[FFmpeg-cvslog] gmc_mmx: enable also for large pictures when emu edge isnt needed.

[Michael Niedermayer]

ffmpeg | branch: master | Michael Niedermayer <michaelni at gmx.at> | Sat Oct  6 23:12:57 2012 +0200| [19a4e103601919d5a385dba9e4a05e1834850259] | committer: Michael Niedermayer

gmc_mmx: enable also for large pictures when emu edge isnt needed.

Signed-off-by: Michael Niedermayer <michaelni at gmx.at>

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=19a4e103601919d5a385dba9e4a05e1834850259
---

 libavcodec/x86/dsputil_mmx.c |    8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/libavcodec/x86/dsputil_mmx.c b/libavcodec/x86/dsputil_mmx.c
index 27e732f..29faf1e 100644
--- a/libavcodec/x86/dsputil_mmx.c
+++ b/libavcodec/x86/dsputil_mmx.c
@@ -1961,12 +1961,15 @@ static av_always_inline void gmc(uint8_t *dst, uint8_t *src,
     const int dyh = (dyy - (1 << (16 + shift))) * (h - 1);
     const int dxh = dxy * (h - 1);
     const int dyw = dyx * (w - 1);
+    int need_emu =  (unsigned)ix >= width  - w ||
+                    (unsigned)iy >= height - h;
+
     if ( // non-constant fullpel offset (3% of blocks)
         ((ox ^ (ox + dxw)) | (ox ^ (ox + dxh)) | (ox ^ (ox + dxw + dxh)) |
          (oy ^ (oy + dyw)) | (oy ^ (oy + dyh)) | (oy ^ (oy + dyw + dyh))) >> (16 + shift)
         // uses more than 16 bits of subpel mv (only at huge resolution)
         || (dxx | dxy | dyx | dyy) & 15
-        || h > MAX_H || stride > MAX_STRIDE) {
+        || (need_emu && (h > MAX_H || stride > MAX_STRIDE))) {
         // FIXME could still use mmx for some of the rows
         ff_gmc_c(dst, src, stride, h, ox, oy, dxx, dxy, dyx, dyy,
                  shift, r, width, height);
@@ -1974,8 +1977,7 @@ static av_always_inline void gmc(uint8_t *dst, uint8_t *src,
     }
 
     src += ix + iy * stride;
-    if ((unsigned)ix >= width  - w ||
-        (unsigned)iy >= height - h) {
+    if (need_emu) {
         emu_edge_fn(edge_buf, src, stride, w + 1, h + 1, ix, iy, width, height);
         src = edge_buf;
     }