Merge pull request #275 from trav90/ffvpx-resync

Resync ffvpx code with 3.4.2-release from upstream

1 files changed, 12 insertions, 9 deletions

diff --git a/media/ffvpx/libavcodec/x86/vp8dsp_init.c b/media/ffvpx/libavcodec/x86/vp8dsp_init.c
index 897d5a0e7..397b2518c 100644
--- a/media/ffvpx/libavcodec/x86/vp8dsp_init.c
+++ b/media/ffvpx/libavcodec/x86/vp8dsp_init.c
@@ -26,7 +26,7 @@
 #include "libavutil/x86/cpu.h"
 #include "libavcodec/vp8dsp.h"
 
-#if HAVE_YASM
+#if HAVE_X86ASM
 
 /*
  * MC functions
@@ -233,6 +233,8 @@ HVBILIN(ssse3, 8, 16, 16)
 
 void ff_vp8_idct_dc_add_mmx(uint8_t *dst, int16_t block[16],
                             ptrdiff_t stride);
+void ff_vp8_idct_dc_add_sse2(uint8_t *dst, int16_t block[16],
+                             ptrdiff_t stride);
 void ff_vp8_idct_dc_add_sse4(uint8_t *dst, int16_t block[16],
                              ptrdiff_t stride);
 void ff_vp8_idct_dc_add4y_mmx(uint8_t *dst, int16_t block[4][16],
@@ -288,7 +290,7 @@ DECLARE_LOOP_FILTER(sse2)
 DECLARE_LOOP_FILTER(ssse3)
 DECLARE_LOOP_FILTER(sse4)
 
-#endif /* HAVE_YASM */
+#endif /* HAVE_X86ASM */
 
 #define VP8_LUMA_MC_FUNC(IDX, SIZE, OPT) \
     c->put_vp8_epel_pixels_tab[IDX][0][2] = ff_put_vp8_epel ## SIZE ## _h6_ ## OPT; \
@@ -316,7 +318,7 @@ DECLARE_LOOP_FILTER(sse4)
 
 av_cold void ff_vp78dsp_init_x86(VP8DSPContext *c)
 {
-#if HAVE_YASM
+#if HAVE_X86ASM
     int cpu_flags = av_get_cpu_flags();
 
     if (EXTERNAL_MMX(cpu_flags)) {
@@ -346,7 +348,7 @@ av_cold void ff_vp78dsp_init_x86(VP8DSPContext *c)
         c->put_vp8_bilinear_pixels_tab[0][0][0] = ff_put_vp8_pixels16_sse;
     }
 
-    if (HAVE_SSE2_EXTERNAL && cpu_flags & (AV_CPU_FLAG_SSE2 | AV_CPU_FLAG_SSE2SLOW)) {
+    if (EXTERNAL_SSE2(cpu_flags) || EXTERNAL_SSE2_SLOW(cpu_flags)) {
         VP8_LUMA_MC_FUNC(0, 16, sse2);
         VP8_MC_FUNC(1, 8, sse2);
         VP8_BILINEAR_MC_FUNC(0, 16, sse2);
@@ -361,18 +363,18 @@ av_cold void ff_vp78dsp_init_x86(VP8DSPContext *c)
         VP8_BILINEAR_MC_FUNC(1, 8, ssse3);
         VP8_BILINEAR_MC_FUNC(2, 4, ssse3);
     }
-#endif /* HAVE_YASM */
+#endif /* HAVE_X86ASM */
 }
 
 av_cold void ff_vp8dsp_init_x86(VP8DSPContext *c)
 {
-#if HAVE_YASM
+#if HAVE_X86ASM
     int cpu_flags = av_get_cpu_flags();
 
     if (EXTERNAL_MMX(cpu_flags)) {
-        c->vp8_idct_dc_add    = ff_vp8_idct_dc_add_mmx;
         c->vp8_idct_dc_add4uv = ff_vp8_idct_dc_add4uv_mmx;
 #if ARCH_X86_32
+        c->vp8_idct_dc_add    = ff_vp8_idct_dc_add_mmx;
         c->vp8_idct_dc_add4y  = ff_vp8_idct_dc_add4y_mmx;
         c->vp8_idct_add       = ff_vp8_idct_add_mmx;
         c->vp8_luma_dc_wht    = ff_vp8_luma_dc_wht_mmx;
@@ -416,7 +418,7 @@ av_cold void ff_vp8dsp_init_x86(VP8DSPContext *c)
         c->vp8_luma_dc_wht                      = ff_vp8_luma_dc_wht_sse;
     }
 
-    if (HAVE_SSE2_EXTERNAL && cpu_flags & (AV_CPU_FLAG_SSE2 | AV_CPU_FLAG_SSE2SLOW)) {
+    if (EXTERNAL_SSE2(cpu_flags) || EXTERNAL_SSE2_SLOW(cpu_flags)) {
         c->vp8_v_loop_filter_simple = ff_vp8_v_loop_filter_simple_sse2;
 
         c->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16y_inner_sse2;
@@ -427,6 +429,7 @@ av_cold void ff_vp8dsp_init_x86(VP8DSPContext *c)
     }
 
     if (EXTERNAL_SSE2(cpu_flags)) {
+        c->vp8_idct_dc_add            = ff_vp8_idct_dc_add_sse2;
         c->vp8_idct_dc_add4y          = ff_vp8_idct_dc_add4y_sse2;
 
         c->vp8_h_loop_filter_simple   = ff_vp8_h_loop_filter_simple_sse2;
@@ -460,5 +463,5 @@ av_cold void ff_vp8dsp_init_x86(VP8DSPContext *c)
         c->vp8_h_loop_filter16y       = ff_vp8_h_loop_filter16y_mbedge_sse4;
         c->vp8_h_loop_filter8uv       = ff_vp8_h_loop_filter8uv_mbedge_sse4;
     }
-#endif /* HAVE_YASM */
+#endif /* HAVE_X86ASM */
 }