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

vp8: change int stride to ptrdiff_t stride. 

On 64bit platforms with 32bit int, this means we won't have to sign-
extend the integer anymore.
tags/n0.11
Ronald S. Bultje 13 years ago
parent
349b7977e4
commit
bd66f073fe
5 changed files with 178 additions and 148 deletions
Split View
Diff Options
Show Stats Download Patch File Download Diff File
  1. +16
    -16
      libavcodec/arm/vp8dsp_init_arm.c
  2. +8
    -8
      libavcodec/ppc/vp8dsp_altivec.c
  3. +22
    -22
      libavcodec/vp8dsp.c
  4. +26
    -19
      libavcodec/vp8dsp.h
  5. +106
    -83
      libavcodec/x86/vp8dsp-init.c

+ 16
- 16
libavcodec/arm/vp8dsp_init_arm.c View File

@@ -22,38 +22,38 @@
void ff_vp8_luma_dc_wht_neon(DCTELEM block[4][4][16], DCTELEM dc[16]);
void ff_vp8_luma_dc_wht_dc_neon(DCTELEM block[4][4][16], DCTELEM dc[16]);

void ff_vp8_idct_add_neon(uint8_t *dst, DCTELEM block[16], int stride);
void ff_vp8_idct_dc_add_neon(uint8_t *dst, DCTELEM block[16], int stride);
void ff_vp8_idct_dc_add4y_neon(uint8_t *dst, DCTELEM block[4][16], int stride);
void ff_vp8_idct_dc_add4uv_neon(uint8_t *dst, DCTELEM block[4][16], int stride);
void ff_vp8_idct_add_neon(uint8_t *dst, DCTELEM block[16], ptrdiff_t stride);
void ff_vp8_idct_dc_add_neon(uint8_t *dst, DCTELEM block[16], ptrdiff_t stride);
void ff_vp8_idct_dc_add4y_neon(uint8_t *dst, DCTELEM block[4][16], ptrdiff_t stride);
void ff_vp8_idct_dc_add4uv_neon(uint8_t *dst, DCTELEM block[4][16], ptrdiff_t stride);

void ff_vp8_v_loop_filter16_neon(uint8_t *dst, int stride,
void ff_vp8_v_loop_filter16_neon(uint8_t *dst, ptrdiff_t stride,
int flim_E, int flim_I, int hev_thresh);
void ff_vp8_h_loop_filter16_neon(uint8_t *dst, int stride,
void ff_vp8_h_loop_filter16_neon(uint8_t *dst, ptrdiff_t stride,
int flim_E, int flim_I, int hev_thresh);
void ff_vp8_v_loop_filter8uv_neon(uint8_t *dstU, uint8_t *dstV, int stride,
void ff_vp8_v_loop_filter8uv_neon(uint8_t *dstU, uint8_t *dstV, ptrdiff_t stride,
int flim_E, int flim_I, int hev_thresh);
void ff_vp8_h_loop_filter8uv_neon(uint8_t *dstU, uint8_t *dstV, int stride,
void ff_vp8_h_loop_filter8uv_neon(uint8_t *dstU, uint8_t *dstV, ptrdiff_t stride,
int flim_E, int flim_I, int hev_thresh);

void ff_vp8_v_loop_filter16_inner_neon(uint8_t *dst, int stride,
void ff_vp8_v_loop_filter16_inner_neon(uint8_t *dst, ptrdiff_t stride,
int flim_E, int flim_I, int hev_thresh);
void ff_vp8_h_loop_filter16_inner_neon(uint8_t *dst, int stride,
void ff_vp8_h_loop_filter16_inner_neon(uint8_t *dst, ptrdiff_t stride,
int flim_E, int flim_I, int hev_thresh);
void ff_vp8_v_loop_filter8uv_inner_neon(uint8_t *dstU, uint8_t *dstV,
int stride, int flim_E, int flim_I,
ptrdiff_t stride, int flim_E, int flim_I,
int hev_thresh);
void ff_vp8_h_loop_filter8uv_inner_neon(uint8_t *dstU, uint8_t *dstV,
int stride, int flim_E, int flim_I,
ptrdiff_t stride, int flim_E, int flim_I,
int hev_thresh);

void ff_vp8_v_loop_filter16_simple_neon(uint8_t *dst, int stride, int flim);
void ff_vp8_h_loop_filter16_simple_neon(uint8_t *dst, int stride, int flim);
void ff_vp8_v_loop_filter16_simple_neon(uint8_t *dst, ptrdiff_t stride, int flim);
void ff_vp8_h_loop_filter16_simple_neon(uint8_t *dst, ptrdiff_t stride, int flim);


#define VP8_MC(n) \
void ff_put_vp8_##n##_neon(uint8_t *dst, int dststride, \
uint8_t *src, int srcstride, \
void ff_put_vp8_##n##_neon(uint8_t *dst, ptrdiff_t dststride, \
uint8_t *src, ptrdiff_t srcstride, \
int h, int x, int y)

#define VP8_EPEL(w) \


+ 8
- 8
libavcodec/ppc/vp8dsp_altivec.c View File

@@ -75,8 +75,8 @@ static const vec_s8 h_subpel_filters_outer[3] =
dstv = vec_sra(dstv, c7)

static av_always_inline
void put_vp8_epel_h_altivec_core(uint8_t *dst, int dst_stride,
uint8_t *src, int src_stride,
void put_vp8_epel_h_altivec_core(uint8_t *dst, ptrdiff_t dst_stride,
uint8_t *src, ptrdiff_t src_stride,
int h, int mx, int w, int is6tap)
{
LOAD_H_SUBPEL_FILTER(mx-1);
@@ -161,8 +161,8 @@ static const vec_u8 v_subpel_filters[7] =
dstv = vec_sra(dstv, c7)

static av_always_inline
void put_vp8_epel_v_altivec_core(uint8_t *dst, int dst_stride,
uint8_t *src, int src_stride,
void put_vp8_epel_v_altivec_core(uint8_t *dst, ptrdiff_t dst_stride,
uint8_t *src, ptrdiff_t src_stride,
int h, int my, int w, int is6tap)
{
LOAD_V_SUBPEL_FILTER(my-1);
@@ -226,19 +226,19 @@ void put_vp8_epel_v_altivec_core(uint8_t *dst, int dst_stride,

#define EPEL_FUNCS(WIDTH, TAPS) \
static av_noinline \
void put_vp8_epel ## WIDTH ## _h ## TAPS ## _altivec(uint8_t *dst, int dst_stride, uint8_t *src, int src_stride, int h, int mx, int my) \
void put_vp8_epel ## WIDTH ## _h ## TAPS ## _altivec(uint8_t *dst, ptrdiff_t dst_stride, uint8_t *src, ptrdiff_t src_stride, int h, int mx, int my) \
{ \
put_vp8_epel_h_altivec_core(dst, dst_stride, src, src_stride, h, mx, WIDTH, TAPS == 6); \
} \
\
static av_noinline \
void put_vp8_epel ## WIDTH ## _v ## TAPS ## _altivec(uint8_t *dst, int dst_stride, uint8_t *src, int src_stride, int h, int mx, int my) \
void put_vp8_epel ## WIDTH ## _v ## TAPS ## _altivec(uint8_t *dst, ptrdiff_t dst_stride, uint8_t *src, ptrdiff_t src_stride, int h, int mx, int my) \
{ \
put_vp8_epel_v_altivec_core(dst, dst_stride, src, src_stride, h, my, WIDTH, TAPS == 6); \
}

#define EPEL_HV(WIDTH, HTAPS, VTAPS) \
static void put_vp8_epel ## WIDTH ## _h ## HTAPS ## v ## VTAPS ## _altivec(uint8_t *dst, int stride, uint8_t *src, int s, int h, int mx, int my) \
static void put_vp8_epel ## WIDTH ## _h ## HTAPS ## v ## VTAPS ## _altivec(uint8_t *dst, ptrdiff_t stride, uint8_t *src, ptrdiff_t s, int h, int mx, int my) \
{ \
DECLARE_ALIGNED(16, uint8_t, tmp)[(2*WIDTH+5)*16]; \
if (VTAPS == 6) { \
@@ -266,7 +266,7 @@ EPEL_HV(4, 4,6)
EPEL_HV(4, 6,4)
EPEL_HV(4, 4,4)

static void put_vp8_pixels16_altivec(uint8_t *dst, int stride, uint8_t *src, int s, int h, int mx, int my)
static void put_vp8_pixels16_altivec(uint8_t *dst, ptrdiff_t stride, uint8_t *src, ptrdiff_t s, int h, int mx, int my)
{
ff_put_pixels16_altivec(dst, src, stride, h);
}


+ 22
- 22
libavcodec/vp8dsp.c View File

@@ -77,7 +77,7 @@ static void vp8_luma_dc_wht_dc_c(DCTELEM block[4][4][16], DCTELEM dc[16])
#define MUL_20091(a) ((((a)*20091) >> 16) + (a))
#define MUL_35468(a) (((a)*35468) >> 16)

static void vp8_idct_add_c(uint8_t *dst, DCTELEM block[16], int stride)
static void vp8_idct_add_c(uint8_t *dst, DCTELEM block[16], ptrdiff_t stride)
{
int i, t0, t1, t2, t3;
uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
@@ -113,7 +113,7 @@ static void vp8_idct_add_c(uint8_t *dst, DCTELEM block[16], int stride)
}
}

static void vp8_idct_dc_add_c(uint8_t *dst, DCTELEM block[16], int stride)
static void vp8_idct_dc_add_c(uint8_t *dst, DCTELEM block[16], ptrdiff_t stride)
{
int i, dc = (block[0] + 4) >> 3;
uint8_t *cm = ff_cropTbl + MAX_NEG_CROP + dc;
@@ -128,7 +128,7 @@ static void vp8_idct_dc_add_c(uint8_t *dst, DCTELEM block[16], int stride)
}
}

static void vp8_idct_dc_add4uv_c(uint8_t *dst, DCTELEM block[4][16], int stride)
static void vp8_idct_dc_add4uv_c(uint8_t *dst, DCTELEM block[4][16], ptrdiff_t stride)
{
vp8_idct_dc_add_c(dst+stride*0+0, block[0], stride);
vp8_idct_dc_add_c(dst+stride*0+4, block[1], stride);
@@ -136,7 +136,7 @@ static void vp8_idct_dc_add4uv_c(uint8_t *dst, DCTELEM block[4][16], int stride)
vp8_idct_dc_add_c(dst+stride*4+4, block[3], stride);
}

static void vp8_idct_dc_add4y_c(uint8_t *dst, DCTELEM block[4][16], int stride)
static void vp8_idct_dc_add4y_c(uint8_t *dst, DCTELEM block[4][16], ptrdiff_t stride)
{
vp8_idct_dc_add_c(dst+ 0, block[0], stride);
vp8_idct_dc_add_c(dst+ 4, block[1], stride);
@@ -157,7 +157,7 @@ static void vp8_idct_dc_add4y_c(uint8_t *dst, DCTELEM block[4][16], int stride)

#define clip_int8(n) (cm[n+0x80]-0x80)

static av_always_inline void filter_common(uint8_t *p, int stride, int is4tap)
static av_always_inline void filter_common(uint8_t *p, ptrdiff_t stride, int is4tap)
{
LOAD_PIXELS
int a, f1, f2;
@@ -188,7 +188,7 @@ static av_always_inline void filter_common(uint8_t *p, int stride, int is4tap)
}
}

static av_always_inline int simple_limit(uint8_t *p, int stride, int flim)
static av_always_inline int simple_limit(uint8_t *p, ptrdiff_t stride, int flim)
{
LOAD_PIXELS
return 2*FFABS(p0-q0) + (FFABS(p1-q1) >> 1) <= flim;
@@ -198,7 +198,7 @@ static av_always_inline int simple_limit(uint8_t *p, int stride, int flim)
* E - limit at the macroblock edge
* I - limit for interior difference
*/
static av_always_inline int normal_limit(uint8_t *p, int stride, int E, int I)
static av_always_inline int normal_limit(uint8_t *p, ptrdiff_t stride, int E, int I)
{
LOAD_PIXELS
return simple_limit(p, stride, E)
@@ -207,13 +207,13 @@ static av_always_inline int normal_limit(uint8_t *p, int stride, int E, int I)
}

// high edge variance
static av_always_inline int hev(uint8_t *p, int stride, int thresh)
static av_always_inline int hev(uint8_t *p, ptrdiff_t stride, int thresh)
{
LOAD_PIXELS
return FFABS(p1-p0) > thresh || FFABS(q1-q0) > thresh;
}

static av_always_inline void filter_mbedge(uint8_t *p, int stride)
static av_always_inline void filter_mbedge(uint8_t *p, ptrdiff_t stride)
{
int a0, a1, a2, w;
uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
@@ -236,7 +236,7 @@ static av_always_inline void filter_mbedge(uint8_t *p, int stride)
}

#define LOOP_FILTER(dir, size, stridea, strideb, maybe_inline) \
static maybe_inline void vp8_ ## dir ## _loop_filter ## size ## _c(uint8_t *dst, int stride,\
static maybe_inline void vp8_ ## dir ## _loop_filter ## size ## _c(uint8_t *dst, ptrdiff_t stride,\
int flim_E, int flim_I, int hev_thresh)\
{\
int i;\
@@ -250,7 +250,7 @@ static maybe_inline void vp8_ ## dir ## _loop_filter ## size ## _c(uint8_t *dst,
}\
}\
\
static maybe_inline void vp8_ ## dir ## _loop_filter ## size ## _inner_c(uint8_t *dst, int stride,\
static maybe_inline void vp8_ ## dir ## _loop_filter ## size ## _inner_c(uint8_t *dst, ptrdiff_t stride,\
int flim_E, int flim_I, int hev_thresh)\
{\
int i;\
@@ -270,13 +270,13 @@ LOOP_FILTER(h, 16, stride, 1,)

#define UV_LOOP_FILTER(dir, stridea, strideb) \
LOOP_FILTER(dir, 8, stridea, strideb, av_always_inline) \
static void vp8_ ## dir ## _loop_filter8uv_c(uint8_t *dstU, uint8_t *dstV, int stride,\
static void vp8_ ## dir ## _loop_filter8uv_c(uint8_t *dstU, uint8_t *dstV, ptrdiff_t stride,\
int fE, int fI, int hev_thresh)\
{\
vp8_ ## dir ## _loop_filter8_c(dstU, stride, fE, fI, hev_thresh);\
vp8_ ## dir ## _loop_filter8_c(dstV, stride, fE, fI, hev_thresh);\
}\
static void vp8_ ## dir ## _loop_filter8uv_inner_c(uint8_t *dstU, uint8_t *dstV, int stride,\
static void vp8_ ## dir ## _loop_filter8uv_inner_c(uint8_t *dstU, uint8_t *dstV, ptrdiff_t stride,\
int fE, int fI, int hev_thresh)\
{\
vp8_ ## dir ## _loop_filter8_inner_c(dstU, stride, fE, fI, hev_thresh);\
@@ -286,7 +286,7 @@ static void vp8_ ## dir ## _loop_filter8uv_inner_c(uint8_t *dstU, uint8_t *dstV,
UV_LOOP_FILTER(v, 1, stride)
UV_LOOP_FILTER(h, stride, 1)

static void vp8_v_loop_filter_simple_c(uint8_t *dst, int stride, int flim)
static void vp8_v_loop_filter_simple_c(uint8_t *dst, ptrdiff_t stride, int flim)
{
int i;

@@ -295,7 +295,7 @@ static void vp8_v_loop_filter_simple_c(uint8_t *dst, int stride, int flim)
filter_common(dst+i, stride, 1);
}

static void vp8_h_loop_filter_simple_c(uint8_t *dst, int stride, int flim)
static void vp8_h_loop_filter_simple_c(uint8_t *dst, ptrdiff_t stride, int flim)
{
int i;

@@ -315,7 +315,7 @@ static const uint8_t subpel_filters[7][6] = {
};

#define PUT_PIXELS(WIDTH) \
static void put_vp8_pixels ## WIDTH ##_c(uint8_t *dst, int dststride, uint8_t *src, int srcstride, int h, int x, int y) { \
static void put_vp8_pixels ## WIDTH ##_c(uint8_t *dst, ptrdiff_t dststride, uint8_t *src, ptrdiff_t srcstride, int h, int x, int y) { \
int i; \
for (i = 0; i < h; i++, dst+= dststride, src+= srcstride) { \
memcpy(dst, src, WIDTH); \
@@ -335,7 +335,7 @@ PUT_PIXELS(4)
F[3]*src[x+1*stride] - F[4]*src[x+2*stride] + 64) >> 7]

#define VP8_EPEL_H(SIZE, TAPS) \
static void put_vp8_epel ## SIZE ## _h ## TAPS ## _c(uint8_t *dst, int dststride, uint8_t *src, int srcstride, int h, int mx, int my) \
static void put_vp8_epel ## SIZE ## _h ## TAPS ## _c(uint8_t *dst, ptrdiff_t dststride, uint8_t *src, ptrdiff_t srcstride, int h, int mx, int my) \
{ \
const uint8_t *filter = subpel_filters[mx-1]; \
uint8_t *cm = ff_cropTbl + MAX_NEG_CROP; \
@@ -349,7 +349,7 @@ static void put_vp8_epel ## SIZE ## _h ## TAPS ## _c(uint8_t *dst, int dststride
} \
}
#define VP8_EPEL_V(SIZE, TAPS) \
static void put_vp8_epel ## SIZE ## _v ## TAPS ## _c(uint8_t *dst, int dststride, uint8_t *src, int srcstride, int h, int mx, int my) \
static void put_vp8_epel ## SIZE ## _v ## TAPS ## _c(uint8_t *dst, ptrdiff_t dststride, uint8_t *src, ptrdiff_t srcstride, int h, int mx, int my) \
{ \
const uint8_t *filter = subpel_filters[my-1]; \
uint8_t *cm = ff_cropTbl + MAX_NEG_CROP; \
@@ -363,7 +363,7 @@ static void put_vp8_epel ## SIZE ## _v ## TAPS ## _c(uint8_t *dst, int dststride
} \
}
#define VP8_EPEL_HV(SIZE, HTAPS, VTAPS) \
static void put_vp8_epel ## SIZE ## _h ## HTAPS ## v ## VTAPS ## _c(uint8_t *dst, int dststride, uint8_t *src, int srcstride, int h, int mx, int my) \
static void put_vp8_epel ## SIZE ## _h ## HTAPS ## v ## VTAPS ## _c(uint8_t *dst, ptrdiff_t dststride, uint8_t *src, ptrdiff_t srcstride, int h, int mx, int my) \
{ \
const uint8_t *filter = subpel_filters[mx-1]; \
uint8_t *cm = ff_cropTbl + MAX_NEG_CROP; \
@@ -416,7 +416,7 @@ VP8_EPEL_HV(8, 6, 6)
VP8_EPEL_HV(4, 6, 6)

#define VP8_BILINEAR(SIZE) \
static void put_vp8_bilinear ## SIZE ## _h_c(uint8_t *dst, int stride, uint8_t *src, int s2, int h, int mx, int my) \
static void put_vp8_bilinear ## SIZE ## _h_c(uint8_t *dst, ptrdiff_t stride, uint8_t *src, ptrdiff_t s2, int h, int mx, int my) \
{ \
int a = 8-mx, b = mx; \
int x, y; \
@@ -428,7 +428,7 @@ static void put_vp8_bilinear ## SIZE ## _h_c(uint8_t *dst, int stride, uint8_t *
src += stride; \
} \
} \
static void put_vp8_bilinear ## SIZE ## _v_c(uint8_t *dst, int stride, uint8_t *src, int s2, int h, int mx, int my) \
static void put_vp8_bilinear ## SIZE ## _v_c(uint8_t *dst, ptrdiff_t stride, uint8_t *src, ptrdiff_t s2, int h, int mx, int my) \
{ \
int c = 8-my, d = my; \
int x, y; \
@@ -441,7 +441,7 @@ static void put_vp8_bilinear ## SIZE ## _v_c(uint8_t *dst, int stride, uint8_t *
} \
} \
\
static void put_vp8_bilinear ## SIZE ## _hv_c(uint8_t *dst, int stride, uint8_t *src, int s2, int h, int mx, int my) \
static void put_vp8_bilinear ## SIZE ## _hv_c(uint8_t *dst, ptrdiff_t stride, uint8_t *src, ptrdiff_t s2, int h, int mx, int my) \
{ \
int a = 8-mx, b = mx; \
int c = 8-my, d = my; \


+ 26
- 19
libavcodec/vp8dsp.h View File

@@ -29,40 +29,44 @@

#include "dsputil.h"

typedef void (*vp8_mc_func)(uint8_t *dst/*align 8*/, int dstStride,
uint8_t *src/*align 1*/, int srcStride,
typedef void (*vp8_mc_func)(uint8_t *dst/*align 8*/, ptrdiff_t dstStride,
uint8_t *src/*align 1*/, ptrdiff_t srcStride,
int h, int x, int y);

typedef struct VP8DSPContext {
void (*vp8_luma_dc_wht)(DCTELEM block[4][4][16], DCTELEM dc[16]);
void (*vp8_luma_dc_wht_dc)(DCTELEM block[4][4][16], DCTELEM dc[16]);
void (*vp8_idct_add)(uint8_t *dst, DCTELEM block[16], int stride);
void (*vp8_idct_dc_add)(uint8_t *dst, DCTELEM block[16], int stride);
void (*vp8_idct_dc_add4y)(uint8_t *dst, DCTELEM block[4][16], int stride);
void (*vp8_idct_dc_add4uv)(uint8_t *dst, DCTELEM block[4][16], int stride);
void (*vp8_idct_add)(uint8_t *dst, DCTELEM block[16], ptrdiff_t stride);
void (*vp8_idct_dc_add)(uint8_t *dst, DCTELEM block[16], ptrdiff_t stride);
void (*vp8_idct_dc_add4y)(uint8_t *dst, DCTELEM block[4][16],
ptrdiff_t stride);
void (*vp8_idct_dc_add4uv)(uint8_t *dst, DCTELEM block[4][16],
ptrdiff_t stride);

// loop filter applied to edges between macroblocks
void (*vp8_v_loop_filter16y)(uint8_t *dst, int stride,
void (*vp8_v_loop_filter16y)(uint8_t *dst, ptrdiff_t stride,
int flim_E, int flim_I, int hev_thresh);
void (*vp8_h_loop_filter16y)(uint8_t *dst, int stride,
void (*vp8_h_loop_filter16y)(uint8_t *dst, ptrdiff_t stride,
int flim_E, int flim_I, int hev_thresh);
void (*vp8_v_loop_filter8uv)(uint8_t *dstU, uint8_t *dstV, int stride,
void (*vp8_v_loop_filter8uv)(uint8_t *dstU, uint8_t *dstV, ptrdiff_t stride,
int flim_E, int flim_I, int hev_thresh);
void (*vp8_h_loop_filter8uv)(uint8_t *dstU, uint8_t *dstV, int stride,
void (*vp8_h_loop_filter8uv)(uint8_t *dstU, uint8_t *dstV, ptrdiff_t stride,
int flim_E, int flim_I, int hev_thresh);

// loop filter applied to inner macroblock edges
void (*vp8_v_loop_filter16y_inner)(uint8_t *dst, int stride,
void (*vp8_v_loop_filter16y_inner)(uint8_t *dst, ptrdiff_t stride,
int flim_E, int flim_I, int hev_thresh);
void (*vp8_h_loop_filter16y_inner)(uint8_t *dst, int stride,
void (*vp8_h_loop_filter16y_inner)(uint8_t *dst, ptrdiff_t stride,
int flim_E, int flim_I, int hev_thresh);
void (*vp8_v_loop_filter8uv_inner)(uint8_t *dstU, uint8_t *dstV, int stride,
void (*vp8_v_loop_filter8uv_inner)(uint8_t *dstU, uint8_t *dstV,
ptrdiff_t stride,
int flim_E, int flim_I, int hev_thresh);
void (*vp8_h_loop_filter8uv_inner)(uint8_t *dstU, uint8_t *dstV, int stride,
void (*vp8_h_loop_filter8uv_inner)(uint8_t *dstU, uint8_t *dstV,
ptrdiff_t stride,
int flim_E, int flim_I, int hev_thresh);

void (*vp8_v_loop_filter_simple)(uint8_t *dst, int stride, int flim);
void (*vp8_h_loop_filter_simple)(uint8_t *dst, int stride, int flim);
void (*vp8_v_loop_filter_simple)(uint8_t *dst, ptrdiff_t stride, int flim);
void (*vp8_h_loop_filter_simple)(uint8_t *dst, ptrdiff_t stride, int flim);

/**
* first dimension: width>>3, height is assumed equal to width
@@ -76,9 +80,12 @@ typedef struct VP8DSPContext {
vp8_mc_func put_vp8_bilinear_pixels_tab[3][3][3];
} VP8DSPContext;

void ff_put_vp8_pixels16_c(uint8_t *dst, uint8_t *src, int stride, int h, int x, int y);
void ff_put_vp8_pixels8_c(uint8_t *dst, uint8_t *src, int stride, int h, int x, int y);
void ff_put_vp8_pixels4_c(uint8_t *dst, uint8_t *src, int stride, int h, int x, int y);
void ff_put_vp8_pixels16_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride,
int h, int x, int y);
void ff_put_vp8_pixels8_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride,
int h, int x, int y);
void ff_put_vp8_pixels4_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride,
int h, int x, int y);

void ff_vp8dsp_init(VP8DSPContext *c);
void ff_vp8dsp_init_x86(VP8DSPContext *c);


+ 106
- 83
libavcodec/x86/vp8dsp-init.c View File

@@ -29,98 +29,98 @@
/*
* MC functions
*/
extern void ff_put_vp8_epel4_h4_mmxext(uint8_t *dst, int dststride,
uint8_t *src, int srcstride,
extern void ff_put_vp8_epel4_h4_mmxext(uint8_t *dst, ptrdiff_t dststride,
uint8_t *src, ptrdiff_t srcstride,
int height, int mx, int my);
extern void ff_put_vp8_epel4_h6_mmxext(uint8_t *dst, int dststride,
uint8_t *src, int srcstride,
extern void ff_put_vp8_epel4_h6_mmxext(uint8_t *dst, ptrdiff_t dststride,
uint8_t *src, ptrdiff_t srcstride,
int height, int mx, int my);
extern void ff_put_vp8_epel4_v4_mmxext(uint8_t *dst, int dststride,
uint8_t *src, int srcstride,
extern void ff_put_vp8_epel4_v4_mmxext(uint8_t *dst, ptrdiff_t dststride,
uint8_t *src, ptrdiff_t srcstride,
int height, int mx, int my);
extern void ff_put_vp8_epel4_v6_mmxext(uint8_t *dst, int dststride,
uint8_t *src, int srcstride,
extern void ff_put_vp8_epel4_v6_mmxext(uint8_t *dst, ptrdiff_t dststride,
uint8_t *src, ptrdiff_t srcstride,
int height, int mx, int my);

extern void ff_put_vp8_epel8_h4_sse2 (uint8_t *dst, int dststride,
uint8_t *src, int srcstride,
extern void ff_put_vp8_epel8_h4_sse2 (uint8_t *dst, ptrdiff_t dststride,
uint8_t *src, ptrdiff_t srcstride,
int height, int mx, int my);
extern void ff_put_vp8_epel8_h6_sse2 (uint8_t *dst, int dststride,
uint8_t *src, int srcstride,
extern void ff_put_vp8_epel8_h6_sse2 (uint8_t *dst, ptrdiff_t dststride,
uint8_t *src, ptrdiff_t srcstride,
int height, int mx, int my);
extern void ff_put_vp8_epel8_v4_sse2 (uint8_t *dst, int dststride,
uint8_t *src, int srcstride,
extern void ff_put_vp8_epel8_v4_sse2 (uint8_t *dst, ptrdiff_t dststride,
uint8_t *src, ptrdiff_t srcstride,
int height, int mx, int my);
extern void ff_put_vp8_epel8_v6_sse2 (uint8_t *dst, int dststride,
uint8_t *src, int srcstride,
extern void ff_put_vp8_epel8_v6_sse2 (uint8_t *dst, ptrdiff_t dststride,
uint8_t *src, ptrdiff_t srcstride,
int height, int mx, int my);

extern void ff_put_vp8_epel4_h4_ssse3 (uint8_t *dst, int dststride,
uint8_t *src, int srcstride,
extern void ff_put_vp8_epel4_h4_ssse3 (uint8_t *dst, ptrdiff_t dststride,
uint8_t *src, ptrdiff_t srcstride,
int height, int mx, int my);
extern void ff_put_vp8_epel4_h6_ssse3 (uint8_t *dst, int dststride,
uint8_t *src, int srcstride,
extern void ff_put_vp8_epel4_h6_ssse3 (uint8_t *dst, ptrdiff_t dststride,
uint8_t *src, ptrdiff_t srcstride,
int height, int mx, int my);
extern void ff_put_vp8_epel4_v4_ssse3 (uint8_t *dst, int dststride,
uint8_t *src, int srcstride,
extern void ff_put_vp8_epel4_v4_ssse3 (uint8_t *dst, ptrdiff_t dststride,
uint8_t *src, ptrdiff_t srcstride,
int height, int mx, int my);
extern void ff_put_vp8_epel4_v6_ssse3 (uint8_t *dst, int dststride,
uint8_t *src, int srcstride,
extern void ff_put_vp8_epel4_v6_ssse3 (uint8_t *dst, ptrdiff_t dststride,
uint8_t *src, ptrdiff_t srcstride,
int height, int mx, int my);
extern void ff_put_vp8_epel8_h4_ssse3 (uint8_t *dst, int dststride,
uint8_t *src, int srcstride,
extern void ff_put_vp8_epel8_h4_ssse3 (uint8_t *dst, ptrdiff_t dststride,
uint8_t *src, ptrdiff_t srcstride,
int height, int mx, int my);
extern void ff_put_vp8_epel8_h6_ssse3 (uint8_t *dst, int dststride,
uint8_t *src, int srcstride,
extern void ff_put_vp8_epel8_h6_ssse3 (uint8_t *dst, ptrdiff_t dststride,
uint8_t *src, ptrdiff_t srcstride,
int height, int mx, int my);
extern void ff_put_vp8_epel8_v4_ssse3 (uint8_t *dst, int dststride,
uint8_t *src, int srcstride,
extern void ff_put_vp8_epel8_v4_ssse3 (uint8_t *dst, ptrdiff_t dststride,
uint8_t *src, ptrdiff_t srcstride,
int height, int mx, int my);
extern void ff_put_vp8_epel8_v6_ssse3 (uint8_t *dst, int dststride,
uint8_t *src, int srcstride,
extern void ff_put_vp8_epel8_v6_ssse3 (uint8_t *dst, ptrdiff_t dststride,
uint8_t *src, ptrdiff_t srcstride,
int height, int mx, int my);

extern void ff_put_vp8_bilinear4_h_mmxext(uint8_t *dst, int dststride,
uint8_t *src, int srcstride,
extern void ff_put_vp8_bilinear4_h_mmxext(uint8_t *dst, ptrdiff_t dststride,
uint8_t *src, ptrdiff_t srcstride,
int height, int mx, int my);
extern void ff_put_vp8_bilinear8_h_sse2 (uint8_t *dst, int dststride,
uint8_t *src, int srcstride,
extern void ff_put_vp8_bilinear8_h_sse2 (uint8_t *dst, ptrdiff_t dststride,
uint8_t *src, ptrdiff_t srcstride,
int height, int mx, int my);
extern void ff_put_vp8_bilinear4_h_ssse3 (uint8_t *dst, int dststride,
uint8_t *src, int srcstride,
extern void ff_put_vp8_bilinear4_h_ssse3 (uint8_t *dst, ptrdiff_t dststride,
uint8_t *src, ptrdiff_t srcstride,
int height, int mx, int my);
extern void ff_put_vp8_bilinear8_h_ssse3 (uint8_t *dst, int dststride,
uint8_t *src, int srcstride,
extern void ff_put_vp8_bilinear8_h_ssse3 (uint8_t *dst, ptrdiff_t dststride,
uint8_t *src, ptrdiff_t srcstride,
int height, int mx, int my);

extern void ff_put_vp8_bilinear4_v_mmxext(uint8_t *dst, int dststride,
uint8_t *src, int srcstride,
extern void ff_put_vp8_bilinear4_v_mmxext(uint8_t *dst, ptrdiff_t dststride,
uint8_t *src, ptrdiff_t srcstride,
int height, int mx, int my);
extern void ff_put_vp8_bilinear8_v_sse2 (uint8_t *dst, int dststride,
uint8_t *src, int srcstride,
extern void ff_put_vp8_bilinear8_v_sse2 (uint8_t *dst, ptrdiff_t dststride,
uint8_t *src, ptrdiff_t srcstride,
int height, int mx, int my);
extern void ff_put_vp8_bilinear4_v_ssse3 (uint8_t *dst, int dststride,
uint8_t *src, int srcstride,
extern void ff_put_vp8_bilinear4_v_ssse3 (uint8_t *dst, ptrdiff_t dststride,
uint8_t *src, ptrdiff_t srcstride,
int height, int mx, int my);
extern void ff_put_vp8_bilinear8_v_ssse3 (uint8_t *dst, int dststride,
uint8_t *src, int srcstride,
extern void ff_put_vp8_bilinear8_v_ssse3 (uint8_t *dst, ptrdiff_t dststride,
uint8_t *src, ptrdiff_t srcstride,
int height, int mx, int my);


extern void ff_put_vp8_pixels8_mmx (uint8_t *dst, int dststride,
uint8_t *src, int srcstride,
extern void ff_put_vp8_pixels8_mmx (uint8_t *dst, ptrdiff_t dststride,
uint8_t *src, ptrdiff_t srcstride,
int height, int mx, int my);
extern void ff_put_vp8_pixels16_mmx(uint8_t *dst, int dststride,
uint8_t *src, int srcstride,
extern void ff_put_vp8_pixels16_mmx(uint8_t *dst, ptrdiff_t dststride,
uint8_t *src, ptrdiff_t srcstride,
int height, int mx, int my);
extern void ff_put_vp8_pixels16_sse(uint8_t *dst, int dststride,
uint8_t *src, int srcstride,
extern void ff_put_vp8_pixels16_sse(uint8_t *dst, ptrdiff_t dststride,
uint8_t *src, ptrdiff_t srcstride,
int height, int mx, int my);

#define TAP_W16(OPT, FILTERTYPE, TAPTYPE) \
static void ff_put_vp8_ ## FILTERTYPE ## 16_ ## TAPTYPE ## _ ## OPT( \
uint8_t *dst, int dststride, uint8_t *src, \
int srcstride, int height, int mx, int my) \
uint8_t *dst, ptrdiff_t dststride, uint8_t *src, \
ptrdiff_t srcstride, int height, int mx, int my) \
{ \
ff_put_vp8_ ## FILTERTYPE ## 8_ ## TAPTYPE ## _ ## OPT( \
dst, dststride, src, srcstride, height, mx, my); \
@@ -129,8 +129,8 @@ static void ff_put_vp8_ ## FILTERTYPE ## 16_ ## TAPTYPE ## _ ## OPT( \
}
#define TAP_W8(OPT, FILTERTYPE, TAPTYPE) \
static void ff_put_vp8_ ## FILTERTYPE ## 8_ ## TAPTYPE ## _ ## OPT( \
uint8_t *dst, int dststride, uint8_t *src, \
int srcstride, int height, int mx, int my) \
uint8_t *dst, ptrdiff_t dststride, uint8_t *src, \
ptrdiff_t srcstride, int height, int mx, int my) \
{ \
ff_put_vp8_ ## FILTERTYPE ## 4_ ## TAPTYPE ## _ ## OPT( \
dst, dststride, src, srcstride, height, mx, my); \
@@ -161,8 +161,8 @@ TAP_W16(ssse3, bilinear, v)

#define HVTAP(OPT, ALIGN, TAPNUMX, TAPNUMY, SIZE, MAXHEIGHT) \
static void ff_put_vp8_epel ## SIZE ## _h ## TAPNUMX ## v ## TAPNUMY ## _ ## OPT( \
uint8_t *dst, int dststride, uint8_t *src, \
int srcstride, int height, int mx, int my) \
uint8_t *dst, ptrdiff_t dststride, uint8_t *src, \
ptrdiff_t srcstride, int height, int mx, int my) \
{ \
DECLARE_ALIGNED(ALIGN, uint8_t, tmp)[SIZE * (MAXHEIGHT + TAPNUMY - 1)]; \
uint8_t *tmpptr = tmp + SIZE * (TAPNUMY / 2 - 1); \
@@ -200,8 +200,8 @@ HVTAP(ssse3, 16, 6, 6, 4, 8)

#define HVBILIN(OPT, ALIGN, SIZE, MAXHEIGHT) \
static void ff_put_vp8_bilinear ## SIZE ## _hv_ ## OPT( \
uint8_t *dst, int dststride, uint8_t *src, \
int srcstride, int height, int mx, int my) \
uint8_t *dst, ptrdiff_t dststride, uint8_t *src, \
ptrdiff_t srcstride, int height, int mx, int my) \
{ \
DECLARE_ALIGNED(ALIGN, uint8_t, tmp)[SIZE * (MAXHEIGHT + 2)]; \
ff_put_vp8_bilinear ## SIZE ## _h_ ## OPT( \
@@ -219,35 +219,58 @@ HVBILIN(ssse3, 8, 4, 8)
HVBILIN(ssse3, 8, 8, 16)
HVBILIN(ssse3, 8, 16, 16)

extern void ff_vp8_idct_dc_add_mmx(uint8_t *dst, DCTELEM block[16], int stride);
extern void ff_vp8_idct_dc_add_sse4(uint8_t *dst, DCTELEM block[16], int stride);
extern void ff_vp8_idct_dc_add4y_mmx(uint8_t *dst, DCTELEM block[4][16], int stride);
extern void ff_vp8_idct_dc_add4y_sse2(uint8_t *dst, DCTELEM block[4][16], int stride);
extern void ff_vp8_idct_dc_add4uv_mmx(uint8_t *dst, DCTELEM block[2][16], int stride);
extern void ff_vp8_idct_dc_add_mmx(uint8_t *dst, DCTELEM block[16],
ptrdiff_t stride);
extern void ff_vp8_idct_dc_add_sse4(uint8_t *dst, DCTELEM block[16],
ptrdiff_t stride);
extern void ff_vp8_idct_dc_add4y_mmx(uint8_t *dst, DCTELEM block[4][16],
ptrdiff_t stride);
extern void ff_vp8_idct_dc_add4y_sse2(uint8_t *dst, DCTELEM block[4][16],
ptrdiff_t stride);
extern void ff_vp8_idct_dc_add4uv_mmx(uint8_t *dst, DCTELEM block[2][16],
ptrdiff_t stride);
extern void ff_vp8_luma_dc_wht_mmx(DCTELEM block[4][4][16], DCTELEM dc[16]);
extern void ff_vp8_luma_dc_wht_sse(DCTELEM block[4][4][16], DCTELEM dc[16]);
extern void ff_vp8_idct_add_mmx(uint8_t *dst, DCTELEM block[16], int stride);
extern void ff_vp8_idct_add_sse(uint8_t *dst, DCTELEM block[16], int stride);
extern void ff_vp8_idct_add_mmx(uint8_t *dst, DCTELEM block[16],
ptrdiff_t stride);
extern void ff_vp8_idct_add_sse(uint8_t *dst, DCTELEM block[16],
ptrdiff_t stride);

#define DECLARE_LOOP_FILTER(NAME)\
extern void ff_vp8_v_loop_filter_simple_ ## NAME(uint8_t *dst, int stride, int flim);\
extern void ff_vp8_h_loop_filter_simple_ ## NAME(uint8_t *dst, int stride, int flim);\
extern void ff_vp8_v_loop_filter16y_inner_ ## NAME (uint8_t *dst, int stride,\
extern void ff_vp8_v_loop_filter_simple_ ## NAME(uint8_t *dst, \
ptrdiff_t stride, \
int flim);\
extern void ff_vp8_h_loop_filter_simple_ ## NAME(uint8_t *dst, \
ptrdiff_t stride, \
int flim);\
extern void ff_vp8_v_loop_filter16y_inner_ ## NAME (uint8_t *dst, \
ptrdiff_t stride,\
int e, int i, int hvt);\
extern void ff_vp8_h_loop_filter16y_inner_ ## NAME (uint8_t *dst, int stride,\
extern void ff_vp8_h_loop_filter16y_inner_ ## NAME (uint8_t *dst, \
ptrdiff_t stride,\
int e, int i, int hvt);\
extern void ff_vp8_v_loop_filter8uv_inner_ ## NAME (uint8_t *dstU, uint8_t *dstV,\
int s, int e, int i, int hvt);\
extern void ff_vp8_h_loop_filter8uv_inner_ ## NAME (uint8_t *dstU, uint8_t *dstV,\
int s, int e, int i, int hvt);\
extern void ff_vp8_v_loop_filter16y_mbedge_ ## NAME(uint8_t *dst, int stride,\
extern void ff_vp8_v_loop_filter8uv_inner_ ## NAME (uint8_t *dstU, \
uint8_t *dstV,\
ptrdiff_t s, \
int e, int i, int hvt);\
extern void ff_vp8_h_loop_filter16y_mbedge_ ## NAME(uint8_t *dst, int stride,\
extern void ff_vp8_h_loop_filter8uv_inner_ ## NAME (uint8_t *dstU, \
uint8_t *dstV,\
ptrdiff_t s, \
int e, int i, int hvt);\
extern void ff_vp8_v_loop_filter8uv_mbedge_ ## NAME(uint8_t *dstU, uint8_t *dstV,\
int s, int e, int i, int hvt);\
extern void ff_vp8_h_loop_filter8uv_mbedge_ ## NAME(uint8_t *dstU, uint8_t *dstV,\
int s, int e, int i, int hvt);
extern void ff_vp8_v_loop_filter16y_mbedge_ ## NAME(uint8_t *dst, \
ptrdiff_t stride,\
int e, int i, int hvt);\
extern void ff_vp8_h_loop_filter16y_mbedge_ ## NAME(uint8_t *dst, \
ptrdiff_t stride,\
int e, int i, int hvt);\
extern void ff_vp8_v_loop_filter8uv_mbedge_ ## NAME(uint8_t *dstU, \
uint8_t *dstV,\
ptrdiff_t s, \
int e, int i, int hvt);\
extern void ff_vp8_h_loop_filter8uv_mbedge_ ## NAME(uint8_t *dstU, \
uint8_t *dstV,\
ptrdiff_t s, \
int e, int i, int hvt);

DECLARE_LOOP_FILTER(mmx)
DECLARE_LOOP_FILTER(mmxext)


Write Preview
Loading…
Cancel
Save