1/*
2 * SIMD-optimized pixel operations
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 "libavcodec/dsputil.h"
23#include "libavcodec/pixblockdsp.h"
24#include "asm.h"
25
26static void get_pixels_mvi(int16_t *restrict block,
27                           const uint8_t *restrict pixels, int line_size)
28{
29    int h = 8;
30
31    do {
32        uint64_t p;
33
34        p = ldq(pixels);
35        stq(unpkbw(p),       block);
36        stq(unpkbw(p >> 32), block + 4);
37
38        pixels += line_size;
39        block += 8;
40    } while (--h);
41}
42
43static void diff_pixels_mvi(int16_t *block, const uint8_t *s1, const uint8_t *s2,
44                            int stride) {
45    int h = 8;
46    uint64_t mask = 0x4040;
47
48    mask |= mask << 16;
49    mask |= mask << 32;
50    do {
51        uint64_t x, y, c, d, a;
52        uint64_t signs;
53
54        x = ldq(s1);
55        y = ldq(s2);
56        c = cmpbge(x, y);
57        d = x - y;
58        a = zap(mask, c);       /* We use 0x4040404040404040 here...  */
59        d += 4 * a;             /* ...so we can use s4addq here.      */
60        signs = zap(-1, c);
61
62        stq(unpkbw(d)       | (unpkbw(signs)       << 8), block);
63        stq(unpkbw(d >> 32) | (unpkbw(signs >> 32) << 8), block + 4);
64
65        s1 += stride;
66        s2 += stride;
67        block += 8;
68    } while (--h);
69}
70
71av_cold void ff_pixblockdsp_init_alpha(PixblockDSPContext *c, AVCodecContext *avctx,
72                                       unsigned high_bit_depth)
73{
74    if (amask(AMASK_MVI) == 0) {
75        if (!high_bit_depth)
76            c->get_pixels = get_pixels_mvi;
77        c->diff_pixels = diff_pixels_mvi;
78    }
79}
80