sync with liba52 0.7.4, patch by Emanuele Giaquinta >emanuele.giaquinta ! gmail * com<

part 3: removed files


git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@18725 b3059339-0415-0410-9bf9-f77b7e298cf2

4 files changed, 0 insertions, 2345 deletions

diff --git a/liba52/config.h b/liba52/config.h
deleted file mode 100644
index 096682bcdd..0000000000
--- a/liba52/config.h
+++ /dev/null
@@ -1,12 +0,0 @@
-
-// mplayer -> liba52 config wrapper mess
-
-#include "../config.h"
-
-#ifdef HAVE_MLIB
-#define LIBA52_MLIB
-#endif
-
-// define for double precision:
-//#define LIBA52_DOUBLE
-
diff --git a/liba52/imdct_mlib.c b/liba52/imdct_mlib.c
deleted file mode 100644
index ceec96e1a4..0000000000
--- a/liba52/imdct_mlib.c
+++ /dev/null
@@ -1,142 +0,0 @@
-/*
- * imdct_mlib.c
- * Copyright (C) 2000-2001 Michel Lespinasse <walken@zoy.org>
- * Copyright (C) 1999-2000 Aaron Holtzman <aholtzma@ess.engr.uvic.ca>
- *
- * This file is part of a52dec, a free ATSC A-52 stream decoder.
- * See http://liba52.sourceforge.net/ for updates.
- *
- * Modified for use with MPlayer, changes contained in liba52_changes.diff.
- * detailed CVS changelog at http://www.mplayerhq.hu/cgi-bin/cvsweb.cgi/main/
- * $Id$
- *
- * a52dec 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.
- *
- * a52dec 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
- */
-
-#include "config.h"
-
-#ifdef LIBA52_MLIB
-
-#include <mlib_types.h>
-#include <mlib_status.h>
-#include <mlib_signal.h>
-#include <string.h>
-#include <inttypes.h>
-
-#include "a52.h"
-#include "a52_internal.h"
-
-extern sample_t imdct_window[];
-
-void
-imdct_do_512_mlib(sample_t data[], sample_t delay[], sample_t bias)
-{
-	sample_t *buf_real;
-	sample_t *buf_imag;
-	sample_t *data_ptr;
-	sample_t *delay_ptr;
-	sample_t *window_ptr;
-	sample_t tmp[256] __attribute__((aligned(16)));
-	int i;
-	
-	memcpy(tmp, data, 256 * sizeof(sample_t));
-	mlib_SignalIMDCT_F32(tmp);
-  
-	buf_real = tmp;
-	buf_imag = tmp + 128;
-	data_ptr = data;
-	delay_ptr = delay;
-	window_ptr = imdct_window;
-
-	/* Window and convert to real valued signal */
-	for(i=0; i< 64; i++) 
-	{ 
-		*data_ptr++ = -buf_imag[64+i]   * *window_ptr++ + *delay_ptr++ + bias; 
-		*data_ptr++ =  buf_real[64-i-1] * *window_ptr++ + *delay_ptr++ + bias; 
-	}
-
-	for(i=0; i< 64; i++) 
-	{ 
-		*data_ptr++ = -buf_real[i]       * *window_ptr++ + *delay_ptr++ + bias; 
-		*data_ptr++ =  buf_imag[128-i-1] * *window_ptr++ + *delay_ptr++ + bias; 
-	}
-	
-	/* The trailing edge of the window goes into the delay line */
-	delay_ptr = delay;
-
-	for(i=0; i< 64; i++) 
-	{ 
-		*delay_ptr++ = -buf_real[64+i]   * *--window_ptr; 
-		*delay_ptr++ =  buf_imag[64-i-1] * *--window_ptr; 
-	}
-
-	for(i=0; i<64; i++) 
-	{
-		*delay_ptr++ =  buf_imag[i]       * *--window_ptr; 
-		*delay_ptr++ = -buf_real[128-i-1] * *--window_ptr; 
-	}
-}
-
-void
-imdct_do_256_mlib(sample_t data[], sample_t delay[], sample_t bias)
-{
-	sample_t *buf1_real, *buf1_imag;
-	sample_t *buf2_real, *buf2_imag;
-	sample_t *data_ptr;
-	sample_t *delay_ptr;
-	sample_t *window_ptr;
-	sample_t tmp[256] __attribute__((aligned(16)));
-	int i;
-	
-	memcpy(tmp, data, 256 * sizeof(sample_t));
-	mlib_SignalIMDCTSplit_F32(tmp);
-  
-	buf1_real = tmp;
-	buf1_imag = tmp + 128 + 64;
-	buf2_real = tmp + 64;
-	buf2_imag = tmp + 128;
-	data_ptr = data;
-	delay_ptr = delay;
-	window_ptr = imdct_window;
-
-	/* Window and convert to real valued signal */
-	for(i=0; i< 64; i++) 
-	{
-		*data_ptr++ = -buf1_imag[i]      * *window_ptr++ + *delay_ptr++ + bias;
-		*data_ptr++ =  buf1_real[64-i-1] * *window_ptr++ + *delay_ptr++ + bias;
-	}
-
-	for(i=0; i< 64; i++) 
-	{
-		*data_ptr++ = -buf1_real[i]      * *window_ptr++ + *delay_ptr++ + bias;
-		*data_ptr++ =  buf1_imag[64-i-1] * *window_ptr++ + *delay_ptr++ + bias;
-	}
-	
-	delay_ptr = delay;
-
-	for(i=0; i< 64; i++) 
-	{
-		*delay_ptr++ = -buf2_real[i]      * *--window_ptr;
-		*delay_ptr++ =  buf2_imag[64-i-1] * *--window_ptr;
-	}
-
-	for(i=0; i< 64; i++) 
-	{
-		*delay_ptr++ =  buf2_imag[i]      * *--window_ptr;
-		*delay_ptr++ = -buf2_real[64-i-1] * *--window_ptr;
-	}
-}
-
-#endif
diff --git a/liba52/liba52_amd64_changes.diff b/liba52/liba52_amd64_changes.diff
deleted file mode 100644
index d0d5869c49..0000000000
--- a/liba52/liba52_amd64_changes.diff
+++ /dev/null
@@ -1,2189 +0,0 @@
-Index: liba52/a52_internal.h
-===================================================================
-RCS file: /cvsroot/mplayer/main/liba52/a52_internal.h,v
-retrieving revision 1.4
-diff -u -r1.4 a52_internal.h
---- liba52/a52_internal.h	22 Mar 2005 23:27:18 -0000	1.4
-+++ liba52/a52_internal.h	31 Jul 2005 21:20:09 -0000
-@@ -41,6 +41,20 @@
- #define DELTA_BIT_NONE (2)
- #define DELTA_BIT_RESERVED (3)
- 
-+#ifdef ARCH_X86_64
-+# define REG_a "rax"
-+# define REG_d "rdx"
-+# define REG_S "rsi"
-+# define REG_D "rdi"
-+# define REG_BP "rbp"
-+#else
-+# define REG_a "eax"
-+# define REG_d "edx"
-+# define REG_S "esi"
-+# define REG_D "edi"
-+# define REG_BP "ebp"
-+#endif
-+
- void bit_allocate (a52_state_t * state, a52_ba_t * ba, int bndstart,
- 		   int start, int end, int fastleak, int slowleak,
- 		   uint8_t * exp, int8_t * bap);
-Index: liba52/downmix.c
-===================================================================
-RCS file: /cvsroot/mplayer/main/liba52/downmix.c,v
-retrieving revision 1.17
-diff -u -r1.17 downmix.c
---- liba52/downmix.c	22 Mar 2005 23:27:18 -0000	1.17
-+++ liba52/downmix.c	31 Jul 2005 21:20:09 -0000
-@@ -56,7 +56,7 @@
- {
-     upmix= upmix_C;
-     downmix= downmix_C;
--#ifdef ARCH_X86    
-+#if defined(ARCH_X86) || defined(ARCH_X86_64)
-     if(mm_accel & MM_ACCEL_X86_MMX) upmix= upmix_MMX;
-     if(mm_accel & MM_ACCEL_X86_SSE) downmix= downmix_SSE;
-     if(mm_accel & MM_ACCEL_X86_3DNOW) downmix= downmix_3dnow;
-@@ -684,27 +684,27 @@
-     }
- }
- 
--#ifdef ARCH_X86
-+#if defined(ARCH_X86) || defined(ARCH_X86_64)
- static void mix2to1_SSE (sample_t * dest, sample_t * src, sample_t bias)
- {
- 	asm volatile(
- 	"movlps %2, %%xmm7		\n\t"
- 	"shufps $0x00, %%xmm7, %%xmm7	\n\t"
--	"movl $-1024, %%esi		\n\t"
-+	"mov $-1024, %%"REG_S"		\n\t"
- 	".balign 16\n\t"
- 	"1:				\n\t"
--	"movaps (%0, %%esi), %%xmm0	\n\t" 
--	"movaps 16(%0, %%esi), %%xmm1	\n\t" 
--	"addps (%1, %%esi), %%xmm0	\n\t" 
--	"addps 16(%1, %%esi), %%xmm1	\n\t" 
-+	"movaps (%0, %%"REG_S"), %%xmm0	\n\t" 
-+	"movaps 16(%0, %%"REG_S"), %%xmm1\n\t" 
-+	"addps (%1, %%"REG_S"), %%xmm0	\n\t" 
-+	"addps 16(%1, %%"REG_S"), %%xmm1\n\t" 
- 	"addps %%xmm7, %%xmm0		\n\t"
- 	"addps %%xmm7, %%xmm1		\n\t"
--	"movaps %%xmm0, (%1, %%esi)	\n\t"
--	"movaps %%xmm1, 16(%1, %%esi)	\n\t"
--	"addl $32, %%esi		\n\t"
-+	"movaps %%xmm0, (%1, %%"REG_S")	\n\t"
-+	"movaps %%xmm1, 16(%1, %%"REG_S")\n\t"
-+	"add $32, %%"REG_S"		\n\t"
- 	" jnz 1b			\n\t"
- 	:: "r" (src+256), "r" (dest+256), "m" (bias)
--	: "%esi"
-+	: "%"REG_S
- 	);
- }
- 
-@@ -713,19 +713,19 @@
- 	asm volatile(
- 	"movlps %1, %%xmm7		\n\t"
- 	"shufps $0x00, %%xmm7, %%xmm7	\n\t"
--	"movl $-1024, %%esi		\n\t"
-+	"mov $-1024, %%"REG_S"		\n\t"
- 	".balign 16\n\t"
- 	"1:				\n\t"
--	"movaps (%0, %%esi), %%xmm0	\n\t" 
--	"movaps 1024(%0, %%esi), %%xmm1	\n\t" 
--	"addps 2048(%0, %%esi), %%xmm0	\n\t" 
-+	"movaps (%0, %%"REG_S"), %%xmm0	\n\t" 
-+	"movaps 1024(%0, %%"REG_S"), %%xmm1\n\t" 
-+	"addps 2048(%0, %%"REG_S"), %%xmm0\n\t" 
- 	"addps %%xmm7, %%xmm1		\n\t"
- 	"addps %%xmm1, %%xmm0		\n\t"
--	"movaps %%xmm0, (%0, %%esi)	\n\t"
--	"addl $16, %%esi		\n\t"
-+	"movaps %%xmm0, (%0, %%"REG_S")	\n\t"
-+	"add $16, %%"REG_S"		\n\t"
- 	" jnz 1b			\n\t"
- 	:: "r" (samples+256), "m" (bias)
--	: "%esi"
-+	: "%"REG_S
- 	);
- }
- 
-@@ -734,20 +734,20 @@
- 	asm volatile(
- 	"movlps %1, %%xmm7		\n\t"
- 	"shufps $0x00, %%xmm7, %%xmm7	\n\t"
--	"movl $-1024, %%esi		\n\t"
-+	"mov $-1024, %%"REG_S"		\n\t"
- 	".balign 16\n\t"
- 	"1:				\n\t"
--	"movaps (%0, %%esi), %%xmm0	\n\t" 
--	"movaps 1024(%0, %%esi), %%xmm1	\n\t" 
--	"addps 2048(%0, %%esi), %%xmm0	\n\t" 
--	"addps 3072(%0, %%esi), %%xmm1	\n\t" 
-+	"movaps (%0, %%"REG_S"), %%xmm0	\n\t" 
-+	"movaps 1024(%0, %%"REG_S"), %%xmm1\n\t" 
-+	"addps 2048(%0, %%"REG_S"), %%xmm0\n\t" 
-+	"addps 3072(%0, %%"REG_S"), %%xmm1\n\t" 
- 	"addps %%xmm7, %%xmm0		\n\t"
- 	"addps %%xmm1, %%xmm0		\n\t"
--	"movaps %%xmm0, (%0, %%esi)	\n\t"
--	"addl $16, %%esi		\n\t"
-+	"movaps %%xmm0, (%0, %%"REG_S")	\n\t"
-+	"add $16, %%"REG_S"		\n\t"
- 	" jnz 1b			\n\t"
- 	:: "r" (samples+256), "m" (bias)
--	: "%esi"
-+	: "%"REG_S
- 	);
- }
- 
-@@ -756,21 +756,21 @@
- 	asm volatile(
- 	"movlps %1, %%xmm7		\n\t"
- 	"shufps $0x00, %%xmm7, %%xmm7	\n\t"
--	"movl $-1024, %%esi		\n\t"
-+	"mov $-1024, %%"REG_S"		\n\t"
- 	".balign 16\n\t"
- 	"1:				\n\t"
--	"movaps (%0, %%esi), %%xmm0	\n\t" 
--	"movaps 1024(%0, %%esi), %%xmm1	\n\t" 
--	"addps 2048(%0, %%esi), %%xmm0	\n\t" 
--	"addps 3072(%0, %%esi), %%xmm1	\n\t" 
-+	"movaps (%0, %%"REG_S"), %%xmm0	\n\t" 
-+	"movaps 1024(%0, %%"REG_S"), %%xmm1\n\t" 
-+	"addps 2048(%0, %%"REG_S"), %%xmm0\n\t" 
-+	"addps 3072(%0, %%"REG_S"), %%xmm1\n\t" 
- 	"addps %%xmm7, %%xmm0		\n\t"
--	"addps 4096(%0, %%esi), %%xmm1	\n\t" 
-+	"addps 4096(%0, %%"REG_S"), %%xmm1\n\t" 
- 	"addps %%xmm1, %%xmm0		\n\t"
--	"movaps %%xmm0, (%0, %%esi)	\n\t"
--	"addl $16, %%esi		\n\t"
-+	"movaps %%xmm0, (%0, %%"REG_S")	\n\t"
-+	"add $16, %%"REG_S"		\n\t"
- 	" jnz 1b			\n\t"
- 	:: "r" (samples+256), "m" (bias)
--	: "%esi"
-+	: "%"REG_S
- 	);
- }
- 
-@@ -779,21 +779,21 @@
- 	asm volatile(
- 	"movlps %1, %%xmm7		\n\t"
- 	"shufps $0x00, %%xmm7, %%xmm7	\n\t"
--	"movl $-1024, %%esi		\n\t"
-+	"mov $-1024, %%"REG_S"		\n\t"
- 	".balign 16\n\t"
- 	"1:				\n\t"
--	"movaps 1024(%0, %%esi), %%xmm0	\n\t" 
-+	"movaps 1024(%0, %%"REG_S"), %%xmm0\n\t" 
- 	"addps %%xmm7, %%xmm0		\n\t" //common
--	"movaps (%0, %%esi), %%xmm1	\n\t" 
--	"movaps 2048(%0, %%esi), %%xmm2	\n\t"
-+	"movaps (%0, %%"REG_S"), %%xmm1	\n\t" 
-+	"movaps 2048(%0, %%"REG_S"), %%xmm2\n\t"
- 	"addps %%xmm0, %%xmm1		\n\t"
- 	"addps %%xmm0, %%xmm2		\n\t"
--	"movaps %%xmm1, (%0, %%esi)	\n\t"
--	"movaps %%xmm2, 1024(%0, %%esi)	\n\t"
--	"addl $16, %%esi		\n\t"
-+	"movaps %%xmm1, (%0, %%"REG_S")	\n\t"
-+	"movaps %%xmm2, 1024(%0, %%"REG_S")\n\t"
-+	"add $16, %%"REG_S"		\n\t"
- 	" jnz 1b			\n\t"
- 	:: "r" (samples+256), "m" (bias)
--	: "%esi"
-+	: "%"REG_S
- 	);
- }
- 
-@@ -802,21 +802,21 @@
- 	asm volatile(
- 		"movlps %2, %%xmm7		\n\t"
- 		"shufps $0x00, %%xmm7, %%xmm7	\n\t"
--		"movl $-1024, %%esi		\n\t"
-+		"mov $-1024, %%"REG_S"		\n\t"
- 		".balign 16\n\t"
- 		"1:				\n\t"
--		"movaps 1024(%1, %%esi), %%xmm0	\n\t" 
-+		"movaps 1024(%1, %%"REG_S"), %%xmm0\n\t" 
- 		"addps %%xmm7, %%xmm0		\n\t" //common
--		"movaps (%0, %%esi), %%xmm1	\n\t" 
--		"movaps (%1, %%esi), %%xmm2	\n\t"
-+		"movaps (%0, %%"REG_S"), %%xmm1	\n\t" 
-+		"movaps (%1, %%"REG_S"), %%xmm2	\n\t"
- 		"addps %%xmm0, %%xmm1		\n\t"
- 		"addps %%xmm0, %%xmm2		\n\t"
--		"movaps %%xmm1, (%0, %%esi)	\n\t"
--		"movaps %%xmm2, (%1, %%esi)	\n\t"
--		"addl $16, %%esi		\n\t"
-+		"movaps %%xmm1, (%0, %%"REG_S")	\n\t"
-+		"movaps %%xmm2, (%1, %%"REG_S")	\n\t"
-+		"add $16, %%"REG_S"		\n\t"
- 		" jnz 1b			\n\t"
- 	:: "r" (left+256), "r" (right+256), "m" (bias)
--	: "%esi"
-+	: "%"REG_S
- 	);
- }
- 
-@@ -825,22 +825,22 @@
- 	asm volatile(
- 		"movlps %1, %%xmm7		\n\t"
- 		"shufps $0x00, %%xmm7, %%xmm7	\n\t"
--		"movl $-1024, %%esi		\n\t"
-+		"mov $-1024, %%"REG_S"		\n\t"
- 		".balign 16\n\t"
- 		"1:				\n\t"
--		"movaps 2048(%0, %%esi), %%xmm0	\n\t"  // surround
--		"movaps (%0, %%esi), %%xmm1	\n\t" 
--		"movaps 1024(%0, %%esi), %%xmm2	\n\t"
-+		"movaps 2048(%0, %%"REG_S"), %%xmm0\n\t"  // surround
-+		"movaps (%0, %%"REG_S"), %%xmm1	\n\t" 
-+		"movaps 1024(%0, %%"REG_S"), %%xmm2\n\t"
- 		"addps %%xmm7, %%xmm1		\n\t"
- 		"addps %%xmm7, %%xmm2		\n\t"
- 		"subps %%xmm0, %%xmm1		\n\t"
- 		"addps %%xmm0, %%xmm2		\n\t"
--		"movaps %%xmm1, (%0, %%esi)	\n\t"
--		"movaps %%xmm2, 1024(%0, %%esi)	\n\t"
--		"addl $16, %%esi		\n\t"
-+		"movaps %%xmm1, (%0, %%"REG_S")	\n\t"
-+		"movaps %%xmm2, 1024(%0, %%"REG_S")\n\t"
-+		"add $16, %%"REG_S"		\n\t"
- 		" jnz 1b			\n\t"
- 	:: "r" (samples+256), "m" (bias)
--	: "%esi"
-+	: "%"REG_S
- 	);
- }
- 
-@@ -849,22 +849,22 @@
- 	asm volatile(
- 		"movlps %1, %%xmm7		\n\t"
- 		"shufps $0x00, %%xmm7, %%xmm7	\n\t"
--		"movl $-1024, %%esi		\n\t"
-+		"mov $-1024, %%"REG_S"		\n\t"
- 		".balign 16\n\t"
- 		"1:				\n\t"
--		"movaps 1024(%0, %%esi), %%xmm0	\n\t"  
--		"addps 3072(%0, %%esi), %%xmm0	\n\t"  
-+		"movaps 1024(%0, %%"REG_S"), %%xmm0\n\t"  
-+		"addps 3072(%0, %%"REG_S"), %%xmm0\n\t"  
- 		"addps %%xmm7, %%xmm0		\n\t" // common
--		"movaps (%0, %%esi), %%xmm1	\n\t" 
--		"movaps 2048(%0, %%esi), %%xmm2	\n\t"
-+		"movaps (%0, %%"REG_S"), %%xmm1	\n\t" 
-+		"movaps 2048(%0, %%"REG_S"), %%xmm2\n\t"
- 		"addps %%xmm0, %%xmm1		\n\t"
- 		"addps %%xmm0, %%xmm2		\n\t"
--		"movaps %%xmm1, (%0, %%esi)	\n\t"
--		"movaps %%xmm2, 1024(%0, %%esi)	\n\t"
--		"addl $16, %%esi		\n\t"
-+		"movaps %%xmm1, (%0, %%"REG_S")	\n\t"
-+		"movaps %%xmm2, 1024(%0, %%"REG_S")\n\t"
-+		"add $16, %%"REG_S"		\n\t"
- 		" jnz 1b			\n\t"
- 	:: "r" (samples+256), "m" (bias)
--	: "%esi"
-+	: "%"REG_S
- 	);
- }
- 
-@@ -873,24 +873,24 @@
- 	asm volatile(
- 		"movlps %1, %%xmm7		\n\t"
- 		"shufps $0x00, %%xmm7, %%xmm7	\n\t"
--		"movl $-1024, %%esi		\n\t"
-+		"mov $-1024, %%"REG_S"		\n\t"
- 		".balign 16\n\t"
- 		"1:				\n\t"
--		"movaps 1024(%0, %%esi), %%xmm0	\n\t"  
--		"movaps 3072(%0, %%esi), %%xmm3	\n\t" // surround
-+		"movaps 1024(%0, %%"REG_S"), %%xmm0\n\t"  
-+		"movaps 3072(%0, %%"REG_S"), %%xmm3\n\t" // surround
- 		"addps %%xmm7, %%xmm0		\n\t" // common
--		"movaps (%0, %%esi), %%xmm1	\n\t" 
--		"movaps 2048(%0, %%esi), %%xmm2	\n\t"
-+		"movaps (%0, %%"REG_S"), %%xmm1	\n\t" 
-+		"movaps 2048(%0, %%"REG_S"), %%xmm2\n\t"
- 		"addps %%xmm0, %%xmm1		\n\t"
- 		"addps %%xmm0, %%xmm2		\n\t"
- 		"subps %%xmm3, %%xmm1		\n\t"
- 		"addps %%xmm3, %%xmm2		\n\t"
--		"movaps %%xmm1, (%0, %%esi)	\n\t"
--		"movaps %%xmm2, 1024(%0, %%esi)	\n\t"
--		"addl $16, %%esi		\n\t"
-+		"movaps %%xmm1, (%0, %%"REG_S")	\n\t"
-+		"movaps %%xmm2, 1024(%0, %%"REG_S")\n\t"
-+		"add $16, %%"REG_S"		\n\t"
- 		" jnz 1b			\n\t"
- 	:: "r" (samples+256), "m" (bias)
--	: "%esi"
-+	: "%"REG_S
- 	);
- }
- 
-@@ -899,23 +899,23 @@
- 	asm volatile(
- 		"movlps %1, %%xmm7		\n\t"
- 		"shufps $0x00, %%xmm7, %%xmm7	\n\t"
--		"movl $-1024, %%esi		\n\t"
-+		"mov $-1024, %%"REG_S"		\n\t"
- 		".balign 16\n\t"
- 		"1:				\n\t"
--		"movaps 2048(%0, %%esi), %%xmm0	\n\t"  
--		"addps 3072(%0, %%esi), %%xmm0	\n\t" // surround
--		"movaps (%0, %%esi), %%xmm1	\n\t" 
--		"movaps 1024(%0, %%esi), %%xmm2	\n\t"
-+		"movaps 2048(%0, %%"REG_S"), %%xmm0\n\t"  
-+		"addps 3072(%0, %%"REG_S"), %%xmm0\n\t" // surround
-+		"movaps (%0, %%"REG_S"), %%xmm1	\n\t" 
-+		"movaps 1024(%0, %%"REG_S"), %%xmm2\n\t"
- 		"addps %%xmm7, %%xmm1		\n\t"
- 		"addps %%xmm7, %%xmm2		\n\t"
- 		"subps %%xmm0, %%xmm1		\n\t"
- 		"addps %%xmm0, %%xmm2		\n\t"
--		"movaps %%xmm1, (%0, %%esi)	\n\t"
--		"movaps %%xmm2, 1024(%0, %%esi)	\n\t"
--		"addl $16, %%esi		\n\t"
-+		"movaps %%xmm1, (%0, %%"REG_S")	\n\t"
-+		"movaps %%xmm2, 1024(%0, %%"REG_S")\n\t"
-+		"add $16, %%"REG_S"		\n\t"
- 		" jnz 1b			\n\t"
- 	:: "r" (samples+256), "m" (bias)
--	: "%esi"
-+	: "%"REG_S
- 	);
- }
- 
-@@ -924,22 +924,22 @@
- 	asm volatile(
- 	"movlps %1, %%xmm7		\n\t"
- 	"shufps $0x00, %%xmm7, %%xmm7	\n\t"
--	"movl $-1024, %%esi		\n\t"
-+	"mov $-1024, %%"REG_S"		\n\t"
- 	".balign 16\n\t"
- 	"1:				\n\t"
--	"movaps 1024(%0, %%esi), %%xmm0	\n\t" 
-+	"movaps 1024(%0, %%"REG_S"), %%xmm0\n\t" 
- 	"addps %%xmm7, %%xmm0		\n\t" // common
- 	"movaps %%xmm0, %%xmm1		\n\t" // common
--	"addps (%0, %%esi), %%xmm0	\n\t" 
--	"addps 2048(%0, %%esi), %%xmm1	\n\t" 
--	"addps 3072(%0, %%esi), %%xmm0	\n\t" 
--	"addps 4096(%0, %%esi), %%xmm1	\n\t" 
--	"movaps %%xmm0, (%0, %%esi)	\n\t"
--	"movaps %%xmm1, 1024(%0, %%esi)	\n\t"
--	"addl $16, %%esi		\n\t"
-+	"addps (%0, %%"REG_S"), %%xmm0	\n\t" 
-+	"addps 2048(%0, %%"REG_S"), %%xmm1\n\t" 
-+	"addps 3072(%0, %%"REG_S"), %%xmm0\n\t" 
-+	"addps 4096(%0, %%"REG_S"), %%xmm1\n\t" 
-+	"movaps %%xmm0, (%0, %%"REG_S")	\n\t"
-+	"movaps %%xmm1, 1024(%0, %%"REG_S")\n\t"
-+	"add $16, %%"REG_S"		\n\t"
- 	" jnz 1b			\n\t"
- 	:: "r" (samples+256), "m" (bias)
--	: "%esi"
-+	: "%"REG_S
- 	);
- }
- 
-@@ -948,25 +948,25 @@
- 	asm volatile(
- 	"movlps %1, %%xmm7		\n\t"
- 	"shufps $0x00, %%xmm7, %%xmm7	\n\t"
--	"movl $-1024, %%esi		\n\t"
-+	"mov $-1024, %%"REG_S"		\n\t"
- 	".balign 16\n\t"
- 	"1:				\n\t"
--	"movaps 1024(%0, %%esi), %%xmm0	\n\t" 
--	"movaps 3072(%0, %%esi), %%xmm2	\n\t" 
-+	"movaps 1024(%0, %%"REG_S"), %%xmm0\n\t" 
-+	"movaps 3072(%0, %%"REG_S"), %%xmm2\n\t" 
- 	"addps %%xmm7, %%xmm0		\n\t" // common
--	"addps 4096(%0, %%esi), %%xmm2	\n\t" // surround	
--	"movaps (%0, %%esi), %%xmm1	\n\t" 
--	"movaps 2048(%0, %%esi), %%xmm3	\n\t" 
-+	"addps 4096(%0, %%"REG_S"), %%xmm2\n\t" // surround	
-+	"movaps (%0, %%"REG_S"), %%xmm1	\n\t" 
-+	"movaps 2048(%0, %%"REG_S"), %%xmm3\n\t" 
- 	"subps %%xmm2, %%xmm1		\n\t"	
- 	"addps %%xmm2, %%xmm3		\n\t"	
- 	"addps %%xmm0, %%xmm1		\n\t"	
- 	"addps %%xmm0, %%xmm3		\n\t"	
--	"movaps %%xmm1, (%0, %%esi)	\n\t"
--	"movaps %%xmm3, 1024(%0, %%esi)	\n\t"
--	"addl $16, %%esi		\n\t"
-+	"movaps %%xmm1, (%0, %%"REG_S")	\n\t"
-+	"movaps %%xmm3, 1024(%0, %%"REG_S")\n\t"
-+	"add $16, %%"REG_S"		\n\t"
- 	" jnz 1b			\n\t"
- 	:: "r" (samples+256), "m" (bias)
--	: "%esi"
-+	: "%"REG_S
- 	);
- }
- 
-@@ -975,40 +975,40 @@
- 	asm volatile(
- 		"movlps %2, %%xmm7		\n\t"
- 		"shufps $0x00, %%xmm7, %%xmm7	\n\t"
--		"movl $-1024, %%esi		\n\t"
-+		"mov $-1024, %%"REG_S"		\n\t"
- 		".balign 16\n\t"
- 		"1:				\n\t"
--		"movaps (%0, %%esi), %%xmm0	\n\t"  
--		"movaps 16(%0, %%esi), %%xmm1	\n\t"  
--		"addps 1024(%0, %%esi), %%xmm0	\n\t"
--		"addps 1040(%0, %%esi), %%xmm1	\n\t"
-+		"movaps (%0, %%"REG_S"), %%xmm0	\n\t"  
-+		"movaps 16(%0, %%"REG_S"), %%xmm1\n\t"  
-+		"addps 1024(%0, %%"REG_S"), %%xmm0\n\t"
-+		"addps 1040(%0, %%"REG_S"), %%xmm1\n\t"
- 		"addps %%xmm7, %%xmm0		\n\t"
- 		"addps %%xmm7, %%xmm1		\n\t"
--		"movaps %%xmm0, (%1, %%esi)	\n\t"
--		"movaps %%xmm1, 16(%1, %%esi)	\n\t"
--		"addl $32, %%esi		\n\t"
-+		"movaps %%xmm0, (%1, %%"REG_S")	\n\t"
-+		"movaps %%xmm1, 16(%1, %%"REG_S")\n\t"
-+		"add $32, %%"REG_S"		\n\t"
- 		" jnz 1b			\n\t"
- 	:: "r" (src+256), "r" (dest+256), "m" (bias)
--	: "%esi"
-+	: "%"REG_S
- 	);
- }
- 
- static void zero_MMX(sample_t * samples)
- {
- 	asm volatile(
--		"movl $-1024, %%esi		\n\t"
-+		"mov $-1024, %%"REG_S"		\n\t"
- 		"pxor %%mm0, %%mm0		\n\t"
- 		".balign 16\n\t"
- 		"1:				\n\t"
--		"movq %%mm0, (%0, %%esi)	\n\t"
--		"movq %%mm0, 8(%0, %%esi)	\n\t"
--		"movq %%mm0, 16(%0, %%esi)	\n\t"
--		"movq %%mm0, 24(%0, %%esi)	\n\t"
--		"addl $32, %%esi		\n\t"
-+		"movq %%mm0, (%0, %%"REG_S")	\n\t"
-+		"movq %%mm0, 8(%0, %%"REG_S")	\n\t"
-+		"movq %%mm0, 16(%0, %%"REG_S")	\n\t"
-+		"movq %%mm0, 24(%0, %%"REG_S")	\n\t"
-+		"add $32, %%"REG_S"		\n\t"
- 		" jnz 1b			\n\t"
- 		"emms"
- 	:: "r" (samples+256)
--	: "%esi"
-+	: "%"REG_S
- 	);
- }
- 
-@@ -1257,29 +1257,29 @@
- 	asm volatile(
- 	"movd  %2, %%mm7	\n\t"
- 	"punpckldq %2, %%mm7	\n\t"
--	"movl  $-1024, %%esi	\n\t"
-+	"mov $-1024, %%"REG_S"	\n\t"
- 	".balign 16\n\t"
- 	"1:			\n\t"
--	"movq  (%0, %%esi), %%mm0	\n\t" 
--	"movq  8(%0, %%esi), %%mm1	\n\t"
--	"movq  16(%0, %%esi), %%mm2	\n\t" 
--	"movq  24(%0, %%esi), %%mm3	\n\t"
--	"pfadd (%1, %%esi), %%mm0	\n\t" 
--	"pfadd 8(%1, %%esi), %%mm1	\n\t"
--	"pfadd 16(%1, %%esi), %%mm2	\n\t" 
--	"pfadd 24(%1, %%esi), %%mm3	\n\t"
-+	"movq  (%0, %%"REG_S"), %%mm0	\n\t" 
-+	"movq  8(%0, %%"REG_S"), %%mm1	\n\t"
-+	"movq  16(%0, %%"REG_S"), %%mm2	\n\t" 
-+	"movq  24(%0, %%"REG_S"), %%mm3	\n\t"
-+	"pfadd (%1, %%"REG_S"), %%mm0	\n\t" 
-+	"pfadd 8(%1, %%"REG_S"), %%mm1	\n\t"
-+	"pfadd 16(%1, %%"REG_S"), %%mm2	\n\t" 
-+	"pfadd 24(%1, %%"REG_S"), %%mm3	\n\t"
- 	"pfadd %%mm7, %%mm0		\n\t"
- 	"pfadd %%mm7, %%mm1		\n\t"
- 	"pfadd %%mm7, %%mm2		\n\t"
- 	"pfadd %%mm7, %%mm3		\n\t"
--	"movq  %%mm0, (%1, %%esi)	\n\t"
--	"movq  %%mm1, 8(%1, %%esi)	\n\t"
--	"movq  %%mm2, 16(%1, %%esi)	\n\t"
--	"movq  %%mm3, 24(%1, %%esi)	\n\t"
--	"addl $32, %%esi		\n\t"
-+	"movq  %%mm0, (%1, %%"REG_S")	\n\t"
-+	"movq  %%mm1, 8(%1, %%"REG_S")	\n\t"
-+	"movq  %%mm2, 16(%1, %%"REG_S")	\n\t"
-+	"movq  %%mm3, 24(%1, %%"REG_S")	\n\t"
-+	"add $32, %%"REG_S"		\n\t"
- 	" jnz 1b			\n\t"
- 	:: "r" (src+256), "r" (dest+256), "m" (bias)
--	: "%esi"
-+	: "%"REG_S
- 	);
- }
- 
-@@ -1288,25 +1288,25 @@
- 	asm volatile(
- 	"movd  %1, %%mm7	\n\t"
- 	"punpckldq %1, %%mm7	\n\t"
--	"movl $-1024, %%esi	\n\t"
-+	"mov $-1024, %%"REG_S"	\n\t"
- 	".balign 16\n\t"
- 	"1:			\n\t"
--	"movq  (%0, %%esi), %%mm0	\n\t" 
--	"movq  8(%0, %%esi), %%mm1	\n\t"
--	"movq  1024(%0, %%esi), %%mm2	\n\t" 
--	"movq  1032(%0, %%esi), %%mm3	\n\t"
--	"pfadd 2048(%0, %%esi), %%mm0	\n\t" 
--	"pfadd 2056(%0, %%esi), %%mm1	\n\t"
-+	"movq  (%0, %%"REG_S"), %%mm0	\n\t" 
-+	"movq  8(%0, %%"REG_S"), %%mm1	\n\t"
-+	"movq  1024(%0, %%"REG_S"), %%mm2\n\t" 
-+	"movq  1032(%0, %%"REG_S"), %%mm3\n\t"
-+	"pfadd 2048(%0, %%"REG_S"), %%mm0\n\t" 
-+	"pfadd 2056(%0, %%"REG_S"), %%mm1\n\t"
- 	"pfadd %%mm7, %%mm0		\n\t"
- 	"pfadd %%mm7, %%mm1		\n\t"
- 	"pfadd %%mm2, %%mm0		\n\t"
- 	"pfadd %%mm3, %%mm1		\n\t"
--	"movq  %%mm0, (%0, %%esi)	\n\t"
--	"movq  %%mm1, 8(%0, %%esi)	\n\t"
--	"addl $16, %%esi		\n\t"
-+	"movq  %%mm0, (%0, %%"REG_S")	\n\t"
-+	"movq  %%mm1, 8(%0, %%"REG_S")	\n\t"
-+	"add $16, %%"REG_S"		\n\t"
- 	" jnz 1b			\n\t"
- 	:: "r" (samples+256), "m" (bias)
--	: "%esi"
-+	: "%"REG_S
- 	);
- }
- 
-@@ -1315,27 +1315,27 @@
- 	asm volatile(
- 	"movd  %1, %%mm7	\n\t"
- 	"punpckldq %1, %%mm7	\n\t"
--	"movl $-1024, %%esi	\n\t"
-+	"mov $-1024, %%"REG_S"	\n\t"
- 	".balign 16\n\t"
- 	"1:			\n\t"
--	"movq  (%0, %%esi), %%mm0	\n\t" 
--	"movq  8(%0, %%esi), %%mm1	\n\t"
--	"movq  1024(%0, %%esi), %%mm2	\n\t" 
--	"movq  1032(%0, %%esi), %%mm3	\n\t"
--	"pfadd 2048(%0, %%esi), %%mm0	\n\t" 
--	"pfadd 2056(%0, %%esi), %%mm1	\n\t"
--	"pfadd 3072(%0, %%esi), %%mm2	\n\t" 
--	"pfadd 3080(%0, %%esi), %%mm3	\n\t"
-+	"movq  (%0, %%"REG_S"), %%mm0	\n\t" 
-+	"movq  8(%0, %%"REG_S"), %%mm1	\n\t"
-+	"movq  1024(%0, %%"REG_S"), %%mm2\n\t" 
-+	"movq  1032(%0, %%"REG_S"), %%mm3\n\t"
-+	"pfadd 2048(%0, %%"REG_S"), %%mm0\n\t" 
-+	"pfadd 2056(%0, %%"REG_S"), %%mm1\n\t"
-+	"pfadd 3072(%0, %%"REG_S"), %%mm2\n\t" 
-+	"pfadd 3080(%0, %%"REG_S"), %%mm3\n\t"
- 	"pfadd %%mm7, %%mm0		\n\t"
- 	"pfadd %%mm7, %%mm1		\n\t"
- 	"pfadd %%mm2, %%mm0		\n\t"
- 	"pfadd %%mm3, %%mm1		\n\t"
--	"movq  %%mm0, (%0, %%esi)	\n\t"
--	"movq  %%mm1, 8(%0, %%esi)	\n\t"
--	"addl $16, %%esi		\n\t"
-+	"movq  %%mm0, (%0, %%"REG_S")	\n\t"
-+	"movq  %%mm1, 8(%0, %%"REG_S")	\n\t"
-+	"add $16, %%"REG_S"		\n\t"
- 	" jnz 1b			\n\t"
- 	:: "r" (samples+256), "m" (bias)
--	: "%esi"
-+	: "%"REG_S
- 	);
- }
- 
-@@ -1344,29 +1344,29 @@
- 	asm volatile(
- 	"movd  %1, %%mm7	\n\t"
- 	"punpckldq %1, %%mm7	\n\t"
--	"movl $-1024, %%esi	\n\t"
-+	"mov $-1024, %%"REG_S"	\n\t"
- 	".balign 16\n\t"
- 	"1:			\n\t"
--	"movq  (%0, %%esi), %%mm0	\n\t" 
--	"movq  8(%0, %%esi), %%mm1	\n\t"
--	"movq  1024(%0, %%esi), %%mm2	\n\t" 
--	"movq  1032(%0, %%esi), %%mm3	\n\t"
--	"pfadd 2048(%0, %%esi), %%mm0	\n\t" 
--	"pfadd 2056(%0, %%esi), %%mm1	\n\t"
--	"pfadd 3072(%0, %%esi), %%mm2	\n\t" 
--	"pfadd 3080(%0, %%esi), %%mm3	\n\t"
-+	"movq  (%0, %%"REG_S"), %%mm0	\n\t" 
-+	"movq  8(%0, %%"REG_S"), %%mm1	\n\t"
-+	"movq  1024(%0, %%"REG_S"), %%mm2\n\t" 
-+	"movq  1032(%0, %%"REG_S"), %%mm3\n\t"
-+	"pfadd 2048(%0, %%"REG_S"), %%mm0\n\t" 
-+	"pfadd 2056(%0, %%"REG_S"), %%mm1\n\t"
-+	"pfadd 3072(%0, %%"REG_S"), %%mm2\n\t" 
-+	"pfadd 3080(%0, %%"REG_S"), %%mm3\n\t"
- 	"pfadd %%mm7, %%mm0		\n\t"
- 	"pfadd %%mm7, %%mm1		\n\t"
--	"pfadd 4096(%0, %%esi), %%mm2	\n\t" 
--	"pfadd 4104(%0, %%esi), %%mm3	\n\t"
-+	"pfadd 4096(%0, %%"REG_S"), %%mm2\n\t" 
-+	"pfadd 4104(%0, %%"REG_S"), %%mm3\n\t"
- 	"pfadd %%mm2, %%mm0		\n\t"
- 	"pfadd %%mm3, %%mm1		\n\t"
--	"movq  %%mm0, (%0, %%esi)	\n\t"
--	"movq  %%mm1, 8(%0, %%esi)	\n\t"
--	"addl $16, %%esi		\n\t"
-+	"movq  %%mm0, (%0, %%"REG_S")	\n\t"
-+	"movq  %%mm1, 8(%0, %%"REG_S")	\n\t"
-+	"add $16, %%"REG_S"		\n\t"
- 	" jnz 1b			\n\t"
- 	:: "r" (samples+256), "m" (bias)
--	: "%esi"
-+	: "%"REG_S
- 	);
- }
- 
-@@ -1375,29 +1375,29 @@
- 	asm volatile(
- 	"movd  %1, %%mm7	\n\t"
- 	"punpckldq %1, %%mm7	\n\t"
--	"movl $-1024, %%esi	\n\t"
-+	"mov $-1024, %%"REG_S"	\n\t"
- 	".balign 16\n\t"
- 	"1:			\n\t"
--	"movq   1024(%0, %%esi), %%mm0	\n\t" 
--	"movq   1032(%0, %%esi), %%mm1	\n\t"
-+	"movq   1024(%0, %%"REG_S"), %%mm0\n\t" 
-+	"movq   1032(%0, %%"REG_S"), %%mm1\n\t"
- 	"pfadd  %%mm7, %%mm0		\n\t" //common
- 	"pfadd  %%mm7, %%mm1		\n\t" //common
--	"movq   (%0, %%esi), %%mm2	\n\t" 
--	"movq   8(%0, %%esi), %%mm3	\n\t"
--	"movq   2048(%0, %%esi), %%mm4	\n\t"
--	"movq   2056(%0, %%esi), %%mm5	\n\t"
-+	"movq   (%0, %%"REG_S"), %%mm2	\n\t" 
-+	"movq   8(%0, %%"REG_S"), %%mm3	\n\t"
-+	"movq   2048(%0, %%"REG_S"), %%mm4\n\t"
-+	"movq   2056(%0, %%"REG_S"), %%mm5\n\t"
- 	"pfadd  %%mm0, %%mm2		\n\t"
- 	"pfadd  %%mm1, %%mm3		\n\t"
- 	"pfadd  %%mm0, %%mm4		\n\t"
- 	"pfadd  %%mm1, %%mm5		\n\t"
--	"movq   %%mm2, (%0, %%esi)	\n\t"
--	"movq   %%mm3, 8(%0, %%esi)	\n\t"
--	"movq   %%mm4, 1024(%0, %%esi)	\n\t"
--	"movq   %%mm5, 1032(%0, %%esi)	\n\t"
--	"addl $16, %%esi		\n\t"
-+	"movq   %%mm2, (%0, %%"REG_S")	\n\t"
-+	"movq   %%mm3, 8(%0, %%"REG_S")	\n\t"
-+	"movq   %%mm4, 1024(%0, %%"REG_S")\n\t"
-+	"movq   %%mm5, 1032(%0, %%"REG_S")\n\t"
-+	"add $16, %%"REG_S"		\n\t"
- 	" jnz 1b			\n\t"
- 	:: "r" (samples+256), "m" (bias)
--	: "%esi"
-+	: "%"REG_S
- 	);
- }
- 
-@@ -1406,29 +1406,29 @@
- 	asm volatile(
- 		"movd  %2, %%mm7	\n\t"
- 		"punpckldq %2, %%mm7	\n\t"
--		"movl $-1024, %%esi	\n\t"
-+		"mov $-1024, %%"REG_S"	\n\t"
- 		".balign 16\n\t"
- 		"1:			\n\t"
--		"movq  1024(%1, %%esi), %%mm0	\n\t" 
--		"movq  1032(%1, %%esi), %%mm1	\n\t"
-+		"movq  1024(%1, %%"REG_S"), %%mm0\n\t" 
-+		"movq  1032(%1, %%"REG_S"), %%mm1\n\t"
- 		"pfadd %%mm7, %%mm0		\n\t" //common
- 		"pfadd %%mm7, %%mm1		\n\t" //common
--		"movq  (%0, %%esi), %%mm2	\n\t" 
--		"movq  8(%0, %%esi), %%mm3	\n\t"
--		"movq  (%1, %%esi), %%mm4	\n\t"
--		"movq  8(%1, %%esi), %%mm5	\n\t"
-+		"movq  (%0, %%"REG_S"), %%mm2	\n\t" 
-+		"movq  8(%0, %%"REG_S"), %%mm3	\n\t"
-+		"movq  (%1, %%"REG_S"), %%mm4	\n\t"
-+		"movq  8(%1, %%"REG_S"), %%mm5	\n\t"
- 		"pfadd %%mm0, %%mm2		\n\t"
- 		"pfadd %%mm1, %%mm3		\n\t"
- 		"pfadd %%mm0, %%mm4		\n\t"
- 		"pfadd %%mm1, %%mm5		\n\t"
--		"movq  %%mm2, (%0, %%esi)	\n\t"
--		"movq  %%mm3, 8(%0, %%esi)	\n\t"
--		"movq  %%mm4, (%1, %%esi)	\n\t"
--		"movq  %%mm5, 8(%1, %%esi)	\n\t"
--		"addl $16, %%esi		\n\t"
-+		"movq  %%mm2, (%0, %%"REG_S")	\n\t"
-+		"movq  %%mm3, 8(%0, %%"REG_S")	\n\t"
-+		"movq  %%mm4, (%1, %%"REG_S")	\n\t"
-+		"movq  %%mm5, 8(%1, %%"REG_S")	\n\t"
-+		"add $16, %%"REG_S"		\n\t"
- 		" jnz 1b			\n\t"
- 	:: "r" (left+256), "r" (right+256), "m" (bias)
--	: "%esi"
-+	: "%"REG_S
- 	);
- }
- 
-@@ -1437,15 +1437,15 @@
- 	asm volatile(
- 		"movd  %1, %%mm7	\n\t"
- 		"punpckldq %1, %%mm7	\n\t"
--		"movl $-1024, %%esi	\n\t"
-+		"mov $-1024, %%"REG_S"	\n\t"
- 		".balign 16\n\t"
- 		"1:			\n\t"
--		"movq  2048(%0, %%esi), %%mm0	\n\t"  // surround
--		"movq  2056(%0, %%esi), %%mm1	\n\t"  // surround
--		"movq  (%0, %%esi), %%mm2	\n\t" 
--		"movq  8(%0, %%esi), %%mm3	\n\t"
--		"movq  1024(%0, %%esi), %%mm4	\n\t"
--		"movq  1032(%0, %%esi), %%mm5	\n\t"
-+		"movq  2048(%0, %%"REG_S"), %%mm0\n\t"  // surround
-+		"movq  2056(%0, %%"REG_S"), %%mm1\n\t"  // surround
-+		"movq  (%0, %%"REG_S"), %%mm2	\n\t" 
-+		"movq  8(%0, %%"REG_S"), %%mm3	\n\t"
-+		"movq  1024(%0, %%"REG_S"), %%mm4\n\t"
-+		"movq  1032(%0, %%"REG_S"), %%mm5\n\t"
- 		"pfadd %%mm7, %%mm2		\n\t"
- 		"pfadd %%mm7, %%mm3		\n\t"
- 		"pfadd %%mm7, %%mm4		\n\t"
-@@ -1454,14 +1454,14 @@
- 		"pfsub %%mm1, %%mm3		\n\t"
- 		"pfadd %%mm0, %%mm4		\n\t"
- 		"pfadd %%mm1, %%mm5		\n\t"
--		"movq  %%mm2, (%0, %%esi)	\n\t"
--		"movq  %%mm3, 8(%0, %%esi)	\n\t"
--		"movq  %%mm4, 1024(%0, %%esi)	\n\t"
--		"movq  %%mm5, 1032(%0, %%esi)	\n\t"
--		"addl $16, %%esi		\n\t"
-+		"movq  %%mm2, (%0, %%"REG_S")	\n\t"
-+		"movq  %%mm3, 8(%0, %%"REG_S")	\n\t"
-+		"movq  %%mm4, 1024(%0, %%"REG_S")\n\t"
-+		"movq  %%mm5, 1032(%0, %%"REG_S")\n\t"
-+		"add $16, %%"REG_S"		\n\t"
- 		" jnz 1b			\n\t"
- 	:: "r" (samples+256), "m" (bias)
--	: "%esi"
-+	: "%"REG_S
- 	);
- }
- 
-@@ -1470,31 +1470,31 @@
- 	asm volatile(
- 		"movd  %1, %%mm7	\n\t"
- 		"punpckldq %1, %%mm7	\n\t"
--		"movl $-1024, %%esi	\n\t"
-+		"mov $-1024, %%"REG_S"	\n\t"
- 		".balign 16\n\t"
- 		"1:			\n\t"
--		"movq  1024(%0, %%esi), %%mm0	\n\t"  
--		"movq  1032(%0, %%esi), %%mm1	\n\t"
--		"pfadd 3072(%0, %%esi), %%mm0	\n\t"  
--		"pfadd 3080(%0, %%esi), %%mm1	\n\t"
-+		"movq  1024(%0, %%"REG_S"), %%mm0\n\t"  
-+		"movq  1032(%0, %%"REG_S"), %%mm1\n\t"
-+		"pfadd 3072(%0, %%"REG_S"), %%mm0\n\t"  
-+		"pfadd 3080(%0, %%"REG_S"), %%mm1\n\t"
- 		"pfadd %%mm7, %%mm0		\n\t" // common
- 		"pfadd %%mm7, %%mm1		\n\t" // common
--		"movq  (%0, %%esi), %%mm2	\n\t" 
--		"movq  8(%0, %%esi), %%mm3	\n\t"
--		"movq  2048(%0, %%esi), %%mm4	\n\t"
--		"movq  2056(%0, %%esi), %%mm5	\n\t"
-+		"movq  (%0, %%"REG_S"), %%mm2	\n\t" 
-+		"movq  8(%0, %%"REG_S"), %%mm3	\n\t"
-+		"movq  2048(%0, %%"REG_S"), %%mm4\n\t"
-+		"movq  2056(%0, %%"REG_S"), %%mm5\n\t"
- 		"pfadd %%mm0, %%mm2		\n\t"
- 		"pfadd %%mm1, %%mm3		\n\t"
- 		"pfadd %%mm0, %%mm4		\n\t"
- 		"pfadd %%mm1, %%mm5		\n\t"
--		"movq  %%mm2, (%0, %%esi)	\n\t"
--		"movq  %%mm3, 8(%0, %%esi)	\n\t"
--		"movq  %%mm4, 1024(%0, %%esi)	\n\t"
--		"movq  %%mm5, 1032(%0, %%esi)	\n\t"
--		"addl $16, %%esi		\n\t"
-+		"movq  %%mm2, (%0, %%"REG_S")	\n\t"
-+		"movq  %%mm3, 8(%0, %%"REG_S")	\n\t"
-+		"movq  %%mm4, 1024(%0, %%"REG_S")\n\t"
-+		"movq  %%mm5, 1032(%0, %%"REG_S")\n\t"
-+		"add $16, %%"REG_S"		\n\t"
- 		" jnz 1b			\n\t"
- 	:: "r" (samples+256), "m" (bias)
--	: "%esi"
-+	: "%"REG_S
- 	);
- }
- 
-@@ -1503,35 +1503,35 @@
- 	asm volatile(
- 		"movd  %1, %%mm7	\n\t"
- 		"punpckldq %1, %%mm7	\n\t"
--		"movl $-1024, %%esi	\n\t"
-+		"mov $-1024, %%"REG_S"	\n\t"
- 		".balign 16\n\t"
- 		"1:			\n\t"
--		"movq   1024(%0, %%esi), %%mm0	\n\t"  
--		"movq   1032(%0, %%esi), %%mm1	\n\t"
-+		"movq   1024(%0, %%"REG_S"), %%mm0\n\t"  
-+		"movq   1032(%0, %%"REG_S"), %%mm1\n\t"
- 		"pfadd  %%mm7, %%mm0		\n\t" // common
- 		"pfadd  %%mm7, %%mm1		\n\t" // common
--		"movq   (%0, %%esi), %%mm2	\n\t" 
--		"movq   8(%0, %%esi), %%mm3	\n\t"
--		"movq   2048(%0, %%esi), %%mm4	\n\t"
--		"movq   2056(%0, %%esi), %%mm5	\n\t"
-+		"movq   (%0, %%"REG_S"), %%mm2	\n\t" 
-+		"movq   8(%0, %%"REG_S"), %%mm3	\n\t"
-+		"movq   2048(%0, %%"REG_S"), %%mm4\n\t"
-+		"movq   2056(%0, %%"REG_S"), %%mm5\n\t"
- 		"pfadd  %%mm0, %%mm2		\n\t"
- 		"pfadd  %%mm1, %%mm3		\n\t"
- 		"pfadd  %%mm0, %%mm4		\n\t"
- 		"pfadd  %%mm1, %%mm5		\n\t"
--		"movq   3072(%0, %%esi), %%mm0	\n\t" // surround
--		"movq   3080(%0, %%esi), %%mm1	\n\t" // surround
-+		"movq   3072(%0, %%"REG_S"), %%mm0\n\t" // surround
-+		"movq   3080(%0, %%"REG_S"), %%mm1\n\t" // surround
- 		"pfsub  %%mm0, %%mm2		\n\t"
- 		"pfsub  %%mm1, %%mm3		\n\t"
- 		"pfadd  %%mm0, %%mm4		\n\t"
- 		"pfadd  %%mm1, %%mm5		\n\t"
--		"movq   %%mm2, (%0, %%esi)	\n\t"
--		"movq   %%mm3, 8(%0, %%esi)	\n\t"
--		"movq   %%mm4, 1024(%0, %%esi)	\n\t"
--		"movq   %%mm5, 1032(%0, %%esi)	\n\t"
--		"addl $16, %%esi		\n\t"
-+		"movq   %%mm2, (%0, %%"REG_S")	\n\t"
-+		"movq   %%mm3, 8(%0, %%"REG_S")	\n\t"
-+		"movq   %%mm4, 1024(%0, %%"REG_S")\n\t"
-+		"movq   %%mm5, 1032(%0, %%"REG_S")\n\t"
-+		"add $16, %%"REG_S"		\n\t"
- 		" jnz 1b			\n\t"
- 	:: "r" (samples+256), "m" (bias)
--	: "%esi"
-+	: "%"REG_S
- 	);
- }
- 
-@@ -1540,17 +1540,17 @@
- 	asm volatile(
- 		"movd  %1, %%mm7	\n\t"
- 		"punpckldq %1, %%mm7	\n\t"
--		"movl $-1024, %%esi	\n\t"
-+		"mov $-1024, %%"REG_S"	\n\t"
- 		".balign 16\n\t"
- 		"1:			\n\t"
--		"movq  2048(%0, %%esi), %%mm0	\n\t"  
--		"movq  2056(%0, %%esi), %%mm1	\n\t"
--		"pfadd 3072(%0, %%esi), %%mm0	\n\t" // surround
--		"pfadd 3080(%0, %%esi), %%mm1	\n\t" // surround
--		"movq  (%0, %%esi), %%mm2	\n\t" 
--		"movq  8(%0, %%esi), %%mm3	\n\t"
--		"movq  1024(%0, %%esi), %%mm4	\n\t"
--		"movq  1032(%0, %%esi), %%mm5	\n\t"
-+		"movq  2048(%0, %%"REG_S"), %%mm0\n\t"  
-+		"movq  2056(%0, %%"REG_S"), %%mm1\n\t"
-+		"pfadd 3072(%0, %%"REG_S"), %%mm0\n\t" // surround