Merge branch 'ronalds_buggy_gt8bit_sws'

[coroid/ffmpeg_saccubus.git] / libswscale / utils.c

diff --git a/libswscale/utils.c b/libswscale/utils.c
index bff9f4d..1879319 100644 (file)
--- a/libswscale/utils.c
+++ b/libswscale/utils.c
@@ -46,6 +46,7 @@
 #include "libavutil/bswap.h"
 #include "libavutil/opt.h"
 #include "libavutil/pixdesc.h"
+#include "libavutil/avassert.h"
 
 unsigned swscale_version(void)
 {
@@ -882,8 +883,15 @@ int sws_init_context(SwsContext *c, SwsFilter *srcFilter, SwsFilter *dstFilter)
         }
     }
 
+    // FIXME it's even nicer if bpp isn't 16, but max({src,dst}formatbpp)
+    c->scalingBpp = FFMAX(av_pix_fmt_descriptors[srcFormat].comp[0].depth_minus1,
+                          av_pix_fmt_descriptors[dstFormat].comp[0].depth_minus1) >= 15 ? 16 : 8;
+
+    if (c->scalingBpp == 16)
+        dst_stride <<= 1;
+    av_assert0(c->scalingBpp<=16);
     FF_ALLOC_OR_GOTO(c, c->formatConvBuffer, FFALIGN(srcW*2+78, 16) * 2, fail);
-    if (HAVE_MMX2 && cpu_flags & AV_CPU_FLAG_MMX2) {
+    if (HAVE_MMX2 && cpu_flags & AV_CPU_FLAG_MMX2 && c->scalingBpp == 8) {
         c->canMMX2BeUsed= (dstW >=srcW && (dstW&31)==0 && (srcW&15)==0) ? 1 : 0;
         if (!c->canMMX2BeUsed && dstW >=srcW && (srcW&15)==0 && (flags&SWS_FAST_BILINEAR)) {
             if (flags&SWS_PRINT_INFO)