- Removed the mlib and mmx yuv2rgb converters

- Added an optimized (non-scaling !) arm4l yuv2rgb (taken from bbplay)

7 files changed, 375 insertions, 1363 deletions

diff --git a/noncore/multimedia/opieplayer2/nullvideo.c b/noncore/multimedia/opieplayer2/nullvideo.c
index 79337c2..bd52869 100644
--- a/noncore/multimedia/opieplayer2/nullvideo.c
+++ b/noncore/multimedia/opieplayer2/nullvideo.c
@@ -335,26 +335,24 @@ static void null_update_frame_format( vo_driver_t* self, vo_frame_t* img,
 				    this->bytes_per_pixel );
 
 	if( format == IMGFMT_YV12 ) {
 	    frame->frame.pitches[0] = 8*((width + 7) / 8);
 	    frame->frame.pitches[1] = 8*((width + 15) / 16);
 	    frame->frame.pitches[2] = 8*((width + 15) / 16);
 	    frame->frame.base[0] = xine_xmalloc_aligned (16, frame->frame.pitches[0] * height,(void **)&frame->chunk[0]);
 	    frame->frame.base[1] = xine_xmalloc_aligned (16, frame->frame.pitches[1] * ((height+ 1)/2), (void **)&frame->chunk[1]);
 	    frame->frame.base[2] = xine_xmalloc_aligned (16, frame->frame.pitches[2] * ((height+ 1)/2), (void **)&frame->chunk[2]);
 
 	}else{
 	    frame->frame.pitches[0] = 8*((width + 3) / 4);
-	    frame->frame.pitches[1] = 8*((width + 3) / 4);
-	    frame->frame.pitches[2] = 8*((width + 3) / 4);
 	    
 	    frame->frame.base[0] = xine_xmalloc_aligned (16, frame->frame.pitches[0] * height,
                                                       (void **)&frame->chunk[0]);
 	    frame->chunk[1] = NULL;
 	    frame->chunk[2] = NULL;
 	}
 
 	frame->format = format;
 	frame->width  = width;
 	frame->height = height;
 
 	frame->stripe_height = 16 * frame->output_height / frame->height;
diff --git a/noncore/multimedia/opieplayer2/opieplayer2.pro b/noncore/multimedia/opieplayer2/opieplayer2.pro
index d8cacd0..fee9242 100644
--- a/noncore/multimedia/opieplayer2/opieplayer2.pro
+++ b/noncore/multimedia/opieplayer2/opieplayer2.pro
@@ -1,23 +1,23 @@
 TEMPLATE  = app
 CONFIG    = qt warn_on release
 #release
 DESTDIR   = $(OPIEDIR)/bin
 HEADERS   = playlistselection.h mediaplayerstate.h xinecontrol.h mediadetect.h\
         videowidget.h audiowidget.h playlistwidget.h mediaplayer.h inputDialog.h \
         frame.h lib.h xinevideowidget.h \
               alphablend.h yuv2rgb.h
 SOURCES   = main.cpp \
         playlistselection.cpp mediaplayerstate.cpp xinecontrol.cpp mediadetect.cpp\
        videowidget.cpp audiowidget.cpp playlistwidget.cpp mediaplayer.cpp inputDialog.cpp \
    frame.cpp lib.cpp nullvideo.c xinevideowidget.cpp \
-         alphablend.c yuv2rgb.c yuv2rgb_mlib.c yuv2rgb_mmx.c
+         alphablend.c yuv2rgb.c yuv2rgb_arm.c yuv2rgb_arm4l.S
 TARGET    = opieplayer2
 INCLUDEPATH += $(OPIEDIR)/include 
 DEPENDPATH  += $(OPIEDIR)/include 
 LIBS            += -lqpe -lpthread -lopie -lxine -lxineutils
 MOC_DIR=qpeobj
 OBJECTS_DIR=qpeobj
 
 INCLUDEPATH += $(OPIEDIR)/include
 DEPENDPATH  += $(OPIEDIR)/include
 
diff --git a/noncore/multimedia/opieplayer2/yuv2rgb.c b/noncore/multimedia/opieplayer2/yuv2rgb.c
index d1d6627..22bb4cb 100644
--- a/noncore/multimedia/opieplayer2/yuv2rgb.c
+++ b/noncore/multimedia/opieplayer2/yuv2rgb.c
@@ -3128,24 +3128,32 @@ yuv2rgb_factory_t* yuv2rgb_factory_init (int mode, int swapped,
       printf ("yuv2rgb: using MMX for colorspace transform\n");
   }
 #endif
 #if HAVE_MLIB
   if (this->yuv2rgb_fun == NULL) {
 
     yuv2rgb_init_mlib (this);
 
     if (this->yuv2rgb_fun != NULL)
       printf ("yuv2rgb: using medialib for colorspace transform\n");
   }
 #endif
+#ifdef __arm__
+  if (this->yuv2rgb_fun == NULL) {
+  yuv2rgb_init_arm ( this );
+  
+  if(this->yuv2rgb_fun != NULL)
+  	printf("yuv2rgb: using arm4l assembler for colorspace transform\n" );
+  }
+#endif
   if (this->yuv2rgb_fun == NULL) {
     printf ("yuv2rgb: no accelerated colorspace conversion found\n");
     yuv2rgb_c_init (this);
   }
 
   /*
    * auto-probe for the best yuy22rgb function
    */
 
   /* FIXME: implement mmx/mlib functions */
   yuy22rgb_c_init (this);
 
diff --git a/noncore/multimedia/opieplayer2/yuv2rgb_arm.c b/noncore/multimedia/opieplayer2/yuv2rgb_arm.c
new file mode 100644
index 0000000..699ee48
--- a/dev/null
+++ b/noncore/multimedia/opieplayer2/yuv2rgb_arm.c
@@ -0,0 +1,174 @@
+/*
+ * yuv2rgb_arm.c
+ * Copyright (C) 2000-2001 Project OPIE.
+ * All Rights Reserved.
+ *
+ * Author: Robert Griebl <sandman@handhelds.org>
+ *
+ * This file is part of OpiePlayer2.
+ *
+ * OpiePlayer2 is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * OpiePlayer2 is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+
+#ifdef __arm__
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <inttypes.h>
+
+#include "yuv2rgb.h"
+#include <xine/xineutils.h>
+
+#define RGB(i)							\
+	U = pu[i];						\
+	V = pv[i];						\
+	r = this->table_rV[V];					\
+	g = (void *) (((uint8_t *)this->table_gU[U]) + this->table_gV[V]);	\
+	b = this->table_bU[U];
+
+#define DST1(i)					\
+	Y = py_1[2*i];                          \
+	dst_1[2*i] = r[Y] + g[Y] + b[Y];	\
+	Y = py_1[2*i+1];			\
+	dst_1[2*i+1] = r[Y] + g[Y] + b[Y];
+
+
+struct dummy {
+	uint8_t *yuv [3];
+	int stride [3];
+};
+
+extern void convert_yuv420_rgb565(struct dummy *picture, unsigned char *results, int w, int h) ;
+
+
+static void arm_rgb16 (yuv2rgb_t *this, uint8_t * _dst,
+               uint8_t * _py, uint8_t * _pu, uint8_t * _pv)
+{
+	if ( !this-> do_scale ) {
+		struct dummy d;
+		d. yuv [0] = _py;
+		d. yuv [1] = _pu;
+		d. yuv [2] = _pv;
+		d. stride [0] = this-> y_stride;
+		d. stride [1] = d. stride [2] =  this-> uv_stride;
+
+//		printf( "calling arm (%dx%d)\n", this-> dest_width, this-> dest_height );
+
+		convert_yuv420_rgb565 ( &d, _dst, this->dest_width, this->dest_height );
+		
+//		printf ( "arm done\n" );
+    }
+    else {
+		int U, V, Y;
+		uint8_t * py_1, * py_2, * pu, * pv;
+		uint16_t * r, * g, * b;
+		uint16_t * dst_1, * dst_2;
+		int width, height, dst_height;
+		int dy;
+
+		  scale_line_func_t scale_line = this->scale_line;
+
+		  scale_line (_pu, this->u_buffer,
+		this->dest_width >> 1, this->step_dx);
+		  scale_line (_pv, this->v_buffer,
+		this->dest_width >> 1, this->step_dx);
+		  scale_line (_py, this->y_buffer, 
+		this->dest_width, this->step_dx);
+
+		  dy = 0;
+		  dst_height = this->dest_height;
+
+		  for (height = 0;; ) {
+		    dst_1 = (uint16_t*)_dst;
+		    py_1  = this->y_buffer;
+		    pu    = this->u_buffer;
+		    pv    = this->v_buffer;
+
+		    width = this->dest_width >> 3;
+
+		    do {
+		 RGB(0);
+		 DST1(0);
+
+		 RGB(1);
+		 DST1(1);
+		    
+		 RGB(2);
+		 DST1(2);
+
+		 RGB(3);
+		 DST1(3);
+
+		 pu += 4;
+		 pv += 4;
+		 py_1 += 8;
+		 dst_1 += 8;
+		    } while (--width);
+
+		    dy += this->step_dy;
+		    _dst += this->rgb_stride;
+
+		    while (--dst_height > 0 && dy < 32768) {
+
+		    xine_fast_memcpy (_dst, (uint8_t*)_dst-this->rgb_stride, this->dest_width*2); 
+
+		    dy += this->step_dy;
+		    _dst += this->rgb_stride;
+		    }
+
+		    if (dst_height <= 0)
+		    break;
+
+		    do {
+		        dy -= 32768;
+		        _py += this->y_stride;
+
+		        scale_line (_py, this->y_buffer, 
+		                    this->dest_width, this->step_dx);
+
+		        if (height & 1) {
+		            _pu += this->uv_stride;
+		            _pv += this->uv_stride;
+		 
+		            scale_line (_pu, this->u_buffer,
+		                        this->dest_width >> 1, this->step_dx);
+		            scale_line (_pv, this->v_buffer,
+		                        this->dest_width >> 1, this->step_dx);
+		 
+		        }
+		        height++;
+		    } while( dy>=32768);
+		  }
+    }
+}
+                       
+
+
+void yuv2rgb_init_arm (yuv2rgb_factory_t *this) {
+
+  if (this->swapped) 
+    return; /*no swapped pixel output upto now*/
+
+  switch (this->mode) {
+  case MODE_16_RGB:
+    this->yuv2rgb_fun = arm_rgb16;
+    break;
+  }
+}
+
+
+
+#endif
diff --git a/noncore/multimedia/opieplayer2/yuv2rgb_arm4l.S b/noncore/multimedia/opieplayer2/yuv2rgb_arm4l.S
new file mode 100644
index 0000000..f4a3395
--- a/dev/null
+++ b/noncore/multimedia/opieplayer2/yuv2rgb_arm4l.S
@@ -0,0 +1,192 @@
+/* WARNING : this function only works when stride_U == stride_V (I use some hacks to
+	     not have to do too many computations at line's end)...
+
+   C-like prototype :
+	void convert_yuv420_rgb565(AVPicture *picture, unsigned char *results, int w, int h) ;
+
+*/
+
+#ifdef __arm__
+
+	.text
+	.align
+	
+	.global convert_yuv420_rgb565
+convert_yuv420_rgb565:
+	stmdb   sp!, { r4 - r12, lr }   @ all callee saved regs
+	ldr r7,  [r0,  #0]       @ Y ptr
+	ldr r9,  [r0,  #4]       @ U ptr
+	ldr r10, [r0,  #8]       @ V ptr
+	subs r10, r10, r9        @ V ptr - U ptr
+	ldr r8,  [r0, #12]
+	add r8, r8, r7           @ Y + stride_Y
+	ldr r4,  [r0, #12]       @ Stride_Y
+	mov r4, r4, lsl #1
+	sub r4, r4, r2           @ (2 * Stride_Y) - width
+	ldr r5,  [r0, #16]       @ Stride_U
+	sub r5, r5, r2, lsr #1   @ Stride_U - (width / 2)
+	ldr r6,  [r0, #20]       @ Stride_V
+	sub r6, r6, r2, lsr #1   @ Stride_V - (width / 2)
+	add r0, r1, r2, lsl #1   @ RGB + 1
+	stmdb   sp!, { r0-r10 }
+	@ Stack description :
+	@ (sp+ 0) RGB + one line
+	@ (sp+ 4) RGB
+	@ (sp+ 8) width (save)
+	@ (sp+12) height
+	@ (sp+16) (2 * stride_Y) - width
+	@ (sp+20) stride_U - (width / 2)
+	@ (sp+24) stride_V - (width / 2) !!! UNUSED !!!
+	@ (sp+28) Y ptr
+	@ (sp+32) Y ptr + one line
+	@ (sp+36) U ptr
+	@ (sp+40) V - U
+	mov lr, r2                         @ Initialize the width counter
+	add r0, pc, #(const_storage-.-8)   @ r0 = base pointer to the constants array
+	ldr r8, [r0, #(4*4)]               @ r8 = multy	
+yuv_loop:
+	add r0, pc, #(const_storage-.-8)   @ r0 = base pointer to the constants array
+	ldr r10, [sp, #28]                 @ r10 = Y
+	ldr r1, [sp, #36]                  @ r1 = U
+	ldrb r9, [r10, #0]                 @ r9 = *Y
+	ldrb r11, [r1]                     @ r11 = *U
+	add r1, r1, #1                     @ r1 = U++
+	ldr r2, [sp, #40]                  @ r2 = V - U
+	str r1, [sp, #36]                  @ store U++
+	add r2, r1, r2                     @ r2 = V+1
+	ldrb r12, [r2, #-1]                @ r12 = *V
+	sub r11, r11, #128                 @ r11 = *U - 128
+	sub r12, r12, #128                 @ r12 = *V - 128
+	ldr r1, [r0, #(4*0)]               @ r1 = crv
+	mov r7, #32768                     @ r7 = 32768 (for additions in MLA)
+	ldr r2, [r0, #(4*3)]               @ r2 = -cgv
+	mla r6, r1, r12, r7                @ r6 = nonyc_r = crv * (*V - 128) + 32768
+	ldr r3, [r0, #(4*1)]               @ r3 = cbu
+	mla r4, r2, r12, r7                @ r4 = - cgv * (*V - 128) + 32768
+	sub r9, r9, #16                    @ r9 = *Y - 16
+	mla r5, r3, r11, r7                @ r5 = nonyc_b = cbu * (*U - 128) + 32768	
+	ldr r0, [r0, #(4*2)]               @ r0 = -cgu
+	mla r7, r8, r9, r6                 @ r7 = (*Y - 16) * multy + nonyc_r
+	add r10, r10, #2                   @ r10 = Y + 2
+	mla r4, r0, r11, r4                @ r4 = nonyc_g = - cgu * (*U - 128) + r4 = - cgu * (*U - 128) - cgv * (*V - 128) + 32768
+	add r0, pc, #(rb_clip-.-8)         @ r0 contains the pointer to the R and B clipping array
+	mla r12, r8, r9, r5                @ r12 = (*Y - 16) * multy + nonyc_b
+	ldrb r7, [r0, r7, asr #(16+3)]     @ r7 = R composant
+	mla r1, r8, r9, r4                 @ r1 = (*Y - 16) * multy + nonyc_g
+	ldrb r9, [r10, #-1]                @ r9 = *(Y+1)
+	str r10, [sp, #28]                 @ save Y + 2
+	ldrb r12, [r0, r12, asr #(16+3)]   @ r12 = B composant (and the start of the RGB word)
+	add r11, pc, #(g_clip-.-8)         @ r11 now contains the pointer to the G clipping array
+	ldrb r1, [r11, r1, asr #(16+2)]    @ r1 contains the G part of the RGB triplet
+	sub r9, r9, #16                    @ r9 = *(Y+1) - 16
+	mla r10, r8, r9, r6                @ r10 is the Red part of the RGB triplet
+	add r12, r12, r7, lsl #11          @ r12 = .GB ...
+	mla r7, r8, r9, r5                 @ r7 is the Blue part of the RGB triplet
+	add r12, r12, r1, lsl #5           @ r12 = RGB ... (ie the first pixel (half-word) is done)
+	mla r2, r8, r9, r4                 @ r2 is the Green part of the RGB triplet
+	ldrb r10, [r0, r10, asr #(16+3)]   @ r10 = R composant
+	ldrb r7, [r0, r7, asr #(16+3)]     @ r7 = B composant
+	ldr r1, [sp, #32]                  @ r1 = Ynext
+	ldrb r2, [r11, r2, asr #(16+2)]    @ r2 = G composant
+	ldrb r9, [r1]                      @ r9 = *Ynext
+	add r12, r12, r2, lsl #(5+16)      @ r12 = RGB .G.
+	sub r9, r9, #16                    @ r9 = *Ynext - 16
+	mla r2, r8, r9, r4                 @ r2 is the Green part of the RGB triplet
+	add r12, r12, r7, lsl #(0+16)      @ r12 = RGB .GB
+	mla r7, r8, r9, r5                 @ r7 is the Blue part of the RGB triplet
+	add r12, r12, r10, lsl #(11+16)    @ r12 = RGB RGB
+	ldr r3, [sp, #4]                   @ r3 = RGB
+	mla r10, r8, r9, r6                @ r10 is the Red part of the RGB triplet
+	str r12, [r3]                      @ store the rgb pixel at *RGB
+	add r3, r3, #4                     @ r3 = RGB++ (ie next double-pixel)
+	str r3, [sp, #4]                   @ store the RGB pointer
+	ldrb r9, [r1, #1]                  @ r9 = *(Ynext+1)
+	add r1, r1, #2                     @ r1 = Ynext + 2
+	sub r9, r9, #16                    @ r9 = *(Ynext+1) - 16
+	ldrb r12, [r0, r7, asr #(16+3)]    @ r12 = ..B ...
+	ldrb r10, [r0, r10, asr #(16+3)]   @ r10 = B composant
+	mla r7, r8, r9, r5                 @ r7 is the Blue part of the RGB triplet
+	add r12, r12, r10, lsl #11         @ r12 = R.B ...
+	ldrb r2, [r11, r2, asr #(16+2)]    @ r2 = G composant
+	mla r10, r8, r9, r6                @ r10 is the Red part of the RGB triplet
+	add r12, r12, r2, lsl #5           @ r12 = RGB ...
+	mla r2, r8, r9, r4                 @ r2 is the Green part of the RGB triplet
+	ldrb r7, [r0, r7, asr #(16+3)]     @ r7 = B composant
+	str r1, [sp, #32]                  @ store the increased Ynext pointer
+	add r12, r12, r7, lsl #(16+0)      @ r12 = RGB ..B
+	ldrb r10, [r0, r10, asr #(16+3)]   @ r10 = R composant
+	ldr r3, [sp, #0]                   @ r3 = RGBnext pointer
+	add r12, r12, r10, lsl #(16+11)    @ r12 = RGB R.B
+	ldrb r2, [r11, r2, asr #(16+2)]    @ r2 = G composant
+	add r3, r3, #4                     @ r3 = next pixel on the RGBnext line
+	add r12, r12, r2, lsl #(16+5)      @ r12 = RGB RGB
+	str r12, [r3, #-4]                 @ store the next pixel
+	str r3, [sp, #0]                   @ store the increased 'next line' pixel pointer
+	subs lr, lr, #2                    @ decrement the line counter
+	bne yuv_loop                       @ and restart if not at the end of the line
+
+	ldr r0, [sp, #8]                   @ r0 = saved width
+	ldr r1, [sp, #0]                   @ r1 = RGBnext pointer
+	mov lr, r0                         @ lr = saved width (to restart the line counter)
+	str r1, [sp, #4]                   @ current RGBnext pointer is next iteration RGB pointer
+	add r1, r1, r0, lsl #1             @ r1 = update RGBnext to next line
+	str r1, [sp, #0]                   @ store updated RGBnext pointer
+
+	ldr r3, [sp, #16]                  @ r3 = (2 * stride_Y) - width
+	ldr r4, [sp, #28]                  @ r4 = Y ptr
+	ldr r5, [sp, #32]                  @ r5 = Ynext ptr
+	add r4, r4, r3                     @ r4 = Y ptr for the next two lines
+	add r5, r5, r3                     @ r5 = Ynext ptr for the next two lines
+	str r4, [sp, #28]                  @ store updated Y pointer
+	str r5, [sp, #32]                  @ store update Ynext pointer
+
+	ldr r1, [sp, #20]                  @ r1 = stride_U - (width / 2)
+	ldr r2, [sp, #36]                  @ r2 = U ptr
+
+	ldr r6, [sp, #12]                  @ get height counter
+	
+	add r2, r2, r1                     @ update U ptr
+	str r2, [sp, #36]                  @ store updated U ptr (and update 'V' at the same time :-) )
+
+	subs r6, r6, #2
+	str r6, [sp, #12]
+	bne yuv_loop
+	
+	@ Exit cleanly :-)
+	add sp, sp, #(11*4)             @ remove all custom things from stack
+	ldmia   sp!, { r4 - r12, pc }   @ restore callee saved regs and return
+
+
+const_storage:
+	@ In order : crv, cbu, - cgu, - cgv, multy
+	.word 0x00019895, 0x00020469, 0xffff9bb5, 0xffff2fe1, 0x00012A15
+rb_clip_dummy:	
+        .byte 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00
+        .byte 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00
+        .byte 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00
+rb_clip:
+        .byte 0x00,0x01,0x02,0x03,0x04,0x05,0x06,0x07,0x08,0x09,0x0a,0x0b,0x0c,0x0d,0x0e,0x0f
+        .byte 0x10,0x11,0x12,0x13,0x14,0x15,0x16,0x17,0x18,0x19,0x1a,0x1b,0x1c,0x1d,0x1e,0x1f
+        .byte 0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f
+        .byte 0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f
+        .byte 0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f,0x1f
+g_clip_dummy:	
+        .byte 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00
+        .byte 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00
+        .byte 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00
+        .byte 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00
+        .byte 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00
+        .byte 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00
+g_clip:	
+        .byte 0x00,0x01,0x02,0x03,0x04,0x05,0x06,0x07,0x08,0x09,0x0a,0x0b,0x0c,0x0d,0x0e,0x0f
+        .byte 0x10,0x11,0x12,0x13,0x14,0x15,0x16,0x17,0x18,0x19,0x1a,0x1b,0x1c,0x1d,0x1e,0x1f
+        .byte 0x20,0x21,0x22,0x23,0x24,0x25,0x26,0x27,0x28,0x29,0x2a,0x2b,0x2c,0x2d,0x2e,0x2f
+        .byte 0x30,0x31,0x32,0x33,0x34,0x35,0x36,0x37,0x38,0x39,0x3a,0x3b,0x3c,0x3d,0x3e,0x3f
+        .byte 0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f
+        .byte 0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f
+        .byte 0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f
+        .byte 0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f
+        .byte 0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f
+        .byte 0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f,0x3f
+
+#endif
diff --git a/noncore/multimedia/opieplayer2/yuv2rgb_mlib.c b/noncore/multimedia/opieplayer2/yuv2rgb_mlib.c
deleted file mode 100644
index 908b439..0000000
--- a/noncore/multimedia/opieplayer2/yuv2rgb_mlib.c
+++ b/dev/null
@@ -1,313 +0,0 @@
-/*
- * yuv2rgb_mlib.c
- * Copyright (C) 2000-2001 Silicon Integrated System Corp.
- * All Rights Reserved.
- *
- * Author: Juergen Keil <jk@tools.de>
- *
- * This file is part of xine, a free unix video player.
- *
- * xine is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2 of the License, or
- * (at your option) any later version.
- *
- * xine is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
- */
-
-
-#if HAVE_MLIB
-
-#include <stdio.h>
-#include <stdlib.h>
-#include <string.h>
-#include <inttypes.h>
-#include <mlib_video.h>
-
-#include "attributes.h"
-#include "yuv2rgb.h"
-
-
-static void scale_line (uint8_t *source, uint8_t *dest,
-			int width, int step) {
-
-  unsigned p1;
-  unsigned p2;
-  int dx;
-
-  p1 = *source++;
-  p2 = *source++;
-  dx = 0;
-
-  while (width) {
-
-    /*
-    printf ("scale_line, width = %d\n", width);
-    printf ("scale_line, dx = %d, p1 = %d, p2 = %d\n", dx, p1, p2);
-    */
- 
-    *dest = (p1 * (32768 - dx) + p2 * dx) / 32768;
-
-    dx += step;
-    while (dx > 32768) {
-      dx -= 32768;
-      p1 = p2;
-      p2 = *source++;
-    }
-
-    dest ++;
-    width --;
-  }
-}
-			
-
-
-static void mlib_yuv420_rgb24 (yuv2rgb_t *this,
-			       uint8_t * image, uint8_t * py,
-			       uint8_t * pu, uint8_t * pv)
-{
-  int dst_height;
-  int dy;
-  mlib_status mlib_stat;
-
-  if (this->do_scale) {
-    dy = 0;
-    dst_height = this->dest_height;
-
-    for (;;) {
-      scale_line (pu, this->u_buffer,
-		  this->dest_width >> 1, this->step_dx);
-      pu += this->uv_stride;
-
-      scale_line (pv, this->v_buffer,
-		  this->dest_width >> 1, this->step_dx);
-      pv += this->uv_stride;
-
-      scale_line (py, this->y_buffer, 
-		  this->dest_width, this->step_dx);
-      py += this->y_stride;
-      scale_line (py, this->y_buffer + this->dest_width, 
-		  this->dest_width, this->step_dx);
-      py += this->y_stride;
-	
-      mlib_stat = mlib_VideoColorYUV2RGB420(image,
-					    this->y_buffer, 
-					    this->u_buffer,
-					    this->v_buffer,
-					    this->dest_width & ~1, 2,
-					    this->rgb_stride,
-					    this->dest_width,
-					    this->dest_width >> 1);
-      dy += this->step_dy;
-      image += this->rgb_stride;
-      
-      while (--dst_height > 0 && dy < 32768) {
-	memcpy (image, (uint8_t*)image-this->rgb_stride, this->dest_width*6);
-	dy += this->step_dy;
-	image += this->rgb_stride;
-      }
-
-      if (dst_height <= 0)
-	break;
-
-      dy -= 32768;
-
-      dy += this->step_dy;
-      image += this->rgb_stride;
-      
-      while (--dst_height > 0 && dy < 32768) {
-	memcpy (image, (uint8_t*)image-this->rgb_stride, this->dest_width*3);
-	dy += this->step_dy;
-	image += this->rgb_stride;
-      }
-
-      if (dst_height <= 0)
-	break;
-
-      dy -= 32768;
-    }
-  } else {
-    mlib_stat = mlib_VideoColorYUV2RGB420(image, py, pu, pv,
-					  this->source_width,
-					  this->source_height,
-					  this->rgb_stride,
-					  this->y_stride,
-					  this->uv_stride);
-  }
-}
-
-static void mlib_yuv420_argb32 (yuv2rgb_t *this,
-				uint8_t * image, uint8_t * py,
-				uint8_t * pu, uint8_t * pv)
-{
-  int dst_height;
-  int dy;
-  mlib_status mlib_stat;
-
-  if (this->do_scale) {
-    dy = 0;
-    dst_height = this->dest_height;
-
-    for (;;) {
-      scale_line (pu, this->u_buffer,
-		  this->dest_width >> 1, this->step_dx);
-      pu += this->uv_stride;
-
-      scale_line (pv, this->v_buffer,
-		  this->dest_width >> 1, this->step_dx);
-      pv += this->uv_stride;
-
-      scale_line (py, this->y_buffer, 
-		  this->dest_width, this->step_dx);
-      py += this->y_stride;
-      scale_line (py, this->y_buffer + this->dest_width, 
-		  this->dest_width, this->step_dx);
-      py += this->y_stride;
-	
-      mlib_stat = mlib_VideoColorYUV2ARGB420(image,
-					     this->y_buffer, 
-					     this->u_buffer,
-					     this->v_buffer,
-					     this->dest_width & ~1, 2,
-					     this->rgb_stride,
-					     this->dest_width,
-					     this->dest_width >> 1);
-      dy += this->step_dy;
-      image += this->rgb_stride;
-      
-      while (--dst_height > 0 && dy < 32768) {
-	memcpy (image, (uint8_t*)image-this->rgb_stride, this->dest_width*8);
-	dy += this->step_dy;
-	image += this->rgb_stride;
-      }
-      
-      if (dst_height <= 0)
-	break;
-
-      dy -= 32768;
-
-      dy += this->step_dy;
-      image += this->rgb_stride;
-      
-      while (--dst_height > 0 && dy < 32768) {
-	memcpy (image, (uint8_t*)image-this->rgb_stride, this->dest_width*4);
-	dy += this->step_dy;
-	image += this->rgb_stride;
-      }
-
-      if (dst_height <= 0)
-	break;
-
-      dy -= 32768;
-    }
-  } else {
-    mlib_stat = mlib_VideoColorYUV2ARGB420(image, py, pu, pv,
-					   this->source_width,
-					   this->source_height,
-					   this->rgb_stride,
-					   this->y_stride,
-					   this->uv_stride);
-  }
-}
-
-static void mlib_yuv420_abgr32 (yuv2rgb_t *this,
-				uint8_t * image, uint8_t * py,
-				uint8_t * pu, uint8_t * pv)
-{
-  int dst_height;
-  int dy;
-  mlib_status mlib_stat;
-
-  if (this->do_scale) {
-    dy = 0;
-    dst_height = this->dest_height;
-
-    for (;;) {
-      scale_line (pu, this->u_buffer,
-		  this->dest_width >> 1, this->step_dx);
-      pu += this->uv_stride;
-
-      scale_line (pv, this->v_buffer,
-		  this->dest_width >> 1, this->step_dx);
-      pv += this->uv_stride;
-
-      scale_line (py, this->y_buffer, 
-		  this->dest_width, this->step_dx);
-      py += this->y_stride;
-      scale_line (py, this->y_buffer + this->dest_width, 
-		  this->dest_width, this->step_dx);
-      py += this->y_stride;
-	
-      mlib_stat = mlib_VideoColorYUV2ABGR420(image,
-					     this->y_buffer, 
-					     this->u_buffer,
-					     this->v_buffer,
-					     this->dest_width & ~1, 2,
-					     this->rgb_stride,
-					     this->dest_width,
-					     this->dest_width >> 1);
-      dy += this->step_dy;
-      image += this->rgb_stride;
-      
-      while (--dst_height > 0 && dy < 32768) {
-	memcpy (image, (uint8_t*)image-this->rgb_stride, this->dest_width*8);
-	dy += this->step_dy;
-	image += this->rgb_stride;
-      }
-
-      if (dst_height <= 0)
-	break;
-
-      dy -= 32768;
-
-      dy += this->step_dy;
-      image += this->rgb_stride;
-      
-      while (--dst_height > 0 && dy < 32768) {
-	memcpy (image, (uint8_t*)image-this->rgb_stride, this->dest_width*4);
-	dy += this->step_dy;
-	image += this->rgb_stride;
-      }
-
-      if (dst_height <= 0)
-	break;
-
-      dy -= 32768;
-    }
-  } else {
-    mlib_stat = mlib_VideoColorYUV2ABGR420(image, py, pu, pv,
-					   this->source_width,
-					   this->source_height,
-					   this->rgb_stride,
-					   this->y_stride,
-					   this->uv_stride);
-  }
-}
-
-
-void yuv2rgb_init_mlib (yuv2rgb_factory_t *this) {
-
-  if (this->swapped) return; /*no swapped pixel output upto now*/
-
-  switch (this->mode) {
-  case MODE_24_RGB:
-    this->yuv2rgb_fun = mlib_yuv420_rgb24;
-    break;
-  case MODE_32_RGB:
-    this->yuv2rgb_fun = mlib_yuv420_argb32;
-    break;
-  case MODE_32_BGR:
-    this->yuv2rgb_fun = mlib_yuv420_abgr32;
-    break;
-  }
-}
-
-
-#endif	/* HAVE_MLIB */
diff --git a/noncore/multimedia/opieplayer2/yuv2rgb_mmx.c b/noncore/multimedia/opieplayer2/yuv2rgb_mmx.c
deleted file mode 100644
index f092e6f..0000000
--- a/noncore/multimedia/opieplayer2/yuv2rgb_mmx.c
+++ b/dev/null
@@ -1,1047 +0,0 @@
-/*
- * yuv2rgb_mmx.c
- * Copyright (C) 2000-2001 Silicon Integrated System Corp.
- * All Rights Reserved.
- *
- * Author: Olie Lho <ollie@sis.com.tw>
- *
- * This file is part of mpeg2dec, a free MPEG-2 video stream decoder.
- *
- * mpeg2dec is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2 of the License, or
- * (at your option) any later version.
- *
- * mpeg2dec is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
- */
-
-
-#ifdef ARCH_X86
-
-#include <stdio.h>
-#include <stdlib.h>
-#include <string.h>
-#include <inttypes.h>
-
-#include "yuv2rgb.h"
-#include "xineutils.h"
-
-#define CPU_MMXEXT 0
-#define CPU_MMX 1
-
-/* CPU_MMXEXT/CPU_MMX adaptation layer */
-
-#define movntq(src,dest)	\
-do {				\
-    if (cpu == CPU_MMXEXT)	\
-	movntq_r2m (src, dest);	\
-    else			\
-	movq_r2m (src, dest);	\
-} while (0)
-
-static mmx_t mmx_subYw = {0x1010101010101010};
-static mmx_t mmx_addYw = {0x0000000000000000};
-
-void mmx_yuv2rgb_set_gamma(int gamma) 
-{
-int a,s,i;
-
-  if( gamma <= 16 ) {
-    a = 0;
-    s = 16 - gamma;
-  } else {
-    a = gamma - 16;
-    s = 0;
-  }
-  
-  for( i = 0; i < 8; i++ ) {
-    *((unsigned char *)&mmx_subYw + i) = s;
-    *((unsigned char *)&mmx_addYw + i) = a;
-  }
-}
-
-static inline void mmx_yuv2rgb (uint8_t * py, uint8_t * pu, uint8_t * pv)
-{
-    static mmx_t mmx_80w = {0x0080008000800080};
-    static mmx_t mmx_U_green = {0xf37df37df37df37d};
-    static mmx_t mmx_U_blue = {0x4093409340934093};
-    static mmx_t mmx_V_red = {0x3312331233123312};
-    static mmx_t mmx_V_green = {0xe5fce5fce5fce5fc};
-    static mmx_t mmx_00ffw = {0x00ff00ff00ff00ff};
-    static mmx_t mmx_Y_coeff = {0x253f253f253f253f};
-
-    movq_m2r (*py, mm6);		// mm6 = Y7 Y6 Y5 Y4 Y3 Y2 Y1 Y0
-    pxor_r2r (mm4, mm4);		// mm4 = 0
-
-    psubusb_m2r (mmx_subYw, mm6);	// Y -= 16
-    paddusb_m2r (mmx_addYw, mm6);
-
-    movd_m2r (*pu, mm0);		// mm0 = 00 00 00 00 u3 u2 u1 u0
-    movq_r2r (mm6, mm7);		// mm7 = Y7 Y6 Y5 Y4 Y3 Y2 Y1 Y0
-
-    pand_m2r (mmx_00ffw, mm6);		// mm6 =    Y6    Y4    Y2    Y0
-    psrlw_i2r (8, mm7);			// mm7 =    Y7    Y5    Y3    Y1
-
-    movd_m2r (*pv, mm1);		// mm1 = 00 00 00 00 v3 v2 v1 v0
-    psllw_i2r (3, mm6);			// promote precision
-
-    pmulhw_m2r (mmx_Y_coeff, mm6);	// mm6 = luma_rgb even
-    psllw_i2r (3, mm7);			// promote precision
-
-    punpcklbw_r2r (mm4, mm0);		// mm0 = u3 u2 u1 u0
-
-    psubsw_m2r (mmx_80w, mm0);		// u -= 128
-    punpcklbw_r2r (mm4, mm1);		// mm1 = v3 v2 v1 v0
-
-    pmulhw_m2r (mmx_Y_coeff, mm7);	// mm7 = luma_rgb odd
-    psllw_i2r (3, mm0);			// promote precision
-
-    psubsw_m2r (mmx_80w, mm1);		// v -= 128
-    movq_r2r (mm0, mm2);		// mm2 = u3 u2 u1 u0
-
-    psllw_i2r (3, mm1);			// promote precision
-
-    movq_r2r (mm1, mm4);		// mm4 = v3 v2 v1 v0
-
-    pmulhw_m2r (mmx_U_blue, mm0);	// mm0 = chroma_b
-
-
-    // slot
-
-
-    // slot
-
-
-    pmulhw_m2r (mmx_V_red, mm1);	// mm1 = chroma_r
-    movq_r2r (mm0, mm3);		// mm3 = chroma_b
-
-    paddsw_r2r (mm6, mm0);		// mm0 = B6 B4 B2 B0
-    paddsw_r2r (mm7, mm3);		// mm3 = B7 B5 B3 B1
-
-    packuswb_r2r (mm0, mm0);		// saturate to 0-255
-
-
-    pmulhw_m2r (mmx_U_green, mm2);	// mm2 = u * u_green
-
-
-    packuswb_r2r (mm3, mm3);		// saturate to 0-255
-
-
-    punpcklbw_r2r (mm3, mm0);		// mm0 = B7 B6 B5 B4 B3 B2 B1 B0
-
-
-    pmulhw_m2r (mmx_V_green, mm4);	// mm4 = v * v_green
-
-    
-    // slot
-
-
-    // slot
-
-
-    paddsw_r2r (mm4, mm2);		// mm2 = chroma_g
-    movq_r2r (mm2, mm5);		// mm5 = chroma_g
-
-
-    movq_r2r (mm1, mm4);		// mm4 = chroma_r
-    paddsw_r2r (mm6, mm2);		// mm2 = G6 G4 G2 G0
-
-
-    packuswb_r2r (mm2, mm2);		// saturate to 0-255
-    paddsw_r2r (mm6, mm1);		// mm1 = R6 R4 R2 R0
-
-    packuswb_r2r (mm1, mm1);		// saturate to 0-255
-    paddsw_r2r (mm7, mm4);		// mm4 = R7 R5 R3 R1
-
-    packuswb_r2r (mm4, mm4);		// saturate to 0-255
-    paddsw_r2r (mm7, mm5);		// mm5 = G7 G5 G3 G1
-
-
-    packuswb_r2r (mm5, mm5);		// saturate to 0-255
-
-
-    punpcklbw_r2r (mm4, mm1);		// mm1 = R7 R6 R5 R4 R3 R2 R1 R0
-
-
-    punpcklbw_r2r (mm5, mm2);		// mm2 = G7 G6 G5 G4 G3 G2 G1 G0
-}
-
-// basic opt
-static inline void mmx_unpack_16rgb (uint8_t * image, int cpu)
-{
-    static mmx_t mmx_bluemask = {0xf8f8f8f8f8f8f8f8};
-    static mmx_t mmx_greenmask = {0xfcfcfcfcfcfcfcfc};
-    static mmx_t mmx_redmask = {0xf8f8f8f8f8f8f8f8};
-
-    /*
-     * convert RGB plane to RGB 16 bits
-     * mm0 -> B, mm1 -> R, mm2 -> G
-     * mm4 -> GB, mm5 -> AR pixel 4-7
-     * mm6 -> GB, mm7 -> AR pixel 0-3
-     */
-
-    pand_m2r (mmx_bluemask, mm0);	// mm0 = b7b6b5b4b3______
-    pxor_r2r (mm4, mm4);		// mm4 = 0
-
-    pand_m2r (mmx_greenmask, mm2);	// mm2 = g7g6g5g4g3g2____
-    psrlq_i2r (3, mm0);			// mm0 = ______b7b6b5b4b3
-
-    movq_r2r (mm2, mm7);		// mm7 = g7g6g5g4g3g2____
-    movq_r2r (mm0, mm5);		// mm5 = ______b7b6b5b4b3
-
-    pand_m2r (mmx_redmask, mm1);	// mm1 = r7r6r5r4r3______
-    punpcklbw_r2r (mm4, mm2);
-
-    punpcklbw_r2r (mm1, mm0);
-
-    psllq_i2r (3, mm2);
-
-    punpckhbw_r2r (mm4, mm7);
-    por_r2r (mm2, mm0);
-
-    psllq_i2r (3, mm7);
-
-    movntq (mm0, *image);
-    punpckhbw_r2r (mm1, mm5);
-
-    por_r2r (mm7, mm5);
-
-    // U
-    // V
-
-    movntq (mm5, *(image+8));
-}
-
-static inline void mmx_unpack_15rgb (uint8_t * image, int cpu)
-{
-    static mmx_t mmx_bluemask = {0xf8f8f8f8f8f8f8f8};
-    static mmx_t mmx_greenmask = {0xf8f8f8f8f8f8f8f8};
-    static mmx_t mmx_redmask = {0xf8f8f8f8f8f8f8f8};
-
-    /*
-     * convert RGB plane to RGB 15 bits
-     * mm0 -> B, mm1 -> R, mm2 -> G
-     * mm4 -> GB, mm5 -> AR pixel 4-7
-     * mm6 -> GB, mm7 -> AR pixel 0-3
-     */
-
-    pand_m2r (mmx_bluemask, mm0);	// mm0 = b7b6b5b4b3______
-    pxor_r2r (mm4, mm4);		// mm4 = 0
-
-    pand_m2r (mmx_greenmask, mm2);	// mm2 = g7g6g5g4g3g2____
-    psrlq_i2r (3, mm0);			// mm0 = ______b7b6b5b4b3
-
-    movq_r2r (mm2, mm7);		// mm7 = g7g6g5g4g3g2____
-    movq_r2r (mm0, mm5);		// mm5 = ______b7b6b5b4b3
-
-    pand_m2r (mmx_redmask, mm1);	// mm1 = r7r6r5r4r3______
-    punpcklbw_r2r (mm4, mm2);
-
-    psrlq_i2r (1, mm1);
-    punpcklbw_r2r (mm1, mm0);
-
-    psllq_i2r (2, mm2);
-
-    punpckhbw_r2r (mm4, mm7);
-    por_r2r (mm2, mm0);
-
-    psllq_i2r (2, mm7);
-
-    movntq (mm0, *image);
-    punpckhbw_r2r (mm1, mm5);
-
-    por_r2r (mm7, mm5);
-
-    // U
-    // V
-
-    movntq (mm5, *(image+8));
-}
-
-static inline void mmx_unpack_32rgb (uint8_t * image, int cpu)
-{
-    /*
-     * convert RGB plane to RGB packed format,
-     * mm0 -> B, mm1 -> R, mm2 -> G, mm3 -> 0,
-     * mm4 -> GB, mm5 -> AR pixel 4-7,
-     * mm6 -> GB, mm7 -> AR pixel 0-3
-     */
-
-    pxor_r2r (mm3, mm3);
-    movq_r2r (mm0, mm6);
-
-    punpcklbw_r2r (mm2, mm6);
-    movq_r2r (mm1, mm7);
-
-    punpcklbw_r2r (mm3, mm7);
-    movq_r2r (mm0, mm4);
-
-    punpcklwd_r2r (mm7, mm6);
-    movq_r2r (mm1, mm5);
-
-    /* scheduling: this is hopeless */
-    movntq (mm6, *image);
-    movq_r2r (mm0, mm6);
-    punpcklbw_r2r (mm2, mm6);
-    punpckhwd_r2r (mm7, mm6);
-    movntq (mm6, *(image+8));
-    punpckhbw_r2r (mm2, mm4);
-    punpckhbw_r2r (mm3, mm5);
-    punpcklwd_r2r (mm5, mm4);
-    movntq (mm4, *(image+16));
-    movq_r2r (mm0, mm4);
-    punpckhbw_r2r (mm2, mm4);
-    punpckhwd_r2r (mm5, mm4);
-    movntq (mm4, *(image+24));
-}
-
-static inline void mmx_unpack_32bgr (uint8_t * image, int cpu)
-{
-    /*
-     * convert RGB plane to RGB packed format,
-     * mm0 -> B, mm1 -> R, mm2 -> G, mm3 -> 0,
-     * mm4 -> GB, mm5 -> AR pixel 4-7,
-     * mm6 -> GB, mm7 -> AR pixel 0-3
-     */
-
-    pxor_r2r (mm3, mm3);
-    movq_r2r (mm1, mm6);
-
-    punpcklbw_r2r (mm2, mm6);
-    movq_r2r (mm0, mm7);
-
-    punpcklbw_r2r (mm3, mm7);
-    movq_r2r (mm1, mm4);
-
-    punpcklwd_r2r (mm7, mm6);
-    movq_r2r (mm0, mm5);
-
-    /* scheduling: this is hopeless */
-    movntq (mm6, *image);
-    movq_r2r (mm0, mm6);
-    punpcklbw_r2r (mm2, mm6);
-    punpckhwd_r2r (mm7, mm6);
-    movntq (mm6, *(image+8));
-    punpckhbw_r2r (mm2, mm4);
-    punpckhbw_r2r (mm3, mm5);
-    punpcklwd_r2r (mm5, mm4);
-    movntq (mm4, *(image+16));
-    movq_r2r (mm0, mm4);
-    punpckhbw_r2r (mm2, mm4);
-    punpckhwd_r2r (mm5, mm4);
-    movntq (mm4, *(image+24));
-}
-
-static inline void mmx_unpack_24rgb (uint8_t * image, int cpu)
-{
-    /*
-     * convert RGB plane to RGB packed format,
-     * mm0 -> B, mm1 -> R, mm2 -> G, mm3 -> 0,
-     * mm4 -> GB, mm5 -> AR pixel 4-7,
-     * mm6 -> GB, mm7 -> AR pixel 0-3
-     */
-
-    pxor_r2r (mm3, mm3);
-    movq_r2r (mm0, mm6);
-
-    punpcklbw_r2r (mm2, mm6);
-    movq_r2r (mm1, mm7);
-
-    punpcklbw_r2r (mm3, mm7);
-    movq_r2r (mm0, mm4);
-
-    punpcklwd_r2r (mm7, mm6);
-    movq_r2r (mm1, mm5);
-
-    /* scheduling: this is hopeless */
-    movntq (mm6, *image);
-    movq_r2r (mm0, mm6);
-    punpcklbw_r2r (mm2, mm6);
-    punpckhwd_r2r (mm7, mm6);
-    movntq (mm6, *(image+8));
-    punpckhbw_r2r (mm2, mm4);
-    punpckhbw_r2r (mm3, mm5);
-    punpcklwd_r2r (mm5, mm4);
-    movntq (mm4, *(image+16));
-}
-
-static inline void yuv420_rgb16 (yuv2rgb_t *this,
-				 uint8_t * image,
-				 uint8_t * py, uint8_t * pu, uint8_t * pv,
-				 int cpu)
-{
-    int i;
-    int rgb_stride = this->rgb_stride;
-    int y_stride   = this->y_stride;
-    int uv_stride  = this->uv_stride;
-    int width      = this->source_width;
-    int height     = this->source_height;
-    int dst_height = this->dest_height;
-    uint8_t *img;
-
-    width >>= 3;
-
-    if (!this->do_scale) {
-      y_stride -= 8 * width;
-      uv_stride -= 4 * width;
-
-      do {
-
-	i = width; img = image;
-	do {
-	  mmx_yuv2rgb (py, pu, pv); 
-	  mmx_unpack_16rgb (img, cpu); 
-	  py += 8;
-	  pu += 4;
-	  pv += 4;
-	  img += 16;
-	} while (--i);
-	
-	py += y_stride;
-	image += rgb_stride;
-	if (height & 1) {
-	  pu += uv_stride;
-	  pv += uv_stride;
-	} else {
-	  pu -= 4 * width;
-	  pv -= 4 * width;
-	}
-      } while (--height);
-
-    } else {
-
-      scale_line_func_t scale_line = this->scale_line;
-      uint8_t *y_buf, *u_buf, *v_buf;
-      int      dy = 0;
-
-      scale_line (pu, this->u_buffer,
-		  this->dest_width >> 1, this->step_dx);
-      scale_line (pv, this->v_buffer,
-		  this->dest_width >> 1, this->step_dx);
-      scale_line (py, this->y_buffer, 
-		  this->dest_width, this->step_dx);
-      for (height = 0;; ) {
-
-	y_buf = this->y_buffer;
-	u_buf = this->u_buffer;
-	v_buf = this->v_buffer;
-
-	i = this->dest_width >> 3; img = image;
-	do {
-	  /* printf ("i : %d\n",i); */
-
-	  mmx_yuv2rgb (y_buf, u_buf, v_buf); 
-	  mmx_unpack_16rgb (img, cpu); 
-	  y_buf += 8;
-	  u_buf += 4;
-	  v_buf += 4;
-	  img += 16;
-	} while (--i);
-	
-	dy += this->step_dy;
-	image += rgb_stride;
-
-	while (--dst_height > 0 && dy < 32768) {
-
-	  xine_fast_memcpy (image, image-rgb_stride, this->dest_width*2); 
-
-	  dy += this->step_dy;
-	  image += rgb_stride;
-	}
-
-	if (dst_height <= 0)
-	  break;
-
-        do {
-            dy -= 32768;
-
-            py += y_stride;
-
-            scale_line (py, this->y_buffer, 
-                        this->dest_width, this->step_dx);
-
-            if (height & 1) {
-                pu += uv_stride;
-                pv += uv_stride;
-	  
-                scale_line (pu, this->u_buffer,
-                            this->dest_width >> 1, this->step_dx);
-                scale_line (pv, this->v_buffer,
-                            this->dest_width >> 1, this->step_dx);
-	  
-            }
-            height++;
-        } while( dy>=32768);
-      }
-    } 
-}
-
-static inline void yuv420_rgb15 (yuv2rgb_t *this,
-				 uint8_t * image,
-				 uint8_t * py, uint8_t * pu, uint8_t * pv,
-				 int cpu)
-{
-    int i;
-    int rgb_stride = this->rgb_stride;
-    int y_stride   = this->y_stride;
-    int uv_stride  = this->uv_stride;
-    int width      = this->source_width;
-    int height     = this->source_height;
-    int dst_height = this->dest_height;
-    uint8_t *img;
-
-    width >>= 3;
-
-    if (!this->do_scale) {
-      y_stride -= 8 * width;
-      uv_stride -= 4 * width;
-
-      do {
-
-	i = width; img = image;
-	do {
-	  mmx_yuv2rgb (py, pu, pv); 
-	  mmx_unpack_15rgb (img, cpu); 
-	  py += 8;
-	  pu += 4;
-	  pv += 4;
-	  img += 16;
-	} while (--i);
-	
-	py += y_stride;
-	image += rgb_stride;
-	if (height & 1) {
-	  pu += uv_stride;
-	  pv += uv_stride;
-	} else {
-	  pu -= 4 * width;
-	  pv -= 4 * width;
-	}
-      } while (--height);
-
-    } else {
-
-      scale_line_func_t scale_line = this->scale_line;
-      uint8_t *y_buf, *u_buf, *v_buf;
-      int      dy = 0;
-
-      scale_line (pu, this->u_buffer,
-		  this->dest_width >> 1, this->step_dx);
-      scale_line (pv, this->v_buffer,
-		  this->dest_width >> 1, this->step_dx);
-      scale_line (py, this->y_buffer, 
-		  this->dest_width, this->step_dx);
-      for (height = 0;; ) {
-
-	y_buf = this->y_buffer;
-	u_buf = this->u_buffer;
-	v_buf = this->v_buffer;
-
-	i = this->dest_width >> 3; img = image;
-	do {
-	  /* printf ("i : %d\n",i); */
-
-	  mmx_yuv2rgb (y_buf, u_buf, v_buf); 
-	  mmx_unpack_15rgb (img, cpu); 
-	  y_buf += 8;
-	  u_buf += 4;
-	  v_buf += 4;
-	  img += 16;
-	} while (--i);
-	
-	dy += this->step_dy;
-	image += rgb_stride;
-
-	while (--dst_height > 0 && dy < 32768) {
-
-	  xine_fast_memcpy (image, image-rgb_stride, this->dest_width*2); 
-
-	  dy += this->step_dy;
-	  image += rgb_stride;
-	}
-
-	if (dst_height <= 0)
-	  break;
-
-        do {
-            dy -= 32768;
-            py += y_stride;
-
-            scale_line (py, this->y_buffer, 
-                        this->dest_width, this->step_dx);
-
-            if (height & 1) {
-                pu += uv_stride;
-                pv += uv_stride;
-	  
-                scale_line (pu, this->u_buffer,
-                            this->dest_width >> 1, this->step_dx);
-                scale_line (pv, this->v_buffer,
-                            this->dest_width >> 1, this->step_dx);
-	  
-            }
-            height++;
-        } while( dy>=32768 );
-      }
-    } 
-}
-
-static inline void yuv420_rgb24 (yuv2rgb_t *this,
-				 uint8_t * image, uint8_t * py,
-				 uint8_t * pu, uint8_t * pv, int cpu)
-{
-    int i;
-    int rgb_stride = this->rgb_stride;
-    int y_stride   = this->y_stride;
-    int uv_stride  = this->uv_stride;
-    int width      = this->source_width;
-    int height     = this->source_height;
-    int dst_height = this->dest_height;
-    uint8_t *img;
-
-    /* rgb_stride -= 4 * this->dest_width; */
-    width >>= 3;
-
-    if (!this->do_scale) {
-      y_stride -= 8 * width;
-      uv_stride -= 4 * width;
-
-      do {
-	i = width; img = image;
-	do {
-	  mmx_yuv2rgb (py, pu, pv);
-	  mmx_unpack_24rgb (img, cpu);
-	  py += 8;
-	  pu += 4;
-	  pv += 4;
-	  img += 24;
-	} while (--i);
-
-	py += y_stride;
-	image += rgb_stride;
-	if (height & 1) {
-	  pu += uv_stride;
-	  pv += uv_stride;
-	} else {
-	  pu -= 4 * width;
-	  pv -= 4 * width;
-	}
-      } while (--height);
-    } else {
-
-      scale_line_func_t scale_line = this->scale_line;
-      uint8_t *y_buf, *u_buf, *v_buf;
-      int      dy = 0;
-
-      scale_line (pu, this->u_buffer,
-		  this->dest_width >> 1, this->step_dx);
-      scale_line (pv, this->v_buffer,
-		  this->dest_width >> 1, this->step_dx);
-      scale_line (py, this->y_buffer, 
-		  this->dest_width, this->step_dx);
-
-      for (height = 0;; ) {
-
-	y_buf = this->y_buffer;
-	u_buf = this->u_buffer;
-	v_buf = this->v_buffer;
-
-
-	i = this->dest_width >> 3; img=image;
-	do {
-	  /* printf ("i : %d\n",i); */
-
-	  mmx_yuv2rgb (y_buf, u_buf, v_buf); 
-	  mmx_unpack_24rgb (img, cpu); 
-	  y_buf += 8;
-	  u_buf += 4;
-	  v_buf += 4;
-	  img += 24;
-	} while (--i);
-	
-	dy += this->step_dy;
-	image += rgb_stride;
-
-	while (--dst_height > 0 && dy < 32768) {
-
-	  xine_fast_memcpy (image, image-rgb_stride, this->dest_width*3);
-
-	  dy += this->step_dy;
-	  image += rgb_stride;
-	}
-
-	if (dst_height <= 0)
-	  break;
-
-        do {
-            dy -= 32768;
-            py += y_stride;
-	
-            scale_line (py, this->y_buffer, 
-                        this->dest_width, this->step_dx);
-
-            if (height & 1) {
-                pu += uv_stride;
-                pv += uv_stride;
-	  
-                scale_line (pu, this->u_buffer,
-                            this->dest_width >> 1, this->step_dx);
-                scale_line (pv, this->v_buffer,
-                            this->dest_width >> 1, this->step_dx);
-            }
-            height++;
-        } while( dy>=32768 );
-
-      }
-      
-    }
-}
-
-static inline void yuv420_argb32 (yuv2rgb_t *this,
-				  uint8_t * image, uint8_t * py,
-				  uint8_t * pu, uint8_t * pv, int cpu)
-{
-    int i;
-    int rgb_stride = this->rgb_stride;
-    int y_stride   = this->y_stride;
-    int uv_stride  = this->uv_stride;
-    int width      = this->source_width;
-    int height     = this->source_height;
-    int dst_height = this->dest_height;
-    uint8_t *img;
-
-    /* rgb_stride -= 4 * this->dest_width; */
-    width >>= 3;
-
-    if (!this->do_scale) {
-      y_stride -= 8 * width;
-      uv_stride -= 4 * width;
-
-      do {
-	i = width; img = image;
-	do {
-	  mmx_yuv2rgb (py, pu, pv);
-	  mmx_unpack_32rgb (img, cpu);
-	  py += 8;
-	  pu += 4;
-	  pv += 4;
-	  img += 32;
-	} while (--i);
-
-	py += y_stride;
-	image += rgb_stride;
-	if (height & 1) {
-	  pu += uv_stride;
-	  pv += uv_stride;
-	} else {
-	  pu -= 4 * width;
-	  pv -= 4 * width;
-	}
-      } while (--height);
-    } else {
-
-      scale_line_func_t scale_line = this->scale_line;
-      uint8_t *y_buf, *u_buf, *v_buf;
-      int      dy = 0;
-
-      scale_line (pu, this->u_buffer,
-		  this->dest_width >> 1, this->step_dx);
-      scale_line (pv, this->v_buffer,
-		  this->dest_width >> 1, this->step_dx);
-      scale_line (py, this->y_buffer, 
-		  this->dest_width, this->step_dx);
-
-      for (height = 0;; ) {
-
-	y_buf = this->y_buffer;
-	u_buf = this->u_buffer;
-	v_buf = this->v_buffer;
-
-
-	i = this->dest_width >> 3; img=image;
-	do {
-	  /* printf ("i : %d\n",i); */
-
-	  mmx_yuv2rgb (y_buf, u_buf, v_buf); 
-	  mmx_unpack_32rgb (img, cpu); 
-	  y_buf += 8;
-	  u_buf += 4;
-	  v_buf += 4;
-	  img += 32;
-	} while (--i);
-	
-	dy += this->step_dy;
-	image += rgb_stride;
-
-	while (--dst_height > 0 && dy < 32768) {
-
-	  xine_fast_memcpy (image, image-rgb_stride, this->dest_width*4); 
-
-	  dy += this->step_dy;
-	  image += rgb_stride;
-	}
-
-	if (dst_height <= 0)
-	  break;
-
-        do {
-            dy -= 32768;
-            py += y_stride;
-	
-            scale_line (py, this->y_buffer, 
-                        this->dest_width, this->step_dx);
-
-            if (height & 1) {
-                pu += uv_stride;
-                pv += uv_stride;
-	  
-                scale_line (pu, this->u_buffer,
-                            this->dest_width >> 1, this->step_dx);
-                scale_line (pv, this->v_buffer,
-                            this->dest_width >> 1, this->step_dx);
-            }
-            height++;
-        } while( dy>=32768 );
-      }
-      
-    }
-}
-
-static inline void yuv420_abgr32 (yuv2rgb_t *this,
-				  uint8_t * image, uint8_t * py,
-				  uint8_t * pu, uint8_t * pv, int cpu)
-{
-    int i;
-    int rgb_stride = this->rgb_stride;
-    int y_stride   = this->y_stride;
-    int uv_stride  = this->uv_stride;
-    int width      = this->source_width;
-    int height     = this->source_height;
-    int dst_height = this->dest_height;
-    uint8_t *img;
-
-    /* rgb_stride -= 4 * this->dest_width; */
-    width >>= 3;
-
-    if (!this->do_scale) {
-      y_stride -= 8 * width;
-      uv_stride -= 4 * width;
-
-      do {
-	i = width; img = image;
-	do {
-	  mmx_yuv2rgb (py, pu, pv);
-	  mmx_unpack_32bgr (img, cpu);
-	  py += 8;
-	  pu += 4;
-	  pv += 4;
-	  img += 32;
-	} while (--i);
-
-	py += y_stride;
-	image += rgb_stride;
-	if (height & 1) {
-	  pu += uv_stride;
-	  pv += uv_stride;
-	} else {
-	  pu -= 4 * width;
-	  pv -= 4 * width;
-	}
-      } while (--height);
-    } else {
-
-      scale_line_func_t scale_line = this->scale_line;
-      uint8_t *y_buf, *u_buf, *v_buf;
-      int      dy = 0;
-
-      scale_line (pu, this->u_buffer,
-		  this->dest_width >> 1, this->step_dx);
-      scale_line (pv, this->v_buffer,
-		  this->dest_width >> 1, this->step_dx);
-      scale_line (py, this->y_buffer, 
-		  this->dest_width, this->step_dx);
-
-      for (height = 0;; ) {
-
-	y_buf = this->y_buffer;
-	u_buf = this->u_buffer;
-	v_buf = this->v_buffer;
-
-
-	i = this->dest_width >> 3; img=image;
-	do {
-	  /* printf ("i : %d\n",i); */
-
-	  mmx_yuv2rgb (y_buf, u_buf, v_buf); 
-	  mmx_unpack_32bgr (img, cpu); 
-	  y_buf += 8;
-	  u_buf += 4;
-	  v_buf += 4;
-	  img += 32;
-	} while (--i);
-	
-	dy += this->step_dy;
-	image += rgb_stride;
-
-	while (--dst_height > 0 && dy < 32768) {
-
-	  xine_fast_memcpy (image, image-rgb_stride, this->dest_width*4); 
-
-	  dy += this->step_dy;
-	  image += rgb_stride;
-	}
-
-	if (dst_height <= 0)
-	  break;
-
-        do {
-            dy -= 32768;
-            py += y_stride;
-	
-            scale_line (py, this->y_buffer, 
-                        this->dest_width, this->step_dx);
-
-            if (height & 1) {
-                pu += uv_stride;
-                pv += uv_stride;
-	  
-                scale_line (pu, this->u_buffer,
-                            this->dest_width >> 1, this->step_dx);
-                scale_line (pv, this->v_buffer,
-                            this->dest_width >> 1, this->step_dx);
-            }
-            height++;
-        } while( dy>=32768 );
-
-      }
-      
-    }
-}
-
-static void mmxext_rgb15 (yuv2rgb_t *this, uint8_t * image,
-			  uint8_t * py, uint8_t * pu, uint8_t * pv)
-{
-    yuv420_rgb15 (this, image, py, pu, pv, CPU_MMXEXT);
-    emms();	/* re-initialize x86 FPU after MMX use */
-}
-
-static void mmxext_rgb16 (yuv2rgb_t *this, uint8_t * image,
-			  uint8_t * py, uint8_t * pu, uint8_t * pv)
-{
-    yuv420_rgb16 (this, image, py, pu, pv, CPU_MMXEXT);
-    emms();	/* re-initialize x86 FPU after MMX use */
-}
-
-static void mmxext_rgb24 (yuv2rgb_t *this, uint8_t * image,
-			   uint8_t * py, uint8_t * pu, uint8_t * pv)
-{
-    yuv420_rgb24 (this, image, py, pu, pv, CPU_MMXEXT);
-    emms();	/* re-initialize x86 FPU after MMX use */
-}
-
-static void mmxext_argb32 (yuv2rgb_t *this, uint8_t * image,
-			   uint8_t * py, uint8_t * pu, uint8_t * pv)
-{
-    yuv420_argb32 (this, image, py, pu, pv, CPU_MMXEXT);
-    emms();	/* re-initialize x86 FPU after MMX use */
-}
-
-static void mmxext_abgr32 (yuv2rgb_t *this, uint8_t * image,
-			   uint8_t * py, uint8_t * pu, uint8_t * pv)
-{
-    yuv420_abgr32 (this, image, py, pu, pv, CPU_MMXEXT);
-    emms();	/* re-initialize x86 FPU after MMX use */
-}
-
-static void mmx_rgb15 (yuv2rgb_t *this, uint8_t * image,
-		       uint8_t * py, uint8_t * pu, uint8_t * pv)
-{
-    yuv420_rgb15 (this, image, py, pu, pv, CPU_MMX);
-    emms();	/* re-initialize x86 FPU after MMX use */
-}
-
-static void mmx_rgb16 (yuv2rgb_t *this, uint8_t * image,
-		       uint8_t * py, uint8_t * pu, uint8_t * pv)
-{
-    yuv420_rgb16 (this, image, py, pu, pv, CPU_MMX);
-    emms();	/* re-initialize x86 FPU after MMX use */
-}
-
-static void mmx_rgb24 (yuv2rgb_t *this, uint8_t * image,
-		       uint8_t * py, uint8_t * pu, uint8_t * pv)
-{
-    yuv420_rgb24 (this, image, py, pu, pv, CPU_MMX);
-    emms();	/* re-initialize x86 FPU after MMX use */
-}
-
-static void mmx_argb32 (yuv2rgb_t *this, uint8_t * image,
-			uint8_t * py, uint8_t * pu, uint8_t * pv)
-{
-    yuv420_argb32 (this, image, py, pu, pv, CPU_MMX);
-    emms();	/* re-initialize x86 FPU after MMX use */
-}
-
-static void mmx_abgr32 (yuv2rgb_t *this, uint8_t * image,
-			uint8_t * py, uint8_t * pu, uint8_t * pv)
-{
-    yuv420_abgr32 (this, image, py, pu, pv, CPU_MMX);
-    emms();	/* re-initialize x86 FPU after MMX use */
-}
-
-void yuv2rgb_init_mmxext (yuv2rgb_factory_t *this) {
-
-  if (this->swapped) 
-    return; /*no swapped pixel output upto now*/
-
-  switch (this->mode) {
-  case MODE_15_RGB:
-    this->yuv2rgb_fun = mmxext_rgb15;
-    break;
-  case MODE_16_RGB:
-    this->yuv2rgb_fun = mmxext_rgb16;
-    break;
-  case MODE_24_RGB:
-    this->yuv2rgb_fun = mmxext_rgb24;
-    break;
-  case MODE_32_RGB:
-    this->yuv2rgb_fun = mmxext_argb32;
-    break;
-  case MODE_32_BGR:
-    this->yuv2rgb_fun = mmxext_abgr32;
-    break;
-  }
-}
-
-void yuv2rgb_init_mmx (yuv2rgb_factory_t *this) {
-
-  if (this->swapped) 
-    return; /*no swapped pixel output upto now*/
-
-  switch (this->mode) {
-  case MODE_15_RGB:
-    this->yuv2rgb_fun = mmx_rgb15;
-    break;
-  case MODE_16_RGB:
-    this->yuv2rgb_fun = mmx_rgb16;
-    break;
-  case MODE_24_RGB:
-    this->yuv2rgb_fun = mmx_rgb24;
-    break;
-  case MODE_32_RGB:
-    this->yuv2rgb_fun = mmx_argb32;
-    break;
-  case MODE_32_BGR:
-    this->yuv2rgb_fun = mmx_abgr32;
-    break;
-  }
-}
-
-
-#endif