generic-library/ffmpeg
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity
f0f3425311
ffmpeg/libavcodec/arm/dsputil_iwmmxt.c
Oskar Arvidsson 8dbe585641 Adds 8-, 9- and 10-bit versions of some of the functions used by the h264 decoder. 
This patch lets e.g. dsputil_init chose dsp functions with respect to
the bit depth to decode. The naming scheme of bit depth dependent
functions is <base name>_<bit depth>[_<prefix>] (i.e. the old
clear_blocks_c is now named clear_blocks_8_c).

Note: Some of the functions for high bit depth is not dependent on the
bit depth, but only on the pixel size. This leaves some room for
optimizing binary size.

Preparatory patch for high bit depth h264 decoding support.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2011-04-10 22:33:42 +02:00

211 lines
8.9 KiB
C
Raw Blame History

/*
* iWMMXt optimized DSP utils
* Copyright (c) 2004 AGAWA Koji
*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "libavutil/cpu.h"
#include "libavcodec/dsputil.h"
#define DEF(x, y) x ## _no_rnd_ ## y ##_iwmmxt
#define SET_RND(regd) __asm__ volatile ("mov r12, #1 \n\t tbcsth " #regd ", r12":::"r12");
#define WAVG2B "wavg2b"
#include "dsputil_iwmmxt_rnd_template.c"
#undef DEF
#undef SET_RND
#undef WAVG2B
#define DEF(x, y) x ## _ ## y ##_iwmmxt
#define SET_RND(regd) __asm__ volatile ("mov r12, #2 \n\t tbcsth " #regd ", r12":::"r12");
#define WAVG2B "wavg2br"
#include "dsputil_iwmmxt_rnd_template.c"
#undef DEF
#undef SET_RND
#undef WAVG2BR
// need scheduling
#define OP(AVG) \
__asm__ volatile ( \
/* alignment */ \
"and r12, %[pixels], #7 \n\t" \
"bic %[pixels], %[pixels], #7 \n\t" \
"tmcr wcgr1, r12 \n\t" \
\
"wldrd wr0, [%[pixels]] \n\t" \
"wldrd wr1, [%[pixels], #8] \n\t" \
"add %[pixels], %[pixels], %[line_size] \n\t" \
"walignr1 wr4, wr0, wr1 \n\t" \
\
"1: \n\t" \
\
"wldrd wr2, [%[pixels]] \n\t" \
"wldrd wr3, [%[pixels], #8] \n\t" \
"add %[pixels], %[pixels], %[line_size] \n\t" \
"pld [%[pixels]] \n\t" \
"walignr1 wr5, wr2, wr3 \n\t" \
AVG " wr6, wr4, wr5 \n\t" \
"wstrd wr6, [%[block]] \n\t" \
"add %[block], %[block], %[line_size] \n\t" \
\
"wldrd wr0, [%[pixels]] \n\t" \
"wldrd wr1, [%[pixels], #8] \n\t" \
"add %[pixels], %[pixels], %[line_size] \n\t" \
"walignr1 wr4, wr0, wr1 \n\t" \
"pld [%[pixels]] \n\t" \
AVG " wr6, wr4, wr5 \n\t" \
"wstrd wr6, [%[block]] \n\t" \
"add %[block], %[block], %[line_size] \n\t" \
\
"subs %[h], %[h], #2 \n\t" \
"bne 1b \n\t" \
: [block]"+r"(block), [pixels]"+r"(pixels), [h]"+r"(h) \
: [line_size]"r"(line_size) \
: "memory", "r12");
void put_pixels8_y2_iwmmxt(uint8_t *block, const uint8_t *pixels, const int line_size, int h)
{
OP("wavg2br");
}
void put_no_rnd_pixels8_y2_iwmmxt(uint8_t *block, const uint8_t *pixels, const int line_size, int h)
{
OP("wavg2b");
}
#undef OP
void add_pixels_clamped_iwmmxt(const DCTELEM *block, uint8_t *pixels, int line_size)
{
uint8_t *pixels2 = pixels + line_size;
__asm__ volatile (
"mov r12, #4 \n\t"
"1: \n\t"
"pld [%[pixels], %[line_size2]] \n\t"
"pld [%[pixels2], %[line_size2]] \n\t"
"wldrd wr4, [%[pixels]] \n\t"
"wldrd wr5, [%[pixels2]] \n\t"
"pld [%[block], #32] \n\t"
"wunpckelub wr6, wr4 \n\t"
"wldrd wr0, [%[block]] \n\t"
"wunpckehub wr7, wr4 \n\t"
"wldrd wr1, [%[block], #8] \n\t"
"wunpckelub wr8, wr5 \n\t"
"wldrd wr2, [%[block], #16] \n\t"
"wunpckehub wr9, wr5 \n\t"
"wldrd wr3, [%[block], #24] \n\t"
"add %[block], %[block], #32 \n\t"
"waddhss wr10, wr0, wr6 \n\t"
"waddhss wr11, wr1, wr7 \n\t"
"waddhss wr12, wr2, wr8 \n\t"
"waddhss wr13, wr3, wr9 \n\t"
"wpackhus wr14, wr10, wr11 \n\t"
"wpackhus wr15, wr12, wr13 \n\t"
"wstrd wr14, [%[pixels]] \n\t"
"add %[pixels], %[pixels], %[line_size2] \n\t"
"subs r12, r12, #1 \n\t"
"wstrd wr15, [%[pixels2]] \n\t"
"add %[pixels2], %[pixels2], %[line_size2] \n\t"
"bne 1b \n\t"
: [block]"+r"(block), [pixels]"+r"(pixels), [pixels2]"+r"(pixels2)
: [line_size2]"r"(line_size << 1)
: "cc", "memory", "r12");
}
static void clear_blocks_iwmmxt(DCTELEM *blocks)
{
__asm__ volatile(
"wzero wr0 \n\t"
"mov r1, #(128 * 6 / 32) \n\t"
"1: \n\t"
"wstrd wr0, [%0] \n\t"
"wstrd wr0, [%0, #8] \n\t"
"wstrd wr0, [%0, #16] \n\t"
"wstrd wr0, [%0, #24] \n\t"
"subs r1, r1, #1 \n\t"
"add %0, %0, #32 \n\t"
"bne 1b \n\t"
: "+r"(blocks)
:
: "r1"
);
}
static void nop(uint8_t *block, const uint8_t *pixels, int line_size, int h)
{
return;
}
/* A run time test is not simple. If this file is compiled in
* then we should install the functions
*/
void ff_dsputil_init_iwmmxt(DSPContext* c, AVCodecContext *avctx)
{
int mm_flags = AV_CPU_FLAG_IWMMXT; /* multimedia extension flags */
const int h264_high_depth = avctx->codec_id == CODEC_ID_H264 && avctx->bits_per_raw_sample > 8;
if (avctx->dsp_mask) {
if (avctx->dsp_mask & AV_CPU_FLAG_FORCE)
mm_flags |= (avctx->dsp_mask & 0xffff);
else
mm_flags &= ~(avctx->dsp_mask & 0xffff);
}
if (!(mm_flags & AV_CPU_FLAG_IWMMXT)) return;
c->add_pixels_clamped = add_pixels_clamped_iwmmxt;
if (!h264_high_depth) {
c->clear_blocks = clear_blocks_iwmmxt;
c->put_pixels_tab[0][0] = put_pixels16_iwmmxt;
c->put_pixels_tab[0][1] = put_pixels16_x2_iwmmxt;
c->put_pixels_tab[0][2] = put_pixels16_y2_iwmmxt;
c->put_pixels_tab[0][3] = put_pixels16_xy2_iwmmxt;
c->put_no_rnd_pixels_tab[0][0] = put_pixels16_iwmmxt;
c->put_no_rnd_pixels_tab[0][1] = put_no_rnd_pixels16_x2_iwmmxt;
c->put_no_rnd_pixels_tab[0][2] = put_no_rnd_pixels16_y2_iwmmxt;
c->put_no_rnd_pixels_tab[0][3] = put_no_rnd_pixels16_xy2_iwmmxt;
c->put_pixels_tab[1][0] = put_pixels8_iwmmxt;
c->put_pixels_tab[1][1] = put_pixels8_x2_iwmmxt;
c->put_pixels_tab[1][2] = put_pixels8_y2_iwmmxt;
c->put_pixels_tab[1][3] = put_pixels8_xy2_iwmmxt;
c->put_no_rnd_pixels_tab[1][0] = put_pixels8_iwmmxt;
c->put_no_rnd_pixels_tab[1][1] = put_no_rnd_pixels8_x2_iwmmxt;
c->put_no_rnd_pixels_tab[1][2] = put_no_rnd_pixels8_y2_iwmmxt;
c->put_no_rnd_pixels_tab[1][3] = put_no_rnd_pixels8_xy2_iwmmxt;
c->avg_pixels_tab[0][0] = avg_pixels16_iwmmxt;
c->avg_pixels_tab[0][1] = avg_pixels16_x2_iwmmxt;
c->avg_pixels_tab[0][2] = avg_pixels16_y2_iwmmxt;
c->avg_pixels_tab[0][3] = avg_pixels16_xy2_iwmmxt;
c->avg_no_rnd_pixels_tab[0][0] = avg_pixels16_iwmmxt;
c->avg_no_rnd_pixels_tab[0][1] = avg_no_rnd_pixels16_x2_iwmmxt;
c->avg_no_rnd_pixels_tab[0][2] = avg_no_rnd_pixels16_y2_iwmmxt;
c->avg_no_rnd_pixels_tab[0][3] = avg_no_rnd_pixels16_xy2_iwmmxt;
c->avg_pixels_tab[1][0] = avg_pixels8_iwmmxt;
c->avg_pixels_tab[1][1] = avg_pixels8_x2_iwmmxt;
c->avg_pixels_tab[1][2] = avg_pixels8_y2_iwmmxt;
c->avg_pixels_tab[1][3] = avg_pixels8_xy2_iwmmxt;
c->avg_no_rnd_pixels_tab[1][0] = avg_no_rnd_pixels8_iwmmxt;
c->avg_no_rnd_pixels_tab[1][1] = avg_no_rnd_pixels8_x2_iwmmxt;
c->avg_no_rnd_pixels_tab[1][2] = avg_no_rnd_pixels8_y2_iwmmxt;
c->avg_no_rnd_pixels_tab[1][3] = avg_no_rnd_pixels8_xy2_iwmmxt;
}
}
Reference in New Issue View Git Blame Copy Permalink