diff --git a/.github/workflows/test_qemu.yml b/.github/workflows/test_qemu.yml
index 3d2aadf..632aff1 100644
--- a/.github/workflows/test_qemu.yml
+++ b/.github/workflows/test_qemu.yml
@@ -30,7 +30,7 @@ jobs:
       uses: actions/checkout@v3
       
     - name: Test
-      run: go test -v -short ./...
+      run: go test -v -short ./cipher/...
       env:
         DISABLE_SM3NI: 1
         DISABLE_SM4NI: 1
diff --git a/sm4/sm4_xts.go b/sm4/sm4_xts.go
index 4da3151..8ca879b 100644
--- a/sm4/sm4_xts.go
+++ b/sm4/sm4_xts.go
@@ -1,5 +1,5 @@
-//go:build amd64 && !purego
-// +build amd64,!purego
+//go:build (amd64 && !purego) || (arm64 && !purego)
+// +build amd64,!purego arm64,!purego
 
 package sm4
 
diff --git a/sm4/xts_arm64.s b/sm4/xts_arm64.s
new file mode 100644
index 0000000..07e7cea
--- /dev/null
+++ b/sm4/xts_arm64.s
@@ -0,0 +1,916 @@
+//go:build arm64 && !purego
+// +build arm64,!purego
+
+#include "textflag.h"
+
+#define B0 V0
+#define B1 V1
+#define B2 V2
+#define B3 V3
+#define B4 V4
+#define B5 V5
+#define B6 V6
+#define B7 V7
+
+#define T0 V8
+#define T1 V9
+#define T2 V10
+#define T3 V11
+#define T4 V12
+#define T5 V13
+#define T6 V14
+#define T7 V15
+
+#define POLY V16
+#define ZERO V17
+#define TW V18
+
+#define K0 V19
+#define K1 V20
+#define K2 V21
+#define K3 V22
+
+#define NIBBLE_MASK V23
+#define INVERSE_SHIFT_ROWS V24
+#define M1L V25
+#define M1H V26 
+#define M2L V27 
+#define M2H V28
+#define R08_MASK V29 
+#define R16_MASK V30
+#define R24_MASK V31
+
+#include "aesni_macros_arm64.s"
+
+#define mul2Inline        \
+	VMOV	TW.D[1], I;                     \
+	ASR	$63, I;                             \
+	VMOV	I, K0.D2;                       \
+	VAND	POLY.B16, K0.B16, K0.B16;       \
+	\
+	VUSHR	$63, TW.D2, K1.D2;              \
+	VEXT	$8, K1.B16, ZERO.B16, K1.B16;   \
+	VSHL	$1, TW.D2, TW.D2;               \
+	VEOR	K0.B16, TW.B16, TW.B16;         \
+	VEOR	K1.B16, TW.B16, TW.B16
+
+#define mul2GBInline        \
+	VREV64 TW.B16, TW.B16;                  \
+	VEXT	$8, TW.B16, TW.B16, TW.B16;     \
+	\
+	VMOV	TW.D[0], I;                     \
+	LSL $63, I;                             \
+	ASR $63, I;                             \
+	VMOV	I, K0.D2;                       \
+	VAND	POLY.B16, K0.B16, K0.B16;       \
+	\
+	VSHL $63, TW.D2, K1.D2;                 \
+	VEXT	$8, ZERO.B16, K1.B16, K1.B16;   \
+	VUSHR	$1, TW.D2, TW.D2;               \
+	VEOR	K0.B16, TW.B16, TW.B16;         \
+	VEOR	K1.B16, TW.B16, TW.B16;         \
+	\
+	VEXT	$8, TW.B16, TW.B16, TW.B16;     \
+	VREV64 TW.B16, TW.B16
+
+#define prepare4Tweaks \
+	VMOV TW.B16, T0.B16;   \
+	mul2Inline;            \
+	VMOV TW.B16, T1.B16;   \
+	mul2Inline;            \
+	VMOV TW.B16, T2.B16;   \
+	mul2Inline;            \
+	VMOV TW.B16, T3.B16;   \
+	mul2Inline
+
+#define prepare8Tweaks \
+	prepare4Tweaks;        \
+	VMOV TW.B16, T4.B16;   \
+	mul2Inline;            \
+	VMOV TW.B16, T5.B16;   \
+	mul2Inline;            \
+	VMOV TW.B16, T6.B16;   \
+	mul2Inline;            \
+	VMOV TW.B16, T7.B16;   \
+	mul2Inline
+
+#define prepareGB4Tweaks \
+	VMOV TW.B16, T0.B16;     \
+	mul2GBInline;            \
+	VMOV TW.B16, T1.B16;     \
+	mul2GBInline;            \
+	VMOV TW.B16, T2.B16;     \
+	mul2GBInline;            \
+	VMOV TW.B16, T3.B16;     \
+	mul2GBInline
+
+#define prepareGB8Tweaks \
+	prepareGB4Tweaks;        \
+	VMOV TW.B16, T4.B16;     \
+	mul2GBInline;            \
+	VMOV TW.B16, T5.B16;     \
+	mul2GBInline;            \
+	VMOV TW.B16, T6.B16;     \
+	mul2GBInline;            \
+	VMOV TW.B16, T7.B16;     \
+	mul2GBInline
+
+#define load8blocks \
+	VLD1.P 64(srcPtr), [B0.S4, B1.S4, B2.S4, B3.S4]; \
+	VEOR T0.B16, B0.B16, B0.B16; \
+	VEOR T1.B16, B1.B16, B1.B16; \
+	VEOR T2.B16, B2.B16, B2.B16; \
+	VEOR T3.B16, B3.B16, B3.B16; \
+	\
+	VLD1.P 64(srcPtr), [B4.S4, B5.S4, B6.S4, B7.S4]; \
+	VEOR T4.B16, B4.B16, B4.B16; \
+	VEOR T5.B16, B5.B16, B5.B16; \
+	VEOR T6.B16, B6.B16, B6.B16; \
+	VEOR T7.B16, B7.B16, B7.B16; \
+	\
+	VREV32 B0.B16, B0.B16; \
+	VREV32 B1.B16, B1.B16; \
+	VREV32 B2.B16, B2.B16; \
+	VREV32 B3.B16, B3.B16; \
+	VREV32 B4.B16, B4.B16; \
+	VREV32 B5.B16, B5.B16; \
+	VREV32 B6.B16, B6.B16; \
+	VREV32 B7.B16, B7.B16; \
+	\
+	PRE_TRANSPOSE_MATRIX(B0, B1, B2, B3, K0, K1, K2, K3); \
+	PRE_TRANSPOSE_MATRIX(B4, B5, B6, B7, K0, K1, K2, K3)
+
+#define store8blocks \
+	VREV32 B0.B16, B0.B16; \
+	VREV32 B1.B16, B1.B16; \
+	VREV32 B2.B16, B2.B16; \
+	VREV32 B3.B16, B3.B16; \
+	TRANSPOSE_MATRIX(B0, B1, B2, B3, K0, K1, K2, K3); \
+	VREV32 B4.B16, B4.B16; \
+	VREV32 B5.B16, B5.B16; \
+	VREV32 B6.B16, B6.B16; \
+	VREV32 B7.B16, B7.B16; \
+	TRANSPOSE_MATRIX(B4, B5, B6, B7, K0, K1, K2, K3); \
+	\
+	VEOR T0.B16, B0.B16, B0.B16; \
+	VEOR T1.B16, B1.B16, B1.B16; \
+	VEOR T2.B16, B2.B16, B2.B16; \
+	VEOR T3.B16, B3.B16, B3.B16; \
+	VEOR T4.B16, B4.B16, B4.B16; \
+	VEOR T5.B16, B5.B16, B5.B16; \
+	VEOR T6.B16, B6.B16, B6.B16; \
+	VEOR T7.B16, B7.B16, B7.B16; \
+	\
+	VST1.P [B0.S4, B1.S4, B2.S4, B3.S4], 64(dstPtr); \
+	VST1.P [B4.S4, B5.S4, B6.S4, B7.S4], 64(dstPtr)
+
+#define load4blocks \
+	VLD1.P 64(srcPtr), [B0.S4, B1.S4, B2.S4, B3.S4]; \
+	VEOR T0.B16, B0.B16, B0.B16; \
+	VEOR T1.B16, B1.B16, B1.B16; \
+	VEOR T2.B16, B2.B16, B2.B16; \
+	VEOR T3.B16, B3.B16, B3.B16; \
+	\
+	VREV32 B0.B16, B0.B16; \
+	VREV32 B1.B16, B1.B16; \
+	VREV32 B2.B16, B2.B16; \
+	VREV32 B3.B16, B3.B16; \
+	PRE_TRANSPOSE_MATRIX(B0, B1, B2, B3, K0, K1, K2, K3)
+
+#define store4blocks \
+	VREV32 B0.B16, B0.B16; \
+	VREV32 B1.B16, B1.B16; \
+	VREV32 B2.B16, B2.B16; \
+	VREV32 B3.B16, B3.B16; \
+	TRANSPOSE_MATRIX(B0, B1, B2, B3, K0, K1, K2, K3); \
+	\
+	VEOR T0.B16, B0.B16, B0.B16; \
+	VEOR T1.B16, B1.B16, B1.B16; \
+	VEOR T2.B16, B2.B16, B2.B16; \
+	VEOR T3.B16, B3.B16, B3.B16; \
+	\
+	VST1.P [B0.S4, B1.S4, B2.S4, B3.S4], 64(dstPtr)
+
+#define loadOneBlock \
+	VLD1.P 16(srcPtr), [B0.S4]; \
+	VEOR TW.B16, B0.B16, B0.B16; \
+	\
+	VREV32 B0.B16, B0.B16; \
+	VMOV B0.S[1], B1.S[0]; \
+	VMOV B0.S[2], B2.S[0]; \
+	VMOV B0.S[3], B3.S[0]
+
+#define storeOneBlock \
+	VMOV B2.S[0], B3.S[1]; \
+	VMOV B1.S[0], B3.S[2]; \
+	VMOV B0.S[0], B3.S[3]; \
+	VREV32 B3.B16, B3.B16; \
+	\
+	VEOR TW.B16, B3.B16, B3.B16; \
+	VST1.P [B3.S4], 16(dstPtr)
+
+#define dstPtr R2
+#define srcPtr R3
+#define rk R0
+#define twPtr R1
+#define srcPtrLen R4
+#define I R5
+#define rkSave R6
+
+// func encryptSm4Xts(xk *uint32, tweak *[BlockSize]byte, dst, src []byte)
+TEXT ·encryptSm4Xts(SB),0,$128-64
+	LOAD_SM4_AESNI_CONSTS()
+	MOVD xk+0(FP), rk
+	MOVD tweak+8(FP), twPtr
+	MOVD dst+16(FP), dstPtr
+	MOVD src+40(FP), srcPtr
+	MOVD src_len+48(FP), srcPtrLen
+
+	VEOR	POLY.B16, POLY.B16, POLY.B16
+	VEOR	ZERO.B16, ZERO.B16, ZERO.B16
+
+	MOVD	$0x87, I
+	VMOV	I, POLY.D[0]
+
+	MOVD rk, rkSave
+	VLD1 (twPtr), [TW.B16]
+
+xtsSm4EncOctets:
+	CMP	$128, srcPtrLen
+	BLT	xtsSm4EncNibbles
+	SUB	$128, srcPtrLen
+
+	prepare8Tweaks
+	load8blocks
+	MOVD rkSave, rk
+	EOR R13, R13
+
+encOctetsEnc8Blocks:
+			SM4_8BLOCKS_ROUND(rk, R19, K0, K1, K2, K3, B0, B1, B2, B3, B4, B5, B6, B7)
+			SM4_8BLOCKS_ROUND(rk, R19, K0, K1, K2, K3, B1, B2, B3, B0, B5, B6, B7, B4)
+			SM4_8BLOCKS_ROUND(rk, R19, K0, K1, K2, K3, B2, B3, B0, B1, B6, B7, B4, B5)
+			SM4_8BLOCKS_ROUND(rk, R19, K0, K1, K2, K3, B3, B0, B1, B2, B7, B4, B5, B6)
+
+		ADD $1, R13
+		CMP $8, R13
+		BNE encOctetsEnc8Blocks
+
+	store8blocks
+	B	xtsSm4EncOctets
+
+xtsSm4EncNibbles:
+	CMP	$64, srcPtrLen
+	BLT	xtsSm4EncSingles
+	SUB	$64, srcPtrLen
+
+	prepare4Tweaks
+	load4blocks
+	MOVD rkSave, rk
+	EOR R13, R13
+
+encNibblesEnc4Blocks:
+		SM4_ROUND(rk, R19, K0, K1, K2, B0, B1, B2, B3)
+		SM4_ROUND(rk, R19, K0, K1, K2, B1, B2, B3, B0)
+		SM4_ROUND(rk, R19, K0, K1, K2, B2, B3, B0, B1)
+		SM4_ROUND(rk, R19, K0, K1, K2, B3, B0, B1, B2)
+
+		ADD $1, R13
+		CMP $8, R13
+		BNE encNibblesEnc4Blocks
+	
+	store4blocks
+
+xtsSm4EncSingles:
+	CMP	$16, srcPtrLen
+	BLT	xtsSm4EncTail
+	SUB	$16, srcPtrLen
+
+	loadOneBlock
+
+	MOVD rkSave, rk
+	EOR R13, R13
+
+encSinglesEnc4Blocks:	
+		SM4_ROUND(rk, R19, K0, K1, K2, B0, B1, B2, B3)
+		SM4_ROUND(rk, R19, K0, K1, K2, B1, B2, B3, B0)
+		SM4_ROUND(rk, R19, K0, K1, K2, B2, B3, B0, B1)
+		SM4_ROUND(rk, R19, K0, K1, K2, B3, B0, B1, B2)
+		ADD $1, R13
+		CMP $8, R13
+		BNE encSinglesEnc4Blocks
+
+	storeOneBlock
+	mul2Inline
+	B	xtsSm4EncSingles
+
+xtsSm4EncTail:
+	CBZ	srcPtrLen, xtsSm4EncDone
+	SUB $16, dstPtr, R7
+	MOVD R7, R9
+	MOVD RSP, R8
+	VLD1 (R7), [B0.B16]
+	VST1 [B0.B16], (R8)
+
+	TBZ	$3, srcPtrLen, less_than8
+	MOVD.P 8(srcPtr), R11
+	MOVD.P R11, 8(R8)
+	MOVD.P 8(R7), R12
+	MOVD.P R12, 8(dstPtr)
+
+less_than8:
+	TBZ	$2, srcPtrLen, less_than4
+	MOVWU.P 4(srcPtr), R11
+	MOVWU.P R11, 4(R8)
+	MOVWU.P 4(R7), R12
+	MOVWU.P R12, 4(dstPtr)
+
+less_than4:
+	TBZ	$1, srcPtrLen, less_than2
+	MOVHU.P 2(srcPtr), R11
+	MOVHU.P R11, 2(R8)
+	MOVHU.P 2(R7), R12
+	MOVHU.P R12, 2(dstPtr)
+
+less_than2:
+	TBZ	$0, srcPtrLen, xtsSm4EncTailEnc
+	MOVBU (srcPtr), R11
+	MOVBU R11, (R8)
+	MOVBU (R7), R12
+	MOVBU R12, (dstPtr)
+
+xtsSm4EncTailEnc:
+	VLD1 (RSP), [B0.B16]
+	VEOR TW.B16, B0.B16, B0.B16
+	VREV32 B0.B16, B0.B16
+	VMOV B0.S[1], B1.S[0]
+	VMOV B0.S[2], B2.S[0]
+	VMOV B0.S[3], B3.S[0]
+
+	MOVD rkSave, rk
+	EOR R13, R13
+
+tailEncLoop:
+		SM4_ROUND(rk, R19, K0, K1, K2, B0, B1, B2, B3)
+		SM4_ROUND(rk, R19, K0, K1, K2, B1, B2, B3, B0)
+		SM4_ROUND(rk, R19, K0, K1, K2, B2, B3, B0, B1)
+		SM4_ROUND(rk, R19, K0, K1, K2, B3, B0, B1, B2)
+		ADD $1, R13
+		CMP $8, R13
+		BNE tailEncLoop
+
+	VMOV B2.S[0], B3.S[1]
+	VMOV B1.S[0], B3.S[2]
+	VMOV B0.S[0], B3.S[3]
+	VREV32 B3.B16, B3.B16
+
+	VEOR TW.B16, B3.B16, B3.B16
+	VST1 [B3.B16], (R9)
+
+xtsSm4EncDone:
+	VST1 [TW.B16], (twPtr)
+	RET
+
+// func encryptSm4XtsGB(xk *uint32, tweak *[BlockSize]byte, dst, src []byte)
+TEXT ·encryptSm4XtsGB(SB),0,$128-64
+	LOAD_SM4_AESNI_CONSTS()
+	MOVD xk+0(FP), rk
+	MOVD tweak+8(FP), twPtr
+	MOVD dst+16(FP), dstPtr
+	MOVD src+40(FP), srcPtr
+	MOVD src_len+48(FP), srcPtrLen
+
+	VEOR	POLY.B16, POLY.B16, POLY.B16
+	VEOR	ZERO.B16, ZERO.B16, ZERO.B16
+
+	MOVD	$0xE1, I
+	LSL	$56, I
+	VMOV	I, POLY.D[1]
+
+	MOVD rk, rkSave
+	VLD1 (twPtr), [TW.B16]
+
+xtsSm4EncOctets:
+	CMP	$128, srcPtrLen
+	BLT	xtsSm4EncNibbles
+	SUB	$128, srcPtrLen
+
+	prepareGB8Tweaks
+	load8blocks
+	MOVD rkSave, rk
+	EOR R13, R13
+
+encOctetsEnc8Blocks:
+			SM4_8BLOCKS_ROUND(rk, R19, K0, K1, K2, K3, B0, B1, B2, B3, B4, B5, B6, B7)
+			SM4_8BLOCKS_ROUND(rk, R19, K0, K1, K2, K3, B1, B2, B3, B0, B5, B6, B7, B4)
+			SM4_8BLOCKS_ROUND(rk, R19, K0, K1, K2, K3, B2, B3, B0, B1, B6, B7, B4, B5)
+			SM4_8BLOCKS_ROUND(rk, R19, K0, K1, K2, K3, B3, B0, B1, B2, B7, B4, B5, B6)
+
+		ADD $1, R13
+		CMP $8, R13
+		BNE encOctetsEnc8Blocks
+
+	store8blocks
+	B	xtsSm4EncOctets
+
+xtsSm4EncNibbles:
+	CMP	$64, srcPtrLen
+	BLT	xtsSm4EncSingles
+	SUB	$64, srcPtrLen
+
+	prepareGB4Tweaks
+	load4blocks
+	MOVD rkSave, rk
+	EOR R13, R13
+
+encNibblesEnc4Blocks:
+		SM4_ROUND(rk, R19, K0, K1, K2, B0, B1, B2, B3)
+		SM4_ROUND(rk, R19, K0, K1, K2, B1, B2, B3, B0)
+		SM4_ROUND(rk, R19, K0, K1, K2, B2, B3, B0, B1)
+		SM4_ROUND(rk, R19, K0, K1, K2, B3, B0, B1, B2)
+
+		ADD $1, R13
+		CMP $8, R13
+		BNE encNibblesEnc4Blocks
+	
+	store4blocks
+
+xtsSm4EncSingles:
+	CMP	$16, srcPtrLen
+	BLT	xtsSm4EncTail
+	SUB	$16, srcPtrLen
+
+	loadOneBlock
+
+	MOVD rkSave, rk
+	EOR R13, R13
+
+encSinglesEnc4Blocks:	
+		SM4_ROUND(rk, R19, K0, K1, K2, B0, B1, B2, B3)
+		SM4_ROUND(rk, R19, K0, K1, K2, B1, B2, B3, B0)
+		SM4_ROUND(rk, R19, K0, K1, K2, B2, B3, B0, B1)
+		SM4_ROUND(rk, R19, K0, K1, K2, B3, B0, B1, B2)
+		ADD $1, R13
+		CMP $8, R13
+		BNE encSinglesEnc4Blocks
+
+	storeOneBlock
+	mul2GBInline
+	B	xtsSm4EncSingles
+
+xtsSm4EncTail:
+	CBZ	srcPtrLen, xtsSm4EncDone
+	SUB $16, dstPtr, R7
+	MOVD R7, R9
+	MOVD RSP, R8
+	VLD1 (R7), [B0.B16]
+	VST1 [B0.B16], (R8)
+
+	TBZ	$3, srcPtrLen, less_than8
+	MOVD.P 8(srcPtr), R11
+	MOVD.P R11, 8(R8)
+	MOVD.P 8(R7), R12
+	MOVD.P R12, 8(dstPtr)
+
+less_than8:
+	TBZ	$2, srcPtrLen, less_than4
+	MOVWU.P 4(srcPtr), R11
+	MOVWU.P R11, 4(R8)
+	MOVWU.P 4(R7), R12
+	MOVWU.P R12, 4(dstPtr)
+
+less_than4:
+	TBZ	$1, srcPtrLen, less_than2
+	MOVHU.P 2(srcPtr), R11
+	MOVHU.P R11, 2(R8)
+	MOVHU.P 2(R7), R12
+	MOVHU.P R12, 2(dstPtr)
+
+less_than2:
+	TBZ	$0, srcPtrLen, xtsSm4EncTailEnc
+	MOVBU (srcPtr), R11
+	MOVBU R11, (R8)
+	MOVBU (R7), R12
+	MOVBU R12, (dstPtr)
+
+xtsSm4EncTailEnc:
+	VLD1 (RSP), [B0.B16]
+	VEOR TW.B16, B0.B16, B0.B16
+	VREV32 B0.B16, B0.B16
+	VMOV B0.S[1], B1.S[0]
+	VMOV B0.S[2], B2.S[0]
+	VMOV B0.S[3], B3.S[0]
+
+	MOVD rkSave, rk
+	EOR R13, R13
+
+tailEncLoop:
+		SM4_ROUND(rk, R19, K0, K1, K2, B0, B1, B2, B3)
+		SM4_ROUND(rk, R19, K0, K1, K2, B1, B2, B3, B0)
+		SM4_ROUND(rk, R19, K0, K1, K2, B2, B3, B0, B1)
+		SM4_ROUND(rk, R19, K0, K1, K2, B3, B0, B1, B2)
+		ADD $1, R13
+		CMP $8, R13
+		BNE tailEncLoop
+
+	VMOV B2.S[0], B3.S[1]
+	VMOV B1.S[0], B3.S[2]
+	VMOV B0.S[0], B3.S[3]
+	VREV32 B3.B16, B3.B16
+
+	VEOR TW.B16, B3.B16, B3.B16
+	VST1 [B3.B16], (R9)
+
+xtsSm4EncDone:
+	VST1 [TW.B16], (twPtr)
+	RET
+
+// func decryptSm4Xts(xk *uint32, tweak *[BlockSize]byte, dst, src []byte)
+TEXT ·decryptSm4Xts(SB),0,$128-64
+	LOAD_SM4_AESNI_CONSTS()
+	MOVD xk+0(FP), rk
+	MOVD tweak+8(FP), twPtr
+	MOVD dst+16(FP), dstPtr
+	MOVD src+40(FP), srcPtr
+	MOVD src_len+48(FP), srcPtrLen
+
+	VEOR	POLY.B16, POLY.B16, POLY.B16
+	VEOR	ZERO.B16, ZERO.B16, ZERO.B16
+
+	MOVD	$0x87, I
+	VMOV	I, POLY.D[0]
+
+	MOVD rk, rkSave
+	VLD1 (twPtr), [TW.B16]
+
+xtsSm4DecOctets:
+	CMP	$128, srcPtrLen
+	BLT	xtsSm4DecNibbles
+	SUB	$128, srcPtrLen
+
+	prepare8Tweaks
+	load8blocks
+	MOVD rkSave, rk
+	EOR R13, R13
+
+decOctetsDec8Blocks:
+			SM4_8BLOCKS_ROUND(rk, R19, K0, K1, K2, K3, B0, B1, B2, B3, B4, B5, B6, B7)
+			SM4_8BLOCKS_ROUND(rk, R19, K0, K1, K2, K3, B1, B2, B3, B0, B5, B6, B7, B4)
+			SM4_8BLOCKS_ROUND(rk, R19, K0, K1, K2, K3, B2, B3, B0, B1, B6, B7, B4, B5)
+			SM4_8BLOCKS_ROUND(rk, R19, K0, K1, K2, K3, B3, B0, B1, B2, B7, B4, B5, B6)
+
+		ADD $1, R13
+		CMP $8, R13
+		BNE decOctetsDec8Blocks
+
+	store8blocks
+	B	xtsSm4DecOctets
+
+xtsSm4DecNibbles:
+	CMP	$64, srcPtrLen
+	BLT	xtsSm4DecSingles
+	SUB	$64, srcPtrLen
+
+	prepare4Tweaks
+	load4blocks
+	MOVD rkSave, rk
+	EOR R13, R13
+
+decNibblesDec4Blocks:
+		SM4_ROUND(rk, R19, K0, K1, K2, B0, B1, B2, B3)
+		SM4_ROUND(rk, R19, K0, K1, K2, B1, B2, B3, B0)
+		SM4_ROUND(rk, R19, K0, K1, K2, B2, B3, B0, B1)
+		SM4_ROUND(rk, R19, K0, K1, K2, B3, B0, B1, B2)
+
+		ADD $1, R13
+		CMP $8, R13
+		BNE decNibblesDec4Blocks
+	
+	store4blocks
+
+xtsSm4DecSingles:
+	CMP	$32, srcPtrLen
+	BLT	xtsSm4DecSingles
+	SUB	$16, srcPtrLen
+
+	loadOneBlock
+
+	MOVD rkSave, rk
+	EOR R13, R13
+
+decSinglesDec4Blocks:	
+		SM4_ROUND(rk, R19, K0, K1, K2, B0, B1, B2, B3)
+		SM4_ROUND(rk, R19, K0, K1, K2, B1, B2, B3, B0)
+		SM4_ROUND(rk, R19, K0, K1, K2, B2, B3, B0, B1)
+		SM4_ROUND(rk, R19, K0, K1, K2, B3, B0, B1, B2)
+		ADD $1, R13
+		CMP $8, R13
+		BNE decSinglesDec4Blocks
+
+	storeOneBlock
+	mul2Inline
+
+	B	xtsSm4DecSingles
+
+xtsSm4DecTail:
+	CBZ	srcPtrLen, xtsSm4DecDone
+	
+	CMP	$16, srcPtrLen
+	BEQ xtsSm4DecLastBlock
+
+	VMOV TW.B16, B4.B16
+	mul2Inline
+	loadOneBlock
+	MOVD rkSave, rk
+	EOR R13, R13
+
+decLastCompleteBlockLoop:	
+		SM4_ROUND(rk, R19, K0, K1, K2, B0, B1, B2, B3)
+		SM4_ROUND(rk, R19, K0, K1, K2, B1, B2, B3, B0)
+		SM4_ROUND(rk, R19, K0, K1, K2, B2, B3, B0, B1)
+		SM4_ROUND(rk, R19, K0, K1, K2, B3, B0, B1, B2)
+		ADD $1, R13
+		CMP $8, R13
+		BNE decLastCompleteBlockLoop
+	storeOneBlock
+	VMOV B4.B16, TW.B16
+	VST1 [B3.B16], (RSP)
+
+	SUB	$16, srcPtrLen
+	SUB $16, dstPtr, R7
+	MOVD R7, R9
+	MOVD RSP, R8
+
+	TBZ	$3, srcPtrLen, less_than8
+	MOVD.P 8(srcPtr), R11
+	MOVD.P R11, 8(R8)
+	MOVD.P 8(R7), R12
+	MOVD.P R12, 8(dstPtr)
+
+less_than8:
+	TBZ	$2, srcPtrLen, less_than4
+	MOVWU.P 4(srcPtr), R11
+	MOVWU.P R11, 4(R8)
+	MOVWU.P 4(R7), R12
+	MOVWU.P R12, 4(dstPtr)
+
+less_than4:
+	TBZ	$1, srcPtrLen, less_than2
+	MOVHU.P 2(srcPtr), R11
+	MOVHU.P R11, 2(R8)
+	MOVHU.P 2(R7), R12
+	MOVHU.P R12, 2(dstPtr)
+
+less_than2:
+	TBZ	$0, srcPtrLen, xtsSm4DecTailEnc
+	MOVBU (srcPtr), R11
+	MOVBU R11, (R8)
+	MOVBU (R7), R12
+	MOVBU R12, (dstPtr)
+
+xtsSm4DecTailEnc:
+	VLD1 (RSP), [B0.B16]
+	VEOR TW.B16, B0.B16, B0.B16
+	VREV32 B0.B16, B0.B16
+	VMOV B0.S[1], B1.S[0]
+	VMOV B0.S[2], B2.S[0]
+	VMOV B0.S[3], B3.S[0]
+
+	MOVD rkSave, rk
+	EOR R13, R13
+
+tailDecLoop:
+		SM4_ROUND(rk, R19, K0, K1, K2, B0, B1, B2, B3)
+		SM4_ROUND(rk, R19, K0, K1, K2, B1, B2, B3, B0)
+		SM4_ROUND(rk, R19, K0, K1, K2, B2, B3, B0, B1)
+		SM4_ROUND(rk, R19, K0, K1, K2, B3, B0, B1, B2)
+		ADD $1, R13
+		CMP $8, R13
+		BNE tailDecLoop
+
+	VMOV B2.S[0], B3.S[1]
+	VMOV B1.S[0], B3.S[2]
+	VMOV B0.S[0], B3.S[3]
+	VREV32 B3.B16, B3.B16
+
+	VEOR TW.B16, B3.B16, B3.B16
+	VST1 [B3.B16], (R9)
+
+	B xtsSm4DecDone
+
+xtsSm4DecLastBlock:
+	loadOneBlock
+
+	MOVD rkSave, rk
+	EOR R13, R13
+
+decLastBlockLoop:	
+		SM4_ROUND(rk, R19, K0, K1, K2, B0, B1, B2, B3)
+		SM4_ROUND(rk, R19, K0, K1, K2, B1, B2, B3, B0)
+		SM4_ROUND(rk, R19, K0, K1, K2, B2, B3, B0, B1)
+		SM4_ROUND(rk, R19, K0, K1, K2, B3, B0, B1, B2)
+		ADD $1, R13
+		CMP $8, R13
+		BNE decLastBlockLoop
+
+	storeOneBlock
+	mul2Inline
+
+xtsSm4DecDone:
+	VST1 [TW.B16], (twPtr)
+	RET
+
+// func decryptSm4XtsGB(xk *uint32, tweak *[BlockSize]byte, dst, src []byte)
+TEXT ·decryptSm4XtsGB(SB),0,$128-64
+	LOAD_SM4_AESNI_CONSTS()
+	MOVD xk+0(FP), rk
+	MOVD tweak+8(FP), twPtr
+	MOVD dst+16(FP), dstPtr
+	MOVD src+40(FP), srcPtr
+	MOVD src_len+48(FP), srcPtrLen
+
+	VEOR	POLY.B16, POLY.B16, POLY.B16
+	VEOR	ZERO.B16, ZERO.B16, ZERO.B16
+
+	MOVD	$0xE1, I
+	LSL	$56, I
+	VMOV	I, POLY.D[1]
+
+	MOVD rk, rkSave
+	VLD1 (twPtr), [TW.B16]
+
+xtsSm4DecOctets:
+	CMP	$128, srcPtrLen
+	BLT	xtsSm4DecNibbles
+	SUB	$128, srcPtrLen
+
+	prepareGB8Tweaks
+	load8blocks
+	MOVD rkSave, rk
+	EOR R13, R13
+
+decOctetsDec8Blocks:
+			SM4_8BLOCKS_ROUND(rk, R19, K0, K1, K2, K3, B0, B1, B2, B3, B4, B5, B6, B7)
+			SM4_8BLOCKS_ROUND(rk, R19, K0, K1, K2, K3, B1, B2, B3, B0, B5, B6, B7, B4)
+			SM4_8BLOCKS_ROUND(rk, R19, K0, K1, K2, K3, B2, B3, B0, B1, B6, B7, B4, B5)
+			SM4_8BLOCKS_ROUND(rk, R19, K0, K1, K2, K3, B3, B0, B1, B2, B7, B4, B5, B6)
+
+		ADD $1, R13
+		CMP $8, R13
+		BNE decOctetsDec8Blocks
+
+	store8blocks
+	B	xtsSm4DecOctets
+
+xtsSm4DecNibbles:
+	CMP	$64, srcPtrLen
+	BLT	xtsSm4DecSingles
+	SUB	$64, srcPtrLen
+
+	prepareGB4Tweaks
+	load4blocks
+	MOVD rkSave, rk
+	EOR R13, R13
+
+decNibblesDec4Blocks:
+		SM4_ROUND(rk, R19, K0, K1, K2, B0, B1, B2, B3)
+		SM4_ROUND(rk, R19, K0, K1, K2, B1, B2, B3, B0)
+		SM4_ROUND(rk, R19, K0, K1, K2, B2, B3, B0, B1)
+		SM4_ROUND(rk, R19, K0, K1, K2, B3, B0, B1, B2)
+
+		ADD $1, R13
+		CMP $8, R13
+		BNE decNibblesDec4Blocks
+	
+	store4blocks
+
+xtsSm4DecSingles:
+	CMP	$32, srcPtrLen
+	BLT	xtsSm4DecSingles
+	SUB	$16, srcPtrLen
+
+	loadOneBlock
+
+	MOVD rkSave, rk
+	EOR R13, R13
+
+decSinglesDec4Blocks:	
+		SM4_ROUND(rk, R19, K0, K1, K2, B0, B1, B2, B3)
+		SM4_ROUND(rk, R19, K0, K1, K2, B1, B2, B3, B0)
+		SM4_ROUND(rk, R19, K0, K1, K2, B2, B3, B0, B1)
+		SM4_ROUND(rk, R19, K0, K1, K2, B3, B0, B1, B2)
+		ADD $1, R13
+		CMP $8, R13
+		BNE decSinglesDec4Blocks
+
+	storeOneBlock
+	mul2GBInline
+
+	B	xtsSm4DecSingles
+
+xtsSm4DecTail:
+	CBZ	srcPtrLen, xtsSm4DecDone
+	
+	CMP	$16, srcPtrLen
+	BEQ xtsSm4DecLastBlock
+
+	VMOV TW.B16, B4.B16
+	mul2GBInline
+	loadOneBlock
+	MOVD rkSave, rk
+	EOR R13, R13
+
+decLastCompleteBlockLoop:	
+		SM4_ROUND(rk, R19, K0, K1, K2, B0, B1, B2, B3)
+		SM4_ROUND(rk, R19, K0, K1, K2, B1, B2, B3, B0)
+		SM4_ROUND(rk, R19, K0, K1, K2, B2, B3, B0, B1)
+		SM4_ROUND(rk, R19, K0, K1, K2, B3, B0, B1, B2)
+		ADD $1, R13
+		CMP $8, R13
+		BNE decLastCompleteBlockLoop
+	storeOneBlock
+	VMOV B4.B16, TW.B16
+	VST1 [B3.B16], (RSP)
+
+	SUB	$16, srcPtrLen
+	SUB $16, dstPtr, R7
+	MOVD R7, R9
+	MOVD RSP, R8
+
+	TBZ	$3, srcPtrLen, less_than8
+	MOVD.P 8(srcPtr), R11
+	MOVD.P R11, 8(R8)
+	MOVD.P 8(R7), R12
+	MOVD.P R12, 8(dstPtr)
+
+less_than8:
+	TBZ	$2, srcPtrLen, less_than4
+	MOVWU.P 4(srcPtr), R11
+	MOVWU.P R11, 4(R8)
+	MOVWU.P 4(R7), R12
+	MOVWU.P R12, 4(dstPtr)
+
+less_than4:
+	TBZ	$1, srcPtrLen, less_than2
+	MOVHU.P 2(srcPtr), R11
+	MOVHU.P R11, 2(R8)
+	MOVHU.P 2(R7), R12
+	MOVHU.P R12, 2(dstPtr)
+
+less_than2:
+	TBZ	$0, srcPtrLen, xtsSm4DecTailEnc
+	MOVBU (srcPtr), R11
+	MOVBU R11, (R8)
+	MOVBU (R7), R12
+	MOVBU R12, (dstPtr)
+
+xtsSm4DecTailEnc:
+	VLD1 (RSP), [B0.B16]
+	VEOR TW.B16, B0.B16, B0.B16
+	VREV32 B0.B16, B0.B16
+	VMOV B0.S[1], B1.S[0]
+	VMOV B0.S[2], B2.S[0]
+	VMOV B0.S[3], B3.S[0]
+
+	MOVD rkSave, rk
+	EOR R13, R13
+
+tailDecLoop:
+		SM4_ROUND(rk, R19, K0, K1, K2, B0, B1, B2, B3)
+		SM4_ROUND(rk, R19, K0, K1, K2, B1, B2, B3, B0)
+		SM4_ROUND(rk, R19, K0, K1, K2, B2, B3, B0, B1)
+		SM4_ROUND(rk, R19, K0, K1, K2, B3, B0, B1, B2)
+		ADD $1, R13
+		CMP $8, R13
+		BNE tailDecLoop
+
+	VMOV B2.S[0], B3.S[1]
+	VMOV B1.S[0], B3.S[2]
+	VMOV B0.S[0], B3.S[3]
+	VREV32 B3.B16, B3.B16
+
+	VEOR TW.B16, B3.B16, B3.B16
+	VST1 [B3.B16], (R9)
+
+	B xtsSm4DecDone
+
+xtsSm4DecLastBlock:
+	loadOneBlock
+
+	MOVD rkSave, rk
+	EOR R13, R13
+
+decLastBlockLoop:	
+		SM4_ROUND(rk, R19, K0, K1, K2, B0, B1, B2, B3)
+		SM4_ROUND(rk, R19, K0, K1, K2, B1, B2, B3, B0)
+		SM4_ROUND(rk, R19, K0, K1, K2, B2, B3, B0, B1)
+		SM4_ROUND(rk, R19, K0, K1, K2, B3, B0, B1, B2)
+		ADD $1, R13
+		CMP $8, R13
+		BNE decLastBlockLoop
+
+	storeOneBlock
+	mul2GBInline
+
+xtsSm4DecDone:
+	VST1 [TW.B16], (twPtr)
+	RET