ref: 0b8cdcaff8eb9d81c7cb24defc1a34d2576b52a3
parent: 50daa8f737fe174d723a10f97fa7ec24f54a6178
author: Guangwei Wang <[email protected]>
date: Sun Apr 17 15:41:57 EDT 2016
extension 32-bit parameters to 64-bit on arm64 assembly function
--- a/codec/decoder/core/arm64/block_add_aarch64_neon.S
+++ b/codec/decoder/core/arm64/block_add_aarch64_neon.S
@@ -121,6 +121,7 @@
WELS_ASM_AARCH64_FUNC_BEGIN WelsBlockZero8x8_AArch64_neon
eor v0.16b, v0.16b, v0.16b
+ SIGN_EXTENSION x1, w1
lsl x1, x1, 1
.rept 8
st1 {v0.16b}, [x0], x1--- a/codec/encoder/core/arm64/reconstruct_aarch64_neon.S
+++ b/codec/encoder/core/arm64/reconstruct_aarch64_neon.S
@@ -469,7 +469,10 @@
st1 {v0.16b, v1.16b}, [x0]WELS_ASM_AARCH64_FUNC_END
+//void WelsDctT4_AArch64_neon (int16_t* pDct, uint8_t* pPixel1, int32_t iStride1, uint8_t* pPixel2, int32_t iStride2);
WELS_ASM_AARCH64_FUNC_BEGIN WelsDctT4_AArch64_neon
+ SIGN_EXTENSION x2, w2
+ SIGN_EXTENSION x4, w4
LOAD_4x4_DATA_FOR_DCT v0, v1, x1, x2, x3, x4
usubl v2.8h, v0.8b, v1.8b
usubl2 v4.8h, v0.16b, v1.16b