yading@10: /*
yading@10:  *  sh4 dsputil
yading@10:  *
yading@10:  * Copyright (c) 2003 BERO <bero@geocities.co.jp>
yading@10:  *
yading@10:  * This file is part of FFmpeg.
yading@10:  *
yading@10:  * FFmpeg is free software; you can redistribute it and/or
yading@10:  * modify it under the terms of the GNU Lesser General Public
yading@10:  * License as published by the Free Software Foundation; either
yading@10:  * version 2.1 of the License, or (at your option) any later version.
yading@10:  *
yading@10:  * FFmpeg is distributed in the hope that it will be useful,
yading@10:  * but WITHOUT ANY WARRANTY; without even the implied warranty of
yading@10:  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
yading@10:  * Lesser General Public License for more details.
yading@10:  *
yading@10:  * You should have received a copy of the GNU Lesser General Public
yading@10:  * License along with FFmpeg; if not, write to the Free Software
yading@10:  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
yading@10:  */
yading@10: 
yading@10: #include "libavutil/attributes.h"
yading@10: #include "libavcodec/avcodec.h"
yading@10: #include "libavcodec/dsputil.h"
yading@10: #include "dsputil_sh4.h"
yading@10: #include "sh4.h"
yading@10: 
yading@10: static void memzero_align8(void *dst,size_t size)
yading@10: {
yading@10:         int fpscr;
yading@10:         fp_single_enter(fpscr);
yading@10:         dst = (char *)dst + size;
yading@10:         size /= 32;
yading@10:         __asm__ volatile (
yading@10:         " fldi0 fr0\n"
yading@10:         " fldi0 fr1\n"
yading@10:         " fschg\n"  // double
yading@10:         "1: \n" \
yading@10:         " dt %1\n"
yading@10:         " fmov  dr0,@-%0\n"
yading@10:         " fmov  dr0,@-%0\n"
yading@10:         " fmov  dr0,@-%0\n"
yading@10:         " bf.s 1b\n"
yading@10:         " fmov  dr0,@-%0\n"
yading@10:         " fschg" //back to single
yading@10:         : "+r"(dst),"+r"(size) :: "memory" );
yading@10:         fp_single_leave(fpscr);
yading@10: }
yading@10: 
yading@10: static void clear_blocks_sh4(int16_t *blocks)
yading@10: {
yading@10:         memzero_align8(blocks,sizeof(int16_t)*6*64);
yading@10: }
yading@10: 
yading@10: static void idct_put(uint8_t *dest, int line_size, int16_t *block)
yading@10: {
yading@10:         int i;
yading@10:         const uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
yading@10:         ff_idct_sh4(block);
yading@10:         for(i=0;i<8;i++) {
yading@10:                 dest[0] = cm[block[0]];
yading@10:                 dest[1] = cm[block[1]];
yading@10:                 dest[2] = cm[block[2]];
yading@10:                 dest[3] = cm[block[3]];
yading@10:                 dest[4] = cm[block[4]];
yading@10:                 dest[5] = cm[block[5]];
yading@10:                 dest[6] = cm[block[6]];
yading@10:                 dest[7] = cm[block[7]];
yading@10:                 dest+=line_size;
yading@10:                 block+=8;
yading@10:         }
yading@10: }
yading@10: static void idct_add(uint8_t *dest, int line_size, int16_t *block)
yading@10: {
yading@10:         int i;
yading@10:         const uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
yading@10:         ff_idct_sh4(block);
yading@10:         for(i=0;i<8;i++) {
yading@10:                 dest[0] = cm[dest[0]+block[0]];
yading@10:                 dest[1] = cm[dest[1]+block[1]];
yading@10:                 dest[2] = cm[dest[2]+block[2]];
yading@10:                 dest[3] = cm[dest[3]+block[3]];
yading@10:                 dest[4] = cm[dest[4]+block[4]];
yading@10:                 dest[5] = cm[dest[5]+block[5]];
yading@10:                 dest[6] = cm[dest[6]+block[6]];
yading@10:                 dest[7] = cm[dest[7]+block[7]];
yading@10:                 dest+=line_size;
yading@10:                 block+=8;
yading@10:         }
yading@10: }
yading@10: 
yading@10: av_cold void ff_dsputil_init_sh4(DSPContext *c, AVCodecContext *avctx)
yading@10: {
yading@10:         const int idct_algo= avctx->idct_algo;
yading@10:         const int high_bit_depth = avctx->bits_per_raw_sample > 8;
yading@10:         ff_dsputil_init_align(c,avctx);
yading@10: 
yading@10:         if (!high_bit_depth)
yading@10:         c->clear_blocks = clear_blocks_sh4;
yading@10:         if (avctx->bits_per_raw_sample <= 8 &&
yading@10:             (idct_algo==FF_IDCT_AUTO || idct_algo==FF_IDCT_SH4)) {
yading@10:                 c->idct_put = idct_put;
yading@10:                 c->idct_add = idct_add;
yading@10:                 c->idct     = ff_idct_sh4;
yading@10:                 c->idct_permutation_type= FF_NO_IDCT_PERM;
yading@10:         }
yading@10: }