falkTX
/
FFmpeg
mirror of https://github.com/falkTX/FFmpeg.git
1
0
Fork 0
Code Issues 0 Releases 338 Wiki Activity
Browse Source

Merge commit 'a8b6015823e628047a45916404c00044c5e80415' 

* commit 'a8b6015823e628047a45916404c00044c5e80415':
  dsputil: convert remaining functions to use ptrdiff_t strides

Conflicts:
	libavcodec/dsputil.h
	libavcodec/dsputil_template.c
	libavcodec/h264qpel_template.c

Merged-by: Michael Niedermayer <michaelni@gmx.at>
tags/n2.0
Michael Niedermayer 12 years ago
parent
849f015406 a8b6015823
commit
db4e4f766c
17 changed files with 442 additions and 283 deletions
Split View
Diff Options
Show Stats Download Patch File Download Diff File
  1. +64
    -64
      libavcodec/arm/h264qpel_init_arm.c
  2. +1
    -1
      libavcodec/arm/rv40dsp_init_arm.c
  3. +30
    -15
      libavcodec/cavsdsp.c
  4. +114
    -57
      libavcodec/dsputil.c
  5. +12
    -12
      libavcodec/dsputil.h
  6. +32
    -16
      libavcodec/h264qpel_template.c
  7. +32
    -16
      libavcodec/ppc/h264_qpel.c
  8. +16
    -8
      libavcodec/rv30dsp.c
  9. +24
    -12
      libavcodec/rv40dsp.c
  10. +11
    -5
      libavcodec/vc1dsp.c
  11. +8
    -4
      libavcodec/x86/cavsdsp.c
  12. +41
    -41
      libavcodec/x86/dsputil_mmx.c
  13. +9
    -9
      libavcodec/x86/dsputil_mmx.h
  14. +39
    -19
      libavcodec/x86/h264_qpel.c
  15. +1
    -1
      libavcodec/x86/rv40dsp_init.c
  16. +1
    -1
      libavcodec/x86/vc1dsp_init.c
  17. +7
    -2
      libavcodec/x86/vc1dsp_mmx.c

+ 64
- 64
libavcodec/arm/h264qpel_init_arm.c View File

@@ -26,73 +26,73 @@
#include "libavutil/arm/cpu.h"
#include "libavcodec/h264qpel.h"

void ff_put_h264_qpel16_mc00_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel16_mc10_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel16_mc20_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel16_mc30_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel16_mc01_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel16_mc11_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel16_mc21_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel16_mc31_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel16_mc02_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel16_mc12_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel16_mc22_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel16_mc32_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel16_mc03_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel16_mc13_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel16_mc23_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel16_mc33_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel16_mc00_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel16_mc10_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel16_mc20_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel16_mc30_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel16_mc01_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel16_mc11_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel16_mc21_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel16_mc31_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel16_mc02_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel16_mc12_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel16_mc22_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel16_mc32_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel16_mc03_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel16_mc13_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel16_mc23_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel16_mc33_neon(uint8_t *, uint8_t *, ptrdiff_t);

void ff_put_h264_qpel8_mc00_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel8_mc10_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel8_mc20_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel8_mc30_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel8_mc01_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel8_mc11_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel8_mc21_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel8_mc31_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel8_mc02_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel8_mc12_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel8_mc22_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel8_mc32_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel8_mc03_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel8_mc13_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel8_mc23_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel8_mc33_neon(uint8_t *, uint8_t *, int);
void ff_put_h264_qpel8_mc00_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel8_mc10_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel8_mc20_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel8_mc30_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel8_mc01_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel8_mc11_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel8_mc21_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel8_mc31_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel8_mc02_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel8_mc12_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel8_mc22_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel8_mc32_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel8_mc03_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel8_mc13_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel8_mc23_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_put_h264_qpel8_mc33_neon(uint8_t *, uint8_t *, ptrdiff_t);

void ff_avg_h264_qpel16_mc00_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel16_mc10_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel16_mc20_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel16_mc30_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel16_mc01_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel16_mc11_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel16_mc21_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel16_mc31_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel16_mc02_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel16_mc12_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel16_mc22_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel16_mc32_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel16_mc03_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel16_mc13_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel16_mc23_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel16_mc33_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel16_mc00_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel16_mc10_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel16_mc20_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel16_mc30_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel16_mc01_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel16_mc11_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel16_mc21_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel16_mc31_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel16_mc02_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel16_mc12_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel16_mc22_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel16_mc32_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel16_mc03_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel16_mc13_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel16_mc23_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel16_mc33_neon(uint8_t *, uint8_t *, ptrdiff_t);

void ff_avg_h264_qpel8_mc00_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel8_mc10_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel8_mc20_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel8_mc30_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel8_mc01_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel8_mc11_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel8_mc21_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel8_mc31_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel8_mc02_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel8_mc12_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel8_mc22_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel8_mc32_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel8_mc03_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel8_mc13_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel8_mc23_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel8_mc33_neon(uint8_t *, uint8_t *, int);
void ff_avg_h264_qpel8_mc00_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel8_mc10_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel8_mc20_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel8_mc30_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel8_mc01_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel8_mc11_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel8_mc21_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel8_mc31_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel8_mc02_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel8_mc12_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel8_mc22_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel8_mc32_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel8_mc03_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel8_mc13_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel8_mc23_neon(uint8_t *, uint8_t *, ptrdiff_t);
void ff_avg_h264_qpel8_mc33_neon(uint8_t *, uint8_t *, ptrdiff_t);

av_cold void ff_h264qpel_init_arm(H264QpelContext *c, int bit_depth)
{


+ 1
- 1
libavcodec/arm/rv40dsp_init_arm.c View File

@@ -27,7 +27,7 @@

#define DECL_QPEL3(type, w, pos) \
void ff_##type##_rv40_qpel##w##_mc##pos##_neon(uint8_t *dst, uint8_t *src,\
int stride)
ptrdiff_t stride)
#define DECL_QPEL2(w, pos) \
DECL_QPEL3(put, w, pos); \
DECL_QPEL3(avg, w, pos)


+ 30
- 15
libavcodec/cavsdsp.c View File

@@ -421,63 +421,78 @@ static void OPNAME ## cavs_filt16_hv_ ## NAME(uint8_t *dst, uint8_t *src1, uint8
}\

#define CAVS_MC(OPNAME, SIZE) \
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc10_c(uint8_t *dst, uint8_t *src, int stride){\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc10_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## cavs_filt ## SIZE ## _h_qpel_l(dst, src, stride, stride);\
}\
\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc20_c(uint8_t *dst, uint8_t *src, int stride){\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc20_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## cavs_filt ## SIZE ## _h_hpel(dst, src, stride, stride);\
}\
\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc30_c(uint8_t *dst, uint8_t *src, int stride){\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc30_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## cavs_filt ## SIZE ## _h_qpel_r(dst, src, stride, stride);\
}\
\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc01_c(uint8_t *dst, uint8_t *src, int stride){\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc01_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## cavs_filt ## SIZE ## _v_qpel_l(dst, src, stride, stride);\
}\
\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc02_c(uint8_t *dst, uint8_t *src, int stride){\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc02_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## cavs_filt ## SIZE ## _v_hpel(dst, src, stride, stride);\
}\
\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc03_c(uint8_t *dst, uint8_t *src, int stride){\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc03_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## cavs_filt ## SIZE ## _v_qpel_r(dst, src, stride, stride);\
}\
\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc22_c(uint8_t *dst, uint8_t *src, int stride){\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc22_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## cavs_filt ## SIZE ## _hv_jj(dst, src, NULL, stride, stride); \
}\
\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc11_c(uint8_t *dst, uint8_t *src, int stride){\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc11_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## cavs_filt ## SIZE ## _hv_egpr(dst, src, src, stride, stride); \
}\
\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc13_c(uint8_t *dst, uint8_t *src, int stride){\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc13_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## cavs_filt ## SIZE ## _hv_egpr(dst, src, src+stride, stride, stride); \
}\
\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc31_c(uint8_t *dst, uint8_t *src, int stride){\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc31_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## cavs_filt ## SIZE ## _hv_egpr(dst, src, src+1, stride, stride); \
}\
\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc33_c(uint8_t *dst, uint8_t *src, int stride){\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc33_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## cavs_filt ## SIZE ## _hv_egpr(dst, src, src+stride+1,stride, stride); \
}\
\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc21_c(uint8_t *dst, uint8_t *src, int stride){\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc21_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## cavs_filt ## SIZE ## _hv_ff(dst, src, src+stride+1,stride, stride); \
}\
\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc12_c(uint8_t *dst, uint8_t *src, int stride){\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc12_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## cavs_filt ## SIZE ## _hv_ii(dst, src, src+stride+1,stride, stride); \
}\
\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc32_c(uint8_t *dst, uint8_t *src, int stride){\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc32_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## cavs_filt ## SIZE ## _hv_kk(dst, src, src+stride+1,stride, stride); \
}\
\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc23_c(uint8_t *dst, uint8_t *src, int stride){\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc23_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## cavs_filt ## SIZE ## _hv_qq(dst, src, src+stride+1,stride, stride); \
}\



+ 114
- 57
libavcodec/dsputil.c View File

@@ -925,23 +925,27 @@ static void OPNAME ## mpeg4_qpel16_v_lowpass(uint8_t *dst, uint8_t *src, int dst
}\
}\
\
static void OPNAME ## qpel8_mc10_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel8_mc10_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t half[64];\
put ## RND ## mpeg4_qpel8_h_lowpass(half, src, 8, stride, 8);\
OPNAME ## pixels8_l2_8(dst, src, half, stride, stride, 8, 8);\
}\
\
static void OPNAME ## qpel8_mc20_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel8_mc20_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## mpeg4_qpel8_h_lowpass(dst, src, stride, stride, 8);\
}\
\
static void OPNAME ## qpel8_mc30_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel8_mc30_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t half[64];\
put ## RND ## mpeg4_qpel8_h_lowpass(half, src, 8, stride, 8);\
OPNAME ## pixels8_l2_8(dst, src+1, half, stride, stride, 8, 8);\
}\
\
static void OPNAME ## qpel8_mc01_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel8_mc01_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[16*9];\
uint8_t half[64];\
copy_block9(full, src, 16, stride, 9);\
@@ -949,20 +953,23 @@ static void OPNAME ## qpel8_mc01_c(uint8_t *dst, uint8_t *src, int stride){\
OPNAME ## pixels8_l2_8(dst, full, half, stride, 16, 8, 8);\
}\
\
static void OPNAME ## qpel8_mc02_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel8_mc02_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[16*9];\
copy_block9(full, src, 16, stride, 9);\
OPNAME ## mpeg4_qpel8_v_lowpass(dst, full, stride, 16);\
}\
\
static void OPNAME ## qpel8_mc03_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel8_mc03_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[16*9];\
uint8_t half[64];\
copy_block9(full, src, 16, stride, 9);\
put ## RND ## mpeg4_qpel8_v_lowpass(half, full, 8, 16);\
OPNAME ## pixels8_l2_8(dst, full+16, half, stride, 16, 8, 8);\
}\
void ff_ ## OPNAME ## qpel8_mc11_old_c(uint8_t *dst, uint8_t *src, int stride){\
void ff_ ## OPNAME ## qpel8_mc11_old_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[16*9];\
uint8_t halfH[72];\
uint8_t halfV[64];\
@@ -973,7 +980,8 @@ void ff_ ## OPNAME ## qpel8_mc11_old_c(uint8_t *dst, uint8_t *src, int stride){\
put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
OPNAME ## pixels8_l4_8(dst, full, halfH, halfV, halfHV, stride, 16, 8, 8, 8, 8);\
}\
static void OPNAME ## qpel8_mc11_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel8_mc11_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[16*9];\
uint8_t halfH[72];\
uint8_t halfHV[64];\
@@ -983,7 +991,8 @@ static void OPNAME ## qpel8_mc11_c(uint8_t *dst, uint8_t *src, int stride){\
put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
OPNAME ## pixels8_l2_8(dst, halfH, halfHV, stride, 8, 8, 8);\
}\
void ff_ ## OPNAME ## qpel8_mc31_old_c(uint8_t *dst, uint8_t *src, int stride){\
void ff_ ## OPNAME ## qpel8_mc31_old_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[16*9];\
uint8_t halfH[72];\
uint8_t halfV[64];\
@@ -994,7 +1003,8 @@ void ff_ ## OPNAME ## qpel8_mc31_old_c(uint8_t *dst, uint8_t *src, int stride){\
put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
OPNAME ## pixels8_l4_8(dst, full+1, halfH, halfV, halfHV, stride, 16, 8, 8, 8, 8);\
}\
static void OPNAME ## qpel8_mc31_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel8_mc31_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[16*9];\
uint8_t halfH[72];\
uint8_t halfHV[64];\
@@ -1004,7 +1014,8 @@ static void OPNAME ## qpel8_mc31_c(uint8_t *dst, uint8_t *src, int stride){\
put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
OPNAME ## pixels8_l2_8(dst, halfH, halfHV, stride, 8, 8, 8);\
}\
void ff_ ## OPNAME ## qpel8_mc13_old_c(uint8_t *dst, uint8_t *src, int stride){\
void ff_ ## OPNAME ## qpel8_mc13_old_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[16*9];\
uint8_t halfH[72];\
uint8_t halfV[64];\
@@ -1015,7 +1026,8 @@ void ff_ ## OPNAME ## qpel8_mc13_old_c(uint8_t *dst, uint8_t *src, int stride){\
put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
OPNAME ## pixels8_l4_8(dst, full+16, halfH+8, halfV, halfHV, stride, 16, 8, 8, 8, 8);\
}\
static void OPNAME ## qpel8_mc13_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel8_mc13_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[16*9];\
uint8_t halfH[72];\
uint8_t halfHV[64];\
@@ -1025,7 +1037,8 @@ static void OPNAME ## qpel8_mc13_c(uint8_t *dst, uint8_t *src, int stride){\
put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
OPNAME ## pixels8_l2_8(dst, halfH+8, halfHV, stride, 8, 8, 8);\
}\
void ff_ ## OPNAME ## qpel8_mc33_old_c(uint8_t *dst, uint8_t *src, int stride){\
void ff_ ## OPNAME ## qpel8_mc33_old_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[16*9];\
uint8_t halfH[72];\
uint8_t halfV[64];\
@@ -1036,7 +1049,8 @@ void ff_ ## OPNAME ## qpel8_mc33_old_c(uint8_t *dst, uint8_t *src, int stride){\
put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
OPNAME ## pixels8_l4_8(dst, full+17, halfH+8, halfV, halfHV, stride, 16, 8, 8, 8, 8);\
}\
static void OPNAME ## qpel8_mc33_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel8_mc33_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[16*9];\
uint8_t halfH[72];\
uint8_t halfHV[64];\
@@ -1046,21 +1060,24 @@ static void OPNAME ## qpel8_mc33_c(uint8_t *dst, uint8_t *src, int stride){\
put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
OPNAME ## pixels8_l2_8(dst, halfH+8, halfHV, stride, 8, 8, 8);\
}\
static void OPNAME ## qpel8_mc21_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel8_mc21_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t halfH[72];\
uint8_t halfHV[64];\
put ## RND ## mpeg4_qpel8_h_lowpass(halfH, src, 8, stride, 9);\
put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
OPNAME ## pixels8_l2_8(dst, halfH, halfHV, stride, 8, 8, 8);\
}\
static void OPNAME ## qpel8_mc23_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel8_mc23_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t halfH[72];\
uint8_t halfHV[64];\
put ## RND ## mpeg4_qpel8_h_lowpass(halfH, src, 8, stride, 9);\
put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
OPNAME ## pixels8_l2_8(dst, halfH+8, halfHV, stride, 8, 8, 8);\
}\
void ff_ ## OPNAME ## qpel8_mc12_old_c(uint8_t *dst, uint8_t *src, int stride){\
void ff_ ## OPNAME ## qpel8_mc12_old_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[16*9];\
uint8_t halfH[72];\
uint8_t halfV[64];\
@@ -1071,7 +1088,8 @@ void ff_ ## OPNAME ## qpel8_mc12_old_c(uint8_t *dst, uint8_t *src, int stride){\
put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
OPNAME ## pixels8_l2_8(dst, halfV, halfHV, stride, 8, 8, 8);\
}\
static void OPNAME ## qpel8_mc12_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel8_mc12_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[16*9];\
uint8_t halfH[72];\
copy_block9(full, src, 16, stride, 9);\
@@ -1079,7 +1097,8 @@ static void OPNAME ## qpel8_mc12_c(uint8_t *dst, uint8_t *src, int stride){\
put ## RND ## pixels8_l2_8(halfH, halfH, full, 8, 8, 16, 9);\
OPNAME ## mpeg4_qpel8_v_lowpass(dst, halfH, stride, 8);\
}\
void ff_ ## OPNAME ## qpel8_mc32_old_c(uint8_t *dst, uint8_t *src, int stride){\
void ff_ ## OPNAME ## qpel8_mc32_old_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[16*9];\
uint8_t halfH[72];\
uint8_t halfV[64];\
@@ -1090,7 +1109,8 @@ void ff_ ## OPNAME ## qpel8_mc32_old_c(uint8_t *dst, uint8_t *src, int stride){\
put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
OPNAME ## pixels8_l2_8(dst, halfV, halfHV, stride, 8, 8, 8);\
}\
static void OPNAME ## qpel8_mc32_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel8_mc32_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[16*9];\
uint8_t halfH[72];\
copy_block9(full, src, 16, stride, 9);\
@@ -1098,29 +1118,34 @@ static void OPNAME ## qpel8_mc32_c(uint8_t *dst, uint8_t *src, int stride){\
put ## RND ## pixels8_l2_8(halfH, halfH, full+1, 8, 8, 16, 9);\
OPNAME ## mpeg4_qpel8_v_lowpass(dst, halfH, stride, 8);\
}\
static void OPNAME ## qpel8_mc22_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel8_mc22_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t halfH[72];\
put ## RND ## mpeg4_qpel8_h_lowpass(halfH, src, 8, stride, 9);\
OPNAME ## mpeg4_qpel8_v_lowpass(dst, halfH, stride, 8);\
}\
\
static void OPNAME ## qpel16_mc10_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel16_mc10_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t half[256];\
put ## RND ## mpeg4_qpel16_h_lowpass(half, src, 16, stride, 16);\
OPNAME ## pixels16_l2_8(dst, src, half, stride, stride, 16, 16);\
}\
\
static void OPNAME ## qpel16_mc20_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel16_mc20_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## mpeg4_qpel16_h_lowpass(dst, src, stride, stride, 16);\
}\
\
static void OPNAME ## qpel16_mc30_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel16_mc30_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t half[256];\
put ## RND ## mpeg4_qpel16_h_lowpass(half, src, 16, stride, 16);\
OPNAME ## pixels16_l2_8(dst, src+1, half, stride, stride, 16, 16);\
}\
\
static void OPNAME ## qpel16_mc01_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel16_mc01_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[24*17];\
uint8_t half[256];\
copy_block17(full, src, 24, stride, 17);\
@@ -1128,20 +1153,23 @@ static void OPNAME ## qpel16_mc01_c(uint8_t *dst, uint8_t *src, int stride){\
OPNAME ## pixels16_l2_8(dst, full, half, stride, 24, 16, 16);\
}\
\
static void OPNAME ## qpel16_mc02_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel16_mc02_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[24*17];\
copy_block17(full, src, 24, stride, 17);\
OPNAME ## mpeg4_qpel16_v_lowpass(dst, full, stride, 24);\
}\
\
static void OPNAME ## qpel16_mc03_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel16_mc03_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[24*17];\
uint8_t half[256];\
copy_block17(full, src, 24, stride, 17);\
put ## RND ## mpeg4_qpel16_v_lowpass(half, full, 16, 24);\
OPNAME ## pixels16_l2_8(dst, full+24, half, stride, 24, 16, 16);\
}\
void ff_ ## OPNAME ## qpel16_mc11_old_c(uint8_t *dst, uint8_t *src, int stride){\
void ff_ ## OPNAME ## qpel16_mc11_old_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[24*17];\
uint8_t halfH[272];\
uint8_t halfV[256];\
@@ -1152,7 +1180,8 @@ void ff_ ## OPNAME ## qpel16_mc11_old_c(uint8_t *dst, uint8_t *src, int stride){
put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
OPNAME ## pixels16_l4_8(dst, full, halfH, halfV, halfHV, stride, 24, 16, 16, 16, 16);\
}\
static void OPNAME ## qpel16_mc11_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel16_mc11_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[24*17];\
uint8_t halfH[272];\
uint8_t halfHV[256];\
@@ -1162,7 +1191,8 @@ static void OPNAME ## qpel16_mc11_c(uint8_t *dst, uint8_t *src, int stride){\
put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
OPNAME ## pixels16_l2_8(dst, halfH, halfHV, stride, 16, 16, 16);\
}\
void ff_ ## OPNAME ## qpel16_mc31_old_c(uint8_t *dst, uint8_t *src, int stride){\
void ff_ ## OPNAME ## qpel16_mc31_old_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[24*17];\
uint8_t halfH[272];\
uint8_t halfV[256];\
@@ -1173,7 +1203,8 @@ void ff_ ## OPNAME ## qpel16_mc31_old_c(uint8_t *dst, uint8_t *src, int stride){
put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
OPNAME ## pixels16_l4_8(dst, full+1, halfH, halfV, halfHV, stride, 24, 16, 16, 16, 16);\
}\
static void OPNAME ## qpel16_mc31_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel16_mc31_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[24*17];\
uint8_t halfH[272];\
uint8_t halfHV[256];\
@@ -1183,7 +1214,8 @@ static void OPNAME ## qpel16_mc31_c(uint8_t *dst, uint8_t *src, int stride){\
put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
OPNAME ## pixels16_l2_8(dst, halfH, halfHV, stride, 16, 16, 16);\
}\
void ff_ ## OPNAME ## qpel16_mc13_old_c(uint8_t *dst, uint8_t *src, int stride){\
void ff_ ## OPNAME ## qpel16_mc13_old_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[24*17];\
uint8_t halfH[272];\
uint8_t halfV[256];\
@@ -1194,7 +1226,8 @@ void ff_ ## OPNAME ## qpel16_mc13_old_c(uint8_t *dst, uint8_t *src, int stride){
put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
OPNAME ## pixels16_l4_8(dst, full+24, halfH+16, halfV, halfHV, stride, 24, 16, 16, 16, 16);\
}\
static void OPNAME ## qpel16_mc13_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel16_mc13_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[24*17];\
uint8_t halfH[272];\
uint8_t halfHV[256];\
@@ -1204,7 +1237,8 @@ static void OPNAME ## qpel16_mc13_c(uint8_t *dst, uint8_t *src, int stride){\
put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
OPNAME ## pixels16_l2_8(dst, halfH+16, halfHV, stride, 16, 16, 16);\
}\
void ff_ ## OPNAME ## qpel16_mc33_old_c(uint8_t *dst, uint8_t *src, int stride){\
void ff_ ## OPNAME ## qpel16_mc33_old_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[24*17];\
uint8_t halfH[272];\
uint8_t halfV[256];\
@@ -1215,7 +1249,8 @@ void ff_ ## OPNAME ## qpel16_mc33_old_c(uint8_t *dst, uint8_t *src, int stride){
put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
OPNAME ## pixels16_l4_8(dst, full+25, halfH+16, halfV, halfHV, stride, 24, 16, 16, 16, 16);\
}\
static void OPNAME ## qpel16_mc33_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel16_mc33_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[24*17];\
uint8_t halfH[272];\
uint8_t halfHV[256];\
@@ -1225,21 +1260,24 @@ static void OPNAME ## qpel16_mc33_c(uint8_t *dst, uint8_t *src, int stride){\
put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
OPNAME ## pixels16_l2_8(dst, halfH+16, halfHV, stride, 16, 16, 16);\
}\
static void OPNAME ## qpel16_mc21_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel16_mc21_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t halfH[272];\
uint8_t halfHV[256];\
put ## RND ## mpeg4_qpel16_h_lowpass(halfH, src, 16, stride, 17);\
put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
OPNAME ## pixels16_l2_8(dst, halfH, halfHV, stride, 16, 16, 16);\
}\
static void OPNAME ## qpel16_mc23_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel16_mc23_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t halfH[272];\
uint8_t halfHV[256];\
put ## RND ## mpeg4_qpel16_h_lowpass(halfH, src, 16, stride, 17);\
put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
OPNAME ## pixels16_l2_8(dst, halfH+16, halfHV, stride, 16, 16, 16);\
}\
void ff_ ## OPNAME ## qpel16_mc12_old_c(uint8_t *dst, uint8_t *src, int stride){\
void ff_ ## OPNAME ## qpel16_mc12_old_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[24*17];\
uint8_t halfH[272];\
uint8_t halfV[256];\
@@ -1250,7 +1288,8 @@ void ff_ ## OPNAME ## qpel16_mc12_old_c(uint8_t *dst, uint8_t *src, int stride){
put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
OPNAME ## pixels16_l2_8(dst, halfV, halfHV, stride, 16, 16, 16);\
}\
static void OPNAME ## qpel16_mc12_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel16_mc12_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[24*17];\
uint8_t halfH[272];\
copy_block17(full, src, 24, stride, 17);\
@@ -1258,7 +1297,8 @@ static void OPNAME ## qpel16_mc12_c(uint8_t *dst, uint8_t *src, int stride){\
put ## RND ## pixels16_l2_8(halfH, halfH, full, 16, 16, 24, 17);\
OPNAME ## mpeg4_qpel16_v_lowpass(dst, halfH, stride, 16);\
}\
void ff_ ## OPNAME ## qpel16_mc32_old_c(uint8_t *dst, uint8_t *src, int stride){\
void ff_ ## OPNAME ## qpel16_mc32_old_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[24*17];\
uint8_t halfH[272];\
uint8_t halfV[256];\
@@ -1269,7 +1309,8 @@ void ff_ ## OPNAME ## qpel16_mc32_old_c(uint8_t *dst, uint8_t *src, int stride){
put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
OPNAME ## pixels16_l2_8(dst, halfV, halfHV, stride, 16, 16, 16);\
}\
static void OPNAME ## qpel16_mc32_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel16_mc32_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[24*17];\
uint8_t halfH[272];\
copy_block17(full, src, 24, stride, 17);\
@@ -1277,7 +1318,8 @@ static void OPNAME ## qpel16_mc32_c(uint8_t *dst, uint8_t *src, int stride){\
put ## RND ## pixels16_l2_8(halfH, halfH, full+1, 16, 16, 24, 17);\
OPNAME ## mpeg4_qpel16_v_lowpass(dst, halfH, stride, 16);\
}\
static void OPNAME ## qpel16_mc22_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## qpel16_mc22_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t halfH[272];\
put ## RND ## mpeg4_qpel16_h_lowpass(halfH, src, 16, stride, 17);\
OPNAME ## mpeg4_qpel16_v_lowpass(dst, halfH, stride, 16);\
@@ -1297,16 +1339,20 @@ QPEL_MC(0, avg_ , _ , op_avg)
#undef op_put
#undef op_put_no_rnd

void ff_put_pixels8x8_c(uint8_t *dst, uint8_t *src, int stride) {
void ff_put_pixels8x8_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
{
put_pixels8_8_c(dst, src, stride, 8);
}
void ff_avg_pixels8x8_c(uint8_t *dst, uint8_t *src, int stride) {
void ff_avg_pixels8x8_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
{
avg_pixels8_8_c(dst, src, stride, 8);
}
void ff_put_pixels16x16_c(uint8_t *dst, uint8_t *src, int stride) {
void ff_put_pixels16x16_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
{
put_pixels16_8_c(dst, src, stride, 16);
}
void ff_avg_pixels16x16_c(uint8_t *dst, uint8_t *src, int stride) {
void ff_avg_pixels16x16_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
{
avg_pixels16_8_c(dst, src, stride, 16);
}

@@ -1336,16 +1382,20 @@ static void wmv2_mspel8_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int
}

#if CONFIG_RV40_DECODER
void ff_put_rv40_qpel16_mc33_c(uint8_t *dst, uint8_t *src, int stride){
void ff_put_rv40_qpel16_mc33_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
{
put_pixels16_xy2_8_c(dst, src, stride, 16);
}
void ff_avg_rv40_qpel16_mc33_c(uint8_t *dst, uint8_t *src, int stride){
void ff_avg_rv40_qpel16_mc33_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
{
avg_pixels16_xy2_8_c(dst, src, stride, 16);
}
void ff_put_rv40_qpel8_mc33_c(uint8_t *dst, uint8_t *src, int stride){
void ff_put_rv40_qpel8_mc33_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
{
put_pixels8_xy2_8_c(dst, src, stride, 8);
}
void ff_avg_rv40_qpel8_mc33_c(uint8_t *dst, uint8_t *src, int stride){
void ff_avg_rv40_qpel8_mc33_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
{
avg_pixels8_xy2_8_c(dst, src, stride, 8);
}
#endif /* CONFIG_RV40_DECODER */
@@ -1424,27 +1474,32 @@ static void wmv2_mspel8_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int
}
}

static void put_mspel8_mc10_c(uint8_t *dst, uint8_t *src, int stride){
static void put_mspel8_mc10_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
{
uint8_t half[64];
wmv2_mspel8_h_lowpass(half, src, 8, stride, 8);
put_pixels8_l2_8(dst, src, half, stride, stride, 8, 8);
}

static void put_mspel8_mc20_c(uint8_t *dst, uint8_t *src, int stride){
static void put_mspel8_mc20_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
{
wmv2_mspel8_h_lowpass(dst, src, stride, stride, 8);
}

static void put_mspel8_mc30_c(uint8_t *dst, uint8_t *src, int stride){
static void put_mspel8_mc30_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
{
uint8_t half[64];
wmv2_mspel8_h_lowpass(half, src, 8, stride, 8);
put_pixels8_l2_8(dst, src+1, half, stride, stride, 8, 8);
}

static void put_mspel8_mc02_c(uint8_t *dst, uint8_t *src, int stride){
static void put_mspel8_mc02_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
{
wmv2_mspel8_v_lowpass(dst, src, stride, stride, 8);
}

static void put_mspel8_mc12_c(uint8_t *dst, uint8_t *src, int stride){
static void put_mspel8_mc12_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
{
uint8_t halfH[88];
uint8_t halfV[64];
uint8_t halfHV[64];
@@ -1453,7 +1508,8 @@ static void put_mspel8_mc12_c(uint8_t *dst, uint8_t *src, int stride){
wmv2_mspel8_v_lowpass(halfHV, halfH+8, 8, 8, 8);
put_pixels8_l2_8(dst, halfV, halfHV, stride, 8, 8, 8);
}
static void put_mspel8_mc32_c(uint8_t *dst, uint8_t *src, int stride){
static void put_mspel8_mc32_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
{
uint8_t halfH[88];
uint8_t halfV[64];
uint8_t halfHV[64];
@@ -1462,7 +1518,8 @@ static void put_mspel8_mc32_c(uint8_t *dst, uint8_t *src, int stride){
wmv2_mspel8_v_lowpass(halfHV, halfH+8, 8, 8, 8);
put_pixels8_l2_8(dst, halfV, halfHV, stride, 8, 8, 8);
}
static void put_mspel8_mc22_c(uint8_t *dst, uint8_t *src, int stride){
static void put_mspel8_mc22_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
{
uint8_t halfH[88];
wmv2_mspel8_h_lowpass(halfH, src-stride, 8, stride, 11);
wmv2_mspel8_v_lowpass(dst, halfH+8, stride, 8, 8);


+ 12
- 12
libavcodec/dsputil.h View File

@@ -50,16 +50,16 @@ extern const uint8_t ff_zigzag248_direct[64];
extern uint32_t ff_squareTbl[512];
extern const uint8_t ff_cropTbl[256 + 2 * MAX_NEG_CROP];

void ff_put_pixels8x8_c(uint8_t *dst, uint8_t *src, int stride);
void ff_avg_pixels8x8_c(uint8_t *dst, uint8_t *src, int stride);
void ff_put_pixels16x16_c(uint8_t *dst, uint8_t *src, int stride);
void ff_avg_pixels16x16_c(uint8_t *dst, uint8_t *src, int stride);
void ff_put_pixels8x8_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride);
void ff_avg_pixels8x8_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride);
void ff_put_pixels16x16_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride);
void ff_avg_pixels16x16_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride);

/* RV40 functions */
void ff_put_rv40_qpel16_mc33_c(uint8_t *dst, uint8_t *src, int stride);
void ff_avg_rv40_qpel16_mc33_c(uint8_t *dst, uint8_t *src, int stride);
void ff_put_rv40_qpel8_mc33_c(uint8_t *dst, uint8_t *src, int stride);
void ff_avg_rv40_qpel8_mc33_c(uint8_t *dst, uint8_t *src, int stride);
void ff_put_rv40_qpel16_mc33_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride);
void ff_avg_rv40_qpel16_mc33_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride);
void ff_put_rv40_qpel8_mc33_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride);
void ff_avg_rv40_qpel8_mc33_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride);

void ff_gmc_c(uint8_t *dst, uint8_t *src, int stride, int h, int ox, int oy,
int dxx, int dxy, int dyx, int dyy, int shift, int r, int width, int height);
@@ -81,14 +81,14 @@ could be reached easily ...
// blocksizes for op_pixels_func are 8x4,8x8 16x8 16x16
//h for op_pixels_func is limited to {width/2, width} but never larger than 16 and never smaller than 4
typedef void (*tpel_mc_func)(uint8_t *block/*align width (8 or 16)*/, const uint8_t *pixels/*align 1*/, int line_size, int w, int h);
typedef void (*qpel_mc_func)(uint8_t *dst/*align width (8 or 16)*/, uint8_t *src/*align 1*/, int stride);
typedef void (*qpel_mc_func)(uint8_t *dst/*align width (8 or 16)*/, uint8_t *src/*align 1*/, ptrdiff_t stride);

typedef void (*op_fill_func)(uint8_t *block/*align width (8 or 16)*/, uint8_t value, int line_size, int h);

#define DEF_OLD_QPEL(name)\
void ff_put_ ## name (uint8_t *dst/*align width (8 or 16)*/, uint8_t *src/*align 1*/, int stride);\
void ff_put_no_rnd_ ## name (uint8_t *dst/*align width (8 or 16)*/, uint8_t *src/*align 1*/, int stride);\
void ff_avg_ ## name (uint8_t *dst/*align width (8 or 16)*/, uint8_t *src/*align 1*/, int stride);
void ff_put_ ## name (uint8_t *dst/*align width (8 or 16)*/, uint8_t *src/*align 1*/, ptrdiff_t stride);\
void ff_put_no_rnd_ ## name (uint8_t *dst/*align width (8 or 16)*/, uint8_t *src/*align 1*/, ptrdiff_t stride);\
void ff_avg_ ## name (uint8_t *dst/*align width (8 or 16)*/, uint8_t *src/*align 1*/, ptrdiff_t stride);

DEF_OLD_QPEL(qpel16_mc11_old_c)
DEF_OLD_QPEL(qpel16_mc31_old_c)


+ 32
- 16
libavcodec/h264qpel_template.c View File

@@ -375,27 +375,32 @@ static void FUNC(OPNAME ## h264_qpel16_hv_lowpass)(uint8_t *dst, pixeltmp *tmp,
}\

#define H264_MC(OPNAME, SIZE) \
static av_unused void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc00)(uint8_t *dst, uint8_t *src, int stride){\
static av_unused void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc00)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
FUNCC(OPNAME ## pixels ## SIZE)(dst, src, stride, SIZE);\
}\
\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc10)(uint8_t *dst, uint8_t *src, int stride){\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc10)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t half[SIZE*SIZE*sizeof(pixel)];\
FUNC(put_h264_qpel ## SIZE ## _h_lowpass)(half, src, SIZE*sizeof(pixel), stride);\
FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, src, half, stride, stride, SIZE*sizeof(pixel), SIZE);\
}\
\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc20)(uint8_t *dst, uint8_t *src, int stride){\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc20)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
FUNC(OPNAME ## h264_qpel ## SIZE ## _h_lowpass)(dst, src, stride, stride);\
}\
\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc30)(uint8_t *dst, uint8_t *src, int stride){\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc30)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t half[SIZE*SIZE*sizeof(pixel)];\
FUNC(put_h264_qpel ## SIZE ## _h_lowpass)(half, src, SIZE*sizeof(pixel), stride);\
FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, src+sizeof(pixel), half, stride, stride, SIZE*sizeof(pixel), SIZE);\
}\
\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc01)(uint8_t *dst, uint8_t *src, int stride){\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc01)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\
uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\
uint8_t half[SIZE*SIZE*sizeof(pixel)];\
@@ -404,14 +409,16 @@ static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc01)(uint8_t *dst, uint8_t *s
FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, full_mid, half, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
}\
\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc02)(uint8_t *dst, uint8_t *src, int stride){\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc02)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\
uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\
FUNC(copy_block ## SIZE )(full, src - stride*2, SIZE*sizeof(pixel), stride, SIZE + 5);\
FUNC(OPNAME ## h264_qpel ## SIZE ## _v_lowpass)(dst, full_mid, stride, SIZE*sizeof(pixel));\
}\
\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc03)(uint8_t *dst, uint8_t *src, int stride){\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc03)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\
uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\
uint8_t half[SIZE*SIZE*sizeof(pixel)];\
@@ -420,7 +427,8 @@ static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc03)(uint8_t *dst, uint8_t *s
FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, full_mid+SIZE*sizeof(pixel), half, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
}\
\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc11)(uint8_t *dst, uint8_t *src, int stride){\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc11)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\
uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\
uint8_t halfH[SIZE*SIZE*sizeof(pixel)];\
@@ -431,7 +439,8 @@ static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc11)(uint8_t *dst, uint8_t *s
FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, halfH, halfV, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
}\
\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc31)(uint8_t *dst, uint8_t *src, int stride){\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc31)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\
uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\
uint8_t halfH[SIZE*SIZE*sizeof(pixel)];\
@@ -442,7 +451,8 @@ static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc31)(uint8_t *dst, uint8_t *s
FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, halfH, halfV, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
}\
\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc13)(uint8_t *dst, uint8_t *src, int stride){\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc13)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\
uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\
uint8_t halfH[SIZE*SIZE*sizeof(pixel)];\
@@ -453,7 +463,8 @@ static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc13)(uint8_t *dst, uint8_t *s
FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, halfH, halfV, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
}\
\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc33)(uint8_t *dst, uint8_t *src, int stride){\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc33)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\
uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\
uint8_t halfH[SIZE*SIZE*sizeof(pixel)];\
@@ -464,12 +475,14 @@ static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc33)(uint8_t *dst, uint8_t *s
FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, halfH, halfV, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
}\
\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc22)(uint8_t *dst, uint8_t *src, int stride){\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc22)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
pixeltmp tmp[SIZE*(SIZE+5)*sizeof(pixel)];\
FUNC(OPNAME ## h264_qpel ## SIZE ## _hv_lowpass)(dst, tmp, src, stride, SIZE*sizeof(pixel), stride);\
}\
\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc21)(uint8_t *dst, uint8_t *src, int stride){\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc21)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
pixeltmp tmp[SIZE*(SIZE+5)*sizeof(pixel)];\
uint8_t halfH[SIZE*SIZE*sizeof(pixel)];\
uint8_t halfHV[SIZE*SIZE*sizeof(pixel)];\
@@ -478,7 +491,8 @@ static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc21)(uint8_t *dst, uint8_t *s
FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, halfH, halfHV, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
}\
\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc23)(uint8_t *dst, uint8_t *src, int stride){\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc23)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
pixeltmp tmp[SIZE*(SIZE+5)*sizeof(pixel)];\
uint8_t halfH[SIZE*SIZE*sizeof(pixel)];\
uint8_t halfHV[SIZE*SIZE*sizeof(pixel)];\
@@ -487,7 +501,8 @@ static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc23)(uint8_t *dst, uint8_t *s
FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, halfH, halfHV, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
}\
\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc12)(uint8_t *dst, uint8_t *src, int stride){\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc12)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\
uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\
pixeltmp tmp[SIZE*(SIZE+5)*sizeof(pixel)];\
@@ -499,7 +514,8 @@ static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc12)(uint8_t *dst, uint8_t *s
FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, halfV, halfHV, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
}\
\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc32)(uint8_t *dst, uint8_t *src, int stride){\
static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc32)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\
uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\
pixeltmp tmp[SIZE*(SIZE+5)*sizeof(pixel)];\


+ 32
- 16
libavcodec/ppc/h264_qpel.c View File

@@ -65,43 +65,51 @@
#undef PREFIX_h264_qpel16_hv_lowpass_num

#define H264_MC(OPNAME, SIZE, CODETYPE) \
static void OPNAME ## h264_qpel ## SIZE ## _mc00_ ## CODETYPE (uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc00_ ## CODETYPE (uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
ff_ ## OPNAME ## pixels ## SIZE ## _ ## CODETYPE(dst, src, stride, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc10_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){ \
static void OPNAME ## h264_qpel ## SIZE ## _mc10_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{ \
DECLARE_ALIGNED(16, uint8_t, half)[SIZE*SIZE];\
put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(half, src, SIZE, stride);\
OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, src, half, stride, stride, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc20_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc20_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(dst, src, stride, stride);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc30_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc30_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
DECLARE_ALIGNED(16, uint8_t, half)[SIZE*SIZE];\
put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(half, src, SIZE, stride);\
OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, src+1, half, stride, stride, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc01_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc01_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
DECLARE_ALIGNED(16, uint8_t, half)[SIZE*SIZE];\
put_h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(half, src, SIZE, stride);\
OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, src, half, stride, stride, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc02_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc02_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(dst, src, stride, stride);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc03_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc03_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
DECLARE_ALIGNED(16, uint8_t, half)[SIZE*SIZE];\
put_h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(half, src, SIZE, stride);\
OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, src+stride, half, stride, stride, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc11_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc11_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
DECLARE_ALIGNED(16, uint8_t, halfH)[SIZE*SIZE];\
DECLARE_ALIGNED(16, uint8_t, halfV)[SIZE*SIZE];\
put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(halfH, src, SIZE, stride);\
@@ -109,7 +117,8 @@ static void OPNAME ## h264_qpel ## SIZE ## _mc11_ ## CODETYPE(uint8_t *dst, uint
OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfH, halfV, stride, SIZE, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc31_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc31_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
DECLARE_ALIGNED(16, uint8_t, halfH)[SIZE*SIZE];\
DECLARE_ALIGNED(16, uint8_t, halfV)[SIZE*SIZE];\
put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(halfH, src, SIZE, stride);\
@@ -117,7 +126,8 @@ static void OPNAME ## h264_qpel ## SIZE ## _mc31_ ## CODETYPE(uint8_t *dst, uint
OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfH, halfV, stride, SIZE, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc13_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc13_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
DECLARE_ALIGNED(16, uint8_t, halfH)[SIZE*SIZE];\
DECLARE_ALIGNED(16, uint8_t, halfV)[SIZE*SIZE];\
put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(halfH, src + stride, SIZE, stride);\
@@ -125,7 +135,8 @@ static void OPNAME ## h264_qpel ## SIZE ## _mc13_ ## CODETYPE(uint8_t *dst, uint
OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfH, halfV, stride, SIZE, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc33_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc33_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
DECLARE_ALIGNED(16, uint8_t, halfH)[SIZE*SIZE];\
DECLARE_ALIGNED(16, uint8_t, halfV)[SIZE*SIZE];\
put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(halfH, src + stride, SIZE, stride);\
@@ -133,12 +144,14 @@ static void OPNAME ## h264_qpel ## SIZE ## _mc33_ ## CODETYPE(uint8_t *dst, uint
OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfH, halfV, stride, SIZE, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc22_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc22_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
DECLARE_ALIGNED(16, int16_t, tmp)[SIZE*(SIZE+8)];\
OPNAME ## h264_qpel ## SIZE ## _hv_lowpass_ ## CODETYPE(dst, tmp, src, stride, SIZE, stride);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc21_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc21_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
DECLARE_ALIGNED(16, uint8_t, halfH)[SIZE*SIZE];\
DECLARE_ALIGNED(16, uint8_t, halfHV)[SIZE*SIZE];\
DECLARE_ALIGNED(16, int16_t, tmp)[SIZE*(SIZE+8)];\
@@ -147,7 +160,8 @@ static void OPNAME ## h264_qpel ## SIZE ## _mc21_ ## CODETYPE(uint8_t *dst, uint
OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfH, halfHV, stride, SIZE, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc23_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc23_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
DECLARE_ALIGNED(16, uint8_t, halfH)[SIZE*SIZE];\
DECLARE_ALIGNED(16, uint8_t, halfHV)[SIZE*SIZE];\
DECLARE_ALIGNED(16, int16_t, tmp)[SIZE*(SIZE+8)];\
@@ -156,7 +170,8 @@ static void OPNAME ## h264_qpel ## SIZE ## _mc23_ ## CODETYPE(uint8_t *dst, uint
OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfH, halfHV, stride, SIZE, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc12_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc12_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
DECLARE_ALIGNED(16, uint8_t, halfV)[SIZE*SIZE];\
DECLARE_ALIGNED(16, uint8_t, halfHV)[SIZE*SIZE];\
DECLARE_ALIGNED(16, int16_t, tmp)[SIZE*(SIZE+8)];\
@@ -165,7 +180,8 @@ static void OPNAME ## h264_qpel ## SIZE ## _mc12_ ## CODETYPE(uint8_t *dst, uint
OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfV, halfHV, stride, SIZE, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc32_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc32_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
DECLARE_ALIGNED(16, uint8_t, halfV)[SIZE*SIZE];\
DECLARE_ALIGNED(16, uint8_t, halfHV)[SIZE*SIZE];\
DECLARE_ALIGNED(16, int16_t, tmp)[SIZE*(SIZE+8)];\


+ 16
- 8
libavcodec/rv30dsp.c View File

@@ -210,35 +210,43 @@ static void OPNAME ## rv30_tpel16_hhvv_lowpass(uint8_t *dst, uint8_t *src, int d
\

#define RV30_MC(OPNAME, SIZE) \
static void OPNAME ## rv30_tpel ## SIZE ## _mc10_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## rv30_tpel ## SIZE ## _mc10_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## rv30_tpel ## SIZE ## _h_lowpass(dst, src, stride, stride, 12, 6);\
}\
\
static void OPNAME ## rv30_tpel ## SIZE ## _mc20_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## rv30_tpel ## SIZE ## _mc20_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## rv30_tpel ## SIZE ## _h_lowpass(dst, src, stride, stride, 6, 12);\
}\
\
static void OPNAME ## rv30_tpel ## SIZE ## _mc01_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## rv30_tpel ## SIZE ## _mc01_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## rv30_tpel ## SIZE ## _v_lowpass(dst, src, stride, stride, 12, 6);\
}\
\
static void OPNAME ## rv30_tpel ## SIZE ## _mc02_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## rv30_tpel ## SIZE ## _mc02_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## rv30_tpel ## SIZE ## _v_lowpass(dst, src, stride, stride, 6, 12);\
}\
\
static void OPNAME ## rv30_tpel ## SIZE ## _mc11_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## rv30_tpel ## SIZE ## _mc11_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## rv30_tpel ## SIZE ## _hv_lowpass(dst, src, stride, stride);\
}\
\
static void OPNAME ## rv30_tpel ## SIZE ## _mc12_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## rv30_tpel ## SIZE ## _mc12_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## rv30_tpel ## SIZE ## _hvv_lowpass(dst, src, stride, stride);\
}\
\
static void OPNAME ## rv30_tpel ## SIZE ## _mc21_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## rv30_tpel ## SIZE ## _mc21_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## rv30_tpel ## SIZE ## _hhv_lowpass(dst, src, stride, stride);\
}\
\
static void OPNAME ## rv30_tpel ## SIZE ## _mc22_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## rv30_tpel ## SIZE ## _mc22_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## rv30_tpel ## SIZE ## _hhvv_lowpass(dst, src, stride, stride);\
}\
\


+ 24
- 12
libavcodec/rv40dsp.c View File

@@ -104,72 +104,84 @@ static void OPNAME ## rv40_qpel16_h_lowpass(uint8_t *dst, uint8_t *src, int dstS
\

#define RV40_MC(OPNAME, SIZE) \
static void OPNAME ## rv40_qpel ## SIZE ## _mc10_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## rv40_qpel ## SIZE ## _mc10_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## rv40_qpel ## SIZE ## _h_lowpass(dst, src, stride, stride, SIZE, 52, 20, 6);\
}\
\
static void OPNAME ## rv40_qpel ## SIZE ## _mc30_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## rv40_qpel ## SIZE ## _mc30_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## rv40_qpel ## SIZE ## _h_lowpass(dst, src, stride, stride, SIZE, 20, 52, 6);\
}\
\
static void OPNAME ## rv40_qpel ## SIZE ## _mc01_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## rv40_qpel ## SIZE ## _mc01_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## rv40_qpel ## SIZE ## _v_lowpass(dst, src, stride, stride, SIZE, 52, 20, 6);\
}\
\
static void OPNAME ## rv40_qpel ## SIZE ## _mc11_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## rv40_qpel ## SIZE ## _mc11_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[SIZE*(SIZE+5)];\
uint8_t * const full_mid = full + SIZE*2;\
put_rv40_qpel ## SIZE ## _h_lowpass(full, src - 2*stride, SIZE, stride, SIZE+5, 52, 20, 6);\
OPNAME ## rv40_qpel ## SIZE ## _v_lowpass(dst, full_mid, stride, SIZE, SIZE, 52, 20, 6);\
}\
\
static void OPNAME ## rv40_qpel ## SIZE ## _mc21_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## rv40_qpel ## SIZE ## _mc21_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[SIZE*(SIZE+5)];\
uint8_t * const full_mid = full + SIZE*2;\
put_rv40_qpel ## SIZE ## _h_lowpass(full, src - 2*stride, SIZE, stride, SIZE+5, 20, 20, 5);\
OPNAME ## rv40_qpel ## SIZE ## _v_lowpass(dst, full_mid, stride, SIZE, SIZE, 52, 20, 6);\
}\
\
static void OPNAME ## rv40_qpel ## SIZE ## _mc31_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## rv40_qpel ## SIZE ## _mc31_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[SIZE*(SIZE+5)];\
uint8_t * const full_mid = full + SIZE*2;\
put_rv40_qpel ## SIZE ## _h_lowpass(full, src - 2*stride, SIZE, stride, SIZE+5, 20, 52, 6);\
OPNAME ## rv40_qpel ## SIZE ## _v_lowpass(dst, full_mid, stride, SIZE, SIZE, 52, 20, 6);\
}\
\
static void OPNAME ## rv40_qpel ## SIZE ## _mc12_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## rv40_qpel ## SIZE ## _mc12_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[SIZE*(SIZE+5)];\
uint8_t * const full_mid = full + SIZE*2;\
put_rv40_qpel ## SIZE ## _h_lowpass(full, src - 2*stride, SIZE, stride, SIZE+5, 52, 20, 6);\
OPNAME ## rv40_qpel ## SIZE ## _v_lowpass(dst, full_mid, stride, SIZE, SIZE, 20, 20, 5);\
}\
\
static void OPNAME ## rv40_qpel ## SIZE ## _mc22_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## rv40_qpel ## SIZE ## _mc22_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[SIZE*(SIZE+5)];\
uint8_t * const full_mid = full + SIZE*2;\
put_rv40_qpel ## SIZE ## _h_lowpass(full, src - 2*stride, SIZE, stride, SIZE+5, 20, 20, 5);\
OPNAME ## rv40_qpel ## SIZE ## _v_lowpass(dst, full_mid, stride, SIZE, SIZE, 20, 20, 5);\
}\
\
static void OPNAME ## rv40_qpel ## SIZE ## _mc32_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## rv40_qpel ## SIZE ## _mc32_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[SIZE*(SIZE+5)];\
uint8_t * const full_mid = full + SIZE*2;\
put_rv40_qpel ## SIZE ## _h_lowpass(full, src - 2*stride, SIZE, stride, SIZE+5, 20, 52, 6);\
OPNAME ## rv40_qpel ## SIZE ## _v_lowpass(dst, full_mid, stride, SIZE, SIZE, 20, 20, 5);\
}\
\
static void OPNAME ## rv40_qpel ## SIZE ## _mc03_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## rv40_qpel ## SIZE ## _mc03_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## rv40_qpel ## SIZE ## _v_lowpass(dst, src, stride, stride, SIZE, 20, 52, 6);\
}\
\
static void OPNAME ## rv40_qpel ## SIZE ## _mc13_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## rv40_qpel ## SIZE ## _mc13_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[SIZE*(SIZE+5)];\
uint8_t * const full_mid = full + SIZE*2;\
put_rv40_qpel ## SIZE ## _h_lowpass(full, src - 2*stride, SIZE, stride, SIZE+5, 52, 20, 6);\
OPNAME ## rv40_qpel ## SIZE ## _v_lowpass(dst, full_mid, stride, SIZE, SIZE, 20, 52, 6);\
}\
\
static void OPNAME ## rv40_qpel ## SIZE ## _mc23_c(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## rv40_qpel ## SIZE ## _mc23_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
uint8_t full[SIZE*(SIZE+5)];\
uint8_t * const full_mid = full + SIZE*2;\
put_rv40_qpel ## SIZE ## _h_lowpass(full, src - 2*stride, SIZE, stride, SIZE+5, 20, 20, 5);\


+ 11
- 5
libavcodec/vc1dsp.c View File

@@ -640,11 +640,17 @@ VC1_MSPEL_MC(op_avg, op4_avg, avg_)
/* pixel functions - really are entry points to vc1_mspel_mc */

#define PUT_VC1_MSPEL(a, b)\
static void put_vc1_mspel_mc ## a ## b ##_c(uint8_t *dst, const uint8_t *src, int stride, int rnd) { \
put_vc1_mspel_mc(dst, src, stride, a, b, rnd); \
}\
static void avg_vc1_mspel_mc ## a ## b ##_c(uint8_t *dst, const uint8_t *src, int stride, int rnd) { \
avg_vc1_mspel_mc(dst, src, stride, a, b, rnd); \
static void put_vc1_mspel_mc ## a ## b ##_c(uint8_t *dst, \
const uint8_t *src, \
ptrdiff_t stride, int rnd) \
{ \
put_vc1_mspel_mc(dst, src, stride, a, b, rnd); \
} \
static void avg_vc1_mspel_mc ## a ## b ##_c(uint8_t *dst, \
const uint8_t *src, \
ptrdiff_t stride, int rnd) \
{ \
avg_vc1_mspel_mc(dst, src, stride, a, b, rnd); \
}

PUT_VC1_MSPEL(1, 0)


+ 8
- 4
libavcodec/x86/cavsdsp.c View File

@@ -409,19 +409,23 @@ static void OPNAME ## cavs_qpel16_h_ ## MMX(uint8_t *dst, uint8_t *src, int dstS
}\

#define CAVS_MC(OPNAME, SIZE, MMX) \
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc20_ ## MMX(uint8_t *dst, uint8_t *src, int stride){\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc20_ ## MMX(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## cavs_qpel ## SIZE ## _h_ ## MMX(dst, src, stride, stride);\
}\
\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc01_ ## MMX(uint8_t *dst, uint8_t *src, int stride){\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc01_ ## MMX(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## cavs_qpel ## SIZE ## _v1_ ## MMX(dst, src, stride, stride);\
}\
\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc02_ ## MMX(uint8_t *dst, uint8_t *src, int stride){\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc02_ ## MMX(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## cavs_qpel ## SIZE ## _v2_ ## MMX(dst, src, stride, stride);\
}\
\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc03_ ## MMX(uint8_t *dst, uint8_t *src, int stride){\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc03_ ## MMX(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
OPNAME ## cavs_qpel ## SIZE ## _v3_ ## MMX(dst, src, stride, stride);\
}\



+ 41
- 41
libavcodec/x86/dsputil_mmx.c View File

@@ -655,13 +655,13 @@ static void draw_edges_mmx(uint8_t *buf, int wrap, int width, int height,
#if HAVE_YASM
#define QPEL_OP(OPNAME, ROUNDER, RND, MMX) \
static void OPNAME ## qpel8_mc00_ ## MMX (uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
ff_ ## OPNAME ## pixels8_ ## MMX(dst, src, stride, 8); \
} \
\
static void OPNAME ## qpel8_mc10_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
uint64_t temp[8]; \
uint8_t * const half = (uint8_t*)temp; \
@@ -672,14 +672,14 @@ static void OPNAME ## qpel8_mc10_ ## MMX(uint8_t *dst, uint8_t *src, \
} \
\
static void OPNAME ## qpel8_mc20_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
ff_ ## OPNAME ## mpeg4_qpel8_h_lowpass_ ## MMX(dst, src, stride, \
stride, 8); \
} \
\
static void OPNAME ## qpel8_mc30_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
uint64_t temp[8]; \
uint8_t * const half = (uint8_t*)temp; \
@@ -690,7 +690,7 @@ static void OPNAME ## qpel8_mc30_ ## MMX(uint8_t *dst, uint8_t *src, \
} \
\
static void OPNAME ## qpel8_mc01_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
uint64_t temp[8]; \
uint8_t * const half = (uint8_t*)temp; \
@@ -701,14 +701,14 @@ static void OPNAME ## qpel8_mc01_ ## MMX(uint8_t *dst, uint8_t *src, \
} \
\
static void OPNAME ## qpel8_mc02_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
ff_ ## OPNAME ## mpeg4_qpel8_v_lowpass_ ## MMX(dst, src, \
stride, stride); \
} \
\
static void OPNAME ## qpel8_mc03_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
uint64_t temp[8]; \
uint8_t * const half = (uint8_t*)temp; \
@@ -719,7 +719,7 @@ static void OPNAME ## qpel8_mc03_ ## MMX(uint8_t *dst, uint8_t *src, \
} \
\
static void OPNAME ## qpel8_mc11_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
uint64_t half[8 + 9]; \
uint8_t * const halfH = ((uint8_t*)half) + 64; \
@@ -734,7 +734,7 @@ static void OPNAME ## qpel8_mc11_ ## MMX(uint8_t *dst, uint8_t *src, \
} \
\
static void OPNAME ## qpel8_mc31_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
uint64_t half[8 + 9]; \
uint8_t * const halfH = ((uint8_t*)half) + 64; \
@@ -749,7 +749,7 @@ static void OPNAME ## qpel8_mc31_ ## MMX(uint8_t *dst, uint8_t *src, \
} \
\
static void OPNAME ## qpel8_mc13_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
uint64_t half[8 + 9]; \
uint8_t * const halfH = ((uint8_t*)half) + 64; \
@@ -764,7 +764,7 @@ static void OPNAME ## qpel8_mc13_ ## MMX(uint8_t *dst, uint8_t *src, \
} \
\
static void OPNAME ## qpel8_mc33_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
uint64_t half[8 + 9]; \
uint8_t * const halfH = ((uint8_t*)half) + 64; \
@@ -779,7 +779,7 @@ static void OPNAME ## qpel8_mc33_ ## MMX(uint8_t *dst, uint8_t *src, \
} \
\
static void OPNAME ## qpel8_mc21_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
uint64_t half[8 + 9]; \
uint8_t * const halfH = ((uint8_t*)half) + 64; \
@@ -792,7 +792,7 @@ static void OPNAME ## qpel8_mc21_ ## MMX(uint8_t *dst, uint8_t *src, \
} \
\
static void OPNAME ## qpel8_mc23_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
uint64_t half[8 + 9]; \
uint8_t * const halfH = ((uint8_t*)half) + 64; \
@@ -805,7 +805,7 @@ static void OPNAME ## qpel8_mc23_ ## MMX(uint8_t *dst, uint8_t *src, \
} \
\
static void OPNAME ## qpel8_mc12_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
uint64_t half[8 + 9]; \
uint8_t * const halfH = ((uint8_t*)half); \
@@ -818,7 +818,7 @@ static void OPNAME ## qpel8_mc12_ ## MMX(uint8_t *dst, uint8_t *src, \
} \
\
static void OPNAME ## qpel8_mc32_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
uint64_t half[8 + 9]; \
uint8_t * const halfH = ((uint8_t*)half); \
@@ -831,7 +831,7 @@ static void OPNAME ## qpel8_mc32_ ## MMX(uint8_t *dst, uint8_t *src, \
} \
\
static void OPNAME ## qpel8_mc22_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
uint64_t half[9]; \
uint8_t * const halfH = ((uint8_t*)half); \
@@ -842,13 +842,13 @@ static void OPNAME ## qpel8_mc22_ ## MMX(uint8_t *dst, uint8_t *src, \
} \
\
static void OPNAME ## qpel16_mc00_ ## MMX (uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
ff_ ## OPNAME ## pixels16_ ## MMX(dst, src, stride, 16); \
} \
\
static void OPNAME ## qpel16_mc10_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
uint64_t temp[32]; \
uint8_t * const half = (uint8_t*)temp; \
@@ -859,14 +859,14 @@ static void OPNAME ## qpel16_mc10_ ## MMX(uint8_t *dst, uint8_t *src, \
} \
\
static void OPNAME ## qpel16_mc20_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
ff_ ## OPNAME ## mpeg4_qpel16_h_lowpass_ ## MMX(dst, src, \
stride, stride, 16);\
} \
\
static void OPNAME ## qpel16_mc30_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
uint64_t temp[32]; \
uint8_t * const half = (uint8_t*)temp; \
@@ -877,7 +877,7 @@ static void OPNAME ## qpel16_mc30_ ## MMX(uint8_t *dst, uint8_t *src, \
} \
\
static void OPNAME ## qpel16_mc01_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
uint64_t temp[32]; \
uint8_t * const half = (uint8_t*)temp; \
@@ -888,14 +888,14 @@ static void OPNAME ## qpel16_mc01_ ## MMX(uint8_t *dst, uint8_t *src, \
} \
\
static void OPNAME ## qpel16_mc02_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
ff_ ## OPNAME ## mpeg4_qpel16_v_lowpass_ ## MMX(dst, src, \
stride, stride); \
} \
\
static void OPNAME ## qpel16_mc03_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
uint64_t temp[32]; \
uint8_t * const half = (uint8_t*)temp; \
@@ -906,7 +906,7 @@ static void OPNAME ## qpel16_mc03_ ## MMX(uint8_t *dst, uint8_t *src, \
} \
\
static void OPNAME ## qpel16_mc11_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
uint64_t half[16 * 2 + 17 * 2]; \
uint8_t * const halfH = ((uint8_t*)half) + 256; \
@@ -922,7 +922,7 @@ static void OPNAME ## qpel16_mc11_ ## MMX(uint8_t *dst, uint8_t *src, \
} \
\
static void OPNAME ## qpel16_mc31_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
uint64_t half[16 * 2 + 17 * 2]; \
uint8_t * const halfH = ((uint8_t*)half) + 256; \
@@ -938,7 +938,7 @@ static void OPNAME ## qpel16_mc31_ ## MMX(uint8_t *dst, uint8_t *src, \
} \
\
static void OPNAME ## qpel16_mc13_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
uint64_t half[16 * 2 + 17 * 2]; \
uint8_t * const halfH = ((uint8_t*)half) + 256; \
@@ -954,7 +954,7 @@ static void OPNAME ## qpel16_mc13_ ## MMX(uint8_t *dst, uint8_t *src, \
} \
\
static void OPNAME ## qpel16_mc33_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
uint64_t half[16 * 2 + 17 * 2]; \
uint8_t * const halfH = ((uint8_t*)half) + 256; \
@@ -970,7 +970,7 @@ static void OPNAME ## qpel16_mc33_ ## MMX(uint8_t *dst, uint8_t *src, \
} \
\
static void OPNAME ## qpel16_mc21_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
uint64_t half[16 * 2 + 17 * 2]; \
uint8_t * const halfH = ((uint8_t*)half) + 256; \
@@ -984,7 +984,7 @@ static void OPNAME ## qpel16_mc21_ ## MMX(uint8_t *dst, uint8_t *src, \
} \
\
static void OPNAME ## qpel16_mc23_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
uint64_t half[16 * 2 + 17 * 2]; \
uint8_t * const halfH = ((uint8_t*)half) + 256; \
@@ -998,7 +998,7 @@ static void OPNAME ## qpel16_mc23_ ## MMX(uint8_t *dst, uint8_t *src, \
} \
\
static void OPNAME ## qpel16_mc12_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
uint64_t half[17 * 2]; \
uint8_t * const halfH = ((uint8_t*)half); \
@@ -1011,7 +1011,7 @@ static void OPNAME ## qpel16_mc12_ ## MMX(uint8_t *dst, uint8_t *src, \
} \
\
static void OPNAME ## qpel16_mc32_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
uint64_t half[17 * 2]; \
uint8_t * const halfH = ((uint8_t*)half); \
@@ -1024,7 +1024,7 @@ static void OPNAME ## qpel16_mc32_ ## MMX(uint8_t *dst, uint8_t *src, \
} \
\
static void OPNAME ## qpel16_mc22_ ## MMX(uint8_t *dst, uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
uint64_t half[17 * 2]; \
uint8_t * const halfH = ((uint8_t*)half); \
@@ -1041,19 +1041,19 @@ QPEL_OP(put_no_rnd_, ff_pw_15, _no_rnd_, mmxext)


#if HAVE_INLINE_ASM
void ff_put_rv40_qpel8_mc33_mmx(uint8_t *dst, uint8_t *src, int stride)
void ff_put_rv40_qpel8_mc33_mmx(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
{
put_pixels8_xy2_mmx(dst, src, stride, 8);
}
void ff_put_rv40_qpel16_mc33_mmx(uint8_t *dst, uint8_t *src, int stride)
void ff_put_rv40_qpel16_mc33_mmx(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
{
put_pixels16_xy2_mmx(dst, src, stride, 16);
}
void ff_avg_rv40_qpel8_mc33_mmx(uint8_t *dst, uint8_t *src, int stride)
void ff_avg_rv40_qpel8_mc33_mmx(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
{
avg_pixels8_xy2_mmx(dst, src, stride, 8);
}
void ff_avg_rv40_qpel16_mc33_mmx(uint8_t *dst, uint8_t *src, int stride)
void ff_avg_rv40_qpel16_mc33_mmx(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
{
avg_pixels16_xy2_mmx(dst, src, stride, 16);
}
@@ -1230,29 +1230,29 @@ void ff_avg_pixels16_sse2(uint8_t *block, const uint8_t *pixels,
#if HAVE_INLINE_ASM

/* CAVS-specific */
void ff_put_cavs_qpel8_mc00_mmxext(uint8_t *dst, uint8_t *src, int stride)
void ff_put_cavs_qpel8_mc00_mmxext(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
{
put_pixels8_mmx(dst, src, stride, 8);
}

void ff_avg_cavs_qpel8_mc00_mmxext(uint8_t *dst, uint8_t *src, int stride)
void ff_avg_cavs_qpel8_mc00_mmxext(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
{
avg_pixels8_mmx(dst, src, stride, 8);
}

void ff_put_cavs_qpel16_mc00_mmxext(uint8_t *dst, uint8_t *src, int stride)
void ff_put_cavs_qpel16_mc00_mmxext(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
{
put_pixels16_mmx(dst, src, stride, 16);
}

void ff_avg_cavs_qpel16_mc00_mmxext(uint8_t *dst, uint8_t *src, int stride)
void ff_avg_cavs_qpel16_mc00_mmxext(uint8_t *dst, uint8_t *src, ptrdiff_t stride)
{
avg_pixels16_mmx(dst, src, stride, 16);
}

/* VC-1-specific */
void ff_put_vc1_mspel_mc00_mmx(uint8_t *dst, const uint8_t *src,
int stride, int rnd)
ptrdiff_t stride, int rnd)
{
put_pixels8_mmx(dst, src, stride, 8);
}


+ 9
- 9
libavcodec/x86/dsputil_mmx.h View File

@@ -82,17 +82,17 @@ void ff_put_signed_pixels_clamped_mmx(const int16_t *block, uint8_t *pixels, int
void ff_avg_pixels8_mmxext(uint8_t *block, const uint8_t *pixels,
ptrdiff_t line_size, int h);

void ff_put_cavs_qpel8_mc00_mmxext(uint8_t *dst, uint8_t *src, int stride);
void ff_avg_cavs_qpel8_mc00_mmxext(uint8_t *dst, uint8_t *src, int stride);
void ff_put_cavs_qpel16_mc00_mmxext(uint8_t *dst, uint8_t *src, int stride);
void ff_avg_cavs_qpel16_mc00_mmxext(uint8_t *dst, uint8_t *src, int stride);
void ff_put_cavs_qpel8_mc00_mmxext(uint8_t *dst, uint8_t *src, ptrdiff_t stride);
void ff_avg_cavs_qpel8_mc00_mmxext(uint8_t *dst, uint8_t *src, ptrdiff_t stride);
void ff_put_cavs_qpel16_mc00_mmxext(uint8_t *dst, uint8_t *src, ptrdiff_t stride);
void ff_avg_cavs_qpel16_mc00_mmxext(uint8_t *dst, uint8_t *src, ptrdiff_t stride);

void ff_put_vc1_mspel_mc00_mmx(uint8_t *dst, const uint8_t *src, int stride, int rnd);
void ff_put_vc1_mspel_mc00_mmx(uint8_t *dst, const uint8_t *src, ptrdiff_t stride, int rnd);

void ff_put_rv40_qpel8_mc33_mmx(uint8_t *block, uint8_t *pixels, int line_size);
void ff_put_rv40_qpel16_mc33_mmx(uint8_t *block, uint8_t *pixels, int line_size);
void ff_avg_rv40_qpel8_mc33_mmx(uint8_t *block, uint8_t *pixels, int line_size);
void ff_avg_rv40_qpel16_mc33_mmx(uint8_t *block, uint8_t *pixels, int line_size);
void ff_put_rv40_qpel8_mc33_mmx(uint8_t *block, uint8_t *pixels, ptrdiff_t stride);
void ff_put_rv40_qpel16_mc33_mmx(uint8_t *block, uint8_t *pixels, ptrdiff_t stride);
void ff_avg_rv40_qpel8_mc33_mmx(uint8_t *block, uint8_t *pixels, ptrdiff_t stride);
void ff_avg_rv40_qpel16_mc33_mmx(uint8_t *block, uint8_t *pixels, ptrdiff_t stride);

void ff_mmx_idct(int16_t *block);
void ff_mmxext_idct(int16_t *block);


+ 39
- 19
libavcodec/x86/h264_qpel.c View File

@@ -250,81 +250,98 @@ H264_MC_V(OPNAME, SIZE, MMX, ALIGN)\
H264_MC_H(OPNAME, SIZE, MMX, ALIGN)\
H264_MC_HV(OPNAME, SIZE, MMX, ALIGN)\

static void put_h264_qpel16_mc00_sse2 (uint8_t *dst, uint8_t *src, int stride){
static void put_h264_qpel16_mc00_sse2 (uint8_t *dst, uint8_t *src,
ptrdiff_t stride)
{
ff_put_pixels16_sse2(dst, src, stride, 16);
}
static void avg_h264_qpel16_mc00_sse2 (uint8_t *dst, uint8_t *src, int stride){
static void avg_h264_qpel16_mc00_sse2 (uint8_t *dst, uint8_t *src,
ptrdiff_t stride)
{
ff_avg_pixels16_sse2(dst, src, stride, 16);
}
#define put_h264_qpel8_mc00_sse2 put_h264_qpel8_mc00_mmxext
#define avg_h264_qpel8_mc00_sse2 avg_h264_qpel8_mc00_mmxext

#define H264_MC_C(OPNAME, SIZE, MMX, ALIGN) \
static void OPNAME ## h264_qpel ## SIZE ## _mc00_ ## MMX (uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc00_ ## MMX (uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
ff_ ## OPNAME ## pixels ## SIZE ## _ ## MMX(dst, src, stride, SIZE);\
}\

#define H264_MC_H(OPNAME, SIZE, MMX, ALIGN) \
static void OPNAME ## h264_qpel ## SIZE ## _mc10_ ## MMX(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc10_ ## MMX(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
ff_ ## OPNAME ## h264_qpel ## SIZE ## _h_lowpass_l2_ ## MMX(dst, src, src, stride, stride);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc20_ ## MMX(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc20_ ## MMX(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
ff_ ## OPNAME ## h264_qpel ## SIZE ## _h_lowpass_ ## MMX(dst, src, stride, stride);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc30_ ## MMX(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc30_ ## MMX(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
ff_ ## OPNAME ## h264_qpel ## SIZE ## _h_lowpass_l2_ ## MMX(dst, src, src+1, stride, stride);\
}\

#define H264_MC_V(OPNAME, SIZE, MMX, ALIGN) \
static void OPNAME ## h264_qpel ## SIZE ## _mc01_ ## MMX(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc01_ ## MMX(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
DECLARE_ALIGNED(ALIGN, uint8_t, temp)[SIZE*SIZE];\
ff_put_h264_qpel ## SIZE ## _v_lowpass_ ## MMX(temp, src, SIZE, stride);\
ff_ ## OPNAME ## pixels ## SIZE ## _l2_ ## MMX(dst, src, temp, stride, stride, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc02_ ## MMX(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc02_ ## MMX(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
ff_ ## OPNAME ## h264_qpel ## SIZE ## _v_lowpass_ ## MMX(dst, src, stride, stride);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc03_ ## MMX(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc03_ ## MMX(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
DECLARE_ALIGNED(ALIGN, uint8_t, temp)[SIZE*SIZE];\
ff_put_h264_qpel ## SIZE ## _v_lowpass_ ## MMX(temp, src, SIZE, stride);\
ff_ ## OPNAME ## pixels ## SIZE ## _l2_ ## MMX(dst, src+stride, temp, stride, stride, SIZE);\
}\

#define H264_MC_HV(OPNAME, SIZE, MMX, ALIGN) \
static void OPNAME ## h264_qpel ## SIZE ## _mc11_ ## MMX(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc11_ ## MMX(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
DECLARE_ALIGNED(ALIGN, uint8_t, temp)[SIZE*SIZE];\
ff_put_h264_qpel ## SIZE ## _v_lowpass_ ## MMX(temp, src, SIZE, stride);\
ff_ ## OPNAME ## h264_qpel ## SIZE ## _h_lowpass_l2_ ## MMX(dst, src, temp, stride, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc31_ ## MMX(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc31_ ## MMX(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
DECLARE_ALIGNED(ALIGN, uint8_t, temp)[SIZE*SIZE];\
ff_put_h264_qpel ## SIZE ## _v_lowpass_ ## MMX(temp, src+1, SIZE, stride);\
ff_ ## OPNAME ## h264_qpel ## SIZE ## _h_lowpass_l2_ ## MMX(dst, src, temp, stride, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc13_ ## MMX(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc13_ ## MMX(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
DECLARE_ALIGNED(ALIGN, uint8_t, temp)[SIZE*SIZE];\
ff_put_h264_qpel ## SIZE ## _v_lowpass_ ## MMX(temp, src, SIZE, stride);\
ff_ ## OPNAME ## h264_qpel ## SIZE ## _h_lowpass_l2_ ## MMX(dst, src+stride, temp, stride, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc33_ ## MMX(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc33_ ## MMX(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
DECLARE_ALIGNED(ALIGN, uint8_t, temp)[SIZE*SIZE];\
ff_put_h264_qpel ## SIZE ## _v_lowpass_ ## MMX(temp, src+1, SIZE, stride);\
ff_ ## OPNAME ## h264_qpel ## SIZE ## _h_lowpass_l2_ ## MMX(dst, src+stride, temp, stride, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc22_ ## MMX(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc22_ ## MMX(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
DECLARE_ALIGNED(ALIGN, uint16_t, temp)[SIZE*(SIZE<8?12:24)];\
ff_ ## OPNAME ## h264_qpel ## SIZE ## _hv_lowpass_ ## MMX(dst, temp, src, stride, SIZE, stride);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc21_ ## MMX(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc21_ ## MMX(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
DECLARE_ALIGNED(ALIGN, uint8_t, temp)[SIZE*(SIZE<8?12:24)*2 + SIZE*SIZE];\
uint8_t * const halfHV= temp;\
int16_t * const halfV= (int16_t*)(temp + SIZE*SIZE);\
@@ -333,7 +350,8 @@ static void OPNAME ## h264_qpel ## SIZE ## _mc21_ ## MMX(uint8_t *dst, uint8_t *
ff_ ## OPNAME ## h264_qpel ## SIZE ## _h_lowpass_l2_ ## MMX(dst, src, halfHV, stride, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc23_ ## MMX(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc23_ ## MMX(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
DECLARE_ALIGNED(ALIGN, uint8_t, temp)[SIZE*(SIZE<8?12:24)*2 + SIZE*SIZE];\
uint8_t * const halfHV= temp;\
int16_t * const halfV= (int16_t*)(temp + SIZE*SIZE);\
@@ -342,7 +360,8 @@ static void OPNAME ## h264_qpel ## SIZE ## _mc23_ ## MMX(uint8_t *dst, uint8_t *
ff_ ## OPNAME ## h264_qpel ## SIZE ## _h_lowpass_l2_ ## MMX(dst, src+stride, halfHV, stride, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc12_ ## MMX(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc12_ ## MMX(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
DECLARE_ALIGNED(ALIGN, uint8_t, temp)[SIZE*(SIZE<8?12:24)*2 + SIZE*SIZE];\
uint8_t * const halfHV= temp;\
int16_t * const halfV= (int16_t*)(temp + SIZE*SIZE);\
@@ -351,7 +370,8 @@ static void OPNAME ## h264_qpel ## SIZE ## _mc12_ ## MMX(uint8_t *dst, uint8_t *
ff_ ## OPNAME ## pixels ## SIZE ## _l2_shift5_mmxext(dst, halfV+2, halfHV, stride, SIZE, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc32_ ## MMX(uint8_t *dst, uint8_t *src, int stride){\
static void OPNAME ## h264_qpel ## SIZE ## _mc32_ ## MMX(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
DECLARE_ALIGNED(ALIGN, uint8_t, temp)[SIZE*(SIZE<8?12:24)*2 + SIZE*SIZE];\
uint8_t * const halfHV= temp;\
int16_t * const halfV= (int16_t*)(temp + SIZE*SIZE);\
@@ -398,7 +418,7 @@ H264_MC_816(H264_MC_HV, ssse3)
//10bit
#define LUMA_MC_OP(OP, NUM, DEPTH, TYPE, OPT) \
void ff_ ## OP ## _h264_qpel ## NUM ## _ ## TYPE ## _ ## DEPTH ## _ ## OPT \
(uint8_t *dst, uint8_t *src, int stride);
(uint8_t *dst, uint8_t *src, ptrdiff_t stride);

#define LUMA_MC_ALL(DEPTH, TYPE, OPT) \
LUMA_MC_OP(put, 4, DEPTH, TYPE, OPT) \


+ 1
- 1
libavcodec/x86/rv40dsp_init.c View File

@@ -71,7 +71,7 @@ DECLARE_WEIGHT(ssse3)
#define QPEL_FUNC_DECL(OP, SIZE, PH, PV, OPT) \
static void OP ## rv40_qpel ##SIZE ##_mc ##PH ##PV ##OPT(uint8_t *dst, \
uint8_t *src, \
int stride) \
ptrdiff_t stride) \
{ \
int i; \
if (PH && PV) { \


+ 1
- 1
libavcodec/x86/vc1dsp_init.c View File

@@ -63,7 +63,7 @@ static void vc1_h_loop_filter16_sse4(uint8_t *src, int stride, int pq)
}

static void avg_vc1_mspel_mc00_mmxext(uint8_t *dst, const uint8_t *src,
int stride, int rnd)
ptrdiff_t stride, int rnd)
{
ff_avg_pixels8_mmxext(dst, src, stride, 8);
}


+ 7
- 2
libavcodec/x86/vc1dsp_mmx.c View File

@@ -462,12 +462,17 @@ VC1_MSPEL_MC(avg_)

/** Macro to ease bicubic filter interpolation functions declarations */
#define DECLARE_FUNCTION(a, b) \
static void put_vc1_mspel_mc ## a ## b ## _mmx(uint8_t *dst, const uint8_t *src, int stride, int rnd) { \
static void put_vc1_mspel_mc ## a ## b ## _mmx(uint8_t *dst, \
const uint8_t *src, \
ptrdiff_t stride, \
int rnd) \
{ \
put_vc1_mspel_mc(dst, src, stride, a, b, rnd); \
}\
static void avg_vc1_mspel_mc ## a ## b ## _mmxext(uint8_t *dst, \
const uint8_t *src, \
int stride, int rnd) \
ptrdiff_t stride, \
int rnd) \
{ \
avg_vc1_mspel_mc(dst, src, stride, a, b, rnd); \
}


Write Preview
Loading…
Cancel
Save