Attachment 147899 Details for Bug 179407 – PATCH removing all textrel's

[patch] PATCH removing all textrel's

SDL-1.2.10-PIC.patch (text/plain), 18.31 KB, created by Hans de Goede on 2007-02-12 15:16:08 UTC

(hide)

Description:

Filename:

MIME Type:

Creator: Hans de Goede

Created: 2007-02-12 15:16:08 UTC

Size: 18.31 KB

patch

obsolete

>diff -ur SDL-1.2.10.orig/src/hermes/mmx_main.asm SDL-1.2.10/src/hermes/mmx_main.asm
>--- SDL-1.2.10.orig/src/hermes/mmx_main.asm	2006-05-01 10:02:38.000000000 +0200
>+++ SDL-1.2.10/src/hermes/mmx_main.asm	2007-02-12 14:57:38.000000000 +0100
>@@ -10,7 +10,6 @@
> BITS 32
> 
> GLOBAL _ConvertMMX
>-GLOBAL _mmxreturn
> 
> SECTION .text
> 		
>@@ -49,9 +48,8 @@
> y_loop:	
> 	mov ecx,[ebp+4]
> 
>-	jmp [ebp+32]
>+	call [ebp+32]
> 
>-_mmxreturn:	
> 	add esi,[ebp+12]
> 	add edi,[ebp+28]
> 	
>diff -ur SDL-1.2.10.orig/src/hermes/mmxp2_32.asm SDL-1.2.10/src/hermes/mmxp2_32.asm
>--- SDL-1.2.10.orig/src/hermes/mmxp2_32.asm	2006-05-01 10:02:38.000000000 +0200
>+++ SDL-1.2.10/src/hermes/mmxp2_32.asm	2007-02-12 14:58:25.000000000 +0100
>@@ -27,8 +27,6 @@
> GLOBAL _ConvertMMXpII32_16RGB555
> GLOBAL _ConvertMMXpII32_16BGR555
> 
>-EXTERN _mmxreturn
>- 
> ;; Macros for conversion routines
> 
> %macro _push_immq_mask 1
>@@ -120,7 +118,7 @@
>         dec ecx
>         jnz .L3
> .L4:
>-        jmp _mmxreturn
>+        return
> 
> 
> 
>@@ -189,7 +187,7 @@
>         jnz .L3
> 
> .L4:
>-	jmp _mmxreturn
>+	retn
> 
> 	
> _ConvertMMXpII32_16BGR565:
>@@ -259,7 +257,7 @@
>         jnz .L3
> 
> .L4:
>-        jmp _mmxreturn
>+        retn
> 
> _ConvertMMXpII32_16BGR555:
> 
>@@ -399,7 +397,7 @@
> 	jnz .L3	
> 
> .L4:		
>-	jmp _mmxreturn
>+	retn
> 
> %ifidn __OUTPUT_FORMAT__,elf
> section .note.GNU-stack noalloc noexec nowrite progbits
>diff -ur SDL-1.2.10.orig/src/hermes/x86_main.asm SDL-1.2.10/src/hermes/x86_main.asm
>--- SDL-1.2.10.orig/src/hermes/x86_main.asm	2006-05-01 10:02:38.000000000 +0200
>+++ SDL-1.2.10/src/hermes/x86_main.asm	2007-02-12 15:19:26.000000000 +0100
>@@ -12,10 +12,7 @@
> BITS 32
> 
> GLOBAL _ConvertX86
>-GLOBAL _x86return
> 	
>-GLOBAL _Hermes_X86_CPU
>-
> SECTION .text
> 		
> ;; _ConvertX86:	 
>@@ -53,9 +50,8 @@
> y_loop:	
> 	mov ecx,[ebp+4]
> 
>-	jmp [ebp+32]
>+	call [ebp+32]
> 
>-_x86return:	
> 	add esi,[ebp+12]
> 	add edi,[ebp+28]
> 	
>@@ -72,49 +68,6 @@
> 
> 
> 
>-;; Hermes_X86_CPU returns the CPUID flags in eax
>-	
>-_Hermes_X86_CPU:
>-	pushfd
>-	pop eax
>-	
>-	mov ecx,eax
>-	
>-	xor eax,040000h
>-	push eax
>-	
>-	popfd
>-	pushfd
>-
>-	pop eax
>-	xor eax,ecx
>-	jz .L1			; Processor is 386
>-
>-	push ecx
>-	popfd
>-
>-	mov eax,ecx
>-	xor eax,200000h
>-
>-	push eax
>-	popfd
>-	pushfd
>-
>-	pop eax
>-	xor eax,ecx
>-	je .L1
>-	
>-	push ebx
>-
>-	mov eax,1
>-	cpuid
>-	mov eax,edx
>-
>-	pop ebx
>-
>-.L1:	
>-	ret
>-
> %ifidn __OUTPUT_FORMAT__,elf
> section .note.GNU-stack noalloc noexec nowrite progbits
> %endif
>diff -ur SDL-1.2.10.orig/src/hermes/x86p_16.asm SDL-1.2.10/src/hermes/x86p_16.asm
>--- SDL-1.2.10.orig/src/hermes/x86p_16.asm	2006-05-01 10:02:38.000000000 +0200
>+++ SDL-1.2.10/src/hermes/x86p_16.asm	2007-02-12 14:57:38.000000000 +0100
>@@ -25,7 +25,6 @@
> GLOBAL _ConvertX86p16_8RGB332
> 
> EXTERN _ConvertX86
>-EXTERN _x86return
> 
> SECTION .text
> 
>@@ -54,7 +53,7 @@
>     dec ecx
>     jnz .L1
> .L2
>-    jmp _x86return
>+    retn
> 
> .L3 ; head
>     mov eax,edi
>@@ -132,7 +131,7 @@
>     add edi,BYTE 2
> 
> .L7
>-    jmp _x86return
>+    retn
> 
> 
> 
>@@ -161,7 +160,7 @@
>     dec ecx
>     jnz .L1
> .L2
>-    jmp _x86return
>+    retn
> 
> .L3 ; head
>     mov eax,edi
>@@ -240,7 +239,7 @@
>     jmp SHORT .L6
> 
> .L7 pop ebp
>-    jmp _x86return
>+    retn
> 
> 
> 
>@@ -274,7 +273,7 @@
>     dec ecx
>     jnz .L1
> .L2
>-    jmp _x86return
>+    retn
> 
> .L3 ; head
>     mov eax,edi
>@@ -358,7 +357,7 @@
>     add edi,BYTE 2
> 
> .L7
>-    jmp _x86return
>+    retn
> 
> 
> 
>@@ -391,7 +390,7 @@
>     dec ecx
>     jnz .L1
> .L2
>-    jmp _x86return
>+    retn
> 
> .L3 mov eax,edi
>     and eax,BYTE 11b
>@@ -489,7 +488,7 @@
>     jnz .L6
> 
> .L7 pop ebp
>-    jmp _x86return
>+    retn
> 
> %ifidn __OUTPUT_FORMAT__,elf
> section .note.GNU-stack noalloc noexec nowrite progbits
>diff -ur SDL-1.2.10.orig/src/hermes/x86p_32.asm SDL-1.2.10/src/hermes/x86p_32.asm
>--- SDL-1.2.10.orig/src/hermes/x86p_32.asm	2006-05-01 10:02:38.000000000 +0200
>+++ SDL-1.2.10/src/hermes/x86p_32.asm	2007-02-12 14:58:36.000000000 +0100
>@@ -22,8 +22,6 @@
> GLOBAL _ConvertX86p32_16RGB555
> GLOBAL _ConvertX86p32_16BGR555
> GLOBAL _ConvertX86p32_8RGB332
>-
>-EXTERN _x86return
> 	
> SECTION .text
> 
>@@ -52,7 +50,7 @@
>     dec ecx
>     jnz .L1
> .L2
>-    jmp _x86return
>+    retn
> 
> .L3 ; save ebp
>     push ebp
>@@ -112,7 +110,7 @@
>     jnz .L5
> 
> .L6 pop ebp
>-    jmp _x86return
>+    retn
> 	
> 
> 	
>@@ -132,7 +130,7 @@
>     dec ecx
>     jnz .L1
> .L2
>-    jmp _x86return
>+    retn
> 
> .L3 ; save ebp
>     push ebp
>@@ -183,7 +181,7 @@
>     jnz .L5
> 
> .L6 pop ebp
>-    jmp _x86return
>+    retn
> 
> 	
> 
>@@ -203,7 +201,7 @@
>     dec ecx
>     jnz .L1
> .L2
>-    jmp _x86return
>+    retn
> 
> .L3 ; save ebp
>     push ebp
>@@ -256,7 +254,7 @@
>     jnz .L5
> 
> .L6 pop ebp
>-    jmp _x86return
>+    retn
> 
> 
> 	
>@@ -281,7 +279,7 @@
> 	dec ecx
> 	jnz .L1
> .L2 
>-	jmp _x86return
>+	retn
> 
> .L3	;	 head
> 	mov edx,edi
>@@ -353,7 +351,7 @@
> 	jnz .L6
> 
> .L7	pop ebp
>-	jmp _x86return
>+	retn
> 
> 
> 
>@@ -379,7 +377,7 @@
> 	dec ecx
> 	jnz .L1
> .L2
>-	jmp _x86return
>+	retn
> 
> .L3 ; head
> 	mov edx,edi
>@@ -454,7 +452,7 @@
> 
> .L7 
> 	pop ebp
>-	jmp _x86return
>+	retn
>  
> 
> 	
>@@ -483,7 +481,7 @@
> 	jnz .L1
> 
> .L2:				; End of short loop
>-	jmp _x86return
>+	retn
> 
> 	
> .L3	; head
>@@ -569,7 +567,7 @@
> 	add edi,BYTE 2
> 
> .L7:	
>-	jmp _x86return
>+	retn
> 
> 
> 
>@@ -598,7 +596,7 @@
> 	dec ecx
> 	jnz .L1
> .L2
>-	jmp _x86return
>+	retn
> 
> .L3	; head
> 	mov ebx,edi
>@@ -683,7 +681,7 @@
> 	add edi,BYTE 2
> 
> .L7 
>-	jmp _x86return
>+	retn
> 
> 
> 	
>@@ -712,7 +710,7 @@
> 	dec ecx
> 	jnz .L1
> .L2
>-	jmp _x86return
>+	retn
> 
> .L3	; head
> 	mov ebx,edi
>@@ -794,7 +792,7 @@
> 	add edi,BYTE 2
> 
> .L7
>-	jmp _x86return
>+	retn
> 
> 
> 
>@@ -824,7 +822,7 @@
> 	dec ecx
> 	jnz .L1
> .L2 
>-	jmp _x86return
>+	retn
> 
> .L3	; head
> 	mov ebx,edi
>@@ -909,7 +907,7 @@
> 	add edi,BYTE 2
> 
> .L7
>-	jmp _x86return
>+	retn
> 
> 
> 
>@@ -1039,7 +1037,7 @@
> 	jnz .L3
> 	
> .L4:	
>-	jmp _x86return
>+	retn
> 
> %ifidn __OUTPUT_FORMAT__,elf
> section .note.GNU-stack noalloc noexec nowrite progbits
>diff -ur SDL-1.2.10.orig/src/video/SDL_yuv_mmx.c SDL-1.2.10/src/video/SDL_yuv_mmx.c
>--- SDL-1.2.10.orig/src/video/SDL_yuv_mmx.c	2006-05-01 10:02:48.000000000 +0200
>+++ SDL-1.2.10/src/video/SDL_yuv_mmx.c	2007-02-12 15:18:34.000000000 +0100
>@@ -26,31 +26,31 @@
> 
> #include "SDL_stdinc.h"
> 
>-#define ASM_ARRAY(x) x[] __asm__("_" #x) __attribute__((used))
>- 
>-static unsigned int  ASM_ARRAY(MMX_0080w)    = {0x00800080, 0x00800080};
>-static unsigned int  ASM_ARRAY(MMX_00FFw)    = {0x00ff00ff, 0x00ff00ff};
>-static unsigned int  ASM_ARRAY(MMX_FF00w)    = {0xff00ff00, 0xff00ff00};
>-
>-static unsigned short ASM_ARRAY(MMX_Ycoeff)  = {0x004a, 0x004a, 0x004a, 0x004a};
>-
>-static unsigned short ASM_ARRAY(MMX_UbluRGB) = {0x0072, 0x0072, 0x0072, 0x0072};
>-static unsigned short ASM_ARRAY(MMX_VredRGB) = {0x0059, 0x0059, 0x0059, 0x0059};
>-static unsigned short ASM_ARRAY(MMX_UgrnRGB) = {0xffea, 0xffea, 0xffea, 0xffea};
>-static unsigned short ASM_ARRAY(MMX_VgrnRGB) = {0xffd2, 0xffd2, 0xffd2, 0xffd2};
>-
>-static unsigned short ASM_ARRAY(MMX_Ublu5x5) = {0x0081, 0x0081, 0x0081, 0x0081};
>-static unsigned short ASM_ARRAY(MMX_Vred5x5) = {0x0066, 0x0066, 0x0066, 0x0066};
>-static unsigned short ASM_ARRAY(MMX_Ugrn555) = {0xffe7, 0xffe7, 0xffe7, 0xffe7};
>-static unsigned short ASM_ARRAY(MMX_Vgrn555) = {0xffcc, 0xffcc, 0xffcc, 0xffcc};
>-static unsigned short ASM_ARRAY(MMX_Ugrn565) = {0xffe8, 0xffe8, 0xffe8, 0xffe8};
>-static unsigned short ASM_ARRAY(MMX_Vgrn565) = {0xffcd, 0xffcd, 0xffcd, 0xffcd};
>-
>-static unsigned short ASM_ARRAY(MMX_red555)  = {0x7c00, 0x7c00, 0x7c00, 0x7c00};
>-static unsigned short ASM_ARRAY(MMX_red565)  = {0xf800, 0xf800, 0xf800, 0xf800};
>-static unsigned short ASM_ARRAY(MMX_grn555)  = {0x03e0, 0x03e0, 0x03e0, 0x03e0};
>-static unsigned short ASM_ARRAY(MMX_grn565)  = {0x07e0, 0x07e0, 0x07e0, 0x07e0};
>-static unsigned short ASM_ARRAY(MMX_blu5x5)  = {0x001f, 0x001f, 0x001f, 0x001f};
>+#include "mmx.h"
>+
>+static mmx_t MMX_0080w    = { .ud = {0x00800080, 0x00800080} };
>+static mmx_t MMX_00FFw    = { .ud = {0x00ff00ff, 0x00ff00ff} };
>+static mmx_t MMX_FF00w    = { .ud = {0xff00ff00, 0xff00ff00} };
>+
>+static mmx_t MMX_Ycoeff   = { .uw = {0x004a, 0x004a, 0x004a, 0x004a} };
>+
>+static mmx_t MMX_UbluRGB  = { .uw = {0x0072, 0x0072, 0x0072, 0x0072} };
>+static mmx_t MMX_VredRGB  = { .uw = {0x0059, 0x0059, 0x0059, 0x0059} };
>+static mmx_t MMX_UgrnRGB  = { .uw = {0xffea, 0xffea, 0xffea, 0xffea} };
>+static mmx_t MMX_VgrnRGB  = { .uw = {0xffd2, 0xffd2, 0xffd2, 0xffd2} };
>+
>+static mmx_t MMX_Ublu5x5  = { .uw = {0x0081, 0x0081, 0x0081, 0x0081} };
>+static mmx_t MMX_Vred5x5  = { .uw = {0x0066, 0x0066, 0x0066, 0x0066} };
>+static mmx_t MMX_Ugrn555  = { .uw = {0xffe7, 0xffe7, 0xffe7, 0xffe7} };
>+static mmx_t MMX_Vgrn555  = { .uw = {0xffcc, 0xffcc, 0xffcc, 0xffcc} };
>+static mmx_t MMX_Ugrn565  = { .uw = {0xffe8, 0xffe8, 0xffe8, 0xffe8} };
>+static mmx_t MMX_Vgrn565  = { .uw = {0xffcd, 0xffcd, 0xffcd, 0xffcd} };
>+
>+static mmx_t MMX_red555   = { .uw = {0x7c00, 0x7c00, 0x7c00, 0x7c00} };
>+static mmx_t MMX_red565   = { .uw = {0xf800, 0xf800, 0xf800, 0xf800} };
>+static mmx_t MMX_grn555   = { .uw = {0x03e0, 0x03e0, 0x03e0, 0x03e0} };
>+static mmx_t MMX_grn565   = { .uw = {0x07e0, 0x07e0, 0x07e0, 0x07e0} };
>+static mmx_t MMX_blu5x5   = { .uw = {0x001f, 0x001f, 0x001f, 0x001f} };
> 
> /**
>    This MMX assembler is my first assembler/MMX program ever.
>@@ -112,12 +112,12 @@
> 		 "movd (%2), %%mm2\n"           //    0  0  0  0 l3 l2 l1 l0
> 		 "punpcklbw %%mm7,%%mm1\n" //         0  v3 0  v2 00 v1 00 v0
> 		 "punpckldq %%mm1,%%mm1\n" //         00 v1 00 v0 00 v1 00 v0
>-		 "psubw _MMX_0080w,%%mm1\n"  // mm1-128:r1 r1 r0 r0 r1 r1 r0 r0 
>+		 "psubw %9,%%mm1\n"        // mm1-128:r1 r1 r0 r0 r1 r1 r0 r0 
> 
> 		 // create Cr_g (result in mm0)
> 		 "movq %%mm1,%%mm0\n"           // r1 r1 r0 r0 r1 r1 r0 r0
>-		 "pmullw _MMX_VgrnRGB,%%mm0\n"// red*-46dec=0.7136*64
>-		 "pmullw _MMX_VredRGB,%%mm1\n"// red*89dec=1.4013*64
>+		 "pmullw %10,%%mm0\n"           // red*-46dec=0.7136*64
>+		 "pmullw %11,%%mm1\n"           // red*89dec=1.4013*64
> 		 "psraw  $6, %%mm0\n"           // red=red/64
> 		 "psraw  $6, %%mm1\n"           // red=red/64
> 		 
>@@ -126,8 +126,8 @@
> 		 "movq (%2,%4),%%mm3\n"         //    0  0  0  0 L3 L2 L1 L0
> 		 "punpckldq %%mm3,%%mm2\n"      //   L3 L2 L1 L0 l3 l2 l1 l0
> 		 "movq %%mm2,%%mm4\n"           //   L3 L2 L1 L0 l3 l2 l1 l0
>-		 "pand _MMX_FF00w,%%mm2\n"      //   L3 0  L1  0 l3  0 l1  0
>-		 "pand _MMX_00FFw,%%mm4\n"      //   0  L2  0 L0  0 l2  0 l0
>+		 "pand %12,%%mm2\n"             //   L3 0  L1  0 l3  0 l1  0
>+		 "pand %13,%%mm4\n"             //   0  L2  0 L0  0 l2  0 l0
> 		 "psrlw $8,%%mm2\n"             //   0  L3  0 L1  0 l3  0 l1
> 
> 		 // create R (result in mm6)
>@@ -144,11 +144,11 @@
> 		 "movd (%1), %%mm1\n"      //         0  0  0  0  u3 u2 u1 u0
> 		 "punpcklbw %%mm7,%%mm1\n" //         0  u3 0  u2 00 u1 00 u0
> 		 "punpckldq %%mm1,%%mm1\n" //         00 u1 00 u0 00 u1 00 u0
>-		 "psubw _MMX_0080w,%%mm1\n"  // mm1-128:u1 u1 u0 u0 u1 u1 u0 u0 
>+		 "psubw %9,%%mm1\n"        // mm1-128:u1 u1 u0 u0 u1 u1 u0 u0 
> 		 // create Cb_g (result in mm5)
> 		 "movq %%mm1,%%mm5\n"            // u1 u1 u0 u0 u1 u1 u0 u0
>-		 "pmullw _MMX_UgrnRGB,%%mm5\n"    // blue*-109dec=1.7129*64
>-		 "pmullw _MMX_UbluRGB,%%mm1\n"    // blue*114dec=1.78125*64
>+		 "pmullw %14,%%mm5\n"            // blue*-109dec=1.7129*64
>+		 "pmullw %15,%%mm1\n"            // blue*114dec=1.78125*64
> 		 "psraw  $6, %%mm5\n"            // blue=red/64
> 		 "psraw  $6, %%mm1\n"            // blue=blue/64
> 
>@@ -230,7 +230,10 @@
> 		 "popl %%ebx\n"
> 		 :
> 		 : "m" (cr), "r"(cb),"r"(lum),
>-		 "r"(row1),"r"(cols),"r"(row2),"m"(x),"m"(y),"m"(mod));
>+		 "r"(row1),"r"(cols),"r"(row2),"m"(x),"m"(y),"m"(mod),
>+		 "m"(MMX_0080w),"m"(MMX_VgrnRGB),"m"(MMX_VredRGB),
>+		 "m"(MMX_FF00w),"m"(MMX_00FFw),"m"(MMX_UgrnRGB),
>+		 "m"(MMX_UbluRGB));
> }
> 
> void Color565DitherYV12MMX1X( int *colortab, Uint32 *rgb_2_pix,
>@@ -259,21 +262,21 @@
>          "movd           (%%ebx),                %%mm1\n" // 4 Cr                0  0  0  0 v3 v2 v1 v0
>          "punpcklbw      %%mm7,                  %%mm0\n" // 4 W cb   0 u3  0 u2  0 u1  0 u0
>          "punpcklbw      %%mm7,                  %%mm1\n" // 4 W cr   0 v3  0 v2  0 v1  0 v0
>-         "psubw          _MMX_0080w,             %%mm0\n"
>-         "psubw          _MMX_0080w,             %%mm1\n"
>+         "psubw          %9,                     %%mm0\n"
>+         "psubw          %9,                     %%mm1\n"
>          "movq           %%mm0,                  %%mm2\n" // Cb                   0 u3  0 u2  0 u1  0 u0
>          "movq           %%mm1,                  %%mm3\n" // Cr
>-         "pmullw         _MMX_Ugrn565,           %%mm2\n" // Cb2green 0 R3  0 R2  0 R1  0 R0
>+         "pmullw         %10,                    %%mm2\n" // Cb2green 0 R3  0 R2  0 R1  0 R0
>          "movq           (%2),                   %%mm6\n" // L1      l7 L6 L5 L4 L3 L2 L1 L0
>-         "pmullw         _MMX_Ublu5x5,           %%mm0\n" // Cb2blue
>-         "pand           _MMX_00FFw,             %%mm6\n" // L1      00 L6 00 L4 00 L2 00 L0
>-         "pmullw         _MMX_Vgrn565,           %%mm3\n" // Cr2green
>+         "pmullw         %11,                    %%mm0\n" // Cb2blue
>+         "pand           %12,                    %%mm6\n" // L1      00 L6 00 L4 00 L2 00 L0
>+         "pmullw         %13,                    %%mm3\n" // Cr2green
>          "movq           (%2),                   %%mm7\n" // L2
>-         "pmullw         _MMX_Vred5x5,           %%mm1\n" // Cr2red
>+         "pmullw         %14,                    %%mm1\n" // Cr2red
>          "psrlw          $8,                     %%mm7\n"        // L2           00 L7 00 L5 00 L3 00 L1
>-         "pmullw         _MMX_Ycoeff,            %%mm6\n" // lum1
>+         "pmullw         %15,                    %%mm6\n" // lum1
>          "paddw          %%mm3,                  %%mm2\n" // Cb2green + Cr2green == green
>-         "pmullw         _MMX_Ycoeff,            %%mm7\n" // lum2
>+         "pmullw         %15,                    %%mm7\n" // lum2
> 
>          "movq           %%mm6,                  %%mm4\n" // lum1
>          "paddw          %%mm0,                  %%mm6\n" // lum1 +blue 00 B6 00 B4 00 B2 00 B0
>@@ -291,11 +294,11 @@
>          "punpcklbw      %%mm4,                  %%mm4\n"
>          "punpcklbw      %%mm5,                  %%mm5\n"
> 
>-         "pand           _MMX_red565,            %%mm4\n"
>+         "pand           %16,                    %%mm4\n"
>          "psllw          $3,                     %%mm5\n" // GREEN       1
>          "punpcklbw      %%mm6,                  %%mm6\n"
>-         "pand           _MMX_grn565,            %%mm5\n"
>-         "pand           _MMX_red565,            %%mm6\n"
>+         "pand           %17,                    %%mm5\n"
>+         "pand           %16,                    %%mm6\n"
>          "por            %%mm5,                  %%mm4\n" //
>          "psrlw          $11,                    %%mm6\n" // BLUE        1
>          "movq           %%mm3,                  %%mm5\n" // lum2
>@@ -309,23 +312,23 @@
>          "packuswb       %%mm3,                  %%mm3\n"
>          "packuswb       %%mm5,                  %%mm5\n"
>          "packuswb       %%mm7,                  %%mm7\n"
>-         "pand           _MMX_00FFw,             %%mm6\n" // L3
>+         "pand           %12,                    %%mm6\n" // L3
>          "punpcklbw      %%mm3,                  %%mm3\n"
>          "punpcklbw      %%mm5,                  %%mm5\n"
>-         "pmullw         _MMX_Ycoeff,            %%mm6\n" // lum3
>+         "pmullw         %15,                    %%mm6\n" // lum3
>          "punpcklbw      %%mm7,                  %%mm7\n"
>          "psllw          $3,                     %%mm5\n" // GREEN 2
>-         "pand           _MMX_red565,            %%mm7\n"
>-         "pand           _MMX_red565,            %%mm3\n"
>+         "pand           %16,                    %%mm7\n"
>+         "pand           %16,                    %%mm3\n"
>          "psrlw          $11,                    %%mm7\n" // BLUE  2
>-         "pand           _MMX_grn565,            %%mm5\n"
>+         "pand           %17,                    %%mm5\n"
>          "por            %%mm7,                  %%mm3\n"
>          "movq           (%2,%4),                %%mm7\n" // L4 load lum2
>          "por            %%mm5,                  %%mm3\n" //
>          "psrlw          $8,                     %%mm7\n" // L4
>          "movq           %%mm4,                  %%mm5\n"
>          "punpcklwd      %%mm3,                  %%mm4\n"
>-         "pmullw         _MMX_Ycoeff,            %%mm7\n" // lum4
>+         "pmullw         %15,                    %%mm7\n" // lum4
>          "punpckhwd      %%mm3,                  %%mm5\n"
> 
>          "movq           %%mm4,                  (%3)\n"  // write row1
>@@ -352,11 +355,11 @@
>          "punpcklbw      %%mm5,                  %%mm5\n"
>          "punpcklbw      %%mm6,                  %%mm6\n"
>          "psllw          $3,                     %%mm5\n" // GREEN 3
>-         "pand           _MMX_red565,            %%mm4\n"
>+         "pand           %16,                    %%mm4\n"
>          "psraw          $6,                     %%mm3\n" // psr 6
>          "psraw          $6,                     %%mm0\n"
>-         "pand           _MMX_red565,            %%mm6\n" // BLUE
>-         "pand           _MMX_grn565,            %%mm5\n"
>+         "pand           %16,                    %%mm6\n" // BLUE
>+         "pand           %17,                    %%mm5\n"
>          "psrlw          $11,                    %%mm6\n" // BLUE  3
>          "por            %%mm5,                  %%mm4\n"
>          "psraw          $6,                     %%mm7\n"
>@@ -367,11 +370,11 @@
>          "punpcklbw      %%mm3,                  %%mm3\n"
>          "punpcklbw      %%mm0,                  %%mm0\n"
>          "punpcklbw      %%mm7,                  %%mm7\n"
>-         "pand           _MMX_red565,            %%mm3\n"
>-         "pand           _MMX_red565,            %%mm7\n" // BLUE
>+         "pand           %16,                    %%mm3\n"
>+         "pand           %16,                    %%mm7\n" // BLUE
>          "psllw          $3,                     %%mm0\n" // GREEN 4
>          "psrlw          $11,                    %%mm7\n"
>-         "pand           _MMX_grn565,            %%mm0\n"
>+         "pand           %17,                    %%mm0\n"
>          "por            %%mm7,                  %%mm3\n"
>          "por            %%mm0,                  %%mm3\n"
> 
>@@ -403,7 +406,10 @@
> 	 "popl %%ebx\n"
>          :
>          :"m" (cr), "r"(cb),"r"(lum),
>-	 "r"(row1),"r"(cols),"r"(row2),"m"(x),"m"(y),"m"(mod));
>+	 "r"(row1),"r"(cols),"r"(row2),"m"(x),"m"(y),"m"(mod),
>+	 "m"(MMX_0080w),"m"(MMX_Ugrn565),"m"(MMX_Ublu5x5),
>+	 "m"(MMX_00FFw),"m"(MMX_Vgrn565),"m"(MMX_Vred5x5),
>+	 "m"(MMX_Ycoeff),"m"(MMX_red565),"m"(MMX_grn565));
> }
> 
> #endif /* GCC i386 inline assembly */

Actions: View | Diff

Attachments on bug 179407: 147899