Merge "convert copy16x16 to rtcd" into experimental

vp8/common/generic/systemdependent.c

@@ -36,7 +36,6 @@ void vp8_machine_specific_config(VP8_COMMON *ctx) {
 #if CONFIG_TX16X16 || CONFIG_HYBRIDTRANSFORM16X16
   rtcd->idct.idct16x16    = vp8_short_idct16x16_c;
 #endif
-  rtcd->recon.copy16x16   = vp8_copy_mem16x16_c;
   rtcd->recon.copy8x8     = vp8_copy_mem8x8_c;
   rtcd->recon.avg16x16    = vp8_avg_mem16x16_c;
   rtcd->recon.avg8x8      = vp8_avg_mem8x8_c;

vp8/common/onyxc_int.h

@@ -14,6 +14,7 @@
 
 #include "vpx_config.h"
 #include "vpx/internal/vpx_codec_internal.h"
+#include "vpx_rtcd.h"
 #include "loopfilter.h"
 #include "entropymv.h"
 #include "entropy.h"

vp8/common/recon.h

@@ -44,11 +44,6 @@ struct vp8_recon_rtcd_vtable;
 #include "arm/recon_arm.h"
 #endif
 
-#ifndef vp8_recon_copy16x16
-#define vp8_recon_copy16x16 vp8_copy_mem16x16_c
-#endif
-extern prototype_copy_block(vp8_recon_copy16x16);
-
 #ifndef vp8_recon_copy8x8
 #define vp8_recon_copy8x8 vp8_copy_mem8x8_c
 #endif

vp8/common/reconinter.c

@@ -662,8 +662,7 @@ void vp8_build_1st_inter16x16_predictors_mby(MACROBLOCKD *xd,
                                 (ymv.as_mv.row & 7) << 1,
                                 dst_y, dst_ystride);
     } else {
-      RECON_INVOKE(&xd->rtcd->recon, copy16x16)
-          (ptr, pre_stride, dst_y, dst_ystride);
+      vp8_copy_mem16x16(ptr, pre_stride, dst_y, dst_ystride);
     }
 }
 

vp8/common/rtcd_defs.sh

@@ -5,6 +5,8 @@ EOF
 }
 forward_decls common_forward_decls
 
+
+
 prototype void vp8_filter_block2d_4x4_8 "const unsigned char *src_ptr, const unsigned int src_stride, const short *HFilter_aligned16, const short *VFilter_aligned16, unsigned char *dst_ptr, unsigned int dst_stride"
 prototype void vp8_filter_block2d_8x4_8 "const unsigned char *src_ptr, const unsigned int src_stride, const short *HFilter_aligned16, const short *VFilter_aligned16, unsigned char *dst_ptr, unsigned int dst_stride"
 prototype void vp8_filter_block2d_8x8_8 "const unsigned char *src_ptr, const unsigned int src_stride, const short *HFilter_aligned16, const short *VFilter_aligned16, unsigned char *dst_ptr, unsigned int dst_stride"
@@ -19,3 +21,12 @@ if [ "$CONFIG_GCC" = "yes" ]; then
     specialize vp8_filter_block2d_8x8_8 sse4_1 sse2
     specialize vp8_filter_block2d_16x16_8 sse4_1 sse2
 fi
+
+
+#
+# RECON
+#
+prototype void vp8_copy_mem16x16 "unsigned char *src, int src_pitch, unsigned char *dst, int dst_pitch"
+specialize vp8_copy_mem16x16 mmx sse2 media neon dspr2
+vp8_copy_mem16x16_media=vp8_copy_mem16x16_v6
+vp8_copy_mem16x16_dspr2=vp8_copy_mem16x16_dspr2

vp8/common/x86/recon_x86.h

@@ -36,9 +36,6 @@ extern prototype_copy_block(vp8_copy_mem16x16_mmx);
 #undef  vp8_recon_copy8x4
 #define vp8_recon_copy8x4 vp8_copy_mem8x4_mmx
 
-#undef  vp8_recon_copy16x16
-#define vp8_recon_copy16x16 vp8_copy_mem16x16_mmx
-
 #endif
 #endif
 

vp8/common/x86/x86_systemdependent.c

@@ -44,7 +44,6 @@ void vp8_arch_x86_common_init(VP8_COMMON *ctx) {
     rtcd->recon.recon       = vp8_recon_b_mmx;
     rtcd->recon.copy8x8     = vp8_copy_mem8x8_mmx;
     rtcd->recon.copy8x4     = vp8_copy_mem8x4_mmx;
-    rtcd->recon.copy16x16   = vp8_copy_mem16x16_mmx;
 
     /* Disabled due to unsupported enhanced interpolation/high_prec mv
     rtcd->subpix.sixtap16x16   = vp8_sixtap_predict16x16_mmx;
@@ -80,7 +79,6 @@ void vp8_arch_x86_common_init(VP8_COMMON *ctx) {
   if (flags & HAS_SSE2) {
     rtcd->recon.recon2      = vp8_recon2b_sse2;
     rtcd->recon.recon4      = vp8_recon4b_sse2;
-    rtcd->recon.copy16x16   = vp8_copy_mem16x16_sse2;
 
     /* these are disable because of unsupported diagonal pred modes
     rtcd->recon.build_intra_predictors_mbuv =

vp8/encoder/encodeframe.c

@@ -309,9 +309,7 @@ static void build_activity_map(VP8_COMP *cpi) {
       recon_yoffset += 16;
 #endif
       // Copy current mb to a buffer
-      RECON_INVOKE(&xd->rtcd->recon, copy16x16)(x->src.y_buffer,
-                                                x->src.y_stride,
-                                                x->thismb, 16);
+      vp8_copy_mem16x16(x->src.y_buffer, x->src.y_stride, x->thismb, 16);
 
       // measure activity
       mb_activity = mb_activity_measure(cpi, x, mb_row, mb_col);
@@ -591,9 +589,7 @@ static void pick_mb_modes(VP8_COMP *cpi,
     xd->dst.v_buffer = cm->yv12_fb[dst_fb_idx].v_buffer + recon_uvoffset;
 
     // Copy current MB to a work buffer
-    RECON_INVOKE(&xd->rtcd->recon, copy16x16)(x->src.y_buffer,
-                                              x->src.y_stride,
-                                              x->thismb, 16);
+    vp8_copy_mem16x16(x->src.y_buffer, x->src.y_stride, x->thismb, 16);
 
     x->rddiv = cpi->RDDIV;
     x->rdmult = cpi->RDMULT;
@@ -776,9 +772,7 @@ static void pick_sb_modes (VP8_COMP *cpi,
   xd->dst.v_buffer = cm->yv12_fb[dst_fb_idx].v_buffer + recon_uvoffset;
 #if 0 // FIXME
   /* Copy current MB to a work buffer */
-  RECON_INVOKE(&xd->rtcd->recon, copy16x16)(x->src.y_buffer,
-                                            x->src.y_stride,
-                                            x->thismb, 16);
+  vp8_copy_mem16x16(x->src.y_buffer, x->src.y_stride, x->thismb, 16);
 #endif
   x->rddiv = cpi->RDDIV;
   x->rdmult = cpi->RDMULT;
@@ -976,9 +970,7 @@ static void encode_sb(VP8_COMP *cpi,
     xd->dst.v_buffer = cm->yv12_fb[dst_fb_idx].v_buffer + recon_uvoffset;
 
     // Copy current MB to a work buffer
-    RECON_INVOKE(&xd->rtcd->recon, copy16x16)(x->src.y_buffer,
-                                              x->src.y_stride,
-                                              x->thismb, 16);
+    vp8_copy_mem16x16(x->src.y_buffer, x->src.y_stride, x->thismb, 16);
 
     if (cpi->oxcf.tuning == VP8_TUNE_SSIM)
       vp8_activity_masking(cpi, x);

vp8/encoder/firstpass.c

@@ -528,7 +528,7 @@ void vp8_first_pass(VP8_COMP *cpi) {
       xd->left_available = (mb_col != 0);
 
       // Copy current mb to a buffer
-      RECON_INVOKE(&xd->rtcd->recon, copy16x16)(x->src.y_buffer, x->src.y_stride, x->thismb, 16);
+      vp8_copy_mem16x16(x->src.y_buffer, x->src.y_stride, x->thismb, 16);
 
       // do intra 16x16 prediction
       this_error = vp8_encode_intra(cpi, x, use_dc_pred);

vp8/encoder/temporal_filter.c

@@ -59,7 +59,7 @@ static void vp8_temporal_filter_predictors_mb_c
     xd->subpixel_predict16x16(yptr, stride,
                              (mv_col & 7) << 1, (mv_row & 7) << 1, &pred[0], 16);
   } else {
-    RECON_INVOKE(&xd->rtcd->recon, copy16x16)(yptr, stride, &pred[0], 16);
+    vp8_copy_mem16x16(yptr, stride, &pred[0], 16);
   }
 
   // U & V