[Mplayerxp-cvslog] SF.net SVN: mplayerxp:[107] mplayerxp
Brought to you by:
olov
Menu
▾
▴
[Mplayerxp-cvslog] SF.net SVN: mplayerxp:[107] mplayerxp
|
From: <nic...@us...> - 2010-01-18 15:55:32
|
Revision: 107
http://mplayerxp.svn.sourceforge.net/mplayerxp/?rev=107&view=rev
Author: nickols_k
Date: 2010-01-18 15:55:23 +0000 (Mon, 18 Jan 2010)
Log Message:
-----------
fastmemcpy-related improvements
Modified Paths:
--------------
DOCS/mplayerxp.1
mplayerxp/cfg-mplayer.h
mplayerxp/dec_ahead.h
mplayerxp/libmpcodecs/ad_mp3.c
mplayerxp/libmpcodecs/dec_video.c
mplayerxp/libmpcodecs/vd_ffmpeg.c
mplayerxp/libvo/aclib.c
mplayerxp/libvo/aclib_template.c
mplayerxp/libvo/dri_vo.h
mplayerxp/libvo/fastmemcpy.h
mplayerxp/libvo/osd.c
mplayerxp/libvo/osd.h
mplayerxp/libvo/osd_template.c
mplayerxp/libvo/video_out.c
mplayerxp/libvo/video_out.h
mplayerxp/libvo/vo_x11.c
mplayerxp/libvo/vosub_vidix.c
mplayerxp/mp_image.h
mplayerxp/postproc/af_export.c
mplayerxp/postproc/af_scaletempo.c
mplayerxp/postproc/libmenu/menu.c
mplayerxp/postproc/vf_delogo.c
mplayerxp/postproc/vf_dint.c
mplayerxp/postproc/vf_down3dright.c
mplayerxp/postproc/vf_expand.c
mplayerxp/postproc/vf_il.c
mplayerxp/postproc/vf_mirror.c
mplayerxp/postproc/vf_noise.c
mplayerxp/postproc/vf_panscan.c
mplayerxp/postproc/vf_rectangle.c
mplayerxp/postproc/vf_scale.c
mplayerxp/postproc/vf_softpulldown.c
mplayerxp/postproc/vf_unsharp.c
mplayerxp/postproc/vf_vo.c
mplayerxp/postproc/vf_yvu9.c
Removed Paths:
-------------
mplayerxp/libvo/aclib_x86_64.h
Modified: DOCS/mplayerxp.1
===================================================================
--- DOCS/mplayerxp.1 2010-01-17 18:46:44 UTC (rev 106)
+++ DOCS/mplayerxp.1 2010-01-18 15:55:23 UTC (rev 107)
@@ -210,7 +210,7 @@
.TP
modprobe dhahelper
.TP
-mplayerxp -vo xvidix -core.xp -video.bm -video.fs -video.zoom videoout.avi
+mplayerxp -vo xvidix -core.xp -video.bm -video.fs -video.aspect-ratio videoout.avi
.SS
.I Another ways to speedup playback:
In general, there are two ways to increase performance of playback -
@@ -666,13 +666,13 @@
-video.monitorpixelaspect=4:3 or 1.3333
.TP
-.B \-video.x\ <x>
-scale image to x width (if driver supports)
+.B \-video.width\ <x>
+scale image to width (if driver supports)
.TP
-.B \-video.y\ <y>
-scale image to y height (if driver supports)
+.B \-video.height\ <y>
+scale image to height (if driver supports)
.TP
-.B \-video.xy\ <factor>
+.B \-video.zoom\ <factor>
scale image by <factor>
.TP
@@ -717,7 +717,7 @@
try to change to a different video mode. dga2, x11 (XF86VidMode) and sdl
output drivers support it.
.TP
-.B \-video.zoom
+.B \-video.aspect-ratio
Keeps aspect ratio on the screen
.I [default]
.TP
@@ -809,7 +809,7 @@
.TP
.B \-video.sws\ <software\ scaler\ type>
this option sets the quality (and speed, respectively) of the software scaler,
-with the -zoom option. For example with x11 or other outputs which lack
+with the -video.aspect-ratio option. For example with x11 or other outputs which lack
hardware scaler. Possible settings are:
0 - fast bilinear (default)
Modified: mplayerxp/cfg-mplayer.h
===================================================================
--- mplayerxp/cfg-mplayer.h 2010-01-17 18:46:44 UTC (rev 106)
+++ mplayerxp/cfg-mplayer.h 2010-01-18 15:55:23 UTC (rev 107)
@@ -132,12 +132,12 @@
*/
static const config_t xpcore_config[]={
- {"xp", &enable_xp, CONF_TYPE_INT, CONF_RANGE, 0, 4, NULL, "starts MPlayerXP in multi-thread and multi-buffer XP mode"},
- {"noxp", &enable_xp, CONF_TYPE_FLAG, 0, 1, 0, NULL, "starts MPlayerXP in single-thread mode"},
+ {"xp", &enable_xp, CONF_TYPE_INT, CONF_RANGE, 0, 4, NULL, "starts MPlayerXP in multi-thread and multi-buffer XP mode"},
+ {"noxp", &enable_xp, CONF_TYPE_FLAG, 0, 1, 0, NULL, "starts MPlayerXP in single-thread mode"},
{"dump", &stream_dump, CONF_TYPE_STRING, 0, 0, 0, NULL, "specifies dump type and name for the dump of stream"},
- {"gomp", &enable_gomp, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables usage of OpenMP extensions"},
- {"nogomp", &enable_gomp, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables usage of OpenMP extensions"},
- {"da_buffs", &vo_da_buffs, CONF_TYPE_INT, CONF_RANGE, 4, 1024, NULL, "specifies number of buffers for decoding-ahead in XP mode"},
+ {"gomp", &enable_gomp, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables usage of OpenMP extensions"},
+ {"nogomp", &enable_gomp, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables usage of OpenMP extensions"},
+ {"da_buffs", &vo_da_buffs, CONF_TYPE_INT, CONF_RANGE, 4, 1024, NULL, "specifies number of buffers for decoding-ahead in XP mode"},
{"double", &vo_doublebuffering, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables double-buffering for single-thread decoding"},
{"nodouble", &vo_doublebuffering, CONF_TYPE_FLAG, 0, 1, 0, NULL, "enables single-buffer for single-thread decoding"},
{"cache", &stream_cache_size, CONF_TYPE_INT, CONF_RANGE, 4, 65536, NULL,"specifies amount of memory for precaching a file/URL"},
@@ -168,32 +168,32 @@
#if defined( ARCH_X86 ) || defined(ARCH_X86_64)
static const config_t cpu_config[]={
- {"mmx", &x86_mmx, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables using of MMX extensions of CPU"},
- {"nommx", &x86_mmx, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables using of MMX extensions of CPU"},
- {"mmx2", &x86_mmx2, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables using of MMX2 extensions of CPU"},
- {"nommx2", &x86_mmx2, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables using of MMX2 extensions of CPU"},
- {"3dnow", &x86_3dnow, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables using of 3DNow! extensions of CPU"},
- {"no3dnow", &x86_3dnow, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables using of 3DNow! extensions of CPU"},
- {"3dnow2", &x86_3dnow2, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables using of 3DNow-2! extensions of CPU"},
- {"no3dnow2", &x86_3dnow2, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables using of 3DNow-2! extensions of CPU"},
- {"sse", &x86_sse, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables using of SSE extensions of CPU"},
- {"nosse", &x86_sse, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables using of SSE extensions of CPU"},
- {"sse2", &x86_sse2, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables using of SSE2 extensions of CPU"},
- {"nosse2", &x86_sse2, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables using of SSE2 extensions of CPU"},
- {"sse3", &x86_sse3, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables using of SSE3 extensions of CPU"},
- {"nosse3", &x86_sse3, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables using of SSE3 extensions of CPU"},
- {"ssse3", &x86_ssse3, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables using of SSSE3 extensions of CPU"},
- {"nossse3", &x86_ssse3, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables using of SSSE3 extensions of CPU"},
- {"sse41", &x86_sse41, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables using of SSE41 extensions of CPU"},
- {"nosse41", &x86_sse41, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables using of SSE41 extensions of CPU"},
- {"sse42", &x86_sse42, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables using of SSE42 extensions of CPU"},
- {"nosse42", &x86_sse42, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables using of SSE42 extensions of CPU"},
- {"aes", &x86_aes, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables using of AES extensions of CPU"},
- {"noaes", &x86_aes, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables using of AES extensions of CPU"},
- {"avx", &x86_avx, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables using of AVX extensions of CPU"},
- {"noavx", &x86_avx, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables using of AVX extensions of CPU"},
- {"fma", &x86_fma, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables using of FMA extensions of CPU"},
- {"nofma", &x86_fma, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables using of FMA extensions of CPU"},
+ {"mmx", &x86_mmx, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables using of MMX extensions of CPU"},
+ {"nommx", &x86_mmx, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables using of MMX extensions of CPU"},
+ {"mmx2", &x86_mmx2, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables using of MMX2 extensions of CPU"},
+ {"nommx2", &x86_mmx2, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables using of MMX2 extensions of CPU"},
+ {"3dnow", &x86_3dnow, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables using of 3DNow! extensions of CPU"},
+ {"no3dnow", &x86_3dnow, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables using of 3DNow! extensions of CPU"},
+ {"3dnow2", &x86_3dnow2, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables using of 3DNow-2! extensions of CPU"},
+ {"no3dnow2", &x86_3dnow2, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables using of 3DNow-2! extensions of CPU"},
+ {"sse", &x86_sse, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables using of SSE extensions of CPU"},
+ {"nosse", &x86_sse, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables using of SSE extensions of CPU"},
+ {"sse2", &x86_sse2, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables using of SSE2 extensions of CPU"},
+ {"nosse2", &x86_sse2, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables using of SSE2 extensions of CPU"},
+ {"sse3", &x86_sse3, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables using of SSE3 extensions of CPU"},
+ {"nosse3", &x86_sse3, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables using of SSE3 extensions of CPU"},
+ {"ssse3", &x86_ssse3, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables using of SSSE3 extensions of CPU"},
+ {"nossse3", &x86_ssse3, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables using of SSSE3 extensions of CPU"},
+ {"sse41", &x86_sse41, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables using of SSE41 extensions of CPU"},
+ {"nosse41", &x86_sse41, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables using of SSE41 extensions of CPU"},
+ {"sse42", &x86_sse42, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables using of SSE42 extensions of CPU"},
+ {"nosse42", &x86_sse42, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables using of SSE42 extensions of CPU"},
+ {"aes", &x86_aes, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables using of AES extensions of CPU"},
+ {"noaes", &x86_aes, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables using of AES extensions of CPU"},
+ {"avx", &x86_avx, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables using of AVX extensions of CPU"},
+ {"noavx", &x86_avx, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables using of AVX extensions of CPU"},
+ {"fma", &x86_fma, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables using of FMA extensions of CPU"},
+ {"nofma", &x86_fma, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables using of FMA extensions of CPU"},
{NULL, NULL, 0, 0, 0, 0, NULL,NULL},
};
#endif
@@ -227,7 +227,6 @@
{"framedrop", &frame_dropping, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables frame-dropping on slow systems: decodes all video frames, but skips displaying some ones"},
/*UD*/ {"hardframedrop", &frame_dropping, CONF_TYPE_FLAG, 0, 0, 2, NULL, "enables hard frame-dropping on slow systems: skips displaying and decoding of some frames"},
{"noframedrop", &frame_dropping, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables frame dropping"},
-
{"pts", &av_sync_pts, CONF_TYPE_FLAG, 0, 0, 1, NULL, "use PTS-based method of A/V synchronization"},
{"nopts", &av_sync_pts, CONF_TYPE_FLAG, 0, 1, 0, NULL, "use BPS-based method of A/V synchronization"},
{"dap", &dapsync, CONF_TYPE_FLAG, 0, 0, 1, NULL, "use alternative method of A/V synchronization"},
@@ -266,7 +265,7 @@
{"nounicode", &sub_unicode, CONF_TYPE_FLAG, 0, 1, 0, NULL, "tells MPlayerXP to handle the subtitle file as non-UNICODE"},
{"utf8", &sub_utf8, CONF_TYPE_FLAG, 0, 0, 1, NULL, "tells MPlayerXP to handle the subtitle file as UTF8"},
{"noutf8", &sub_utf8, CONF_TYPE_FLAG, 0, 1, 0, NULL, "tells MPlayerXP to handle the subtitle file as non-UTF8"},
- {"pos",&sub_pos, CONF_TYPE_INT, CONF_RANGE, 0, 100, NULL, "specifies vertical shift of subtitles"},
+ {"pos",&sub_pos, CONF_TYPE_INT, CONF_RANGE, 0, 100, NULL, "specifies vertical shift of subtitles"},
#endif
{"cc", &subcc_enabled, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enable DVD Closed Caption (CC) subtitles"},
{"nocc", &subcc_enabled, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disable DVD Closed Caption (CC) subtitles"},
@@ -291,7 +290,7 @@
static const config_t audio_config[]={
{"on", &has_audio, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables audio-steam playback"},
{"off", &has_audio, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables audio-stream playback"},
- {"mixer", &oss_mixer_device, CONF_TYPE_STRING, 0, 0, 0, NULL, "select audio-mixer device"},
+ {"mixer", &oss_mixer_device, CONF_TYPE_STRING, 0, 0, 0, NULL, "select audio-mixer device"},
{"channels", &audio_output_channels, CONF_TYPE_INT, CONF_RANGE, 2, 8, NULL, "select number of audio output channels to be used"},
{"rate", &force_srate, CONF_TYPE_INT, CONF_RANGE, 1000, 8*48000, NULL, "specifies Hz for audio playback"},
{"lang", &audio_lang, CONF_TYPE_STRING, 0, 0, 0, NULL, "specifies language of DVD-audio stream as two-letter country code(s)"},
@@ -304,28 +303,28 @@
};
static const config_t video_config[]={
- {"x", &opt_screen_size_x, CONF_TYPE_INT, CONF_RANGE, 0, 4096, NULL, "scale output image to x width (if driver supports)"},
- {"y", &opt_screen_size_y, CONF_TYPE_INT, CONF_RANGE, 0, 4096, NULL, "scale output image to y height (if driver supports)"},
- {"xy", &screen_size_xy, CONF_TYPE_FLOAT, CONF_RANGE, 0, 4096, NULL, "scale output image by given factor"},
+ {"width", &opt_screen_size_x, CONF_TYPE_INT, CONF_RANGE, 0, 4096, NULL, "scale output image to width (if driver supports)"},
+ {"height", &opt_screen_size_y, CONF_TYPE_INT, CONF_RANGE, 0, 4096, NULL, "scale output image to height (if driver supports)"},
+ {"zoom", &screen_size_xy, CONF_TYPE_FLOAT, CONF_RANGE, 0, 4096, NULL, "scale output image by given factor"},
{"screenw", &vo_screenwidth, CONF_TYPE_INT, CONF_RANGE, 0, 4096, NULL, "specifies the horizontal resolution of the screen (if supported)"},
{"screenh", &vo_screenheight, CONF_TYPE_INT, CONF_RANGE, 0, 4096, NULL, "specifies the vertical resolution of the screen (if supported)"},
{"speed", &playbackspeed_factor, CONF_TYPE_FLOAT, CONF_RANGE, 0.01, 100.0, NULL, "sets playback speed factor"},
{"aspect", &movie_aspect, CONF_TYPE_FLOAT, CONF_RANGE, 0.2, 3.0, NULL, "sets aspect-ratio of movies (autodetect)"},
{"noaspect", &movie_aspect, CONF_TYPE_FLAG, 0, 0, 0, NULL, "unsets aspect-ratio of movies"},
+ {"aspect-ratio", &softzoom, CONF_TYPE_FLAG, 0, 0, 1, NULL, "keeps aspect-ratio of the movie during window resize"},
+ {"noaspect-ratio", &softzoom, CONF_TYPE_FLAG, 0, 1, 0, NULL, "render movie to the user-defined window's geometry"},
{"monitorpixelaspect", &monitor_pixel_aspect, CONF_TYPE_FLOAT, CONF_RANGE, 0.2, 9.0, NULL, "sets the aspect-ratio of a single pixel of TV screen"},
- {"vm", &vidmode, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables video-mode changing during playback"},
+ {"vm", &vidmode, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables video-mode changing during playback"},
{"novm", &vidmode, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables video-mode changing during playback"},
{"fs", &fullscreen, CONF_TYPE_FLAG, 0, 0, 1, NULL, "fullscreen playback"},
{"nofs", &fullscreen, CONF_TYPE_FLAG, 0, 1, 0, NULL, "windowed playback"},
{"fsmode", &vo_fsmode, CONF_TYPE_INT, CONF_RANGE, 0, 15, NULL, "enables workaround for some fullscreen related problems"},
- {"zoom", &softzoom, CONF_TYPE_FLAG, 0, 0, 1, NULL, "keeps aspect-ratio of the movie during window resize"},
- {"nozoom", &softzoom, CONF_TYPE_FLAG, 0, 1, 0, NULL, "render movie to the user-defined window's geometry"},
- {"flip", &flip, CONF_TYPE_FLAG, 0, -1, 1, NULL, "flip output image upside-down"},
- {"noflip", &flip, CONF_TYPE_FLAG, 0, -1, 0, NULL, "render output image as is"},
- {"bpp", &vo_dbpp, CONF_TYPE_INT, CONF_RANGE, 0, 32, NULL, "use different color depth than autodetect"},
- {"bm", &vo_use_bm, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables using of bus-mastering (if it available for given OS/videocard)"},
- {"bm2", &vo_use_bm, CONF_TYPE_FLAG, 0, 0, 2, NULL, "enables using of bus-mastering to store all decoded-ahead frames in video-memory"},
- {"nobm", &vo_use_bm, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables using of bus-mastering"},
+ {"flip", &flip, CONF_TYPE_FLAG, 0, -1, 1, NULL, "flip output image upside-down"},
+ {"noflip", &flip, CONF_TYPE_FLAG, 0, -1, 0, NULL, "render output image as is"},
+ {"bpp", &vo_dbpp, CONF_TYPE_INT, CONF_RANGE, 0, 32, NULL, "use different color depth than autodetect"},
+ {"bm", &vo_use_bm, CONF_TYPE_FLAG, 0, 0, 1, NULL, "enables using of bus-mastering (if it available for given OS/videocard)"},
+ {"bm2", &vo_use_bm, CONF_TYPE_FLAG, 0, 0, 2, NULL, "enables using of bus-mastering to store all decoded-ahead frames in video-memory"},
+ {"nobm", &vo_use_bm, CONF_TYPE_FLAG, 0, 1, 0, NULL, "disables using of bus-mastering"},
{"id", &video_id, CONF_TYPE_INT, CONF_RANGE, 0, 255, NULL, "selects video channel"},
{"pp", &npp_options, CONF_TYPE_STRING, 0, 0, 0, NULL, "specifies options of post-processing"},
{"sws", &sws_flags, CONF_TYPE_INT, 0, 0, 2, NULL, "specifies the quality of the software scaler"},
@@ -336,7 +335,7 @@
{"noxv", &sdl_noxv, CONF_TYPE_FLAG, 0, 0, 1, NULL, "disable XVideo hardware acceleration for SDL"},
{"forcexv", &sdl_forcexv, CONF_TYPE_FLAG, 0, 0, 1, NULL, "force XVideo hardware acceleration for SDL"},
{"forcegl", &sdl_forcegl, CONF_TYPE_FLAG, 0, 0, 1, NULL, "force OpenGL hardware acceleration for SDL"},
-#endif
+#endif
{"eq",&veq_config, CONF_TYPE_SUBCONFIG, 0, 0, 0, NULL, "Video-equalizer specific options"},
{NULL, NULL, 0, 0, 0, 0, NULL,NULL},
};
@@ -346,8 +345,8 @@
{"ss", &seek_to_sec, CONF_TYPE_STRING, CONF_MIN, 0, 0, NULL, "seek to given time position before playback"},
{"loop", &loop_times, CONF_TYPE_INT, CONF_RANGE, -1, 10000, NULL, "loops movie playback given number of times. 0 means forever"},
{"noloop", &loop_times, CONF_TYPE_FLAG, 0, 0, -1, NULL, "disable loop of playback"},
- {"shuffle",&shuffle_playback, CONF_TYPE_FLAG, 0, 0, 1, NULL, "play files in random order"},
- {"noshuffle",&shuffle_playback, CONF_TYPE_FLAG, 0, 1, 0, NULL, "play files in regular order"},
+ {"shuffle",&shuffle_playback, CONF_TYPE_FLAG, 0, 0, 1, NULL, "play files in random order"},
+ {"noshuffle",&shuffle_playback, CONF_TYPE_FLAG, 0, 1, 0, NULL, "play files in regular order"},
{"list", NULL, CONF_TYPE_STRING, 0, 0, 0, NULL, "specifies playlist (1 file/row or Winamp or ASX format)"},
{"frames", &play_n_frames, CONF_TYPE_INT, CONF_MIN, 0, 0, NULL, "play given number of frames and exit"},
{NULL, NULL, 0, 0, 0, 0, NULL,NULL},
Modified: mplayerxp/dec_ahead.h
===================================================================
--- mplayerxp/dec_ahead.h 2010-01-17 18:46:44 UTC (rev 106)
+++ mplayerxp/dec_ahead.h 2010-01-18 15:55:23 UTC (rev 107)
@@ -60,6 +60,8 @@
#define LOCK_VIDEO_DECODE() { MSG_D(DA_PREFIX"LOCK_VIDEO_DECODE\n"); pthread_mutex_lock(&video_decode_mutex); }
#define UNLOCK_VIDEO_DECODE() { MSG_D(DA_PREFIX"UNLOCK_VIDEO_DECODE\n"); pthread_mutex_unlock(&video_decode_mutex); }
+#define __MP_ATOMIC(OP) { pthread_mutex_t loc_mutex; pthread_mutex_lock(&loc_mutex); OP; pthread_mutex_unlock(&loc_mutex); }
+
typedef struct sh_video_attr
{
int eof; /* indicates last frame in stream */
Modified: mplayerxp/libmpcodecs/ad_mp3.c
===================================================================
--- mplayerxp/libmpcodecs/ad_mp3.c 2010-01-17 18:46:44 UTC (rev 106)
+++ mplayerxp/libmpcodecs/ad_mp3.c 2010-01-18 15:55:23 UTC (rev 107)
@@ -361,8 +361,10 @@
if(!((err==MPG123_OK)||(err==MPG123_NEED_MORE))) {
MSG_ERR("mpg123_read = %s done = %u minlen = %u\n",mpg123_plain_strerror(err),done,minlen);
}
- else
+ else {
+ MSG_DBG2("ad_mp3.decode: copy %u bytes from %p\n",done,outdata);
memcpy(buf,outdata,done);
+ }
if(err==MPG123_NEED_MORE) {
indata_size=ds_get_packet_r(sh->ds,&indata,pts);
if(indata_size<0) return 0;
Modified: mplayerxp/libmpcodecs/dec_video.c
===================================================================
--- mplayerxp/libmpcodecs/dec_video.c 2010-01-17 18:46:44 UTC (rev 106)
+++ mplayerxp/libmpcodecs/dec_video.c 2010-01-18 15:55:23 UTC (rev 107)
@@ -80,12 +80,10 @@
sh_video->inited=0;
}
-#ifdef _OPENMP
#define MPDEC_THREAD_COND (VF_FLAGS_THREADS|VF_FLAGS_SLICES)
static unsigned smp_num_cpus=1;
static unsigned use_vf_threads=0;
extern int enable_gomp;
-#endif
extern char *video_codec;
int init_video(sh_video_t *sh_video,const char* codecname,const char * vfm,int status){
@@ -207,12 +205,12 @@
for(j=0;j<num_slices;j+=smp_num_cpus) {
#pragma omp parallel for shared(vf) private(i)
for(i=j;i<smp_num_cpus;i++) {
- MSG_DBG2("Put slice[%u %u] in threads\n",ampi[i].y,ampi[i].h);
+ MSG_DBG2("parallel: dec_video.put_slice[%ux%u] %i %i %i %i\n",ampi[i].width,ampi[i].height,ampi[i].x,ampi[i].y,ampi[i].w,ampi[i].h);
vf->put_slice(vf,&i[i]);
}
}
for(;j<num_slices;j++) {
- MSG_DBG2("Put slice[%u %u] in threads\n",ampi[j].y,h_step);
+ MSG_DBG2("par_tail: dec_video.put_slice[%ux%u] %i %i %i %i\n",ampi[i].width,ampi[i].height,ampi[i].x,ampi[i].y,ampi[i].w,ampi[i].h);
vf->put_slice(vf,&i[j]);
}
}
@@ -221,12 +219,12 @@
{
/* execute slices instead of whole frame make faster multiple filters */
for(i=0;i<num_slices;i++) {
- MSG_DBG2("vf(%s) Put slice[%u %u] in threads\n",vf->info->name,ampi[i].y,ampi[i].h);
+ MSG_DBG2("dec_video.put_slice[%ux%u] %i %i %i %i\n",ampi[i].width,ampi[i].height,ampi[i].x,ampi[i].y,ampi[i].w,ampi[i].h);
vf->put_slice(vf,&i[i]);
}
}
} else {
- MSG_DBG2("Put whole frame\n");
+ MSG_DBG2("Put whole frame[%ux%u]\n",mpi->width,mpi->height);
vf->put_slice(vf,mpi);
}
}
Modified: mplayerxp/libmpcodecs/vd_ffmpeg.c
===================================================================
--- mplayerxp/libmpcodecs/vd_ffmpeg.c 2010-01-17 18:46:44 UTC (rev 106)
+++ mplayerxp/libmpcodecs/vd_ffmpeg.c 2010-01-18 15:55:23 UTC (rev 107)
@@ -5,6 +5,7 @@
#include <dlfcn.h> /* GLIBC specific. Exists under cygwin too! */
#include "mp_config.h"
+#include "../dec_ahead.h"
#ifdef HAVE_GOMP
#include <omp.h>
#endif
@@ -578,17 +579,18 @@
priv_t *vdff_ctx=sh->context;
mp_image_t *mpi;
if(vdff_ctx->use_dr1) { MSG_DBG2("Ignoring draw_slice due dr1\n"); return; } /* we may call vo_start_slice() here */
- mpi=mpcodecs_get_image(sh,MP_IMGTYPE_EXPORT, MP_IMGFLAG_ACCEPT_STRIDE|MP_IMGFLAG_DRAW_CALLBACK|MP_IMGFLAG_DIRECT,s->width,height);
+ mpi=mpcodecs_get_image(sh,MP_IMGTYPE_EXPORT, MP_IMGFLAG_ACCEPT_STRIDE|MP_IMGFLAG_DRAW_CALLBACK|MP_IMGFLAG_DIRECT,s->width,s->height);
mpi->stride[0]=src->linesize[0];
mpi->stride[1]=src->linesize[1];
mpi->stride[2]=src->linesize[2];
- mpi->planes[0] = src->base[0]+offset[0];
- mpi->planes[1] = src->base[1]+offset[1];
- mpi->planes[2] = src->base[2]+offset[2];
+ mpi->planes[0] = src->data[0];
+ mpi->planes[1] = src->data[1];
+ mpi->planes[2] = src->data[2];
mpi->w=s->width;
mpi->y=y;
mpi->h=height;
+ mpi->chroma_height = height >> mpi->chroma_y_shift;
/* provide info for pp */
mpi->qscale=(QP_STORE_T *)vdff_ctx->lavc_picture->qscale_table;
mpi->qstride=vdff_ctx->lavc_picture->qstride;
@@ -607,14 +609,10 @@
mpi->stride[2]=mpi->stride[1];
mpi->stride[1]=ls;
}
- MSG_DBG2("ff_draw_callback %i %i %i %i\n",mpi->x,mpi->y,mpi->w,mpi->h);
- pthread_mutex_lock(&sh->mutex);
- sh->active_slices++;
- pthread_mutex_unlock(&sh->mutex);
+ MSG_DBG2("ff_draw_callback[%ux%u] %i %i %i %i\n",mpi->width,mpi->height,mpi->x,mpi->y,mpi->w,mpi->h);
+ __MP_ATOMIC(sh->active_slices++);
mpcodecs_draw_slice (sh, mpi);
- pthread_mutex_lock(&sh->mutex);
- sh->active_slices--;
- pthread_mutex_unlock(&sh->mutex);
+ __MP_ATOMIC(sh->active_slices--);
}
/* copypaste from demux_real.c - it should match to get it working!*/
Modified: mplayerxp/libvo/aclib.c
===================================================================
--- mplayerxp/libvo/aclib.c 2010-01-17 18:46:44 UTC (rev 106)
+++ mplayerxp/libvo/aclib.c 2010-01-18 15:55:23 UTC (rev 107)
@@ -8,110 +8,81 @@
#if defined(USE_FASTMEMCPY)
#include "fastmemcpy.h"
#include "../cpudetect.h"
-/*
- aclib - advanced C library ;)
- This file contains functions which improve and expand standard C-library
- see aclib_template.c ... this file only contains runtime cpu detection and config options stuff
- runtime cpu detection by michael niedermayer (mic...@gm...) is under GPL
-*/
-#if defined( CAN_COMPILE_MMX ) && defined (ARCH_X86)
#define BLOCK_SIZE 4096
#define CONFUSION_FACTOR 0
-//Feel free to fine-tune the above 2, it might be possible to get some speedup with them :)
-//#define STATISTICS
+/* generic version */
+#undef OPTIMIZE_AVX
+#undef OPTIMIZE_SSE4
+#undef OPTIMIZE_SSSE3
+#undef OPTIMIZE_SSE3
+#undef OPTIMIZE_SSE2
+#undef OPTIMIZE_SSE
+#undef OPTIMIZE_MMX2
+#undef OPTIMIZE_MMX
-#if defined( ARCH_X86 )
-#define CAN_COMPILE_X86_ASM
-#endif
-
-//Note: we have MMX, MMX2, 3DNOW version there is no 3DNOW+MMX2 one
-//Plain C versions
-//#if !defined (HAVE_MMX) || defined (RUNTIME_CPUDETECT)
-//#define COMPILE_C
-//#endif
-
-#ifdef CAN_COMPILE_X86_ASM
-
-#undef HAVE_MMX
-#undef HAVE_MMX2
-#undef HAVE_3DNOW
-#undef HAVE_SSE
-
-//MMX versions
-#ifdef CAN_COMPILE_MMX
-#undef RENAME
+#ifndef __x86_64__
+#ifdef __MMX__
+#define OPTIMIZE_MMX
#undef CL_SIZE
#define CL_SIZE 32
-#define HAVE_MMX
-#undef HAVE_MMX2
-#undef HAVE_3DNOW
+#undef RENAME
#define RENAME(a) a ## _MMX
#include "aclib_template.c"
#endif
-
-//MMX2 versions 32-byte cache-line size
-#ifdef CAN_COMPILE_MMX2
-#undef RENAME
+#ifdef __MMX2__
+#define OPTIMIZE_MMX2
#undef CL_SIZE
#define CL_SIZE 32
-#define HAVE_MMX
-#define HAVE_MMX2
-#undef HAVE_3DNOW
+#undef RENAME
#define RENAME(a) a ## _MMX2_CL32
#include "aclib_template.c"
#endif
-
-//MMX2 versions 64-byte cache-line size
-#ifdef CAN_COMPILE_MMX2
-#undef RENAME
+#ifdef __MMX2__
+#define OPTIMIZE_MMX2
#undef CL_SIZE
#define CL_SIZE 64
-#define HAVE_MMX
-#define HAVE_MMX2
-#undef HAVE_3DNOW
+#undef RENAME
#define RENAME(a) a ## _MMX2_CL64
#include "aclib_template.c"
#endif
-
-//MMX2 versions 128-byte cache-line size
-#ifdef CAN_COMPILE_MMX2
-#undef RENAME
+#ifdef __MMX2__
+#define OPTIMIZE_MMX2
#undef CL_SIZE
#define CL_SIZE 128
-#define HAVE_MMX
-#define HAVE_MMX2
-#undef HAVE_3DNOW
+#undef RENAME
#define RENAME(a) a ## _MMX2_CL128
#include "aclib_template.c"
#endif
-
-//3DNOW versions all K6 have 32-bit cache-line size
-#ifdef CAN_COMPILE_3DNOW
+#endif // __x86_64__
+#ifdef __SSE2__
+#define OPTIMIZE_SSE2
#undef RENAME
#undef CL_SIZE
-#define CL_SIZE 32
-#define HAVE_MMX
-#undef HAVE_MMX2
-#define HAVE_3DNOW
-#define RENAME(a) a ## _3DNow
+#define CL_SIZE 128
+#define RENAME(a) a ## _SSE2
#include "aclib_template.c"
#endif
-#endif // CAN_COMPILE_X86_ASM
+/*
+ aclib - advanced C library ;)
+ This file contains functions which improve and expand standard C-library
+ see aclib_template.c ... this file only contains runtime cpu detection and config options stuff
+ runtime cpu detection by michael niedermayer (mic...@gm...) is under GPL
+*/
-#elif defined( ARCH_X86_64 )
-#define RENAME(a) a ## _x86_64
-#include "aclib_x86_64.h"
-#endif
-
static void * init_fast_memcpy(void * to, const void * from, size_t len)
{
-#if defined( ARCH_X86_64 ) && defined( USE_FASTMEMCPY )
- fast_memcpy_ptr = fast_memcpy_x86_64;
-#elif defined( CAN_COMPILE_X86_ASM )
- // ordered per speed fasterst first
-#ifdef CAN_COMPILE_MMX2
+#ifdef __SSE2__
+ if(gCpuCaps.hasSSE2)
+ {
+ MSG_V("Using SSE2 optimized memcpy\n");
+ fast_memcpy_ptr = fast_memcpy_SSE2;
+ }
+ else
+#endif
+#ifndef __x86_64__
+#ifdef __MMX2__
if(gCpuCaps.hasMMX2)
{
MSG_V("Using MMX2 optimized memcpy\n");
@@ -123,14 +94,6 @@
}
else
#endif
-#ifdef CAN_COMPILE_3DNOW
- if(gCpuCaps.has3DNow)
- {
- MSG_V("Using 3DNow optimized memcpy\n");
- fast_memcpy_ptr = fast_memcpy_3DNow;
- }
- else
-#endif
#ifdef CAN_COMPILE_MMX
if(gCpuCaps.hasMMX)
{
@@ -139,60 +102,54 @@
}
else
#endif
-#else
+#endif
{
MSG_V("Using generic memcpy\n");
fast_memcpy_ptr = memcpy; /* prior to mmx we use the standart memcpy */
}
-#endif
return (*fast_memcpy_ptr)(to,from,len);
}
-static void * init_mem2agpcpy(void * to, const void * from, size_t len)
+static void * init_stream_copy(void * to, const void * from, size_t len)
{
-#if defined( ARCH_X86_64 ) && defined( USE_FASTMEMCPY )
- mem2agpcpy_ptr = mem2agpcpy_x86_64;
-#elif defined ( CAN_COMPILE_X86_ASM )
- // ordered per speed fasterst first
-#ifdef CAN_COMPILE_MMX2
+#ifdef __SSE2__
+ if(gCpuCaps.hasSSE2)
+ {
+ MSG_V("Using SSE2 optimized agpcpy\n");
+ fast_stream_copy_ptr = fast_stream_copy_SSE2;
+ }
+#endif
+#ifndef __x86_64__
+#ifdef __MMX2__
if(gCpuCaps.hasMMX2)
{
MSG_V("Using MMX2 optimized agpcpy\n");
- if(gCpuCaps.cl_size >= 128) mem2agpcpy_ptr = mem2agpcpy_MMX2_CL128;
+ if(gCpuCaps.cl_size >= 128) fast_stream_copy_ptr = fast_stream_copy_MMX2_CL128;
else
- if(gCpuCaps.cl_size == 64) mem2agpcpy_ptr = mem2agpcpy_MMX2_CL64;
+ if(gCpuCaps.cl_size == 64) fast_stream_copy_ptr = fast_stream_copy_MMX2_CL64;
else
- mem2agpcpy_ptr = mem2agpcpy_MMX2_CL32;
+ fast_stream_copy_ptr = fast_stream_copy_MMX2_CL32;
}
else
#endif
-#ifdef CAN_COMPILE_3DNOW
- if(gCpuCaps.has3DNow)
- {
- MSG_V("Using 3DNow optimized agpcpy\n");
- mem2agpcpy_ptr = mem2agpcpy_3DNow;
- }
- else
-#endif
-#ifdef CAN_COMPILE_MMX
+#ifdef __MMX__
if(gCpuCaps.hasMMX)
{
MSG_V("Using MMX optimized agpcpy\n");
- mem2agpcpy_ptr = mem2agpcpy_MMX;
+ fast_stream_copy_ptr = fast_stream_copy_MMX;
}
else
#endif
-#else
+#endif
{
MSG_V("Using generic optimized agpcpy\n");
- mem2agpcpy_ptr = memcpy; /* prior to mmx we use the standart memcpy */
+ fast_stream_copy_ptr = memcpy; /* prior to mmx we use the standart memcpy */
}
-#endif
- return (*mem2agpcpy_ptr)(to,from,len);
+ return (*fast_stream_copy_ptr)(to,from,len);
}
void *(*fast_memcpy_ptr)(void * to, const void * from, size_t len) = init_fast_memcpy;
-void *(*mem2agpcpy_ptr)(void * to, const void * from, size_t len) = init_mem2agpcpy;
+void *(*fast_stream_copy_ptr)(void * to, const void * from, size_t len) = init_stream_copy;
#endif /* use fastmemcpy */
Modified: mplayerxp/libvo/aclib_template.c
===================================================================
--- mplayerxp/libvo/aclib_template.c 2010-01-17 18:46:44 UTC (rev 106)
+++ mplayerxp/libvo/aclib_template.c 2010-01-18 15:55:23 UTC (rev 107)
@@ -2,102 +2,70 @@
aclib - advanced C library ;)
This file contains functions which improve and expand standard C-library
*/
+#include "pvector/pvector.h"
-#ifndef HAVE_SSE2
-/*
- P3 processor has only one SSE decoder so can execute only 1 sse insn per
- cpu clock, but it has 3 mmx decoders (include load/store unit)
- and executes 3 mmx insns per cpu clock.
- P4 processor has some chances, but after reading:
- http://www.emulators.com/pentium4.htm
- I have doubts. Anyway SSE2 version of this code can be written better.
-*/
-#undef HAVE_SSE
-#endif
-
-
-/*
- This part of code was taken by me from Linux-2.4.3 and slightly modified
-for MMX, MMX2, SSE instruction set. I have done it since linux uses page aligned
-blocks but mplayer uses weakly ordered data and original sources can not
-speedup them. Only using PREFETCHNTA and MOVNTQ together have effect!
-
->From IA-32 Intel Architecture Software Developer's Manual Volume 1,
-
-Order Number 245470:
-"10.4.6. Cacheability Control, Prefetch, and Memory Ordering Instructions"
-
-Data referenced by a program can be temporal (data will be used again) or
-non-temporal (data will be referenced once and not reused in the immediate
-future). To make efficient use of the processor's caches, it is generally
-desirable to cache temporal data and not cache non-temporal data. Overloading
-the processor's caches with non-temporal data is sometimes referred to as
-"polluting the caches".
-The non-temporal data is written to memory with Write-Combining semantics.
-
-The PREFETCHh instructions permits a program to load data into the processor
-at a suggested cache level, so that it is closer to the processors load and
-store unit when it is needed. If the data is already present in a level of
-the cache hierarchy that is closer to the processor, the PREFETCHh instruction
-will not result in any data movement.
-But we should you PREFETCHNTA: Non-temporal data fetch data into location
-close to the processor, minimizing cache pollution.
-
-The MOVNTQ (store quadword using non-temporal hint) instruction stores
-packed integer data from an MMX register to memory, using a non-temporal hint.
-The MOVNTPS (store packed single-precision floating-point values using
-non-temporal hint) instruction stores packed floating-point data from an
-XMM register to memory, using a non-temporal hint.
-
-The SFENCE (Store Fence) instruction controls write ordering by creating a
-fence for memory store operations. This instruction guarantees that the results
-of every store instruction that precedes the store fence in program order is
-globally visible before any store instruction that follows the fence. The
-SFENCE instruction provides an efficient way of ensuring ordering between
-procedures that produce weakly-ordered data and procedures that consume that
-data.
-
-If you have questions please contact with me: Nickols_K <nic...@ma...>.
-*/
-
-/* 3dnow memcpy support from kernel 2.4.2
- by Pontscho/fresh!mindworkz */
-
-
/* for small memory blocks (<256 bytes) this version is faster */
+#ifdef __x86_64__
#define small_memcpy(to,from,n)\
{\
+register unsigned long int siz;\
register unsigned long int dummy;\
+ siz=n&0x7; n>>=3;\
+ if(siz)\
__asm__ __volatile__(\
"rep; movsb"\
:"=&D"(to), "=&S"(from), "=&c"(dummy)\
/* It's most portable way to notify compiler */\
/* that edi, esi and ecx are clobbered in asm block. */\
/* Thanks to A'rpi for hint!!! */\
+ :"0" (to), "1" (from),"2" (siz)\
+ : "memory","cc");\
+ if(n)\
+__asm__ __volatile__(\
+ "rep; movsq"\
+ :"=&D"(to), "=&S"(from), "=&c"(dummy)\
+/* It's most portable way to notify compiler */\
+/* that edi, esi and ecx are clobbered in asm block. */\
+/* Thanks to A'rpi for hint!!! */\
:"0" (to), "1" (from),"2" (n)\
: "memory","cc");\
}
+#else
+#define small_memcpy(to,from,n)\
+{\
+register unsigned long int dummy;\
+__asm__ __volatile__(\
+ "rep; movsb"\
+ :"=&D"(to), "=&S"(from), "=&c"(dummy)\
+/* It's most portable way to notify compiler */\
+/* that edi, esi and ecx are clobbered in asm block. */\
+/* Thanks to A'rpi for hint!!! */\
+ :"0" (to), "1" (from),"2" (n)\
+ : "memory","cc");\
+}
+#endif
-#include "../mmx_defs.h"
#undef MMREG_SIZE
-#ifdef HAVE_SSE
+#ifdef OPTIMIZE_SSE2
#define MMREG_SIZE 16
#else
#define MMREG_SIZE 64 //8
#endif
#undef MIN_LEN
-#ifdef HAVE_MMX1
+#ifndef OPTIMIZE_MMX2
#define MIN_LEN 0x800 /* 2K blocks */
#else
#define MIN_LEN 0x40 /* 64-byte blocks */
#endif
-static inline void * RENAME(fast_memcpy)(void * to, const void * from, size_t len)
+static inline void * RENAME(fast_memory_copy)(void * to, const void * from, size_t len,int final)
{
void *retval;
const unsigned char *cfrom=from;
unsigned char *tto=to;
+ const unsigned ivec_block_size = 8*__IVEC_SIZE;
+ __ivec iarr[8];
size_t i;
retval = to;
if(!len) return retval;
@@ -114,45 +82,41 @@
MSG_V("freq < %8d %4d\n", 1<<i, freq[i]);
}
#endif
-#ifndef HAVE_MMX1
- /* PREFETCH has effect even for MOVSB instruction ;) */
- __asm__ __volatile__ (
- PREFETCH" (%0)\n"
+
+ _ivec_prefetch(cfrom);
#if CL_SIZE == 32
- PREFETCH" 32(%0)\n"
+ _ivec_prefetch(&cfrom[32]);
#endif
#if CL_SIZE < 128
- PREFETCH" 64(%0)\n"
+ _ivec_prefetch(&cfrom[64]);
#endif
#if CL_SIZE == 32
- PREFETCH" 96(%0)\n"
+ _ivec_prefetch(&cfrom[96]);
#endif
- PREFETCH" 128(%0)\n"
+ _ivec_prefetch(&cfrom[128]);
#if CL_SIZE == 32
- PREFETCH" 160(%0)\n"
+ _ivec_prefetch(&cfrom[160]);
#endif
#if CL_SIZE < 128
- PREFETCH" 192(%0)\n"
+ _ivec_prefetch(&cfrom[192]);
#endif
#if CL_SIZE == 32
- PREFETCH" 224(%0)\n"
+ _ivec_prefetch(&cfrom[224]);
#endif
- PREFETCH" 256(%0)\n"
- : : "r" (cfrom) );
-#endif
- if(len >= MIN_LEN)
- {
- register unsigned long int delta;
- /* Align destinition to cache-line size -boundary */
- delta = ((unsigned long int)tto)&(CL_SIZE-1);
- if(delta)
- {
+ _ivec_prefetch(&cfrom[256]);
+
+ if(len >= MIN_LEN)
+ {
+ register unsigned long int delta;
+ /* Align destinition to cache-line size -boundary */
+ delta = ((unsigned long int)tto)&(CL_SIZE-1);
+ if(delta) {
delta=MMREG_SIZE-delta;
len -= delta;
small_memcpy(tto, cfrom, delta);
- }
- i = len >> 6; /* len/64 */
- len&=63;
+ }
+ i = len/ivec_block_size;
+ len&=(ivec_block_size-1);
/*
This algorithm is top effective when the code consequently
reads and writes blocks which have size of cache line.
@@ -162,340 +126,73 @@
perform reading and writing to be multiple to a number of
processor's decoders, but it's not always possible.
*/
-#ifdef HAVE_SSE /* Only P3 (may be Cyrix3) */
- if(((unsigned long)cfrom) & 15)
- /* if SRC is misaligned */
for(; i>0; i--)
{
- __asm__ __volatile__ (
- PREFETCH" 320(%0)\n"
+ _ivec_prefetch(&cfrom[320]);
#if CL_SIZE == 32
- PREFETCH" 352(%0)\n"
+ _ivec_prefetch(&cfrom[352]);
#endif
- "movups (%0), %%xmm0\n"
- "movups 16(%0), %%xmm1\n"
- "movups 32(%0), %%xmm2\n"
- "movups 48(%0), %%xmm3\n"
- "movntps %%xmm0, (%1)\n"
- "movntps %%xmm1, 16(%1)\n"
- "movntps %%xmm2, 32(%1)\n"
- "movntps %%xmm3, 48(%1)\n"
- :: "r" (cfrom), "r" (tto):
- "memory"
-#ifdef SSE_CLOBBERED
- ,SSE_CLOBBERED
-#endif
- );
- cfrom+=64;
- tto+=64;
+ if(((unsigned long)cfrom) & 15) {
+ /* if SRC is misaligned */
+ iarr[0] = _ivec_loadu(&cfrom[__IVEC_SIZE*0]);
+ iarr[1] = _ivec_loadu(&cfrom[__IVEC_SIZE*1]);
+ iarr[2] = _ivec_loadu(&cfrom[__IVEC_SIZE*2]);
+ iarr[3] = _ivec_loadu(&cfrom[__IVEC_SIZE*3]);
+ iarr[4] = _ivec_loadu(&cfrom[__IVEC_SIZE*4]);
+ iarr[5] = _ivec_loadu(&cfrom[__IVEC_SIZE*5]);
+ iarr[6] = _ivec_loadu(&cfrom[__IVEC_SIZE*6]);
+ iarr[7] = _ivec_loadu(&cfrom[__IVEC_SIZE*7]);
+ } else {
+ iarr[0] = _ivec_loada(&cfrom[__IVEC_SIZE*0]);
+ iarr[1] = _ivec_loada(&cfrom[__IVEC_SIZE*1]);
+ iarr[2] = _ivec_loada(&cfrom[__IVEC_SIZE*2]);
+ iarr[3] = _ivec_loada(&cfrom[__IVEC_SIZE*3]);
+ iarr[4] = _ivec_loada(&cfrom[__IVEC_SIZE*4]);
+ iarr[5] = _ivec_loada(&cfrom[__IVEC_SIZE*5]);
+ iarr[6] = _ivec_loada(&cfrom[__IVEC_SIZE*6]);
+ iarr[7] = _ivec_loada(&cfrom[__IVEC_SIZE*7]);
+ }
+ if(final) {
+ _ivec_stream(&tto[__IVEC_SIZE*0],iarr[0]);
+ _ivec_stream(&tto[__IVEC_SIZE*1],iarr[1]);
+ _ivec_stream(&tto[__IVEC_SIZE*2],iarr[2]);
+ _ivec_stream(&tto[__IVEC_SIZE*3],iarr[3]);
+ _ivec_stream(&tto[__IVEC_SIZE*4],iarr[4]);
+ _ivec_stream(&tto[__IVEC_SIZE*5],iarr[5]);
+ _ivec_stream(&tto[__IVEC_SIZE*6],iarr[6]);
+ _ivec_stream(&tto[__IVEC_SIZE*7],iarr[7]);
+ } else {
+ _ivec_storea(&tto[__IVEC_SIZE*0],iarr[0]);
+ _ivec_storea(&tto[__IVEC_SIZE*1],iarr[1]);
+ _ivec_storea(&tto[__IVEC_SIZE*2],iarr[2]);
+ _ivec_storea(&tto[__IVEC_SIZE*3],iarr[3]);
+ _ivec_storea(&tto[__IVEC_SIZE*4],iarr[4]);
+ _ivec_storea(&tto[__IVEC_SIZE*5],iarr[5]);
+ _ivec_storea(&tto[__IVEC_SIZE*6],iarr[6]);
+ _ivec_storea(&tto[__IVEC_SIZE*7],iarr[7]);
+ }
+ cfrom+=ivec_block_size;
+ tto+=ivec_block_size;
}
- else
- /*
- Only if SRC is aligned on 16-byte boundary.
- It allows to use movaps instead of movups, which required data
- to be aligned or a general-protection exception (#GP) is generated.
- */
- for(; i>0; i--)
- {
- __asm__ __volatile__ (
- PREFETCH" 320(%0)\n"
-#if CL_SIZE == 32
- PREFETCH" 352(%0)\n"
-#endif
- "movaps (%0), %%xmm0\n"
- "movaps 16(%0), %%xmm1\n"
- "movaps 32(%0), %%xmm2\n"
- "movaps 48(%0), %%xmm3\n"
- "movntps %%xmm0, (%1)\n"
- "movntps %%xmm1, 16(%1)\n"
- "movntps %%xmm2, 32(%1)\n"
- "movntps %%xmm3, 48(%1)\n"
- :: "r" (cfrom), "r" (tto)
- :"memory"
-#ifdef SSE_CLOBBERED
- ,SSE_CLOBBERED
-#endif
- );
- cfrom+=64;
- tto+=64;
- }
-#else
- // Align destination at BLOCK_SIZE boundary
- for(; ((int)tto & (BLOCK_SIZE-1)) && i>0; i--)
- {
- __asm__ __volatile__ (
-#ifndef HAVE_MMX1
- PREFETCH" 320(%0)\n"
-#if CL_SIZE == 32
- PREFETCH" 352(%0)\n"
-#endif
-#endif
- "movq (%0), %%mm0\n"
- "movq 8(%0), %%mm1\n"
- "movq 16(%0), %%mm2\n"
- "movq 24(%0), %%mm3\n"
- "movq 32(%0), %%mm4\n"
- "movq 40(%0), %%mm5\n"
- "movq 48(%0), %%mm6\n"
- "movq 56(%0), %%mm7\n"
- MOVNTQ" %%mm0, (%1)\n"
- MOVNTQ" %%mm1, 8(%1)\n"
- MOVNTQ" %%mm2, 16(%1)\n"
- MOVNTQ" %%mm3, 24(%1)\n"
- MOVNTQ" %%mm4, 32(%1)\n"
- MOVNTQ" %%mm5, 40(%1)\n"
- MOVNTQ" %%mm6, 48(%1)\n"
- MOVNTQ" %%mm7, 56(%1)\n"
- :: "r" (cfrom), "r" (tto)
- : "memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
- cfrom+=64;
- tto+=64;
- }
-
- // Pure Assembly cuz gcc is a bit unpredictable ;)
- if(i>=BLOCK_SIZE/64)
- asm volatile(
- "xorl %%"REG_a", %%"REG_a" \n\t"
- ".balign 16 \n\t"
- "1: \n\t"
- "movl (%0, %%"REG_a"), %%"REG_b" \n\t"
-#if CL_SIZE == 32
- "movl 32(%0, %%"REG_a"), %%"REG_b" \n\t"
-#endif
-#if CL_SIZE < 128
- "movl 64(%0, %%"REG_a"), %%"REG_b" \n\t"
-#endif
-#if CL_SIZE == 32
- "movl 96(%0, %%"REG_a"), %%"REG_b" \n\t"
-#endif
- "addl $128, %%"REG_a" \n\t"
- "cmpl %3, %%"REG_a" \n\t"
- " jb 1b \n\t"
-
- "xorl %%"REG_a", %%"REG_a" \n\t"
-
- ".balign 16 \n\t"
- "2: \n\t"
- "movq (%0, %%"REG_a"), %%mm0\n"
- "movq 8(%0, %%"REG_a"), %%mm1\n"
- "movq 16(%0, %%"REG_a"), %%mm2\n"
- "movq 24(%0, %%"REG_a"), %%mm3\n"
- "movq 32(%0, %%"REG_a"), %%mm4\n"
- "movq 40(%0, %%"REG_a"), %%mm5\n"
- "movq 48(%0, %%"REG_a"), %%mm6\n"
- "movq 56(%0, %%"REG_a"), %%mm7\n"
- MOVNTQ" %%mm0, (%1, %%"REG_a")\n"
- MOVNTQ" %%mm1, 8(%1, %%"REG_a")\n"
- MOVNTQ" %%mm2, 16(%1, %%"REG_a")\n"
- MOVNTQ" %%mm3, 24(%1, %%"REG_a")\n"
- MOVNTQ" %%mm4, 32(%1, %%"REG_a")\n"
- MOVNTQ" %%mm5, 40(%1, %%"REG_a")\n"
- MOVNTQ" %%mm6, 48(%1, %%"REG_a")\n"
- MOVNTQ" %%mm7, 56(%1, %%"REG_a")\n"
- "addl $64, %%"REG_a" \n\t"
- "cmpl %3, %%"REG_a" \n\t"
- "jb 2b \n\t"
-
-#if CONFUSION_FACTOR > 0
- // a few percent speedup on out of order executing CPUs
- "movl %5, %%"REG_a" \n\t"
- "2: \n\t"
- "movl (%0), %%"REG_b" \n\t"
- "movl (%0), %%"REG_b" \n\t"
- "movl (%0), %%"REG_b" \n\t"
- "movl (%0), %%"REG_b" \n\t"
- "decl %%"REG_a" \n\t"
- " jnz 2b \n\t"
-#endif
-
- "xorl %%"REG_a", %%"REG_a" \n\t"
- "addl %3, %0 \n\t"
- "addl %3, %1 \n\t"
- "subl %4, %2 \n\t"
- "cmpl %4, %2 \n\t"
- " jae 1b \n\t"
- : "+r" (cfrom), "+r" (tto), "+r" (i)
- : "r" (BLOCK_SIZE), "i" (BLOCK_SIZE/64), "i" (CONFUSION_FACTOR)
- : "%"REG_a, "%"REG_b, "memory", "cc"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
-
- for(; i>0; i--)
- {
- __asm__ __volatile__ (
-#ifndef HAVE_MMX1
- PREFETCH" 320(%0)\n"
-#if CL_SIZE == 32
- PREFETCH" 352(%0)\n"
-#endif
-#endif
- "movq (%0), %%mm0\n"
- "movq 8(%0), %%mm1\n"
- "movq 16(%0), %%mm2\n"
- "movq 24(%0), %%mm3\n"
- "movq 32(%0), %%mm4\n"
- "movq 40(%0), %%mm5\n"
- "movq 48(%0), %%mm6\n"
- "movq 56(%0), %%mm7\n"
- MOVNTQ" %%mm0, (%1)\n"
- MOVNTQ" %%mm1, 8(%1)\n"
- MOVNTQ" %%mm2, 16(%1)\n"
- MOVNTQ" %%mm3, 24(%1)\n"
- MOVNTQ" %%mm4, 32(%1)\n"
- MOVNTQ" %%mm5, 40(%1)\n"
- MOVNTQ" %%mm6, 48(%1)\n"
- MOVNTQ" %%mm7, 56(%1)\n"
- :: "r" (cfrom), "r" (tto)
- : "memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
- cfrom+=64;
- tto+=64;
- }
-
-#endif /* Have SSE */
-#ifdef HAVE_MMX2
- /* since movntq is weakly-ordered, a "sfence"
- * is needed to become ordered again. */
- __asm__ __volatile__ ("sfence":::"memory");
-#endif
-#ifndef HAVE_SSE
- /* enables to use FPU */
- __asm__ __volatile__ (EMMS::
- :"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
-#endif
- }
- /*
- * Now do the tail of the block
- */
- if(len) small_memcpy(tto, cfrom, len);
- return retval;
+ _ivec_sfence();
+ _ivec_empty();
+ }
+ /*
+ * Now do the tail of the block
+ */
+ if(len) small_memcpy(tto, cfrom, len);
+ return retval;
}
/**
* special copy routine for mem -> agp/pci copy (based upon fast_memcpy)
*/
-static inline void * RENAME(mem2agpcpy)(void * to, const void * from, size_t len)
+static inline void * RENAME(fast_memcpy)(void * to, const void * from, size_t len)
{
- void *retval;
- const unsigned char *cfrom=from;
- unsigned char *tto=to;
- size_t i;
- retval = to;
- if(!len) return retval;
-#ifdef STATISTICS
- {
- static int freq[33];
- static int t=0;
- int i;
- for(i=0; len>(1<<i); i++);
- freq[i]++;
- t++;
- if(1024*1024*1024 % t == 0)
- for(i=0; i<32; i++)
- MSG_V("mem2agp freq < %8d %4d\n", 1<<i, freq[i]);
- }
-#endif
- if(len >= MIN_LEN)
- {
- register unsigned long int delta;
- /* Align destinition to cache-line size -boundary */
- delta = ((unsigned long int)tto)&(CL_SIZE-1);
- if(delta)
- {
- delta=8-delta;
- len -= delta;
- small_memcpy(tto, cfrom, delta);
- }
- i = len >> 6; /* len/64 */
- len &= 63;
- /*
- This algorithm is top effective when the code consequently
- reads and writes blocks which have size of cache line.
- Size of cache line is processor-dependent.
- It will, however, be a minimum of 32 bytes on any processors.
- It would be better to have a number of instructions which
- perform reading and writing to be multiple to a number of
- processor's decoders, but it's not always possible.
- */
- for(; i>0; i--)
- {
- __asm__ __volatile__ (
- PREFETCH" 320(%0)\n"
-#if CL_SIZE == 32
- PREFETCH" 352(%0)\n"
-#endif
- "movq (%0), %%mm0\n"
- "movq 8(%0), %%mm1\n"
- "movq 16(%0), %%mm2\n"
- "movq 24(%0), %%mm3\n"
- "movq 32(%0), %%mm4\n"
- "movq 40(%0), %%mm5\n"
- "movq 48(%0), %%mm6\n"
- "movq 56(%0), %%mm7\n"
- MOVNTQ" %%mm0, (%1)\n"
- MOVNTQ" %%mm1, 8(%1)\n"
- MOVNTQ" %%mm2, 16(%1)\n"
- MOVNTQ" %%mm3, 24(%1)\n"
- MOVNTQ" %%mm4, 32(%1)\n"
- MOVNTQ" %%mm5, 40(%1)\n"
- MOVNTQ" %%mm6, 48(%1)\n"
- MOVNTQ" %%mm7, 56(%1)\n"
- :: "r" (cfrom), "r" (tto)
- : "memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
- cfrom+=64;
- tto+=64;
- }
-#ifdef HAVE_MMX2
- /* since movntq is weakly-ordered, a "sfence"
- * is needed to become ordered again. */
- __asm__ __volatile__ ("sfence":::"memory");
-#endif
- /* enables to use FPU */
- __asm__ __volatile__ (EMMS::
- :"memory"
-#ifdef FPU_CLOBBERED
- ,FPU_CLOBBERED
-#endif
-#ifdef MMX_CLOBBERED
- ,MMX_CLOBBERED
-#endif
- );
- }
- /*
- * Now do the tail of the block
- */
- if(len) small_memcpy(tto, cfrom, len);
- return retval;
+ return RENAME(fast_memory_copy)(to,from,len,0);
}
+
+static inline void * RENAME(fast_stream_copy)(void * to, const void * from, size_t len)
+{
+ return RENAME(fast_memory_copy)(to,from,len,1);
+}
Deleted: mplayerxp/libvo/aclib_x86_64.h
===================================================================
--- mplayerxp/libvo/aclib_x86_64.h 2010-01-17 18:46:44 UTC (rev 106)
+++ mplayerxp/libvo/aclib_x86_64.h 2010-01-18 15:55:23 UTC (rev 107)
@@ -1,191 +0,0 @@
-/*
- aclib - advanced C library ;)
- This file contains functions which improve and expand standard C-library
-*/
-
-/* for small memory blocks (<256 bytes) this version is faster */
-#define small_memcpy(to,from,n)\
-{\
-register unsigned long int siz;\
-register unsigned long int dummy;\
- siz=n&0x7; n>>=3;\
- if(siz)\
-__asm__ __volatile__(\
- "rep; movsb"\
- :"=&D"(to), "=&S"(from), "=&c"(dummy)\
-/* It's most portable way to notify compiler */\
-/* that edi, esi and ecx are clobbered in asm block. */\
-/* Thanks to A'rpi for hint!!! */\
- :"0" (to), "1" (from),"2" (siz)\
- : "memory","cc");\
- if(n)\
-__asm__ __volatile__(\
- "rep; movsq"\
- :"=&D"(to), "=&S"(from), "=&c"(dummy)\
-/* It's most portable way to notify compiler */\
-/* that edi, esi and ecx are clobbered in asm block. */\
-/* Thanks to A'rpi for hint!!! */\
- :"0" (to), "1" (from),"2" (n)\
- : "memory","cc");\
-}
-
-
-#define MMREG_SIZE 16ULL
-#define MIN_LEN 257ULL
-#define CL_SIZE 256ULL /*always align on 256 byte boundary */
-
-static inline void * RENAME(fast_memcpy)(void * to, const void * from, size_t len)
-{
- void *retval;
- const unsigned char *cfrom=from;
- unsigned char *tto=to;
- size_t i=0;
- retval = to;
- if(!len) return retval;
- /* PREFETCH has effect even for MOVSB instruction ;) */
- __asm__ __volatile__ (
- "prefetcht0 (%0)\n"
- "prefetcht0 64(%0)\n"
- "prefetcht0 128(%0)\n"
- "prefetcht0 192(%0)\n"
- :: "r" (cfrom));
- if(len >= MIN_LEN)
- {
- register unsigned long int delta;
- /* Align destinition to cache-line size -boundary */
- delta = ((unsigned long int)tto)&(CL_SIZE-1ULL);
- if(delta)
- {
- delta=CL_SIZE-delta;
- len -=delta;
- small_memcpy(tto, cfrom, delta);
- }
- i = len>>8; /* len/256 */
- len=len-(i<<8);
- }
- if(i) {
- /*
- This algorithm is top effective when the code consequently
- reads and writes blocks which have size of cache line.
- Size of cache line is processor-dependent.
- It will, however, be a minimum of 32 bytes on any processors.
- It would be better to have a number of instructions which
- perform reading and writing to be multiple to a number of
- processor's decoders, but it's not always possible.
- */
- if(((unsigned long)cfrom) & 15)
- /* if SRC is misaligned */
- for(; i>0; i--)
- {
- __asm__ __volatile__ (
- "prefetcht0 256(%0)\n"
- "prefetcht0 320(%0)\n"
- "movdqu (%0), %%xmm0\n"
- "movdqu 16(%0), %%xmm1\n"
- "movdqu 32(%0), %%xmm2\n"
- "movdqu 48(%0), %%xmm3\n"
- "movdqu 64(%0), %%xmm4\n"
- "movdqu 80(%0), %%xmm5\n"
- "movdqu 96(%0), %%xmm6\n"
- "movdqu 112(%0), %%xmm7\n"
- "prefetcht0 384(%0)\n"
- "prefetcht0 448(%0)\n"
- "movdqu 128(%0), %%xmm8\n"
- "movdqu 144(%0), %%xmm9\n"
- "movdqu 160(%0), %%xmm10\n"
- "movdqu 176(%0), %%xmm11\n"
- "movdqu 192(%0), %%xmm12\n"
- "movdqu 208(%0), %%xmm13\n"
- "movdqu 224(%0), %%xmm14\n"
- "movdqu 240(%0), %%xmm15\n"
- "movntdq %%xmm0, (%1)\n"
- "movntdq %%xmm1, 16(%1)\n"
- "movntdq %%xmm2, 32(%1)\n"
- "movntdq %%xmm3, 48(%1)\n"
- "movntdq %%xmm4, 64(%1)\n"
- "movntdq %%xmm5, 80(%1)\n"
- "movntdq %%xmm6, 96(%1)\n"
- "movntdq %%xmm7, 112(%1)\n"
- "movntdq %%xmm8, 128(%1)\n"
- "movntdq %%xmm9, 144(%1)\n"
- "movntdq %%xmm10, 160(%1)\n"
- "movntdq %%xmm11, 176(%1)\n"
- "movntdq %%xmm12, 192(%1)\n"
- "movntdq %%xmm13, 208(%1)\n"
- "movntdq %%xmm14, 224(%1)\n"
- "movntdq %%xmm15, 240(%1)\n"
- :: "r" (cfrom), "r" (tto):
- "memory"
- ,"xmm0","xmm1","xmm2","xmm3","xmm4","xmm5","xmm6","xmm7","xmm8","xmm9","xmm10","xmm11","xmm12","xmm13","xmm14","xmm15"
- );
- cfrom+=256ULL;
- tto+=256ULL;
- }
- else
- /*
- Only if SRC is aligned on 16-byte boundary.
- It allows to use movdqa instead of movdqu, which required data
- to be aligned or a general-protection exception (#GP) is generated.
- */
- for(; i>0; i--)
- {
- __asm__ __volatile__ (
- "prefetcht0 256(%0)\n"
- "prefetcht0 320(%0)\n"
- "movdqa (%0), %%xmm0\n"
- "movdqa 16(%0), %%xmm1\n"
- "movdqa 32(%0), %%xmm2\n"
- "movdqa 48(%0), %%xmm3\n"
- "movdqa 64(%0), %%xmm4\n"
- "movdqa 80(%0), %%xmm5\n"
- "movdqa 96(%0), %%xmm6\n"
- "movdqa 112(%0), %%xmm7\n"
- "prefetcht0 384(%0)\n"
- "prefetcht0 448(%0)\n"
- "movdqa 128(%0), %%xmm8\n"
- "movdqa 144(%0), %%xmm9\n"
- "movdqa 160(%0), %%xmm10\n"
- "movdqa 176(%0), %%xmm11\n"
- "movdqa 192(%0), %%xmm12\n"
- "movdqa 208(%0), %%xmm13\n"
- "movdqa 224(%0), %%xmm14\n"
- "movdqa 240(%0), %%xmm15\n"
- "movntdq %%xmm0, (%1)\n"
- "movntdq %%xmm1, 16(%1)\n"
- "movntdq %%xmm2, 32(%1)\n"
- "movntdq %%xmm3, 48(%1)\n"
- "movntdq %%xmm4, 64(%1)\n"
- "movntdq %%xmm5, 80(%1)\n"
- "movntdq %%xmm6, 96(%1)\n"
- "movntdq %%xmm7, 112(%1)\n"
- "movntdq %%xmm8, 128(%1)\n"
- "movntdq %%xmm9, 144(%1)\n"
- "movntdq %%xmm10, 160(%1)\n"
- "movntdq %%xmm11, 176(%1)\n"
- "movntdq %%xmm12, 192(%1)\n"
- "movntdq %%xmm13, 208(%1)\n"
- "movntdq %%xmm14, 224(%1)\n"
- "movntdq %%xmm15, 240(%1)\n"
- :: "r" (cfrom), "r" (tto):
- "memory"
- ,"xmm0","xmm1","xmm2","xmm3","xmm4","xmm5","xmm6","xmm7","xmm8","xmm9","xmm10","xmm11","xmm12","xmm13","xmm14","xmm15"
- );
- cfrom+=256ULL;
- tto+=256ULL;
- }
- __asm__ __volatile__ ("sfence":::"memory");
- }
- /*
- * Now do the tail of the block
- */
- if(len) small_memcpy(tto, cfrom, len);
- return retval;
-}
-
-/**
- * special copy routine for mem -> agp/pci copy (based upon fast_memcpy)
- */
-static inline void * RENAME(mem2agpcpy)(void * to, const void * from, size_t len)
-{
- return memcpy(to,from,len);
-}
Modified: mplayerxp/libvo/dri_vo.h
===================================================================
--- mplayerxp/libvo/dri_vo.h 2010-01-17 18:46:44 UTC (rev 106)
+++ mplayerxp/libvo/dri_vo.h 2010-01-18 15:55:23 UTC (rev 107)
@@ -20,6 +20,7 @@
#define DRI_CAP_HORZSCALER 0x00000040UL /**< Driver supports horizontal scaling */
#define DRI_CAP_VERTSCALER 0x00000080UL /**< Driver supports vertical scaling */
#define DRI_CAP_HWOSD 0x00000100UL /**< Driver supports OSD painting */
+#define DRI_CAP_BUSMASTERING 0x80000000UL /**< Means: final video buffer but allocated in RAM */
typedef struct dri_surface_cap_s
{
Modified: mplayerxp/libvo/fastmemcpy.h
===================================================================
--- mplayerxp/libvo/fastmemcpy.h 2010-01-17 18:46:44 UTC (rev 106)
+++ mplayerxp/libvo/fastmemcpy.h 2010-01-18 15:55:23 UTC (rev 107)
@@ -7,24 +7,24 @@
#include <stddef.h>
#include <string.h> /* memcpy prototypes */
extern void * (*fast_memcpy_ptr)(void * to, const void * from, size_t len);
-extern void * (*mem2agpcpy_ptr)(void * to, const void * from, size_t len);
+extern void * (*fast_stream_copy_ptr)(void * to, const void * from, size_t len);
#define memcpy(a,b,c) (*fast_memcpy_ptr)(a,b,c)
-#define mem2agpcpy(a,b,c) (*mem2agpcpy_ptr)(a,b,c)
+#define stream_copy(a,b,c) (*fast_stream_copy_ptr)(a,b,c)
#else
-#define mem2agpcpy(a,b,c) memcpy(a,b,c)
+#define stream_copy(a,b,c) memcpy(a,b,c)
#endif
-static inline void * mem2agpcpy_pic(void * dst, const void * src, int bytesPerLine, int height, int dstStride, int srcStride)
+static inline void * stream_copy_pic(void * dst, const void * src, int bytesPerLine, int height, int dstStride, int srcStride)
{
int i;
void *retval=dst;
- if(dstStride == srcStride) mem2agpcpy(dst, src, srcStride*height);
+ if(dstStride == srcStride) stream_copy(dst, src, srcStride*height);
else
{
for(i=0; i<height; i++)
{
- mem2agpcpy(dst, src, bytesPerLine);
+ stream_copy(dst, src, bytesPerLine);
src+= srcStride;
dst+= dstStride;
}
Modified: mplayerxp/libvo/osd.c
===================================================================
--- mplayerxp/libvo/osd.c 2010-01-17 18:46:44 UTC (rev 106)
+++ mplayerxp/libvo/osd.c 2010-01-18 15:55:23 UTC (rev 107)
@@ -73,7 +73,7 @@
static unsigned short fast_osd_16bpp_table[256];
#endif
-static void __FASTCALL__ vo_draw_alpha_rgb15_C(int w,int h, const unsigned char* src, const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride){
+static void __FASTCALL__ vo_draw_alpha_rgb15_C(int w,int h, const unsigned char* src, const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride,int finalize){
int y;
for(y=0;y<h;y++){
register unsigned short *dst = (unsigned short*) dstbase;
@@ -105,7 +105,7 @@
return;
}
-static void __FASTCALL__ vo_draw_alpha_rgb16_C(int w,int h, const unsigned char* src, const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride){
+static void __FASTCALL__ vo_draw_alpha_rgb16_C(int w,int h, const unsigned char* src, const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride,int finalize){
int y;
for(y=0;y<h;y++){
register unsigned short *dst = (unsigned short*) dstbase;
@@ -136,8 +136,8 @@
return;
}
-static void __FASTCALL__ vo_draw_alpha_uyvy_C(int w,int h, const unsigned char* src, const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride){
- (*vo_draw_alpha_yuy2_ptr)(w,h,src,srca,srcstride,dstbase+1,dststride);
+static void __FASTCALL__ vo_draw_alpha_uyvy_C(int w,int h, const unsigned char* src, const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride,int finalize){
+ (*vo_draw_alpha_yuy2_ptr)(w,h,src,srca,srcstride,dstbase+1,dststride,finalize);
}
draw_alpha_f vo_draw_alpha_yv12_ptr=NULL;
Modified: mplayerxp/libvo/osd.h
===================================================================
--- mplayerxp/libvo/osd.h 2010-01-17 18:46:44 UTC (rev 106)
+++ mplayerxp/libvo/osd.h 2010-01-18 15:55:23 UTC (rev 107)
@@ -6,7 +6,7 @@
extern void vo_draw_alpha_init( void ); /* build tables */
-typedef void (* __FASTCALL__ draw_alpha_f)(int w,int h, const unsigned char* src, const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride);
+typedef void (* __FASTCALL__ draw_alpha_f)(int w,int h, const unsigned char* src, const unsigned char *srca, int srcstride, unsigned char* dstbase,int dststride,int finalize);
extern draw_alpha_f vo_draw_alpha_yv12_ptr;
extern draw_alpha_f vo_draw_alpha_yuy2_ptr;
@@ -15,11 +15,11 @@
extern draw_alpha_f vo_draw_alpha_rgb32_ptr;
extern draw_alpha_f vo_draw_alpha_rgb15_ptr;
extern draw_alpha_f vo_draw_alpha_rgb16_ptr;
-#define vo_draw_alpha_yv12(w,h,src,srca,srcstride,dstbase,dstrstride) (*vo_draw_alpha_yv12_ptr)(w,h,src,srca,srcstride,dstbase,dstrstride)
-#define vo_draw_alpha_yuy2(w,h,src,srca,srcstride,dstbase,dstrstride) (*vo_draw_alpha_yuy2_ptr)(w,h,src,srca,srcstride,dstbase,dstrstride)
-#define vo_draw_alpha_uyvy(w,h,src,srca,srcstride,dstbase,dstrstride) (*vo_draw_alpha_uyvy_ptr)(w,h,src,srca,srcstride,dstbase,dstrstride)
-#define vo_draw_alpha_rgb24(w,h,src,srca,srcstride,dstbase,dstrstride) (*vo_draw_alpha_rgb24_ptr)(w,h,src,srca,srcstride,dstbase,dstrstride)
-#define vo_draw_alpha_rgb32(w,h,src,srca,srcstride,dstbase,dstrstride) (*vo_draw_alpha_rgb32_ptr)(w,h,src,srca,srcstride,dstbase,dstrstride)
-#define vo_draw_alpha_rgb15(w,h,src,srca,srcstride,dstbase,dstrstride) (*vo_draw_alpha_rgb15_ptr)(w,h,src,srca,srcstride,dstbase,dstrstride)
-#define vo_draw_alpha_rgb16(w,h,src,srca,srcstride,dstbase,dstrstride) (*vo_draw_alpha_rgb16_ptr)(w,h,src,srca,srcstride,dstbase,dstrstride)
+#define vo_draw_alpha_yv12(w,h,src,srca,srcstride,dstbase,dstrstride,finalize) (*vo_draw_alpha_yv12_ptr)(w,h,src,srca,srcstride,dstbase,dstrstride,finalize)
+#define vo_draw_alpha_yuy2(w,h,src,srca,srcstride,dstbase,dstrstride,finalize) (*vo_draw_alpha_yuy2_ptr)(w,h,src,srca,srcstride,dstbase,dstrstride,finalize)
+#define vo_draw_alpha_uyvy(w,h,src,srca,srcstride,dstbase,dstrstride,finalize) (*vo_draw_alpha_uyvy_ptr)(w,h,src,srca,srcstride,dstbase,dstrstride,finalize)
+#define vo_draw_alpha_rgb24(w,h,src,srca,srcstride,dstbase,dstrstride,finalize) (*vo_draw_alpha_rgb24_ptr)(w,h,src,srca,srcstride,dstbase,dstrstride,finalize)
+#define vo_draw_alpha_rgb32(w,h,src,srca,srcstride,dstbase,dstrstride,finalize) (*vo_draw_alpha_rgb32_ptr)(w,h,src,srca,srcstride,dstbase,dstrstride,finalize)
+#define vo_draw_alpha_rgb15(w,h,src,srca,srcstride,dstbase,dstrstride,finalize) (*vo_draw_alpha_rgb15_ptr)(w,h,src,srca,srcstride,dstbase,dstrstride,finalize)
+#define vo_draw_alpha_rgb16(w,h,src,srca,srcstride,dstbase,dstrstride,finalize) (*vo_draw_alpha_rgb16_ptr)(w,h,src,srca,srcstride,dstbase,dstrstride,finalize)
#endif
Modified: mplayerxp/libvo/osd_template.c
===================================================================
--- mplayerxp/libvo/osd_template.c 2010-01-17 18:46:44 UTC (rev 106)
+++ mplayerxp/libvo/osd_template.c 2010-01-18 15:55:23 UTC (rev 107)
@@ -50,7 +50,7 @@
#endif
-static inline void RENAME(vo_draw_alpha_yv12)(int w,int h,const unsigned char* src,const unsig...
[truncated message content] |
