sh4 optimized idct & bswap patch by (BERO <bero at geocities dot co dot jp>)
[libav.git] / libavcodec / sh4 / dsputil_sh4.c
1 /*
2 * sh4 dsputil
3 *
4 * Copyright (c) 2003 BERO <bero@geocities.co.jp>
5 *
6 * This library 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 of the License, or (at your option) any later version.
10 *
11 * This library 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 this library; if not, write to the Free Software
18 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
19 */
20
21 #include "../avcodec.h"
22 #include "../dsputil.h"
23
24 static void memzero_align8(void *dst,size_t size)
25 {
26 #if defined(__SH4__) || defined(__SH4_SINGLE__) || defined(__SH4_SINGLE_ONLY__)
27 (char*)dst+=size;
28 size/=8*4;
29 asm(
30 #if defined(__SH4__)
31 " fschg\n" //single float mode
32 #endif
33 " fldi0 fr0\n"
34 " fldi0 fr1\n"
35 " fschg\n" // double
36 "1: \n" \
37 " dt %1\n"
38 " fmov dr0,@-%0\n"
39 " fmov dr0,@-%0\n"
40 " fmov dr0,@-%0\n"
41 " bf.s 1b\n"
42 " fmov dr0,@-%0\n"
43 #if defined(__SH4_SINGLE__) || defined(__SH4_SINGLE_ONLY__)
44 " fschg" //back to single
45 #endif
46 : : "r"(dst),"r"(size): "memory" );
47 #else
48 double *d = dst;
49 size/=8*4;
50 do {
51 d[0] = 0.0;
52 d[1] = 0.0;
53 d[2] = 0.0;
54 d[3] = 0.0;
55 d+=4;
56 } while(--size);
57 #endif
58 }
59
60 static void clear_blocks_sh4(DCTELEM *blocks)
61 {
62 // if (((int)blocks&7)==0)
63 memzero_align8(blocks,sizeof(DCTELEM)*6*64);
64 }
65
66 extern void idct_sh4(DCTELEM *block);
67 static void idct_put(uint8_t *dest, int line_size, DCTELEM *block)
68 {
69 idct_sh4(block);
70 int i;
71 uint8_t *cm = cropTbl + MAX_NEG_CROP;
72 for(i=0;i<8;i++) {
73 dest[0] = cm[block[0]];
74 dest[1] = cm[block[1]];
75 dest[2] = cm[block[2]];
76 dest[3] = cm[block[3]];
77 dest[4] = cm[block[4]];
78 dest[5] = cm[block[5]];
79 dest[6] = cm[block[6]];
80 dest[7] = cm[block[7]];
81 dest+=line_size;
82 block+=8;
83 }
84 }
85 static void idct_add(uint8_t *dest, int line_size, DCTELEM *block)
86 {
87 idct_sh4(block);
88 int i;
89 uint8_t *cm = cropTbl + MAX_NEG_CROP;
90 for(i=0;i<8;i++) {
91 dest[0] = cm[dest[0]+block[0]];
92 dest[1] = cm[dest[1]+block[1]];
93 dest[2] = cm[dest[2]+block[2]];
94 dest[3] = cm[dest[3]+block[3]];
95 dest[4] = cm[dest[4]+block[4]];
96 dest[5] = cm[dest[5]+block[5]];
97 dest[6] = cm[dest[6]+block[6]];
98 dest[7] = cm[dest[7]+block[7]];
99 dest+=line_size;
100 block+=8;
101 }
102 }
103
104 extern void dsputil_init_align(DSPContext* c, AVCodecContext *avctx);
105
106 void dsputil_init_sh4(DSPContext* c, AVCodecContext *avctx)
107 {
108 dsputil_init_align(c,avctx);
109
110 c->clear_blocks = clear_blocks_sh4;
111 c->idct_put = idct_put;
112 c->idct_add = idct_add;
113 c->idct_permutation_type= FF_NO_IDCT_PERM; //FF_SIMPLE_IDCT_PERM; //FF_LIBMPEG2_IDCT_PERM;
114 }