Commit 97217256 authored by Guillaume Poirier's avatar Guillaume Poirier

Fix build issues on Linux. Only gcc-4.x is supported, as on OSX.

Cleans up a few inconsistencies in the code too.


git-svn-id: svn://svn.videolan.org/x264/trunk@621 df754926-b1dd-0310-bc7b-ec298dee348c
parent c9cd0fce
......@@ -60,8 +60,8 @@ void x264_sub4x4_dct_altivec( int16_t dct[4][4],
permHighv = (vec_u8_t) CV(0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17);
VEC_DCT( dct0v, dct1v, dct2v, dct3v, tmp0v, tmp1v, tmp2v, tmp3v );
vec_st(vec_perm(tmp0v, tmp1v, permHighv), 0, dct);
vec_st(vec_perm(tmp2v, tmp3v, permHighv), 16, dct);
vec_st(vec_perm(tmp0v, tmp1v, permHighv), 0, (int16_t*)dct);
vec_st(vec_perm(tmp2v, tmp3v, permHighv), 16, (int16_t*)dct);
}
void x264_sub8x8_dct_altivec( int16_t dct[4][4][4],
......@@ -94,14 +94,14 @@ void x264_sub8x8_dct_altivec( int16_t dct[4][4][4],
VEC_DCT( dct0v, dct1v, dct2v, dct3v, tmp0v, tmp1v, tmp2v, tmp3v );
VEC_DCT( dct4v, dct5v, dct6v, dct7v, tmp4v, tmp5v, tmp6v, tmp7v );
vec_st(vec_perm(tmp0v, tmp1v, permHighv), 0, dct);
vec_st(vec_perm(tmp2v, tmp3v, permHighv), 16, dct);
vec_st(vec_perm(tmp4v, tmp5v, permHighv), 32, dct);
vec_st(vec_perm(tmp6v, tmp7v, permHighv), 48, dct);
vec_st(vec_perm(tmp0v, tmp1v, permLowv), 64, dct);
vec_st(vec_perm(tmp2v, tmp3v, permLowv), 80, dct);
vec_st(vec_perm(tmp4v, tmp5v, permLowv), 96, dct);
vec_st(vec_perm(tmp6v, tmp7v, permLowv), 112, dct);
vec_st(vec_perm(tmp0v, tmp1v, permHighv), 0, (int16_t*)dct);
vec_st(vec_perm(tmp2v, tmp3v, permHighv), 16, (int16_t*)dct);
vec_st(vec_perm(tmp4v, tmp5v, permHighv), 32, (int16_t*)dct);
vec_st(vec_perm(tmp6v, tmp7v, permHighv), 48, (int16_t*)dct);
vec_st(vec_perm(tmp0v, tmp1v, permLowv), 64, (int16_t*)dct);
vec_st(vec_perm(tmp2v, tmp3v, permLowv), 80, (int16_t*)dct);
vec_st(vec_perm(tmp4v, tmp5v, permLowv), 96, (int16_t*)dct);
vec_st(vec_perm(tmp6v, tmp7v, permLowv), 112, (int16_t*)dct);
}
void x264_sub16x16_dct_altivec( int16_t dct[16][4][4],
......@@ -311,8 +311,8 @@ void x264_sub16x16_dct8_altivec( int16_t dct[4][8][8], uint8_t *pix1, uint8_t *p
void x264_add8x8_idct8_altivec( uint8_t *dst, int16_t dct[8][8] )
{
vec_s16_t onev = vec_splat_s16(1);
vec_s16_t twov = vec_splat_s16(2);
vec_u16_t onev = vec_splat_s16(1);
vec_u16_t twov = vec_splat_s16(2);
dct[0][0] += 32; // rounding for the >>6 at the end
......@@ -341,7 +341,7 @@ void x264_add8x8_idct8_altivec( uint8_t *dst, int16_t dct[8][8] )
vec_u8_t perm_ldv = vec_lvsl(0, dst);
vec_u8_t perm_stv = vec_lvsr(8, dst);
vec_s16_t sixv = vec_splat_s16(6);
vec_u16_t sixv = vec_splat_s16(6);
const vec_u8_t sel = (vec_u8_t) CV(0,0,0,0,0,0,0,0,-1,-1,-1,-1,-1,-1,-1,-1);
LOAD_ZERO;
......
......@@ -18,10 +18,6 @@
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111, USA.
*****************************************************************************/
#ifdef HAVE_ALTIVEC_H
#include <altivec.h>
#endif
#include "common/common.h"
#include "ppccommon.h"
#include "quant.h"
......@@ -53,31 +49,31 @@ temp2v = (vec_s16_t) vec_packs(vec_mergeh(multEvenvB, multOddvB), vec_mergel(mul
temp1v = vec_xor(temp1v, mskA); \
temp2v = vec_xor(temp2v, mskB); \
temp1v = vec_adds(temp1v, vec_and(mskA, one)); \
vec_st(temp1v, (dct0), dct); \
vec_st(temp1v, (dct0), (int16_t*)dct); \
temp2v = vec_adds(temp2v, vec_and(mskB, one)); \
vec_st(temp2v, (dct1), dct);
vec_st(temp2v, (dct1), (int16_t*)dct);
void x264_quant_4x4_altivec( int16_t dct[4][4], int quant_mf[4][4], int const i_qbits, int const f ) {
vector bool short mskA;
vec_s32_t i_qbitsv;
vec_u32_t i_qbitsv;
vec_u16_t coefvA;
vec_u32_t multEvenvA, multOddvA;
vec_u32_t mfvA;
vec_u16_t mfvA;
vec_s16_t zerov, one;
vec_s32_t fV;
vec_u32_t fV;
vector bool short mskB;
vec_u16_t coefvB;
vec_u32_t multEvenvB, multOddvB;
vec_u32_t mfvB;
vec_u16_t mfvB;
vec_s16_t temp1v, temp2v;
vect_sint_u qbits_u;
vect_int_u qbits_u;
qbits_u.s[0]=i_qbits;
i_qbitsv = vec_splat(qbits_u.v, 0);
vect_sint_u f_u;
vect_int_u f_u;
f_u.s[0]=f;
fV = vec_splat(f_u.v, 0);
......@@ -113,18 +109,18 @@ temp2v = (vec_s16_t) vec_packs(vec_mergeh(multEvenvB, multOddvB), vec_mergel(mul
temp1v = vec_xor(temp1v, mskA); \
temp2v = vec_xor(temp2v, mskB); \
temp1v = vec_add(temp1v, vec_and(mskA, one)); \
vec_st(temp1v, (dct0), dct); \
vec_st(temp1v, (dct0), (int16_t*)dct); \
temp2v = vec_add(temp2v, vec_and(mskB, one)); \
vec_st(temp2v, (dct1), dct);
vec_st(temp2v, (dct1), (int16_t*)dct);
void x264_quant_4x4_dc_altivec( int16_t dct[4][4], int i_quant_mf, int const i_qbits, int const f ) {
vector bool short mskA;
vec_s32_t i_qbitsv;
vec_u32_t i_qbitsv;
vec_u16_t coefvA;
vec_u32_t multEvenvA, multOddvA;
vec_s16_t zerov, one;
vec_s32_t fV;
vec_u32_t fV;
vector bool short mskB;
vec_u16_t coefvB;
......@@ -132,17 +128,16 @@ void x264_quant_4x4_dc_altivec( int16_t dct[4][4], int i_quant_mf, int const i_q
vec_s16_t temp1v, temp2v;
vec_u32_t mfv;
vect_int_u mf_u;
vec_u16_t mfv;
vect_ushort_u mf_u;
mf_u.s[0]=i_quant_mf;
mfv = vec_splat( mf_u.v, 0 );
mfv = vec_packs( mfv, mfv);
vect_sint_u qbits_u;
vect_int_u qbits_u;
qbits_u.s[0]=i_qbits;
i_qbitsv = vec_splat(qbits_u.v, 0);
vect_sint_u f_u;
vect_int_u f_u;
f_u.s[0]=f;
fV = vec_splat(f_u.v, 0);
......@@ -155,15 +150,17 @@ void x264_quant_4x4_dc_altivec( int16_t dct[4][4], int i_quant_mf, int const i_q
void x264_quant_8x8_altivec( int16_t dct[8][8], int quant_mf[8][8], int const i_qbits, int const f ) {
vector bool short mskA;
vec_s32_t i_qbitsv;
vec_u32_t i_qbitsv;
vec_u16_t coefvA;
vec_s32_t multEvenvA, multOddvA, mfvA;
vec_u32_t multEvenvA, multOddvA;
vec_u16_t mfvA;
vec_s16_t zerov, one;
vec_s32_t fV;
vec_u32_t fV;
vector bool short mskB;
vec_u16_t coefvB;
vec_u32_t multEvenvB, multOddvB, mfvB;
vec_u32_t multEvenvB, multOddvB;
vec_u16_t mfvB;
vec_s16_t temp1v, temp2v;
......@@ -171,7 +168,7 @@ void x264_quant_8x8_altivec( int16_t dct[8][8], int quant_mf[8][8], int const i_
qbits_u.s[0]=i_qbits;
i_qbitsv = vec_splat(qbits_u.v, 0);
vect_sint_u f_u;
vect_int_u f_u;
f_u.s[0]=f;
fV = vec_splat(f_u.v, 0);
......
......@@ -18,6 +18,10 @@
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111, USA.
*****************************************************************************/
#ifdef SYS_LINUX
#include <altivec.h>
#endif
#ifndef _PPC_QUANT_H
#define _PPC_QUANT_H 1
......@@ -27,10 +31,9 @@ typedef union {
} vect_int_u;
typedef union {
signed int s[4];
vector signed int v;
} vect_sint_u;
unsigned short s[8];
vector unsigned short v;
} vect_ushort_u;
void x264_quant_4x4_altivec( int16_t dct[4][4], int quant_mf[4][4], int const i_qbits, int const f );
void x264_quant_8x8_altivec( int16_t dct[8][8], int quant_mf[8][8], int const i_qbits, int const f );
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment