/*****************************************************************************
* Preamble
*****************************************************************************/
-#include <string.h> /* strerror() */
-#include <stdlib.h> /* malloc(), free() */
-#include <vlc/vlc.h>
+#ifdef HAVE_CONFIG_H
+# include "config.h"
+#endif
+
+#include <vlc_common.h>
+#include <vlc_plugin.h>
+#include <vlc_filter.h>
#include <vlc_vout.h>
#if defined (MODULE_NAME_IS_i420_yuy2_altivec) && defined(HAVE_ALTIVEC_H)
*****************************************************************************/
static int Activate ( vlc_object_t * );
-static void I420_YUY2 ( vout_thread_t *, picture_t *, picture_t * );
-static void I420_YVYU ( vout_thread_t *, picture_t *, picture_t * );
-static void I420_UYVY ( vout_thread_t *, picture_t *, picture_t * );
+static void I420_YUY2 ( filter_t *, picture_t *, picture_t * );
+static void I420_YVYU ( filter_t *, picture_t *, picture_t * );
+static void I420_UYVY ( filter_t *, picture_t *, picture_t * );
+static picture_t *I420_YUY2_Filter ( filter_t *, picture_t * );
+static picture_t *I420_YVYU_Filter ( filter_t *, picture_t * );
+static picture_t *I420_UYVY_Filter ( filter_t *, picture_t * );
#if !defined (MODULE_NAME_IS_i420_yuy2_altivec)
-static void I420_IUYV ( vout_thread_t *, picture_t *, picture_t * );
-static void I420_cyuv ( vout_thread_t *, picture_t *, picture_t * );
+static void I420_IUYV ( filter_t *, picture_t *, picture_t * );
+static void I420_cyuv ( filter_t *, picture_t *, picture_t * );
+static picture_t *I420_IUYV_Filter ( filter_t *, picture_t * );
+static picture_t *I420_cyuv_Filter ( filter_t *, picture_t * );
#endif
#if defined (MODULE_NAME_IS_i420_yuy2)
-static void I420_Y211 ( vout_thread_t *, picture_t *, picture_t * );
+static void I420_Y211 ( filter_t *, picture_t *, picture_t * );
+static picture_t *I420_Y211_Filter ( filter_t *, picture_t * );
#endif
#ifdef MODULE_NAME_IS_i420_yuy2_mmx
*****************************************************************************/
vlc_module_begin();
#if defined (MODULE_NAME_IS_i420_yuy2)
- set_description( _("Conversions from " SRC_FOURCC " to " DEST_FOURCC) );
- set_capability( "chroma", 80 );
+ set_description( N_("Conversions from " SRC_FOURCC " to " DEST_FOURCC) );
+ set_capability( "video filter2", 80 );
#elif defined (MODULE_NAME_IS_i420_yuy2_mmx)
- set_description( _("MMX conversions from " SRC_FOURCC " to " DEST_FOURCC) );
- set_capability( "chroma", 100 );
+ set_description( N_("MMX conversions from " SRC_FOURCC " to " DEST_FOURCC) );
+ set_capability( "video filter2", 100 );
add_requirement( MMX );
#elif defined (MODULE_NAME_IS_i420_yuy2_sse2)
- set_description( _("SSE2 conversions from " SRC_FOURCC " to " DEST_FOURCC) );
- set_capability( "chroma", 120 );
+ set_description( N_("SSE2 conversions from " SRC_FOURCC " to " DEST_FOURCC) );
+ set_capability( "video filter2", 120 );
add_requirement( SSE2 );
#elif defined (MODULE_NAME_IS_i420_yuy2_altivec)
set_description(
_("AltiVec conversions from " SRC_FOURCC " to " DEST_FOURCC) );
- set_capability( "chroma", 100 );
+ set_capability( "video filter2", 100 );
add_requirement( ALTIVEC );
#endif
set_callbacks( Activate, NULL );
*****************************************************************************/
static int Activate( vlc_object_t *p_this )
{
- vout_thread_t *p_vout = (vout_thread_t *)p_this;
+ filter_t *p_filter = (filter_t *)p_this;
- if( p_vout->render.i_width & 1 || p_vout->render.i_height & 1 )
+ if( p_filter->fmt_in.video.i_width & 1
+ || p_filter->fmt_in.video.i_height & 1 )
{
return -1;
}
- switch( p_vout->render.i_chroma )
+ switch( p_filter->fmt_in.video.i_chroma )
{
case VLC_FOURCC('Y','V','1','2'):
case VLC_FOURCC('I','4','2','0'):
case VLC_FOURCC('I','Y','U','V'):
- switch( p_vout->output.i_chroma )
+ switch( p_filter->fmt_out.video.i_chroma )
{
case VLC_FOURCC('Y','U','Y','2'):
case VLC_FOURCC('Y','U','N','V'):
- p_vout->chroma.pf_convert = I420_YUY2;
+ p_filter->pf_video_filter = I420_YUY2_Filter;
break;
case VLC_FOURCC('Y','V','Y','U'):
- p_vout->chroma.pf_convert = I420_YVYU;
+ p_filter->pf_video_filter = I420_YVYU_Filter;
break;
case VLC_FOURCC('U','Y','V','Y'):
case VLC_FOURCC('U','Y','N','V'):
case VLC_FOURCC('Y','4','2','2'):
- p_vout->chroma.pf_convert = I420_UYVY;
+ p_filter->pf_video_filter = I420_UYVY_Filter;
break;
#if !defined (MODULE_NAME_IS_i420_yuy2_altivec)
case VLC_FOURCC('I','U','Y','V'):
- p_vout->chroma.pf_convert = I420_IUYV;
+ p_filter->pf_video_filter = I420_IUYV_Filter;
break;
case VLC_FOURCC('c','y','u','v'):
- p_vout->chroma.pf_convert = I420_cyuv;
+ p_filter->pf_video_filter = I420_cyuv_Filter;
break;
#endif
#if defined (MODULE_NAME_IS_i420_yuy2)
case VLC_FOURCC('Y','2','1','1'):
- p_vout->chroma.pf_convert = I420_Y211;
+ p_filter->pf_video_filter = I420_Y211_Filter;
break;
#endif
#endif
/* Following functions are local */
+
+VIDEO_FILTER_WRAPPER( I420_YUY2 )
+VIDEO_FILTER_WRAPPER( I420_YVYU )
+VIDEO_FILTER_WRAPPER( I420_UYVY )
+#if !defined (MODULE_NAME_IS_i420_yuy2_altivec)
+VIDEO_FILTER_WRAPPER( I420_IUYV )
+#endif
+#if defined (MODULE_NAME_IS_i420_yuy2)
+VIDEO_FILTER_WRAPPER( I420_Y211 )
+#endif
+
/*****************************************************************************
* I420_YUY2: planar YUV 4:2:0 to packed YUYV 4:2:2
*****************************************************************************/
-static void I420_YUY2( vout_thread_t *p_vout, picture_t *p_source,
- picture_t *p_dest )
+static void I420_YUY2( filter_t *p_filter, picture_t *p_source,
+ picture_t *p_dest )
{
uint8_t *p_line1, *p_line2 = p_dest->p->p_pixels;
uint8_t *p_y1, *p_y2 = p_source->Y_PIXELS;
vector unsigned char uv_vec;
vector unsigned char y_vec;
- if( !( ( p_vout->render.i_width % 32 ) |
- ( p_vout->render.i_height % 2 ) ) )
+ if( !( ( p_filter->fmt_in.video.i_width % 32 ) |
+ ( p_filter->fmt_in.video.i_height % 2 ) ) )
{
/* Width is a multiple of 32, we take 2 lines at a time */
- for( i_y = p_vout->render.i_height / 2 ; i_y-- ; )
+ for( i_y = p_filter->fmt_in.video.i_height / 2 ; i_y-- ; )
{
VEC_NEXT_LINES( );
- for( i_x = p_vout->render.i_width / 32 ; i_x-- ; )
+ for( i_x = p_filter->fmt_in.video.i_width / 32 ; i_x-- ; )
{
VEC_LOAD_UV( );
VEC_MERGE( vec_mergeh );
}
}
}
- else if( !( ( p_vout->render.i_width % 16 ) |
- ( p_vout->render.i_height % 4 ) ) )
+ else if( !( ( p_filter->fmt_in.video.i_width % 16 ) |
+ ( p_filter->fmt_in.video.i_height % 4 ) ) )
{
/* Width is only a multiple of 16, we take 4 lines at a time */
- for( i_y = p_vout->render.i_height / 4 ; i_y-- ; )
+ for( i_y = p_filter->fmt_in.video.i_height / 4 ; i_y-- ; )
{
/* Line 1 and 2, pixels 0 to ( width - 16 ) */
VEC_NEXT_LINES( );
- for( i_x = p_vout->render.i_width / 32 ; i_x-- ; )
+ for( i_x = p_fiter->fmt_in.video.i_width / 32 ; i_x-- ; )
{
VEC_LOAD_UV( );
VEC_MERGE( vec_mergeh );
VEC_MERGE( vec_mergel );
/* Line 3 and 4, pixels 16 to ( width ) */
- for( i_x = p_vout->render.i_width / 32 ; i_x-- ; )
+ for( i_x = p_filter->fmt_in.video.i_width / 32 ; i_x-- ; )
{
VEC_LOAD_UV( );
VEC_MERGE( vec_mergeh );
- p_dest->p->i_visible_pitch;
#if !defined(MODULE_NAME_IS_i420_yuy2_sse2)
- for( i_y = p_vout->render.i_height / 2 ; i_y-- ; )
+ for( i_y = p_filter->fmt_in.video.i_height / 2 ; i_y-- ; )
{
p_line1 = p_line2;
p_line2 += p_dest->p->i_pitch;
p_y2 += p_source->p[Y_PLANE].i_pitch;
#if !defined (MODULE_NAME_IS_i420_yuy2_mmx)
- for( i_x = p_vout->render.i_width / 8; i_x-- ; )
+ for( i_x = p_filter->fmt_in.video.i_width / 8; i_x-- ; )
{
C_YUV420_YUYV( );
C_YUV420_YUYV( );
C_YUV420_YUYV( );
}
#else
- for( i_x = p_vout->render.i_width / 8 ; i_x-- ; )
+ for( i_x = p_filter->fmt_in.video.i_width / 8 ; i_x-- ; )
{
MMX_CALL( MMX_YUV420_YUYV );
}
#endif
- for( i_x = ( p_vout->render.i_width % 8 ) / 2; i_x-- ; )
+ for( i_x = ( p_filter->fmt_in.video.i_width % 8 ) / 2; i_x-- ; )
{
C_YUV420_YUYV( );
}
#if defined (MODULE_NAME_IS_i420_yuy2_mmx)
/* re-enable FPU registers */
- __asm__ __volatile__ ( "emms" );
+ MMX_END;
#endif
#if defined (MODULE_NAME_IS_i420_yuy2_altivec)
#else // defined(MODULE_NAME_IS_i420_yuy2_sse2)
/*
- ** SSE2 128 bits fetch/store instructions are faster
+ ** SSE2 128 bits fetch/store instructions are faster
** if memory access is 16 bytes aligned
*/
if( 0 == (15 & (p_source->p[Y_PLANE].i_pitch|p_dest->p->i_pitch|
- ((int)p_line2|(int)p_y2))) )
+ ((intptr_t)p_line2|(intptr_t)p_y2))) )
{
/* use faster SSE2 aligned fetch and store */
- for( i_y = p_vout->render.i_height / 2 ; i_y-- ; )
+ for( i_y = p_filter->fmt_in.video.i_height / 2 ; i_y-- ; )
{
p_line1 = p_line2;
p_line2 += p_dest->p->i_pitch;
p_y1 = p_y2;
p_y2 += p_source->p[Y_PLANE].i_pitch;
- for( i_x = p_vout->render.i_width / 16 ; i_x-- ; )
+ for( i_x = p_filter->fmt_in.video.i_width / 16 ; i_x-- ; )
{
SSE2_CALL( SSE2_YUV420_YUYV_ALIGNED );
}
- for( i_x = ( p_vout->render.i_width % 16 ) / 2; i_x-- ; )
+ for( i_x = ( p_filter->fmt_in.video.i_width % 16 ) / 2; i_x-- ; )
{
C_YUV420_YUYV( );
}
p_line1 += i_dest_margin;
p_line2 += i_dest_margin;
}
- /* make sure all SSE2 stores are visible thereafter */
- __asm__ __volatile__ ( "sfence" );
}
else
{
/* use slower SSE2 unaligned fetch and store */
- for( i_y = p_vout->render.i_height / 2 ; i_y-- ; )
+ for( i_y = p_filter->fmt_in.video.i_height / 2 ; i_y-- ; )
{
p_line1 = p_line2;
p_line2 += p_dest->p->i_pitch;
p_y1 = p_y2;
p_y2 += p_source->p[Y_PLANE].i_pitch;
- for( i_x = p_vout->render.i_width / 16 ; i_x-- ; )
+ for( i_x = p_filter->fmt_in.video.i_width / 16 ; i_x-- ; )
{
SSE2_CALL( SSE2_YUV420_YUYV_UNALIGNED );
}
- for( i_x = ( p_vout->render.i_width % 16 ) / 2; i_x-- ; )
+ for( i_x = ( p_filter->fmt_in.video.i_width % 16 ) / 2; i_x-- ; )
{
C_YUV420_YUYV( );
}
p_line2 += i_dest_margin;
}
}
+ /* make sure all SSE2 stores are visible thereafter */
+ SSE2_END;
#endif // defined(MODULE_NAME_IS_i420_yuy2_sse2)
}
/*****************************************************************************
* I420_YVYU: planar YUV 4:2:0 to packed YVYU 4:2:2
*****************************************************************************/
-static void I420_YVYU( vout_thread_t *p_vout, picture_t *p_source,
- picture_t *p_dest )
+static void I420_YVYU( filter_t *p_filter, picture_t *p_source,
+ picture_t *p_dest )
{
uint8_t *p_line1, *p_line2 = p_dest->p->p_pixels;
uint8_t *p_y1, *p_y2 = p_source->Y_PIXELS;
vector unsigned char vu_vec;
vector unsigned char y_vec;
- if( !( ( p_vout->render.i_width % 32 ) |
- ( p_vout->render.i_height % 2 ) ) )
+ if( !( ( p_filter->fmt_in.video.i_width % 32 ) |
+ ( p_filter->fmt_in.video.i_height % 2 ) ) )
{
/* Width is a multiple of 32, we take 2 lines at a time */
- for( i_y = p_vout->render.i_height / 2 ; i_y-- ; )
+ for( i_y = p_filter->fmt_in.video.i_height / 2 ; i_y-- ; )
{
VEC_NEXT_LINES( );
- for( i_x = p_vout->render.i_width / 32 ; i_x-- ; )
+ for( i_x = p_filter->fmt_in.video.i_width / 32 ; i_x-- ; )
{
VEC_LOAD_UV( );
VEC_MERGE( vec_mergeh );
}
}
}
- else if( !( ( p_vout->render.i_width % 16 ) |
- ( p_vout->render.i_height % 4 ) ) )
+ else if( !( ( p_filter->fmt_in.video.i_width % 16 ) |
+ ( p_filter->fmt_in.video.i_height % 4 ) ) )
{
/* Width is only a multiple of 16, we take 4 lines at a time */
- for( i_y = p_vout->render.i_height / 4 ; i_y-- ; )
+ for( i_y = p_filter->fmt_in.video.i_height / 4 ; i_y-- ; )
{
/* Line 1 and 2, pixels 0 to ( width - 16 ) */
VEC_NEXT_LINES( );
- for( i_x = p_vout->render.i_width / 32 ; i_x-- ; )
+ for( i_x = p_filter->fmt_in.video.i_width / 32 ; i_x-- ; )
{
VEC_LOAD_UV( );
VEC_MERGE( vec_mergeh );
VEC_MERGE( vec_mergel );
/* Line 3 and 4, pixels 16 to ( width ) */
- for( i_x = p_vout->render.i_width / 32 ; i_x-- ; )
+ for( i_x = p_filter->fmt_in.video.i_width / 32 ; i_x-- ; )
{
VEC_LOAD_UV( );
VEC_MERGE( vec_mergeh );
- p_dest->p->i_visible_pitch;
#if !defined(MODULE_NAME_IS_i420_yuy2_sse2)
- for( i_y = p_vout->render.i_height / 2 ; i_y-- ; )
+ for( i_y = p_filter->fmt_in.video.i_height / 2 ; i_y-- ; )
{
p_line1 = p_line2;
p_line2 += p_dest->p->i_pitch;
p_y1 = p_y2;
p_y2 += p_source->p[Y_PLANE].i_pitch;
- for( i_x = p_vout->render.i_width / 8 ; i_x-- ; )
+ for( i_x = p_filter->fmt_in.video.i_width / 8 ; i_x-- ; )
{
#if !defined (MODULE_NAME_IS_i420_yuy2_mmx)
C_YUV420_YVYU( );
MMX_CALL( MMX_YUV420_YVYU );
#endif
}
- for( i_x = ( p_vout->render.i_width % 8 ) / 2; i_x-- ; )
+ for( i_x = ( p_filter->fmt_in.video.i_width % 8 ) / 2; i_x-- ; )
{
C_YUV420_YVYU( );
}
#if defined (MODULE_NAME_IS_i420_yuy2_mmx)
/* re-enable FPU registers */
- __asm__ __volatile__ ( "emms" );
+ MMX_END;
#endif
#if defined (MODULE_NAME_IS_i420_yuy2_altivec)
#else // defined(MODULE_NAME_IS_i420_yuy2_sse2)
/*
- ** SSE2 128 bits fetch/store instructions are faster
+ ** SSE2 128 bits fetch/store instructions are faster
** if memory access is 16 bytes aligned
*/
if( 0 == (15 & (p_source->p[Y_PLANE].i_pitch|p_dest->p->i_pitch|
- ((int)p_line2|(int)p_y2))) )
+ ((intptr_t)p_line2|(intptr_t)p_y2))) )
{
/* use faster SSE2 aligned fetch and store */
- for( i_y = p_vout->render.i_height / 2 ; i_y-- ; )
+ for( i_y = p_filter->fmt_in.video.i_height / 2 ; i_y-- ; )
{
p_line1 = p_line2;
p_line2 += p_dest->p->i_pitch;
p_y1 = p_y2;
p_y2 += p_source->p[Y_PLANE].i_pitch;
- for( i_x = p_vout->render.i_width / 16 ; i_x-- ; )
+ for( i_x = p_filter->fmt_in.video.i_width / 16 ; i_x-- ; )
{
SSE2_CALL( SSE2_YUV420_YVYU_ALIGNED );
}
- for( i_x = ( p_vout->render.i_width % 16 ) / 2; i_x-- ; )
+ for( i_x = ( p_filter->fmt_in.video.i_width % 16 ) / 2; i_x-- ; )
{
C_YUV420_YVYU( );
}
p_line1 += i_dest_margin;
p_line2 += i_dest_margin;
}
- /* make sure all SSE2 stores are visible thereafter */
- __asm__ __volatile__ ( "sfence" );
}
else
{
/* use slower SSE2 unaligned fetch and store */
- for( i_y = p_vout->render.i_height / 2 ; i_y-- ; )
+ for( i_y = p_filter->fmt_in.video.i_height / 2 ; i_y-- ; )
{
p_line1 = p_line2;
p_line2 += p_dest->p->i_pitch;
p_y1 = p_y2;
p_y2 += p_source->p[Y_PLANE].i_pitch;
- for( i_x = p_vout->render.i_width / 16 ; i_x-- ; )
+ for( i_x = p_filter->fmt_in.video.i_width / 16 ; i_x-- ; )
{
SSE2_CALL( SSE2_YUV420_YVYU_UNALIGNED );
}
- for( i_x = ( p_vout->render.i_width % 16 ) / 2; i_x-- ; )
+ for( i_x = ( p_filter->fmt_in.video.i_width % 16 ) / 2; i_x-- ; )
{
C_YUV420_YVYU( );
}
p_line2 += i_dest_margin;
}
}
+ /* make sure all SSE2 stores are visible thereafter */
+ SSE2_END;
#endif // defined(MODULE_NAME_IS_i420_yuy2_sse2)
}
/*****************************************************************************
* I420_UYVY: planar YUV 4:2:0 to packed UYVY 4:2:2
*****************************************************************************/
-static void I420_UYVY( vout_thread_t *p_vout, picture_t *p_source,
- picture_t *p_dest )
+static void I420_UYVY( filter_t *p_filter, picture_t *p_source,
+ picture_t *p_dest )
{
uint8_t *p_line1, *p_line2 = p_dest->p->p_pixels;
uint8_t *p_y1, *p_y2 = p_source->Y_PIXELS;
vector unsigned char uv_vec;
vector unsigned char y_vec;
- if( !( ( p_vout->render.i_width % 32 ) |
- ( p_vout->render.i_height % 2 ) ) )
+ if( !( ( p_filter->fmt_in.video.i_width % 32 ) |
+ ( p_filter->fmt_in.video.i_height % 2 ) ) )
{
/* Width is a multiple of 32, we take 2 lines at a time */
- for( i_y = p_vout->render.i_height / 2 ; i_y-- ; )
+ for( i_y = p_filter->fmt_in.video.i_height / 2 ; i_y-- ; )
{
VEC_NEXT_LINES( );
- for( i_x = p_vout->render.i_width / 32 ; i_x-- ; )
+ for( i_x = p_filter->fmt_in.video.i_width / 32 ; i_x-- ; )
{
VEC_LOAD_UV( );
VEC_MERGE( vec_mergeh );
}
}
}
- else if( !( ( p_vout->render.i_width % 16 ) |
- ( p_vout->render.i_height % 4 ) ) )
+ else if( !( ( p_filter->fmt_in.video.i_width % 16 ) |
+ ( p_filter->fmt_in.video.i_height % 4 ) ) )
{
/* Width is only a multiple of 16, we take 4 lines at a time */
- for( i_y = p_vout->render.i_height / 4 ; i_y-- ; )
+ for( i_y = p_filter->fmt_in.video.i_height / 4 ; i_y-- ; )
{
/* Line 1 and 2, pixels 0 to ( width - 16 ) */
VEC_NEXT_LINES( );
- for( i_x = p_vout->render.i_width / 32 ; i_x-- ; )
+ for( i_x = p_filter->fmt_in.video.i_width / 32 ; i_x-- ; )
{
VEC_LOAD_UV( );
VEC_MERGE( vec_mergeh );
VEC_MERGE( vec_mergel );
/* Line 3 and 4, pixels 16 to ( width ) */
- for( i_x = p_vout->render.i_width / 32 ; i_x-- ; )
+ for( i_x = p_filter->fmt_in.video.i_width / 32 ; i_x-- ; )
{
VEC_LOAD_UV( );
VEC_MERGE( vec_mergeh );
- p_dest->p->i_visible_pitch;
#if !defined(MODULE_NAME_IS_i420_yuy2_sse2)
- for( i_y = p_vout->render.i_height / 2 ; i_y-- ; )
+ for( i_y = p_filter->fmt_in.video.i_height / 2 ; i_y-- ; )
{
p_line1 = p_line2;
p_line2 += p_dest->p->i_pitch;
p_y1 = p_y2;
p_y2 += p_source->p[Y_PLANE].i_pitch;
- for( i_x = p_vout->render.i_width / 8 ; i_x-- ; )
+ for( i_x = p_filter->fmt_in.video.i_width / 8 ; i_x-- ; )
{
#if !defined (MODULE_NAME_IS_i420_yuy2_mmx)
C_YUV420_UYVY( );
MMX_CALL( MMX_YUV420_UYVY );
#endif
}
- for( i_x = ( p_vout->render.i_width % 8 ) / 2; i_x--; )
+ for( i_x = ( p_filter->fmt_in.video.i_width % 8 ) / 2; i_x--; )
{
C_YUV420_UYVY( );
}
#if defined (MODULE_NAME_IS_i420_yuy2_mmx)
/* re-enable FPU registers */
- __asm__ __volatile__ ( "emms" );
+ MMX_END;
#endif
#if defined (MODULE_NAME_IS_i420_yuy2_altivec)
#else // defined(MODULE_NAME_IS_i420_yuy2_sse2)
/*
- ** SSE2 128 bits fetch/store instructions are faster
+ ** SSE2 128 bits fetch/store instructions are faster
** if memory access is 16 bytes aligned
*/
if( 0 == (15 & (p_source->p[Y_PLANE].i_pitch|p_dest->p->i_pitch|
- ((int)p_line2|(int)p_y2))) )
+ ((intptr_t)p_line2|(intptr_t)p_y2))) )
{
/* use faster SSE2 aligned fetch and store */
- for( i_y = p_vout->render.i_height / 2 ; i_y-- ; )
+ for( i_y = p_filter->fmt_in.video.i_height / 2 ; i_y-- ; )
{
p_line1 = p_line2;
p_line2 += p_dest->p->i_pitch;
p_y1 = p_y2;
p_y2 += p_source->p[Y_PLANE].i_pitch;
- for( i_x = p_vout->render.i_width / 16 ; i_x-- ; )
+ for( i_x = p_filter->fmt_in.video.i_width / 16 ; i_x-- ; )
{
SSE2_CALL( SSE2_YUV420_UYVY_ALIGNED );
}
- for( i_x = ( p_vout->render.i_width % 16 ) / 2; i_x-- ; )
+ for( i_x = ( p_filter->fmt_in.video.i_width % 16 ) / 2; i_x-- ; )
{
C_YUV420_UYVY( );
}
p_line1 += i_dest_margin;
p_line2 += i_dest_margin;
}
- /* make sure all SSE2 stores are visible thereafter */
- __asm__ __volatile__ ( "sfence" );
}
else
{
/* use slower SSE2 unaligned fetch and store */
- for( i_y = p_vout->render.i_height / 2 ; i_y-- ; )
+ for( i_y = p_filter->fmt_in.video.i_height / 2 ; i_y-- ; )
{
p_line1 = p_line2;
p_line2 += p_dest->p->i_pitch;
p_y1 = p_y2;
p_y2 += p_source->p[Y_PLANE].i_pitch;
- for( i_x = p_vout->render.i_width / 16 ; i_x-- ; )
+ for( i_x = p_filter->fmt_in.video.i_width / 16 ; i_x-- ; )
{
SSE2_CALL( SSE2_YUV420_UYVY_UNALIGNED );
}
- for( i_x = ( p_vout->render.i_width % 16 ) / 2; i_x-- ; )
+ for( i_x = ( p_filter->fmt_in.video.i_width % 16 ) / 2; i_x-- ; )
{
C_YUV420_UYVY( );
}
p_line2 += i_dest_margin;
}
}
+ /* make sure all SSE2 stores are visible thereafter */
+ SSE2_END;
#endif // defined(MODULE_NAME_IS_i420_yuy2_sse2)
}
/*****************************************************************************
* I420_IUYV: planar YUV 4:2:0 to interleaved packed UYVY 4:2:2
*****************************************************************************/
-static void I420_IUYV( vout_thread_t *p_vout, picture_t *p_source,
- picture_t *p_dest )
+static void I420_IUYV( filter_t *p_filter, picture_t *p_source,
+ picture_t *p_dest )
{
+ VLC_UNUSED(p_source); VLC_UNUSED(p_dest);
/* FIXME: TODO ! */
- msg_Err( p_vout, "I420_IUYV unimplemented, please harass <sam@zoy.org>" );
+ msg_Err( p_filter, "I420_IUYV unimplemented, please harass <sam@zoy.org>" );
}
/*****************************************************************************
* I420_cyuv: planar YUV 4:2:0 to upside-down packed UYVY 4:2:2
*****************************************************************************/
-static void I420_cyuv( vout_thread_t *p_vout, picture_t *p_source,
- picture_t *p_dest )
+static void I420_cyuv( filter_t *p_filter, picture_t *p_source,
+ picture_t *p_dest )
{
uint8_t *p_line1 = p_dest->p->p_pixels +
p_dest->p->i_visible_lines * p_dest->p->i_pitch
- p_dest->p->i_visible_pitch;
#if !defined(MODULE_NAME_IS_i420_yuy2_sse2)
- for( i_y = p_vout->render.i_height / 2 ; i_y-- ; )
+ for( i_y = p_filter->fmt_in.video.i_height / 2 ; i_y-- ; )
{
p_line1 -= 3 * p_dest->p->i_pitch;
p_line2 -= 3 * p_dest->p->i_pitch;
p_y1 = p_y2;
p_y2 += p_source->p[Y_PLANE].i_pitch;
- for( i_x = p_vout->render.i_width / 8 ; i_x-- ; )
+ for( i_x = p_filter->fmt_in.video.i_width / 8 ; i_x-- ; )
{
#if !defined (MODULE_NAME_IS_i420_yuy2_mmx)
C_YUV420_UYVY( );
MMX_CALL( MMX_YUV420_UYVY );
#endif
}
- for( i_x = ( p_vout->render.i_width % 8 ) / 2; i_x-- ; )
+ for( i_x = ( p_filter->fmt_in.video.i_width % 8 ) / 2; i_x-- ; )
{
C_YUV420_UYVY( );
}
#if defined (MODULE_NAME_IS_i420_yuy2_mmx)
/* re-enable FPU registers */
- __asm__ __volatile__ ( "emms" );
+ MMX_END;
#endif
#else // defined(MODULE_NAME_IS_i420_yuy2_sse2)
/*
- ** SSE2 128 bits fetch/store instructions are faster
+ ** SSE2 128 bits fetch/store instructions are faster
** if memory access is 16 bytes aligned
*/
if( 0 == (15 & (p_source->p[Y_PLANE].i_pitch|p_dest->p->i_pitch|
- ((int)p_line2|(int)p_y2))) )
+ ((intptr_t)p_line2|(intptr_t)p_y2))) )
{
/* use faster SSE2 aligned fetch and store */
- for( i_y = p_vout->render.i_height / 2 ; i_y-- ; )
+ for( i_y = p_filter->fmt_in.video.i_height / 2 ; i_y-- ; )
{
p_line1 = p_line2;
p_line2 += p_dest->p->i_pitch;
p_y1 = p_y2;
p_y2 += p_source->p[Y_PLANE].i_pitch;
- for( i_x = p_vout->render.i_width / 16 ; i_x-- ; )
+ for( i_x = p_filter->fmt_in.video.i_width / 16 ; i_x-- ; )
{
SSE2_CALL( SSE2_YUV420_UYVY_ALIGNED );
}
- for( i_x = ( p_vout->render.i_width % 16 ) / 2; i_x-- ; )
+ for( i_x = ( p_filter->fmt_in.video.i_width % 16 ) / 2; i_x-- ; )
{
C_YUV420_UYVY( );
}
p_line1 += i_dest_margin;
p_line2 += i_dest_margin;
}
- /* make sure all SSE2 stores are visible thereafter */
- __asm__ __volatile__ ( "sfence" );
}
else
{
/* use slower SSE2 unaligned fetch and store */
- for( i_y = p_vout->render.i_height / 2 ; i_y-- ; )
+ for( i_y = p_filter->fmt_in.video.i_height / 2 ; i_y-- ; )
{
p_line1 = p_line2;
p_line2 += p_dest->p->i_pitch;
p_y1 = p_y2;
p_y2 += p_source->p[Y_PLANE].i_pitch;
- for( i_x = p_vout->render.i_width / 16 ; i_x-- ; )
+ for( i_x = p_filter->fmt_in.video.i_width / 16 ; i_x-- ; )
{
SSE2_CALL( SSE2_YUV420_UYVY_UNALIGNED );
}
- for( i_x = ( p_vout->render.i_width % 16 ) / 2; i_x-- ; )
+ for( i_x = ( p_filter->fmt_in.video.i_width % 16 ) / 2; i_x-- ; )
{
C_YUV420_UYVY( );
}
p_line2 += i_dest_margin;
}
}
+ /* make sure all SSE2 stores are visible thereafter */
+ SSE2_END;
#endif // defined(MODULE_NAME_IS_i420_yuy2_sse2)
}
#endif // !defined (MODULE_NAME_IS_i420_yuy2_altivec)
* I420_Y211: planar YUV 4:2:0 to packed YUYV 2:1:1
*****************************************************************************/
#if defined (MODULE_NAME_IS_i420_yuy2)
-static void I420_Y211( vout_thread_t *p_vout, picture_t *p_source,
- picture_t *p_dest )
+static void I420_Y211( filter_t *p_filter, picture_t *p_source,
+ picture_t *p_dest )
{
uint8_t *p_line1, *p_line2 = p_dest->p->p_pixels;
uint8_t *p_y1, *p_y2 = p_source->Y_PIXELS;
const int i_dest_margin = p_dest->p->i_pitch
- p_dest->p->i_visible_pitch;
- for( i_y = p_vout->render.i_height / 2 ; i_y-- ; )
+ for( i_y = p_filter->fmt_in.video.i_height / 2 ; i_y-- ; )
{
p_line1 = p_line2;
p_line2 += p_dest->p->i_pitch;
p_y1 = p_y2;
p_y2 += p_source->p[Y_PLANE].i_pitch;
- for( i_x = p_vout->render.i_width / 8 ; i_x-- ; )
+ for( i_x = p_filter->fmt_in.video.i_width / 8 ; i_x-- ; )
{
C_YUV420_Y211( );
C_YUV420_Y211( );
projects / vlc / blobdiff