[Mplayerxp-cvslog] SF.net SVN: mplayerxp:[19] mplayerxp
Brought to you by:
olov
Menu
▾
▴
[Mplayerxp-cvslog] SF.net SVN: mplayerxp:[19] mplayerxp
|
From: <nic...@us...> - 2009-12-06 12:02:00
|
Revision: 19
http://mplayerxp.svn.sourceforge.net/mplayerxp/?rev=19&view=rev
Author: nickols_k
Date: 2009-12-06 12:01:50 +0000 (Sun, 06 Dec 2009)
Log Message:
-----------
remove duplicated stuff from MPlayerXP
Modified Paths:
--------------
mplayerxp/libmpcodecs/ad_dshow.c
mplayerxp/libmpcodecs/codecs_ld.c
mplayerxp/libvo/vo_x11.c
mplayerxp/mplayer.c
mplayerxp/postproc/Makefile
mplayerxp/postproc/postprocess.c
mplayerxp/postproc/swscale.c
mplayerxp/postproc/swscale.h
mplayerxp/postproc/vf_1bpp.c
mplayerxp/postproc/vf_down3dright.c
mplayerxp/postproc/vf_mirror.c
mplayerxp/postproc/vf_palette.c
mplayerxp/postproc/vf_raw.c
mplayerxp/postproc/vf_rgb2bgr.c
mplayerxp/postproc/vf_rotate.c
mplayerxp/postproc/vf_test.c
mplayerxp/postproc/vf_yuy2.c
mplayerxp/postproc/vf_yvu9.c
Removed Paths:
-------------
mplayerxp/postproc/rgb2rgb.c
mplayerxp/postproc/rgb2rgb.h
mplayerxp/postproc/rgb2rgb_template.c
mplayerxp/postproc/swscale_altivec_template.c
mplayerxp/postproc/swscale_internal.h
mplayerxp/postproc/swscale_template.c
mplayerxp/postproc/yuv2rgb.c
mplayerxp/postproc/yuv2rgb_altivec.c
mplayerxp/postproc/yuv2rgb_mlib.c
mplayerxp/postproc/yuv2rgb_template.c
Modified: mplayerxp/libmpcodecs/ad_dshow.c
===================================================================
--- mplayerxp/libmpcodecs/ad_dshow.c 2009-12-06 09:53:07 UTC (rev 18)
+++ mplayerxp/libmpcodecs/ad_dshow.c 2009-12-06 12:01:50 UTC (rev 19)
@@ -40,10 +40,10 @@
static int load_lib( const char *libname )
{
if(!(dll_handle=ld_codec(libname,NULL))) return 0;
- DS_AudioDecoder_Open_ptr = dlsym(dll_handle,"DS_AudioDecoder_Open");
- DS_AudioDecoder_Destroy_ptr = dlsym(dll_handle,"DS_AudioDecoder_Destroy");
- DS_AudioDecoder_Convert_ptr = dlsym(dll_handle,"DS_AudioDecoder_Convert");
- DS_AudioDecoder_GetSrcSize_ptr = dlsym(dll_handle,"DS_AudioDecoder_GetSrcSize");
+ DS_AudioDecoder_Open_ptr = ld_sym(dll_handle,"DS_AudioDecoder_Open");
+ DS_AudioDecoder_Destroy_ptr = ld_sym(dll_handle,"DS_AudioDecoder_Destroy");
+ DS_AudioDecoder_Convert_ptr = ld_sym(dll_handle,"DS_AudioDecoder_Convert");
+ DS_AudioDecoder_GetSrcSize_ptr = ld_sym(dll_handle,"DS_AudioDecoder_GetSrcSize");
return DS_AudioDecoder_Open_ptr && DS_AudioDecoder_Convert_ptr &&
DS_AudioDecoder_GetSrcSize_ptr && DS_AudioDecoder_Destroy_ptr;
}
Modified: mplayerxp/libmpcodecs/codecs_ld.c
===================================================================
--- mplayerxp/libmpcodecs/codecs_ld.c 2009-12-06 09:53:07 UTC (rev 18)
+++ mplayerxp/libmpcodecs/codecs_ld.c 2009-12-06 12:01:50 UTC (rev 19)
@@ -43,7 +43,8 @@
void * ld_sym(void *handle,const char *sym_name)
{
void *rval;
- if(!(rval=dlsym(handle,sym_name)))
+ if(!(rval=dlsym(handle,sym_name))) {
MSG_ERR(MSGTR_CODEC_DLL_SYM_ERR,sym_name);
+ }
return rval;
}
Modified: mplayerxp/libvo/vo_x11.c
===================================================================
--- mplayerxp/libvo/vo_x11.c 2009-12-06 09:53:07 UTC (rev 18)
+++ mplayerxp/libvo/vo_x11.c 2009-12-06 12:01:50 UTC (rev 19)
@@ -48,7 +48,7 @@
#define MSG_D(args...)
#define LOCK_VDECODING() { MSG_D(DA_PREFIX"LOCK_VDECODING\n"); pthread_mutex_lock(&vdecoding_mutex); }
-#include "../postproc/rgb2rgb.h" /* for MODE_RGB(BGR) definitions */
+#include "../postproc/swscale.h" /* for MODE_RGB(BGR) definitions */
#include "video_out_internal.h"
#include "dri_vo.h"
#include "../mp_image.h"
Modified: mplayerxp/mplayer.c
===================================================================
--- mplayerxp/mplayer.c 2009-12-06 09:53:07 UTC (rev 18)
+++ mplayerxp/mplayer.c 2009-12-06 12:01:50 UTC (rev 19)
@@ -655,7 +655,7 @@
pinfo[xp_id].current_module="uninit_xp";
uninit_dec_ahead(0);
}
-
+
if (mask&INITED_SPUDEC){
inited_flags&=~INITED_SPUDEC;
pinfo[xp_id].current_module="uninit_spudec";
@@ -746,6 +746,7 @@
pinfo[xp_id].current_module="exit_player";
+ sws_uninit();
if(how) MSG_HINT(MSGTR_Exiting,how);
MSG_DBG2("max framesize was %d bytes\n",max_framesize);
mp_msg_uninit();
@@ -2048,7 +2049,6 @@
memset(&vstat,0,sizeof(video_stat_t));
mp_msg_init(MSGL_STATUS);
-
MSG_INFO("%s",banner_text);
// memset(&vtune,0,sizeof(vo_tune_info_t));
/* Test for cpu capabilities (and corresponding OS support) for optimizing */
@@ -2065,8 +2065,12 @@
xp_num_cpu=get_number_cpu();
#if defined( ARCH_X86 ) || defined(ARCH_X86_64)
- get_mmx_optimizations();
+ get_mmx_optimizations();
#endif
+ if(!sws_init()) {
+ MSG_ERR("MPlayerXP requires working copy of libswscaler\n");
+ return 0;
+ }
if(shuffle_playback) playtree->flags|=PLAY_TREE_RND;
else playtree->flags&=~PLAY_TREE_RND;
playtree = play_tree_cleanup(playtree);
@@ -2090,7 +2094,7 @@
if(!filename){
// no file/vcd/dvd -> show HELP:
MSG_INFO("%s",help_text);
- exit(0);
+ return 0;
}
// Many users forget to include command line in bugreports...
Modified: mplayerxp/postproc/Makefile
===================================================================
--- mplayerxp/postproc/Makefile 2009-12-06 09:53:07 UTC (rev 18)
+++ mplayerxp/postproc/Makefile 2009-12-06 12:01:50 UTC (rev 19)
@@ -6,7 +6,7 @@
SUBDIRS=libmenu
DO_MAKE = @ for i in $(SUBDIRS); do $(MAKE) -C $$i $@ || exit; done
-SRCS=postprocess.c swscale.c rgb2rgb.c yuv2rgb.c
+SRCS=postprocess.c swscale.c
SRCS+=af.c af_ao2.c af_crystality.c af_mp.c af_dummy.c af_delay.c af_channels.c af_format.c af_resample.c af_volume.c af_equalizer.c af_tools.c af_comp.c af_gate.c af_pan.c af_surround.c af_sub.c af_export.c af_volnorm.c af_extrastereo.c af_lp.c af_dyn.c af_echo3d.c af_hrtf.c af_ffenc.c af_raw.c af_karaoke.c af_center.c af_sinesuppress.c af_scaletempo.c
SRCS+=dsp.c
SRCS+=vf.c vf_vo.c vf_expand.c vf_format.c vf_yuy2.c vf_rgb2bgr.c vf_rotate.c vf_mirror.c vf_palette.c vf_test.c vf_noise.c vf_yvu9.c vf_rectangle.c vf_eq.c vf_dint.c vf_1bpp.c vf_unsharp.c vf_swapuv.c vf_il.c vf_smartblur.c vf_perspective.c vf_down3dright.c vf_denoise3d.c vf_aspect.c vf_softpulldown.c vf_delogo.c vf_yuvcsp.c vf_pp.c vf_scale.c vf_panscan.c vf_raw.c vf_ow.c vf_2xsai.c vf_framestep.c vf_menu.c
Modified: mplayerxp/postproc/postprocess.c
===================================================================
--- mplayerxp/postproc/postprocess.c 2009-12-06 09:53:07 UTC (rev 18)
+++ mplayerxp/postproc/postprocess.c 2009-12-06 12:01:50 UTC (rev 19)
@@ -30,12 +30,12 @@
static int load_dll(const char *libname)
{
if(!(dll_handle=ld_codec(libname,"http://ffmpeg.sf.net"))) return 0;
- pp_postprocess_ptr=dlsym(dll_handle,"pp_postprocess");
- pp_get_context_ptr=dlsym(dll_handle,"pp_get_context");
- pp_free_context_ptr=dlsym(dll_handle,"pp_free_context");
- pp_get_mode_by_name_and_quality_ptr=dlsym(dll_handle,"pp_get_mode_by_name_and_quality");
- pp_free_mode_ptr=dlsym(dll_handle,"pp_free_mode");
- pp_help_ptr=dlsym(dll_handle,"pp_help");
+ pp_postprocess_ptr=ld_sym(dll_handle,"pp_postprocess");
+ pp_get_context_ptr=ld_sym(dll_handle,"pp_get_context");
+ pp_free_context_ptr=ld_sym(dll_handle,"pp_free_context");
+ pp_get_mode_by_name_and_quality_ptr=ld_sym(dll_handle,"pp_get_mode_by_name_and_quality");
+ pp_free_mode_ptr=ld_sym(dll_handle,"pp_free_mode");
+ pp_help_ptr=ld_sym(dll_handle,"pp_help");
return pp_postprocess_ptr && pp_get_context_ptr && pp_free_context_ptr &&
pp_get_mode_by_name_and_quality_ptr && pp_free_mode_ptr;
}
@@ -48,8 +48,8 @@
{
MSG_ERR("Detected error during loading libpostproc"SLIBSUFFIX"! Try to upgrade this codec\n");
return 0;
- }
- return 1;
+ }
+ return 1;
}
extern void exit_player(char *);
Deleted: mplayerxp/postproc/rgb2rgb.c
===================================================================
--- mplayerxp/postproc/rgb2rgb.c 2009-12-06 09:53:07 UTC (rev 18)
+++ mplayerxp/postproc/rgb2rgb.c 2009-12-06 12:01:50 UTC (rev 19)
@@ -1,642 +0,0 @@
-/*
- *
- * rgb2rgb.c, Software RGB to RGB convertor
- * pluralize by Software PAL8 to RGB convertor
- * Software YUV to YUV convertor
- * Software YUV to RGB convertor
- * Written by Nickols_K.
- * palette & yuv & runtime cpu stuff by Michael (mic...@gm...) (under GPL)
- */
-#include <inttypes.h>
-#include "../config.h"
-#include "rgb2rgb.h"
-#include "swscale.h"
-#include "../mangle.h"
-#include "../bswap.h"
-#include "../libvo/fastmemcpy.h"
-#include "../cpudetect.h"
-
-#define FAST_BGR2YV12 // use 7 bit coeffs instead of 15bit
-
-rgb2rgb_f rgb24to32;
-rgb2rgb_f rgb24to16;
-rgb2rgb_f rgb24to15;
-rgb2rgb_f rgb32to24;
-rgb2rgb_f rgb32to16;
-rgb2rgb_f rgb32to15;
-rgb2rgb_f rgb15to16;
-rgb2rgb_f rgb15to24;
-rgb2rgb_f rgb15to32;
-rgb2rgb_f rgb16to15;
-rgb2rgb_f rgb16to24;
-rgb2rgb_f rgb16to32;
-rgb2rgb_f rgb24tobgr24;
-rgb2rgb_f rgb24tobgr16;
-rgb2rgb_f rgb24tobgr15;
-rgb2rgb_f rgb32tobgr32;
-rgb2rgb_f rgb32tobgr16;
-rgb2rgb_f rgb32tobgr15;
-
-planar2packet_f yv12toyuy2;
-planar2packet_f yv12touyvy;
-planar2packet_f yuv422ptoyuy2;
-packet2planar_f yuy2toyv12;
-packet2planar_f rgb24toyv12;
-
-void (* __FASTCALL__ planar2x)(const uint8_t *src, uint8_t *dst, int width, int height, int srcStride, int dstStride);
-void (* __FASTCALL__ interleaveBytes)(uint8_t *src1, uint8_t *src2, uint8_t *dst,
- unsigned width, unsigned height, int src1Stride,
- int src2Stride, int dstStride);
-void (* __FASTCALL__ vu9_to_vu12)(const uint8_t *src1, const uint8_t *src2,
- uint8_t *dst1, uint8_t *dst2,
- unsigned width, unsigned height,
- int srcStride1, int srcStride2,
- int dstStride1, int dstStride2);
-void (* __FASTCALL__ yvu9_to_yuy2)(const uint8_t *src1, const uint8_t *src2, const uint8_t *src3,
- uint8_t *dst,
- unsigned width, unsigned height,
- int srcStride1, int srcStride2,
- int srcStride3, int dstStride);
-void (* __FASTCALL__ yvu9toyv12)(const uint8_t *ysrc, const uint8_t *usrc, const uint8_t *vsrc,
- uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
- unsigned int width, unsigned int height, int lumStride, int chromStride);
-#if defined( ARCH_X86 ) || defined(ARCH_X86_64)
-static const uint64_t mmx_null __attribute__((used)) __attribute__((aligned(8))) = 0x0000000000000000ULL;
-static const uint64_t mmx_one __attribute__((used)) __attribute__((aligned(8))) = 0xFFFFFFFFFFFFFFFFULL;
-static const uint64_t mask32b __attribute__((used)) __attribute__((aligned(8))) = 0x000000FF000000FFULL;
-static const uint64_t mask32g __attribute__((used)) __attribute__((aligned(8))) = 0x0000FF000000FF00ULL;
-static const uint64_t mask32r __attribute__((used)) __attribute__((aligned(8))) = 0x00FF000000FF0000ULL;
-static const uint64_t mask32 __attribute__((used)) __attribute__((aligned(8))) = 0x00FFFFFF00FFFFFFULL;
-static const uint64_t mask3216br __attribute__((used)) __attribute__((aligned(8)))=0x00F800F800F800F8ULL;
-static const uint64_t mask3216g __attribute__((used)) __attribute__((aligned(8)))=0x0000FC000000FC00ULL;
-static const uint64_t mask3215g __attribute__((used)) __attribute__((aligned(8)))=0x0000F8000000F800ULL;
-static const uint64_t mul3216 __attribute__((used)) __attribute__((aligned(8))) = 0x2000000420000004ULL;
-static const uint64_t mul3215 __attribute__((used)) __attribute__((aligned(8))) = 0x2000000820000008ULL;
-static const uint64_t mask24b __attribute__((used)) __attribute__((aligned(8))) = 0x00FF0000FF0000FFULL;
-static const uint64_t mask24g __attribute__((used)) __attribute__((aligned(8))) = 0xFF0000FF0000FF00ULL;
-static const uint64_t mask24r __attribute__((used)) __attribute__((aligned(8))) = 0x0000FF0000FF0000ULL;
-static const uint64_t mask24l __attribute__((used)) __attribute__((aligned(8))) = 0x0000000000FFFFFFULL;
-static const uint64_t mask24h __attribute__((used)) __attribute__((aligned(8))) = 0x0000FFFFFF000000ULL;
-static const uint64_t mask24hh __attribute__((used)) __attribute__((aligned(8))) = 0xffff000000000000ULL;
-static const uint64_t mask24hhh __attribute__((used)) __attribute__((aligned(8))) = 0xffffffff00000000ULL;
-static const uint64_t mask24hhhh __attribute__((used)) __attribute__((aligned(8))) = 0xffffffffffff0000ULL;
-static const uint64_t mask15b __attribute__((used)) __attribute__((aligned(8))) = 0x001F001F001F001FULL; /* 00000000 00011111 xxB */
-static const uint64_t mask15rg __attribute__((used)) __attribute__((aligned(8))) = 0x7FE07FE07FE07FE0ULL; /* 01111111 11100000 RGx */
-static const uint64_t mask15s __attribute__((used)) __attribute__((aligned(8))) = 0xFFE0FFE0FFE0FFE0ULL;
-static const uint64_t mask15g __attribute__((used)) __attribute__((aligned(8))) = 0x03E003E003E003E0ULL;
-static const uint64_t mask15r __attribute__((used)) __attribute__((aligned(8))) = 0x7C007C007C007C00ULL;
-#define mask16b mask15b
-static const uint64_t mask16g __attribute__((used)) __attribute__((aligned(8))) = 0x07E007E007E007E0ULL;
-static const uint64_t mask16r __attribute__((used)) __attribute__((aligned(8))) = 0xF800F800F800F800ULL;
-static const uint64_t red_16mask __attribute__((used)) __attribute__((aligned(8))) = 0x0000f8000000f800ULL;
-static const uint64_t green_16mask __attribute__((used)) __attribute__((aligned(8)))= 0x000007e0000007e0ULL;
-static const uint64_t blue_16mask __attribute__((used)) __attribute__((aligned(8))) = 0x0000001f0000001fULL;
-static const uint64_t red_15mask __attribute__((used)) __attribute__((aligned(8))) = 0x00007c000000f800ULL;
-static const uint64_t green_15mask __attribute__((used)) __attribute__((aligned(8)))= 0x000003e0000007e0ULL;
-static const uint64_t blue_15mask __attribute__((used)) __attribute__((aligned(8))) = 0x0000001f0000001fULL;
-
-#ifdef FAST_BGR2YV12
-static const uint64_t bgr2YCoeff __attribute__((used)) __attribute__((aligned(8))) = 0x000000210041000DULL;
-static const uint64_t bgr2UCoeff __attribute__((used)) __attribute__((aligned(8))) = 0x0000FFEEFFDC0038ULL;
-static const uint64_t bgr2VCoeff __attribute__((used)) __attribute__((aligned(8))) = 0x00000038FFD2FFF8ULL;
-#else
-static const uint64_t bgr2YCoeff __attribute__((used)) __attribute__((aligned(8))) = 0x000020E540830C8BULL;
-static const uint64_t bgr2UCoeff __attribute__((used)) __attribute__((aligned(8))) = 0x0000ED0FDAC23831ULL;
-static const uint64_t bgr2VCoeff __attribute__((used)) __attribute__((aligned(8))) = 0x00003831D0E6F6EAULL;
-#endif
-static const uint64_t bgr2YOffset __attribute__((used)) __attribute__((aligned(8))) = 0x1010101010101010ULL;
-static const uint64_t bgr2UVOffset __attribute__((used)) __attribute__((aligned(8)))= 0x8080808080808080ULL;
-static const uint64_t w1111 __attribute__((used)) __attribute__((aligned(8))) = 0x0001000100010001ULL;
-
-#if 0
-static volatile uint64_t __attribute__((aligned(8))) b5Dither;
-static volatile uint64_t __attribute__((aligned(8))) g5Dither;
-static volatile uint64_t __attribute__((aligned(8))) g6Dither;
-static volatile uint64_t __attribute__((aligned(8))) r5Dither;
-
-static uint64_t __attribute__((aligned(8))) dither4[2]={
- 0x0103010301030103LL,
- 0x0200020002000200LL,};
-
-static uint64_t __attribute__((aligned(8))) dither8[2]={
- 0x0602060206020602LL,
- 0x0004000400040004LL,};
-#endif
-#endif
-
-#define RGB2YUV_SHIFT 8
-#define BY ((int)( 0.098*(1<<RGB2YUV_SHIFT)+0.5))
-#define BV ((int)(-0.071*(1<<RGB2YUV_SHIFT)+0.5))
-#define BU ((int)( 0.439*(1<<RGB2YUV_SHIFT)+0.5))
-#define GY ((int)( 0.504*(1<<RGB2YUV_SHIFT)+0.5))
-#define GV ((int)(-0.368*(1<<RGB2YUV_SHIFT)+0.5))
-#define GU ((int)(-0.291*(1<<RGB2YUV_SHIFT)+0.5))
-#define RY ((int)( 0.257*(1<<RGB2YUV_SHIFT)+0.5))
-#define RV ((int)( 0.439*(1<<RGB2YUV_SHIFT)+0.5))
-#define RU ((int)(-0.148*(1<<RGB2YUV_SHIFT)+0.5))
-
-//Note: we have C, MMX, MMX2, 3DNOW version therse no 3DNOW+MMX2 one
-//Plain C versions
-#undef HAVE_MMX
-#undef HAVE_MMX2
-#undef HAVE_3DNOW
-#undef HAVE_SSE2
-#define RENAME(a) a ## _C
-#include "rgb2rgb_template.c"
-
-#if defined( ARCH_X86 ) || defined(ARCH_X86_64)
-
-//MMX versions
-#ifdef CAN_COMPILE_MMX
-#undef RENAME
-#define HAVE_MMX
-#undef HAVE_MMX2
-#undef HAVE_3DNOW
-#undef HAVE_SSE2
-#define RENAME(a) a ## _MMX
-#include "rgb2rgb_template.c"
-#endif
-
-//MMX2 versions
-#ifdef CAN_COMPILE_MMX2
-#undef RENAME
-#define HAVE_MMX
-#define HAVE_MMX2
-#undef HAVE_3DNOW
-#undef HAVE_SSE2
-#define RENAME(a) a ## _MMX2
-#include "rgb2rgb_template.c"
-#endif
-
-//3DNOW versions
-#ifdef CAN_COMPILE_3DNOW
-#undef RENAME
-#define HAVE_MMX
-#undef HAVE_MMX2
-#define HAVE_3DNOW
-#undef HAVE_SSE2
-#define RENAME(a) a ## _3DNOW
-#include "rgb2rgb_template.c"
-#endif
-
-#endif //ARCH_X86
-
-/*
- rgb15->rgb16 Original by Strepto/Astral
- ported to gcc & bugfixed : A'rpi
- MMX2, 3DNOW optimization by Nickols_K
- 32bit c version, and and&add trick by Michael Niedermayer
-*/
-
-void sws_rgb2rgb_init(int flags){
-#if defined( ARCH_X86 ) || defined(ARCH_X86_64)
-#ifdef CAN_COMPILE_MMX2
- if(flags & SWS_CPU_CAPS_MMX2){
- rgb15to16= rgb15to16_MMX2;
- rgb15to24= rgb15to24_MMX2;
- rgb15to32= rgb15to32_MMX2;
- rgb16to24= rgb16to24_MMX2;
- rgb16to32= rgb16to32_MMX2;
- rgb16to15= rgb16to15_MMX2;
- rgb24to16= rgb24to16_MMX2;
- rgb24to15= rgb24to15_MMX2;
- rgb24to32= rgb24to32_MMX2;
- rgb32to16= rgb32to16_MMX2;
- rgb32to15= rgb32to15_MMX2;
- rgb32to24= rgb32to24_MMX2;
- rgb24tobgr15= rgb24tobgr15_MMX2;
- rgb24tobgr16= rgb24tobgr16_MMX2;
- rgb24tobgr24= rgb24tobgr24_MMX2;
- rgb32tobgr32= rgb32tobgr32_MMX2;
- rgb32tobgr16= rgb32tobgr16_MMX2;
- rgb32tobgr15= rgb32tobgr15_MMX2;
- yv12toyuy2= yv12toyuy2_MMX2;
- yv12touyvy= yv12touyvy_MMX2;
- yuv422ptoyuy2= yuv422ptoyuy2_MMX2;
- yuy2toyv12= yuy2toyv12_MMX2;
-// uyvytoyv12= uyvytoyv12_MMX2;
- yvu9toyv12= yvu9toyv12_MMX2;
- planar2x= planar2x_MMX2;
- rgb24toyv12= rgb24toyv12_MMX2;
- interleaveBytes= interleaveBytes_MMX2;
- vu9_to_vu12= vu9_to_vu12_MMX2;
- yvu9_to_yuy2= yvu9_to_yuy2_MMX2;
- }else
-#endif
-#ifdef CAN_COMPILE_3DNOW
- if(flags & SWS_CPU_CAPS_3DNOW){
- rgb15to16= rgb15to16_3DNOW;
- rgb15to24= rgb15to24_3DNOW;
- rgb15to32= rgb15to32_3DNOW;
- rgb16to24= rgb16to24_3DNOW;
- rgb16to32= rgb16to32_3DNOW;
- rgb16to15= rgb16to15_3DNOW;
- rgb24to16= rgb24to16_3DNOW;
- rgb24to15= rgb24to15_3DNOW;
- rgb24to32= rgb24to32_3DNOW;
- rgb32to16= rgb32to16_3DNOW;
- rgb32to15= rgb32to15_3DNOW;
- rgb32to24= rgb32to24_3DNOW;
- rgb24tobgr15= rgb24tobgr15_3DNOW;
- rgb24tobgr16= rgb24tobgr16_3DNOW;
- rgb24tobgr24= rgb24tobgr24_3DNOW;
- rgb32tobgr32= rgb32tobgr32_3DNOW;
- rgb32tobgr16= rgb32tobgr16_3DNOW;
- rgb32tobgr15= rgb32tobgr15_3DNOW;
- yv12toyuy2= yv12toyuy2_3DNOW;
- yv12touyvy= yv12touyvy_3DNOW;
- yuv422ptoyuy2= yuv422ptoyuy2_3DNOW;
- yuy2toyv12= yuy2toyv12_3DNOW;
-// uyvytoyv12= uyvytoyv12_3DNOW;
- yvu9toyv12= yvu9toyv12_3DNOW;
- planar2x= planar2x_3DNOW;
- rgb24toyv12= rgb24toyv12_3DNOW;
- interleaveBytes= interleaveBytes_3DNOW;
- vu9_to_vu12= vu9_to_vu12_3DNOW;
- yvu9_to_yuy2= yvu9_to_yuy2_3DNOW;
- }else
-#endif
-#ifdef CAN_COMPILE_MMX
- if(flags & SWS_CPU_CAPS_MMX){
- rgb15to16= rgb15to16_MMX;
- rgb15to24= rgb15to24_MMX;
- rgb15to32= rgb15to32_MMX;
- rgb16to24= rgb16to24_MMX;
- rgb16to32= rgb16to32_MMX;
- rgb16to15= rgb16to15_MMX;
- rgb24to16= rgb24to16_MMX;
- rgb24to15= rgb24to15_MMX;
- rgb24to32= rgb24to32_MMX;
- rgb32to16= rgb32to16_MMX;
- rgb32to15= rgb32to15_MMX;
- rgb32to24= rgb32to24_MMX;
- rgb24tobgr15= rgb24tobgr15_MMX;
- rgb24tobgr16= rgb24tobgr16_MMX;
- rgb24tobgr24= rgb24tobgr24_MMX;
- rgb32tobgr32= rgb32tobgr32_MMX;
- rgb32tobgr16= rgb32tobgr16_MMX;
- rgb32tobgr15= rgb32tobgr15_MMX;
- yv12toyuy2= yv12toyuy2_MMX;
- yv12touyvy= yv12touyvy_MMX;
- yuv422ptoyuy2= yuv422ptoyuy2_MMX;
- yuy2toyv12= yuy2toyv12_MMX;
-// uyvytoyv12= uyvytoyv12_MMX;
- yvu9toyv12= yvu9toyv12_MMX;
- planar2x= planar2x_MMX;
- rgb24toyv12= rgb24toyv12_MMX;
- interleaveBytes= interleaveBytes_MMX;
- vu9_to_vu12= vu9_to_vu12_MMX;
- yvu9_to_yuy2= yvu9_to_yuy2_MMX;
- }else
-#endif
-#endif
- {
- rgb15to16= rgb15to16_C;
- rgb15to24= rgb15to24_C;
- rgb15to32= rgb15to32_C;
- rgb16to24= rgb16to24_C;
- rgb16to32= rgb16to32_C;
- rgb16to15= rgb16to15_C;
- rgb24to16= rgb24to16_C;
- rgb24to15= rgb24to15_C;
- rgb24to32= rgb24to32_C;
- rgb32to16= rgb32to16_C;
- rgb32to15= rgb32to15_C;
- rgb32to24= rgb32to24_C;
- rgb24tobgr15= rgb24tobgr15_C;
- rgb24tobgr16= rgb24tobgr16_C;
- rgb24tobgr24= rgb24tobgr24_C;
- rgb32tobgr32= rgb32tobgr32_C;
- rgb32tobgr16= rgb32tobgr16_C;
- rgb32tobgr15= rgb32tobgr15_C;
- yv12toyuy2= yv12toyuy2_C;
- yv12touyvy= yv12touyvy_C;
- yuv422ptoyuy2= yuv422ptoyuy2_C;
- yuy2toyv12= yuy2toyv12_C;
-// uyvytoyv12= uyvytoyv12_C;
- yvu9toyv12= yvu9toyv12_C;
- planar2x= planar2x_C;
- rgb24toyv12= rgb24toyv12_C;
- interleaveBytes= interleaveBytes_C;
- vu9_to_vu12= vu9_to_vu12_C;
- yvu9_to_yuy2= yvu9_to_yuy2_C;
- }
-}
-
-/**
- * Pallete is assumed to contain bgr32
- */
-void __FASTCALL__ palette8torgb32(const uint8_t *src, uint8_t *dst, unsigned num_pixels, const uint8_t *palette)
-{
- unsigned i;
-
-/*
- for(i=0; i<num_pixels; i++)
- ((unsigned *)dst)[i] = ((unsigned *)palette)[ src[i] ];
-*/
-
- for(i=0; i<num_pixels; i++)
- {
-#ifdef WORDS_BIGENDIAN
- dst[3]= palette[ src[i]*4+2 ];
- dst[2]= palette[ src[i]*4+1 ];
- dst[1]= palette[ src[i]*4+0 ];
-#else
- //FIXME slow?
- dst[0]= palette[ src[i]*4+2 ];
- dst[1]= palette[ src[i]*4+1 ];
- dst[2]= palette[ src[i]*4+0 ];
-// dst[3]= 0; /* do we need this cleansing? */
-#endif
- dst+= 4;
- }
-}
-
-void __FASTCALL__ palette8tobgr32(const uint8_t *src, uint8_t *dst, unsigned num_pixels, const uint8_t *palette)
-{
- unsigned i;
- for(i=0; i<num_pixels; i++)
- {
-#ifdef WORDS_BIGENDIAN
- dst[3]= palette[ src[i]*4+0 ];
- dst[2]= palette[ src[i]*4+1 ];
- dst[1]= palette[ src[i]*4+2 ];
-#else
- //FIXME slow?
- dst[0]= palette[ src[i]*4+0 ];
- dst[1]= palette[ src[i]*4+1 ];
- dst[2]= palette[ src[i]*4+2 ];
-// dst[3]= 0; /* do we need this cleansing? */
-#endif
- dst+= 4;
- }
-}
-
-/**
- * Pallete is assumed to contain bgr32
- */
-void __FASTCALL__ palette8torgb24(const uint8_t *src, uint8_t *dst, unsigned num_pixels, const uint8_t *palette)
-{
- unsigned i;
-/*
- writes 1 byte o much and might cause alignment issues on some architectures?
- for(i=0; i<num_pixels; i++)
- ((unsigned *)(&dst[i*3])) = ((unsigned *)palette)[ src[i] ];
-*/
- for(i=0; i<num_pixels; i++)
- {
- //FIXME slow?
- dst[0]= palette[ src[i]*4+2 ];
- dst[1]= palette[ src[i]*4+1 ];
- dst[2]= palette[ src[i]*4+0 ];
- dst+= 3;
- }
-}
-
-void __FASTCALL__ palette8tobgr24(const uint8_t *src, uint8_t *dst, unsigned num_pixels, const uint8_t *palette)
-{
- unsigned i;
-/*
- writes 1 byte o much and might cause alignment issues on some architectures?
- for(i=0; i<num_pixels; i++)
- ((unsigned *)(&dst[i*3])) = ((unsigned *)palette)[ src[i] ];
-*/
- for(i=0; i<num_pixels; i++)
- {
- //FIXME slow?
- dst[0]= palette[ src[i]*4+0 ];
- dst[1]= palette[ src[i]*4+1 ];
- dst[2]= palette[ src[i]*4+2 ];
- dst+= 3;
- }
-}
-
-/**
- * Palette is assumed to contain bgr16, see rgb32to16 to convert the palette
- */
-void __FASTCALL__ palette8torgb16(const uint8_t *src, uint8_t *dst, unsigned num_pixels, const uint8_t *palette)
-{
- unsigned i;
- for(i=0; i<num_pixels; i++)
- {
- unsigned char r,g,b;
- b=palette[src[i]*4+0];
- g=palette[src[i]*4+1];
- r=palette[src[i]*4+2];
- ((uint16_t *)dst)[i]=((b&0xF8)<<8)+((g&0xFC)<<3)+(r>>3);
- }
-}
-void __FASTCALL__ palette8tobgr16(const uint8_t *src, uint8_t *dst, unsigned num_pixels, const uint8_t *palette)
-{
- unsigned i;
- for(i=0; i<num_pixels; i++)
- {
- unsigned char r,g,b;
- r=palette[src[i]*4+0];
- g=palette[src[i]*4+1];
- b=palette[src[i]*4+2];
- ((uint16_t *)dst)[i]=((b&0xF8)<<8)+((g&0xFC)<<3)+(r>>3);
- }
-}
-
-/**
- * Pallete is assumed to contain bgr15, see rgb32to15 to convert the palette
- */
-void __FASTCALL__ palette8torgb15(const uint8_t *src, uint8_t *dst, unsigned num_pixels, const uint8_t *palette)
-{
- unsigned i;
- for(i=0; i<num_pixels; i++)
- {
- unsigned char r,g,b;
- b=palette[src[i]*4+0];
- g=palette[src[i]*4+1];
- r=palette[src[i]*4+2];
- ((uint16_t *)dst)[i]=((b&0xF8)<<7)+((g&0xF8)<<2)+(r>>3);
- }
-}
-void __FASTCALL__ palette8tobgr15(const uint8_t *src, uint8_t *dst, unsigned num_pixels, const uint8_t *palette)
-{
- unsigned i;
- for(i=0; i<num_pixels; i++)
- {
- unsigned char r,g,b;
- r=palette[src[i]*4+0];
- g=palette[src[i]*4+1];
- b=palette[src[i]*4+2];
- ((uint16_t *)dst)[i]=((b&0xF8)<<7)+((g&0xF8)<<2)+(r>>3);
- }
-}
-
-void __FASTCALL__ rgb32tobgr24(const uint8_t *src, uint8_t *dst, unsigned int src_size)
-{
- unsigned i;
- unsigned num_pixels = src_size >> 2;
- for(i=0; i<num_pixels; i++)
- {
- dst[3*i + 0] = src[4*i + 2];
- dst[3*i + 1] = src[4*i + 1];
- dst[3*i + 2] = src[4*i + 0];
- }
-}
-
-void __FASTCALL__ rgb24tobgr32(const uint8_t *src, uint8_t *dst, unsigned int src_size)
-{
- unsigned i;
- for(i=0; 3*i<src_size; i++)
- {
- dst[4*i + 0] = src[3*i + 2];
- dst[4*i + 1] = src[3*i + 1];
- dst[4*i + 2] = src[3*i + 0];
- dst[4*i + 3] = 0;
- }
-}
-
-void __FASTCALL__ rgb16tobgr32(const uint8_t *src, uint8_t *dst, unsigned int src_size)
-{
- const uint16_t *end;
- uint8_t *d = (uint8_t *)dst;
- const uint16_t *s = (uint16_t *)src;
- end = s + src_size/2;
- while(s < end)
- {
- register uint16_t bgr;
- bgr = *s++;
- *d++ = (bgr&0xF800)>>8;
- *d++ = (bgr&0x7E0)>>3;
- *d++ = (bgr&0x1F)<<3;
- *d++ = 0;
- }
-}
-
-void __FASTCALL__ rgb16tobgr24(const uint8_t *src, uint8_t *dst, unsigned int src_size)
-{
- const uint16_t *end;
- uint8_t *d = (uint8_t *)dst;
- const uint16_t *s = (const uint16_t *)src;
- end = s + src_size/2;
- while(s < end)
- {
- register uint16_t bgr;
- bgr = *s++;
- *d++ = (bgr&0xF800)>>8;
- *d++ = (bgr&0x7E0)>>3;
- *d++ = (bgr&0x1F)<<3;
- }
-}
-
-void __FASTCALL__ rgb16tobgr16(const uint8_t *src, uint8_t *dst, unsigned int src_size)
-{
- unsigned i;
- unsigned num_pixels = src_size >> 1;
-
- for(i=0; i<num_pixels; i++)
- {
- unsigned b,g,r;
- register uint16_t rgb;
- rgb = src[2*i];
- r = rgb&0x1F;
- g = (rgb&0x7E0)>>5;
- b = (rgb&0xF800)>>11;
- dst[2*i] = (b&0x1F) | ((g&0x3F)<<5) | ((r&0x1F)<<11);
- }
-}
-
-void __FASTCALL__ rgb16tobgr15(const uint8_t *src, uint8_t *dst, unsigned int src_size)
-{
- unsigned i;
- unsigned num_pixels = src_size >> 1;
-
- for(i=0; i<num_pixels; i++)
- {
- unsigned b,g,r;
- register uint16_t rgb;
- rgb = src[2*i];
- r = rgb&0x1F;
- g = (rgb&0x7E0)>>5;
- b = (rgb&0xF800)>>11;
- dst[2*i] = (b&0x1F) | ((g&0x1F)<<5) | ((r&0x1F)<<10);
- }
-}
-
-void __FASTCALL__ rgb15tobgr32(const uint8_t *src, uint8_t *dst, unsigned int src_size)
-{
- const uint16_t *end;
- uint8_t *d = (uint8_t *)dst;
- const uint16_t *s = (const uint16_t *)src;
- end = s + src_size/2;
- while(s < end)
- {
- register uint16_t bgr;
- bgr = *s++;
- *d++ = (bgr&0x7C00)>>7;
- *d++ = (bgr&0x3E0)>>2;
- *d++ = (bgr&0x1F)<<3;
- *d++ = 0;
- }
-}
-
-void __FASTCALL__ rgb15tobgr24(const uint8_t *src, uint8_t *dst, unsigned int src_size)
-{
- const uint16_t *end;
- uint8_t *d = (uint8_t *)dst;
- const uint16_t *s = (uint16_t *)src;
- end = s + src_size/2;
- while(s < end)
- {
- register uint16_t bgr;
- bgr = *s++;
- *d++ = (bgr&0x7C00)>>7;
- *d++ = (bgr&0x3E0)>>2;
- *d++ = (bgr&0x1F)<<3;
- }
-}
-
-void __FASTCALL__ rgb15tobgr16(const uint8_t *src, uint8_t *dst, unsigned int src_size)
-{
- unsigned i;
- unsigned num_pixels = src_size >> 1;
-
- for(i=0; i<num_pixels; i++)
- {
- unsigned b,g,r;
- register uint16_t rgb;
- rgb = src[2*i];
- r = rgb&0x1F;
- g = (rgb&0x3E0)>>5;
- b = (rgb&0x7C00)>>10;
- dst[2*i] = (b&0x1F) | ((g&0x3F)<<5) | ((r&0x1F)<<11);
- }
-}
-
-void __FASTCALL__ rgb15tobgr15(const uint8_t *src, uint8_t *dst, unsigned int src_size)
-{
- unsigned i;
- unsigned num_pixels = src_size >> 1;
-
- for(i=0; i<num_pixels; i++)
- {
- unsigned b,g,r;
- register uint16_t rgb;
- rgb = src[2*i];
- r = rgb&0x1F;
- g = (rgb&0x3E0)>>5;
- b = (rgb&0x7C00)>>10;
- dst[2*i] = (b&0x1F) | ((g&0x1F)<<5) | ((r&0x1F)<<10);
- }
-}
-
-void __FASTCALL__ rgb8tobgr8(const uint8_t *src, uint8_t *dst, unsigned int src_size)
-{
- unsigned i;
- unsigned num_pixels = src_size;
- for(i=0; i<num_pixels; i++)
- {
- unsigned b,g,r;
- register uint8_t rgb;
- rgb = src[i];
- r = (rgb&0x07);
- g = (rgb&0x38)>>3;
- b = (rgb&0xC0)>>6;
- dst[i] = ((b<<1)&0x07) | ((g&0x07)<<3) | ((r&0x03)<<6);
- }
-}
Deleted: mplayerxp/postproc/rgb2rgb.h
===================================================================
--- mplayerxp/postproc/rgb2rgb.h 2009-12-06 09:53:07 UTC (rev 18)
+++ mplayerxp/postproc/rgb2rgb.h 2009-12-06 12:01:50 UTC (rev 19)
@@ -1,134 +0,0 @@
-/*
- *
- * rgb2rgb.h, Software RGB to RGB convertor
- * pluralize by Software PAL8 to RGB convertor
- * Software YUV to YUV convertor
- * Software YUV to RGB convertor
- */
-
-#ifndef RGB2RGB_INCLUDED
-#define RGB2RGB_INCLUDED
-
-// Note: do not fix the dependence on stdio.h
-
-/* A full collection of rgb to rgb(bgr) convertors */
-typedef void (* __FASTCALL__ rgb2rgb_f)(const uint8_t *src,uint8_t *dst,unsigned src_size);
-
-extern rgb2rgb_f rgb24to32;
-extern rgb2rgb_f rgb24to16;
-extern rgb2rgb_f rgb24to15;
-extern rgb2rgb_f rgb32to24;
-extern rgb2rgb_f rgb32to16;
-extern rgb2rgb_f rgb32to15;
-extern rgb2rgb_f rgb15to16;
-extern rgb2rgb_f rgb15to24;
-extern rgb2rgb_f rgb15to32;
-extern rgb2rgb_f rgb16to15;
-extern rgb2rgb_f rgb16to24;
-extern rgb2rgb_f rgb16to32;
-extern rgb2rgb_f rgb24tobgr24;
-extern rgb2rgb_f rgb24tobgr16;
-extern rgb2rgb_f rgb24tobgr15;
-extern rgb2rgb_f rgb32tobgr32;
-extern rgb2rgb_f rgb32tobgr16;
-extern rgb2rgb_f rgb32tobgr15;
-
-extern void __FASTCALL__ rgb24tobgr32(const uint8_t *src, uint8_t *dst, unsigned src_size);
-extern void __FASTCALL__ rgb32tobgr24(const uint8_t *src, uint8_t *dst, unsigned src_size);
-extern void __FASTCALL__ rgb16tobgr32(const uint8_t *src, uint8_t *dst, unsigned src_size);
-extern void __FASTCALL__ rgb16tobgr24(const uint8_t *src, uint8_t *dst, unsigned src_size);
-extern void __FASTCALL__ rgb16tobgr16(const uint8_t *src, uint8_t *dst, unsigned src_size);
-extern void __FASTCALL__ rgb16tobgr15(const uint8_t *src, uint8_t *dst, unsigned src_size);
-extern void __FASTCALL__ rgb15tobgr32(const uint8_t *src, uint8_t *dst, unsigned src_size);
-extern void __FASTCALL__ rgb15tobgr24(const uint8_t *src, uint8_t *dst, unsigned src_size);
-extern void __FASTCALL__ rgb15tobgr16(const uint8_t *src, uint8_t *dst, unsigned src_size);
-extern void __FASTCALL__ rgb15tobgr15(const uint8_t *src, uint8_t *dst, unsigned src_size);
-extern void __FASTCALL__ rgb8tobgr8(const uint8_t *src, uint8_t *dst, unsigned src_size);
-
-
-extern void __FASTCALL__ palette8torgb32(const uint8_t *src, uint8_t *dst, unsigned num_pixels, const uint8_t *palette);
-extern void __FASTCALL__ palette8tobgr32(const uint8_t *src, uint8_t *dst, unsigned num_pixels, const uint8_t *palette);
-extern void __FASTCALL__ palette8torgb24(const uint8_t *src, uint8_t *dst, unsigned num_pixels, const uint8_t *palette);
-extern void __FASTCALL__ palette8tobgr24(const uint8_t *src, uint8_t *dst, unsigned num_pixels, const uint8_t *palette);
-extern void __FASTCALL__ palette8torgb16(const uint8_t *src, uint8_t *dst, unsigned num_pixels, const uint8_t *palette);
-extern void __FASTCALL__ palette8tobgr16(const uint8_t *src, uint8_t *dst, unsigned num_pixels, const uint8_t *palette);
-extern void __FASTCALL__ palette8torgb15(const uint8_t *src, uint8_t *dst, unsigned num_pixels, const uint8_t *palette);
-extern void __FASTCALL__ palette8tobgr15(const uint8_t *src, uint8_t *dst, unsigned num_pixels, const uint8_t *palette);
-
-/**
- *
- * height should be a multiple of 2 and width should be a multiple of 16 (if this is a
- * problem for anyone then tell me, and ill fix it)
- * chrominance data is only taken from every secound line others are ignored FIXME write HQ version
- */
-//void uyvytoyv12(const uint8_t *src, uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
-
-/**
- *
- * height should be a multiple of 2 and width should be a multiple of 16 (if this is a
- * problem for anyone then tell me, and ill fix it)
- */
-
-typedef void (* __FASTCALL__ planar2packet_f)(const uint8_t *ysrc, const uint8_t *usrc,
- const uint8_t *vsrc, uint8_t *dst,
- unsigned int width, unsigned int height,
- int lumStride, int chromStride, int dstStride);
-typedef void (* __FASTCALL__ packet2planar_f)(const uint8_t *src, uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
- unsigned int width, unsigned int height,
- int lumStride, int chromStride, int srcStride);
-/**
- *
- * width should be a multiple of 16
- */
-extern planar2packet_f yv12toyuy2;
-extern planar2packet_f yuv422ptoyuy2;
-
-/**
- *
- * height should be a multiple of 2 and width should be a multiple of 16 (if this is a
- * problem for anyone then tell me, and ill fix it)
- */
-extern packet2planar_f yuy2toyv12;
-
-/**
- *
- * height should be a multiple of 2 and width should be a multiple of 16 (if this is a
- * problem for anyone then tell me, and ill fix it)
- */
-extern planar2packet_f yv12touyvy;
-
-/**
- *
- * height should be a multiple of 2 and width should be a multiple of 2 (if this is a
- * problem for anyone then tell me, and ill fix it)
- * chrominance data is only taken from every secound line others are ignored FIXME write HQ version
- */
-extern packet2planar_f rgb24toyv12;
-extern void (* __FASTCALL__ planar2x)(const uint8_t *src, uint8_t *dst, int width, int height, int srcStride, int dstStride);
-
-extern void (* __FASTCALL__ interleaveBytes)(uint8_t *src1, uint8_t *src2, uint8_t *dst,
- unsigned width, unsigned height, int src1Stride,
- int src2Stride, int dstStride);
-
-extern void (* __FASTCALL__ yvu9toyv12)(const uint8_t *ysrc, const uint8_t *usrc, const uint8_t *vsrc,
- uint8_t *ydst, uint8_t *udst, uint8_t *vdst,
- unsigned int width, unsigned int height, int lumStride, int chromStride);
-
-extern void (* __FASTCALL__ vu9_to_vu12)(const uint8_t *src1, const uint8_t *src2,
- uint8_t *dst1, uint8_t *dst2,
- unsigned width, unsigned height,
- int srcStride1, int srcStride2,
- int dstStride1, int dstStride2);
-
-extern void (* __FASTCALL__ yvu9_to_yuy2)(const uint8_t *src1, const uint8_t *src2, const uint8_t *src3,
- uint8_t *dst,
- unsigned width, unsigned height,
- int srcStride1, int srcStride2,
- int srcStride3, int dstStride);
-
-void sws_rgb2rgb_init(int flags);
-
-#define MODE_RGB 0x1
-#define MODE_BGR 0x2
-
-#endif
Deleted: mplayerxp/postproc/rgb2rgb_template.c
===================================================================
--- mplayerxp/postproc/rgb2rgb_template.c 2009-12-06 09:53:07 UTC (rev 18)
+++ mplayerxp/postproc/rgb2rgb_template.c 2009-12-06 12:01:50 UTC (rev 19)
@@ -1,3202 +0,0 @@
-/*
- *
- * rgb2rgb.c, Software RGB to RGB convertor
- * pluralize by Software PAL8 to RGB convertor
- * Software YUV to YUV convertor
- * Software YUV to RGB convertor
- * Written by Nickols_K.
- * palette & yuv & runtime cpu stuff by Michael (mic...@gm...) (under GPL)
- * lot of big-endian byteorder fixes by Alex Beregszaszi
- */
-
-#include <stddef.h>
-#include <inttypes.h> /* for __WORDSIZE */
-
-#ifndef __WORDSIZE
-// #warning You have misconfigured system and probably will lose performance!
-#define __WORDSIZE MP_WORDSIZE
-#endif
-
-#undef PREFETCH
-#undef MOVNTQ
-#undef EMMS
-#undef SFENCE
-#undef MMREG_SIZE
-#undef PREFETCHW
-#undef PAVGB
-
-#ifdef HAVE_SSE2
-#define MMREG_SIZE 16
-#else
-#define MMREG_SIZE 8
-#endif
-
-#ifdef HAVE_3DNOW
-#define PREFETCH "prefetch"
-#define PREFETCHW "prefetchw"
-#define PAVGB "pavgusb"
-#elif defined ( HAVE_MMX2 )
-#define PREFETCH "prefetchnta"
-#define PREFETCHW "prefetcht0"
-#define PAVGB "pavgb"
-#else
-#define PREFETCH "/nop"
-#define PREFETCHW "/nop"
-#endif
-
-#ifdef HAVE_3DNOW
-/* On K6 femms is faster of emms. On K7 femms is directly mapped on emms. */
-#define EMMS "femms"
-#else
-#define EMMS "emms"
-#endif
-
-#ifdef HAVE_MMX2
-#define MOVNTQ "movntq"
-#define SFENCE "sfence"
-#else
-#define MOVNTQ "movq"
-#define SFENCE "/nop"
-#endif
-
-static void __FASTCALL__ RENAME(rgb24to32)(const uint8_t *src,uint8_t *dst,unsigned src_size)
-{
- uint8_t *dest = dst;
- const uint8_t *s = src;
- const uint8_t *end;
-#ifdef HAVE_MMX
- const uint8_t *mm_end;
-#endif
- end = s + src_size;
-#ifdef HAVE_MMX
- __asm __volatile(PREFETCH" %0"::"m"(*s):"memory");
- mm_end = end - 23;
- __asm __volatile("movq %0, %%mm7"::"m"(mask32)
- :"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
- while(s < mm_end)
- {
- __asm __volatile(
- PREFETCH" 32%1\n\t"
- "movd %1, %%mm0\n\t"
- "punpckldq 3%1, %%mm0\n\t"
- "movd 6%1, %%mm1\n\t"
- "punpckldq 9%1, %%mm1\n\t"
- "movd 12%1, %%mm2\n\t"
- "punpckldq 15%1, %%mm2\n\t"
- "movd 18%1, %%mm3\n\t"
- "punpckldq 21%1, %%mm3\n\t"
- "pand %%mm7, %%mm0\n\t"
- "pand %%mm7, %%mm1\n\t"
- "pand %%mm7, %%mm2\n\t"
- "pand %%mm7, %%mm3\n\t"
- MOVNTQ" %%mm0, %0\n\t"
- MOVNTQ" %%mm1, 8%0\n\t"
- MOVNTQ" %%mm2, 16%0\n\t"
- MOVNTQ" %%mm3, 24%0"
- :"=m"(*dest)
- :"m"(*s)
- :"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
- dest += 32;
- s += 24;
- }
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
-#endif
- while(s < end)
- {
-#ifdef WORDS_BIGENDIAN
- *dest++ = 0;
- *dest++ = *s++;
- *dest++ = *s++;
- *dest++ = *s++;
-#else
- *dest++ = *s++;
- *dest++ = *s++;
- *dest++ = *s++;
- *dest++ = 0;
-#endif
- }
-}
-
-static void __FASTCALL__ RENAME(rgb32to24)(const uint8_t *src,uint8_t *dst,unsigned src_size)
-{
- uint8_t *dest = dst;
- const uint8_t *s = src;
- const uint8_t *end;
-#ifdef HAVE_MMX
- const uint8_t *mm_end;
-#endif
- end = s + src_size;
-#ifdef HAVE_MMX
- __asm __volatile(PREFETCH" %0"::"m"(*s):"memory");
- mm_end = end - 31;
- while(s < mm_end)
- {
- __asm __volatile(
- PREFETCH" 32%1\n\t"
- "movq %1, %%mm0\n\t"
- "movq 8%1, %%mm1\n\t"
- "movq 16%1, %%mm4\n\t"
- "movq 24%1, %%mm5\n\t"
- "movq %%mm0, %%mm2\n\t"
- "movq %%mm1, %%mm3\n\t"
- "movq %%mm4, %%mm6\n\t"
- "movq %%mm5, %%mm7\n\t"
- "psrlq $8, %%mm2\n\t"
- "psrlq $8, %%mm3\n\t"
- "psrlq $8, %%mm6\n\t"
- "psrlq $8, %%mm7\n\t"
- "pand %2, %%mm0\n\t"
- "pand %2, %%mm1\n\t"
- "pand %2, %%mm4\n\t"
- "pand %2, %%mm5\n\t"
- "pand %3, %%mm2\n\t"
- "pand %3, %%mm3\n\t"
- "pand %3, %%mm6\n\t"
- "pand %3, %%mm7\n\t"
- "por %%mm2, %%mm0\n\t"
- "por %%mm3, %%mm1\n\t"
- "por %%mm6, %%mm4\n\t"
- "por %%mm7, %%mm5\n\t"
-
- "movq %%mm1, %%mm2\n\t"
- "movq %%mm4, %%mm3\n\t"
- "psllq $48, %%mm2\n\t"
- "psllq $32, %%mm3\n\t"
- "pand %4, %%mm2\n\t"
- "pand %5, %%mm3\n\t"
- "por %%mm2, %%mm0\n\t"
- "psrlq $16, %%mm1\n\t"
- "psrlq $32, %%mm4\n\t"
- "psllq $16, %%mm5\n\t"
- "por %%mm3, %%mm1\n\t"
- "pand %6, %%mm5\n\t"
- "por %%mm5, %%mm4\n\t"
-
- MOVNTQ" %%mm0, %0\n\t"
- MOVNTQ" %%mm1, 8%0\n\t"
- MOVNTQ" %%mm4, 16%0"
- :"=m"(*dest)
- :"m"(*s),"m"(mask24l),
- "m"(mask24h),"m"(mask24hh),"m"(mask24hhh),"m"(mask24hhhh)
- :"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
- dest += 24;
- s += 32;
- }
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
-#endif
- while(s < end)
- {
-#ifdef WORDS_BIGENDIAN
- s++;
- *dest++ = *s++;
- *dest++ = *s++;
- *dest++ = *s++;
-#else
- *dest++ = *s++;
- *dest++ = *s++;
- *dest++ = *s++;
- s++;
-#endif
- }
-}
-
-/*
- Original by Strepto/Astral
- ported to gcc & bugfixed : A'rpi
- MMX2, 3DNOW optimization by Nickols_K
- 32bit c version, and and&add trick by Michael Niedermayer
-*/
-static void __FASTCALL__ RENAME(rgb15to16)(const uint8_t *src,uint8_t *dst,unsigned src_size)
-{
- register const uint8_t* s=src;
- register uint8_t* d=dst;
- register const uint8_t *end;
- const uint8_t *mm_end;
- end = s + src_size;
-#ifdef HAVE_MMX
- __asm __volatile(PREFETCH" %0"::"m"(*s));
- __asm __volatile("movq %0, %%mm4"::"m"(mask15s)
- :"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
- mm_end = end - 15;
- while(s<mm_end)
- {
- __asm __volatile(
- PREFETCH" 32%1\n\t"
- "movq %1, %%mm0\n\t"
- "movq 8%1, %%mm2\n\t"
- "movq %%mm0, %%mm1\n\t"
- "movq %%mm2, %%mm3\n\t"
- "pand %%mm4, %%mm0\n\t"
- "pand %%mm4, %%mm2\n\t"
- "paddw %%mm1, %%mm0\n\t"
- "paddw %%mm3, %%mm2\n\t"
- MOVNTQ" %%mm0, %0\n\t"
- MOVNTQ" %%mm2, 8%0"
- :"=m"(*d)
- :"m"(*s)
- :"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
- d+=16;
- s+=16;
- }
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
-#endif
- mm_end = end - 3;
- while(s < mm_end)
- {
- register unsigned x= *((uint32_t *)s);
- *((uint32_t *)d) = (x&0x7FFF7FFF) + (x&0x7FE07FE0);
- d+=4;
- s+=4;
- }
- if(s < end)
- {
- register unsigned short x= *((uint16_t *)s);
- *((uint16_t *)d) = (x&0x7FFF) + (x&0x7FE0);
- }
-}
-
-static void __FASTCALL__ RENAME(rgb16to15)(const uint8_t *src,uint8_t *dst,unsigned src_size)
-{
- register const uint8_t* s=src;
- register uint8_t* d=dst;
- register const uint8_t *end;
- const uint8_t *mm_end;
- end = s + src_size;
-#ifdef HAVE_MMX
- __asm __volatile(PREFETCH" %0"::"m"(*s));
- __asm __volatile( "movq %0, %%mm7\n\t"
- "movq %1, %%mm6"::"m"(mask15rg),"m"(mask15b)
- :"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
- mm_end = end - 15;
- while(s<mm_end)
- {
- __asm __volatile(
- PREFETCH" 32%1\n\t"
- "movq %1, %%mm0\n\t"
- "movq 8%1, %%mm2\n\t"
- "movq %%mm0, %%mm1\n\t"
- "movq %%mm2, %%mm3\n\t"
- "psrlq $1, %%mm0\n\t"
- "psrlq $1, %%mm2\n\t"
- "pand %%mm7, %%mm0\n\t"
- "pand %%mm7, %%mm2\n\t"
- "pand %%mm6, %%mm1\n\t"
- "pand %%mm6, %%mm3\n\t"
- "por %%mm1, %%mm0\n\t"
- "por %%mm3, %%mm2\n\t"
- MOVNTQ" %%mm0, %0\n\t"
- MOVNTQ" %%mm2, 8%0"
- :"=m"(*d)
- :"m"(*s)
- :"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
- d+=16;
- s+=16;
- }
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
-#endif
- mm_end = end - 3;
- while(s < mm_end)
- {
- register uint32_t x= *((uint32_t *)s);
- *((uint32_t *)d) = ((x>>1)&0x7FE07FE0) | (x&0x001F001F);
- s+=4;
- d+=4;
- }
- if(s < end)
- {
- register uint16_t x= *((uint16_t *)s);
- *((uint16_t *)d) = ((x>>1)&0x7FE0) | (x&0x001F);
- s+=2;
- d+=2;
- }
-}
-
-static void __FASTCALL__ RENAME(rgb32to16)(const uint8_t *src, uint8_t *dst, unsigned src_size)
-{
- const uint8_t *s = src;
- const uint8_t *end;
-#ifdef HAVE_MMX
- const uint8_t *mm_end;
-#endif
- uint16_t *d = (uint16_t *)dst;
- end = s + src_size;
-#ifdef HAVE_MMX
- mm_end = end - 15;
-#if 1 //is faster only if multiplies are reasonable fast (FIXME figure out on which cpus this is faster, on Athlon its slightly faster)
- asm volatile(
- "movq %3, %%mm5 \n\t"
- "movq %4, %%mm6 \n\t"
- "movq %5, %%mm7 \n\t"
- ".balign 16 \n\t"
- "1: \n\t"
- PREFETCH" 32(%1) \n\t"
- "movd (%1), %%mm0 \n\t"
- "movd 4(%1), %%mm3 \n\t"
- "punpckldq 8(%1), %%mm0 \n\t"
- "punpckldq 12(%1), %%mm3 \n\t"
- "movq %%mm0, %%mm1 \n\t"
- "movq %%mm3, %%mm4 \n\t"
- "pand %%mm6, %%mm0 \n\t"
- "pand %%mm6, %%mm3 \n\t"
- "pmaddwd %%mm7, %%mm0 \n\t"
- "pmaddwd %%mm7, %%mm3 \n\t"
- "pand %%mm5, %%mm1 \n\t"
- "pand %%mm5, %%mm4 \n\t"
- "por %%mm1, %%mm0 \n\t"
- "por %%mm4, %%mm3 \n\t"
- "psrld $5, %%mm0 \n\t"
- "pslld $11, %%mm3 \n\t"
- "por %%mm3, %%mm0 \n\t"
- MOVNTQ" %%mm0, (%0) \n\t"
- "add $16, %1 \n\t"
- "add $8, %0 \n\t"
- "cmp %2, %1 \n\t"
- " jb 1b \n\t"
- : "+r" (d), "+r"(s)
- : "r" (mm_end), "m" (mask3216g), "m" (mask3216br), "m" (mul3216)
- : "memory", "cc"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
-#else
- __asm __volatile(PREFETCH" %0"::"m"(*src):"memory");
- __asm __volatile(
- "movq %0, %%mm7\n\t"
- "movq %1, %%mm6\n\t"
- ::"m"(red_16mask),"m"(green_16mask)
- :"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
- while(s < mm_end)
- {
- __asm __volatile(
- PREFETCH" 32%1\n\t"
- "movd %1, %%mm0\n\t"
- "movd 4%1, %%mm3\n\t"
- "punpckldq 8%1, %%mm0\n\t"
- "punpckldq 12%1, %%mm3\n\t"
- "movq %%mm0, %%mm1\n\t"
- "movq %%mm0, %%mm2\n\t"
- "movq %%mm3, %%mm4\n\t"
- "movq %%mm3, %%mm5\n\t"
- "psrlq $3, %%mm0\n\t"
- "psrlq $3, %%mm3\n\t"
- "pand %2, %%mm0\n\t"
- "pand %2, %%mm3\n\t"
- "psrlq $5, %%mm1\n\t"
- "psrlq $5, %%mm4\n\t"
- "pand %%mm6, %%mm1\n\t"
- "pand %%mm6, %%mm4\n\t"
- "psrlq $8, %%mm2\n\t"
- "psrlq $8, %%mm5\n\t"
- "pand %%mm7, %%mm2\n\t"
- "pand %%mm7, %%mm5\n\t"
- "por %%mm1, %%mm0\n\t"
- "por %%mm4, %%mm3\n\t"
- "por %%mm2, %%mm0\n\t"
- "por %%mm5, %%mm3\n\t"
- "psllq $16, %%mm3\n\t"
- "por %%mm3, %%mm0\n\t"
- MOVNTQ" %%mm0, %0\n\t"
- :"=m"(*d):"m"(*s),"m"(blue_16mask)
- :"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
- d += 4;
- s += 16;
- }
-#endif
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
-#endif
- while(s < end)
- {
- // FIXME on bigendian
- const int src= *s; s += 4;
- *d++ = ((src&0xFF)>>3) + ((src&0xFC00)>>5) + ((src&0xF80000)>>8);
-// *d++ = ((src>>3)&0x1F) + ((src>>5)&0x7E0) + ((src>>8)&0xF800);
- }
-}
-
-static void __FASTCALL__ RENAME(rgb32tobgr16)(const uint8_t *src, uint8_t *dst, unsigned int src_size)
-{
- const uint8_t *s = src;
- const uint8_t *end;
-#ifdef HAVE_MMX
- const uint8_t *mm_end;
-#endif
- uint16_t *d = (uint16_t *)dst;
- end = s + src_size;
-#ifdef HAVE_MMX
- __asm __volatile(PREFETCH" %0"::"m"(*src):"memory");
- __asm __volatile(
- "movq %0, %%mm7\n\t"
- "movq %1, %%mm6\n\t"
- ::"m"(red_16mask),"m"(green_16mask)
- :"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
- mm_end = end - 15;
- while(s < mm_end)
- {
- __asm __volatile(
- PREFETCH" 32%1\n\t"
- "movd %1, %%mm0\n\t"
- "movd 4%1, %%mm3\n\t"
- "punpckldq 8%1, %%mm0\n\t"
- "punpckldq 12%1, %%mm3\n\t"
- "movq %%mm0, %%mm1\n\t"
- "movq %%mm0, %%mm2\n\t"
- "movq %%mm3, %%mm4\n\t"
- "movq %%mm3, %%mm5\n\t"
- "psllq $8, %%mm0\n\t"
- "psllq $8, %%mm3\n\t"
- "pand %%mm7, %%mm0\n\t"
- "pand %%mm7, %%mm3\n\t"
- "psrlq $5, %%mm1\n\t"
- "psrlq $5, %%mm4\n\t"
- "pand %%mm6, %%mm1\n\t"
- "pand %%mm6, %%mm4\n\t"
- "psrlq $19, %%mm2\n\t"
- "psrlq $19, %%mm5\n\t"
- "pand %2, %%mm2\n\t"
- "pand %2, %%mm5\n\t"
- "por %%mm1, %%mm0\n\t"
- "por %%mm4, %%mm3\n\t"
- "por %%mm2, %%mm0\n\t"
- "por %%mm5, %%mm3\n\t"
- "psllq $16, %%mm3\n\t"
- "por %%mm3, %%mm0\n\t"
- MOVNTQ" %%mm0, %0\n\t"
- :"=m"(*d):"m"(*s),"m"(blue_16mask)
- :"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
- d += 4;
- s += 16;
- }
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
-#endif
- while(s < end)
- {
- // FIXME on bigendian
- const int src= *s; s += 4;
- *d++ = ((src&0xF8)<<8) + ((src&0xFC00)>>5) + ((src&0xF80000)>>19);
- }
-}
-
-static void __FASTCALL__ RENAME(rgb32to15)(const uint8_t *src, uint8_t *dst, unsigned src_size)
-{
- const uint8_t *s = src;
- const uint8_t *end;
-#ifdef HAVE_MMX
- const uint8_t *mm_end;
-#endif
- uint16_t *d = (uint16_t *)dst;
- end = s + src_size;
-#ifdef HAVE_MMX
- mm_end = end - 15;
-#if 1 //is faster only if multiplies are reasonable fast (FIXME figure out on which cpus this is faster, on Athlon its slightly faster)
- asm volatile(
- "movq %3, %%mm5 \n\t"
- "movq %4, %%mm6 \n\t"
- "movq %5, %%mm7 \n\t"
- ".balign 16 \n\t"
- "1: \n\t"
- PREFETCH" 32(%1) \n\t"
- "movd (%1), %%mm0 \n\t"
- "movd 4(%1), %%mm3 \n\t"
- "punpckldq 8(%1), %%mm0 \n\t"
- "punpckldq 12(%1), %%mm3 \n\t"
- "movq %%mm0, %%mm1 \n\t"
- "movq %%mm3, %%mm4 \n\t"
- "pand %%mm6, %%mm0 \n\t"
- "pand %%mm6, %%mm3 \n\t"
- "pmaddwd %%mm7, %%mm0 \n\t"
- "pmaddwd %%mm7, %%mm3 \n\t"
- "pand %%mm5, %%mm1 \n\t"
- "pand %%mm5, %%mm4 \n\t"
- "por %%mm1, %%mm0 \n\t"
- "por %%mm4, %%mm3 \n\t"
- "psrld $6, %%mm0 \n\t"
- "pslld $10, %%mm3 \n\t"
- "por %%mm3, %%mm0 \n\t"
- MOVNTQ" %%mm0, (%0) \n\t"
- "add $16, %1 \n\t"
- "add $8, %0 \n\t"
- "cmp %2, %1 \n\t"
- " jb 1b \n\t"
- : "+r" (d), "+r"(s)
- : "r" (mm_end), "m" (mask3215g), "m" (mask3216br), "m" (mul3215)
- : "memory", "cc"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
-#else
- __asm __volatile(PREFETCH" %0"::"m"(*src):"memory");
- __asm __volatile(
- "movq %0, %%mm7\n\t"
- "movq %1, %%mm6\n\t"
- ::"m"(red_15mask),"m"(green_15mask)
- :"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
- while(s < mm_end)
- {
- __asm __volatile(
- PREFETCH" 32%1\n\t"
- "movd %1, %%mm0\n\t"
- "movd 4%1, %%mm3\n\t"
- "punpckldq 8%1, %%mm0\n\t"
- "punpckldq 12%1, %%mm3\n\t"
- "movq %%mm0, %%mm1\n\t"
- "movq %%mm0, %%mm2\n\t"
- "movq %%mm3, %%mm4\n\t"
- "movq %%mm3, %%mm5\n\t"
- "psrlq $3, %%mm0\n\t"
- "psrlq $3, %%mm3\n\t"
- "pand %2, %%mm0\n\t"
- "pand %2, %%mm3\n\t"
- "psrlq $6, %%mm1\n\t"
- "psrlq $6, %%mm4\n\t"
- "pand %%mm6, %%mm1\n\t"
- "pand %%mm6, %%mm4\n\t"
- "psrlq $9, %%mm2\n\t"
- "psrlq $9, %%mm5\n\t"
- "pand %%mm7, %%mm2\n\t"
- "pand %%mm7, %%mm5\n\t"
- "por %%mm1, %%mm0\n\t"
- "por %%mm4, %%mm3\n\t"
- "por %%mm2, %%mm0\n\t"
- "por %%mm5, %%mm3\n\t"
- "psllq $16, %%mm3\n\t"
- "por %%mm3, %%mm0\n\t"
- MOVNTQ" %%mm0, %0\n\t"
- :"=m"(*d):"m"(*s),"m"(blue_15mask)
- :"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
- d += 4;
- s += 16;
- }
-#endif
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
-#endif
- while(s < end)
- {
- // FIXME on bigendian
- const int src= *s; s += 4;
- *d++ = ((src&0xFF)>>3) + ((src&0xF800)>>6) + ((src&0xF80000)>>9);
- }
-}
-
-static void __FASTCALL__ RENAME(rgb32tobgr15)(const uint8_t *src, uint8_t *dst, unsigned src_size)
-{
- const uint8_t *s = src;
- const uint8_t *end;
-#ifdef HAVE_MMX
- const uint8_t *mm_end;
-#endif
- uint16_t *d = (uint16_t *)dst;
- end = s + src_size;
-#ifdef HAVE_MMX
- __asm __volatile(PREFETCH" %0"::"m"(*src):"memory");
- __asm __volatile(
- "movq %0, %%mm7\n\t"
- "movq %1, %%mm6\n\t"
- ::"m"(red_15mask),"m"(green_15mask)
- :"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
- mm_end = end - 15;
- while(s < mm_end)
- {
- __asm __volatile(
- PREFETCH" 32%1\n\t"
- "movd %1, %%mm0\n\t"
- "movd 4%1, %%mm3\n\t"
- "punpckldq 8%1, %%mm0\n\t"
- "punpckldq 12%1, %%mm3\n\t"
- "movq %%mm0, %%mm1\n\t"
- "movq %%mm0, %%mm2\n\t"
- "movq %%mm3, %%mm4\n\t"
- "movq %%mm3, %%mm5\n\t"
- "psllq $7, %%mm0\n\t"
- "psllq $7, %%mm3\n\t"
- "pand %%mm7, %%mm0\n\t"
- "pand %%mm7, %%mm3\n\t"
- "psrlq $6, %%mm1\n\t"
- "psrlq $6, %%mm4\n\t"
- "pand %%mm6, %%mm1\n\t"
- "pand %%mm6, %%mm4\n\t"
- "psrlq $19, %%mm2\n\t"
- "psrlq $19, %%mm5\n\t"
- "pand %2, %%mm2\n\t"
- "pand %2, %%mm5\n\t"
- "por %%mm1, %%mm0\n\t"
- "por %%mm4, %%mm3\n\t"
- "por %%mm2, %%mm0\n\t"
- "por %%mm5, %%mm3\n\t"
- "psllq $16, %%mm3\n\t"
- "por %%mm3, %%mm0\n\t"
- MOVNTQ" %%mm0, %0\n\t"
- :"=m"(*d):"m"(*s),"m"(blue_15mask)
- :"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
- d += 4;
- s += 16;
- }
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
-#endif
- while(s < end)
- {
- // FIXME on bigendian
- const int src= *s; s += 4;
- *d++ = ((src&0xF8)<<7) + ((src&0xF800)>>6) + ((src&0xF80000)>>19);
- }
-}
-
-static void __FASTCALL__ RENAME(rgb24to16)(const uint8_t *src, uint8_t *dst, unsigned src_size)
-{
- const uint8_t *s = src;
- const uint8_t *end;
-#ifdef HAVE_MMX
- const uint8_t *mm_end;
-#endif
- uint16_t *d = (uint16_t *)dst;
- end = s + src_size;
-#ifdef HAVE_MMX
- __asm __volatile(PREFETCH" %0"::"m"(*src):"memory");
- __asm __volatile(
- "movq %0, %%mm7\n\t"
- "movq %1, %%mm6\n\t"
- ::"m"(red_16mask),"m"(green_16mask)
- :"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
- mm_end = end - 11;
- while(s < mm_end)
- {
- __asm __volatile(
- PREFETCH" 32%1\n\t"
- "movd %1, %%mm0\n\t"
- "movd 3%1, %%mm3\n\t"
- "punpckldq 6%1, %%mm0\n\t"
- "punpckldq 9%1, %%mm3\n\t"
- "movq %%mm0, %%mm1\n\t"
- "movq %%mm0, %%mm2\n\t"
- "movq %%mm3, %%mm4\n\t"
- "movq %%mm3, %%mm5\n\t"
- "psrlq $3, %%mm0\n\t"
- "psrlq $3, %%mm3\n\t"
- "pand %2, %%mm0\n\t"
- "pand %2, %%mm3\n\t"
- "psrlq $5, %%mm1\n\t"
- "psrlq $5, %%mm4\n\t"
- "pand %%mm6, %%mm1\n\t"
- "pand %%mm6, %%mm4\n\t"
- "psrlq $8, %%mm2\n\t"
- "psrlq $8, %%mm5\n\t"
- "pand %%mm7, %%mm2\n\t"
- "pand %%mm7, %%mm5\n\t"
- "por %%mm1, %%mm0\n\t"
- "por %%mm4, %%mm3\n\t"
- "por %%mm2, %%mm0\n\t"
- "por %%mm5, %%mm3\n\t"
- "psllq $16, %%mm3\n\t"
- "por %%mm3, %%mm0\n\t"
- MOVNTQ" %%mm0, %0\n\t"
- :"=m"(*d):"m"(*s),"m"(blue_16mask)
- :"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
- d += 4;
- s += 12;
- }
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
-#endif
- while(s < end)
- {
- const int b= *s++;
- const int g= *s++;
- const int r= *s++;
- *d++ = (b>>3) | ((g&0xFC)<<3) | ((r&0xF8)<<8);
- }
-}
-
-static void __FASTCALL__ RENAME(rgb24tobgr16)(const uint8_t *src, uint8_t *dst, unsigned int src_size)
-{
- const uint8_t *s = src;
- const uint8_t *end;
-#ifdef HAVE_MMX
- const uint8_t *mm_end;
-#endif
- uint16_t *d = (uint16_t *)dst;
- end = s + src_size;
-#ifdef HAVE_MMX
- __asm __volatile(PREFETCH" %0"::"m"(*src):"memory");
- __asm __volatile(
- "movq %0, %%mm7\n\t"
- "movq %1, %%mm6\n\t"
- ::"m"(red_16mask),"m"(green_16mask)
- :"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
- mm_end = end - 15;
- while(s < mm_end)
- {
- __asm __volatile(
- PREFETCH" 32%1\n\t"
- "movd %1, %%mm0\n\t"
- "movd 3%1, %%mm3\n\t"
- "punpckldq 6%1, %%mm0\n\t"
- "punpckldq 9%1, %%mm3\n\t"
- "movq %%mm0, %%mm1\n\t"
- "movq %%mm0, %%mm2\n\t"
- "movq %%mm3, %%mm4\n\t"
- "movq %%mm3, %%mm5\n\t"
- "psllq $8, %%mm0\n\t"
- "psllq $8, %%mm3\n\t"
- "pand %%mm7, %%mm0\n\t"
- "pand %%mm7, %%mm3\n\t"
- "psrlq $5, %%mm1\n\t"
- "psrlq $5, %%mm4\n\t"
- "pand %%mm6, %%mm1\n\t"
- "pand %%mm6, %%mm4\n\t"
- "psrlq $19, %%mm2\n\t"
- "psrlq $19, %%mm5\n\t"
- "pand %2, %%mm2\n\t"
- "pand %2, %%mm5\n\t"
- "por %%mm1, %%mm0\n\t"
- "por %%mm4, %%mm3\n\t"
- "por %%mm2, %%mm0\n\t"
- "por %%mm5, %%mm3\n\t"
- "psllq $16, %%mm3\n\t"
- "por %%mm3, %%mm0\n\t"
- MOVNTQ" %%mm0, %0\n\t"
- :"=m"(*d):"m"(*s),"m"(blue_16mask)
- :"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
- d += 4;
- s += 12;
- }
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
-#endif
- while(s < end)
- {
- const int r= *s++;
- const int g= *s++;
- const int b= *s++;
- *d++ = (b>>3) | ((g&0xFC)<<3) | ((r&0xF8)<<8);
- }
-}
-
-static void __FASTCALL__ RENAME(rgb24to15)(const uint8_t *src, uint8_t *dst, unsigned src_size)
-{
- const uint8_t *s = src;
- const uint8_t *end;
-#ifdef HAVE_MMX
- const uint8_t *mm_end;
-#endif
- uint16_t *d = (uint16_t *)dst;
- end = s + src_size;
-#ifdef HAVE_MMX
- __asm __volatile(PREFETCH" %0"::"m"(*src):"memory");
- __asm __volatile(
- "movq %0, %%mm7\n\t"
- "movq %1, %%mm6\n\t"
- ::"m"(red_15mask),"m"(green_15mask)
- :"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
- mm_end = end - 11;
- while(s < mm_end)
- {
- __asm __volatile(
- PREFETCH" 32%1\n\t"
- "movd %1, %%mm0\n\t"
- "movd 3%1, %%mm3\n\t"
- "punpckldq 6%1, %%mm0\n\t"
- "punpckldq 9%1, %%mm3\n\t"
- "movq %%mm0, %%mm1\n\t"
- "movq %%mm0, %%mm2\n\t"
- "movq %%mm3, %%mm4\n\t"
- "movq %%mm3, %%mm5\n\t"
- "psrlq $3, %%mm0\n\t"
- "psrlq $3, %%mm3\n\t"
- "pand %2, %%mm0\n\t"
- "pand %2, %%mm3\n\t"
- "psrlq $6, %%mm1\n\t"
- "psrlq $6, %%mm4\n\t"
- "pand %%mm6, %%mm1\n\t"
- "pand %%mm6, %%mm4\n\t"
- "psrlq $9, %%mm2\n\t"
- "psrlq $9, %%mm5\n\t"
- "pand %%mm7, %%mm2\n\t"
- "pand %%mm7, %%mm5\n\t"
- "por %%mm1, %%mm0\n\t"
- "por %%mm4, %%mm3\n\t"
- "por %%mm2, %%mm0\n\t"
- "por %%mm5, %%mm3\n\t"
- "psllq $16, %%mm3\n\t"
- "por %%mm3, %%mm0\n\t"
- MOVNTQ" %%mm0, %0\n\t"
- :"=m"(*d):"m"(*s),"m"(blue_15mask)
- :"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
- d += 4;
- s += 12;
- }
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
-#endif
- while(s < end)
- {
- const int b= *s++;
- const int g= *s++;
- const int r= *s++;
- *d++ = (b>>3) | ((g&0xF8)<<2) | ((r&0xF8)<<7);
- }
-}
-
-static void __FASTCALL__ RENAME(rgb24tobgr15)(const uint8_t *src, uint8_t *dst, unsigned src_size)
-{
- const uint8_t *s = src;
- const uint8_t *end;
-#ifdef HAVE_MMX
- const uint8_t *mm_end;
-#endif
- uint16_t *d = (uint16_t *)dst;
- end = s + src_size;
-#ifdef HAVE_MMX
- __asm __volatile(PREFETCH" %0"::"m"(*src):"memory");
- __asm __volatile(
- "movq %0, %%mm7\n\t"
- "movq %1, %%mm6\n\t"
- ::"m"(red_15mask),"m"(green_15mask)
- :"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
- mm_end = end - 15;
- while(s < mm_end)
- {
- __asm __volatile(
- PREFETCH" 32%1\n\t"
- "movd %1, %%mm0\n\t"
- "movd 3%1, %%mm3\n\t"
- "punpckldq 6%1, %%mm0\n\t"
- "punpckldq 9%1, %%mm3\n\t"
- "movq %%mm0, %%mm1\n\t"
- "movq %%mm0, %%mm2\n\t"
- "movq %%mm3, %%mm4\n\t"
- "movq %%mm3, %%mm5\n\t"
- "psllq $7, %%mm0\n\t"
- "psllq $7, %%mm3\n\t"
- "pand %%mm7, %%mm0\n\t"
- "pand %%mm7, %%mm3\n\t"
- "psrlq $6, %%mm1\n\t"
- "psrlq $6, %%mm4\n\t"
- "pand %%mm6, %%mm1\n\t"
- "pand %%mm6, %%mm4\n\t"
- "psrlq $19, %%mm2\n\t"
- "psrlq $19, %%mm5\n\t"
- "pand %2, %%mm2\n\t"
- "pand %2, %%mm5\n\t"
- "por %%mm1, %%mm0\n\t"
- "por %%mm4, %%mm3\n\t"
- "por %%mm2, %%mm0\n\t"
- "por %%mm5, %%mm3\n\t"
- "psllq $16, %%mm3\n\t"
- "por %%mm3, %%mm0\n\t"
- MOVNTQ" %%mm0, %0\n\t"
- :"=m"(*d):"m"(*s),"m"(blue_15mask)
- :"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
- d += 4;
- s += 12;
- }
- __asm __volatile(SFENCE:::"memory");
- __asm __volatile(EMMS:::"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
-#endif
- while(s < end)
- {
- const int r= *s++;
- const int g= *s++;
- const int b= *s++;
- *d++ = (b>>3) | ((g&0xF8)<<2) | ((r&0xF8)<<7);
- }
-}
-
-/*
- I use here less accurate approximation by simply
- left-shifti...
[truncated message content] |
