mirror of https://github.com/mpv-player/mpv
mangle for win32 in liba52 (includes dummy mangle.h pointing to the one in main)
git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@4248 b3059339-0415-0410-9bf9-f77b7e298cf2
This commit is contained in:
parent
e6f0381dc2
commit
4e3181a168
|
@ -37,6 +37,7 @@
|
|||
#include "a52.h"
|
||||
#include "a52_internal.h"
|
||||
#include "mm_accel.h"
|
||||
#include "mangle.h"
|
||||
|
||||
#ifdef RUNTIME_CPUDETECT
|
||||
#undef HAVE_3DNOWEX
|
||||
|
@ -660,7 +661,7 @@ imdct_do_512_sse(sample_t data[],sample_t delay[], sample_t bias)
|
|||
/* Bit reversed shuffling */
|
||||
asm volatile(
|
||||
"xorl %%esi, %%esi \n\t"
|
||||
"leal bit_reverse_512, %%eax \n\t"
|
||||
"leal "MANGLE(bit_reverse_512)", %%eax \n\t"
|
||||
"movl $1008, %%edi \n\t"
|
||||
"pushl %%ebp \n\t" //use ebp without telling gcc
|
||||
".balign 16 \n\t"
|
||||
|
@ -670,10 +671,10 @@ imdct_do_512_sse(sample_t data[],sample_t delay[], sample_t bias)
|
|||
"movlps 8(%0, %%esi), %%xmm1 \n\t" // XXXi
|
||||
"movhps (%0, %%edi), %%xmm1 \n\t" // rXXi
|
||||
"shufps $0x33, %%xmm1, %%xmm0 \n\t" // irIR
|
||||
"movaps sseSinCos1c(%%esi), %%xmm2 \n\t"
|
||||
"movaps "MANGLE(sseSinCos1c)"(%%esi), %%xmm2\n\t"
|
||||
"mulps %%xmm0, %%xmm2 \n\t"
|
||||
"shufps $0xB1, %%xmm0, %%xmm0 \n\t" // riRI
|
||||
"mulps sseSinCos1d(%%esi), %%xmm0 \n\t"
|
||||
"mulps "MANGLE(sseSinCos1d)"(%%esi), %%xmm0\n\t"
|
||||
"subps %%xmm0, %%xmm2 \n\t"
|
||||
"movzbl (%%eax), %%edx \n\t"
|
||||
"movzbl 1(%%eax), %%ebp \n\t"
|
||||
|
@ -741,7 +742,7 @@ imdct_do_512_sse(sample_t data[],sample_t delay[], sample_t bias)
|
|||
/* 2. iteration */
|
||||
// Note w[1]={{1,0}, {0,-1}}
|
||||
asm volatile(
|
||||
"movaps ps111_1, %%xmm7 \n\t" // 1,1,1,-1
|
||||
"movaps "MANGLE(ps111_1)", %%xmm7\n\t" // 1,1,1,-1
|
||||
"movl %0, %%esi \n\t"
|
||||
".balign 16 \n\t"
|
||||
"1: \n\t"
|
||||
|
@ -769,8 +770,8 @@ imdct_do_512_sse(sample_t data[],sample_t delay[], sample_t bias)
|
|||
Note sseW2+48={1,-1,sqrt(2),-sqrt(2))
|
||||
*/
|
||||
asm volatile(
|
||||
"movaps 48+sseW2, %%xmm6 \n\t"
|
||||
"movaps 16+sseW2, %%xmm7 \n\t"
|
||||
"movaps 48+"MANGLE(sseW2)", %%xmm6\n\t"
|
||||
"movaps 16+"MANGLE(sseW2)", %%xmm7\n\t"
|
||||
"xorps %%xmm5, %%xmm5 \n\t"
|
||||
"xorps %%xmm2, %%xmm2 \n\t"
|
||||
"movl %0, %%esi \n\t"
|
||||
|
@ -778,8 +779,8 @@ imdct_do_512_sse(sample_t data[],sample_t delay[], sample_t bias)
|
|||
"1: \n\t"
|
||||
"movaps 32(%%esi), %%xmm2 \n\t" //r4,i4,r5,i5
|
||||
"movaps 48(%%esi), %%xmm3 \n\t" //r6,i6,r7,i7
|
||||
"movaps sseW2, %%xmm4 \n\t" //r4,i4,r5,i5
|
||||
"movaps 32+sseW2, %%xmm5 \n\t" //r6,i6,r7,i7
|
||||
"movaps "MANGLE(sseW2)", %%xmm4 \n\t" //r4,i4,r5,i5
|
||||
"movaps 32+"MANGLE(sseW2)", %%xmm5\n\t" //r6,i6,r7,i7
|
||||
"mulps %%xmm2, %%xmm4 \n\t"
|
||||
"mulps %%xmm3, %%xmm5 \n\t"
|
||||
"shufps $0xB1, %%xmm2, %%xmm2 \n\t" //i4,r4,i5,r5
|
||||
|
@ -844,14 +845,14 @@ imdct_do_512_sse(sample_t data[],sample_t delay[], sample_t bias)
|
|||
|
||||
/* Post IFFT complex multiply plus IFFT complex conjugate*/
|
||||
asm volatile(
|
||||
"movl $-1024, %%esi \n\t"
|
||||
"movl $-1024, %%esi \n\t"
|
||||
".balign 16 \n\t"
|
||||
"1: \n\t"
|
||||
"movaps (%0, %%esi), %%xmm0 \n\t"
|
||||
"movaps (%0, %%esi), %%xmm1 \n\t"
|
||||
"shufps $0xB1, %%xmm0, %%xmm0 \n\t"
|
||||
"mulps 1024+sseSinCos1c(%%esi), %%xmm1 \n\t"
|
||||
"mulps 1024+sseSinCos1d(%%esi), %%xmm0 \n\t"
|
||||
"mulps 1024+"MANGLE(sseSinCos1c)"(%%esi), %%xmm1\n\t"
|
||||
"mulps 1024+"MANGLE(sseSinCos1d)"(%%esi), %%xmm0\n\t"
|
||||
"addps %%xmm1, %%xmm0 \n\t"
|
||||
"movaps %%xmm0, (%0, %%esi) \n\t"
|
||||
"addl $16, %%esi \n\t"
|
||||
|
@ -878,7 +879,7 @@ imdct_do_512_sse(sample_t data[],sample_t delay[], sample_t bias)
|
|||
"movhps -16(%0, %%edi), %%xmm1 \n\t" // ? D C ?
|
||||
"movhps -8(%0, %%edi), %%xmm0 \n\t" // ? B A ?
|
||||
"shufps $0x99, %%xmm1, %%xmm0 \n\t" // D C B A
|
||||
"mulps sseWindow(%%esi), %%xmm0 \n\t"
|
||||
"mulps "MANGLE(sseWindow)"(%%esi), %%xmm0\n\t"
|
||||
"addps (%2, %%esi), %%xmm0 \n\t"
|
||||
"addps %%xmm2, %%xmm0 \n\t"
|
||||
"movaps %%xmm0, (%1, %%esi) \n\t"
|
||||
|
@ -905,7 +906,7 @@ imdct_do_512_sse(sample_t data[],sample_t delay[], sample_t bias)
|
|||
"movhps -16(%0, %%edi), %%xmm1 \n\t" // D ? ? C
|
||||
"movhps -8(%0, %%edi), %%xmm0 \n\t" // B ? ? A
|
||||
"shufps $0xCC, %%xmm1, %%xmm0 \n\t" // D C B A
|
||||
"mulps 512+sseWindow(%%esi), %%xmm0 \n\t"
|
||||
"mulps 512+"MANGLE(sseWindow)"(%%esi), %%xmm0\n\t"
|
||||
"addps (%2, %%esi), %%xmm0 \n\t"
|
||||
"addps %%xmm2, %%xmm0 \n\t"
|
||||
"movaps %%xmm0, (%1, %%esi) \n\t"
|
||||
|
@ -932,7 +933,7 @@ imdct_do_512_sse(sample_t data[],sample_t delay[], sample_t bias)
|
|||
"movhps -16(%0, %%edi), %%xmm1 \n\t" // D ? ? C
|
||||
"movhps -8(%0, %%edi), %%xmm0 \n\t" // B ? ? A
|
||||
"shufps $0xCC, %%xmm1, %%xmm0 \n\t" // D C B A
|
||||
"mulps 1024+sseWindow(%%esi), %%xmm0 \n\t"
|
||||
"mulps 1024+"MANGLE(sseWindow)"(%%esi), %%xmm0\n\t"
|
||||
"movaps %%xmm0, (%1, %%esi) \n\t"
|
||||
"addl $16, %%esi \n\t"
|
||||
"subl $16, %%edi \n\t"
|
||||
|
@ -954,7 +955,7 @@ imdct_do_512_sse(sample_t data[],sample_t delay[], sample_t bias)
|
|||
"movhps -16(%0, %%edi), %%xmm1 \n\t" // ? D C ?
|
||||
"movhps -8(%0, %%edi), %%xmm0 \n\t" // ? B A ?
|
||||
"shufps $0x99, %%xmm1, %%xmm0 \n\t" // D C B A
|
||||
"mulps 1536+sseWindow(%%esi), %%xmm0 \n\t"
|
||||
"mulps 1536+"MANGLE(sseWindow)"(%%esi), %%xmm0\n\t"
|
||||
"movaps %%xmm0, (%1, %%esi) \n\t"
|
||||
"addl $16, %%esi \n\t"
|
||||
"subl $16, %%edi \n\t"
|
||||
|
|
|
@ -0,0 +1,2 @@
|
|||
#include "../mangle.h"
|
||||
|
|
@ -9,6 +9,7 @@
|
|||
#include "a52.h"
|
||||
#include "mm_accel.h"
|
||||
#include "../config.h"
|
||||
#include "mangle.h"
|
||||
|
||||
int (* a52_resample) (float * _f, int16_t * s16)=NULL;
|
||||
|
||||
|
|
|
@ -17,10 +17,10 @@ static int a52_resample_MONO_to_5_MMX(float * _f, int16_t * s16){
|
|||
int32_t * f = (int32_t *) _f;
|
||||
asm volatile(
|
||||
"movl $-512, %%esi \n\t"
|
||||
"movq magicF2W, %%mm7 \n\t"
|
||||
"movq wm1100, %%mm3 \n\t"
|
||||
"movq wm0101, %%mm4 \n\t"
|
||||
"movq wm1010, %%mm5 \n\t"
|
||||
"movq "MANGLE(magicF2W)", %%mm7 \n\t"
|
||||
"movq "MANGLE(wm1100)", %%mm3 \n\t"
|
||||
"movq "MANGLE(wm0101)", %%mm4 \n\t"
|
||||
"movq "MANGLE(wm1010)", %%mm5 \n\t"
|
||||
"pxor %%mm6, %%mm6 \n\t"
|
||||
"1: \n\t"
|
||||
"movq (%1, %%esi, 2), %%mm0 \n\t"
|
||||
|
@ -73,7 +73,7 @@ static int a52_resample_STEREO_to_2_MMX(float * _f, int16_t * s16){
|
|||
);*/
|
||||
asm volatile(
|
||||
"movl $-1024, %%esi \n\t"
|
||||
"movq magicF2W, %%mm7 \n\t"
|
||||
"movq "MANGLE(magicF2W)", %%mm7 \n\t"
|
||||
"1: \n\t"
|
||||
"movq (%1, %%esi), %%mm0 \n\t"
|
||||
"movq 8(%1, %%esi), %%mm1 \n\t"
|
||||
|
@ -104,7 +104,7 @@ static int a52_resample_3F_to_5_MMX(float * _f, int16_t * s16){
|
|||
int32_t * f = (int32_t *) _f;
|
||||
asm volatile(
|
||||
"movl $-1024, %%esi \n\t"
|
||||
"movq magicF2W, %%mm7 \n\t"
|
||||
"movq "MANGLE(magicF2W)", %%mm7 \n\t"
|
||||
"pxor %%mm6, %%mm6 \n\t"
|
||||
"movq %%mm7, %%mm5 \n\t"
|
||||
"punpckldq %%mm6, %%mm5 \n\t"
|
||||
|
@ -160,7 +160,7 @@ static int a52_resample_2F_2R_to_4_MMX(float * _f, int16_t * s16){
|
|||
int32_t * f = (int32_t *) _f;
|
||||
asm volatile(
|
||||
"movl $-1024, %%esi \n\t"
|
||||
"movq magicF2W, %%mm7 \n\t"
|
||||
"movq "MANGLE(magicF2W)", %%mm7 \n\t"
|
||||
"1: \n\t"
|
||||
"movq (%1, %%esi), %%mm0 \n\t"
|
||||
"movq 8(%1, %%esi), %%mm1 \n\t"
|
||||
|
@ -212,7 +212,7 @@ static int a52_resample_3F_2R_to_5_MMX(float * _f, int16_t * s16){
|
|||
int32_t * f = (int32_t *) _f;
|
||||
asm volatile(
|
||||
"movl $-1024, %%esi \n\t"
|
||||
"movq magicF2W, %%mm7 \n\t"
|
||||
"movq "MANGLE(magicF2W)", %%mm7 \n\t"
|
||||
"1: \n\t"
|
||||
"movd (%1, %%esi), %%mm0 \n\t"
|
||||
"punpckldq 2048(%1, %%esi), %%mm0\n\t"
|
||||
|
@ -272,7 +272,7 @@ static int a52_resample_MONO_LFE_to_6_MMX(float * _f, int16_t * s16){
|
|||
int32_t * f = (int32_t *) _f;
|
||||
asm volatile(
|
||||
"movl $-1024, %%esi \n\t"
|
||||
"movq magicF2W, %%mm7 \n\t"
|
||||
"movq "MANGLE(magicF2W)", %%mm7 \n\t"
|
||||
"pxor %%mm6, %%mm6 \n\t"
|
||||
"1: \n\t"
|
||||
"movq 1024(%1, %%esi), %%mm0 \n\t"
|
||||
|
@ -313,7 +313,7 @@ static int a52_resample_STEREO_LFE_to_6_MMX(float * _f, int16_t * s16){
|
|||
int32_t * f = (int32_t *) _f;
|
||||
asm volatile(
|
||||
"movl $-1024, %%esi \n\t"
|
||||
"movq magicF2W, %%mm7 \n\t"
|
||||
"movq "MANGLE(magicF2W)", %%mm7 \n\t"
|
||||
"pxor %%mm6, %%mm6 \n\t"
|
||||
"1: \n\t"
|
||||
"movq 1024(%1, %%esi), %%mm0 \n\t"
|
||||
|
@ -352,7 +352,7 @@ static int a52_resample_3F_LFE_to_6_MMX(float * _f, int16_t * s16){
|
|||
int32_t * f = (int32_t *) _f;
|
||||
asm volatile(
|
||||
"movl $-1024, %%esi \n\t"
|
||||
"movq magicF2W, %%mm7 \n\t"
|
||||
"movq "MANGLE(magicF2W)", %%mm7 \n\t"
|
||||
"pxor %%mm6, %%mm6 \n\t"
|
||||
"1: \n\t"
|
||||
"movq 1024(%1, %%esi), %%mm0 \n\t"
|
||||
|
@ -393,7 +393,7 @@ static int a52_resample_2F_2R_LFE_to_6_MMX(float * _f, int16_t * s16){
|
|||
int32_t * f = (int32_t *) _f;
|
||||
asm volatile(
|
||||
"movl $-1024, %%esi \n\t"
|
||||
"movq magicF2W, %%mm7 \n\t"
|
||||
"movq "MANGLE(magicF2W)", %%mm7 \n\t"
|
||||
// "pxor %%mm6, %%mm6 \n\t"
|
||||
"1: \n\t"
|
||||
"movq 1024(%1, %%esi), %%mm0 \n\t"
|
||||
|
@ -440,7 +440,7 @@ static int a52_resample_3F_2R_LFE_to_6_MMX(float * _f, int16_t * s16){
|
|||
int32_t * f = (int32_t *) _f;
|
||||
asm volatile(
|
||||
"movl $-1024, %%esi \n\t"
|
||||
"movq magicF2W, %%mm7 \n\t"
|
||||
"movq "MANGLE(magicF2W)", %%mm7 \n\t"
|
||||
// "pxor %%mm6, %%mm6 \n\t"
|
||||
"1: \n\t"
|
||||
"movq 1024(%1, %%esi), %%mm0 \n\t"
|
||||
|
|
Loading…
Reference in New Issue