ## OPNAME ## _cavs_qpel8_h_sse2(uint8_t *dst, const uint8_t *src, ptrdiff_t stride, int h); \
52 void ff_ ## OPNAME ## _cavs_qpel8_v2_sse2(uint8_t *dst, const uint8_t *src, ptrdiff_t stride, int h);\
53 void ff_ ## OPNAME ## _cavs_qpel8_v3_sse2(uint8_t *dst, const uint8_t *src, ptrdiff_t stride, int h);\
54
55 DEF_QPEL(put)
56 DEF_QPEL(avg)
57
58 #define QPEL_CAVS_XMM(OPNAME, XMM) \
59 static void OPNAME ## _cavs_qpel16_mc02_ ## XMM(uint8_t *dst, const uint8_t *src, ptrdiff_t stride) \
60 { \
61 ff_ ## OPNAME ## _cavs_qpel8_v2_ ## XMM(dst, src, stride, 16); \
62 ff_ ## OPNAME ## _cavs_qpel8_v2_ ## XMM(dst + 8, src + 8, stride, 16); \
63 } \
64 static void OPNAME ## _cavs_qpel16_mc03_ ## XMM(uint8_t *dst, const uint8_t *src, ptrdiff_t stride) \
65 { \
66 ff_ ## OPNAME ## _cavs_qpel8_v3_ ## XMM(dst, src, stride, 16); \
67 ff_ ## OPNAME ## _cavs_qpel8_v3_ ## XMM(dst + 8, src + 8, stride, 16); \
68 } \
69 static void OPNAME ## _cavs_qpel8_mc01_ ## XMM(uint8_t *dst, const uint8_t *src, ptrdiff_t stride) \
70 { \
71 ff_ ## OPNAME ## _cavs_qpel8_mc03_ ## XMM(dst + 7 * stride, src + 8 * stride, -stride); \
72 } \
73 static void OPNAME ## _cavs_qpel16_mc01_ ## XMM(uint8_t *dst, const uint8_t *src, ptrdiff_t stride) \
74 { \
75 OPNAME ## _cavs_qpel16_mc03_ ## XMM(dst + 15 * stride, src + 16 * stride, -stride); \
76 } \
77 static void OPNAME ## _cavs_qpel16_mc20_ ## XMM(uint8_t *dst, const uint8_t *src, ptrdiff_t stride) \
78 { \
79 ff_ ## OPNAME ## _cavs_qpel8_h_ ## XMM(dst, src, stride, 16); \
80 ff_ ## OPNAME ## _cavs_qpel8_h_ ## XMM(dst + 8, src + 8, stride, 16); \
81 }
82
83 24 QPEL_CAVS_XMM(put, sse2)
84 24 QPEL_CAVS_XMM(avg, sse2)
85 #endif
86
87 20 av_cold void ff_cavsdsp_init_x86(CAVSDSPContext *c)
88 {
89 #if HAVE_SSE2_EXTERNAL
90 20 int cpu_flags = av_get_cpu_flags();
91
92
2/2
✓ Branch 0 taken 12 times.
✓ Branch 1 taken 8 times.
 20 if (EXTERNAL_SSE2(cpu_flags)) {
93 12 c->put_cavs_qpel_pixels_tab[0][ 0] = ff_put_pixels16x16_sse2;
94 12 c->put_cavs_qpel_pixels_tab[0][ 2] = put_cavs_qpel16_mc20_sse2;
95 12 c->put_cavs_qpel_pixels_tab[0][ 4] = put_cavs_qpel16_mc01_sse2;
96 12 c->put_cavs_qpel_pixels_tab[0][ 8] = put_cavs_qpel16_mc02_sse2;
97 12 c->put_cavs_qpel_pixels_tab[0][12] = put_cavs_qpel16_mc03_sse2;
98 12 c->put_cavs_qpel_pixels_tab[1][ 0] = ff_put_pixels8x8_sse2;
99 12 c->put_cavs_qpel_pixels_tab[1][ 2] = ff_put_cavs_qpel8_mc20_sse2;
100 12 c->put_cavs_qpel_pixels_tab[1][ 4] = put_cavs_qpel8_mc01_sse2;
101 12 c->put_cavs_qpel_pixels_tab[1][ 8] = ff_put_cavs_qpel8_mc02_sse2;
102 12 c->put_cavs_qpel_pixels_tab[1][12] = ff_put_cavs_qpel8_mc03_sse2;
103
104 12 c->avg_cavs_qpel_pixels_tab[0][ 0] = ff_avg_pixels16x16_sse2;
105 12 c->avg_cavs_qpel_pixels_tab[0][ 2] = avg_cavs_qpel16_mc20_sse2;
106 12 c->avg_cavs_qpel_pixels_tab[0][ 4] = avg_cavs_qpel16_mc01_sse2;
107 12 c->avg_cavs_qpel_pixels_tab[0][ 8] = avg_cavs_qpel16_mc02_sse2;
108 12 c->avg_cavs_qpel_pixels_tab[0][12] = avg_cavs_qpel16_mc03_sse2;
109 12 c->avg_cavs_qpel_pixels_tab[1][ 0] = ff_avg_pixels8x8_sse2;
110 12 c->avg_cavs_qpel_pixels_tab[1][ 2] = ff_avg_cavs_qpel8_mc20_sse2;
111 12 c->avg_cavs_qpel_pixels_tab[1][ 4] = avg_cavs_qpel8_mc01_sse2;
112 12 c->avg_cavs_qpel_pixels_tab[1][ 8] = ff_avg_cavs_qpel8_mc02_sse2;
113 12 c->avg_cavs_qpel_pixels_tab[1][12] = ff_avg_cavs_qpel8_mc03_sse2;
114
115 12 c->cavs_idct8_add = cavs_idct8_add_sse2;
116 12 c->idct_perm = FF_IDCT_PERM_TRANSPOSE;
117 }
118 #endif
119 20 }
120