VP9 common for ARMv8 by using NEON intrinsics 06

Add vp9_iht8x8_add_neon.c
- vp9_iht8x8_64_add_neon

The assembly did not previously implement tx_type 0
BUG=716

Change-Id: Icfc99dd24f3d59047f9184a7d0c761ba7e3de934
Signed-off-by: James Yu <james.yu@linaro.org>

test/fdct8x8_test.cc

@@ -699,6 +699,9 @@ INSTANTIATE_TEST_CASE_P(
     ::testing::Values(
         make_tuple(&vp9_fdct8x8_neon, &vp9_idct8x8_64_add_neon, 0,
                    VPX_BITS_8)));
+#endif  // HAVE_NEON_ASM && !CONFIG_VP9_HIGHBITDEPTH && !CONFIG_EMULATE_HARDWARE
+
+#if HAVE_NEON && !CONFIG_VP9_HIGHBITDEPTH && !CONFIG_EMULATE_HARDWARE
 INSTANTIATE_TEST_CASE_P(
     NEON, FwdTrans8x8HT,
     ::testing::Values(
@@ -706,7 +709,7 @@ INSTANTIATE_TEST_CASE_P(
         make_tuple(&vp9_fht8x8_c, &vp9_iht8x8_64_add_neon, 1, VPX_BITS_8),
         make_tuple(&vp9_fht8x8_c, &vp9_iht8x8_64_add_neon, 2, VPX_BITS_8),
         make_tuple(&vp9_fht8x8_c, &vp9_iht8x8_64_add_neon, 3, VPX_BITS_8)));
-#endif  // HAVE_NEON_ASM && !CONFIG_VP9_HIGHBITDEPTH && !CONFIG_EMULATE_HARDWARE
+#endif  // HAVE_NEON && !CONFIG_VP9_HIGHBITDEPTH && !CONFIG_EMULATE_HARDWARE
 
 #if HAVE_SSE2 && !CONFIG_VP9_HIGHBITDEPTH && !CONFIG_EMULATE_HARDWARE
 // TODO(jingning): re-enable after these handle the expanded range [0, 65535]

vp9/common/vp9_rtcd_defs.pl

@@ -460,8 +460,7 @@ if (vpx_config("CONFIG_VP9_HIGHBITDEPTH") eq "yes") {
     specialize qw/vp9_iht4x4_16_add sse2 neon dspr2/;
 
     add_proto qw/void vp9_iht8x8_64_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
-    specialize qw/vp9_iht8x8_64_add sse2 neon_asm dspr2/;
-    $vp9_iht8x8_64_add_neon_asm=vp9_iht8x8_64_add_neon;
+    specialize qw/vp9_iht8x8_64_add sse2 neon dspr2/;
 
     add_proto qw/void vp9_iht16x16_256_add/, "const tran_low_t *input, uint8_t *output, int pitch, int tx_type";
     specialize qw/vp9_iht16x16_256_add sse2 dspr2/;

vp9/vp9_common.mk

@@ -135,12 +135,12 @@ endif
 
 VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_loopfilter_16_neon_asm$(ASM)
 VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_dc_only_idct_add_neon$(ASM)
-VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_iht8x8_add_neon$(ASM)
 VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_mb_lpf_neon$(ASM)
 VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_save_reg_neon$(ASM)
 VP9_COMMON_SRCS-$(HAVE_NEON_ASM) += common/arm/neon/vp9_reconintra_neon$(ASM)
 
 VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_iht4x4_add_neon.c
+VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_iht8x8_add_neon.c
 
 # neon with assembly and intrinsics implementations. If both are available
 # prefer assembly.