--- trunk/xvidcore/src/plugins/x86_asm/plugin_ssim-a.asm	2008/11/14 15:43:28	1794
+++ trunk/xvidcore/src/plugins/x86_asm/plugin_ssim-a.asm	2008/11/26 01:04:34	1795
@@ -24,40 +24,20 @@
 
 BITS 32
 
-%macro cglobal 1
-	%ifdef PREFIX
-		%ifdef MARK_FUNCS
-			global _%1:function %1.endfunc-%1
-			%define %1 _%1:function %1.endfunc-%1
-			%define ENDFUNC .endfunc
-		%else
-			global _%1
-			%define %1 _%1
-			%define ENDFUNC
-		%endif
-	%else
-		%ifdef MARK_FUNCS
-			global %1:function %1.endfunc-%1
-			%define ENDFUNC .endfunc
-		%else
-			global %1
-			%define ENDFUNC
-		%endif
-	%endif
-%endmacro
+%include "nasm.inc"
 
 %macro ACC_ROW 2
-	movq %1,[    ecx]
-	movq %2,[ecx+edx]
+	movq %1,[    TMP0]
+	movq %2,[TMP0+TMP1]
 	psadbw %1,mm0
 	psadbw %2,mm0
-        lea ecx, [ecx+2*edx]
+        lea TMP0, [TMP0+2*TMP1]
         paddw  %1, %2
 %endmacro
 
 %macro CONSIM_1x8_SSE2 0
-	movdqu xmm0,[ecx]
-	movdqu xmm1,[edx]
+	movdqu xmm0,[TMP0]
+	movdqu xmm1,[TMP1]
 
 	;unpack to words
 	punpcklbw xmm0,xmm2
@@ -76,8 +56,8 @@
 %endmacro
 
 %macro CONSIM_1x8_MMX 0
- 	movq mm0,[ecx];orig
- 	movq mm1,[edx];comp
+ 	movq mm0,[TMP0];orig
+ 	movq mm1,[TMP1];comp
 
  	;unpack low half of qw to words
  	punpcklbw mm0,mm2
@@ -94,8 +74,8 @@
 	pmaddwd mm1,mm0
 	paddd mm7,mm1
 
- 	movq mm0,[ecx];orig
- 	movq mm1,[edx];comp
+ 	movq mm0,[TMP0];orig
+ 	movq mm1,[TMP1];comp
 
  	;unpack high half of qw to words
  	punpckhbw mm0,mm2
@@ -114,39 +94,39 @@
 %endmacro
 
 %macro CONSIM_WRITEOUT 3
-	mov eax,[esp + 16];lumo
-	mul eax; lumo^2
-        add eax, 32
-	shr eax,6; 64*lum0^2
-	movd ecx,%1
-	sub ecx,eax
-
-	mov edx,[esp + 24]; pdevo
-	mov [edx],ecx
-
-	mov eax,[esp + 20];lumc
-	mul eax; lumc^2
-        add eax, 32
-	shr eax,6; 64*lumc^2
-	movd ecx,%2
-	sub ecx,eax
-
-	mov edx,[esp + 28]; pdevc
-	mov [edx],ecx
-
-	mov eax,[esp + 16];lumo
-	mul dword [esp + 20]; lumo*lumc, should fit in eax
-        add eax, 32
-	shr eax,6; 64*lumo*lumc
-	movd ecx,%3
-	sub ecx,eax
+	mov _EAX,prm4;lumo
+	mul _EAX; lumo^2
+        add _EAX, 32
+	shr _EAX,6; 64*lum0^2
+	movd TMP0d,%1
+	sub TMP0,_EAX
+
+	mov TMP1,prm6; pdevo
+	mov [TMP1],TMP0
+
+	mov eax,prm5d ;lumc
+	mul _EAX; lumc^2
+        add _EAX, 32
+	shr _EAX,6; 64*lumc^2
+	movd TMP0d,%2
+	sub TMP0,_EAX
+
+	mov TMP1,prm7; pdevc
+	mov [TMP1],TMP0
+
+	mov _EAX,prm4;lumo
+	mul prm5d; lumo*lumc, should fit in _EAX
+        add _EAX, 32
+	shr _EAX,6; 64*lumo*lumc
+	movd TMP0d,%3
+	sub TMP0,_EAX
 
-	mov edx,[esp + 32]; pcorr
-	mov [edx],ecx
+	mov TMP1,prm8; pcorr
+	mov [TMP1],TMP0
 %endmacro
 
 
-SECTION .text
+SECTION .rotext align=SECTION_ALIGN
 
 cglobal lum_8x8_mmx
 cglobal consim_sse2
@@ -154,10 +134,10 @@
 
 ;int lum_8x8_c(uint8_t* ptr, uint32_t stride)
 
-ALIGN 16
+ALIGN SECTION_ALIGN
 lum_8x8_mmx:
-	mov ecx, [esp + 4] ;ptr
-	mov edx, [esp + 8];stride
+	mov TMP0, prm1 ;ptr
+	mov TMP1, prm2 ;stride
 
 	pxor mm0,mm0
 
@@ -177,11 +157,11 @@
 	ret
 ENDFUNC
 
-ALIGN 16
+ALIGN SECTION_ALIGN
 consim_sse2:
-	mov ecx,[esp+4] ;ptro
-	mov edx,[esp+8] ;ptrc
-	mov eax,[esp+12];stride
+	mov TMP0,prm1 ;ptro
+	mov TMP1,prm2 ;ptrc
+	mov _EAX, prm3 ;stride
 
 	pxor xmm2,xmm2;null vektor
 	pxor xmm5,xmm5;devo
@@ -197,26 +177,26 @@
 	pshufd xmm7,xmm7,00000000b
 
 	CONSIM_1x8_SSE2
-	add ecx,eax
-	add edx,eax
+	add TMP0,_EAX
+	add TMP1,_EAX
 	CONSIM_1x8_SSE2
-	add ecx,eax
-	add edx,eax
+	add TMP0,_EAX
+	add TMP1,_EAX
 	CONSIM_1x8_SSE2
-	add ecx,eax
-	add edx,eax
+	add TMP0,_EAX
+	add TMP1,_EAX
 	CONSIM_1x8_SSE2
-	add ecx,eax
-	add edx,eax
+	add TMP0,_EAX
+	add TMP1,_EAX
 	CONSIM_1x8_SSE2
-	add ecx,eax
-	add edx,eax
+	add TMP0,_EAX
+	add TMP1,_EAX
 	CONSIM_1x8_SSE2
-	add ecx,eax
-	add edx,eax
+	add TMP0,_EAX
+	add TMP1,_EAX
 	CONSIM_1x8_SSE2
-	add ecx,eax
-	add edx,eax
+	add TMP0,_EAX
+	add TMP1,_EAX
 	CONSIM_1x8_SSE2
 
 	;accumulate xmm5-7
@@ -243,37 +223,37 @@
 
 
 
-ALIGN 16
+ALIGN SECTION_ALIGN
 consim_mmx:
-	mov ecx,[esp+4] ;ptro
-	mov edx,[esp+8] ;ptrc
-	mov eax,[esp+12];stride
+	mov TMP0,prm1 ;ptro
+	mov TMP1,prm2 ;ptrc
+	mov _EAX,prm3;stride
 	pxor mm2,mm2;null
 	pxor mm5,mm5;devo
 	pxor mm6,mm6;devc
 	pxor mm7,mm7;corr
 
 	CONSIM_1x8_MMX
-	add ecx,eax
-	add edx,eax
+	add TMP0,_EAX
+	add TMP1,_EAX
 	CONSIM_1x8_MMX
-	add ecx,eax
-	add edx,eax
+	add TMP0,_EAX
+	add TMP1,_EAX
 	CONSIM_1x8_MMX
-	add ecx,eax
-	add edx,eax
+	add TMP0,_EAX
+	add TMP1,_EAX
 	CONSIM_1x8_MMX
-	add ecx,eax
-	add edx,eax
+	add TMP0,_EAX
+	add TMP1,_EAX
 	CONSIM_1x8_MMX
-	add ecx,eax
-	add edx,eax
+	add TMP0,_EAX
+	add TMP1,_EAX
 	CONSIM_1x8_MMX
-	add ecx,eax
-	add edx,eax
+	add TMP0,_EAX
+	add TMP1,_EAX
 	CONSIM_1x8_MMX
-	add ecx,eax
-	add edx,eax
+	add TMP0,_EAX
+	add TMP1,_EAX
 	CONSIM_1x8_MMX
 
 	movq mm0,mm5