1
diff -ruN -x Makefile.in -x configure SDL-1.2.7.orig/src/audio/SDL_mixer_MMX.c SDL-1.2.7/src/audio/SDL_mixer_MMX.c
2
--- SDL-1.2.7.orig/src/audio/SDL_mixer_MMX.c 2002-11-09 01:13:28.000000000 -0500
3
+++ SDL-1.2.7/src/audio/SDL_mixer_MMX.c 2004-06-08 13:26:50.000000000 -0400
8
-" movl %0,%%edi\n" // edi = dst
9
-" movl %1,%%esi\n" // esi = src
10
" movl %3,%%eax\n" // eax = volume
12
-" movl %2,%%ebx\n" // ebx = size
13
+" movl %2,%%edx\n" // edx = size
15
-" shrl $4,%%ebx\n" // process 16 bytes per iteration = 8 samples
16
+" shrl $4,%%edx\n" // process 16 bytes per iteration = 8 samples
24
-" movq (%%esi),%%mm1\n" // mm1 = a|b|c|d
25
+" movq (%1),%%mm1\n" // mm1 = a|b|c|d
27
" movq %%mm1,%%mm2\n" // mm2 = a|b|c|d
29
-" movq 8(%%esi),%%mm4\n" // mm4 = e|f|g|h
30
+" movq 8(%1),%%mm4\n" // mm4 = e|f|g|h
32
// pr� charger le buffer dst dans mm7
33
-" movq (%%edi),%%mm7\n" // mm7 = dst[0]"
34
+" movq (%0),%%mm7\n" // mm7 = dst[0]"
36
// multiplier par le volume
37
" pmullw %%mm0,%%mm1\n" // mm1 = l(a*v)|l(b*v)|l(c*v)|l(d*v)
39
" punpcklwd %%mm5,%%mm6\n" // mm6 = g*v|h*v
41
// pr� charger le buffer dst dans mm5
42
-" movq 8(%%edi),%%mm5\n" // mm5 = dst[1]
43
+" movq 8(%0),%%mm5\n" // mm5 = dst[1]
46
" psrad $7,%%mm1\n" // mm1 = a*v/128|b*v/128 , 128 = SDL_MIX_MAXVOLUME
50
" psrad $7,%%mm3\n" // mm3 = c*v/128|d*v/128
54
// mm4 = le sample avec le volume modifi�
55
" packssdw %%mm4,%%mm6\n" // mm6 = s(e*v|f*v|g*v|h*v)
56
-" movq %%mm3,(%%edi)\n"
59
" paddsw %%mm5,%%mm6\n" // mm6 = adjust_volume(src)+dst
61
-" movq %%mm6,8(%%edi)\n"
62
+" movq %%mm6,8(%0)\n"
76
- : "m" (dst), "m"(src),"m"(size),
77
+ : "r" (dst), "r"(src),"m"(size),
79
- : "eax","ebx", "esi", "edi","memory"
80
+ : "eax","edx","memory"
86
__asm__ __volatile__ (
88
-" movl %0,%%edi\n" // edi = dst
89
-" movl %1,%%esi\n" // esi = src
90
" movl %3,%%eax\n" // eax = volume
92
-" movd %%ebx,%%mm0\n"
93
+" movd %%edx,%%mm0\n"
101
-" movl %2,%%ebx\n" // ebx = size
102
-" shr $3,%%ebx\n" // process 8 bytes per iteration = 8 samples
103
+" movl %2,%%edx\n" // edx = size
104
+" shr $3,%%edx\n" // process 8 bytes per iteration = 8 samples
113
" pxor %%mm2,%%mm2\n" // mm2 = 0
114
-" movq (%%esi),%%mm1\n" // mm1 = a|b|c|d|e|f|g|h
115
+" movq (%1),%%mm1\n" // mm1 = a|b|c|d|e|f|g|h
117
" movq %%mm1,%%mm3\n" // mm3 = a|b|c|d|e|f|g|h
119
@@ -152,10 +148,10 @@
120
" punpckhbw %%mm2,%%mm1\n" // mm1 = 0|a|0|b|0|c|0|d
122
" punpcklbw %%mm2,%%mm3\n" // mm3 = 0|e|0|f|0|g|0|h
123
-" movq (%%edi),%%mm2\n" // mm2 = destination
124
+" movq (%0),%%mm2\n" // mm2 = destination
126
" pmullw %%mm0,%%mm1\n" // mm1 = v*a|v*b|v*c|v*d
130
" pmullw %%mm0,%%mm3\n" // mm3 = v*e|v*f|v*g|v*h
131
" psraw $7,%%mm1\n" // mm1 = v*a/128|v*b/128|v*c/128|v*d/128
132
@@ -166,19 +162,19 @@
134
" paddsb %%mm2,%%mm3\n" // add to destination buffer
136
-" movq %%mm3,(%%edi)\n" // store back to ram
138
+" movq %%mm3,(%0)\n" // store back to ram
149
- : "m" (dst), "m"(src),"m"(size),
150
+ : "r" (dst), "r"(src),"m"(size),
152
- : "eax","ebx", "esi", "edi","memory"
153
+ : "eax","edx","memory"
157
diff -ruN -x Makefile.in -x configure SDL-1.2.7.orig/src/cpuinfo/SDL_cpuinfo.c SDL-1.2.7/src/cpuinfo/SDL_cpuinfo.c
158
--- SDL-1.2.7.orig/src/cpuinfo/SDL_cpuinfo.c 2004-02-10 10:31:35.000000000 -0500
159
+++ SDL-1.2.7/src/cpuinfo/SDL_cpuinfo.c 2004-06-08 13:26:51.000000000 -0400
164
- "@(#) $Id: SDL_cpuinfo.c,v 1.14 2004/02/10 15:31:35 slouken Exp $";
165
+ "@(#) $Id: SDL_cpuinfo.c,v 1.16 2004/05/16 17:19:48 slouken Exp $";
168
/* CPU feature detection for SDL */
173
+#elif defined(__GNUC__) && defined(__x86_64__)
174
+/* Technically, if this is being compiled under __x86_64__ then it has
175
+CPUid by definition. But it's nice to be able to prove it. :) */
177
+" pushfq # Get original EFLAGS \n"
179
+" movq %%rax,%%rcx \n"
180
+" xorl $0x200000,%%eax # Flip ID bit in EFLAGS \n"
181
+" pushq %%rax # Save new EFLAGS value on stack \n"
182
+" popfq # Replace current EFLAGS value \n"
183
+" pushfq # Get new EFLAGS \n"
184
+" popq %%rax # Store new EFLAGS in EAX \n"
185
+" xorl %%ecx,%%eax # Can not toggle ID bit, \n"
186
+" jz 1f # Processor=80486 \n"
187
+" movl $1,%0 # We have CPUID support \n"
193
#elif defined(_MSC_VER)
195
pushfd ; Get original EFLAGS
197
static __inline__ int CPU_getCPUIDFeatures()
200
-#if defined(__GNUC__) && defined(i386)
201
+#if defined(__GNUC__) && ( defined(i386) || defined(__x86_64__) )
203
" movl %%ebx,%%edi\n"
204
" xorl %%eax,%%eax # Set up for CPUID instruction \n"
206
" movl %%edi,%%ebx\n"
209
- : "%eax", "%ebx", "%ecx", "%edx", "%edi"
210
+ : "%eax", "%ecx", "%edx", "%edi"
212
#elif defined(_MSC_VER)
215
static __inline__ int CPU_getCPUIDFeaturesExt()
218
-#if defined(__GNUC__) && defined(i386)
219
+#if defined(__GNUC__) && (defined(i386) || defined (__x86_64__) )
221
" movl %%ebx,%%edi\n"
222
" movl $0x80000000,%%eax # Query for extended functions \n"
224
" movl %%edi,%%ebx\n"
227
- : "%eax", "%ebx", "%ecx", "%edx", "%edi"
228
+ : "%eax", "%ecx", "%edx", "%edi"
230
#elif defined(_MSC_VER)
232
diff -ruN -x Makefile.in -x configure SDL-1.2.7.orig/src/video/SDL_yuv_mmx.c SDL-1.2.7/src/video/SDL_yuv_mmx.c
233
--- SDL-1.2.7.orig/src/video/SDL_yuv_mmx.c 2004-02-18 12:22:04.000000000 -0500
234
+++ SDL-1.2.7/src/video/SDL_yuv_mmx.c 2004-06-08 13:26:51.000000000 -0400
239
- "@(#) $Id: SDL_yuv_mmx.c,v 1.5 2004/01/04 16:49:22 slouken Exp $";
240
+ "@(#) $Id: SDL_yuv_mmx.c,v 1.6 2004/05/16 17:19:48 slouken Exp $";
244
@@ -120,12 +120,12 @@
245
"movd (%2), %%mm2\n" // 0 0 0 0 l3 l2 l1 l0
246
"punpcklbw %%mm7,%%mm1\n" // 0 v3 0 v2 00 v1 00 v0
247
"punpckldq %%mm1,%%mm1\n" // 00 v1 00 v0 00 v1 00 v0
248
- "psubw _MMX_0080w,%%mm1\n" // mm1-128:r1 r1 r0 r0 r1 r1 r0 r0
249
+ "psubw %[_MMX_0080w],%%mm1\n" // mm1-128:r1 r1 r0 r0 r1 r1 r0 r0
251
// create Cr_g (result in mm0)
252
"movq %%mm1,%%mm0\n" // r1 r1 r0 r0 r1 r1 r0 r0
253
- "pmullw _MMX_VgrnRGB,%%mm0\n"// red*-46dec=0.7136*64
254
- "pmullw _MMX_VredRGB,%%mm1\n"// red*89dec=1.4013*64
255
+ "pmullw %[_MMX_VgrnRGB],%%mm0\n"// red*-46dec=0.7136*64
256
+ "pmullw %[_MMX_VredRGB],%%mm1\n"// red*89dec=1.4013*64
257
"psraw $6, %%mm0\n" // red=red/64
258
"psraw $6, %%mm1\n" // red=red/64
261
"movq (%2,%4),%%mm3\n" // 0 0 0 0 L3 L2 L1 L0
262
"punpckldq %%mm3,%%mm2\n" // L3 L2 L1 L0 l3 l2 l1 l0
263
"movq %%mm2,%%mm4\n" // L3 L2 L1 L0 l3 l2 l1 l0
264
- "pand _MMX_FF00w,%%mm2\n" // L3 0 L1 0 l3 0 l1 0
265
- "pand _MMX_00FFw,%%mm4\n" // 0 L2 0 L0 0 l2 0 l0
266
+ "pand %[_MMX_FF00w],%%mm2\n" // L3 0 L1 0 l3 0 l1 0
267
+ "pand %[_MMX_00FFw],%%mm4\n" // 0 L2 0 L0 0 l2 0 l0
268
"psrlw $8,%%mm2\n" // 0 L3 0 L1 0 l3 0 l1
270
// create R (result in mm6)
271
@@ -152,11 +152,11 @@
272
"movd (%1), %%mm1\n" // 0 0 0 0 u3 u2 u1 u0
273
"punpcklbw %%mm7,%%mm1\n" // 0 u3 0 u2 00 u1 00 u0
274
"punpckldq %%mm1,%%mm1\n" // 00 u1 00 u0 00 u1 00 u0
275
- "psubw _MMX_0080w,%%mm1\n" // mm1-128:u1 u1 u0 u0 u1 u1 u0 u0
276
+ "psubw %[_MMX_0080w],%%mm1\n" // mm1-128:u1 u1 u0 u0 u1 u1 u0 u0
277
// create Cb_g (result in mm5)
278
"movq %%mm1,%%mm5\n" // u1 u1 u0 u0 u1 u1 u0 u0
279
- "pmullw _MMX_UgrnRGB,%%mm5\n" // blue*-109dec=1.7129*64
280
- "pmullw _MMX_UbluRGB,%%mm1\n" // blue*114dec=1.78125*64
281
+ "pmullw %[_MMX_UgrnRGB],%%mm5\n" // blue*-109dec=1.7129*64
282
+ "pmullw %[_MMX_UbluRGB],%%mm1\n" // blue*114dec=1.78125*64
283
"psraw $6, %%mm5\n" // blue=red/64
284
"psraw $6, %%mm1\n" // blue=blue/64
289
: "m" (cr), "r"(cb),"r"(lum),
290
- "r"(row1),"r"(cols),"r"(row2),"m"(x),"m"(y),"m"(mod)
292
+ "r"(row1),"r"(cols),"r"(row2),"m"(x),"m"(y),"m"(mod),
293
+ [_MMX_0080w] "m" (*_MMX_0080w),
294
+ [_MMX_00FFw] "m" (*_MMX_00FFw),
295
+ [_MMX_FF00w] "m" (*_MMX_FF00w),
296
+ [_MMX_VgrnRGB] "m" (*_MMX_VgrnRGB),
297
+ [_MMX_VredRGB] "m" (*_MMX_VredRGB),
298
+ [_MMX_UgrnRGB] "m" (*_MMX_UgrnRGB),
299
+ [_MMX_UbluRGB] "m" (*_MMX_UbluRGB)
306
:"m" (cr), "r"(cb),"r"(lum),
307
- "r"(row1),"r"(cols),"r"(row2),"m"(x),"m"(y),"m"(mod)
309
+ "r"(row1),"r"(cols),"r"(row2),"m"(x),"m"(y),"m"(mod),
310
+ [_MMX_0080w] "m" (*_MMX_0080w),
311
+ [_MMX_Ugrn565] "m" (*_MMX_Ugrn565),
312
+ [_MMX_Ublu5x5] "m" (*_MMX_Ublu5x5),
313
+ [_MMX_00FFw] "m" (*_MMX_00FFw),
314
+ [_MMX_Vgrn565] "m" (*_MMX_Vgrn565),
315
+ [_MMX_Vred5x5] "m" (*_MMX_Vred5x5),
316
+ [_MMX_Ycoeff] "m" (*_MMX_Ycoeff),
317
+ [_MMX_red565] "m" (*_MMX_red565),
318
+ [_MMX_grn565] "m" (*_MMX_grn565)