[FFmpeg-devel] [PATCH 4/4] avcodec/utils: Optimize ff_color_frame() using memcpy()

[Michael Niedermayer]

4650975 -> 4493240 dezicycles

This optimizes lines 2 and later. Line 1 still uses av_memcpy_backptr()
This change originally fixed ossfuzz 10790 but this is now fixed by other
optimizations already

Signed-off-by: Michael Niedermayer <michael at niedermayer.cc>
---
 libavcodec/utils.c | 18 ++++++++++++------
 1 file changed, 12 insertions(+), 6 deletions(-)

diff --git a/libavcodec/utils.c b/libavcodec/utils.c
index 132f887343..1ac13f6bc3 100644
--- a/libavcodec/utils.c
+++ b/libavcodec/utils.c
@@ -421,13 +421,19 @@ void ff_color_frame(AVFrame *frame, const int c[4])
         int is_chroma = p == 1 || p == 2;
         int bytes  = is_chroma ? AV_CEIL_RSHIFT(frame->width,  desc->log2_chroma_w) : frame->width;
         int height = is_chroma ? AV_CEIL_RSHIFT(frame->height, desc->log2_chroma_h) : frame->height;
-        for (y = 0; y < height; y++) {
-            if (desc->comp[0].depth >= 9) {
-                ((uint16_t*)dst)[0] = c[p];
-                av_memcpy_backptr(dst + 2, 2, bytes - 2);
-            }else
-                memset(dst, c[p], bytes);
+        if (desc->comp[0].depth >= 9) {
+            ((uint16_t*)dst)[0] = c[p];
+            av_memcpy_backptr(dst + 2, 2, bytes - 2);
             dst += frame->linesize[p];
+            for (y = 1; y < height; y++) {
+                memcpy(dst, frame->data[p], 2*bytes);
+                dst += frame->linesize[p];
+            }
+        } else {
+            for (y = 0; y < height; y++) {
+                memset(dst, c[p], bytes);
+                dst += frame->linesize[p];
+            }
         }
     }
 }
-- 
2.23.0