1 /*
2 * Copyright © 2018, VideoLAN and dav1d authors
3 * All rights reserved.
4 *
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions are met:
7 *
8 * 1. Redistributions of source code must retain the above copyright notice, this
9 * list of conditions and the following disclaimer.
10 *
11 * 2. Redistributions in binary form must reproduce the above copyright notice,
12 * this list of conditions and the following disclaimer in the documentation
13 * and/or other materials provided with the distribution.
14 *
15 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
16 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
17 * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
18 * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
19 * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
20 * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
21 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
22 * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
23 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
24 * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
25 */
26
27 #include "src/cpu.h"
28 #include "src/cdef.h"
29
30 decl_cdef_dir_fn(BF(dav1d_cdef_find_dir, neon));
31
32 void BF(dav1d_cdef_padding4, neon)(uint16_t *tmp, const pixel *src,
33 ptrdiff_t src_stride, const pixel (*left)[2],
34 const pixel *const top,
35 const pixel *const bottom, int h,
36 enum CdefEdgeFlags edges);
37 void BF(dav1d_cdef_padding8, neon)(uint16_t *tmp, const pixel *src,
38 ptrdiff_t src_stride, const pixel (*left)[2],
39 const pixel *const top,
40 const pixel *const bottom, int h,
41 enum CdefEdgeFlags edges);
42
43 // Passing edges to this function, to allow it to switch to a more
44 // optimized version for fully edged cases. Using size_t for edges,
45 // to avoid ABI differences for passing more than one argument on the stack.
46 void BF(dav1d_cdef_filter4, neon)(pixel *dst, ptrdiff_t dst_stride,
47 const uint16_t *tmp, int pri_strength,
48 int sec_strength, int dir, int damping, int h,
49 size_t edges HIGHBD_DECL_SUFFIX);
50 void BF(dav1d_cdef_filter8, neon)(pixel *dst, ptrdiff_t dst_stride,
51 const uint16_t *tmp, int pri_strength,
52 int sec_strength, int dir, int damping, int h,
53 size_t edges HIGHBD_DECL_SUFFIX);
54
55 #define DEFINE_FILTER(w, h, tmp_stride) \
56 static void \
57 cdef_filter_##w##x##h##_neon(pixel *dst, const ptrdiff_t stride, \
58 const pixel (*left)[2], \
59 const pixel *const top, \
60 const pixel *const bottom, \
61 const int pri_strength, const int sec_strength, \
62 const int dir, const int damping, \
63 const enum CdefEdgeFlags edges \
64 HIGHBD_DECL_SUFFIX) \
65 { \
66 ALIGN_STK_16(uint16_t, tmp_buf, 12 * tmp_stride + 8,); \
67 uint16_t *tmp = tmp_buf + 2 * tmp_stride + 8; \
68 BF(dav1d_cdef_padding##w, neon)(tmp, dst, stride, \
69 left, top, bottom, h, edges); \
70 BF(dav1d_cdef_filter##w, neon)(dst, stride, tmp, pri_strength, \
71 sec_strength, dir, damping, h, edges \
72 HIGHBD_TAIL_SUFFIX); \
73 }
74
75 DEFINE_FILTER(8, 8, 16)
76 DEFINE_FILTER(4, 8, 8)
77 DEFINE_FILTER(4, 4, 8)
78
cdef_dsp_init_arm(Dav1dCdefDSPContext * const c)79 static ALWAYS_INLINE void cdef_dsp_init_arm(Dav1dCdefDSPContext *const c) {
80 const unsigned flags = dav1d_get_cpu_flags();
81
82 if (!(flags & DAV1D_ARM_CPU_FLAG_NEON)) return;
83
84 c->dir = BF(dav1d_cdef_find_dir, neon);
85 c->fb[0] = cdef_filter_8x8_neon;
86 c->fb[1] = cdef_filter_4x8_neon;
87 c->fb[2] = cdef_filter_4x4_neon;
88 }
89