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 '458446acfa1441d283dacf9e6e545beb083b8bb0' 

* commit '458446acfa1441d283dacf9e6e545beb083b8bb0':
  lavc: Edge emulation with dst/src linesize

Conflicts:
	libavcodec/cavs.c
	libavcodec/h264.c
	libavcodec/hevc.c
	libavcodec/mpegvideo_enc.c
	libavcodec/mpegvideo_motion.c
	libavcodec/rv34.c
	libavcodec/svq3.c
	libavcodec/vc1dec.c
	libavcodec/videodsp.h
	libavcodec/videodsp_template.c
	libavcodec/vp3.c
	libavcodec/vp8.c
	libavcodec/wmv2.c
	libavcodec/x86/videodsp.asm
	libavcodec/x86/videodsp_init.c

Changes to the asm are not merged, they are left for volunteers or
in their absence for later.
The changes this merge introduces are reordering of the function
arguments

See: face578d56
Merged-by: Michael Niedermayer <michaelni@gmx.at>
tags/n2.2-rc1
Michael Niedermayer 11 years ago
parent
e28130bcaf 458446acfa
commit
91e00c4a78
20 changed files with 226 additions and 172 deletions
Split View
Diff Options
Show Stats Download Patch File Download Diff File
  1. +7
    -6
      libavcodec/cavs.c
  2. +2
    -2
      libavcodec/diracdec.c
  3. +10
    -8
      libavcodec/h264.c
  4. +6
    -3
      libavcodec/hevc.c
  5. +22
    -19
      libavcodec/mpegvideo.c
  6. +12
    -6
      libavcodec/mpegvideo_enc.c
  7. +26
    -30
      libavcodec/mpegvideo_motion.c
  8. +12
    -6
      libavcodec/rv34.c
  9. +2
    -1
      libavcodec/snow.c
  10. +4
    -4
      libavcodec/svq3.c
  11. +36
    -22
      libavcodec/vc1dec.c
  12. +9
    -6
      libavcodec/videodsp.h
  13. +15
    -14
      libavcodec/videodsp_template.c
  14. +4
    -3
      libavcodec/vp3.c
  15. +6
    -5
      libavcodec/vp56.c
  16. +9
    -6
      libavcodec/vp8.c
  17. +8
    -6
      libavcodec/vp9.c
  18. +15
    -9
      libavcodec/wmv2.c
  19. +4
    -3
      libavcodec/x86/dsputil_mmx.c
  20. +17
    -13
      libavcodec/x86/videodsp_init.c

+ 7
- 6
libavcodec/cavs.c View File

@@ -407,8 +407,9 @@ static inline void mc_dir_part(AVSContext *h, AVFrame *pic, int chroma_height,
full_my < 0 - extra_height ||
full_mx + 16 /* FIXME */ > pic_width + extra_width ||
full_my + 16 /* FIXME */ > pic_height + extra_height) {
h->vdsp.emulated_edge_mc(h->edge_emu_buffer, h->l_stride,
src_y - 2 - 2 * h->l_stride, h->l_stride,
h->vdsp.emulated_edge_mc(h->edge_emu_buffer,
src_y - 2 - 2 * h->l_stride,
h->l_stride, h->l_stride,
16 + 5, 16 + 5 /* FIXME */,
full_mx - 2, full_my - 2,
pic_width, pic_height);
@@ -420,8 +421,8 @@ static inline void mc_dir_part(AVSContext *h, AVFrame *pic, int chroma_height,
qpix_op[luma_xy](dest_y, src_y, h->l_stride);

if (emu) {
h->vdsp.emulated_edge_mc(h->edge_emu_buffer, h->c_stride, src_cb,
h->c_stride,
h->vdsp.emulated_edge_mc(h->edge_emu_buffer, src_cb,
h->c_stride, h->c_stride,
9, 9 /* FIXME */,
mx >> 3, my >> 3,
pic_width >> 1, pic_height >> 1);
@@ -430,8 +431,8 @@ static inline void mc_dir_part(AVSContext *h, AVFrame *pic, int chroma_height,
chroma_op(dest_cb, src_cb, h->c_stride, chroma_height, mx & 7, my & 7);

if (emu) {
h->vdsp.emulated_edge_mc(h->edge_emu_buffer, h->c_stride, src_cr,
h->c_stride,
h->vdsp.emulated_edge_mc(h->edge_emu_buffer, src_cr,
h->c_stride, h->c_stride,
9, 9 /* FIXME */,
mx >> 3, my >> 3,
pic_width >> 1, pic_height >> 1);


+ 2
- 2
libavcodec/diracdec.c View File

@@ -1414,8 +1414,8 @@ static int mc_subpel(DiracContext *s, DiracBlock *block, const uint8_t *src[5],
y + p->yblen > p->height+EDGE_WIDTH/2 ||
x < 0 || y < 0) {
for (i = 0; i < nplanes; i++) {
ff_emulated_edge_mc(s->edge_emu_buffer[i], p->stride,
src[i], p->stride,
ff_emulated_edge_mc(s->edge_emu_buffer[i], src[i],
p->stride, p->stride,
p->xblen, p->yblen, x, y,
p->width+EDGE_WIDTH/2, p->height+EDGE_WIDTH/2);
src[i] = s->edge_emu_buffer[i];


+ 10
- 8
libavcodec/h264.c View File

@@ -934,9 +934,9 @@ static av_always_inline void mc_dir_part(H264Context *h, Picture *pic,
full_my < 0 - extra_height ||
full_mx + 16 /*FIXME*/ > pic_width + extra_width ||
full_my + 16 /*FIXME*/ > pic_height + extra_height) {
h->vdsp.emulated_edge_mc(h->edge_emu_buffer, h->mb_linesize,
h->vdsp.emulated_edge_mc(h->edge_emu_buffer,
src_y - (2 << pixel_shift) - 2 * h->mb_linesize,
h->mb_linesize,
h->mb_linesize, h->mb_linesize,
16 + 5, 16 + 5 /*FIXME*/, full_mx - 2,
full_my - 2, pic_width, pic_height);
src_y = h->edge_emu_buffer + (2 << pixel_shift) + 2 * h->mb_linesize;
@@ -953,9 +953,9 @@ static av_always_inline void mc_dir_part(H264Context *h, Picture *pic,
if (chroma_idc == 3 /* yuv444 */) {
src_cb = pic->f.data[1] + offset;
if (emu) {
h->vdsp.emulated_edge_mc(h->edge_emu_buffer, h->mb_linesize,
h->vdsp.emulated_edge_mc(h->edge_emu_buffer,
src_cb - (2 << pixel_shift) - 2 * h->mb_linesize,
h->mb_linesize,
h->mb_linesize, h->mb_linesize,
16 + 5, 16 + 5 /*FIXME*/,
full_mx - 2, full_my - 2,
pic_width, pic_height);
@@ -967,9 +967,9 @@ static av_always_inline void mc_dir_part(H264Context *h, Picture *pic,

src_cr = pic->f.data[2] + offset;
if (emu) {
h->vdsp.emulated_edge_mc(h->edge_emu_buffer, h->mb_linesize,
h->vdsp.emulated_edge_mc(h->edge_emu_buffer,
src_cr - (2 << pixel_shift) - 2 * h->mb_linesize,
h->mb_linesize,
h->mb_linesize, h->mb_linesize,
16 + 5, 16 + 5 /*FIXME*/,
full_mx - 2, full_my - 2,
pic_width, pic_height);
@@ -994,7 +994,8 @@ static av_always_inline void mc_dir_part(H264Context *h, Picture *pic,
(my >> ysh) * h->mb_uvlinesize;

if (emu) {
h->vdsp.emulated_edge_mc(h->edge_emu_buffer, h->mb_uvlinesize, src_cb, h->mb_uvlinesize,
h->vdsp.emulated_edge_mc(h->edge_emu_buffer, src_cb,
h->mb_uvlinesize, h->mb_uvlinesize,
9, 8 * chroma_idc + 1, (mx >> 3), (my >> ysh),
pic_width >> 1, pic_height >> (chroma_idc == 1 /* yuv420 */));
src_cb = h->edge_emu_buffer;
@@ -1004,7 +1005,8 @@ static av_always_inline void mc_dir_part(H264Context *h, Picture *pic,
mx & 7, (my << (chroma_idc == 2 /* yuv422 */)) & 7);

if (emu) {
h->vdsp.emulated_edge_mc(h->edge_emu_buffer, h->mb_uvlinesize, src_cr, h->mb_uvlinesize,
h->vdsp.emulated_edge_mc(h->edge_emu_buffer, src_cr,
h->mb_uvlinesize, h->mb_uvlinesize,
9, 8 * chroma_idc + 1, (mx >> 3), (my >> ysh),
pic_width >> 1, pic_height >> (chroma_idc == 1 /* yuv420 */));
src_cr = h->edge_emu_buffer;


+ 6
- 3
libavcodec/hevc.c View File

@@ -1026,7 +1026,8 @@ static void luma_mc(HEVCContext *s, int16_t *dst, ptrdiff_t dststride,
y_off >= pic_height - block_h - ff_hevc_qpel_extra_after[my]) {
int offset = extra_top * srcstride + (extra_left << s->sps->pixel_shift);

s->vdsp.emulated_edge_mc(lc->edge_emu_buffer, srcstride, src - offset, srcstride,
s->vdsp.emulated_edge_mc(lc->edge_emu_buffer, src - offset,
srcstride, srcstride,
block_w + ff_hevc_qpel_extra[mx],
block_h + ff_hevc_qpel_extra[my],
x_off - extra_left, y_off - extra_top,
@@ -1077,7 +1078,8 @@ static void chroma_mc(HEVCContext *s, int16_t *dst1, int16_t *dst2,
int offset1 = EPEL_EXTRA_BEFORE * (src1stride + (1 << s->sps->pixel_shift));
int offset2 = EPEL_EXTRA_BEFORE * (src2stride + (1 << s->sps->pixel_shift));

s->vdsp.emulated_edge_mc(lc->edge_emu_buffer, src1stride, src1 - offset1, src1stride,
s->vdsp.emulated_edge_mc(lc->edge_emu_buffer, src1 - offset1,
src1stride, src1stride,
block_w + EPEL_EXTRA, block_h + EPEL_EXTRA,
x_off - EPEL_EXTRA_BEFORE,
y_off - EPEL_EXTRA_BEFORE,
@@ -1087,7 +1089,8 @@ static void chroma_mc(HEVCContext *s, int16_t *dst1, int16_t *dst2,
s->hevcdsp.put_hevc_epel[!!my][!!mx](dst1, dststride, src1, src1stride,
block_w, block_h, mx, my, lc->mc_buffer);

s->vdsp.emulated_edge_mc(lc->edge_emu_buffer, src2stride, src2 - offset2, src2stride,
s->vdsp.emulated_edge_mc(lc->edge_emu_buffer, src2 - offset2,
src2stride, src2stride,
block_w + EPEL_EXTRA, block_h + EPEL_EXTRA,
x_off - EPEL_EXTRA_BEFORE,
y_off - EPEL_EXTRA_BEFORE,


+ 22
- 19
libavcodec/mpegvideo.c View File

@@ -2227,12 +2227,11 @@ static inline int hpel_motion_lowres(MpegEncContext *s,

if ((unsigned)src_x > FFMAX( h_edge_pos - (!!sx) - w, 0) ||
(unsigned)src_y > FFMAX((v_edge_pos >> field_based) - (!!sy) - h, 0)) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, s->linesize,
src, s->linesize, w + 1,
(h + 1) << field_based, src_x,
src_y << field_based,
h_edge_pos,
v_edge_pos);
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, src,
s->linesize, s->linesize,
w + 1, (h + 1) << field_based,
src_x, src_y << field_based,
h_edge_pos, v_edge_pos);
src = s->edge_emu_buffer;
emu = 1;
}
@@ -2330,21 +2329,22 @@ static av_always_inline void mpeg_motion_lowres(MpegEncContext *s,

if ((unsigned) src_x > FFMAX( h_edge_pos - (!!sx) - 2 * block_s, 0) || uvsrc_y<0 ||
(unsigned) src_y > FFMAX((v_edge_pos >> field_based) - (!!sy) - h, 0)) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, linesize >> field_based, ptr_y,
linesize >> field_based, 17, 17 + field_based,
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, ptr_y,
linesize >> field_based, linesize >> field_based,
17, 17 + field_based,
src_x, src_y << field_based, h_edge_pos,
v_edge_pos);
ptr_y = s->edge_emu_buffer;
if (!CONFIG_GRAY || !(s->flags & CODEC_FLAG_GRAY)) {
uint8_t *uvbuf = s->edge_emu_buffer + 18 * s->linesize;
s->vdsp.emulated_edge_mc(uvbuf, uvlinesize >> field_based,
ptr_cb, uvlinesize >> field_based, 9,
9 + field_based,
s->vdsp.emulated_edge_mc(uvbuf, ptr_cb,
uvlinesize >> field_based, uvlinesize >> field_based,
9, 9 + field_based,
uvsrc_x, uvsrc_y << field_based,
h_edge_pos >> 1, v_edge_pos >> 1);
s->vdsp.emulated_edge_mc(uvbuf + 16, uvlinesize >> field_based,
ptr_cr, uvlinesize >> field_based, 9,
9 + field_based,
s->vdsp.emulated_edge_mc(uvbuf + 16, ptr_cr,
uvlinesize >> field_based,uvlinesize >> field_based,
9, 9 + field_based,
uvsrc_x, uvsrc_y << field_based,
h_edge_pos >> 1, v_edge_pos >> 1);
ptr_cb = uvbuf;
@@ -2416,8 +2416,10 @@ static inline void chroma_4mv_motion_lowres(MpegEncContext *s,
ptr = ref_picture[1] + offset;
if ((unsigned) src_x > FFMAX(h_edge_pos - (!!sx) - block_s, 0) ||
(unsigned) src_y > FFMAX(v_edge_pos - (!!sy) - block_s, 0)) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, s->uvlinesize, ptr, s->uvlinesize,
9, 9, src_x, src_y, h_edge_pos, v_edge_pos);
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, ptr,
s->uvlinesize, s->uvlinesize,
9, 9,
src_x, src_y, h_edge_pos, v_edge_pos);
ptr = s->edge_emu_buffer;
emu = 1;
}
@@ -2427,9 +2429,10 @@ static inline void chroma_4mv_motion_lowres(MpegEncContext *s,

ptr = ref_picture[2] + offset;
if (emu) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, s->uvlinesize,
ptr, s->uvlinesize, 9, 9,
src_x, src_y, h_edge_pos, v_edge_pos);
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, ptr,
s->uvlinesize, s->uvlinesize,
9, 9,
src_x, src_y, h_edge_pos, v_edge_pos);
ptr = s->edge_emu_buffer;
}
pix_op[op_index](dest_cr, ptr, s->uvlinesize, block_s, sx, sy);


+ 12
- 6
libavcodec/mpegvideo_enc.c View File

@@ -1837,15 +1837,21 @@ static av_always_inline void encode_mb_internal(MpegEncContext *s,
uint8_t *ebuf = s->edge_emu_buffer + 32;
int cw = (s->width + s->chroma_x_shift) >> s->chroma_x_shift;
int ch = (s->height + s->chroma_y_shift) >> s->chroma_y_shift;
s->vdsp.emulated_edge_mc(ebuf, wrap_y, ptr_y, wrap_y, 16, 16, mb_x * 16,
mb_y * 16, s->width, s->height);
s->vdsp.emulated_edge_mc(ebuf, ptr_y,
wrap_y, wrap_y,
16, 16, mb_x * 16, mb_y * 16,
s->width, s->height);
ptr_y = ebuf;
s->vdsp.emulated_edge_mc(ebuf + 18 * wrap_y, wrap_c, ptr_cb, wrap_c, mb_block_width,
mb_block_height, mb_x * mb_block_width, mb_y * mb_block_height,
s->vdsp.emulated_edge_mc(ebuf + 18 * wrap_y, ptr_cb,
wrap_c, wrap_c,
mb_block_width, mb_block_height,
mb_x * mb_block_width, mb_y * mb_block_height,
cw, ch);
ptr_cb = ebuf + 18 * wrap_y;
s->vdsp.emulated_edge_mc(ebuf + 18 * wrap_y + 16, wrap_c, ptr_cr, wrap_c, mb_block_width,
mb_block_height, mb_x * mb_block_width, mb_y * mb_block_height,
s->vdsp.emulated_edge_mc(ebuf + 18 * wrap_y + 16, ptr_cr,
wrap_c, wrap_c,
mb_block_width, mb_block_height,
mb_x * mb_block_width, mb_y * mb_block_height,
cw, ch);
ptr_cr = ebuf + 18 * wrap_y + 16;
}


+ 26
- 30
libavcodec/mpegvideo_motion.c View File

@@ -62,8 +62,8 @@ static void gmc1_motion(MpegEncContext *s,

if ((unsigned)src_x >= FFMAX(s->h_edge_pos - 17, 0) ||
(unsigned)src_y >= FFMAX(s->v_edge_pos - 17, 0)) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, linesize, ptr,
linesize,
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, ptr,
linesize, linesize,
17, 17,
src_x, src_y,
s->h_edge_pos, s->v_edge_pos);
@@ -106,8 +106,8 @@ static void gmc1_motion(MpegEncContext *s,
ptr = ref_picture[1] + offset;
if ((unsigned)src_x >= FFMAX((s->h_edge_pos >> 1) - 9, 0) ||
(unsigned)src_y >= FFMAX((s->v_edge_pos >> 1) - 9, 0)) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, uvlinesize,
ptr, uvlinesize,
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, ptr,
uvlinesize, uvlinesize,
9, 9,
src_x, src_y,
s->h_edge_pos >> 1, s->v_edge_pos >> 1);
@@ -119,8 +119,8 @@ static void gmc1_motion(MpegEncContext *s,

ptr = ref_picture[2] + offset;
if (emu) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, uvlinesize,
ptr, uvlinesize,
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, ptr,
uvlinesize, uvlinesize,
9, 9,
src_x, src_y,
s->h_edge_pos >> 1, s->v_edge_pos >> 1);
@@ -212,8 +212,8 @@ static inline int hpel_motion(MpegEncContext *s,
if (s->flags & CODEC_FLAG_EMU_EDGE) {
if ((unsigned)src_x > FFMAX(s->h_edge_pos - (motion_x & 1) - 8, 0) ||
(unsigned)src_y > FFMAX(s->v_edge_pos - (motion_y & 1) - 8, 0)) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, s->linesize, src,
s->linesize,
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, src,
s->linesize, s->linesize,
9, 9,
src_x, src_y,
s->h_edge_pos, s->v_edge_pos);
@@ -317,21 +317,21 @@ void mpeg_motion_internal(MpegEncContext *s,
src_y);
return;
}
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, s->linesize, ptr_y,
s->linesize,
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, ptr_y,
s->linesize, s->linesize,
17, 17 + field_based,
src_x, src_y << field_based,
s->h_edge_pos, s->v_edge_pos);
ptr_y = s->edge_emu_buffer;
if (!CONFIG_GRAY || !(s->flags & CODEC_FLAG_GRAY)) {
uint8_t *uvbuf = s->edge_emu_buffer + 18 * s->linesize;
s->vdsp.emulated_edge_mc(uvbuf, s->uvlinesize, ptr_cb,
s->uvlinesize,
s->vdsp.emulated_edge_mc(uvbuf, ptr_cb,
s->uvlinesize, s->uvlinesize,
9, 9 + field_based,
uvsrc_x, uvsrc_y << field_based,
s->h_edge_pos >> 1, s->v_edge_pos >> 1);
s->vdsp.emulated_edge_mc(uvbuf + 16, s->uvlinesize, ptr_cr,
s->uvlinesize,
s->vdsp.emulated_edge_mc(uvbuf + 16, ptr_cr,
s->uvlinesize, s->uvlinesize,
9, 9 + field_based,
uvsrc_x, uvsrc_y << field_based,
s->h_edge_pos >> 1, s->v_edge_pos >> 1);
@@ -538,24 +538,21 @@ static inline void qpel_motion(MpegEncContext *s,

if ((unsigned)src_x > FFMAX(s->h_edge_pos - (motion_x & 3) - 16, 0) ||
(unsigned)src_y > FFMAX( v_edge_pos - (motion_y & 3) - h, 0)) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer,
s->linesize,
ptr_y, s->linesize,
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, ptr_y,
s->linesize, s->linesize,
17, 17 + field_based,
src_x, src_y << field_based,
s->h_edge_pos, s->v_edge_pos);
ptr_y = s->edge_emu_buffer;
if (!CONFIG_GRAY || !(s->flags & CODEC_FLAG_GRAY)) {
uint8_t *uvbuf = s->edge_emu_buffer + 18 * s->linesize;
s->vdsp.emulated_edge_mc(uvbuf, s->uvlinesize,
ptr_cb,
s->uvlinesize,
s->vdsp.emulated_edge_mc(uvbuf, ptr_cb,
s->uvlinesize, s->uvlinesize,
9, 9 + field_based,
uvsrc_x, uvsrc_y << field_based,
s->h_edge_pos >> 1, s->v_edge_pos >> 1);
s->vdsp.emulated_edge_mc(uvbuf + 16, s->uvlinesize,
ptr_cr,
s->uvlinesize,
s->vdsp.emulated_edge_mc(uvbuf + 16, ptr_cr,
s->uvlinesize, s->uvlinesize,
9, 9 + field_based,
uvsrc_x, uvsrc_y << field_based,
s->h_edge_pos >> 1, s->v_edge_pos >> 1);
@@ -625,8 +622,8 @@ static void chroma_4mv_motion(MpegEncContext *s,
if (s->flags & CODEC_FLAG_EMU_EDGE) {
if ((unsigned)src_x > FFMAX((s->h_edge_pos >> 1) - (dxy & 1) - 8, 0) ||
(unsigned)src_y > FFMAX((s->v_edge_pos >> 1) - (dxy >> 1) - 8, 0)) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, s->uvlinesize,
ptr, s->uvlinesize,
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, ptr,
s->uvlinesize, s->uvlinesize,
9, 9, src_x, src_y,
s->h_edge_pos >> 1, s->v_edge_pos >> 1);
ptr = s->edge_emu_buffer;
@@ -637,8 +634,8 @@ static void chroma_4mv_motion(MpegEncContext *s,

ptr = ref_picture[2] + offset;
if (emu) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, s->uvlinesize,
ptr, s->uvlinesize,
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, ptr,
s->uvlinesize, s->uvlinesize,
9, 9, src_x, src_y,
s->h_edge_pos >> 1, s->v_edge_pos >> 1);
ptr = s->edge_emu_buffer;
@@ -785,9 +782,8 @@ static inline void apply_8x8(MpegEncContext *s,
if (s->flags & CODEC_FLAG_EMU_EDGE) {
if ((unsigned)src_x > FFMAX(s->h_edge_pos - (motion_x & 3) - 8, 0) ||
(unsigned)src_y > FFMAX(s->v_edge_pos - (motion_y & 3) - 8, 0)) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer,
s->linesize, ptr,
s->linesize,
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, ptr,
s->linesize, s->linesize,
9, 9,
src_x, src_y,
s->h_edge_pos,


+ 12
- 6
libavcodec/rv34.c View File

@@ -724,15 +724,21 @@ static inline void rv34_mc(RV34DecContext *r, const int block_type,
uint8_t *uvbuf = s->edge_emu_buffer + 22 * s->linesize;

srcY -= 2 + 2*s->linesize;
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, s->linesize, srcY, s->linesize,
(width<<3)+6, (height<<3)+6, src_x - 2, src_y - 2,
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, srcY,
s->linesize, s->linesize,
(width << 3) + 6, (height << 3) + 6,
src_x - 2, src_y - 2,
s->h_edge_pos, s->v_edge_pos);
srcY = s->edge_emu_buffer + 2 + 2*s->linesize;
s->vdsp.emulated_edge_mc(uvbuf, s->uvlinesize, srcU, s->uvlinesize,
(width<<2)+1, (height<<2)+1, uvsrc_x, uvsrc_y,
s->vdsp.emulated_edge_mc(uvbuf, srcU,
s->uvlinesize, s->uvlinesize,
(width << 2) + 1, (height << 2) + 1,
uvsrc_x, uvsrc_y,
s->h_edge_pos >> 1, s->v_edge_pos >> 1);
s->vdsp.emulated_edge_mc(uvbuf + 16, s->uvlinesize, srcV, s->uvlinesize,
(width<<2)+1, (height<<2)+1, uvsrc_x, uvsrc_y,
s->vdsp.emulated_edge_mc(uvbuf + 16, srcV,
s->uvlinesize, s->uvlinesize,
(width << 2) + 1, (height << 2) + 1,
uvsrc_x, uvsrc_y,
s->h_edge_pos >> 1, s->v_edge_pos >> 1);
srcU = uvbuf;
srcV = uvbuf + 16;


+ 2
- 1
libavcodec/snow.c View File

@@ -349,7 +349,8 @@ void ff_snow_pred_block(SnowContext *s, uint8_t *dst, uint8_t *tmp, ptrdiff_t st
src += sx + sy*stride;
if( (unsigned)sx >= FFMAX(w - b_w - (HTAPS_MAX-2), 0)
|| (unsigned)sy >= FFMAX(h - b_h - (HTAPS_MAX-2), 0)){
s->vdsp.emulated_edge_mc(tmp + MB_SIZE, stride, src, stride,
s->vdsp.emulated_edge_mc(tmp + MB_SIZE, src,
stride, stride,
b_w+HTAPS_MAX-1, b_h+HTAPS_MAX-1,
sx, sy, w, h);
src= tmp + MB_SIZE;


+ 4
- 4
libavcodec/svq3.c View File

@@ -317,8 +317,8 @@ static inline void svq3_mc_dir_part(SVQ3Context *s,
src = pic->f.data[0] + mx + my * h->linesize;

if (emu) {
h->vdsp.emulated_edge_mc(h->edge_emu_buffer, h->linesize,
src, h->linesize,
h->vdsp.emulated_edge_mc(h->edge_emu_buffer, src,
h->linesize, h->linesize,
width + 1, height + 1,
mx, my, s->h_edge_pos, s->v_edge_pos);
src = h->edge_emu_buffer;
@@ -344,8 +344,8 @@ static inline void svq3_mc_dir_part(SVQ3Context *s,
src = pic->f.data[i] + mx + my * h->uvlinesize;

if (emu) {
h->vdsp.emulated_edge_mc(h->edge_emu_buffer, h->uvlinesize,
src, h->uvlinesize,
h->vdsp.emulated_edge_mc(h->edge_emu_buffer, src,
h->uvlinesize, h->uvlinesize,
width + 1, height + 1,
mx, my, (s->h_edge_pos >> 1),
s->v_edge_pos >> 1);


+ 36
- 22
libavcodec/vc1dec.c View File

@@ -454,17 +454,21 @@ static void vc1_mc_1mv(VC1Context *v, int dir)
uint8_t *uvbuf = s->edge_emu_buffer + 19 * s->linesize;

srcY -= s->mspel * (1 + s->linesize);
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, s->linesize,
srcY, s->linesize,
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, srcY,
s->linesize, s->linesize,
17 + s->mspel * 2, 17 + s->mspel * 2,
src_x - s->mspel, src_y - s->mspel,
s->h_edge_pos, v_edge_pos);
srcY = s->edge_emu_buffer;
s->vdsp.emulated_edge_mc(uvbuf, s->uvlinesize, srcU, s->uvlinesize,
8 + 1, 8 + 1, uvsrc_x, uvsrc_y,
s->vdsp.emulated_edge_mc(uvbuf, srcU,
s->uvlinesize, s->uvlinesize,
8 + 1, 8 + 1,
uvsrc_x, uvsrc_y,
s->h_edge_pos >> 1, v_edge_pos >> 1);
s->vdsp.emulated_edge_mc(uvbuf + 16, s->uvlinesize, srcV, s->uvlinesize,
8 + 1, 8 + 1, uvsrc_x, uvsrc_y,
s->vdsp.emulated_edge_mc(uvbuf + 16, srcV,
s->uvlinesize, s->uvlinesize,
8 + 1, 8 + 1,
uvsrc_x, uvsrc_y,
s->h_edge_pos >> 1, v_edge_pos >> 1);
srcU = uvbuf;
srcV = uvbuf + 16;
@@ -700,7 +704,8 @@ static void vc1_mc_4mv_luma(VC1Context *v, int n, int dir, int avg)
|| (unsigned)(src_y - (s->mspel << fieldmv)) > v_edge_pos - (my & 3) - ((8 + s->mspel * 2) << fieldmv)) {
srcY -= s->mspel * (1 + (s->linesize << fieldmv));
/* check emulate edge stride and offset */
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, s->linesize, srcY, s->linesize,
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, srcY,
s->linesize, s->linesize,
9 + s->mspel * 2, (9 + s->mspel * 2) << fieldmv,
src_x - s->mspel, src_y - (s->mspel << fieldmv),
s->h_edge_pos, v_edge_pos);
@@ -915,11 +920,13 @@ static void vc1_mc_4mv_chroma(VC1Context *v, int dir)
|| s->h_edge_pos < 18 || v_edge_pos < 18
|| (unsigned)uvsrc_x > (s->h_edge_pos >> 1) - 9
|| (unsigned)uvsrc_y > (v_edge_pos >> 1) - 9) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, s->uvlinesize, srcU,
s->uvlinesize, 8 + 1, 8 + 1, uvsrc_x, uvsrc_y,
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, srcU,
s->uvlinesize, s->uvlinesize,
8 + 1, 8 + 1, uvsrc_x, uvsrc_y,
s->h_edge_pos >> 1, v_edge_pos >> 1);
s->vdsp.emulated_edge_mc(s->edge_emu_buffer + 16, s->uvlinesize, srcV,
s->uvlinesize, 8 + 1, 8 + 1, uvsrc_x, uvsrc_y,
s->vdsp.emulated_edge_mc(s->edge_emu_buffer + 16, srcV,
s->uvlinesize, s->uvlinesize,
8 + 1, 8 + 1, uvsrc_x, uvsrc_y,
s->h_edge_pos >> 1, v_edge_pos >> 1);
srcU = s->edge_emu_buffer;
srcV = s->edge_emu_buffer + 16;
@@ -1036,12 +1043,14 @@ static void vc1_mc_4mv_chroma4(VC1Context *v, int dir, int dir2, int avg)
|| s->h_edge_pos < 10 || v_edge_pos < (5 << fieldmv)
|| (unsigned)uvsrc_x > (s->h_edge_pos >> 1) - 5
|| (unsigned)uvsrc_y > v_edge_pos - (5 << fieldmv)) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, s->uvlinesize, srcU,
s->uvlinesize, 5, (5 << fieldmv), uvsrc_x,
uvsrc_y, s->h_edge_pos >> 1, v_edge_pos);
s->vdsp.emulated_edge_mc(s->edge_emu_buffer + 16, s->uvlinesize, srcV,
s->uvlinesize, 5, (5 << fieldmv), uvsrc_x,
uvsrc_y, s->h_edge_pos >> 1, v_edge_pos);
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, srcU,
s->uvlinesize, s->uvlinesize,
5, (5 << fieldmv), uvsrc_x, uvsrc_y,
s->h_edge_pos >> 1, v_edge_pos);
s->vdsp.emulated_edge_mc(s->edge_emu_buffer + 16, srcV,
s->uvlinesize, s->uvlinesize,
5, (5 << fieldmv), uvsrc_x, uvsrc_y,
s->h_edge_pos >> 1, v_edge_pos);
srcU = s->edge_emu_buffer;
srcV = s->edge_emu_buffer + 16;

@@ -1968,16 +1977,21 @@ static void vc1_interp_mc(VC1Context *v)
uint8_t *uvbuf = s->edge_emu_buffer + 19 * s->linesize;

srcY -= s->mspel * (1 + s->linesize);
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, s->linesize, srcY, s->linesize,
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, srcY,
s->linesize, s->linesize,
17 + s->mspel * 2, 17 + s->mspel * 2,
src_x - s->mspel, src_y - s->mspel,
s->h_edge_pos, v_edge_pos);
srcY = s->edge_emu_buffer;
s->vdsp.emulated_edge_mc(uvbuf, s->uvlinesize, srcU, s->uvlinesize,
8 + 1, 8 + 1, uvsrc_x, uvsrc_y,
s->vdsp.emulated_edge_mc(uvbuf, srcU,
s->uvlinesize, s->uvlinesize,
8 + 1, 8 + 1,
uvsrc_x, uvsrc_y,
s->h_edge_pos >> 1, v_edge_pos >> 1);
s->vdsp.emulated_edge_mc(uvbuf + 16, s->uvlinesize, srcV, s->uvlinesize,
8 + 1, 8 + 1, uvsrc_x, uvsrc_y,
s->vdsp.emulated_edge_mc(uvbuf + 16, srcV,
s->uvlinesize, s->uvlinesize,
8 + 1, 8 + 1,
uvsrc_x, uvsrc_y,
s->h_edge_pos >> 1, v_edge_pos >> 1);
srcU = uvbuf;
srcV = uvbuf + 16;


+ 9
- 6
libavcodec/videodsp.h View File

@@ -30,8 +30,8 @@
#include <stdint.h>

#define EMULATED_EDGE(depth) \
void ff_emulated_edge_mc_ ## depth(uint8_t *dst, ptrdiff_t dst_stride, \
const uint8_t *src, ptrdiff_t src_stride, \
void ff_emulated_edge_mc_ ## depth(uint8_t *dst, const uint8_t *src, \
ptrdiff_t dst_stride, ptrdiff_t src_stride, \
int block_w, int block_h,\
int src_x, int src_y, int w, int h);

@@ -47,8 +47,10 @@ typedef struct VideoDSPContext {
* @param dst_stride number of bytes between 2 vertically adjacent samples
* in destination buffer
* @param src source buffer
* @param src_stride number of bytes between 2 vertically adjacent samples
* in source buffer
* @param dst_linesize number of bytes between 2 vertically adjacent
* samples in the destination buffer
* @param src_linesize number of bytes between 2 vertically adjacent
* samples in both the source buffer
* @param block_w width of block
* @param block_h height of block
* @param src_x x coordinate of the top left sample of the block in the
@@ -58,8 +60,9 @@ typedef struct VideoDSPContext {
* @param w width of the source buffer
* @param h height of the source buffer
*/
void (*emulated_edge_mc)(uint8_t *dst, ptrdiff_t dst_stride,
const uint8_t *src, ptrdiff_t src_stride,
void (*emulated_edge_mc)(uint8_t *dst, const uint8_t *src,
ptrdiff_t dst_linesize,
ptrdiff_t src_linesize,
int block_w, int block_h,
int src_x, int src_y, int w, int h);



+ 15
- 14
libavcodec/videodsp_template.c View File

@@ -20,8 +20,9 @@
*/

#include "bit_depth_template.c"
void FUNC(ff_emulated_edge_mc)(uint8_t *buf, ptrdiff_t buf_stride,
const uint8_t *src, ptrdiff_t src_stride,
void FUNC(ff_emulated_edge_mc)(uint8_t *buf, const uint8_t *src,
ptrdiff_t buf_linesize,
ptrdiff_t src_linesize,
int block_w, int block_h,
int src_x, int src_y, int w, int h)
{
@@ -32,12 +33,12 @@ void FUNC(ff_emulated_edge_mc)(uint8_t *buf, ptrdiff_t buf_stride,
return;

if (src_y >= h) {
src -= src_y * src_stride;
src += (h - 1) * src_stride;
src -= src_y * src_linesize;
src += (h - 1) * src_linesize;
src_y = h - 1;
} else if (src_y <= -block_h) {
src -= src_y * src_stride;
src += (1 - block_h) * src_stride;
src -= src_y * src_linesize;
src += (1 - block_h) * src_linesize;
src_y = 1 - block_h;
}
if (src_x >= w) {
@@ -56,30 +57,30 @@ void FUNC(ff_emulated_edge_mc)(uint8_t *buf, ptrdiff_t buf_stride,
av_assert2(start_x < end_x && block_w);

w = end_x - start_x;
src += start_y * src_stride + start_x * sizeof(pixel);
src += start_y * src_linesize + start_x * sizeof(pixel);
buf += start_x * sizeof(pixel);

// top
for (y = 0; y < start_y; y++) {
memcpy(buf, src, w * sizeof(pixel));
buf += buf_stride;
buf += buf_linesize;
}

// copy existing part
for (; y < end_y; y++) {
memcpy(buf, src, w * sizeof(pixel));
src += src_stride;
buf += buf_stride;
src += src_linesize;
buf += buf_linesize;
}

// bottom
src -= src_stride;
src -= src_linesize;
for (; y < block_h; y++) {
memcpy(buf, src, w * sizeof(pixel));
buf += buf_stride;
buf += buf_linesize;
}

buf -= block_h * buf_stride + start_x * sizeof(pixel);
buf -= block_h * buf_linesize + start_x * sizeof(pixel);
while (block_h--) {
pixel *bufp = (pixel *) buf;

@@ -92,6 +93,6 @@ void FUNC(ff_emulated_edge_mc)(uint8_t *buf, ptrdiff_t buf_stride,
for (x = end_x; x < block_w; x++) {
bufp[x] = bufp[end_x - 1];
}
buf += buf_stride;
buf += buf_linesize;
}
}

+ 4
- 3
libavcodec/vp3.c View File

@@ -1549,10 +1549,11 @@ static void render_slice(Vp3DecodeContext *s, int slice)
uint8_t *temp= s->edge_emu_buffer;
if(stride<0) temp -= 8*stride;

s->vdsp.emulated_edge_mc(temp, stride,
motion_source, stride,
s->vdsp.emulated_edge_mc(temp, motion_source,
stride, stride,
9, 9, src_x, src_y,
plane_width, plane_height);
plane_width,
plane_height);
motion_source= temp;
}
}


+ 6
- 5
libavcodec/vp56.c View File

@@ -339,11 +339,12 @@ static void vp56_mc(VP56Context *s, int b, int plane, uint8_t *src,

if (x<0 || x+12>=s->plane_width[plane] ||
y<0 || y+12>=s->plane_height[plane]) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, stride,
src + s->block_offset[b] + (dy-2)*stride + (dx-2),
stride, 12, 12, x, y,
s->plane_width[plane],
s->plane_height[plane]);
s->vdsp.emulated_edge_mc(s->edge_emu_buffer,
src + s->block_offset[b] + (dy-2)*stride + (dx-2),
stride, stride,
12, 12, x, y,
s->plane_width[plane],
s->plane_height[plane]);
src_block = s->edge_emu_buffer;
src_offset = 2 + 2*stride;
} else if (deblock_filtering) {


+ 9
- 6
libavcodec/vp8.c View File

@@ -1196,8 +1196,9 @@ void vp8_mc_luma(VP8Context *s, VP8ThreadData *td, uint8_t *dst,
src += y_off * linesize + x_off;
if (x_off < mx_idx || x_off >= width - block_w - subpel_idx[2][mx] ||
y_off < my_idx || y_off >= height - block_h - subpel_idx[2][my]) {
s->vdsp.emulated_edge_mc(td->edge_emu_buffer, 32,
src - my_idx * linesize - mx_idx, linesize,
s->vdsp.emulated_edge_mc(td->edge_emu_buffer,
src - my_idx * linesize - mx_idx,
32, linesize,
block_w + subpel_idx[1][mx],
block_h + subpel_idx[1][my],
x_off - mx_idx, y_off - my_idx, width, height);
@@ -1249,16 +1250,18 @@ void vp8_mc_chroma(VP8Context *s, VP8ThreadData *td, uint8_t *dst1, uint8_t *dst
ff_thread_await_progress(ref, (3 + y_off + block_h + subpel_idx[2][my]) >> 3, 0);
if (x_off < mx_idx || x_off >= width - block_w - subpel_idx[2][mx] ||
y_off < my_idx || y_off >= height - block_h - subpel_idx[2][my]) {
s->vdsp.emulated_edge_mc(td->edge_emu_buffer, 32,
src1 - my_idx * linesize - mx_idx, linesize,
s->vdsp.emulated_edge_mc(td->edge_emu_buffer,
src1 - my_idx * linesize - mx_idx,
32, linesize,
block_w + subpel_idx[1][mx],
block_h + subpel_idx[1][my],
x_off - mx_idx, y_off - my_idx, width, height);
src1 = td->edge_emu_buffer + mx_idx + 32 * my_idx;
mc_func[my_idx][mx_idx](dst1, linesize, src1, 32, block_h, mx, my);

s->vdsp.emulated_edge_mc(td->edge_emu_buffer, 32,
src2 - my_idx * linesize - mx_idx, linesize,
s->vdsp.emulated_edge_mc(td->edge_emu_buffer,
src2 - my_idx * linesize - mx_idx,
32, linesize,
block_w + subpel_idx[1][mx],
block_h + subpel_idx[1][my],
x_off - mx_idx, y_off - my_idx, width, height);


+ 8
- 6
libavcodec/vp9.c View File

@@ -2223,9 +2223,9 @@ static av_always_inline void mc_luma_dir(VP9Context *s, vp9_mc_func (*mc)[2],
// FIXME bilinear filter only needs 0/1 pixels, not 3/4
if (x < !!mx * 3 || y < !!my * 3 ||
x + !!mx * 4 > w - bw || y + !!my * 4 > h - bh) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, 80,
s->vdsp.emulated_edge_mc(s->edge_emu_buffer,
ref - !!my * 3 * ref_stride - !!mx * 3,
ref_stride,
80, ref_stride,
bw + !!mx * 7, bh + !!my * 7,
x - !!mx * 3, y - !!my * 3, w, h);
ref = s->edge_emu_buffer + !!my * 3 * 80 + !!mx * 3;
@@ -2253,15 +2253,17 @@ static av_always_inline void mc_chroma_dir(VP9Context *s, vp9_mc_func (*mc)[2],
// FIXME bilinear filter only needs 0/1 pixels, not 3/4
if (x < !!mx * 3 || y < !!my * 3 ||
x + !!mx * 4 > w - bw || y + !!my * 4 > h - bh) {
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, 80,
ref_u - !!my * 3 * src_stride_u - !!mx * 3, src_stride_u,
s->vdsp.emulated_edge_mc(s->edge_emu_buffer,
ref_u - !!my * 3 * src_stride_u - !!mx * 3,
80, src_stride_u,
bw + !!mx * 7, bh + !!my * 7,
x - !!mx * 3, y - !!my * 3, w, h);
ref_u = s->edge_emu_buffer + !!my * 3 * 80 + !!mx * 3;
mc[!!mx][!!my](dst_u, dst_stride, ref_u, 80, bh, mx, my);

s->vdsp.emulated_edge_mc(s->edge_emu_buffer, 80,
ref_v - !!my * 3 * src_stride_v - !!mx * 3, src_stride_v,
s->vdsp.emulated_edge_mc(s->edge_emu_buffer,
ref_v - !!my * 3 * src_stride_v - !!mx * 3,
80, src_stride_v,
bw + !!mx * 7, bh + !!my * 7,
x - !!mx * 3, y - !!my * 3, w, h);
ref_v = s->edge_emu_buffer + !!my * 3 * 80 + !!mx * 3;


+ 15
- 9
libavcodec/wmv2.c View File

@@ -119,9 +119,11 @@ void ff_mspel_motion(MpegEncContext *s,

if(src_x<1 || src_y<1 || src_x + 17 >= s->h_edge_pos
|| src_y + h+1 >= v_edge_pos){
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, s->linesize,
ptr - 1 - s->linesize, s->linesize, 19, 19,
src_x-1, src_y-1,
s->vdsp.emulated_edge_mc(s->edge_emu_buffer,
ptr - 1 - s->linesize,
s->linesize, s->linesize,
19, 19,
src_x - 1, src_y - 1,
s->h_edge_pos, s->v_edge_pos);
ptr= s->edge_emu_buffer + 1 + s->linesize;
emu=1;
@@ -161,18 +163,22 @@ void ff_mspel_motion(MpegEncContext *s,
offset = (src_y * uvlinesize) + src_x;
ptr = ref_picture[1] + offset;
if(emu){
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, s->uvlinesize,
ptr, s->uvlinesize, 9, 9, src_x, src_y,
s->h_edge_pos>>1, s->v_edge_pos>>1);
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, ptr,
s->uvlinesize, s->uvlinesize,
9, 9,
src_x, src_y,
s->h_edge_pos >> 1, s->v_edge_pos >> 1);
ptr= s->edge_emu_buffer;
}
pix_op[1][dxy](dest_cb, ptr, uvlinesize, h >> 1);

ptr = ref_picture[2] + offset;
if(emu){
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, s->uvlinesize,
ptr, s->uvlinesize, 9, 9, src_x, src_y,
s->h_edge_pos>>1, s->v_edge_pos>>1);
s->vdsp.emulated_edge_mc(s->edge_emu_buffer, ptr,
s->uvlinesize, s->uvlinesize,
9, 9,
src_x, src_y,
s->h_edge_pos >> 1, s->v_edge_pos >> 1);
ptr= s->edge_emu_buffer;
}
pix_op[1][dxy](dest_cr, ptr, uvlinesize, h >> 1);


+ 4
- 3
libavcodec/x86/dsputil_mmx.c View File

@@ -367,8 +367,9 @@ void ff_draw_edges_mmx(uint8_t *buf, int wrap, int width, int height,
}
}

typedef void emulated_edge_mc_func(uint8_t *dst, ptrdiff_t dst_stride,
const uint8_t *src, ptrdiff_t src_linesize,
typedef void emulated_edge_mc_func(uint8_t *dst, const uint8_t *src,
ptrdiff_t dst_stride,
ptrdiff_t src_linesize,
int block_w, int block_h,
int src_x, int src_y, int w, int h);

@@ -417,7 +418,7 @@ static av_always_inline void gmc(uint8_t *dst, uint8_t *src,

src += ix + iy * stride;
if (need_emu) {
emu_edge_fn(edge_buf, stride, src, stride, w + 1, h + 1, ix, iy, width, height);
emu_edge_fn(edge_buf, src, stride, stride, w + 1, h + 1, ix, iy, width, height);
src = edge_buf;
}



+ 17
- 13
libavcodec/x86/videodsp_init.c View File

@@ -129,8 +129,9 @@ static emu_edge_hfix_func *hfixtbl_sse2[11] = {
};
extern emu_edge_hvar_func ff_emu_edge_hvar_sse2;

static av_always_inline void emulated_edge_mc(uint8_t *dst, ptrdiff_t dst_stride,
const uint8_t *src, ptrdiff_t src_stride,
static av_always_inline void emulated_edge_mc(uint8_t *dst, const uint8_t *src,
ptrdiff_t dst_stride,
ptrdiff_t src_stride,
x86_reg block_w, x86_reg block_h,
x86_reg src_x, x86_reg src_y,
x86_reg w, x86_reg h,
@@ -141,7 +142,7 @@ static av_always_inline void emulated_edge_mc(uint8_t *dst, ptrdiff_t dst_stride
{
x86_reg start_y, start_x, end_y, end_x, src_y_add = 0, p;

if(!w || !h)
if (!w || !h)
return;

if (src_y >= h) {
@@ -203,34 +204,37 @@ static av_always_inline void emulated_edge_mc(uint8_t *dst, ptrdiff_t dst_stride
}

#if ARCH_X86_32
static av_noinline void emulated_edge_mc_mmx(uint8_t *buf, ptrdiff_t buf_stride,
const uint8_t *src, ptrdiff_t src_stride,
static av_noinline void emulated_edge_mc_mmx(uint8_t *buf, const uint8_t *src,
ptrdiff_t buf_stride,
ptrdiff_t src_stride,
int block_w, int block_h,
int src_x, int src_y, int w, int h)
{
emulated_edge_mc(buf, buf_stride, src, src_stride, block_w, block_h,
emulated_edge_mc(buf, src, buf_stride, src_stride, block_w, block_h,
src_x, src_y, w, h, vfixtbl_mmx, &ff_emu_edge_vvar_mmx,
hfixtbl_mmx, &ff_emu_edge_hvar_mmx);
}

static av_noinline void emulated_edge_mc_sse(uint8_t *buf, ptrdiff_t buf_stride,
const uint8_t *src, ptrdiff_t src_stride,
static av_noinline void emulated_edge_mc_sse(uint8_t *buf, const uint8_t *src,
ptrdiff_t buf_stride,
ptrdiff_t src_stride,
int block_w, int block_h,
int src_x, int src_y, int w, int h)
{
emulated_edge_mc(buf, buf_stride, src, src_stride, block_w, block_h,
emulated_edge_mc(buf, src, buf_stride, src_stride, block_w, block_h,
src_x, src_y, w, h, vfixtbl_sse, &ff_emu_edge_vvar_sse,
hfixtbl_mmx, &ff_emu_edge_hvar_mmx);
}
#endif

static av_noinline void emulated_edge_mc_sse2(uint8_t *buf, ptrdiff_t buf_stride,
const uint8_t *src, ptrdiff_t src_stride,
static av_noinline void emulated_edge_mc_sse2(uint8_t *buf, const uint8_t *src,
ptrdiff_t buf_stride,
ptrdiff_t src_stride,
int block_w, int block_h,
int src_x, int src_y, int w, int h)
{
emulated_edge_mc(buf, buf_stride, src, src_stride, block_w, block_h, src_x,
src_y, w, h, vfixtbl_sse, &ff_emu_edge_vvar_sse,
emulated_edge_mc(buf, src, buf_stride, src_stride, block_w, block_h,
src_x, src_y, w, h, vfixtbl_sse, &ff_emu_edge_vvar_sse,
hfixtbl_sse2, &ff_emu_edge_hvar_sse2);
}
#endif /* HAVE_YASM */


Write Preview
Loading…
Cancel
Save