- Move v5 and v6 initialisation to separate files. - Move NEON IDCT selection to ff_dsputil_init_neon() Originally committed as revision 20163 to svn://svn.ffmpeg.org/ffmpeg/trunktags/v0.6
@@ -487,11 +487,13 @@ OBJS-$(ARCH_ARM) += arm/dsputil_arm.o \ | |||||
arm/mpegvideo_arm.o \ | arm/mpegvideo_arm.o \ | ||||
arm/simple_idct_arm.o \ | arm/simple_idct_arm.o \ | ||||
OBJS-$(HAVE_ARMV5TE) += arm/mpegvideo_armv5te.o \ | |||||
OBJS-$(HAVE_ARMV5TE) += arm/dsputil_init_armv5te.o \ | |||||
arm/mpegvideo_armv5te.o \ | |||||
arm/mpegvideo_armv5te_s.o \ | arm/mpegvideo_armv5te_s.o \ | ||||
arm/simple_idct_armv5te.o \ | arm/simple_idct_armv5te.o \ | ||||
OBJS-$(HAVE_ARMV6) += arm/simple_idct_armv6.o \ | |||||
OBJS-$(HAVE_ARMV6) += arm/dsputil_init_armv6.o \ | |||||
arm/simple_idct_armv6.o \ | |||||
OBJS-$(HAVE_ARMVFP) += arm/dsputil_vfp.o \ | OBJS-$(HAVE_ARMVFP) += arm/dsputil_vfp.o \ | ||||
arm/float_arm_vfp.o \ | arm/float_arm_vfp.o \ | ||||
@@ -20,30 +20,11 @@ | |||||
*/ | */ | ||||
#include "libavcodec/dsputil.h" | #include "libavcodec/dsputil.h" | ||||
void dsputil_init_iwmmxt(DSPContext* c, AVCodecContext *avctx); | |||||
void ff_float_init_arm_vfp(DSPContext* c, AVCodecContext *avctx); | |||||
void ff_dsputil_init_neon(DSPContext *c, AVCodecContext *avctx); | |||||
#include "dsputil_arm.h" | |||||
void j_rev_dct_ARM(DCTELEM *data); | void j_rev_dct_ARM(DCTELEM *data); | ||||
void simple_idct_ARM(DCTELEM *data); | void simple_idct_ARM(DCTELEM *data); | ||||
void simple_idct_armv5te(DCTELEM *data); | |||||
void simple_idct_put_armv5te(uint8_t *dest, int line_size, DCTELEM *data); | |||||
void simple_idct_add_armv5te(uint8_t *dest, int line_size, DCTELEM *data); | |||||
void ff_simple_idct_armv6(DCTELEM *data); | |||||
void ff_simple_idct_put_armv6(uint8_t *dest, int line_size, DCTELEM *data); | |||||
void ff_simple_idct_add_armv6(uint8_t *dest, int line_size, DCTELEM *data); | |||||
void ff_simple_idct_neon(DCTELEM *data); | |||||
void ff_simple_idct_put_neon(uint8_t *dest, int line_size, DCTELEM *data); | |||||
void ff_simple_idct_add_neon(uint8_t *dest, int line_size, DCTELEM *data); | |||||
void ff_vp3_idct_neon(DCTELEM *data); | |||||
void ff_vp3_idct_put_neon(uint8_t *dest, int line_size, DCTELEM *data); | |||||
void ff_vp3_idct_add_neon(uint8_t *dest, int line_size, DCTELEM *data); | |||||
/* XXX: local hack */ | /* XXX: local hack */ | ||||
static void (*ff_put_pixels_clamped)(const DCTELEM *block, uint8_t *pixels, int line_size); | static void (*ff_put_pixels_clamped)(const DCTELEM *block, uint8_t *pixels, int line_size); | ||||
static void (*ff_add_pixels_clamped)(const DCTELEM *block, uint8_t *pixels, int line_size); | static void (*ff_add_pixels_clamped)(const DCTELEM *block, uint8_t *pixels, int line_size); | ||||
@@ -59,8 +40,6 @@ void put_no_rnd_pixels8_xy2_arm(uint8_t *block, const uint8_t *pixels, int line_ | |||||
void put_pixels16_arm(uint8_t *block, const uint8_t *pixels, int line_size, int h); | void put_pixels16_arm(uint8_t *block, const uint8_t *pixels, int line_size, int h); | ||||
void ff_prefetch_arm(void *mem, int stride, int h); | |||||
CALL_2X_PIXELS(put_pixels16_x2_arm , put_pixels8_x2_arm , 8) | CALL_2X_PIXELS(put_pixels16_x2_arm , put_pixels8_x2_arm , 8) | ||||
CALL_2X_PIXELS(put_pixels16_y2_arm , put_pixels8_y2_arm , 8) | CALL_2X_PIXELS(put_pixels16_y2_arm , put_pixels8_y2_arm , 8) | ||||
CALL_2X_PIXELS(put_pixels16_xy2_arm, put_pixels8_xy2_arm, 8) | CALL_2X_PIXELS(put_pixels16_xy2_arm, put_pixels8_xy2_arm, 8) | ||||
@@ -101,61 +80,21 @@ int mm_support(void) | |||||
void dsputil_init_arm(DSPContext* c, AVCodecContext *avctx) | void dsputil_init_arm(DSPContext* c, AVCodecContext *avctx) | ||||
{ | { | ||||
int idct_algo= avctx->idct_algo; | |||||
ff_put_pixels_clamped = c->put_pixels_clamped; | ff_put_pixels_clamped = c->put_pixels_clamped; | ||||
ff_add_pixels_clamped = c->add_pixels_clamped; | ff_add_pixels_clamped = c->add_pixels_clamped; | ||||
if (avctx->lowres == 0) { | if (avctx->lowres == 0) { | ||||
if(idct_algo == FF_IDCT_AUTO){ | |||||
#if HAVE_NEON | |||||
idct_algo = FF_IDCT_SIMPLENEON; | |||||
#elif HAVE_ARMV6 | |||||
idct_algo = FF_IDCT_SIMPLEARMV6; | |||||
#elif HAVE_ARMV5TE | |||||
idct_algo = FF_IDCT_SIMPLEARMV5TE; | |||||
#else | |||||
idct_algo = FF_IDCT_ARM; | |||||
#endif | |||||
} | |||||
if(idct_algo==FF_IDCT_ARM){ | |||||
if(avctx->idct_algo == FF_IDCT_AUTO || | |||||
avctx->idct_algo == FF_IDCT_ARM){ | |||||
c->idct_put= j_rev_dct_ARM_put; | c->idct_put= j_rev_dct_ARM_put; | ||||
c->idct_add= j_rev_dct_ARM_add; | c->idct_add= j_rev_dct_ARM_add; | ||||
c->idct = j_rev_dct_ARM; | c->idct = j_rev_dct_ARM; | ||||
c->idct_permutation_type= FF_LIBMPEG2_IDCT_PERM; | c->idct_permutation_type= FF_LIBMPEG2_IDCT_PERM; | ||||
} else if (idct_algo==FF_IDCT_SIMPLEARM){ | |||||
} else if (avctx->idct_algo==FF_IDCT_SIMPLEARM){ | |||||
c->idct_put= simple_idct_ARM_put; | c->idct_put= simple_idct_ARM_put; | ||||
c->idct_add= simple_idct_ARM_add; | c->idct_add= simple_idct_ARM_add; | ||||
c->idct = simple_idct_ARM; | c->idct = simple_idct_ARM; | ||||
c->idct_permutation_type= FF_NO_IDCT_PERM; | c->idct_permutation_type= FF_NO_IDCT_PERM; | ||||
#if HAVE_ARMV6 | |||||
} else if (idct_algo==FF_IDCT_SIMPLEARMV6){ | |||||
c->idct_put= ff_simple_idct_put_armv6; | |||||
c->idct_add= ff_simple_idct_add_armv6; | |||||
c->idct = ff_simple_idct_armv6; | |||||
c->idct_permutation_type= FF_LIBMPEG2_IDCT_PERM; | |||||
#endif | |||||
#if HAVE_ARMV5TE | |||||
} else if (idct_algo==FF_IDCT_SIMPLEARMV5TE){ | |||||
c->idct_put= simple_idct_put_armv5te; | |||||
c->idct_add= simple_idct_add_armv5te; | |||||
c->idct = simple_idct_armv5te; | |||||
c->idct_permutation_type = FF_NO_IDCT_PERM; | |||||
#endif | |||||
#if HAVE_NEON | |||||
} else if (idct_algo==FF_IDCT_SIMPLENEON){ | |||||
c->idct_put= ff_simple_idct_put_neon; | |||||
c->idct_add= ff_simple_idct_add_neon; | |||||
c->idct = ff_simple_idct_neon; | |||||
c->idct_permutation_type = FF_PARTTRANS_IDCT_PERM; | |||||
} else if ((CONFIG_VP3_DECODER || CONFIG_VP5_DECODER || CONFIG_VP6_DECODER) && | |||||
idct_algo==FF_IDCT_VP3){ | |||||
c->idct_put= ff_vp3_idct_put_neon; | |||||
c->idct_add= ff_vp3_idct_add_neon; | |||||
c->idct = ff_vp3_idct_neon; | |||||
c->idct_permutation_type = FF_TRANSPOSE_IDCT_PERM; | |||||
#endif | |||||
} | } | ||||
} | } | ||||
@@ -176,9 +115,8 @@ void dsputil_init_arm(DSPContext* c, AVCodecContext *avctx) | |||||
c->put_no_rnd_pixels_tab[1][2] = put_no_rnd_pixels8_y2_arm; | c->put_no_rnd_pixels_tab[1][2] = put_no_rnd_pixels8_y2_arm; | ||||
c->put_no_rnd_pixels_tab[1][3] = put_no_rnd_pixels8_xy2_arm; | c->put_no_rnd_pixels_tab[1][3] = put_no_rnd_pixels8_xy2_arm; | ||||
#if HAVE_ARMV5TE | |||||
c->prefetch = ff_prefetch_arm; | |||||
#endif | |||||
if (HAVE_ARMV5TE) ff_dsputil_init_armv5te(c, avctx); | |||||
if (HAVE_ARMV6) ff_dsputil_init_armv6(c, avctx); | |||||
#if HAVE_IWMMXT | #if HAVE_IWMMXT | ||||
dsputil_init_iwmmxt(c, avctx); | dsputil_init_iwmmxt(c, avctx); | ||||
@@ -0,0 +1,30 @@ | |||||
/* | |||||
* Copyright (c) 2009 Mans Rullgard <mans@mansr.com> | |||||
* | |||||
* 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 | |||||
*/ | |||||
#ifndef AVCODEC_ARM_DSPUTIL_H | |||||
#define AVCODEC_ARM_DSPUTIL_H | |||||
void ff_dsputil_init_armv5te(DSPContext* c, AVCodecContext *avctx); | |||||
void ff_dsputil_init_armv6(DSPContext* c, AVCodecContext *avctx); | |||||
void ff_float_init_arm_vfp(DSPContext* c, AVCodecContext *avctx); | |||||
void ff_dsputil_init_neon(DSPContext *c, AVCodecContext *avctx); | |||||
void dsputil_init_iwmmxt(DSPContext* c, AVCodecContext *avctx); | |||||
#endif |
@@ -0,0 +1,41 @@ | |||||
/* | |||||
* Copyright (c) 2009 Mans Rullgard <mans@mansr.com> | |||||
* | |||||
* 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 "libavcodec/dsputil.h" | |||||
#include "dsputil_arm.h" | |||||
void simple_idct_armv5te(DCTELEM *data); | |||||
void simple_idct_put_armv5te(uint8_t *dest, int line_size, DCTELEM *data); | |||||
void simple_idct_add_armv5te(uint8_t *dest, int line_size, DCTELEM *data); | |||||
void ff_prefetch_arm(void *mem, int stride, int h); | |||||
void av_cold ff_dsputil_init_armv5te(DSPContext* c, AVCodecContext *avctx) | |||||
{ | |||||
if (!avctx->lowres && (avctx->idct_algo == FF_IDCT_AUTO || | |||||
avctx->idct_algo == FF_IDCT_SIMPLEARMV5TE)) { | |||||
c->idct_put= simple_idct_put_armv5te; | |||||
c->idct_add= simple_idct_add_armv5te; | |||||
c->idct = simple_idct_armv5te; | |||||
c->idct_permutation_type = FF_NO_IDCT_PERM; | |||||
} | |||||
c->prefetch = ff_prefetch_arm; | |||||
} |
@@ -0,0 +1,37 @@ | |||||
/* | |||||
* Copyright (c) 2009 Mans Rullgard <mans@mansr.com> | |||||
* | |||||
* 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 "libavcodec/dsputil.h" | |||||
#include "dsputil_arm.h" | |||||
void ff_simple_idct_armv6(DCTELEM *data); | |||||
void ff_simple_idct_put_armv6(uint8_t *dest, int line_size, DCTELEM *data); | |||||
void ff_simple_idct_add_armv6(uint8_t *dest, int line_size, DCTELEM *data); | |||||
void av_cold ff_dsputil_init_armv6(DSPContext* c, AVCodecContext *avctx) | |||||
{ | |||||
if (!avctx->lowres && (avctx->idct_algo == FF_IDCT_AUTO || | |||||
avctx->idct_algo == FF_IDCT_SIMPLEARMV6)) { | |||||
c->idct_put= ff_simple_idct_put_armv6; | |||||
c->idct_add= ff_simple_idct_add_armv6; | |||||
c->idct = ff_simple_idct_armv6; | |||||
c->idct_permutation_type= FF_LIBMPEG2_IDCT_PERM; | |||||
} | |||||
} |
@@ -23,6 +23,15 @@ | |||||
#include "libavcodec/avcodec.h" | #include "libavcodec/avcodec.h" | ||||
#include "libavcodec/dsputil.h" | #include "libavcodec/dsputil.h" | ||||
#include "dsputil_arm.h" | |||||
void ff_simple_idct_neon(DCTELEM *data); | |||||
void ff_simple_idct_put_neon(uint8_t *dest, int line_size, DCTELEM *data); | |||||
void ff_simple_idct_add_neon(uint8_t *dest, int line_size, DCTELEM *data); | |||||
void ff_vp3_idct_neon(DCTELEM *data); | |||||
void ff_vp3_idct_put_neon(uint8_t *dest, int line_size, DCTELEM *data); | |||||
void ff_vp3_idct_add_neon(uint8_t *dest, int line_size, DCTELEM *data); | |||||
void ff_put_pixels16_neon(uint8_t *, const uint8_t *, int, int); | void ff_put_pixels16_neon(uint8_t *, const uint8_t *, int, int); | ||||
void ff_put_pixels16_x2_neon(uint8_t *, const uint8_t *, int, int); | void ff_put_pixels16_x2_neon(uint8_t *, const uint8_t *, int, int); | ||||
@@ -185,6 +194,23 @@ void ff_vorbis_inverse_coupling_neon(float *mag, float *ang, int blocksize); | |||||
void ff_dsputil_init_neon(DSPContext *c, AVCodecContext *avctx) | void ff_dsputil_init_neon(DSPContext *c, AVCodecContext *avctx) | ||||
{ | { | ||||
if (!avctx->lowres) { | |||||
if (avctx->idct_algo == FF_IDCT_AUTO || | |||||
avctx->idct_algo == FF_IDCT_SIMPLENEON) { | |||||
c->idct_put= ff_simple_idct_put_neon; | |||||
c->idct_add= ff_simple_idct_add_neon; | |||||
c->idct = ff_simple_idct_neon; | |||||
c->idct_permutation_type = FF_PARTTRANS_IDCT_PERM; | |||||
} else if ((CONFIG_VP3_DECODER || CONFIG_VP5_DECODER || | |||||
CONFIG_VP6_DECODER) && | |||||
avctx->idct_algo == FF_IDCT_VP3) { | |||||
c->idct_put= ff_vp3_idct_put_neon; | |||||
c->idct_add= ff_vp3_idct_add_neon; | |||||
c->idct = ff_vp3_idct_neon; | |||||
c->idct_permutation_type = FF_TRANSPOSE_IDCT_PERM; | |||||
} | |||||
} | |||||
c->put_pixels_tab[0][0] = ff_put_pixels16_neon; | c->put_pixels_tab[0][0] = ff_put_pixels16_neon; | ||||
c->put_pixels_tab[0][1] = ff_put_pixels16_x2_neon; | c->put_pixels_tab[0][1] = ff_put_pixels16_x2_neon; | ||||
c->put_pixels_tab[0][2] = ff_put_pixels16_y2_neon; | c->put_pixels_tab[0][2] = ff_put_pixels16_y2_neon; | ||||
@@ -19,6 +19,7 @@ | |||||
*/ | */ | ||||
#include "libavcodec/dsputil.h" | #include "libavcodec/dsputil.h" | ||||
#include "dsputil_arm.h" | |||||
void ff_vector_fmul_vfp(float *dst, const float *src, int len); | void ff_vector_fmul_vfp(float *dst, const float *src, int len); | ||||
void ff_vector_fmul_reverse_vfp(float *dst, const float *src0, | void ff_vector_fmul_reverse_vfp(float *dst, const float *src0, | ||||