23 files changed, 225 insertions, 138 deletions

diff --git a/build/make/Android.mk b/build/make/Android.mk
index 369c2a568..826ff2f81 100644
--- a/build/make/Android.mk
+++ b/build/make/Android.mk
@@ -61,6 +61,9 @@ ifeq ($(TARGET_ARCH_ABI),armeabi-v7a)
 else ifeq  ($(TARGET_ARCH_ABI),armeabi)
   include $(CONFIG_DIR)libs-armv5te-android-gcc.mk
   LOCAL_ARM_MODE := arm
+else ifeq  ($(TARGET_ARCH_ABI),arm64-v8a)
+  include $(CONFIG_DIR)libs-armv8-android-gcc.mk
+  LOCAL_ARM_MODE := arm
 else ifeq ($(TARGET_ARCH_ABI),x86)
   include $(CONFIG_DIR)libs-x86-android-gcc.mk
 else ifeq ($(TARGET_ARCH_ABI),mips)
@@ -126,7 +129,7 @@ endef
 ifeq ($(CONFIG_VP8_ENCODER), yes)
   ASM_CNV_OFFSETS_DEPEND += $(ASM_CNV_PATH)/vp8_asm_enc_offsets.asm
 endif
-ifeq ($(HAVE_NEON), yes)
+ifeq ($(HAVE_NEON_ASM), yes)
   ASM_CNV_OFFSETS_DEPEND += $(ASM_CNV_PATH)/vpx_scale_asm_offsets.asm
 endif
 
@@ -153,7 +156,11 @@ LOCAL_NEON_SRCS_C = $(filter %_neon.c, $(CODEC_SRCS_C))
 LOCAL_CODEC_SRCS_C = $(filter-out vpx_config.c %_neon.c, $(CODEC_SRCS_C))
 
 LOCAL_SRC_FILES += $(foreach file, $(LOCAL_CODEC_SRCS_C), libvpx/$(file))
-LOCAL_SRC_FILES += $(foreach file, $(LOCAL_NEON_SRCS_C), libvpx/$(file).neon)
+ifeq ($(TARGET_ARCH_ABI),armeabi-v7a)
+  LOCAL_SRC_FILES += $(foreach file, $(LOCAL_NEON_SRCS_C), libvpx/$(file).neon)
+else # If there are neon sources then we are building for arm64 and do not need to specify .neon
+  LOCAL_SRC_FILES += $(foreach file, $(LOCAL_NEON_SRCS_C), libvpx/$(file))
+endif
 
 # Pull out assembly files, splitting NEON from the rest.  This is
 # done to specify that the NEON assembly files use NEON assembler flags.
diff --git a/build/make/configure.sh b/build/make/configure.sh
index 4c3b05f66..a2a64c2d5 100755
--- a/build/make/configure.sh
+++ b/build/make/configure.sh
@@ -792,8 +792,12 @@ process_common_toolchain() {
     arm*)
         # on arm, isa versions are supersets
         case ${tgt_isa} in
+        armv8)
+            soft_enable neon
+            ;;
         armv7)
             soft_enable neon
+            soft_enable neon_asm
             soft_enable media
             soft_enable edsp
             soft_enable fast_unaligned
diff --git a/build/make/rtcd.pl b/build/make/rtcd.pl
index 18ee80d36..f5f59b146 100755
--- a/build/make/rtcd.pl
+++ b/build/make/rtcd.pl
@@ -272,6 +272,9 @@ sub arm() {
   # Assign the helper variable for each enabled extension
   foreach my $opt (@ALL_ARCHS) {
     my $opt_uc = uc $opt;
+    # Enable neon assembly based on HAVE_NEON logic instead of adding new
+    # HAVE_NEON_ASM logic
+    if ($opt eq 'neon_asm') { $opt_uc = 'NEON' }
     eval "\$have_${opt}=\"flags & HAS_${opt_uc}\"";
   }
 
@@ -381,7 +384,10 @@ if ($opts{arch} eq 'x86') {
   @ALL_ARCHS = filter(qw/edsp media/);
   arm;
 } elsif ($opts{arch} eq 'armv7') {
-  @ALL_ARCHS = filter(qw/edsp media neon/);
+  @ALL_ARCHS = filter(qw/edsp media neon_asm neon/);
+  arm;
+} elsif ($opts{arch} eq 'armv8') {
+  @ALL_ARCHS = filter(qw/neon/);
   arm;
 } else {
   unoptimized;
diff --git a/configure b/configure
index 690ac48fa..af3de7bb1 100755
--- a/configure
+++ b/configure
@@ -239,6 +239,7 @@ ARCH_EXT_LIST="
     edsp
     media
     neon
+    neon_asm
 
     mips32
     dspr2
diff --git a/test/convolve_test.cc b/test/convolve_test.cc
index 37ee0ef2a..cbb403677 100644
--- a/test/convolve_test.cc
+++ b/test/convolve_test.cc
@@ -634,7 +634,7 @@ INSTANTIATE_TEST_CASE_P(SSSE3, ConvolveTest, ::testing::Values(
     make_tuple(64, 64, &convolve8_ssse3)));
 #endif
 
-#if HAVE_NEON
+#if HAVE_NEON_ASM
 const ConvolveFunctions convolve8_neon(
     vp9_convolve8_horiz_neon, vp9_convolve8_avg_horiz_neon,
     vp9_convolve8_vert_neon, vp9_convolve8_avg_vert_neon,
diff --git a/test/dct16x16_test.cc b/test/dct16x16_test.cc
index cb5562ec1..143a26726 100644
--- a/test/dct16x16_test.cc
+++ b/test/dct16x16_test.cc
@@ -512,7 +512,7 @@ INSTANTIATE_TEST_CASE_P(
         make_tuple(&vp9_fht16x16_c, &vp9_iht16x16_256_add_c, 2),
         make_tuple(&vp9_fht16x16_c, &vp9_iht16x16_256_add_c, 3)));
 
-#if HAVE_NEON
+#if HAVE_NEON_ASM
 INSTANTIATE_TEST_CASE_P(
     NEON, Trans16x16DCT,
     ::testing::Values(
diff --git a/test/dct32x32_test.cc b/test/dct32x32_test.cc
index 013f4513e..72c0bd69d 100644
--- a/test/dct32x32_test.cc
+++ b/test/dct32x32_test.cc
@@ -248,7 +248,7 @@ INSTANTIATE_TEST_CASE_P(
         make_tuple(&vp9_fdct32x32_c, &vp9_idct32x32_1024_add_c, 0),
         make_tuple(&vp9_fdct32x32_rd_c, &vp9_idct32x32_1024_add_c, 1)));
 
-#if HAVE_NEON
+#if HAVE_NEON_ASM
 INSTANTIATE_TEST_CASE_P(
     NEON, Trans32x32Test,
     ::testing::Values(
diff --git a/test/fdct4x4_test.cc b/test/fdct4x4_test.cc
index 0a8fb9016..030665eda 100644
--- a/test/fdct4x4_test.cc
+++ b/test/fdct4x4_test.cc
@@ -338,7 +338,7 @@ INSTANTIATE_TEST_CASE_P(
     ::testing::Values(
         make_tuple(&vp9_fwht4x4_c, &vp9_iwht4x4_16_add_c, 0)));
 
-#if HAVE_NEON
+#if HAVE_NEON_ASM
 INSTANTIATE_TEST_CASE_P(
     NEON, Trans4x4DCT,
     ::testing::Values(
diff --git a/test/fdct8x8_test.cc b/test/fdct8x8_test.cc
index e6e270914..c7cf1640c 100644
--- a/test/fdct8x8_test.cc
+++ b/test/fdct8x8_test.cc
@@ -313,7 +313,7 @@ INSTANTIATE_TEST_CASE_P(
         make_tuple(&vp9_fht8x8_c, &vp9_iht8x8_64_add_c, 2),
         make_tuple(&vp9_fht8x8_c, &vp9_iht8x8_64_add_c, 3)));
 
-#if HAVE_NEON
+#if HAVE_NEON_ASM
 INSTANTIATE_TEST_CASE_P(
     NEON, FwdTrans8x8DCT,
     ::testing::Values(
diff --git a/test/partial_idct_test.cc b/test/partial_idct_test.cc
index 5c33c017d..79ef52109 100644
--- a/test/partial_idct_test.cc
+++ b/test/partial_idct_test.cc
@@ -140,7 +140,7 @@ INSTANTIATE_TEST_CASE_P(
         make_tuple(&vp9_idct4x4_16_add_c,
                    &vp9_idct4x4_1_add_c,
                    TX_4X4, 1)));
-#if HAVE_NEON
+#if HAVE_NEON_ASM
 INSTANTIATE_TEST_CASE_P(
     NEON, PartialIDctTest,
     ::testing::Values(
diff --git a/test/register_state_check.h b/test/register_state_check.h
index 5987fe31e..1ee149b0c 100644
--- a/test/register_state_check.h
+++ b/test/register_state_check.h
@@ -82,8 +82,8 @@ class RegisterStateCheck {
 
 }  // namespace libvpx_test
 
-#elif defined(CONFIG_SHARED) && defined(HAVE_NEON) && defined(CONFIG_VP9) \
-      && !CONFIG_SHARED && HAVE_NEON && CONFIG_VP9
+#elif defined(CONFIG_SHARED) && defined(HAVE_NEON_ASM) && defined(CONFIG_VP9) \
+      && !CONFIG_SHARED && HAVE_NEON_ASM && CONFIG_VP9
 
 #include "vpx/vpx_integer.h"
 
diff --git a/test/subtract_test.cc b/test/subtract_test.cc
index 3efb9553c..63e999d6e 100644
--- a/test/subtract_test.cc
+++ b/test/subtract_test.cc
@@ -105,7 +105,7 @@ TEST_P(SubtractBlockTest, SimpleSubtract) {
 INSTANTIATE_TEST_CASE_P(C, SubtractBlockTest,
                         ::testing::Values(vp8_subtract_b_c));
 
-#if HAVE_NEON
+#if HAVE_NEON_ASM
 INSTANTIATE_TEST_CASE_P(NEON, SubtractBlockTest,
                         ::testing::Values(vp8_subtract_b_neon));
 #endif
diff --git a/vp8/common/arm/dequantize_arm.c b/vp8/common/arm/dequantize_arm.c
index 70e72aa47..3e37e0892 100644
--- a/vp8/common/arm/dequantize_arm.c
+++ b/vp8/common/arm/dequantize_arm.c
@@ -12,7 +12,7 @@
 #include "vpx_config.h"
 #include "vp8/common/blockd.h"
 
-#if HAVE_NEON
+#if HAVE_NEON_ASM
 extern void vp8_dequantize_b_loop_neon(short *Q, short *DQC, short *DQ);
 #endif
 
@@ -20,7 +20,7 @@ extern void vp8_dequantize_b_loop_neon(short *Q, short *DQC, short *DQ);
 extern void vp8_dequantize_b_loop_v6(short *Q, short *DQC, short *DQ);
 #endif
 
-#if HAVE_NEON
+#if HAVE_NEON_ASM
 
 void vp8_dequantize_b_neon(BLOCKD *d, short *DQC)
 {
diff --git a/vp8/common/arm/loopfilter_arm.c b/vp8/common/arm/loopfilter_arm.c
index 3bdc9675e..5d693c6d5 100644
--- a/vp8/common/arm/loopfilter_arm.c
+++ b/vp8/common/arm/loopfilter_arm.c
@@ -25,7 +25,7 @@ extern prototype_loopfilter(vp8_mbloop_filter_horizontal_edge_armv6);
 extern prototype_loopfilter(vp8_mbloop_filter_vertical_edge_armv6);
 #endif
 
-#if HAVE_NEON
+#if HAVE_NEON_ASM
 typedef void loopfilter_y_neon(unsigned char *src, int pitch,
         unsigned char blimit, unsigned char limit, unsigned char thresh);
 typedef void loopfilter_uv_neon(unsigned char *u, int pitch,
@@ -118,7 +118,7 @@ void vp8_loop_filter_bvs_armv6(unsigned char *y_ptr, int y_stride,
 }
 #endif
 
-#if HAVE_NEON
+#if HAVE_NEON_ASM
 /* NEON loopfilter functions */
 /* Horizontal MB filtering */
 void vp8_loop_filter_mbh_neon(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
diff --git a/vp8/common/arm/reconintra_arm.c b/vp8/common/arm/reconintra_arm.c
index 2874896e8..765fc3ad8 100644
--- a/vp8/common/arm/reconintra_arm.c
+++ b/vp8/common/arm/reconintra_arm.c
@@ -14,7 +14,7 @@
 #include "vp8/common/blockd.h"
 #include "vpx_mem/vpx_mem.h"
 
-#if HAVE_NEON
+#if HAVE_NEON_ARM
 extern void vp8_build_intra_predictors_mby_neon_func(
     unsigned char *y_buffer,
     unsigned char *ypred_ptr,
diff --git a/vp8/common/arm/variance_arm.c b/vp8/common/arm/variance_arm.c
index 467a50942..e3f7083b6 100644
--- a/vp8/common/arm/variance_arm.c
+++ b/vp8/common/arm/variance_arm.c
@@ -95,7 +95,7 @@ unsigned int vp8_sub_pixel_variance16x16_armv6
 #endif /* HAVE_MEDIA */
 
 
-#if HAVE_NEON
+#if HAVE_NEON_ASM
 
 extern unsigned int vp8_sub_pixel_variance16x16_neon_func
 (
diff --git a/vp8/common/rtcd_defs.pl b/vp8/common/rtcd_defs.pl
index 130d96535..dcf5b8e26 100644
--- a/vp8/common/rtcd_defs.pl
+++ b/vp8/common/rtcd_defs.pl
@@ -29,8 +29,9 @@ $vp8_clear_system_state_mmx=vpx_reset_mmx_state;
 # Dequant
 #
 add_proto qw/void vp8_dequantize_b/, "struct blockd*, short *dqc";
-specialize qw/vp8_dequantize_b mmx media neon/;
+specialize qw/vp8_dequantize_b mmx media neon_asm/;
 $vp8_dequantize_b_media=vp8_dequantize_b_v6;
+$vp8_dequantize_b_neon_asm=vp8_dequantize_b_neon;
 
 add_proto qw/void vp8_dequant_idct_add/, "short *input, short *dq, unsigned char *output, int stride";
 specialize qw/vp8_dequant_idct_add mmx media neon dspr2/;
@@ -38,70 +39,76 @@ $vp8_dequant_idct_add_media=vp8_dequant_idct_add_v6;
 $vp8_dequant_idct_add_dspr2=vp8_dequant_idct_add_dspr2;
 
 add_proto qw/void vp8_dequant_idct_add_y_block/, "short *q, short *dq, unsigned char *dst, int stride, char *eobs";
-specialize qw/vp8_dequant_idct_add_y_block mmx sse2 media neon dspr2/;
+specialize qw/vp8_dequant_idct_add_y_block mmx sse2 media neon_asm dspr2/;
 $vp8_dequant_idct_add_y_block_media=vp8_dequant_idct_add_y_block_v6;
+$vp8_dequant_idct_add_y_block_neon_asm=vp8_dequant_idct_add_y_block_neon;
 $vp8_dequant_idct_add_y_block_dspr2=vp8_dequant_idct_add_y_block_dspr2;
 
 add_proto qw/void vp8_dequant_idct_add_uv_block/, "short *q, short *dq, unsigned char *dst_u, unsigned char *dst_v, int stride, char *eobs";
-specialize qw/vp8_dequant_idct_add_uv_block mmx sse2 media neon dspr2/;
+specialize qw/vp8_dequant_idct_add_uv_block mmx sse2 media neon_asm dspr2/;
 $vp8_dequant_idct_add_uv_block_media=vp8_dequant_idct_add_uv_block_v6;
+$vp8_dequant_idct_add_uv_block_neon_asm=vp8_dequant_idct_add_uv_block_neon;
 $vp8_dequant_idct_add_y_block_dspr2=vp8_dequant_idct_add_y_block_dspr2;
 
 #
 # Loopfilter
 #
 add_proto qw/void vp8_loop_filter_mbv/, "unsigned char *y, unsigned char *u, unsigned char *v, int ystride, int uv_stride, struct loop_filter_info *lfi";
-specialize qw/vp8_loop_filter_mbv mmx sse2 media neon dspr2/;
+specialize qw/vp8_loop_filter_mbv mmx sse2 media neon_asm dspr2/;
 $vp8_loop_filter_mbv_media=vp8_loop_filter_mbv_armv6;
+$vp8_loop_filter_mbv_neon_asm=vp8_loop_filter_mbv_neon;
 $vp8_loop_filter_mbv_dspr2=vp8_loop_filter_mbv_dspr2;
 
 add_proto qw/void vp8_loop_filter_bv/, "unsigned char *y, unsigned char *u, unsigned char *v, int ystride, int uv_stride, struct loop_filter_info *lfi";
-specialize qw/vp8_loop_filter_bv mmx sse2 media neon dspr2/;
+specialize qw/vp8_loop_filter_bv mmx sse2 media neon_asm dspr2/;
 $vp8_loop_filter_bv_media=vp8_loop_filter_bv_armv6;
+$vp8_loop_filter_bv_neon_asm=vp8_loop_filter_bv_neon;
 $vp8_loop_filter_bv_dspr2=vp8_loop_filter_bv_dspr2;
 
 add_proto qw/void vp8_loop_filter_mbh/, "unsigned char *y, unsigned char *u, unsigned char *v, int ystride, int uv_stride, struct loop_filter_info *lfi";
-specialize qw/vp8_loop_filter_mbh mmx sse2 media neon dspr2/;
+specialize qw/vp8_loop_filter_mbh mmx sse2 media neon_asm dspr2/;
 $vp8_loop_filter_mbh_media=vp8_loop_filter_mbh_armv6;
+$vp8_loop_filter_mbh_neon_asm=vp8_loop_filter_mbh_neon;
 $vp8_loop_filter_mbh_dspr2=vp8_loop_filter_mbh_dspr2;
 
 add_proto qw/void vp8_loop_filter_bh/, "unsigned char *y, unsigned char *u, unsigned char *v, int ystride, int uv_stride, struct loop_filter_info *lfi";
-specialize qw/vp8_loop_filter_bh mmx sse2 media neon dspr2/;
+specialize qw/vp8_loop_filter_bh mmx sse2 media neon_asm dspr2/;
 $vp8_loop_filter_bh_media=vp8_loop_filter_bh_armv6;
+$vp8_loop_filter_bh_neon_asm=vp8_loop_filter_bh_neon;
 $vp8_loop_filter_bh_dspr2=vp8_loop_filter_bh_dspr2;
 
 
 add_proto qw/void vp8_loop_filter_simple_mbv/, "unsigned char *y, int ystride, const unsigned char *blimit";
-specialize qw/vp8_loop_filter_simple_mbv mmx sse2 media neon/;
+specialize qw/vp8_loop_filter_simple_mbv mmx sse2 media neon_asm/;
 $vp8_loop_filter_simple_mbv_c=vp8_loop_filter_simple_vertical_edge_c;
 $vp8_loop_filter_simple_mbv_mmx=vp8_loop_filter_simple_vertical_edge_mmx;
 $vp8_loop_filter_simple_mbv_sse2=vp8_loop_filter_simple_vertical_edge_sse2;
 $vp8_loop_filter_simple_mbv_media=vp8_loop_filter_simple_vertical_edge_armv6;
-$vp8_loop_filter_simple_mbv_neon=vp8_loop_filter_mbvs_neon;
+$vp8_loop_filter_simple_mbv_neon_asm=vp8_loop_filter_mbvs_neon;
 
 add_proto qw/void vp8_loop_filter_simple_mbh/, "unsigned char *y, int ystride, const unsigned char *blimit";
-specialize qw/vp8_loop_filter_simple_mbh mmx sse2 media neon/;
+specialize qw/vp8_loop_filter_simple_mbh mmx sse2 media neon_asm/;
 $vp8_loop_filter_simple_mbh_c=vp8_loop_filter_simple_horizontal_edge_c;
 $vp8_loop_filter_simple_mbh_mmx=vp8_loop_filter_simple_horizontal_edge_mmx;
 $vp8_loop_filter_simple_mbh_sse2=vp8_loop_filter_simple_horizontal_edge_sse2;
 $vp8_loop_filter_simple_mbh_media=vp8_loop_filter_simple_horizontal_edge_armv6;
-$vp8_loop_filter_simple_mbh_neon=vp8_loop_filter_mbhs_neon;
+$vp8_loop_filter_simple_mbh_neon_asm=vp8_loop_filter_mbhs_neon;
 
 add_proto qw/void vp8_loop_filter_simple_bv/, "unsigned char *y, int ystride, const unsigned char *blimit";
-specialize qw/vp8_loop_filter_simple_bv mmx sse2 media neon/;
+specialize qw/vp8_loop_filter_simple_bv mmx sse2 media neon_asm/;
 $vp8_loop_filter_simple_bv_c=vp8_loop_filter_bvs_c;
 $vp8_loop_filter_simple_bv_mmx=vp8_loop_filter_bvs_mmx;
 $vp8_loop_filter_simple_bv_sse2=vp8_loop_filter_bvs_sse2;
 $vp8_loop_filter_simple_bv_media=vp8_loop_filter_bvs_armv6;
-$vp8_loop_filter_simple_bv_neon=vp8_loop_filter_bvs_neon;
+$vp8_loop_filter_simple_bv_neon_asm=vp8_loop_filter_bvs_neon;
 
 add_proto qw/void vp8_loop_filter_simple_bh/, "unsigned char *y, int ystride, const unsigned char *blimit";
-specialize qw/vp8_loop_filter_simple_bh mmx sse2 media neon/;
+specialize qw/vp8_loop_filter_simple_bh mmx sse2 media neon_asm/;
 $vp8_loop_filter_simple_bh_c=vp8_loop_filter_bhs_c;
 $vp8_loop_filter_simple_bh_mmx=vp8_loop_filter_bhs_mmx;
 $vp8_loop_filter_simple_bh_sse2=vp8_loop_filter_bhs_sse2;
 $vp8_loop_filter_simple_bh_media=vp8_loop_filter_bhs_armv6;
-$vp8_loop_filter_simple_bh_neon=vp8_loop_filter_bhs_neon;
+$vp8_loop_filter_simple_bh_neon_asm=vp8_loop_filter_bhs_neon;
 
 #
 # IDCT
@@ -269,9 +276,10 @@ specialize qw/vp8_sub_pixel_variance4x4 mmx sse2/;
 $vp8_sub_pixel_variance4x4_sse2=vp8_sub_pixel_variance4x4_wmt;
 
 add_proto qw/unsigned int vp8_sub_pixel_variance8x8/, "const unsigned char  *src_ptr, int  source_stride, int  xoffset, int  yoffset, const unsigned char *ref_ptr, int Refstride, unsigned int *sse";
-specialize qw/vp8_sub_pixel_variance8x8 mmx sse2 media neon/;
+specialize qw/vp8_sub_pixel_variance8x8 mmx sse2 media neon_asm/;
 $vp8_sub_pixel_variance8x8_sse2=vp8_sub_pixel_variance8x8_wmt;
 $vp8_sub_pixel_variance8x8_media=vp8_sub_pixel_variance8x8_armv6;
+$vp8_sub_pixel_variance8x8_neon_asm=vp8_sub_pixel_variance8x8_neon;
 
 add_proto qw/unsigned int vp8_sub_pixel_variance8x16/, "const unsigned char  *src_ptr, int  source_stride, int  xoffset, int  yoffset, const unsigned char *ref_ptr, int Refstride, unsigned int *sse";
 specialize qw/vp8_sub_pixel_variance8x16 mmx sse2/;
@@ -282,24 +290,28 @@ specialize qw/vp8_sub_pixel_variance16x8 mmx sse2 ssse3/;
 $vp8_sub_pixel_variance16x8_sse2=vp8_sub_pixel_variance16x8_wmt;
 
 add_proto qw/unsigned int vp8_sub_pixel_variance16x16/, "const unsigned char  *src_ptr, int  source_stride, int  xoffset, int  yoffset, const unsigned char *ref_ptr, int Refstride, unsigned int *sse";
-specialize qw/vp8_sub_pixel_variance16x16 mmx sse2 ssse3 media neon/;
+specialize qw/vp8_sub_pixel_variance16x16 mmx sse2 ssse3 media neon_asm/;
 $vp8_sub_pixel_variance16x16_sse2=vp8_sub_pixel_variance16x16_wmt;
 $vp8_sub_pixel_variance16x16_media=vp8_sub_pixel_variance16x16_armv6;
+$vp8_sub_pixel_variance16x16_neon_asm=vp8_sub_pixel_variance16x16_neon;
 
 add_proto qw/unsigned int vp8_variance_halfpixvar16x16_h/, "const unsigned char *src_ptr, int source_stride, const unsigned char *ref_ptr, int  ref_stride, unsigned int *sse";
-specialize qw/vp8_variance_halfpixvar16x16_h mmx sse2 media neon/;
+specialize qw/vp8_variance_halfpixvar16x16_h mmx sse2 media neon_asm/;
 $vp8_variance_halfpixvar16x16_h_sse2=vp8_variance_halfpixvar16x16_h_wmt;
 $vp8_variance_halfpixvar16x16_h_media=vp8_variance_halfpixvar16x16_h_armv6;
+$vp8_variance_halfpixvar16x16_h_neon_asm=vp8_variance_halfpixvar16x16_h_neon;
 
 add_proto qw/unsigned int vp8_variance_halfpixvar16x16_v/, "const unsigned char *src_ptr, int source_stride, const unsigned char *ref_ptr, int  ref_stride, unsigned int *sse";
-specialize qw/vp8_variance_halfpixvar16x16_v mmx sse2 media neon/;
+specialize qw/vp8_variance_halfpixvar16x16_v mmx sse2 media neon_asm/;
 $vp8_variance_halfpixvar16x16_v_sse2=vp8_variance_halfpixvar16x16_v_wmt;
 $vp8_variance_halfpixvar16x16_v_media=vp8_variance_halfpixvar16x16_v_armv6;
+$vp8_variance_halfpixvar16x16_v_neon_asm=vp8_variance_halfpixvar16x16_v_neon;
 
 add_proto qw/unsigned int vp8_variance_halfpixvar16x16_hv/, "const unsigned char *src_ptr, int source_stride, const unsigned char *ref_ptr, int  ref_stride, unsigned int *sse";
-specialize qw/vp8_variance_halfpixvar16x16_hv mmx sse2 media neon/;
+specialize qw/vp8_variance_halfpixvar16x16_hv mmx sse2 media neon_asm/;
 $vp8_variance_halfpixvar16x16_hv_sse2=vp8_variance_halfpixvar16x16_hv_wmt;
 $vp8_variance_halfpixvar16x16_hv_media=vp8_variance_halfpixvar16x16_hv_armv6;
+$vp8_variance_halfpixvar16x16_hv_neon_asm=vp8_variance_halfpixvar16x16_hv_neon;
 
 #
 # Single block SAD
@@ -402,12 +414,14 @@ specialize qw/vp8_sub_pixel_mse16x16 mmx sse2/;
 $vp8_sub_pixel_mse16x16_sse2=vp8_sub_pixel_mse16x16_wmt;
 
 add_proto qw/unsigned int vp8_mse16x16/, "const unsigned char *src_ptr, int source_stride, const unsigned char *ref_ptr, int  ref_stride, unsigned int *sse";
-specialize qw/vp8_mse16x16 mmx sse2 media neon/;
+specialize qw/vp8_mse16x16 mmx sse2 media neon_asm/;
 $vp8_mse16x16_sse2=vp8_mse16x16_wmt;
 $vp8_mse16x16_media=vp8_mse16x16_armv6;
+$vp8_mse16x16_neon_asm=vp8_mse16x16_neon;
 
 add_proto qw/unsigned int vp8_get4x4sse_cs/, "const unsigned char *src_ptr, int source_stride, const unsigned char *ref_ptr, int  ref_stride";
-specialize qw/vp8_get4x4sse_cs mmx neon/;
+specialize qw/vp8_get4x4sse_cs mmx neon_asm/;
+$vp8_get4x4sse_cs_neon_asm=vp8_get4x4sse_cs_neon;
 
 #
 # Block copy
@@ -434,16 +448,19 @@ if (vpx_config("CONFIG_INTERNAL_STATS") eq "yes") {
 # Forward DCT
 #
 add_proto qw/void vp8_short_fdct4x4/, "short *input, short *output, int pitch";
-specialize qw/vp8_short_fdct4x4 mmx sse2 media neon/;
+specialize qw/vp8_short_fdct4x4 mmx sse2 media neon_asm/;
 $vp8_short_fdct4x4_media=vp8_short_fdct4x4_armv6;
+$vp8_short_fdct4x4_neon_asm=vp8_short_fdct4x4_neon;
 
 add_proto qw/void vp8_short_fdct8x4/, "short *input, short *output, int pitch";
-specialize qw/vp8_short_fdct8x4 mmx sse2 media neon/;
+specialize qw/vp8_short_fdct8x4 mmx sse2 media neon_asm/;
 $vp8_short_fdct8x4_media=vp8_short_fdct8x4_armv6;
+$vp8_short_fdct8x4_neon_asm=vp8_short_fdct8x4_neon;
 
 add_proto qw/void vp8_short_walsh4x4/, "short *input, short *output, int pitch";
-specialize qw/vp8_short_walsh4x4 sse2 media neon/;
+specialize qw/vp8_short_walsh4x4 sse2 media neon_asm/;
 $vp8_short_walsh4x4_media=vp8_short_walsh4x4_armv6;
+$vp8_short_walsh4x4_neon_asm=vp8_short_walsh4x4_neon;
 
 #
 # Quantizer
@@ -454,14 +471,16 @@ specialize qw/vp8_regular_quantize_b sse2/;
 #$vp8_regular_quantize_b_sse4_1=vp8_regular_quantize_b_sse4;
 
 add_proto qw/void vp8_fast_quantize_b/, "struct block *, struct blockd *";
-specialize qw/vp8_fast_quantize_b sse2 ssse3 media neon/;
+specialize qw/vp8_fast_quantize_b sse2 ssse3 media neon_asm/;
 $vp8_fast_quantize_b_media=vp8_fast_quantize_b_armv6;
+$vp8_fast_quantize_b_neon_asm=vp8_fast_quantize_b_neon;
 
 add_proto qw/void vp8_regular_quantize_b_pair/, "struct block *b1, struct block *b2, struct blockd *d1, struct blockd *d2";
 # no asm yet
 
 add_proto qw/void vp8_fast_quantize_b_pair/, "struct block *b1, struct block *b2, struct blockd *d1, struct blockd *d2";
-specialize qw/vp8_fast_quantize_b_pair neon/;
+specialize qw/vp8_fast_quantize_b_pair neon_asm/;
+$vp8_fast_quantize_b_pair_neon_asm=vp8_fast_quantize_b_pair_neon;
 
 add_proto qw/void vp8_quantize_mb/, "struct macroblock *";
 specialize qw/vp8_quantize_mb neon/;
@@ -488,16 +507,19 @@ specialize qw/vp8_mbuverror mmx sse2/;
 $vp8_mbuverror_sse2=vp8_mbuverror_xmm;
 
 add_proto qw/void vp8_subtract_b/, "struct block *be, struct blockd *bd, int pitch";
-specialize qw/vp8_subtract_b mmx sse2 media neon/;
+specialize qw/vp8_subtract_b mmx sse2 media neon_asm/;
 $vp8_subtract_b_media=vp8_subtract_b_armv6;
+$vp8_subtract_b_neon_asm=vp8_subtract_b_neon;
 
 add_proto qw/void vp8_subtract_mby/, "short *diff, unsigned char *src, int src_stride, unsigned char *pred, int pred_stride";
-specialize qw/vp8_subtract_mby mmx sse2 media neon/;
+specialize qw/vp8_subtract_mby mmx sse2 media neon_asm/;
 $vp8_subtract_mby_media=vp8_subtract_mby_armv6;
+$vp8_subtract_mby_neon_asm=vp8_subtract_mby_neon;
 
 add_proto qw/void vp8_subtract_mbuv/, "short *diff, unsigned char *usrc, unsigned char *vsrc, int src_stride, unsigned char *upred, unsigned char *vpred, int pred_stride";
-specialize qw/vp8_subtract_mbuv mmx sse2 media neon/;
+specialize qw/vp8_subtract_mbuv mmx sse2 media neon_asm/;
 $vp8_subtract_mbuv_media=vp8_subtract_mbuv_armv6;
+$vp8_subtract_mbuv_neon_asm=vp8_subtract_mbuv_neon;
 
 #
 # Motion search
@@ -526,7 +548,8 @@ if (vpx_config("CONFIG_REALTIME_ONLY") ne "yes") {
 # Pick Loopfilter
 #
 add_proto qw/void vp8_yv12_copy_partial_frame/, "struct yv12_buffer_config *src_ybc, struct yv12_buffer_config *dst_ybc";
-specialize qw/vp8_yv12_copy_partial_frame neon/;
+specialize qw/vp8_yv12_copy_partial_frame neon_asm/;
+$vp8_yv12_copy_partial_frame_neon_asm=vp8_yv12_copy_partial_frame_neon;
 
 #
 # Denoiser filter
diff --git a/vp8/vp8_common.mk b/vp8/vp8_common.mk
index 2812111e9..e58cdad4b 100644
--- a/vp8/vp8_common.mk
+++ b/vp8/vp8_common.mk
@@ -129,7 +129,6 @@ VP8_COMMON_SRCS-$(HAVE_DSPR2)  += common/mips/dspr2/dequantize_dspr2.c
 # common (c)
 VP8_COMMON_SRCS-$(ARCH_ARM)  += common/arm/filter_arm.c
 VP8_COMMON_SRCS-$(ARCH_ARM)  += common/arm/loopfilter_arm.c
-VP8_COMMON_SRCS-$(ARCH_ARM)  += common/arm/reconintra_arm.c
 VP8_COMMON_SRCS-$(ARCH_ARM)  += common/arm/dequantize_arm.c
 VP8_COMMON_SRCS-$(ARCH_ARM)  += common/arm/variance_arm.c
 
@@ -159,13 +158,14 @@ VP8_COMMON_SRCS-$(HAVE_MEDIA)  += common/arm/armv6/vp8_variance_halfpixvar16x16_
 VP8_COMMON_SRCS-$(HAVE_MEDIA)  += common/arm/armv6/vp8_variance_halfpixvar16x16_hv_armv6$(ASM)
 
 # common (neon)
-VP8_COMMON_SRCS-$(HAVE_NEON)  += common/arm/neon/loopfilter_neon$(ASM)
-VP8_COMMON_SRCS-$(HAVE_NEON)  += common/arm/neon/loopfiltersimpleverticaledge_neon$(ASM)
-VP8_COMMON_SRCS-$(HAVE_NEON)  += common/arm/neon/buildintrapredictorsmby_neon$(ASM)
-VP8_COMMON_SRCS-$(HAVE_NEON)  += common/arm/neon/idct_blk_neon.c
-VP8_COMMON_SRCS-$(HAVE_NEON)  += common/arm/neon/vp8_subpixelvariance8x8_neon$(ASM)
-VP8_COMMON_SRCS-$(HAVE_NEON)  += common/arm/neon/vp8_subpixelvariance16x16_neon$(ASM)
-VP8_COMMON_SRCS-$(HAVE_NEON)  += common/arm/neon/vp8_subpixelvariance16x16s_neon$(ASM)
+VP8_COMMON_SRCS-$(ARCH_NEON_ASM)  += common/arm/reconintra_arm.c
+VP8_COMMON_SRCS-$(HAVE_NEON_ASM)  += common/arm/neon/loopfilter_neon$(ASM)
+VP8_COMMON_SRCS-$(HAVE_NEON_ASM)  += common/arm/neon/loopfiltersimpleverticaledge_neon$(ASM)
+VP8_COMMON_SRCS-$(HAVE_NEON_ASM)  += common/arm/neon/buildintrapredictorsmby_neon$(ASM)
+VP8_COMMON_SRCS-$(HAVE_NEON_ASM)  += common/arm/neon/idct_blk_neon.c
+VP8_COMMON_SRCS-$(HAVE_NEON_ASM)  += common/arm/neon/vp8_subpixelvariance8x8_neon$(ASM)
+VP8_COMMON_SRCS-$(HAVE_NEON_ASM)  += common/arm/neon/vp8_subpixelvariance16x16_neon$(ASM)
+VP8_COMMON_SRCS-$(HAVE_NEON_ASM)  += common/arm/neon/vp8_subpixelvariance16x16s_neon$(ASM)
 
 # common (neon intrinsics)
 VP8_COMMON_SRCS-$(HAVE_NEON)  += common/arm/neon/bilinearpredict_neon.c
diff --git a/vp8/vp8cx_arm.mk b/vp8/vp8cx_arm.mk
index 398172a2b..573304863 100644
--- a/vp8/vp8cx_arm.mk
+++ b/vp8/vp8cx_arm.mk
@@ -35,11 +35,12 @@ VP8_CX_SRCS-$(HAVE_MEDIA)  += encoder/arm/armv6/walsh_v6$(ASM)
 
 #File list for neon
 # encoder
-VP8_CX_SRCS-$(HAVE_NEON)  += encoder/arm/neon/fastquantizeb_neon$(ASM)
-VP8_CX_SRCS-$(HAVE_NEON)  += encoder/arm/neon/picklpf_arm.c
+VP8_CX_SRCS-$(HAVE_NEON_ASM)  += encoder/arm/neon/fastquantizeb_neon$(ASM)
+VP8_CX_SRCS-$(HAVE_NEON_ASM)  += encoder/arm/neon/picklpf_arm.c
+VP8_CX_SRCS-$(HAVE_NEON_ASM)  += encoder/arm/neon/shortfdct_neon$(ASM)
+VP8_CX_SRCS-$(HAVE_NEON_ASM)  += encoder/arm/neon/subtract_neon$(ASM)
+VP8_CX_SRCS-$(HAVE_NEON_ASM)  += encoder/arm/neon/vp8_mse16x16_neon$(ASM)
+VP8_CX_SRCS-$(HAVE_NEON_ASM)  += encoder/arm/neon/vp8_memcpy_neon$(ASM)
+VP8_CX_SRCS-$(HAVE_NEON_ASM)  += encoder/arm/neon/vp8_shortwalsh4x4_neon$(ASM)
+
 VP8_CX_SRCS-$(HAVE_NEON)  += encoder/arm/neon/denoising_neon.c
-VP8_CX_SRCS-$(HAVE_NEON)  += encoder/arm/neon/shortfdct_neon$(ASM)
-VP8_CX_SRCS-$(HAVE_NEON)  += encoder/arm/neon/subtract_neon$(ASM)
-VP8_CX_SRCS-$(HAVE_NEON)  += encoder/arm/neon/vp8_mse16x16_neon$(ASM)
-VP8_CX_SRCS-$(HAVE_NEON)  += encoder/arm/neon/vp8_memcpy_neon$(ASM)
-VP8_CX_SRCS-$(HAVE_NEON)  += encoder/arm/neon/vp8_shortwalsh4x4_neon$(ASM)
diff --git a/vp9/common/vp9_rtcd_defs.pl b/vp9/common/vp9_rtcd_defs.pl
index 5d7da4674..7754763a0 100644
--- a/vp9/common/vp9_rtcd_defs.pl
+++ b/vp9/common/vp9_rtcd_defs.pl
@@ -58,7 +58,8 @@ add_proto qw/void vp9_d63_predictor_4x4/, "uint8_t *dst, ptrdiff_t y_stride, con
 specialize qw/vp9_d63_predictor_4x4/, "$ssse3_x86inc";
 
 add_proto qw/void vp9_h_predictor_4x4/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vp9_h_predictor_4x4 neon dspr2/, "$ssse3_x86inc";
+specialize qw/vp9_h_predictor_4x4 neon_asm dspr2/, "$ssse3_x86inc";
+$vp9_h_predictor_4x4_neon_asm=vp9_h_predictor_4x4_neon;
 
 add_proto qw/void vp9_d117_predictor_4x4/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
 specialize qw/vp9_d117_predictor_4x4/;
@@ -70,10 +71,12 @@ add_proto qw/void vp9_d153_predictor_4x4/, "uint8_t *dst, ptrdiff_t y_stride, co
 specialize qw/vp9_d153_predictor_4x4/, "$ssse3_x86inc";
 
 add_proto qw/void vp9_v_predictor_4x4/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vp9_v_predictor_4x4 neon/, "$sse_x86inc";
+specialize qw/vp9_v_predictor_4x4 neon_asm/, "$sse_x86inc";
+$vp9_v_predictor_4x4_neon_asm=vp9_v_predictor_4x4_neon;
 
 add_proto qw/void vp9_tm_predictor_4x4/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vp9_tm_predictor_4x4 neon dspr2/, "$sse_x86inc";
+specialize qw/vp9_tm_predictor_4x4 neon_asm dspr2/, "$sse_x86inc";
+$vp9_tm_predictor_4x4_neon_asm=vp9_tm_predictor_4x4_neon;
 
 add_proto qw/void vp9_dc_predictor_4x4/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
 specialize qw/vp9_dc_predictor_4x4 dspr2/, "$sse_x86inc";
@@ -97,7 +100,8 @@ add_proto qw/void vp9_d63_predictor_8x8/, "uint8_t *dst, ptrdiff_t y_stride, con
 specialize qw/vp9_d63_predictor_8x8/, "$ssse3_x86inc";
 
 add_proto qw/void vp9_h_predictor_8x8/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vp9_h_predictor_8x8 neon dspr2/, "$ssse3_x86inc";
+specialize qw/vp9_h_predictor_8x8 neon_asm dspr2/, "$ssse3_x86inc";
+$vp9_h_predictor_8x8_neon_asm=vp9_h_predictor_8x8_neon;
 
 add_proto qw/void vp9_d117_predictor_8x8/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
 specialize qw/vp9_d117_predictor_8x8/;
@@ -109,10 +113,12 @@ add_proto qw/void vp9_d153_predictor_8x8/, "uint8_t *dst, ptrdiff_t y_stride, co
 specialize qw/vp9_d153_predictor_8x8/, "$ssse3_x86inc";
 
 add_proto qw/void vp9_v_predictor_8x8/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vp9_v_predictor_8x8 neon/, "$sse_x86inc";
+specialize qw/vp9_v_predictor_8x8 neon_asm/, "$sse_x86inc";
+$vp9_v_predictor_8x8_neon_asm=vp9_v_predictor_8x8_neon;
 
 add_proto qw/void vp9_tm_predictor_8x8/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vp9_tm_predictor_8x8 neon dspr2/, "$sse2_x86inc";
+specialize qw/vp9_tm_predictor_8x8 neon_asm dspr2/, "$sse2_x86inc";
+$vp9_tm_predictor_8x8_neon_asm=vp9_tm_predictor_8x8_neon;
 
 add_proto qw/void vp9_dc_predictor_8x8/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
 specialize qw/vp9_dc_predictor_8x8 dspr2/, "$sse_x86inc";
@@ -136,7 +142,8 @@ add_proto qw/void vp9_d63_predictor_16x16/, "uint8_t *dst, ptrdiff_t y_stride, c
 specialize qw/vp9_d63_predictor_16x16/, "$ssse3_x86inc";
 
 add_proto qw/void vp9_h_predictor_16x16/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vp9_h_predictor_16x16 neon dspr2/, "$ssse3_x86inc";
+specialize qw/vp9_h_predictor_16x16 neon_asm dspr2/, "$ssse3_x86inc";
+$vp9_h_predictor_16x16_neon_asm=vp9_h_predictor_16x16_neon;
 
 add_proto qw/void vp9_d117_predictor_16x16/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
 specialize qw/vp9_d117_predictor_16x16/;
@@ -148,10 +155,12 @@ add_proto qw/void vp9_d153_predictor_16x16/, "uint8_t *dst, ptrdiff_t y_stride,
 specialize qw/vp9_d153_predictor_16x16/, "$ssse3_x86inc";
 
 add_proto qw/void vp9_v_predictor_16x16/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vp9_v_predictor_16x16 neon/, "$sse2_x86inc";
+specialize qw/vp9_v_predictor_16x16 neon_asm/, "$sse2_x86inc";
+$vp9_v_predictor_16x16_neon_asm=vp9_v_predictor_16x16_neon;
 
 add_proto qw/void vp9_tm_predictor_16x16/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vp9_tm_predictor_16x16 neon/, "$sse2_x86inc";
+specialize qw/vp9_tm_predictor_16x16 neon_asm/, "$sse2_x86inc";
+$vp9_tm_predictor_16x16_neon_asm=vp9_tm_predictor_16x16_neon;
 
 add_proto qw/void vp9_dc_predictor_16x16/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
 specialize qw/vp9_dc_predictor_16x16 dspr2/, "$sse2_x86inc";
@@ -175,7 +184,8 @@ add_proto qw/void vp9_d63_predictor_32x32/, "uint8_t *dst, ptrdiff_t y_stride, c
 specialize qw/vp9_d63_predictor_32x32/, "$ssse3_x86inc";
 
 add_proto qw/void vp9_h_predictor_32x32/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vp9_h_predictor_32x32 neon/, "$ssse3_x86inc";
+specialize qw/vp9_h_predictor_32x32 neon_asm/, "$ssse3_x86inc";
+$vp9_h_predictor_32x32_neon_asm=vp9_h_predictor_32x32_neon;
 
 add_proto qw/void vp9_d117_predictor_32x32/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
 specialize qw/vp9_d117_predictor_32x32/;
@@ -187,10 +197,12 @@ add_proto qw/void vp9_d153_predictor_32x32/, "uint8_t *dst, ptrdiff_t y_stride,
 specialize qw/vp9_d153_predictor_32x32/;
 
 add_proto qw/void vp9_v_predictor_32x32/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vp9_v_predictor_32x32 neon/, "$sse2_x86inc";
+specialize qw/vp9_v_predictor_32x32 neon_asm/, "$sse2_x86inc";
+$vp9_v_predictor_32x32_neon_asm=vp9_v_predictor_32x32_neon;
 
 add_proto qw/void vp9_tm_predictor_32x32/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
-specialize qw/vp9_tm_predictor_32x32 neon/, "$sse2_x86_64";
+specialize qw/vp9_tm_predictor_32x32 neon_asm/, "$sse2_x86_64";
+$vp9_tm_predictor_32x32_neon_asm=vp9_tm_predictor_32x32_neon;
 
 add_proto qw/void vp9_dc_predictor_32x32/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
 specialize qw/vp9_dc_predictor_32x32/, "$sse2_x86inc";
@@ -208,37 +220,48 @@ specialize qw/vp9_dc_128_predictor_32x32/;
 # Loopfilter
 #
 add_proto qw/void vp9_lpf_vertical_16/, "uint8_t *s, int pitch, const uint8_t *blimit, const uint8_t *limit, const uint8_t *thresh";
-specialize qw/vp9_lpf_vertical_16 sse2 neon dspr2/;
+specialize qw/vp9_lpf_vertical_16 sse2 neon_asm dspr2/;
+$vp9_lpf_vertical_16_neon_asm=vp9_lpf_vertical_16_neon;
 
 add_proto qw/void vp9_lpf_vertical_16_dual/, "uint8_t *s, int pitch, const uint8_t *blimit, const uint8_t *limit, const uint8_t *thresh";
-specialize qw/vp9_lpf_vertical_16_dual sse2 neon dspr2/;
+specialize qw/vp9_lpf_vertical_16_dual sse2 neon_asm dspr2/;
+$vp9_lpf_vertical_16_dual_neon_asm=vp9_lpf_vertical_16_dual_neon;
 
 add_proto qw/void vp9_lpf_vertical_8/, "uint8_t *s, int pitch, const uint8_t *blimit, const uint8_t *limit, const uint8_t *thresh, int count";
-specialize qw/vp9_lpf_vertical_8 sse2 neon dspr2/;
+specialize qw/vp9_lpf_vertical_8 sse2 neon_asm dspr2/;
+$vp9_lpf_vertical_8_neon_asm=vp9_lpf_vertical_8_neon;
 
 add_proto qw/void vp9_lpf_vertical_8_dual/, "uint8_t *s, int pitch, const uint8_t *blimit0, const uint8_t *limit0, const uint8_t *thresh0, const uint8_t *blimit1, const uint8_t *limit1, const uint8_t *thresh1";
-specialize qw/vp9_lpf_vertical_8_dual sse2 neon dspr2/;
+specialize qw/vp9_lpf_vertical_8_dual sse2 neon_asm dspr2/;
+$vp9_lpf_vertical_8_dual_neon_asm=vp9_lpf_vertical_8_dual_neon;
 
 add_proto qw/void vp9_lpf_vertical_4/, "uint8_t *s, int pitch, const uint8_t *blimit, const uint8_t *limit, const uint8_t *thresh, int count";
-specialize qw/vp9_lpf_vertical_4 mmx neon dspr2/;
+specialize qw/vp9_lpf_vertical_4 mmx neon_asm dspr2/;
+$vp9_lpf_vertical_4_neon_asm=vp9_lpf_vertical_4_neon;
 
 add_proto qw/void vp9_lpf_vertical_4_dual/, "uint8_t *s, int pitch, const uint8_t *blimit0, const uint8_t *limit0, const uint8_t *thresh0, const uint8_t *blimit1, const uint8_t *limit1, const uint8_t *thresh1";
-specialize qw/vp9_lpf_vertical_4_dual sse2 neon dspr2/;
+specialize qw/vp9_lpf_vertical_4_dual sse2 neon_asm dspr2/;
+$vp9_lpf_vertical_4_dual_neon_asm=vp9_lpf_vertical_4_dual_neon;
 
 add_proto qw/void vp9_lpf_horizontal_16/, "uint8_t *s, int pitch, const uint8_t *blimit, const uint8_t *limit, const uint8_t *thresh, int count";
-specialize qw/vp9_lpf_horizontal_16 sse2 avx2 neon dspr2/;
+specialize qw/vp9_lpf_horizontal_16 sse2 avx2 neon_asm dspr2/;
+$vp9_lpf_horizontal_16_neon_asm=vp9_lpf_horizontal_16_neon;
 
 add_proto qw/void vp9_lpf_horizontal_8/, "uint8_t *s, int pitch, const uint8_t *blimit, const uint8_t *limit, const uint8_t *thresh, int count";
-specialize qw/vp9_lpf_horizontal_8 sse2 neon dspr2/;
+specialize qw/vp9_lpf_horizontal_8 sse2 neon_asm dspr2/;
+$vp9_lpf_horizontal_8_neon_asm=vp9_lpf_horizontal_8_neon;
 
 add_proto qw/void vp9_lpf_horizontal_8_dual/, "uint8_t *s, int pitch, const uint8_t *blimit0, const uint8_t *limit0, const uint8_t *thresh0, const uint8_t *blimit1, const uint8_t *limit1, const uint8_t *thresh1";
-specialize qw/vp9_lpf_horizontal_8_dual sse2 neon dspr2/;
+specialize qw/vp9_lpf_horizontal_8_dual sse2 neon_asm dspr2/;
+$vp9_lpf_horizontal_8_dual_neon_asm=vp9_lpf_horizontal_8_dual_neon;
 
 add_proto qw/void vp9_lpf_horizontal_4/, "uint8_t *s, int pitch, const uint8_t *blimit, const uint8_t *limit, const uint8_t *thresh, int count";
-specialize qw/vp9_lpf_horizontal_4 mmx neon dspr2/;
+specialize qw/vp9_lpf_horizontal_4 mmx neon_asm dspr2/;
+$vp9_lpf_horizontal_4_neon_asm=vp9_lpf_horizontal_4_neon;
 
 add_proto qw/void vp9_lpf_horizontal_4_dual/, "uint8_t *s, int pitch, const uint8_t *blimit0, const uint8_t *limit0, const uint8_t *thresh0, const uint8_t *blimit1, const uint8_t *limit1, const uint8_t *thresh1";
-specialize qw/vp9_lpf_horizontal_4_dual sse2 neon dspr2/;
+specialize qw/vp9_lpf_horizontal_4_dual sse2 neon_asm dspr2/;
+$vp9_lpf_horizontal_4_dual_neon_asm=vp9_lpf_horizontal_4_dual_neon;
 
 #
 # post proc
@@ -274,71 +297,91 @@ specialize qw/vp9_blend_b/;
 # Sub Pixel Filters
 #
 add_proto qw/void vp9_convolve_copy/, "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h";
-specialize qw/vp9_convolve_copy neon dspr2/, "$sse2_x86inc";
+specialize qw/vp9_convolve_copy neon_asm dspr2/, "$sse2_x86inc";
+$vp9_convolve_copy_neon_asm=vp9_convolve_copy_neon;
 
 add_proto qw/void vp9_convolve_avg/, "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h";
-specialize qw/vp9_convolve_avg neon dspr2/, "$sse2_x86inc";
+specialize qw/vp9_convolve_avg neon_asm dspr2/, "$sse2_x86inc";
+$vp9_convolve_avg_neon_asm=vp9_convolve_avg_neon;
 
 add_proto qw/void vp9_convolve8/, "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h";
-specialize qw/vp9_convolve8 sse2 ssse3 avx2 neon dspr2/;
+specialize qw/vp9_convolve8 sse2 ssse3 avx2 neon_asm dspr2/;
+$vp9_convolve8_neon_asm=vp9_convolve8_neon;
 
 add_proto qw/void vp9_convolve8_horiz/, "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h";
-specialize qw/vp9_convolve8_horiz sse2 ssse3 avx2 neon dspr2/;
+specialize qw/vp9_convolve8_horiz sse2 ssse3 avx2 neon_asm dspr2/;
+$vp9_convolve8_horiz_neon_asm=vp9_convolve8_horiz_neon;
 
 add_proto qw/void vp9_convolve8_vert/, "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h";
-specialize qw/vp9_convolve8_vert sse2 ssse3 avx2 neon dspr2/;
+specialize qw/vp9_convolve8_vert sse2 ssse3 avx2 neon_asm dspr2/;
+$vp9_convolve8_vert_neon_asm=vp9_convolve8_vert_neon;
 
 add_proto qw/void vp9_convolve8_avg/, "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h";
-specialize qw/vp9_convolve8_avg sse2 ssse3 neon dspr2/;
+specialize qw/vp9_convolve8_avg sse2 ssse3 neon_asm dspr2/;
+$vp9_convolve8_avg_neon_asm=vp9_convolve8_avg_neon;
 
 add_proto qw/void vp9_convolve8_avg_horiz/, "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h";
-specialize qw/vp9_convolve8_avg_horiz sse2 ssse3 neon dspr2/;
+specialize qw/vp9_convolve8_avg_horiz sse2 ssse3 neon_asm dspr2/;
+$vp9_convolve8_avg_horiz_neon_asm=vp9_convolve8_avg_horiz_neon;
 
 add_proto qw/void vp9_convolve8_avg_vert/, "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h";
-specialize qw/vp9_convolve8_avg_vert sse2 ssse3 neon dspr2/;
+specialize qw/vp9_convolve8_avg_vert sse2 ssse3 neon_asm dspr2/;
+$vp9_convolve8_avg_vert_neon_asm=vp9_convolve8_avg_vert_neon;
 
 #
 # dct
 #
 add_proto qw/void vp9_idct4x4_1_add/, "const int16_t *input, uint8_t *dest, int dest_stride";
-specialize qw/vp9_idct4x4_1_add sse2 neon dspr2/;
+specialize qw/vp9_idct4x4_1_add sse2 neon_asm dspr2/;
+$vp9_idct4x4_1_add_neon_asm=vp9_idct4x4_1_add_neon;
 
 add_proto qw/void vp9_idct4x4_16_add/, "const int16_t *input, uint8_t *dest, int dest_stride";
-specialize qw/vp9_idct4x4_16_add sse2 neon dspr2/;
+specialize qw/vp9_idct4x4_16_add sse2 neon_asm dspr2/;
+$vp9_idct4x4_16_add_neon_asm=vp9_idct4x4_16_add_neon;
 
 add_proto qw/void vp9_idct8x8_1_add/, "const int16_t *input, uint8_t *dest, int dest_stride";
-specialize qw/vp9_idct8x8_1_add sse2 neon dspr2/;
+specialize qw/vp9_idct8x8_1_add sse2 neon_asm dspr2/;
+$vp9_idct8x8_1_add_neon_asm=vp9_idct8x8_1_add_neon;
 
 add_proto qw/void vp9_idct8x8_64_add/, "const int16_t *input, uint8_t *dest, int dest_stride";
-specialize qw/vp9_idct8x8_64_add sse2 neon dspr2/, "$ssse3_x86_64";
+specialize qw/vp9_idct8x8_64_add sse2 neon_asm dspr2/, "$ssse3_x86_64";
+$vp9_idct8x8_64_add_neon_asm=vp9_idct8x8_64_add_neon;
 
 add_proto qw/void vp9_idct8x8_12_add/, "const int16_t *input, uint8_t *dest, int dest_stride";
-specialize qw/vp9_idct8x8_12_add sse2 neon dspr2/, "$ssse3_x86_64";
+specialize qw/vp9_idct8x8_12_add sse2 neon_asm dspr2/, "$ssse3_x86_64";
+$vp9_idct8x8_12_add_neon_asm=vp9_idct8x8_12_add_neon;
 
 add_proto qw/void vp9_idct16x16_1_add/, "const int16_t *input, uint8_t *dest, int dest_stride";
-specialize qw/vp9_idct16x16_1_add sse2 neon dspr2/;
+specialize qw/vp9_idct16x16_1_add sse2 neon_asm dspr2/;
+$vp9_idct16x16_1_add_neon_asm=vp9_idct16x16_1_add_neon;
 
 add_proto qw/void vp9_idct16x16_256_add/, "const int16_t *input, uint8_t *dest, int dest_stride";
-specialize qw/vp9_idct16x16_256_add sse2 neon dspr2/;
+specialize qw/vp9_idct16x16_256_add sse2 neon_asm dspr2/;
+$vp9_idct16x16_256_add_neon_asm=vp9_idct16x16_256_add_neon;
 
 add_proto qw/void vp9_idct16x16_10_add/, "const int16_t *input, uint8_t *dest, int dest_stride";
-specialize qw/vp9_idct16x16_10_add sse2 neon dspr2/;
+specialize qw/vp9_idct16x16_10_add sse2 neon_asm dspr2/;
+$vp9_idct16x16_10_add_neon_asm=vp9_idct16x16_10_add_neon;
 
 add_proto qw/void vp9_idct32x32_1024_add/, "const int16_t *input, uint8_t *dest, int dest_stride";
-specialize qw/vp9_idct32x32_1024_add sse2 neon dspr2/;
+specialize qw/vp9_idct32x32_1024_add sse2 neon_asm dspr2/;
+$vp9_idct32x32_1024_add_neon_asm=vp9_idct32x32_1024_add_neon;
 
 add_proto qw/void vp9_idct32x32_34_add/, "const int16_t *input, uint8_t *dest, int dest_stride";
-specialize qw/vp9_idct32x32_34_add sse2 neon dspr2/;
-$vp9_idct32x32_34_add_neon=vp9_idct32x32_1024_add_neon;
+specialize qw/vp9_idct32x32_34_add sse2 neon_asm dspr2/;
+$vp9_idct32x32_34_add_neon_asm=vp9_idct32x32_1024_add_neon;
 
 add_proto qw/void vp9_idct32x32_1_add/, "const int16_t *input, uint8_t *dest, int dest_stride";
-specialize qw/vp9_idct32x32_1_add sse2 neon dspr2/;
+specialize qw/vp9_idct32x32_1_add sse2 neon_asm dspr2/;
+$vp9_idct32x32_1_add_neon_asm=vp9_idct32x32_1_add_neon;
 
 add_proto qw/void vp9_iht4x4_16_add/, "const int16_t *input, uint8_t *dest, int dest_stride, int tx_type";
-specialize qw/vp9_iht4x4_16_add sse2 neon dspr2/;
+specialize qw/vp9_iht4x4_16_add sse2 neon_asm dspr2/;
+$vp9_iht4x4_16_add_neon_asm=vp9_iht4x4_16_add_neon;
 
 add_proto qw/void vp9_iht8x8_64_add/, "const int16_t *input, uint8_t *dest, int dest_stride, int tx_type";
-specialize qw/vp9_iht8x8_64_add sse2 neon dspr2/;
+specialize qw/vp9_iht8x8_64_add sse2 neon_asm dspr2/;
+$vp9_iht8x8_64_add_neon_asm=vp9_iht8x8_64_add_neon;
 
 add_proto qw/void vp9_iht16x16_256_add/, "const int16_t *input, uint8_t *output, int pitch, int tx_type";
 specialize qw/vp9_iht16x16_256_add sse2 dspr2/;
diff --git a/vp9/vp9_common.mk b/vp9/vp9_common.mk
index eaff60aa0..3b4d6b901 100644
--- a/vp9/vp9_common.mk
+++ b/vp9/vp9_common.mk
@@ -124,28 +124,28 @@ ifeq ($(ARCH_X86_64), yes)
 VP9_COMMON_SRCS-$(HAVE_SSSE3) += common/x86/vp9_idct_ssse3.asm
 endif
 
-VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_convolve_neon.c
-VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_idct16x16_neon.c
-VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_loopfilter_16_neon.c
-VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_convolve8_neon$(ASM)
-VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_convolve8_avg_neon$(ASM)
-VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_loopfilter_neon$(ASM)
-VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_loopfilter_16_neon$(ASM)
-VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_dc_only_idct_add_neon$(ASM)
-VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_idct4x4_1_add_neon$(ASM)
-VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_idct4x4_add_neon$(ASM)
-VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_idct8x8_1_add_neon$(ASM)
-VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_idct8x8_add_neon$(ASM)
-VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_idct16x16_1_add_neon$(ASM)
-VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_idct16x16_add_neon$(ASM)
-VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_idct32x32_1_add_neon$(ASM)
-VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_idct32x32_add_neon$(ASM)
-VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_iht4x4_add_neon$(ASM)
-VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_iht8x8_add_neon$(ASM)
-VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_mb_lpf_neon$(ASM)
-VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_copy_neon$(ASM)
-VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_avg_neon$(ASM)
-VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_save_reg_neon$(ASM)
-VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_reconintra_neon$(ASM)
+VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_convolve_neon.c
+VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_idct16x16_neon.c
+VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_loopfilter_16_neon.c
+VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_convolve8_neon$(ASM)
+VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_convolve8_avg_neon$(ASM)
+VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_loopfilter_neon$(ASM)
+VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_loopfilter_16_neon$(ASM)
+VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_dc_only_idct_add_neon$(ASM)
+VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_idct4x4_1_add_neon$(ASM)
+VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_idct4x4_add_neon$(ASM)
+VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_idct8x8_1_add_neon$(ASM)
+VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_idct8x8_add_neon$(ASM)
+VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_idct16x16_1_add_neon$(ASM)
+VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_idct16x16_add_neon$(ASM)
+VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_idct32x32_1_add_neon$(ASM)
+VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_idct32x32_add_neon$(ASM)
+VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_iht4x4_add_neon$(ASM)
+VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_iht8x8_add_neon$(ASM)
+VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_mb_lpf_neon$(ASM)
+VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_copy_neon$(ASM)
+VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_avg_neon$(ASM)
+VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_save_reg_neon$(ASM)
+VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_reconintra_neon$(ASM)
 
 $(eval $(call rtcd_h_template,vp9_rtcd,vp9/common/vp9_rtcd_defs.pl))
diff --git a/vpx_scale/vpx_scale.mk b/vpx_scale/vpx_scale.mk
index ded8e0b76..95e7483e6 100644
--- a/vpx_scale/vpx_scale.mk
+++ b/vpx_scale/vpx_scale.mk
@@ -10,10 +10,10 @@ SCALE_SRCS-yes += vpx_scale_rtcd.c
 SCALE_SRCS-yes += vpx_scale_rtcd.pl
 
 #neon
-SCALE_SRCS-$(HAVE_NEON)  += arm/neon/vp8_vpxyv12_copyframe_func_neon$(ASM)
-SCALE_SRCS-$(HAVE_NEON)  += arm/neon/vp8_vpxyv12_copysrcframe_func_neon$(ASM)
-SCALE_SRCS-$(HAVE_NEON)  += arm/neon/vp8_vpxyv12_extendframeborders_neon$(ASM)
-SCALE_SRCS-$(HAVE_NEON)  += arm/neon/yv12extend_arm.c
+SCALE_SRCS-$(HAVE_NEON_ASM)  += arm/neon/vp8_vpxyv12_copyframe_func_neon$(ASM)
+SCALE_SRCS-$(HAVE_NEON_ASM)  += arm/neon/vp8_vpxyv12_copysrcframe_func_neon$(ASM)
+SCALE_SRCS-$(HAVE_NEON_ASM)  += arm/neon/vp8_vpxyv12_extendframeborders_neon$(ASM)
+SCALE_SRCS-$(HAVE_NEON_ASM)  += arm/neon/yv12extend_arm.c
 
 #mips(dspr2)
 SCALE_SRCS-$(HAVE_DSPR2)  += mips/dspr2/yv12extend_dspr2.c
diff --git a/vpx_scale/vpx_scale_rtcd.pl b/vpx_scale/vpx_scale_rtcd.pl
index 8c9257052..2e3f1ffbe 100644
--- a/vpx_scale/vpx_scale_rtcd.pl
+++ b/vpx_scale/vpx_scale_rtcd.pl
@@ -17,10 +17,12 @@ if (vpx_config("CONFIG_SPATIAL_RESAMPLING") eq "yes") {
 }
 
 add_proto qw/void vp8_yv12_extend_frame_borders/, "struct yv12_buffer_config *ybf";
-specialize qw/vp8_yv12_extend_frame_borders neon/;
+specialize qw/vp8_yv12_extend_frame_borders neon_asm/;
+$vp8_yv12_extend_frame_borders_neon_asm=vp8_yv12_extend_frame_borders_neon;
 
 add_proto qw/void vp8_yv12_copy_frame/, "const struct yv12_buffer_config *src_ybc, struct yv12_buffer_config *dst_ybc";
-specialize qw/vp8_yv12_copy_frame neon/;
+specialize qw/vp8_yv12_copy_frame neon_asm/;
+$vp8_yv12_copy_frame_neon_asm=vp8_yv12_copy_frame_neon;
 
 add_proto qw/void vpx_yv12_copy_y/, "const struct yv12_buffer_config *src_ybc, struct yv12_buffer_config *dst_ybc";