Merge commit 'de41b555cdea2dcacbe98ee9edc83a8c15c73c4c'

[ffmpeg] / libavcodec / x86 / vc1dsp_init.c

diff --git a/libavcodec/x86/vc1dsp_init.c b/libavcodec/x86/vc1dsp_init.c
index 2bef5f5fb59341549d42a7a39b4f5b2c9561c7c0..1747305d1fb36680ab8701fc5362724f4e87e627 100644 (file)
--- a/libavcodec/x86/vc1dsp_init.c
+++ b/libavcodec/x86/vc1dsp_init.c
@@ -63,16 +63,22 @@ 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);
-}
-static void avg_vc1_mspel_mc00_16_sse2(uint8_t *dst, const uint8_t *src,
-                                       ptrdiff_t stride, int rnd)
-{
-    ff_avg_pixels16_sse2(dst, src, stride, 16);
-}
+
+#define DECLARE_FUNCTION(OP, DEPTH, INSN)                       \
+    static void OP##vc1_mspel_mc00_##DEPTH##INSN(uint8_t *dst,          \
+                             const uint8_t *src, ptrdiff_t stride, int rnd) \
+    {                                                                       \
+        ff_ ## OP ## pixels ## DEPTH ## INSN(dst, src, stride, DEPTH);     \
+    }
+
+DECLARE_FUNCTION(put_,  8, _mmx)
+DECLARE_FUNCTION(put_, 16, _mmx)
+DECLARE_FUNCTION(avg_,  8, _mmx)
+DECLARE_FUNCTION(avg_, 16, _mmx)
+DECLARE_FUNCTION(avg_,  8, _mmxext)
+DECLARE_FUNCTION(avg_, 16, _mmxext)
+DECLARE_FUNCTION(put_, 16, _sse2)
+DECLARE_FUNCTION(avg_, 16, _sse2)
 
 #endif /* HAVE_YASM */
 
@@ -92,10 +98,10 @@ av_cold void ff_vc1dsp_init_x86(VC1DSPContext *dsp)
 {
     int cpu_flags = av_get_cpu_flags();
 
-    if (HAVE_6REGS && INLINE_MMX(cpu_flags))
+    if (HAVE_6REGS && INLINE_MMX(cpu_flags) && EXTERNAL_MMX(cpu_flags))
         ff_vc1dsp_init_mmx(dsp);
 
-    if (HAVE_6REGS && INLINE_MMXEXT(cpu_flags))
+    if (HAVE_6REGS && INLINE_MMXEXT(cpu_flags) && EXTERNAL_MMXEXT(cpu_flags))
         ff_vc1dsp_init_mmxext(dsp);
 
 #define ASSIGN_LF(EXT) \
@@ -109,6 +115,11 @@ av_cold void ff_vc1dsp_init_x86(VC1DSPContext *dsp)
 #if HAVE_YASM
     if (EXTERNAL_MMX(cpu_flags)) {
         dsp->put_no_rnd_vc1_chroma_pixels_tab[0] = ff_put_vc1_chroma_mc8_nornd_mmx;
+
+        dsp->put_vc1_mspel_pixels_tab[1][0]      = put_vc1_mspel_mc00_8_mmx;
+        dsp->put_vc1_mspel_pixels_tab[0][0]      = put_vc1_mspel_mc00_16_mmx;
+        dsp->avg_vc1_mspel_pixels_tab[1][0]      = avg_vc1_mspel_mc00_8_mmx;
+        dsp->avg_vc1_mspel_pixels_tab[0][0]      = avg_vc1_mspel_mc00_16_mmx;
     }
     if (EXTERNAL_AMD3DNOW(cpu_flags)) {
         dsp->avg_no_rnd_vc1_chroma_pixels_tab[0] = ff_avg_vc1_chroma_mc8_nornd_3dnow;
@@ -117,13 +128,16 @@ av_cold void ff_vc1dsp_init_x86(VC1DSPContext *dsp)
         ASSIGN_LF(mmxext);
         dsp->avg_no_rnd_vc1_chroma_pixels_tab[0] = ff_avg_vc1_chroma_mc8_nornd_mmxext;
 
-        dsp->avg_vc1_mspel_pixels_tab[1][0]      = avg_vc1_mspel_mc00_mmxext;
+        dsp->avg_vc1_mspel_pixels_tab[1][0]      = avg_vc1_mspel_mc00_8_mmxext;
+        dsp->avg_vc1_mspel_pixels_tab[0][0]      = avg_vc1_mspel_mc00_16_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;
+
+        dsp->put_vc1_mspel_pixels_tab[0][0]      = put_vc1_mspel_mc00_16_sse2;
         dsp->avg_vc1_mspel_pixels_tab[0][0]      = avg_vc1_mspel_mc00_16_sse2;
     }
     if (EXTERNAL_SSSE3(cpu_flags)) {