X-Git-Url: https://git.sesse.net/?a=blobdiff_plain;f=libavcodec%2Fx86%2Fvc1dsp_init.c;h=9bd5ae3666eaa78dc697e1cab4b37f5357ba9be9;hb=08e3ea60ff4059341b74be04a428a38f7c3630b0;hp=d2548fc348dabae237d70272f6eba1f54e9a331b;hpb=7e522859fc463a14875f78957bb57d2f33607d19;p=ffmpeg

diff --git a/libavcodec/x86/vc1dsp_init.c b/libavcodec/x86/vc1dsp_init.c
index d2548fc348d..9bd5ae3666e 100644
--- a/libavcodec/x86/vc1dsp_init.c
+++ b/libavcodec/x86/vc1dsp_init.c
@@ -24,9 +24,11 @@
  * OTHER DEALINGS IN THE SOFTWARE.
  */
 
+#include "libavutil/attributes.h"
 #include "libavutil/cpu.h"
 #include "libavutil/x86/cpu.h"
 #include "libavcodec/vc1dsp.h"
+#include "dsputil_x86.h"
 #include "vc1dsp.h"
 #include "config.h"
 
@@ -49,7 +51,7 @@ static void vc1_h_loop_filter16_ ## EXT(uint8_t *src, int stride, int pq) \
 }
 
 #if HAVE_YASM
-LOOP_FILTER(mmx2)
+LOOP_FILTER(mmxext)
 LOOP_FILTER(sse2)
 LOOP_FILTER(ssse3)
 
@@ -60,28 +62,34 @@ static void vc1_h_loop_filter16_sse4(uint8_t *src, int stride, int pq)
     ff_vc1_h_loop_filter8_sse4(src,          stride, pq);
     ff_vc1_h_loop_filter8_sse4(src+8*stride, stride, pq);
 }
+
+static void avg_vc1_mspel_mc00_mmxext(uint8_t *dst, const uint8_t *src,
+                                      ptrdiff_t stride, int rnd)
+{
+    ff_avg_pixels8_mmxext(dst, src, stride, 8);
+}
 #endif /* HAVE_YASM */
 
-void ff_put_vc1_chroma_mc8_mmx_nornd  (uint8_t *dst, uint8_t *src,
+void ff_put_vc1_chroma_mc8_nornd_mmx  (uint8_t *dst, uint8_t *src,
                                        int stride, int h, int x, int y);
-void ff_avg_vc1_chroma_mc8_mmx2_nornd (uint8_t *dst, uint8_t *src,
+void ff_avg_vc1_chroma_mc8_nornd_mmxext(uint8_t *dst, uint8_t *src,
+                                        int stride, int h, int x, int y);
+void ff_avg_vc1_chroma_mc8_nornd_3dnow(uint8_t *dst, uint8_t *src,
                                        int stride, int h, int x, int y);
-void ff_avg_vc1_chroma_mc8_3dnow_nornd(uint8_t *dst, uint8_t *src,
+void ff_put_vc1_chroma_mc8_nornd_ssse3(uint8_t *dst, uint8_t *src,
                                        int stride, int h, int x, int y);
-void ff_put_vc1_chroma_mc8_ssse3_nornd(uint8_t *dst, uint8_t *src,
-                                       int stride, int h, int x, int y);
-void ff_avg_vc1_chroma_mc8_ssse3_nornd(uint8_t *dst, uint8_t *src,
+void ff_avg_vc1_chroma_mc8_nornd_ssse3(uint8_t *dst, uint8_t *src,
                                        int stride, int h, int x, int y);
 
 
 av_cold void ff_vc1dsp_init_x86(VC1DSPContext *dsp)
 {
-    int mm_flags = av_get_cpu_flags();
+    int cpu_flags = av_get_cpu_flags();
 
-    if (INLINE_MMX(mm_flags))
+    if (INLINE_MMX(cpu_flags))
         ff_vc1dsp_init_mmx(dsp);
 
-    if (INLINE_MMXEXT(mm_flags))
+    if (INLINE_MMXEXT(cpu_flags))
         ff_vc1dsp_init_mmxext(dsp);
 
 #define ASSIGN_LF(EXT) \
@@ -93,29 +101,30 @@ av_cold void ff_vc1dsp_init_x86(VC1DSPContext *dsp)
         dsp->vc1_h_loop_filter16 = vc1_h_loop_filter16_ ## EXT
 
 #if HAVE_YASM
-    if (mm_flags & AV_CPU_FLAG_MMX) {
-        dsp->put_no_rnd_vc1_chroma_pixels_tab[0]= ff_put_vc1_chroma_mc8_mmx_nornd;
+    if (EXTERNAL_MMX(cpu_flags)) {
+        dsp->put_no_rnd_vc1_chroma_pixels_tab[0] = ff_put_vc1_chroma_mc8_nornd_mmx;
     }
-
-    if (mm_flags & AV_CPU_FLAG_MMXEXT) {
-        ASSIGN_LF(mmx2);
-        dsp->avg_no_rnd_vc1_chroma_pixels_tab[0]= ff_avg_vc1_chroma_mc8_mmx2_nornd;
-    } else if (mm_flags & AV_CPU_FLAG_3DNOW) {
-        dsp->avg_no_rnd_vc1_chroma_pixels_tab[0]= ff_avg_vc1_chroma_mc8_3dnow_nornd;
+    if (EXTERNAL_AMD3DNOW(cpu_flags)) {
+        dsp->avg_no_rnd_vc1_chroma_pixels_tab[0] = ff_avg_vc1_chroma_mc8_nornd_3dnow;
     }
+    if (EXTERNAL_MMXEXT(cpu_flags)) {
+        ASSIGN_LF(mmxext);
+        dsp->avg_no_rnd_vc1_chroma_pixels_tab[0] = ff_avg_vc1_chroma_mc8_nornd_mmxext;
 
-    if (mm_flags & AV_CPU_FLAG_SSE2) {
+        dsp->avg_vc1_mspel_pixels_tab[0]         = avg_vc1_mspel_mc00_mmxext;
+    }
+    if (EXTERNAL_SSE2(cpu_flags)) {
         dsp->vc1_v_loop_filter8  = ff_vc1_v_loop_filter8_sse2;
         dsp->vc1_h_loop_filter8  = ff_vc1_h_loop_filter8_sse2;
         dsp->vc1_v_loop_filter16 = vc1_v_loop_filter16_sse2;
         dsp->vc1_h_loop_filter16 = vc1_h_loop_filter16_sse2;
     }
-    if (mm_flags & AV_CPU_FLAG_SSSE3) {
+    if (EXTERNAL_SSSE3(cpu_flags)) {
         ASSIGN_LF(ssse3);
-        dsp->put_no_rnd_vc1_chroma_pixels_tab[0]= ff_put_vc1_chroma_mc8_ssse3_nornd;
-        dsp->avg_no_rnd_vc1_chroma_pixels_tab[0]= ff_avg_vc1_chroma_mc8_ssse3_nornd;
+        dsp->put_no_rnd_vc1_chroma_pixels_tab[0] = ff_put_vc1_chroma_mc8_nornd_ssse3;
+        dsp->avg_no_rnd_vc1_chroma_pixels_tab[0] = ff_avg_vc1_chroma_mc8_nornd_ssse3;
     }
-    if (mm_flags & AV_CPU_FLAG_SSE4) {
+    if (EXTERNAL_SSE4(cpu_flags)) {
         dsp->vc1_h_loop_filter8  = ff_vc1_h_loop_filter8_sse4;
         dsp->vc1_h_loop_filter16 = vc1_h_loop_filter16_sse4;
     }