1 /*
2  * Copyright (c) 2015 Parag Salasakar (parag.salasakar@imgtec.com)
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/mips/generic_macros_msa.h"
22 #include "blockdsp_mips.h"
23 
copy_8bit_value_width8_msa(uint8_t * src,uint8_t val,int32_t src_stride,int32_t height)24 static void copy_8bit_value_width8_msa(uint8_t *src, uint8_t val,
25                                        int32_t src_stride, int32_t height)
26 {
27     int32_t cnt;
28     uint64_t dst0;
29     v16u8 val0;
30 
31     val0 = (v16u8) __msa_fill_b(val);
32     dst0 = __msa_copy_u_d((v2i64) val0, 0);
33 
34     for (cnt = (height >> 2); cnt--;) {
35         SD4(dst0, dst0, dst0, dst0, src, src_stride);
36         src += (4 * src_stride);
37     }
38 }
39 
copy_8bit_value_width16_msa(uint8_t * src,uint8_t val,int32_t src_stride,int32_t height)40 static void copy_8bit_value_width16_msa(uint8_t *src, uint8_t val,
41                                         int32_t src_stride, int32_t height)
42 {
43     int32_t cnt;
44     v16u8 val0;
45 
46     val0 = (v16u8) __msa_fill_b(val);
47 
48     for (cnt = (height >> 3); cnt--;) {
49         ST_UB8(val0, val0, val0, val0, val0, val0, val0, val0, src, src_stride);
50         src += (8 * src_stride);
51     }
52 }
53 
memset_zero_16width_msa(uint8_t * src,int32_t stride,int32_t height)54 static void memset_zero_16width_msa(uint8_t *src, int32_t stride,
55                                     int32_t height)
56 {
57     int8_t cnt;
58     v16u8 zero = { 0 };
59 
60     for (cnt = (height / 2); cnt--;) {
61         ST_UB(zero, src);
62         src += stride;
63         ST_UB(zero, src);
64         src += stride;
65     }
66 }
67 
ff_fill_block16_msa(uint8_t * src,uint8_t val,ptrdiff_t stride,int height)68 void ff_fill_block16_msa(uint8_t *src, uint8_t val, ptrdiff_t stride, int height)
69 {
70     copy_8bit_value_width16_msa(src, val, stride, height);
71 }
72 
ff_fill_block8_msa(uint8_t * src,uint8_t val,ptrdiff_t stride,int height)73 void ff_fill_block8_msa(uint8_t *src, uint8_t val, ptrdiff_t stride, int height)
74 {
75     copy_8bit_value_width8_msa(src, val, stride, height);
76 }
77 
ff_clear_block_msa(int16_t * block)78 void ff_clear_block_msa(int16_t *block)
79 {
80     memset_zero_16width_msa((uint8_t *) block, 16, 8);
81 }
82 
ff_clear_blocks_msa(int16_t * block)83 void ff_clear_blocks_msa(int16_t *block)
84 {
85     memset_zero_16width_msa((uint8_t *) block, 16, 8 * 6);
86 }
87