diff --git a/sm4/gcm_arm64.s b/sm4/gcm_arm64.s
index a4171af..bad2d71 100644
--- a/sm4/gcm_arm64.s
+++ b/sm4/gcm_arm64.s
@@ -107,76 +107,77 @@ TEXT ·gcmSm4Finish(SB),NOSPLIT,$0
 #undef dlen
 
 #define SM4_SBOX(x, y, z, z1, z2) \
-	VMOV $0x0F0F0F0F0F0F0F0F, z1.D2;               \ // nibble mask
-  VAND x.B16, z1.B16, z2.B16;                    \
+	MOVD $0x0F0F0F0F0F0F0F0F, R19;                 \
+	VMOV R19, z1.D2;                                \ // nibble mask
+	VAND x.B16, z1.B16, z2.B16;                    \
 	MOVD $0x9197E2E474720701, R19;                 \
 	VMOV R19, z.D[0];                              \ 
 	MOVD $0xC7C1B4B222245157, R19;                 \
 	VMOV R19, z.D[1];                              \ // m1 low	
-  VTBL z2.B16, [z.B16], y.B16;                   \
-  VUSHR $4, x.D2, x.D2;                          \
-  VAND x.B16, z1.B16, z2.B16;                    \
+	VTBL z2.B16, [z.B16], y.B16;                   \
+	VUSHR $4, x.D2, x.D2;                          \
+	VAND x.B16, z1.B16, z2.B16;                    \
 	MOVD $0xE240AB09EB49A200, R19;                 \
 	VMOV R19, z.D[0];                              \
 	MOVD $0xF052B91BF95BB012, R19;                 \
 	VMOV R19, z.D[1];                              \ // m1 high		
-  VTBL z2.B16, [z.B16], z2.B16;                  \
-  VEOR y.B16, z2.B16, x.B16;                     \
+	VTBL z2.B16, [z.B16], z2.B16;                  \
+	VEOR y.B16, z2.B16, x.B16;                     \
 	MOVD $0x0B0E0104070A0D00, R19;                 \ 
 	VMOV R19, z.D[0];                              \
 	MOVD $0x0306090C0F020508, R19;                 \
 	VMOV R19, z.D[1];                              \ // inverse shift row			
-  VTBL z.B16, [x.B16], x.B16;                    \
-  AESE ZERO.B16, x.B16;                          \	
-  VAND x.B16, z1.B16, z2.B16;                    \
+	VTBL z.B16, [x.B16], x.B16;                    \
+	AESE ZERO.B16, x.B16;                          \	
+	VAND x.B16, z1.B16, z2.B16;                    \
 	MOVD $0x5B67F2CEA19D0834, R19;                 \
 	VMOV R19, z.D[0];                              \
 	MOVD $0xEDD14478172BBE82, R19;                 \
 	VMOV R19, z.D[1];                              \ // m2 low		
-  VTBL z2.B16, [z.B16], y.B16;                   \
-  VUSHR $4, x.D2, x.D2;                          \
-  VAND x.B16, z1.B16, z2.B16;                    \
+	VTBL z2.B16, [z.B16], y.B16;                   \
+	VUSHR $4, x.D2, x.D2;                          \
+	VAND x.B16, z1.B16, z2.B16;                    \
 	MOVD $0xAE7201DD73AFDC00, R19;                 \
 	VMOV R19, z.D[0];                              \
 	MOVD $0x11CDBE62CC1063BF, R19;                 \
 	VMOV R19, z.D[1];                              \ // m2 high		
-  VTBL z2.B16, [z.B16], z2.B16;                  \
-  VEOR y.B16, z2.B16, x.B16
+	VTBL z2.B16, [z.B16], z2.B16;                  \
+	VEOR y.B16, z2.B16, x.B16
 
 #define SM4_TAO_L1(x, y, z, z1, z2)         \
-  SM4_SBOX(x, y, z, z1, z2);                   \
-  ;                                            \
+	SM4_SBOX(x, y, z, z1, z2);                   \
+	;                                            \
 	MOVD $0x0605040702010003, R19;               \
 	VMOV R19, z.D[0];                            \
 	MOVD $0x0E0D0C0F0A09080B, R19;               \
 	VMOV R19, z.D[1];                            \ // r08 mask
-  VTBL z.B16, [x.B16], y.B16;                  \
-  VEOR y.B16, x.B16, y.B16;                    \
+	VTBL z.B16, [x.B16], y.B16;                  \
+	VEOR y.B16, x.B16, y.B16;                    \
 	MOVD $0x0504070601000302, R19;               \
 	VMOV R19, z.D[0];                            \
 	MOVD $0x0D0C0F0E09080B0A, R19;               \
 	VMOV R19, z.D[1];                            \ // r16 mask
-  VTBL z.B16, [x.B16], z.B16;                  \
-  VEOR z.B16, y.B16, y.B16;                    \
-  VSHL $2, y.S4, z.S4;                         \
-  VUSHR $30, y.S4, y.S4;                       \
-  VORR y.B16, z.B16, y.B16;                    \
+	VTBL z.B16, [x.B16], z.B16;                  \
+	VEOR z.B16, y.B16, y.B16;                    \
+	VSHL $2, y.S4, z.S4;                         \
+	VUSHR $30, y.S4, y.S4;                       \
+	VORR y.B16, z.B16, y.B16;                    \
 	MOVD $0x0407060500030201, R19;               \
 	VMOV R19, z.D[0];                            \
 	MOVD $0x0C0F0E0D080B0A09, R19;               \
 	VMOV R19, z.D[1];                            \ // r24 mask
-  VTBL z.B16, [x.B16], z.B16;                  \
-  VEOR z.B16, x.B16, x.B16;                    \
-  VEOR y.B16, x.B16, x.B16
+	VTBL z.B16, [x.B16], z.B16;                  \
+	VEOR z.B16, x.B16, x.B16;                    \
+	VEOR y.B16, x.B16, x.B16
 
 #define SM4_ROUND(RK, x, y, z, z1, z2, t0, t1, t2, t3)  \ 
-  MOVW.P 4(RK), R19;                                \
-  VMOV R19, x.S4;                                   \
-  VEOR t1.B16, x.B16, x.B16;                        \
-  VEOR t2.B16, x.B16, x.B16;                        \
-  VEOR t3.B16, x.B16, x.B16;                        \
-  SM4_TAO_L1(x, y, z, z1, z2);                      \
-  VEOR x.B16, t0.B16, t0.B16
+	MOVW.P 4(RK), R19;                                \
+	VMOV R19, x.S4;                                   \
+	VEOR t1.B16, x.B16, x.B16;                        \
+	VEOR t2.B16, x.B16, x.B16;                        \
+	VEOR t3.B16, x.B16, x.B16;                        \
+	SM4_TAO_L1(x, y, z, z1, z2);                      \
+	VEOR x.B16, t0.B16, t0.B16
 
 // func gcmSm4Init(productTable *[256]byte, rk []uint32)
 TEXT ·gcmSm4Init(SB),NOSPLIT,$0
@@ -194,7 +195,7 @@ TEXT ·gcmSm4Init(SB),NOSPLIT,$0
 	VMOV	I, POLY.D[1]
 	VEOR	ZERO.B16, ZERO.B16, ZERO.B16
 
-  // Encrypt block 0 with the SM4 keys to generate the hash key H
+	// Encrypt block 0 with the SM4 keys to generate the hash key H
 	VEOR	B0.B16, B0.B16, B0.B16
 	VEOR	B1.B16, B1.B16, B1.B16
 	VEOR	B2.B16, B2.B16, B2.B16
@@ -207,9 +208,9 @@ sm4InitEncLoop:
 	SM4_ROUND(RK, K0, K1, K2, K3, K4, B2, B3, B0, B1)
 	SM4_ROUND(RK, K0, K1, K2, K3, K4, B3, B0, B1, B2)
 
-  ADD $16, R3
-  CMP $128, R3
-  BNE sm4InitEncLoop
+	ADD $16, R3
+	CMP $128, R3
+	BNE sm4InitEncLoop
 
 	VMOV B1.S[0], B0.S[1]
 	VMOV B2.S[0], B0.S[2]