Mercurial > mplayer.hg
changeset 18104:7b408d60de9e
add support for intel mac. mp3lib is not fixed yet.
| author | nplourde |
|---|---|
| date | Sat, 15 Apr 2006 20:46:54 +0000 |
| parents | 26ea12332e67 |
| children | 954eac69b532 |
| files | asmalign.h configure liba52/downmix.c liba52/imdct.c libmpcodecs/pullup.c libmpcodecs/vf_decimate.c libmpcodecs/vf_divtc.c libmpcodecs/vf_eq.c libmpcodecs/vf_eq2.c libmpcodecs/vf_fspp.c libmpcodecs/vf_halfpack.c libmpcodecs/vf_ilpack.c libmpcodecs/vf_ivtc.c libmpcodecs/vf_noise.c mangle.h postproc/rgb2rgb_template.c postproc/swscale_template.c |
| diffstat | 17 files changed, 123 insertions(+), 88 deletions(-) [+] |
line wrap: on
line diff
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/asmalign.h Sat Apr 15 20:46:54 2006 +0000 @@ -0,0 +1,7 @@ +#ifdef SYS_DARWIN +#define ASMALIGN8 ".align 3\n\t" +#define ASMALIGN16 ".align 4\n\t" +#else +#define ASMALIGN8 ".balign 8\n\t" +#define ASMALIGN16 ".balign 16\n\t" +#endif
--- a/configure Sat Apr 15 16:47:25 2006 +0000 +++ b/configure Sat Apr 15 20:46:54 2006 +0000 @@ -6972,7 +6972,10 @@ echores "$_crash_debug" if darwin ; then - CFLAGS="$CFLAGS -mdynamic-no-pic -falign-loops=16 -DSYS_DARWIN" + CFLAGS="$CFLAGS -mdynamic-no-pic -falign-loops=16 -DSYS_DARWIN -DCONFIG_DARWIN" + if x86 ; then + CFLAGS="$CFLAGS -arch i386 -isysroot /Developer/SDKs/MacOSX10.4u.sdk" + fi if [ "$_cc_major" = 3 ] && [ "$_cc_minor" -lt 1 ]; then CFLAGS="$CFLAGS -no-cpp-precomp" fi
--- a/liba52/downmix.c Sat Apr 15 16:47:25 2006 +0000 +++ b/liba52/downmix.c Sat Apr 15 20:46:54 2006 +0000 @@ -28,6 +28,7 @@ */ #include "config.h" +#include "asmalign.h" #include <string.h> #include <inttypes.h> @@ -691,7 +692,7 @@ "movlps %2, %%xmm7 \n\t" "shufps $0x00, %%xmm7, %%xmm7 \n\t" "mov $-1024, %%"REG_S" \n\t" - ".balign 16\n\t" + ASMALIGN16 "1: \n\t" "movaps (%0, %%"REG_S"), %%xmm0 \n\t" "movaps 16(%0, %%"REG_S"), %%xmm1\n\t" @@ -714,7 +715,7 @@ "movlps %1, %%xmm7 \n\t" "shufps $0x00, %%xmm7, %%xmm7 \n\t" "mov $-1024, %%"REG_S" \n\t" - ".balign 16\n\t" + ASMALIGN16 "1: \n\t" "movaps (%0, %%"REG_S"), %%xmm0 \n\t" "movaps 1024(%0, %%"REG_S"), %%xmm1\n\t" @@ -735,7 +736,7 @@ "movlps %1, %%xmm7 \n\t" "shufps $0x00, %%xmm7, %%xmm7 \n\t" "mov $-1024, %%"REG_S" \n\t" - ".balign 16\n\t" + ASMALIGN16 "1: \n\t" "movaps (%0, %%"REG_S"), %%xmm0 \n\t" "movaps 1024(%0, %%"REG_S"), %%xmm1\n\t" @@ -757,7 +758,7 @@ "movlps %1, %%xmm7 \n\t" "shufps $0x00, %%xmm7, %%xmm7 \n\t" "mov $-1024, %%"REG_S" \n\t" - ".balign 16\n\t" + ASMALIGN16 "1: \n\t" "movaps (%0, %%"REG_S"), %%xmm0 \n\t" "movaps 1024(%0, %%"REG_S"), %%xmm1\n\t" @@ -780,7 +781,7 @@ "movlps %1, %%xmm7 \n\t" "shufps $0x00, %%xmm7, %%xmm7 \n\t" "mov $-1024, %%"REG_S" \n\t" - ".balign 16\n\t" + ASMALIGN16 "1: \n\t" "movaps 1024(%0, %%"REG_S"), %%xmm0\n\t" "addps %%xmm7, %%xmm0 \n\t" //common @@ -803,7 +804,7 @@ "movlps %2, %%xmm7 \n\t" "shufps $0x00, %%xmm7, %%xmm7 \n\t" "mov $-1024, %%"REG_S" \n\t" - ".balign 16\n\t" + ASMALIGN16 "1: \n\t" "movaps 1024(%1, %%"REG_S"), %%xmm0\n\t" "addps %%xmm7, %%xmm0 \n\t" //common @@ -826,7 +827,7 @@ "movlps %1, %%xmm7 \n\t" "shufps $0x00, %%xmm7, %%xmm7 \n\t" "mov $-1024, %%"REG_S" \n\t" - ".balign 16\n\t" + ASMALIGN16 "1: \n\t" "movaps 2048(%0, %%"REG_S"), %%xmm0\n\t" // surround "movaps (%0, %%"REG_S"), %%xmm1 \n\t" @@ -850,7 +851,7 @@ "movlps %1, %%xmm7 \n\t" "shufps $0x00, %%xmm7, %%xmm7 \n\t" "mov $-1024, %%"REG_S" \n\t" - ".balign 16\n\t" + ASMALIGN16 "1: \n\t" "movaps 1024(%0, %%"REG_S"), %%xmm0\n\t" "addps 3072(%0, %%"REG_S"), %%xmm0\n\t" @@ -874,7 +875,7 @@ "movlps %1, %%xmm7 \n\t" "shufps $0x00, %%xmm7, %%xmm7 \n\t" "mov $-1024, %%"REG_S" \n\t" - ".balign 16\n\t" + ASMALIGN16 "1: \n\t" "movaps 1024(%0, %%"REG_S"), %%xmm0\n\t" "movaps 3072(%0, %%"REG_S"), %%xmm3\n\t" // surround @@ -900,7 +901,7 @@ "movlps %1, %%xmm7 \n\t" "shufps $0x00, %%xmm7, %%xmm7 \n\t" "mov $-1024, %%"REG_S" \n\t" - ".balign 16\n\t" + ASMALIGN16 "1: \n\t" "movaps 2048(%0, %%"REG_S"), %%xmm0\n\t" "addps 3072(%0, %%"REG_S"), %%xmm0\n\t" // surround @@ -925,7 +926,7 @@ "movlps %1, %%xmm7 \n\t" "shufps $0x00, %%xmm7, %%xmm7 \n\t" "mov $-1024, %%"REG_S" \n\t" - ".balign 16\n\t" + ASMALIGN16 "1: \n\t" "movaps 1024(%0, %%"REG_S"), %%xmm0\n\t" "addps %%xmm7, %%xmm0 \n\t" // common @@ -949,7 +950,7 @@ "movlps %1, %%xmm7 \n\t" "shufps $0x00, %%xmm7, %%xmm7 \n\t" "mov $-1024, %%"REG_S" \n\t" - ".balign 16\n\t" + ASMALIGN16 "1: \n\t" "movaps 1024(%0, %%"REG_S"), %%xmm0\n\t" "movaps 3072(%0, %%"REG_S"), %%xmm2\n\t" @@ -976,7 +977,7 @@ "movlps %2, %%xmm7 \n\t" "shufps $0x00, %%xmm7, %%xmm7 \n\t" "mov $-1024, %%"REG_S" \n\t" - ".balign 16\n\t" + ASMALIGN16 "1: \n\t" "movaps (%0, %%"REG_S"), %%xmm0 \n\t" "movaps 16(%0, %%"REG_S"), %%xmm1\n\t" @@ -998,7 +999,7 @@ asm volatile( "mov $-1024, %%"REG_S" \n\t" "pxor %%mm0, %%mm0 \n\t" - ".balign 16\n\t" + ASMALIGN16 "1: \n\t" "movq %%mm0, (%0, %%"REG_S") \n\t" "movq %%mm0, 8(%0, %%"REG_S") \n\t" @@ -1258,7 +1259,7 @@ "movd %2, %%mm7 \n\t" "punpckldq %2, %%mm7 \n\t" "mov $-1024, %%"REG_S" \n\t" - ".balign 16\n\t" + ASMALIGN16 "1: \n\t" "movq (%0, %%"REG_S"), %%mm0 \n\t" "movq 8(%0, %%"REG_S"), %%mm1 \n\t" @@ -1289,7 +1290,7 @@ "movd %1, %%mm7 \n\t" "punpckldq %1, %%mm7 \n\t" "mov $-1024, %%"REG_S" \n\t" - ".balign 16\n\t" + ASMALIGN16 "1: \n\t" "movq (%0, %%"REG_S"), %%mm0 \n\t" "movq 8(%0, %%"REG_S"), %%mm1 \n\t" @@ -1316,7 +1317,7 @@ "movd %1, %%mm7 \n\t" "punpckldq %1, %%mm7 \n\t" "mov $-1024, %%"REG_S" \n\t" - ".balign 16\n\t" + ASMALIGN16 "1: \n\t" "movq (%0, %%"REG_S"), %%mm0 \n\t" "movq 8(%0, %%"REG_S"), %%mm1 \n\t" @@ -1345,7 +1346,7 @@ "movd %1, %%mm7 \n\t" "punpckldq %1, %%mm7 \n\t" "mov $-1024, %%"REG_S" \n\t" - ".balign 16\n\t" + ASMALIGN16 "1: \n\t" "movq (%0, %%"REG_S"), %%mm0 \n\t" "movq 8(%0, %%"REG_S"), %%mm1 \n\t" @@ -1376,7 +1377,7 @@ "movd %1, %%mm7 \n\t" "punpckldq %1, %%mm7 \n\t" "mov $-1024, %%"REG_S" \n\t" - ".balign 16\n\t" + ASMALIGN16 "1: \n\t" "movq 1024(%0, %%"REG_S"), %%mm0\n\t" "movq 1032(%0, %%"REG_S"), %%mm1\n\t" @@ -1407,7 +1408,7 @@ "movd %2, %%mm7 \n\t" "punpckldq %2, %%mm7 \n\t" "mov $-1024, %%"REG_S" \n\t" - ".balign 16\n\t" + ASMALIGN16 "1: \n\t" "movq 1024(%1, %%"REG_S"), %%mm0\n\t" "movq 1032(%1, %%"REG_S"), %%mm1\n\t" @@ -1438,7 +1439,7 @@ "movd %1, %%mm7 \n\t" "punpckldq %1, %%mm7 \n\t" "mov $-1024, %%"REG_S" \n\t" - ".balign 16\n\t" + ASMALIGN16 "1: \n\t" "movq 2048(%0, %%"REG_S"), %%mm0\n\t" // surround "movq 2056(%0, %%"REG_S"), %%mm1\n\t" // surround @@ -1471,7 +1472,7 @@ "movd %1, %%mm7 \n\t" "punpckldq %1, %%mm7 \n\t" "mov $-1024, %%"REG_S" \n\t" - ".balign 16\n\t" + ASMALIGN16 "1: \n\t" "movq 1024(%0, %%"REG_S"), %%mm0\n\t" "movq 1032(%0, %%"REG_S"), %%mm1\n\t" @@ -1504,7 +1505,7 @@ "movd %1, %%mm7 \n\t" "punpckldq %1, %%mm7 \n\t" "mov $-1024, %%"REG_S" \n\t" - ".balign 16\n\t" + ASMALIGN16 "1: \n\t" "movq 1024(%0, %%"REG_S"), %%mm0\n\t" "movq 1032(%0, %%"REG_S"), %%mm1\n\t" @@ -1541,7 +1542,7 @@ "movd %1, %%mm7 \n\t" "punpckldq %1, %%mm7 \n\t" "mov $-1024, %%"REG_S" \n\t" - ".balign 16\n\t" + ASMALIGN16 "1: \n\t" "movq 2048(%0, %%"REG_S"), %%mm0\n\t" "movq 2056(%0, %%"REG_S"), %%mm1\n\t" @@ -1576,7 +1577,7 @@ "movd %1, %%mm7 \n\t" "punpckldq %1, %%mm7 \n\t" "mov $-1024, %%"REG_S" \n\t" - ".balign 16\n\t" + ASMALIGN16 "1: \n\t" "movq 1024(%0, %%"REG_S"), %%mm0\n\t" "movq 1032(%0, %%"REG_S"), %%mm1\n\t" @@ -1608,7 +1609,7 @@ { asm volatile( "mov $-1024, %%"REG_S" \n\t" - ".balign 16\n\t" + ASMALIGN16 "1: \n\t" "movd %1, %%mm7 \n\t" "punpckldq %1, %%mm7 \n\t" @@ -1649,7 +1650,7 @@ "movd %2, %%mm7 \n\t" "punpckldq %2, %%mm7 \n\t" "mov $-1024, %%"REG_S" \n\t" - ".balign 16\n\t" + ASMALIGN16 "1: \n\t" "movq (%0, %%"REG_S"), %%mm0 \n\t" "movq 8(%0, %%"REG_S"), %%mm1 \n\t"
--- a/liba52/imdct.c Sat Apr 15 16:47:25 2006 +0000 +++ b/liba52/imdct.c Sat Apr 15 20:46:54 2006 +0000 @@ -31,6 +31,7 @@ */ #include "config.h" +#include "asmalign.h" #include <math.h> #include <stdio.h> @@ -792,7 +793,7 @@ "lea "MANGLE(bit_reverse_512)", %%"REG_a"\n\t" "mov $1008, %%"REG_D" \n\t" "push %%"REG_BP" \n\t" //use ebp without telling gcc - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" "movlps (%0, %%"REG_S"), %%xmm0 \n\t" // XXXI "movhps 8(%0, %%"REG_D"), %%xmm0 \n\t" // RXXI @@ -851,7 +852,7 @@ "xorps %%xmm1, %%xmm1 \n\t" "xorps %%xmm2, %%xmm2 \n\t" "mov %0, %%"REG_S" \n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" "movlps (%%"REG_S"), %%xmm0\n\t" //buf[p] "movlps 8(%%"REG_S"), %%xmm1\n\t" //buf[q] @@ -872,7 +873,7 @@ asm volatile( "movaps "MANGLE(ps111_1)", %%xmm7\n\t" // 1,1,1,-1 "mov %0, %%"REG_S" \n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" "movaps 16(%%"REG_S"), %%xmm2 \n\t" //r2,i2,r3,i3 "shufps $0xB4, %%xmm2, %%xmm2 \n\t" //r2,i2,i3,r3 @@ -903,7 +904,7 @@ "xorps %%xmm5, %%xmm5 \n\t" "xorps %%xmm2, %%xmm2 \n\t" "mov %0, %%"REG_S" \n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" "movaps 32(%%"REG_S"), %%xmm2 \n\t" //r4,i4,r5,i5 "movaps 48(%%"REG_S"), %%xmm3 \n\t" //r6,i6,r7,i7 @@ -944,7 +945,7 @@ buf_offset = buf+128; asm volatile( "mov %0, %%"REG_S" \n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" "xor %%"REG_D", %%"REG_D" \n\t" // k "lea (%%"REG_S", %3), %%"REG_d" \n\t" @@ -976,7 +977,7 @@ /* Post IFFT complex multiply plus IFFT complex conjugate*/ asm volatile( "mov $-1024, %%"REG_S" \n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" "movaps (%0, %%"REG_S"), %%xmm0 \n\t" "movaps (%0, %%"REG_S"), %%xmm1 \n\t" @@ -1002,7 +1003,7 @@ "xor %%"REG_S", %%"REG_S" \n\t" // 0 "movss %3, %%xmm2 \n\t" // bias "shufps $0x00, %%xmm2, %%xmm2 \n\t" // bias, bias, ... - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" "movlps (%0, %%"REG_S"), %%xmm0 \n\t" // ? ? A ? "movlps 8(%0, %%"REG_S"), %%xmm1 \n\t" // ? ? C ? @@ -1029,7 +1030,7 @@ "xor %%"REG_S", %%"REG_S" \n\t" // 0 "movss %3, %%xmm2 \n\t" // bias "shufps $0x00, %%xmm2, %%xmm2 \n\t" // bias, bias, ... - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" "movlps (%0, %%"REG_S"), %%xmm0 \n\t" // ? ? ? A "movlps 8(%0, %%"REG_S"), %%xmm1 \n\t" // ? ? ? C @@ -1056,7 +1057,7 @@ asm volatile( "xor %%"REG_D", %%"REG_D" \n\t" // 0 "xor %%"REG_S", %%"REG_S" \n\t" // 0 - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" "movlps (%0, %%"REG_S"), %%xmm0 \n\t" // ? ? ? A "movlps 8(%0, %%"REG_S"), %%xmm1 \n\t" // ? ? ? C @@ -1078,7 +1079,7 @@ asm volatile( "mov $1024, %%"REG_D" \n\t" // 1024 "xor %%"REG_S", %%"REG_S" \n\t" // 0 - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" "movlps (%0, %%"REG_S"), %%xmm0 \n\t" // ? ? A ? "movlps 8(%0, %%"REG_S"), %%xmm1 \n\t" // ? ? C ?
--- a/libmpcodecs/pullup.c Sat Apr 15 16:47:25 2006 +0000 +++ b/libmpcodecs/pullup.c Sat Apr 15 20:46:54 2006 +0000 @@ -5,6 +5,7 @@ #include <string.h> #include "pullup.h" #include "config.h" +#include "asmalign.h" @@ -18,7 +19,7 @@ "pxor %%mm4, %%mm4 \n\t" "pxor %%mm7, %%mm7 \n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" "movq (%%esi), %%mm0 \n\t" @@ -67,7 +68,7 @@ "pxor %%mm7, %%mm7 \n\t" "subl %%eax, %%edi \n\t" - ".balign 16 \n\t" + ASMALIGN16 "2: \n\t" "movq (%%esi), %%mm0 \n\t" @@ -156,7 +157,7 @@ "pxor %%mm4, %%mm4 \n\t" "pxor %%mm7, %%mm7 \n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" "movq (%%esi), %%mm0 \n\t"
--- a/libmpcodecs/vf_decimate.c Sat Apr 15 16:47:25 2006 +0000 +++ b/libmpcodecs/vf_decimate.c Sat Apr 15 20:46:54 2006 +0000 @@ -5,6 +5,7 @@ #include "config.h" #include "mp_msg.h" #include "cpudetect.h" +#include "asmalign.h" #include "img_format.h" #include "mp_image.h" @@ -28,7 +29,7 @@ "pxor %%mm4, %%mm4 \n\t" "pxor %%mm7, %%mm7 \n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" "movq (%%"REG_S"), %%mm0 \n\t"
--- a/libmpcodecs/vf_divtc.c Sat Apr 15 16:47:25 2006 +0000 +++ b/libmpcodecs/vf_divtc.c Sat Apr 15 20:46:54 2006 +0000 @@ -8,6 +8,7 @@ #include "mp_msg.h" #include "cpudetect.h" #include "bswap.h" +#include "asmalign.h" #include "img_format.h" #include "mp_image.h" @@ -41,7 +42,7 @@ "pxor %%mm4, %%mm4 \n\t" "pxor %%mm7, %%mm7 \n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" "movq (%%"REG_S"), %%mm0 \n\t"
--- a/libmpcodecs/vf_eq.c Sat Apr 15 16:47:25 2006 +0000 +++ b/libmpcodecs/vf_eq.c Sat Apr 15 20:46:54 2006 +0000 @@ -6,6 +6,7 @@ #include "config.h" #include "mp_msg.h" #include "cpudetect.h" +#include "asmalign.h" #include "img_format.h" #include "mp_image.h" @@ -51,7 +52,7 @@ "movq (%6), %%mm4 \n\t" "pxor %%mm0, %%mm0 \n\t" "movl %4, %%eax\n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" "movq (%0), %%mm1 \n\t" "movq (%0), %%mm2 \n\t"
--- a/libmpcodecs/vf_eq2.c Sat Apr 15 16:47:25 2006 +0000 +++ b/libmpcodecs/vf_eq2.c Sat Apr 15 20:46:54 2006 +0000 @@ -18,6 +18,7 @@ #include "config.h" #include "mp_msg.h" #include "cpudetect.h" +#include "asmalign.h" #include "img_format.h" #include "mp_image.h" @@ -135,7 +136,7 @@ "movq (%6), %%mm4 \n\t" "pxor %%mm0, %%mm0 \n\t" "movl %4, %%eax\n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" "movq (%0), %%mm1 \n\t" "movq (%0), %%mm2 \n\t"
--- a/libmpcodecs/vf_fspp.c Sat Apr 15 16:47:25 2006 +0000 +++ b/libmpcodecs/vf_fspp.c Sat Apr 15 20:46:54 2006 +0000 @@ -37,6 +37,7 @@ #include <math.h> #include "config.h" +#include "asmalign.h" #include "mp_msg.h" #include "cpudetect.h" @@ -888,7 +889,7 @@ static void column_fidct_mmx(int16_t* thr_adr, DCTELEM *data, DCTELEM *output, int cnt) { asm volatile( - ".align 16 \n\t" + ASMALIGN16 "1: \n\t" "movq "DCTSIZE_S"*0*2(%%"REG_S"), %%mm1 \n\t" //
--- a/libmpcodecs/vf_halfpack.c Sat Apr 15 16:47:25 2006 +0000 +++ b/libmpcodecs/vf_halfpack.c Sat Apr 15 20:46:54 2006 +0000 @@ -6,6 +6,7 @@ #include "config.h" #include "mp_msg.h" #include "cpudetect.h" +#include "asmalign.h" #include "img_format.h" #include "mp_image.h" @@ -40,7 +41,7 @@ for (h/=2; h; h--) { asm ( "pxor %%mm0, %%mm0 \n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" "movq (%0), %%mm1 \n\t" "movq (%0), %%mm2 \n\t"
--- a/libmpcodecs/vf_ilpack.c Sat Apr 15 16:47:25 2006 +0000 +++ b/libmpcodecs/vf_ilpack.c Sat Apr 15 20:46:54 2006 +0000 @@ -6,6 +6,7 @@ #include "config.h" #include "mp_msg.h" #include "cpudetect.h" +#include "asmalign.h" #include "img_format.h" #include "mp_image.h" @@ -66,7 +67,7 @@ { int j; asm volatile ("" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" "movq (%0), %%mm1 \n\t" "movq (%0), %%mm2 \n\t" @@ -105,7 +106,7 @@ #endif "pxor %%mm0, %%mm0 \n\t" - ".balign 16 \n\t" + ASMALIGN16 ".Lli0: \n\t" "movq (%%"REG_S"), %%mm1 \n\t" "movq (%%"REG_S"), %%mm2 \n\t" @@ -213,7 +214,7 @@ #endif "pxor %%mm0, %%mm0 \n\t" - ".balign 16 \n\t" + ASMALIGN16 ".Lli1: \n\t" "movq (%%"REG_S"), %%mm1 \n\t" "movq (%%"REG_S"), %%mm2 \n\t"
--- a/libmpcodecs/vf_ivtc.c Sat Apr 15 16:47:25 2006 +0000 +++ b/libmpcodecs/vf_ivtc.c Sat Apr 15 20:46:54 2006 +0000 @@ -5,6 +5,7 @@ #include "config.h" #include "mp_msg.h" #include "cpudetect.h" +#include "asmalign.h" #include "img_format.h" #include "mp_image.h" @@ -67,7 +68,7 @@ "pxor %%mm5, %%mm5 \n\t" // 4 odd difference sums "pxor %%mm7, %%mm7 \n\t" // all zeros - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" // Even difference @@ -127,7 +128,7 @@ "pxor %%mm5, %%mm5 \n\t" // Temporal noise "pxor %%mm6, %%mm6 \n\t" // Current spacial noise - ".balign 16 \n\t" + ASMALIGN16 "2: \n\t" "movq (%%"REG_S"), %%mm0 \n\t" @@ -181,7 +182,7 @@ "pxor %%mm5, %%mm5 \n\t" "pxor %%mm6, %%mm6 \n\t" - ".balign 16 \n\t" + ASMALIGN16 "3: \n\t" "movq (%%"REG_S"), %%mm0 \n\t"
--- a/libmpcodecs/vf_noise.c Sat Apr 15 16:47:25 2006 +0000 +++ b/libmpcodecs/vf_noise.c Sat Apr 15 20:46:54 2006 +0000 @@ -25,6 +25,7 @@ #include "config.h" #include "mp_msg.h" #include "cpudetect.h" +#include "asmalign.h" #ifdef HAVE_MALLOC_H #include <malloc.h> @@ -153,7 +154,7 @@ "pcmpeqb %%mm7, %%mm7 \n\t" "psllw $15, %%mm7 \n\t" "packsswb %%mm7, %%mm7 \n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" "movq (%0, %%"REG_a"), %%mm0 \n\t" "movq (%1, %%"REG_a"), %%mm1 \n\t" @@ -182,7 +183,7 @@ "pcmpeqb %%mm7, %%mm7 \n\t" "psllw $15, %%mm7 \n\t" "packsswb %%mm7, %%mm7 \n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" "movq (%0, %%"REG_a"), %%mm0 \n\t" "movq (%1, %%"REG_a"), %%mm1 \n\t" @@ -220,7 +221,7 @@ asm volatile( "mov %5, %%"REG_a" \n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" "movq (%1, %%"REG_a"), %%mm1 \n\t" "movq (%0, %%"REG_a"), %%mm0 \n\t"
--- a/mangle.h Sat Apr 15 16:47:25 2006 +0000 +++ b/mangle.h Sat Apr 15 20:46:54 2006 +0000 @@ -9,7 +9,7 @@ /* Feel free to add more to the list, eg. a.out IMO */ #if defined(__CYGWIN__) || defined(__MINGW32__) || defined(__OS2__) || \ - (defined(__OpenBSD__) && !defined(__ELF__)) + (defined(__OpenBSD__) && !defined(__ELF__)) || defined(__APPLE__) #define MANGLE(a) "_" #a #else #define MANGLE(a) #a
--- a/postproc/rgb2rgb_template.c Sat Apr 15 16:47:25 2006 +0000 +++ b/postproc/rgb2rgb_template.c Sat Apr 15 20:46:54 2006 +0000 @@ -12,6 +12,8 @@ #include <stddef.h> #include <inttypes.h> /* for __WORDSIZE */ +#include "asmalign.h" + #ifndef __WORDSIZE // #warning You have misconfigured system and probably will lose performance! #define __WORDSIZE MP_WORDSIZE @@ -40,9 +42,14 @@ #define PREFETCHW "prefetcht0" #define PAVGB "pavgb" #else +#ifdef __APPLE__ +#define PREFETCH "#" +#define PREFETCHW "#" +#elif #define PREFETCH "/nop" #define PREFETCHW "/nop" #endif +#endif #ifdef HAVE_3DNOW /* On K6 femms is faster of emms. On K7 femms is directly mapped on emms. */ @@ -56,8 +63,12 @@ #define SFENCE "sfence" #else #define MOVNTQ "movq" +#ifdef __APPLE__ +#define SFENCE "#" +#elif #define SFENCE "/nop" #endif +#endif static inline void RENAME(rgb24to32)(const uint8_t *src,uint8_t *dst,long src_size) { @@ -332,7 +343,7 @@ "movq %3, %%mm5 \n\t" "movq %4, %%mm6 \n\t" "movq %5, %%mm7 \n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" PREFETCH" 32(%1) \n\t" "movd (%1), %%mm0 \n\t" @@ -489,7 +500,7 @@ "movq %3, %%mm5 \n\t" "movq %4, %%mm6 \n\t" "movq %5, %%mm7 \n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" PREFETCH" 32(%1) \n\t" "movd (%1), %%mm0 \n\t" @@ -1344,7 +1355,7 @@ /* TODO: unroll this loop */ asm volatile ( "xor %%"REG_a", %%"REG_a" \n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" PREFETCH" 32(%0, %%"REG_a") \n\t" "movq (%0, %%"REG_a"), %%mm0 \n\t" @@ -1394,7 +1405,7 @@ "movq "MANGLE(mask24r)", %%mm5 \n\t" "movq "MANGLE(mask24g)", %%mm6 \n\t" "movq "MANGLE(mask24b)", %%mm7 \n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" PREFETCH" 32(%1, %%"REG_a") \n\t" "movq (%1, %%"REG_a"), %%mm0 \n\t" // BGR BGR BG @@ -1464,7 +1475,7 @@ //FIXME handle 2 lines a once (fewer prefetch, reuse some chrom, but very likely limited by mem anyway) asm volatile( "xor %%"REG_a", %%"REG_a" \n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" PREFETCH" 32(%1, %%"REG_a", 2) \n\t" PREFETCH" 32(%2, %%"REG_a") \n\t" @@ -1617,7 +1628,7 @@ //FIXME handle 2 lines a once (fewer prefetch, reuse some chrom, but very likely limited by mem anyway) asm volatile( "xor %%"REG_a", %%"REG_a" \n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" PREFETCH" 32(%1, %%"REG_a", 2) \n\t" PREFETCH" 32(%2, %%"REG_a") \n\t" @@ -1741,7 +1752,7 @@ "xor %%"REG_a", %%"REG_a" \n\t" "pcmpeqw %%mm7, %%mm7 \n\t" "psrlw $8, %%mm7 \n\t" // FF,00,FF,00... - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" PREFETCH" 64(%0, %%"REG_a", 4) \n\t" "movq (%0, %%"REG_a", 4), %%mm0 \n\t" // YUYV YUYV(0) @@ -1794,7 +1805,7 @@ asm volatile( "xor %%"REG_a", %%"REG_a" \n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" PREFETCH" 64(%0, %%"REG_a", 4) \n\t" "movq (%0, %%"REG_a", 4), %%mm0 \n\t" // YUYV YUYV(0) @@ -1979,7 +1990,7 @@ "xorl %%eax, %%eax \n\t" "pcmpeqw %%mm7, %%mm7 \n\t" "psrlw $8, %%mm7 \n\t" // FF,00,FF,00... - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" PREFETCH" 64(%0, %%eax, 4) \n\t" "movq (%0, %%eax, 4), %%mm0 \n\t" // UYVY UYVY(0) @@ -2032,7 +2043,7 @@ asm volatile( "xorl %%eax, %%eax \n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" PREFETCH" 64(%0, %%eax, 4) \n\t" "movq (%0, %%eax, 4), %%mm0 \n\t" // YUYV YUYV(0) @@ -2110,7 +2121,7 @@ "movq "MANGLE(w1111)", %%mm5 \n\t" "pxor %%mm7, %%mm7 \n\t" "lea (%%"REG_a", %%"REG_a", 2), %%"REG_b"\n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" PREFETCH" 64(%0, %%"REG_b") \n\t" "movd (%0, %%"REG_b"), %%mm0 \n\t" @@ -2184,7 +2195,7 @@ "pxor %%mm7, %%mm7 \n\t" "lea (%%"REG_a", %%"REG_a", 2), %%"REG_b"\n\t" "add %%"REG_b", %%"REG_b" \n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" PREFETCH" 64(%0, %%"REG_b") \n\t" PREFETCH" 64(%1, %%"REG_b") \n\t"
--- a/postproc/swscale_template.c Sat Apr 15 16:47:25 2006 +0000 +++ b/postproc/swscale_template.c Sat Apr 15 20:46:54 2006 +0000 @@ -16,6 +16,8 @@ Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA */ +#include "asmalign.h" + #undef REAL_MOVNTQ #undef MOVNTQ #undef PAVGB @@ -71,7 +73,7 @@ "movq %%mm3, %%mm4 \n\t"\ "lea " offset "(%0), %%"REG_d" \n\t"\ "mov (%%"REG_d"), %%"REG_S" \n\t"\ - ".balign 16 \n\t" /* FIXME Unroll? */\ + ASMALIGN16 /* FIXME Unroll? */\ "1: \n\t"\ "movq 8(%%"REG_d"), %%mm0 \n\t" /* filterCoeff */\ "movq " #x "(%%"REG_S", %%"REG_a", 2), %%mm2\n\t" /* srcData */\ @@ -98,7 +100,7 @@ #define YSCALEYUV2YV121 \ "mov %2, %%"REG_a" \n\t"\ - ".balign 16 \n\t" /* FIXME Unroll? */\ + ASMALIGN16 /* FIXME Unroll? */\ "1: \n\t"\ "movq (%0, %%"REG_a", 2), %%mm0 \n\t"\ "movq 8(%0, %%"REG_a", 2), %%mm1\n\t"\ @@ -118,14 +120,14 @@ */ #define YSCALEYUV2PACKEDX \ "xor %%"REG_a", %%"REG_a" \n\t"\ - ".balign 16 \n\t"\ + ASMALIGN16\ "nop \n\t"\ "1: \n\t"\ "lea "CHR_MMX_FILTER_OFFSET"(%0), %%"REG_d"\n\t"\ "mov (%%"REG_d"), %%"REG_S" \n\t"\ "movq "VROUNDER_OFFSET"(%0), %%mm3\n\t"\ "movq %%mm3, %%mm4 \n\t"\ - ".balign 16 \n\t"\ + ASMALIGN16\ "2: \n\t"\ "movq 8(%%"REG_d"), %%mm0 \n\t" /* filterCoeff */\ "movq (%%"REG_S", %%"REG_a"), %%mm2 \n\t" /* UsrcData */\ @@ -143,7 +145,7 @@ "mov (%%"REG_d"), %%"REG_S" \n\t"\ "movq "VROUNDER_OFFSET"(%0), %%mm1\n\t"\ "movq %%mm1, %%mm7 \n\t"\ - ".balign 16 \n\t"\ + ASMALIGN16\ "2: \n\t"\ "movq 8(%%"REG_d"), %%mm0 \n\t" /* filterCoeff */\ "movq (%%"REG_S", %%"REG_a", 2), %%mm2 \n\t" /* Y1srcData */\ @@ -205,7 +207,7 @@ "punpcklwd %%mm5, %%mm5 \n\t"\ "punpcklwd %%mm5, %%mm5 \n\t"\ "xor %%"REG_a", %%"REG_a" \n\t"\ - ".balign 16 \n\t"\ + ASMALIGN16\ "1: \n\t"\ "movq (%0, %%"REG_a", 2), %%mm0 \n\t" /*buf0[eax]*/\ "movq (%1, %%"REG_a", 2), %%mm1 \n\t" /*buf1[eax]*/\ @@ -258,7 +260,7 @@ "movq %%mm0, "CHR_MMX_FILTER_OFFSET"+8("#c")\n\t"\ "movq %%mm1, "LUM_MMX_FILTER_OFFSET"+8("#c")\n\t"\ "xor "#index", "#index" \n\t"\ - ".balign 16 \n\t"\ + ASMALIGN16\ "1: \n\t"\ "movq (%2, "#index"), %%mm2 \n\t" /* uvbuf0[eax]*/\ "movq (%3, "#index"), %%mm3 \n\t" /* uvbuf1[eax]*/\ @@ -290,7 +292,7 @@ #define REAL_YSCALEYUV2RGB(index, c) \ "xor "#index", "#index" \n\t"\ - ".balign 16 \n\t"\ + ASMALIGN16\ "1: \n\t"\ "movq (%2, "#index"), %%mm2 \n\t" /* uvbuf0[eax]*/\ "movq (%3, "#index"), %%mm3 \n\t" /* uvbuf1[eax]*/\ @@ -356,7 +358,7 @@ #define REAL_YSCALEYUV2PACKED1(index, c) \ "xor "#index", "#index" \n\t"\ - ".balign 16 \n\t"\ + ASMALIGN16\ "1: \n\t"\ "movq (%2, "#index"), %%mm3 \n\t" /* uvbuf0[eax]*/\ "movq 4096(%2, "#index"), %%mm4 \n\t" /* uvbuf0[eax+2048]*/\ @@ -371,7 +373,7 @@ #define REAL_YSCALEYUV2RGB1(index, c) \ "xor "#index", "#index" \n\t"\ - ".balign 16 \n\t"\ + ASMALIGN16\ "1: \n\t"\ "movq (%2, "#index"), %%mm3 \n\t" /* uvbuf0[eax]*/\ "movq 4096(%2, "#index"), %%mm4 \n\t" /* uvbuf0[eax+2048]*/\ @@ -420,7 +422,7 @@ #define REAL_YSCALEYUV2PACKED1b(index, c) \ "xor "#index", "#index" \n\t"\ - ".balign 16 \n\t"\ + ASMALIGN16\ "1: \n\t"\ "movq (%2, "#index"), %%mm2 \n\t" /* uvbuf0[eax]*/\ "movq (%3, "#index"), %%mm3 \n\t" /* uvbuf1[eax]*/\ @@ -439,7 +441,7 @@ // do vertical chrominance interpolation #define REAL_YSCALEYUV2RGB1b(index, c) \ "xor "#index", "#index" \n\t"\ - ".balign 16 \n\t"\ + ASMALIGN16\ "1: \n\t"\ "movq (%2, "#index"), %%mm2 \n\t" /* uvbuf0[eax]*/\ "movq (%3, "#index"), %%mm3 \n\t" /* uvbuf1[eax]*/\ @@ -1662,7 +1664,7 @@ "movq "MANGLE(w1111)", %%mm5 \n\t" "pxor %%mm7, %%mm7 \n\t" "lea (%%"REG_a", %%"REG_a", 2), %%"REG_b"\n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" PREFETCH" 64(%0, %%"REG_b") \n\t" "movd (%0, %%"REG_b"), %%mm0 \n\t" @@ -1748,7 +1750,7 @@ "pxor %%mm7, %%mm7 \n\t" "lea (%%"REG_a", %%"REG_a", 2), %%"REG_b" \n\t" "add %%"REG_b", %%"REG_b" \n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" PREFETCH" 64(%0, %%"REG_b") \n\t" PREFETCH" 64(%1, %%"REG_b") \n\t" @@ -2055,7 +2057,7 @@ "movq "MANGLE(w02)", %%mm6 \n\t" "push %%"REG_BP" \n\t" // we use 7 regs here ... "mov %%"REG_a", %%"REG_BP" \n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" "movzwl (%2, %%"REG_BP"), %%eax \n\t" "movzwl 2(%2, %%"REG_BP"), %%ebx\n\t" @@ -2093,7 +2095,7 @@ "movq "MANGLE(w02)", %%mm6 \n\t" "push %%"REG_BP" \n\t" // we use 7 regs here ... "mov %%"REG_a", %%"REG_BP" \n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" "movzwl (%2, %%"REG_BP"), %%eax \n\t" "movzwl 2(%2, %%"REG_BP"), %%ebx\n\t" @@ -2142,7 +2144,7 @@ asm volatile( "pxor %%mm7, %%mm7 \n\t" "movq "MANGLE(w02)", %%mm6 \n\t" - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" "mov %2, %%"REG_c" \n\t" "movzwl (%%"REG_c", %0), %%eax \n\t" @@ -2326,7 +2328,7 @@ "xor %%"REG_a", %%"REG_a" \n\t" // i "xor %%"REG_b", %%"REG_b" \n\t" // xx "xorl %%ecx, %%ecx \n\t" // 2*xalpha - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" "movzbl (%0, %%"REG_b"), %%edi \n\t" //src[xx] "movzbl 1(%0, %%"REG_b"), %%esi \n\t" //src[xx+1] @@ -2523,7 +2525,7 @@ "xor %%"REG_a", %%"REG_a" \n\t" // i "xor %%"REG_b", %%"REG_b" \n\t" // xx "xorl %%ecx, %%ecx \n\t" // 2*xalpha - ".balign 16 \n\t" + ASMALIGN16 "1: \n\t" "mov %0, %%"REG_S" \n\t" "movzbl (%%"REG_S", %%"REG_b"), %%edi \n\t" //src[xx]